Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

RadioAstron orbit determination and evaluation of its results using correlation of space-VLBI observations

RadioAstron orbit determination and evaluation of its results using correlation of space-VLBI... A crucial part of a space mission for very-long baseline interferometery (VLBI), which is the technique capable of providing the highest resolution images in astronomy, is orbit determination of the missions space radio telescope(s). In order to successfully detect interference fringes that result from correlation of the signals recorded by a ground-based and a space-borne radio telescope, the propagation delays experienced in the near-Earth space by radio waves emitted by the source and the relativity e ects on each telescopes clock need to be evaluated, which requires accurate knowledge of position and velocity of the space radio telescope. In this paper we describe our approach to orbit determination (OD) of the RadioAstron spacecraft of the RadioAstron space-VLBI mission. Determining RadioAstrons orbit is complicated due to several factors: strong solar radiation pressure, a highly eccentric orbit, and frequent orbit perturbations caused by the attitude control system. We show that in order to maintain the OD accuracy required for processing space-VLBI observations at cm-wavelengths it is required to take into account the additional data on thruster rings, reaction wheel rotation rates, and attitude of the spacecraft.We also investigate into using the unique orbit data available only for a space-VLBI spacecraft, i.e. the residual delays and delay rates that result from VLBI data processing, as a means to evaluate the achieved OD accuracy. We present the results of the rst experience of OD accuracy evaluation of this kind, using more than 5,000 residual values obtained as a result of space-VLBI observations performed over 7 years of the RadioAstron mission operations. Keywords: RadioAstron, orbit determination, space-VLBI 1. Introduction transmits the science data being collected in real time to a tracking station on Earth using its 1.5-m high-gain dish The RadioAstron spacecraft, equipped with a 10-m antenna via a 15 GHz downlink. The science payload of space radio telescope (SRT), was launched into a highly the SRT includes two hydrogen maser frequency standards elliptical Earth orbit in July 2011 by means of a Zenit-3F (H-masers) with only one of the two allowed to operate rocket and Fregat-SB upper stage. The main purpose of at a time. Shortly after launch one of the two H-masers the spacecraft is to conduct observations of galactic and ex- was identi ed as impaired, while the other fully functional. tragalactic radio sources in conjunction with ground-based The latter H-maser was thus used since the beginning of radio telescopes forming a multi-antenna ground-space ra- the mission operations in 2011 till August 2017, when it ex- dio interferometer with extremely long baselines Karda- hausted its hydrogen supply and was switched o , thereby shev et al. (2013). The spacecraft also made it possible to exceeding its expected lifetime by a factor of two. The test the Einstein Equivalence Principle thanks to its ec- science payload of the SRT and its radio downlinks may centric orbit and a highly stable on-board hydrogen maser be synchronized either to the on-board H-maser signal or frequency standard Litvinov et al. (2018); Nunes et al. to a ground H-maser signal uplinked from one of the Ra- (2019). dioAstron mission's tracking stations. While the on-board The SRT observes at the four frequency bands: P-, L-, H-maser was operational, its signal, being slightly more C- and K. During observations the RadioAstron spacecraft stable than that of the uplink, was used as a reference both for the science payload and the downlink signals, in- Corresponding author at: Keldysh Institute of Applied Math- cluding those analyzed in this paper. ematics, Russian Academy of Sciences, Miusskaya sq. 4, 125047 Moscow, Russia. A C-band radio link is used for telemetry and command Email address: zakhvatkin@kiam1.rssi.ru (M. V. Zakhvatkin) Preprint submitted to Elsevier May 8, 2019 arXiv:1812.01623v4 [astro-ph.IM] 6 May 2019 as well as to obtain range and Doppler tracking data for the initially aroused from the necessity of keeping the phase orbit determination needs. Routine tracking is performed error of a K-band signal over integration interval of 10 every 2{3 days by the two antennas located in Russia: the minutes within a fraction of a radian. However the ac- 6 2 64-m antenna at Bear Lakes (near Moscow) and the 70-m celeration error within 1:5  10 m/s window can be antenna at Ussuriysk (Primorsky Krai). RadioAstron is compensated by the correlator during the processing. also equipped with a retrore ector array, which makes it For an earlier, and so far the only other, realized space- possible to perform satellite laser ranging on it. VLBI mission of VSOP/HALCA the OD requirements were RadioAstron's orbit is highly elliptic, with the geocen- similar (1- errors are given): position | 80 m, velocity | 8 2 tric distance varying from 7,000 to 351,600 km and an 0.43 cm/s, acceleration | 6 10 m/s You et al. (1998). average period of 8.6 day. The orbit of the spacecraft sig- The similarity is easily understandable as VSOP/HALCA ni cantly evolves with time due to the gravitational pull of was designed to observe at comparable wavelengths, 1.35, third bodies. The orbital parameters change periodically 6 and 18 cm, and recorded the data at the same bit rate with a period of about 2 years, which provides for rich op- as RadioAstron (128 Mbps). Despite the fact that scien- portunities of observing various radio sources distributed ti c observations with HALCA were routinely taken only all over the sky. at 1.6 and 5 GHz because of sensitivity problems at 22 VLBI data processing involves the so-called fringe search GHz, its OD requirements were formulated for all three procedure, i.e. the search for a maximum of the cross- bands. Slightly more stringent requirements of VSOP to correlation of the signals recorded at di erent telescopes. position and velocity determination accuracies are due to This procedure requires the delay in arrival times of the a di erent correlator design. wavefront at each pair of telescopes to be accurately known Orbit determination of the RadioAstron spacecraft is for the duration of the observation. The initial guess for complicated by limited tracking support and signi cant the delay is calculated according to a model which utilizes non-gravitational perturbations caused by solar radiation various kinds of data: trajectories of the phase centers of pressure (SRP) and autonomous rings of thrusters of the the participating telescopes in an inertial reference frame, spacecraft attitude control system. A unique feature of o sets and drifts of their clocks, propagation media param- a space-VLBI spacecraft is the possibility of verifying its eters, etc. In order to be able to perform the fringe search OD accuracy by using the so-called residual delays and in a reasonable amount of time, e.g. tens of minutes, us- delay rates obtained as a result of successful detections of ing a modern computing cluster, a priori uncertainty in interference fringes and post-correlation analysis. In this the parameters of the model should be relatively small work we summarize the rst experience of OD accuracy and known (the required level of uncertainties depends on evaluation based on using residual data of space-VLBI ob- many factors, a few examples are given below). servations. Contrary to the line-of-sight tracking obser- In space-VLBI the major uncertainty in the modeled vations this kind of data allows to measure errors of the delay and its derivatives comes from orbit reconstruction spacecraft position and velocity projected on di erent di- errors of the spacecraft(s) carrying the radio telescope(s). rections. This analysis allowed us to quantitatively evalu- For the RadioAstron spacecraft the following requirements ate the accuracy of the two versions of the OD algorithm to the OD accuracy were established prior the launch: used. position error less than 600 m; 2. Orbit and dynamics velocity error less than 2 cm/s; The RadioAstron spacecraft was launched into a highly 8 2 acceleration error less than 10 m/s . elliptic orbit with apogee distance varying between 280 000 and 351 000 km, orbital period between 8.1 and 10.2 days These requirements are determined by several factors, in- and perigee height above 650 km. The orbit evolves sig- cluding the following: the wavelength of the observation, ni cantly with time in both eccentricity and orientation of the data bit rate, the number of participating telescopes, the orbital plane because of luni-solar gravitational pertur- the correlator design, the available computational resources, bations and due to the in uence of Earth's gravitational etc. eld during occasional low perigees (Table 1). An impor- The requirement on the position accuracy formally re- tant aspect of such orbit for space-VLBI is that it provides ects the delay search window with 128 spectral chan- for observing a large sample of galactic and extragalac- nels. The actual window size and the number of chan- tic radio sources at a wide range of projected baselines. nels used by the correlator is signi cantly larger, e.g. 2048 Only three trajectory correction maneuvers have been per- spectral channels for the survey of active galactic nuclei formed so far. Their goal was to prolong the lifetime of (AGN), which allows to account for di erent kinds of er- the orbit by preventing reentry while passing an upcoming rors. The velocity error less than 2 cm/s allows to per- local minimum of the perigee height and also to prevent form a fringe search on the smallest wavelength (1.35 cm) unacceptably long shadowing of the spacecraft. using the integration time of only 1/8 s in the correlator. Except for the rare moments when the spacecraft is Rather stringent requirement on the acceleration accuracy moving near a perigee and the perigee height is near its 2 Table 1: Evolution of the orbital elements of the RadioAstron spacecraft. Epoch 2011/08/21 2012/10/13 2014/01/27 2015/03/23 2016/06/24 2017/07/31 2018/02/17 Perigee height, 10 km 6.213 69.265 1.075 67.666 0.654 75.716 3.699 Apogee height, 10 km 336.526 278.126 338.611 280.593 336.278 306.435 329.026 i, deg 56.354 76.804 9.199 56.692 43.866 68.527 51.309 , deg 331.682 295.970 152.397 107.007 0.930 306.140 297.661 local minimum, the major perturbing accelerations are due the force and torque due to SRP in the spacecraft- xed to gravitational pull of third bodies and non-gravitational frame are accelerations. The latter are the main source of errors d s F = (1  )F (s) +  F (s) when modeling the dynamics of the RadioAstron space- SRP 1 1 1 1 A;B A;B craft. Of the non-gravitational perturbations SRP has the +(1 )F (s) A;B greatest impact on the orbit. s a + F (s) + (1 )F (s); (1) 2 2 SP SP Because of its 10-m on-board antenna, the area-to- d s M = (1  )M (s) +  M (s) mass ratio of the RadioAstron spacecraft is as high as 0.03 SRP 1 1 1 1 A;B A;B kg/m . Depending on the spacecraft orientation with re- +(1 )M (s) A;B spect to the Sun this results in accelerations due to SRP as s a 7 2 + M (s) + (1 )M (s): (2) 2 2 SP SP high as 210 m/s . Accounting for SRP is additionally complicated by the fact that the spacecraft has to change Here, s is the Sun vector, F() and M() on the right-hand its attitude according to the direction to the radio source side of Equations (1) and (2) represent the contributions to be observed, thus the SRP may vary signi cantly over to, respectively, the force and torque from various parts a time span of several hours. of the spacecraft, which we denoted by the following sub- In addition to direct acceleration, SRP produces the scripts: A; B | the antenna and the spacecraft bus, SP | major part of the perturbing torque. Attitude control of solar panels. The superscripts d, s and a denote, respec- the RadioAstron spacecraft, which includes compensation tively, di use re ection, specular re ection and absorption of external torques, is implemented using reaction wheels. of the solar radiation. The spacecraft attitude in the inertial reference frame can The surface of the spacecraft bus and the solar panels be roughly described as piece-wise constant | most of the in the model consists of a number of rectangular elements. time the spacecraft is not rotating, with external torques Provided that all restrictions on the spacecraft attitude compensated by the rotation of the reaction wheels, the are met, these elements do not shadow each other from changes of attitude occur relatively quickly, with their du- the Sun, which facilitates the calculation of the force and ration of order of a few minutes. torque components. Surface of the parabolic antenna of Several times a day the reaction wheels are desatu- the SRT is represented with 4096 at triangular elements. rated. This happens either due to the angular momen- Each element of the antenna surface is considered sunlit if tum accumulated by the reaction wheels reaching its max- other parts of the antenna or rectangular elements of the imum allowed value, so that they cannot parry the external spacecraft bus or solar panels do not shadow the geometric torque any further, or due to the need to perform a signi - center of the element. Thus the surface of the spacecraft cant attitude change which would not be possible using the in the model consists of at elements, for each of which reaction wheels alone in their current state. Desaturating the SRP force due to absorption, specular re ection and the reaction wheels implies stopping them almost com- di use re ection of radiation is calculated using the area of pletely with resulting rotation of the spacecraft prevented the element, its normal vector and unit vector towards the by means of thrusters. Every desaturation produces a net Sun. Summation over all sunlit elements provides F() and Delta V of 3{7 mm/s. Total e ect of the desaturations M() components in the right-hand sides of Equations (1) on the orbit is comparable to the one of direct SRP and and (2). These components depend only on the position therefore must be taken into account. of the Sun, s, therefore, they are tabulated for various Sun In order to properly take the SRP in uence into ac- angles to speed up the integration of equations of motion. count, we have developed an adjustable model, which al- Although this model describes the direct SRP pertur- lows us to calculate both the perturbing acceleration and bation, we believe that adjustable coecients allow it to torque for a given attitude with respect to the Sun Za- also account for the major part of the force due to thermal khvatkin et al. (2014). The SRP model utilizes a simpli ed radiation from the spacecraft surface. model of the spacecraft surface, which consists of the 10-m We do not use the torque estimation provided by Equa- parabolic dish antenna, the solar panels and the spacecraft tion (2) in the model of motion of the spacecraft's center bus (Fig. 1). The coecients of re ectivity and specu- of mass. However, unlike the SRP force the torque, as larity  are associated with sunlit surfaces of the antenna will be shown below, can be measured directly. Since the and spacecraft bus, the coecient of re ectivity | with calculated value of the torque depends on the SRP co- the surface of the solar panels. According to the model, ecients, observations of the torque can be used in the 3 The downlink carrier can be synchronized either to the highly stable signal of the on-board H-maser or the up- linked signal of a ground H-maser. The RadioAstron mis- sion is served by two tracking stations capable of receiving the scienti c data from the satellite and transmitting an uplink reference signal to it: one at Pushchino (Moscow Region, Russia) and another at Green Bank (West Vir- ginia, USA). Additionally, these tracking stations perform Doppler measurements of the downlink signal during each scienti c observation conducted by the satellite. The qual- ity of these measurements is usually much higher than that of the standard radio tracking at the C-band (see Table 3 below). Since the scienti c observations are usually per- Figure 1: The approximation of the RadioAstron spacecraft surface formed several times a day and last for about an hour, used in the SRP model these measurements provide very valuable data for OD. A comparison of a posteriori estimated accuracy of ra- dio tracking observations is shown in Table 3. One-way OD to improve the estimate of the SRP coecients. The Doppler data provided by the Pushchino and Green Bank SRP coecients are assumed constant during the whole stations are much more accurate than the two-way Doppler orbit determination interval, which usually spans for 20{ data obtained by the regular tracking stations of Bear 30 days. Lakes and Ussuriysk. Relatively large two-way Doppler The complete speci cation of the dynamic model we noise at Bear Lakes and Ussuriysk is explained by the per- use to describe the orbital motion of the RadioAstron formance of the C-band telemetry, tracking and control spacecraft is given in Table 2. Between the events of re- system used in the project. However, these stations are action wheel desaturation (momentum dumps) the space- also capable of performing range measurements and their craft moves passively and every desaturation event is as- two-way Doppler data are free from the contribution of sociated with a certain Delta V vector. A priori values the a priori unknown frequency o set of the on-board H- of velocity change vectors are calculated using the atti- maser. tude data available via telemetry and the target values of Radio tracking of RadioAstron is supplemented by op- Delta V reported by the attitude control system for each tical tracking. Routine optical angular observations are thruster ring event. The model for the atmospheric drag conducted by a number of observatories, including the ones is signi cantly simpler than the SRP model, because most of the ISON collaboration, Roscosmos facilities, the MAS- of the time it has close to zero impact on the motion. For TER network and many others. More than 40 di erent the whole duration of the mission the perigee height was telescopes have observed the spacecraft so far and provided below 1500 km (the nominal height of the atmosphere in optical measurements of right ascension and declination of the density model used) only several times. the spacecraft. Laser ranging to RadioAstron is also possible and pro- 3. Tracking and on-board observations vides distance measurements with cm-level accuracy. How- ever, such observations are performed only occasionally for Several types of observational and telemetry data are two reasons. First, the retrore ector array is mounted on used to determine RadioAstron's orbit. Standard radio the spacecraft in such a way that obtaining laser echoes tracking consists of range and Doppler measurements at from it is possible only when the spacecraft is in a speci c the C-band. This tracking is performed in the two-way orientation (or oriented no more than approximately 10 mode with the 64-m antenna at Bear Lakes and the 70- degrees away from it). Because of this, it is almost impos- m antenna at Ussuriysk. Each tracking session is usually sible to perform laser ranging simultaneously with scien- several-hour long, and the observations are separated by ti c observations, with the exception of the gravity exper- 1{3 days. This time interval is much longer than the typ- iment. Second, the spacecraft spends most of the time at ical separation between successive desaturations of the re- distances exceeding 100,000 km, which makes it reachable action wheels (several hours), which makes it dicult to only to the most powerful laser ranging stations, usually achieve the required accuracy of orbit determination using those capable of ranging to the Moon. So far success- this type of data alone. ful laser ranging observations have been performed by the There is another highly valuable type of radio tracking following observatories: Grasse (France), Kavkaz (Russia), data, which can be obtained when the SRT is performing Yarragadee, Mt. Stromlo (both in Australia) and Wettzell a scienti c observation. The RadioAstron spacecraft does (Germany). not store any of the radio astronomy data collected by the A test of applying the Planetary Radio Interferome- SRT and transmits it in real time to Earth using its 1.5-m try and Doppler Experiment (PRIDE), which is an OD high-gain antenna during every experiment. technique based on observing spacecrafts using VLBI, was 4 Table 2: Description of the RadioAstron dynamic model. Perturbation Model Gravity Static EGM-96 up to 75th degree/order Third bodies JPL DE-405 Sun, Moon and planets Earth tides IERS-2003 conv. General relativity IERS-2010 conv. Surface forces Solar radiation 3 parameters ,  and 1 1 2 Earth radiation Applied (constant albedo coe .) GOST R 25645.166-2004 density model, Atmospheric drag cannonball force model with one solve-for parameter Desaturation of reaction wheels Delta V Direction Telemetry data from the on-board star sensors and attitude le Delta V magnitude Telemetry data for the duration of thruster ring and a priori thrust model scopes. This allows us to estimate the perturbing torque Table 3: A posteriori estimates of radio tracking errors based on OD as follows. Given that the spacecraft attitude with respect results from September 2016 to April 2017. Doppler observations at Bear Lakes and Ussuriysk are received over 1 second integration to an inertial frame is constant on the time interval be- interval, at Pushchino and Green Bank | normal points with 1 tween t and t , the change of the angular momentum can 1 2 minute averaging were employed. be represented in a body- xed frame as Range bias Doppler Tracking station (RMS), m noise, mm/s 2 I a ( (t ) (t )) = M(t)dt; (3) Bear Lakes, 64-m 2.9 1.21 i i i 2 i 1 i=1 Ussuriysk, 70-m 13.4 4.78 Pushchino, 22-m N/A 0.19 where I is the moment of inertia of the i-th wheel, a i i Green Bank, 140-ft N/A 0.04 is its axis of rotation, (t) is its angular velocity, and M(t) is the perturbing torque. SRP is usually the only major source of perturbing torque which is responsible for conducted during the in-orbit checkout of the RadioAstron the change of angular momentum of the reaction wheels. mission Duev et al. (2015). The technique showed the po- Therefore, Equation (3) allows one to obtain the observed tential for improving the accuracy of estimating the space- value of the SRP torque in a body- xed frame as long craft state vector. However, PRIDE operations for Ra- as the SRP variations on the time interval of (t ; t ) are 1 2 dioAstron would require massive involvement of ground- negligible based radio telescopes capable of observing at the X- and Ku-bands as well as signi cant data processing resources. M = I a ( (t ) (t )): (4) Such e orts were deemed unnecessary since other \tradi- SRP;o i i i 2 i 1 t t 2 1 i=1 tional" OD techniques proved sucient. Several types of data collected on board and available The corresponding computed value of the SRP torque is via telemetry proved helpful in modeling RadioAstron's given in Equation (2). The computed value depends on the dynamics and determining its orbit. The information on SRP coecients introduced in Section 2, thus the reaction the spacecraft attitude and rings of the attitude control wheel rotation rates contain valuable information, which thrusters is essential for modeling the perturbations due to otherwise may only be obtained from the orbital dynamics SRP and Delta V's that result from momentum dumping of RadioAstron. of the reaction wheels. Finally, also available via telemetry are the rotation rates of the reaction wheels, which implic- 4. Orbit determination technique itly contain information on the dynamics of the spacecraft center of mass. This information can be extracted in the According to the dynamic model described above, Ra- following way. dioAstron's orbital motion is determined by the following Dynamics of the center of mass and attitude dynamics vector of parameters Q = fX(t ); ;  ; ; v ; : : : ; v g, d 0 1 1 2 1 n are related to each other by the action of SRP. The atti- which are, respectively, the initial state vector of the space- tude control system of the spacecraft uses reaction wheels craft, the three SRP coecients and the n vectors of ve- instead of more commonly applied control moment gyro- locity changes due to desaturations of the reaction wheels, 5 0.00002 and where n is the number of desaturation events in the 0.00001 orbit determination time interval. 0.00000 The kinematic parameters that a ect only the com- 0.00001 puted values of observables are the session-wise range bi- 0.0004 ases and the frequency o set of the on-board H-maser, also assumed to be piece-wise constant in a way described in 0.0002 Section 6. The latter a ects the one-way Doppler observ- 0.0000 ables obtained by the Pushchino and Green Bank tracking stations. Standard orbit determination interval includes 0.00002 at least two full orbits of the spacecraft and usually does 0.00000 not exceed 30 days. Orbit determination is performed using the batch least squares estimator. The observations of the torque and a priori Delta V vectors are assumed to be independent Figure 2: Components of the SRP torque on a 48 hr interval in from each other and from the tracking data. Therefore, January 2015: observed (red), computed (blue). the functional to be minimized can be represented as =( ) P( )+ o c o c matrix,  is the error in the estimated magnitude of v, and  is the error in the estimate of v components in X d j j T SRP j j + (M M ) P (M M )+ the plane orthogonal to e. The typical value of magnitude SRP;o SRP j SRP;o SRP j=1 error  used in the OD is 5% of v, the value of the or- X thogonal component,  , corresponds to a 0.25 deg error 0 T 0 + (v v ) P (v v ): (5) i i i i i in direction or 4:36 10 of v. i=1 In reality the functional is not limited to the one in Equation (5). It accounts also for the a priori information The rst term in Equation (5) gives the contribution of the on range biases and the on-board H-maser frequency o set. tracking data with corresponding computed values o c and inverted covariance estimate of as weight matrix P. 5. ASC correlator overview The second term in Equation (5) contains the sum of squares of residual torques, the observed and computed Orbital solutions obtained as a result of the OD pro- values of which were described in the previous section cedure outlined above are used mostly for scienti c appli- (Equations (4) and (2)). cations, primarily for correlation of VLBI data collected The covariance matrices and their inverses, P , can by the mission. In this section we review the operations SRP be estimated using Equation (4), given the estimation er- of one of the correlators capable of processing RadioAs- rors of the rotation rates of the reaction wheels and the ori- tron data, the ASC Correlator. This FX correlator has entations of their axes are known. In practice, however, the been developed by the Astro Space Center of the Lebe- weights have to be adjusted to an accuracy level of about dev Physical Institute (ASC LPI) Likhachev et al. (2017) 2 10 Nm, since the perturbing torque model described speci cally to support the RadioAstron mission. It is used above is relatively simple and does not always match the to process the majority of observations performed by Ra- accuracy of observations. Fig. 2 shows the agreement be- dioAstron (95%) and thus provides the largest data set of tween the modeled torque and the observational data. It is the so-called residual delays and delay rates, which we will clear that for the X and Z components the degree of agree- use in the next section to evaluate the orbit determination ment, in relative terms, is signi cantly lower than that for accuracy. the Y component, but it is the latter that contains the Correlation of space-VLBI data is performed by the dominant part of the torque due to SRP. ASC Correlator in two steps, with the second step being The last term of Equation (5) takes into account the a necessary due to the orbit determination uncertainties. In priori information on Delta V's due to desaturations. For order to decrease the residual delay, delay rate and delay each desaturation event the value of net v is calculated acceleration, the rst step, or pass, is performed in the so- using information on every thruster ring that occurred called \wide" window mode. The window size along the during that event, including its duration, the fuel pressure delay is determined by the number of spectral channels, and the a priori thrust model. Each of the weight matrices, while along the fringe rate it is determined by the integra- P , in (5) is the inverse of the respective covariance matrix i tion time. Both parameters are set before the correlation given by process is started and can be easily adjusted. After a fringe was found, the second correlation pass is performed in a 2 T 2 T C =  (E e e ) +  e e : d m smaller window taking into account the residual delays ob- Here, e is the unit vector of v which has the same di- tained in the rst pass. rection as X-axis of the spacecraft, E is a 3  3 identity 02, 00:00 02, 06:00 02, 12:00 02, 18:00 03, 00:00 03, 06:00 03, 12:00 03, 18:00 04, 00:00 Mx N m M N m My N m RadioAstron's capability to simultaneously observe at observations performed since 2014 and a signi cant por- two di erent wavelengths is of great value for performing tion of earlier observations were correlated using orbits the fringe search. In case an observation was performed produced with this version of the algorithm. At the early at two di erent wavelengths, the residual values obtained stages of the mission, i.e. for experiments conducted be- as a result of the successful fringe search at the longer fore 2014, a previous, less sophisticated version of the OD wavelength can usually be used to signi cantly simplify algorithm was used. the fringe search at the shorter one. A delay model for space-VLBI is naturally more com- 1e 10 plicated than those used for ground-based VLBI. Apart ver. 2 ver. 1 from issues related to the orbit, it has to take into account the fact that the delay rate depends not only on relative ve- locities of the telescopes but also on the frequency o set of the uplink signal, which is used on board as a reference for the on-board scienti c equipment (for the two-way phase- 1 locked loop mode of synchronization), or on the frequency o set of the on-board H-maser (for the one-way mode). The latter, as it turned out, could result in delay rates as high as 30 35 ps/s. A distinctive feature of the ASC Correlator is its de- lay model ORBITA2012. This delay model was developed speci cally for the ASC Correlator and is capable of calcu- Date lating the delay up to O(c ) terms Vlasov et al. (2012), which provides for using smaller correlation window sizes at the rst pass of the correlation process. Moreover, this Figure 3: Residual delay rates obtained by the ASC Correlator using delay model is capable of taking into account the delay two versions of the orbit: generic algorithm, version 1 (red dots); new acceleration. algorithm, version 2 (blue dots). The post-correlation data reduction for all experiments considered in this paper was performed using the PIMA A statistically signi cant correlation detected between package Petrov et al. (2011). This included baseline fringe signals recorded by the RadioAstron space radio telescope tting, bandpass and amplitude calibration, and averag- and ground-based radio telescopes yields a set of three ing the data over time and frequency. The primary goal of residual values, that of delay, delay rate and delay acceler- the post-correlation processing is to allow for studying the ation. In this paper we focus exclusively on the delay rates physical properties of the observed objects. The residual for the following reasons. First, data collected by the space values of group delay, delay rate and delay acceleration radio telescope are not time-tagged on board, i.e. synchro- are a byproduct of this processing. The total delay and its nization of the on-board H-maser time scale with that of derivatives can be further used for studying the observed UTC is possible only indirectly, by means of time-tagging objects, e.g. in imaging experiments. This requires accu- the data stream owing from the satellite by the receiving rate calibration of a number of uncertainties, e.g. the rates tracking station. Second, there is a non-zero probability of the telescope clocks, their trajectories (including that of of inconsistent synchronization of ground radio telescopes the SRT), coordinates of the observed sources, propagation with UTC. And nally, as shown below, there are a num- media parameters, etc. ber of signi cant orbit-unrelated factors, which can con- The residual delays and their derivatives, which we tribute to the residual delay rate also a ecting the residual used in this analysis, were obtained as a result of correla- delay. All this makes it dicult to separate the contribu- tion of the observations of the RadioAstron AGN survey tion to residual delay due to errors in the orbit from those science program and were not used in the OD procedure. due to the involved clocks. On the other hand, possible The sources observed in this program are compact enough, contributions to the residual delay rate are much more so that the contribution to residual values due to non-point predictable. Time synchronization by means of a tracking structure of the sources is negligible. station does not a ect the delay rate, since any change in the delay it may introduce remains constant during every \scan" ( le) of the recorded data. Drifts of ground-based 6. Orbit determination results clocks are usually much smaller than the residual delay rates we observe and, moreover, when an experiment in- The orbit determination algorithm outlined in this pa- volves more than one ground radio telescope, relative drifts per has been implemented and used by the Keldysh In- of the ground-based clocks can be estimated by perform- stitute of Applied Mathematics (KIAM) to provide Ra- ing fringe searches on ground-only baselines. Finally, the dioAstron users with a posteriori orbits required to corre- clock drift of the spacecraft clock can be estimated inde- late VLBI data collected by the mission. The majority of pendently in the process of orbit determination. Residual delay rate, s/s Fig. 3 shows the residual delay rates obtained by the 1e 11 ASC Correlator as a result of processing the observations ARECIBO EVPTRIYA KALYAZIN TIANMA65 performed by the RadioAstron mission up to early 2017. ATCA-104 GBT-VLBA MEDICINA TORUN BADARY HARTRAO MOPRA WSTRB-07 Each data point in Fig. 3 corresponds to a value of the CEDUNA HOBART26 NOTO YEBES40M DSS63 IRBENE PARKES ZELENCHK residual delay rate determined by the correlator as a result EFLSBERG JODRELL2 SVETLOE of correlating two data scans, i.e. units of data typically of 10 min duration, one of which was recorded by the Ra- 0 dioAstron space radio telescope and the other by a ground radio telescope that participated in the experiment. Blue dots denote the residual delay rates obtained with version 4 2 orbits, i.e. those determined with the algorithm outlined in Section 4. Red dots denote the values obtained with verion 1 orbits, i.e. determined with a generic OD algo- rithm and a simpler dynamic model of the spacecraft. The Date earlier version of the algorithm uses a simple cannonball model to estimate the SRP and thus ignores its attitude dependence. Also, the velocity changes due to momentum Figure 4: The residual delay rates obtained using version 2 orbits. dumps are not estimated in the generic algorithm but set Color indicates the ground radio telescope of the baseline. to their a priori values, v . For each data point of Fig. 3 it is assumed that RadioAstron is the rst antenna while accurate to at least 10 : the ground radio telescope is the second one, i.e. the de- d 1 picted residual delay rates are relative to the space-based = k v (t ); (6) 0 1 2 antenna. dt c All the observations used to obtain the residual delay where k is a unit vector pointing in the direction of the rates of Fig. 3 were conducted in the so-called \one-way" source (the aberration factor can be omitted provided our mode, which is characterized by the following two condi- accuracy requirements). tions: a) the on-board scienti c equipment is synchronized The delay understood by the correlator is a di erence to the reference signal of the on-board H-maser; b) the car- of proper time intervals counted by the station clocks from rier of the downlink signal used to transfer the science data the moment of synchronization to the moment of signal re- from the spacecraft to a tracking station is also synchro- ception. Thus the estimate for the TCG delay rate given nized to the reference H-maser signal. by Equation (6) may be insuciently accurate, especially It is clear from Fig. 3 that the variation of the residual in the case when one of the stations is orbiting the Earth. delay rates obtained with version 2 of the OD algorithm is Assuming the clocks were synchronized at t the delay substantially smaller than that of the delay rates obtained should be as follows with the generic algorithm. The RMS of the residual de- Z   Z t 2 t 2 2 1 1 v 1 v 11 10 2 1 lay rates are 2:463 10 for version 2 and 1:097 10 =  + + U (r ) dt + U (r ) 0 2 1 2 2 c 2 c 2 for version 1. It is also clear that the residual delay rates t t s s obtained with version 2 orbits are o set from zero, with @V (R ) +V (R ) V (R ) r dt; 1  1 the mean value equal to 1:663  10 and also reveal a @R long-term linear trend (Fig. 4). The latter is clearly an in- (7) dication of a systematic e ect which cannot be attributed where  is the TCG delay, t and t are, respectively, the 0 1 2 to an OD error since these residual values were obtained moments of signal reception by station 1 and station 2, using a large number of independent orbit solutions and an U (r) is the Newtonian potential of the Earth at position even greater number of observations of radio sources scat- r, V is the sum of the Newtonian potentials of all bodies of tered all over the sky and observed at various projected the Solar System excluding the Earth, R is the barycen- baselines. tric position of the spacecraft, and R is the barycentric Let us consider the impact of the orbit errors on the position of the geocenter. The tidal potential is neglected calculated value of the group delay reported by the cor- for the ground-based station (i.e. station 2), following the relator. Denote the spacecraft state vector in a geocen- recommendations of Petit and Luzum (2010). tric inertial reference frame according to the orbital so- T T T The change of the delay given by Equation (7) due to lution by X(t) = (r (t) ; v (t) ) , the true state vector 1 1 T T T errors in the spacecraft position, r , and velocity, v , 1 1 by X(t) = (r (t); v (t)) and the di erence between the 1 1 T T T can be represented as two by X(t) = X(t) X(t) = (r (t) ; v (t) ) . Pro- 1 1 vided that the requirements to OD accuracy stated in Sec- (r ; v ) = (r ; v )) 1 1 0 1 1 tion 1 are met, the error in the Geocentric Coordinate 1 r  r 1 1 Time (TCG) delay rate Vlasov et al. (2012) due to OD v  v U dt; (8) 1 1 2 2 c r errors can be given by the following equation, which is s 1 Residual delay rate, s/s P P 3 3 1 @ V where the tidal term, which can be approximated by r r , i j 2 i=1 j=1 @R @R i j 1e 11 has been neglected since for RadioAstron's orbit and the expected magnitude of OD errors, the contribution of this term to the delay rate is less than 10 . Therefore the correction to the delay rate can be written as d d v  v r  r 1 1 1 1 =  U : (9) 2 2 dt dt c c r Provided that OD errors do not exceed the values stated in Section 1, the last two terms on the right-hand side of Equation (9) contribute at most 10 . Therefore, if we aim at evaluating the correction to the delay rate with Date an accuracy of 10 , these two terms can be neglected. Adding to this the correction due to the on-board H-maser Figure 5: Evolution of the RadioAstron on-board H-maser fractional clock drift, h (t) and that of the ground-based station, frequency bias. Each value is obtained as part of an orbital solu- h (t), we arrive at the following equation tion. The density of points before and after October 2014 is di erent because of a change of the frequency bias estimation strategy. d 1 ((t t ) (t t ) ) = k v (t ) h (t ) + h (t ); 2 s p 1 s p 1 2 1 1 2 2 dt c (10) i.e. when both the on-board science equipment and the downlink carrier are locked to the on-board H-maser ref- which can be used to investigate the spacecraft velocity erence signal. Therefore, if the on-board H-maser output error, provided it is larger than 10 but within the frequency is biased, i.e. there is a non-zero clock drift, both limit given in Section 1, i.e. less than 10 . The two the science data, which eventually results in the residual additional assumptions needed to justify the use of Equa- delay rates used in our analysis, and the downlink carrier tion (10) for this purpose are: a) the position and velocity frequency measurements, which are used in OD, will be of the ground-based station (#2) are known much better a ected by this same frequency bias. than those of the spacecraft and b) the error of model- In order to properly t one-way Doppler measurements ing the contribution of the propagation media to the delay in the OD process, frequency bias of the tracking station rate is less than 10 . While the rst assumption is un- doubtedly true, the second may be violated in rare cases H-maser and that of the on-board H-maser have to be of extreme weather conditions. taken into account. For each of the two tracking stations we used a priori values of the frequency biases, which we The delay model implemented by the ASC Correlator derived from long-term series of measured di erences be- assumes that the spacecraft and station clocks are ideal, tween the station's local time, synchronized to its H-maser, i.e. h (t)  0 and h (t)  0. Therefore, any instrumental 1 2 and the GPS time. We assume the on-board H-maser frac- e ect a ecting the clock drift, if present, would manifest tional frequency bias, f=f , to be identical to that of the itself in an extra residual delay rate reported by the cor- downlink carrier, and estimate the latter in the OD pro- relator. cess. If we assume that no other systematic e ects are present, The on-board H-maser frequency biases obtained this each data point shown in Fig. 3 corresponds to the right- way are shown in Fig. 5. The bias evolution is almost linear hand side of Equation (10) evaluated at a certain time over more than 4 years of the analyzed data. In late Octo- using a certain orbital solution plus an error of the delay ber 2014 an accident occurred on board after which the on- rate determination inherent to the fringe search procedure. board H-maser started experiencing a power surge every Therefore, if we are able to evaluate the clock drifts of the time one of the telemetry and command (T&C) transpon- space and ground-based station clocks, and apply the cor- der was turned on. Before that accident our approach responding corrections to each data point of Fig. 3, the to estimating the on-board H-maser frequency bias was result will be the spacecraft velocity errors projected onto to treat it as constant over the whole OD interval. Af- the directions to the observed celestial sources. ter the accident we changed our strategy and estimated We assume that the ground-based station clock drift, the bias independently for each interval between the T&C h (t), is always well below the level of 10 and thus link transponder on/o switches, which usually coincided can be neglected, which is justi ed by the results of fringe with radio tracking from Bear Lakes or Ussuriysk stations. searches at ground-only baselines. Under this assumption, This change of strategy explains the di erent density of any clock drift of the ground station, if present, will be attributed to the error in the spacecraft velocity. The drift data points in Fig. 5 before and after October 2014. Using the above assumption that the downlink carrier of the on-board clock, h (t), is estimated as part of the OD fractional frequency bias, f=f , is equal to that of the process in the way outlined below. on-board H-maser, i.e to the on-board clock drift, h , we All the observations used for the present analysis were can substract it from the residual delay rates using Equa- performed, as noted above, in the so-called one-way mode, Fractional frequency bias tion (10). The residual delay rates obtained by applying We are working on developing a full understanding of this correction are shown in Fig. 6 (cf. Fig. 4). With the why the correction of Equation (11) is required. How- correction applied the RMS of the residual delay rates re- ever, its origin is likely due to either an inconsistency be- duces to 1:239  10 and its mean value reduces from tween the models of orbital motion or the delay model used 11 12 1:663 10 to 3:735 10 . by the correlator, or both. This can be concluded from the following reasoning. Orbital solutions are provided in the geocentric coordinate system and the Terrestrial Time 1e 11 (TT) scale, as the correlator delay model expects. In that ARECIBO EVPTRIYA KALYAZIN TIANMA65 case the e ect of the solar gravitational potential on the ATCA-104 GBT-VLBA MEDICINA TORUN BADARY HARTRAO MOPRA USUDA64 delay is described by the last three terms in the second CEDUNA HOBART26 NOTO WSTRB-07 DSS63 IRBENE PARKES YEBES40M integral on the right-hand side of Equation (7), with the EFLSBERG JODRELL2 SVETLOE ZELENCHK potential V replaced by U . For the delay rate it di ers from the correction in Equation (11) by the third term, @U (R )=@R r =c , which is approximately equal to the expression in Equation (11) and makes the tidal e ect of the solar gravitational potential almost negligible in the case of RadioAstron. Therefore, the correction in the form of Equation (11) cannot be due to an unaccounted e ect of the solar gravitational potential but can only arise due to an inconsistency in the computation of delays using the Date provided orbit solutions. The residual delay rates after application of the cor- rection of Equation (11) are shown in Fig. 9. In addition Figure 6: The residual delay rates obtained using version 2 orbits and to overall reduction of delay rate variation, it is now clear corrected for the on-board clock drift. Color indicates the ground radio telescope of the baseline. that part of the data points, i.e. those that correspond to RadioAstron{Torun baselines, have an appreciable bias of According to Equation (10), we expect each value of about 1:9  10 , which can reasonably be attributed the residual delay rate corrected for the spacecraft clock to the clock drift of this ground radio telescope. drift to be equal to the corresponding spacecraft velocity Excluding the Torun data, which constitute 8.2% of the dataset, we arrive at the mean value of the residual delay error projected onto the direction of the observed radio 13 12 rates of just 4:387 10 and its RMS of 9:573 10 . source. Therefore, we expect them to be largely uncorre- The distribution of the data also becomes more symmetric lated and symmetrically distributed with respect to zero. (Fig. 8). However, the values shown in Fig. 6 still reveal signi cant In order to investigate the statistical properties of the systematic e ects, i.e. the average value is 1.12 mm/s in residual delay rates shown in Fig. 9 further, we exclude terms of velocity and there are noticeable variations with the data of experiments conducted in the summer peri- a period of about 1 year. The histogram of the data is ods, i.e. between the 1st of June and the 1st of Septem- shown on the left-hand side of Fig. 8. ber. These experiments constitute 10.4% of data, but their A more detailed analysis of the data shown in Fig. 6 RMS, equal to 2:307 10 , is more than 5 times larger using the Lomb-Scargle periodogram method reveals that, than that of the rest of the data. Degrade orbit deter- in addition to the annual variation, the residual delay rates are modulated with a period of about 8.8 day (Fig. 7), mination accuracy is de nitely one of the main reasons which is close to the average orbital period. why the summer experiments exhibit larger residual delay The periodic variations visible in Fig. 6 and con rmed rates. The number of scienti c observations drops signif- on the left panel of Fig. 7 are present in the original data icantly during summer, mainly due to constrains on the and not due to the applied clock correction h (t). spacecraft attitude with respect to the Sun, which reduce In an attempt to understand the origin of these pe- the visibility of radio sources of interest. The lack of ex- riodicities we tried several hypotheses. We were able to periments reduces the amount of Doppler data from the determine that the periodic patterns in the residual delay tracking stations which, in turn, a ects the orbit accuracy. rates can be noticeably reduced by applying the following After excluding the summer experiments and ltering correction out outliers, which, for the 5 threshold, constitute 1.15% of the original data, we obtain a dataset which contains _ = (U (R) U (R )) : (11) 88.4% of the initial number of residual delay rates. For this subset of the data the mean residual delay rate is 4:428 13 12 Here, U is the Newtonian potential of the Sun, R is the 10 and the RMS is 4:255 10 . This corresponds to barycentric position of the RadioAstron spacecraft during the standard deviation of the projected spacecraft velocity the experiment and R is the barycentric position of the error, k v , of 1.275 mm/s. geocenter. To summarize, we obtained a dataset of the spacecraft Residual delay rate, s/s Figure 7: A Lomb-Scargle periodogram of residual delay rates corrected for the on-board H-maser drift before (left) and after (right) application of the correction of Equation (11). Signi cant spikes are annotated with the periods (in days) corresponding to their frequencies. 1e 11 0.5 0.175 ARECIBO EVPTRIYA KALYAZIN TIANMA65 ATCA-104 GBT-VLBA MEDICINA TORUN BADARY HARTRAO MOPRA USUDA64 CEDUNA HOBART26 NOTO WSTRB-07 0.150 0.4 DSS63 IRBENE PARKES YEBES40M EFLSBERG JODRELL2 SVETLOE ZELENCHK 0.125 0.3 0.100 0.075 0.2 0.050 0.1 0.025 0.000 0.0 4 2 0 2 4 6 8 4 2 0 2 4 Date k· δv mm/s k· δv mm/s 1 1 Figure 8: Histograms of the residual delay rates after application Figure 9: The residual delay rates obtained using version 2 orbits of the correction for the on-board clock drift. Left: the additional corrected for the on-board clock drift and for the correction of Equa- correction of Equation (11) is not applied. Right: the correction of tion (11). Color indicates the ground radio telescope of the baseline. Equation (11) is applied. Experiments with the Torun ground radio telescope are excluded. to the tracking station that was receiving data from the spacecraft. velocity error estimates projected onto the directions of the Since every science observation is accompanied by si- observed sources, k v . We assumed that clock drifts of multaneous Doppler measurements performed by the track- ground-based VLBI stations that provided the ground legs ing station used to receive the science data, the largest of the baselines are negligible, the on-board clock drift is errors are to be expected in the velocity components that equal to the estimated value of the downlink carrier frac- belong to the plane which is normal to the tracking station tional frequency o set, and we applied the correction of direction. However, contrary to this expectation, Fig. 10 Equation (11) to remove systematic patterns in the resid- does not reveal any signi cant dependence of the projected ual delay rates due to suspected inconsistencies between velocity error, k v , on . A possible reason for this is the orbit and correlator delay models. Now, in order to that the corrected residual delay rate, and thus kv , con- obtain information on the spacecraft velocity error itself, tains errors due to inaccurate estimation of the on-board v , and not only on its projection, kv , we represent the 1 1 clock drift, h (t), and the assumption on the ground sta- residual delay rates as a function of the angle, , between tion clocks that h (t)  0, both of which do not depend the direction to the observed source, k, and the direction Residual delay rate, s/s on the source direction, k, and may provide signi cant vided by the discontinuity analysis and those obtained us- contributions to the residual delay rates. ing the residual delay rates are similar. However, since the former contain, as noted above, additional prediction errors, we expected them to be higher. This may be an indication of the fact that even after the corrections to the residual delay rates have been applied, clock-related errors still give a signi cant contribution to these data, resulting in overestimated values of velocity errors. 7. Discussion We obtained two independent estimates of the veloc- 2 ity errors, the one using the comparison of adjacent inde- pendent orbits and the other via the residual delay rate analysis. The obtained results agree very well. However, in order to obtain the velocity error estimates using the residual delay rate data, we had to apply two non-trivial 0 25 50 75 100 125 150 175 corrections to the data. Angle, α degrees The rst correction is to take into account an instru- mental e ect of the spacecraft clock drift. The signi cance Figure 10: Dependence of the residual delay rates on the angle be- of the on-board H-maser clock drift is evident from the tween the direction to observed source and the direction from the combined analysis of two-way range and Doppler track- spacecraft to the tracking station which was used to receive the data of that observation ing data together with the one-way Doppler measurements of the downlink carrier synchronized to the on-board H- We can estimate the velocity error along the \worst" maser. Since the ASC Correlator assumes the clocks to directions using the dependence of the projected veloc- be ideal, i.e. any non-zero clock drift goes directly into ity error, k  v , on . For the experiments, in which the residual delay rate, the necessity to correct for the 75   105 , the RMS of k  v is equal to 1.359 on-board clock drift is well justi ed. This correction, how- mm/s. Assuming the spacecraft velocity is determined ever, introduces additional errors into the residual delay not worse along other directions the error ellipsoid of v rate data due to the way it is estimated in the process of is con ned by the one corresponding to the covariance ma- OD. 2 2 2 2 2 2 trix diag(1:85 mm =s ; 1:85 mm =s ; 1:85 mm =s ). The necessity of the second correction, that of Equa- In order to obtain an independent estimate of the ve- tion (11), is also clear as it signi cantly improves the resid- locity error, we compared 55 independent adjacent orbit ual delay rates by cleaning them from arti cial periodic solutions from 2014 to 2018 for discontinuity in velocity. If patterns. However, its nature is currently not as well un- the rst solution was obtained for, e.g., the time interval derstood. We suspect the necessity of this correction is of (t ; t ) and the second one for (t ; t ), the spacecraft 1 2 2 3 due to an inconsistency between the models of orbital mo- state vector is compared at t . The time of comparison, tion or the delay model used by the correlator, or both. t , is thus either the beginning or the end of an OD inter- Investigation of the physical cause behind this corrections val. Unlike the case of the residual delay rates, these times is beyond the scope of the present work and would require do not correspond to any tracking activity and, therefore, a deeper insight in the experimental data accumulated by the state vectors evaluated at these moments contain ad- the mission. ditional prediction errors. The residual delay rate data provide a unique means The 3D RMS of the 54 velocity di erences computed of estimating velocity errors of a space-VLBI spacecraft. this way is 4.38 mm/s. Just like the residual delay rates, If the mission performs observations of radio sources dis- these velocity di erences exhibit signi cant increase in mag- tributed all over the sky, the method e ectively provides nitude in the summer months due to the decrease of the estimates of all components of the spacecraft velocity er- amount of tracking data. If we exclude the summer data, ror, as shown in Fig. 10. we arrive at a dataset of 42 points and 3D RMS of the The OD accuracy evaluation of the only previous space- velocity di erences of 3.52 mm/s. VLBI mission of VSOP/HALCA, given by You et al. (1998), However, we should take into account that such dif- did not include an analysis of the residual delay rate data. ferences contain errors of two independent orbit solutions, However, the overall OD accuracy of the HALCA space- therefore, their variance is as sum of variances of two orbit craft, obtained with traditional methods, was better than solutions. Thus, we can conclude that the 3D RMS of the that of RadioAstron. This can be easily understood as velocity error of a single solution is 3.10 mm/s for the full the HALCA spacecraft was tracked by a larger number of dataset and 2.53 mm/s for its summer-free subset. tracking stations, each possibly with better instrumental The RMS values of the spacecraft velocity error pro- stability than those involved in RadioAstron. k· δv , mm/s 1 8. Conclusion Bank Observatory is a facility of the National Science Foundation operated under cooperative agreement by As- We outlined our approach to orbit determination of sociated Universities, Inc. The Arecibo Observatory is op- the RadioAstron spacecraft. The method includes an in- erated by SRI International under a cooperative agreement house developed SRP model and an algorithm to take into with the National Science Foundation (AST-1100968), and account the accumulated angular momentum of reaction in alliance with Ana G. Mendez-Universidad Metropoli- wheels to improve our knowledge of the dynamics of the tana, and the Universities Space Research Association. spacecraft center of mass. We tested the performance of The Australia Telescope Compact Array (Parkes radio tele- this OD method using the unique \tracking data" available scope / Mopra radio telescope / Long Baseline Array) is only for a space-VLBI spacecraft, i.e. the residual delay part of the Australia Telescope National Facility which rates, which in our case were obtained by the ASC Corre- is funded by the Commonwealth of Australia for oper- lator of the RadioAstron mission as a result of processing ation as a National Facility managed by CSIRO. This the data of VLBI observations of celestial radio sources work is based in part on observations carried out using performed by the RadioAstron spacecraft together with the 32-meter radio telescope operated by Torun Centre ground radio telescopes. This analysis allowed us to con- for Astronomy of Nicolaus Copernicus University in Torun clude that the OD method we developed provides up to (Poland) and supported by the Polish Ministry of Science 11 times more accurate orbital solutions in terms of veloc- and Higher Education SpUB grant. Results of optical ity and residual delay rates as compared to a generic OD positioning measurements of the Spektr-R spacecraft by algorithm. the global MASTER Robotic Net Lipunov et al. (2010), Using the residual delay rate data we have obtained ISON collaboration, and Kourovka observatory were used an estimate of the standard deviation for every compo- for spacecraft orbit determination in addition to mission nent of the spacecraft velocity error. The residual delay facilities. rates obtained in non-summer observations, which consti- tute 88.4% of all data, exhibit a standard deviation of References the spacecraft velocity error of 1.4 mm/s or less for each velocity component. This result is consistent with an in- Duev, D.A., Zakhvatkin, M.V., Stepanyants, V.A., et al., 2015. Ra- dependent a posteriori estimate of the spacecraft velocity dioAstron as a target and as an instrument: Enhancing the Space VLBI mission's scienti c output. Astronomy & Astrophysics 573, error obtained using the analysis of velocity di erences A99. doi:10.1051/0004-6361/201424940, arXiv:1411.4576. computed at the boundaries of adjacent orbital solutions. Kardashev, N.S., Khartov, V.V., Abramov, V.V., et al., 2013. \RadioAstron"-A telescope with a size of 300 000 km: Main pa- rameters and rst observational results. Astronomy Reports 57, 9. Acknowledgements 153{194. doi:10.1134/S1063772913030025. Likhachev, S.F., Kostenko, V.I., Girin, I.A., et al., 2017. Soft- The RadioAstron project is led by the Astro Space ware Correlator for Radioastron Mission. Journal of Astro- Center of the Lebedev Physical Institute of the Russian nomical Instrumentation 6, 17. doi:10.1142/S2251171717500040, arXiv:1706.06320. Academy of Sciences and the Lavochkin Scienti c and Pro- Lipunov, V., Kornilov, V., Gorbovskoy, E., et al., 2010. Master duction Association under a contract with the Russian Robotic Net. Advances in Astronomy 2010, 349171. doi:10.1155/ Federal Space Agency, in collaboration with partner orga- 2010/349171, arXiv:0907.0827. Litvinov, D., Rudenko, V., Alakoz, A., et al., 2018. Probing the nizations in Russia and other countries including Keldysh gravitational redshift with an Earth-orbiting satellite. Physics Institute of Applied Mathematics of the Russian Academy Letters A 382, 2192 { 2198. doi:10.1016/j.physleta.2017.09. of Sciences. Partly based on observations performed with 014. Special Issue in memory of Professor V.B. Braginsky. radio telescopes of IAA RAS (Federal State Budget Scien- Nunes, N., Bartel, N., Bietenholz, M., et al., 2019. The gravita- tional redshift monitored with RadioAstron from near Earth up ti c Organization Institue of Applied Astronomy of Rus- to 350,000 km. Advances in Space Research doi:10.1016/j.asr. sian Academy of Sciences). Partly based on the Evpato- 2019.03.012, arXiv:1904.01060. ria RT-70 radio telescope (Ukraine) observations carried Petit, G., Luzum, B., 2010. IERS Conventions (2010). Technical Re- out by the Institute of Radio Astronomy of the National port. Verlag des Bundesamts fr Kartographie und Geodsie. IERS Technical Note ; 36. Academy of Sciences of Ukraine under a contract with Petrov, L., Kovalev, Y.Y., Fomalont, E.B., et al., 2011. The Very the State Space Agency of Ukraine and by the National Long Baseline Array Galactic Plane Survey { VGaPS. The As- Space Facilities Control and Test Center with technical tronomical Journal 142, 35. doi:10.1088/0004-6256/142/2/35, support by Astro Space Center of Lebedev Physical In- arXiv:1101.1460. Vlasov, I.Y., Zharov, V.E., Sazhin, M.V., 2012. Signal delay in the stitute, Russian Academy of Sciences. Partly based on RadioAstron ground-space interferometer. Astronomy Reports 56, observations with the 100-m telescope of the MPIfR (Max- 984{987. doi:10.1134/S1063772912120086. Planck-Institute for Radio Astronomy) at E elsberg. Partly You, T.H., Ellis, J., Mottinger, N., 1998. Navigation of the space based on observations with the Medicina (Noto) telescope VLBI mission-HALCA, in: Stengle, T. (Ed.), AAS/GSFC 13th International Symposium on Space Flight Dynamics, pp. 841{856. operated by INAF - Istituto di Radioastronomia. The Na- Zakhvatkin, M.V., Ponomarev, Y.N., Stepanyants, V.A., et al., 2014. tional Radio Astronomy Observatory is a facility of the Navigation support for the RadioAstron mission. Cosmic Research National Science Foundation operated under cooperative 52, 342{352. doi:10.1134/S0010952514050128. agreement by Associated Universities, Inc. The Green http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Astrophysics arXiv (Cornell University)

Loading next page...
 
/lp/arxiv-cornell-university/radioastron-orbit-determination-and-evaluation-of-its-results-using-2KrhDVNB5c
ISSN
0273-1177
eISSN
ARCH-3330
DOI
10.1016/j.asr.2019.05.007
Publisher site
See Article on Publisher Site

Abstract

A crucial part of a space mission for very-long baseline interferometery (VLBI), which is the technique capable of providing the highest resolution images in astronomy, is orbit determination of the missions space radio telescope(s). In order to successfully detect interference fringes that result from correlation of the signals recorded by a ground-based and a space-borne radio telescope, the propagation delays experienced in the near-Earth space by radio waves emitted by the source and the relativity e ects on each telescopes clock need to be evaluated, which requires accurate knowledge of position and velocity of the space radio telescope. In this paper we describe our approach to orbit determination (OD) of the RadioAstron spacecraft of the RadioAstron space-VLBI mission. Determining RadioAstrons orbit is complicated due to several factors: strong solar radiation pressure, a highly eccentric orbit, and frequent orbit perturbations caused by the attitude control system. We show that in order to maintain the OD accuracy required for processing space-VLBI observations at cm-wavelengths it is required to take into account the additional data on thruster rings, reaction wheel rotation rates, and attitude of the spacecraft.We also investigate into using the unique orbit data available only for a space-VLBI spacecraft, i.e. the residual delays and delay rates that result from VLBI data processing, as a means to evaluate the achieved OD accuracy. We present the results of the rst experience of OD accuracy evaluation of this kind, using more than 5,000 residual values obtained as a result of space-VLBI observations performed over 7 years of the RadioAstron mission operations. Keywords: RadioAstron, orbit determination, space-VLBI 1. Introduction transmits the science data being collected in real time to a tracking station on Earth using its 1.5-m high-gain dish The RadioAstron spacecraft, equipped with a 10-m antenna via a 15 GHz downlink. The science payload of space radio telescope (SRT), was launched into a highly the SRT includes two hydrogen maser frequency standards elliptical Earth orbit in July 2011 by means of a Zenit-3F (H-masers) with only one of the two allowed to operate rocket and Fregat-SB upper stage. The main purpose of at a time. Shortly after launch one of the two H-masers the spacecraft is to conduct observations of galactic and ex- was identi ed as impaired, while the other fully functional. tragalactic radio sources in conjunction with ground-based The latter H-maser was thus used since the beginning of radio telescopes forming a multi-antenna ground-space ra- the mission operations in 2011 till August 2017, when it ex- dio interferometer with extremely long baselines Karda- hausted its hydrogen supply and was switched o , thereby shev et al. (2013). The spacecraft also made it possible to exceeding its expected lifetime by a factor of two. The test the Einstein Equivalence Principle thanks to its ec- science payload of the SRT and its radio downlinks may centric orbit and a highly stable on-board hydrogen maser be synchronized either to the on-board H-maser signal or frequency standard Litvinov et al. (2018); Nunes et al. to a ground H-maser signal uplinked from one of the Ra- (2019). dioAstron mission's tracking stations. While the on-board The SRT observes at the four frequency bands: P-, L-, H-maser was operational, its signal, being slightly more C- and K. During observations the RadioAstron spacecraft stable than that of the uplink, was used as a reference both for the science payload and the downlink signals, in- Corresponding author at: Keldysh Institute of Applied Math- cluding those analyzed in this paper. ematics, Russian Academy of Sciences, Miusskaya sq. 4, 125047 Moscow, Russia. A C-band radio link is used for telemetry and command Email address: zakhvatkin@kiam1.rssi.ru (M. V. Zakhvatkin) Preprint submitted to Elsevier May 8, 2019 arXiv:1812.01623v4 [astro-ph.IM] 6 May 2019 as well as to obtain range and Doppler tracking data for the initially aroused from the necessity of keeping the phase orbit determination needs. Routine tracking is performed error of a K-band signal over integration interval of 10 every 2{3 days by the two antennas located in Russia: the minutes within a fraction of a radian. However the ac- 6 2 64-m antenna at Bear Lakes (near Moscow) and the 70-m celeration error within 1:5  10 m/s window can be antenna at Ussuriysk (Primorsky Krai). RadioAstron is compensated by the correlator during the processing. also equipped with a retrore ector array, which makes it For an earlier, and so far the only other, realized space- possible to perform satellite laser ranging on it. VLBI mission of VSOP/HALCA the OD requirements were RadioAstron's orbit is highly elliptic, with the geocen- similar (1- errors are given): position | 80 m, velocity | 8 2 tric distance varying from 7,000 to 351,600 km and an 0.43 cm/s, acceleration | 6 10 m/s You et al. (1998). average period of 8.6 day. The orbit of the spacecraft sig- The similarity is easily understandable as VSOP/HALCA ni cantly evolves with time due to the gravitational pull of was designed to observe at comparable wavelengths, 1.35, third bodies. The orbital parameters change periodically 6 and 18 cm, and recorded the data at the same bit rate with a period of about 2 years, which provides for rich op- as RadioAstron (128 Mbps). Despite the fact that scien- portunities of observing various radio sources distributed ti c observations with HALCA were routinely taken only all over the sky. at 1.6 and 5 GHz because of sensitivity problems at 22 VLBI data processing involves the so-called fringe search GHz, its OD requirements were formulated for all three procedure, i.e. the search for a maximum of the cross- bands. Slightly more stringent requirements of VSOP to correlation of the signals recorded at di erent telescopes. position and velocity determination accuracies are due to This procedure requires the delay in arrival times of the a di erent correlator design. wavefront at each pair of telescopes to be accurately known Orbit determination of the RadioAstron spacecraft is for the duration of the observation. The initial guess for complicated by limited tracking support and signi cant the delay is calculated according to a model which utilizes non-gravitational perturbations caused by solar radiation various kinds of data: trajectories of the phase centers of pressure (SRP) and autonomous rings of thrusters of the the participating telescopes in an inertial reference frame, spacecraft attitude control system. A unique feature of o sets and drifts of their clocks, propagation media param- a space-VLBI spacecraft is the possibility of verifying its eters, etc. In order to be able to perform the fringe search OD accuracy by using the so-called residual delays and in a reasonable amount of time, e.g. tens of minutes, us- delay rates obtained as a result of successful detections of ing a modern computing cluster, a priori uncertainty in interference fringes and post-correlation analysis. In this the parameters of the model should be relatively small work we summarize the rst experience of OD accuracy and known (the required level of uncertainties depends on evaluation based on using residual data of space-VLBI ob- many factors, a few examples are given below). servations. Contrary to the line-of-sight tracking obser- In space-VLBI the major uncertainty in the modeled vations this kind of data allows to measure errors of the delay and its derivatives comes from orbit reconstruction spacecraft position and velocity projected on di erent di- errors of the spacecraft(s) carrying the radio telescope(s). rections. This analysis allowed us to quantitatively evalu- For the RadioAstron spacecraft the following requirements ate the accuracy of the two versions of the OD algorithm to the OD accuracy were established prior the launch: used. position error less than 600 m; 2. Orbit and dynamics velocity error less than 2 cm/s; The RadioAstron spacecraft was launched into a highly 8 2 acceleration error less than 10 m/s . elliptic orbit with apogee distance varying between 280 000 and 351 000 km, orbital period between 8.1 and 10.2 days These requirements are determined by several factors, in- and perigee height above 650 km. The orbit evolves sig- cluding the following: the wavelength of the observation, ni cantly with time in both eccentricity and orientation of the data bit rate, the number of participating telescopes, the orbital plane because of luni-solar gravitational pertur- the correlator design, the available computational resources, bations and due to the in uence of Earth's gravitational etc. eld during occasional low perigees (Table 1). An impor- The requirement on the position accuracy formally re- tant aspect of such orbit for space-VLBI is that it provides ects the delay search window with 128 spectral chan- for observing a large sample of galactic and extragalac- nels. The actual window size and the number of chan- tic radio sources at a wide range of projected baselines. nels used by the correlator is signi cantly larger, e.g. 2048 Only three trajectory correction maneuvers have been per- spectral channels for the survey of active galactic nuclei formed so far. Their goal was to prolong the lifetime of (AGN), which allows to account for di erent kinds of er- the orbit by preventing reentry while passing an upcoming rors. The velocity error less than 2 cm/s allows to per- local minimum of the perigee height and also to prevent form a fringe search on the smallest wavelength (1.35 cm) unacceptably long shadowing of the spacecraft. using the integration time of only 1/8 s in the correlator. Except for the rare moments when the spacecraft is Rather stringent requirement on the acceleration accuracy moving near a perigee and the perigee height is near its 2 Table 1: Evolution of the orbital elements of the RadioAstron spacecraft. Epoch 2011/08/21 2012/10/13 2014/01/27 2015/03/23 2016/06/24 2017/07/31 2018/02/17 Perigee height, 10 km 6.213 69.265 1.075 67.666 0.654 75.716 3.699 Apogee height, 10 km 336.526 278.126 338.611 280.593 336.278 306.435 329.026 i, deg 56.354 76.804 9.199 56.692 43.866 68.527 51.309 , deg 331.682 295.970 152.397 107.007 0.930 306.140 297.661 local minimum, the major perturbing accelerations are due the force and torque due to SRP in the spacecraft- xed to gravitational pull of third bodies and non-gravitational frame are accelerations. The latter are the main source of errors d s F = (1  )F (s) +  F (s) when modeling the dynamics of the RadioAstron space- SRP 1 1 1 1 A;B A;B craft. Of the non-gravitational perturbations SRP has the +(1 )F (s) A;B greatest impact on the orbit. s a + F (s) + (1 )F (s); (1) 2 2 SP SP Because of its 10-m on-board antenna, the area-to- d s M = (1  )M (s) +  M (s) mass ratio of the RadioAstron spacecraft is as high as 0.03 SRP 1 1 1 1 A;B A;B kg/m . Depending on the spacecraft orientation with re- +(1 )M (s) A;B spect to the Sun this results in accelerations due to SRP as s a 7 2 + M (s) + (1 )M (s): (2) 2 2 SP SP high as 210 m/s . Accounting for SRP is additionally complicated by the fact that the spacecraft has to change Here, s is the Sun vector, F() and M() on the right-hand its attitude according to the direction to the radio source side of Equations (1) and (2) represent the contributions to be observed, thus the SRP may vary signi cantly over to, respectively, the force and torque from various parts a time span of several hours. of the spacecraft, which we denoted by the following sub- In addition to direct acceleration, SRP produces the scripts: A; B | the antenna and the spacecraft bus, SP | major part of the perturbing torque. Attitude control of solar panels. The superscripts d, s and a denote, respec- the RadioAstron spacecraft, which includes compensation tively, di use re ection, specular re ection and absorption of external torques, is implemented using reaction wheels. of the solar radiation. The spacecraft attitude in the inertial reference frame can The surface of the spacecraft bus and the solar panels be roughly described as piece-wise constant | most of the in the model consists of a number of rectangular elements. time the spacecraft is not rotating, with external torques Provided that all restrictions on the spacecraft attitude compensated by the rotation of the reaction wheels, the are met, these elements do not shadow each other from changes of attitude occur relatively quickly, with their du- the Sun, which facilitates the calculation of the force and ration of order of a few minutes. torque components. Surface of the parabolic antenna of Several times a day the reaction wheels are desatu- the SRT is represented with 4096 at triangular elements. rated. This happens either due to the angular momen- Each element of the antenna surface is considered sunlit if tum accumulated by the reaction wheels reaching its max- other parts of the antenna or rectangular elements of the imum allowed value, so that they cannot parry the external spacecraft bus or solar panels do not shadow the geometric torque any further, or due to the need to perform a signi - center of the element. Thus the surface of the spacecraft cant attitude change which would not be possible using the in the model consists of at elements, for each of which reaction wheels alone in their current state. Desaturating the SRP force due to absorption, specular re ection and the reaction wheels implies stopping them almost com- di use re ection of radiation is calculated using the area of pletely with resulting rotation of the spacecraft prevented the element, its normal vector and unit vector towards the by means of thrusters. Every desaturation produces a net Sun. Summation over all sunlit elements provides F() and Delta V of 3{7 mm/s. Total e ect of the desaturations M() components in the right-hand sides of Equations (1) on the orbit is comparable to the one of direct SRP and and (2). These components depend only on the position therefore must be taken into account. of the Sun, s, therefore, they are tabulated for various Sun In order to properly take the SRP in uence into ac- angles to speed up the integration of equations of motion. count, we have developed an adjustable model, which al- Although this model describes the direct SRP pertur- lows us to calculate both the perturbing acceleration and bation, we believe that adjustable coecients allow it to torque for a given attitude with respect to the Sun Za- also account for the major part of the force due to thermal khvatkin et al. (2014). The SRP model utilizes a simpli ed radiation from the spacecraft surface. model of the spacecraft surface, which consists of the 10-m We do not use the torque estimation provided by Equa- parabolic dish antenna, the solar panels and the spacecraft tion (2) in the model of motion of the spacecraft's center bus (Fig. 1). The coecients of re ectivity and specu- of mass. However, unlike the SRP force the torque, as larity  are associated with sunlit surfaces of the antenna will be shown below, can be measured directly. Since the and spacecraft bus, the coecient of re ectivity | with calculated value of the torque depends on the SRP co- the surface of the solar panels. According to the model, ecients, observations of the torque can be used in the 3 The downlink carrier can be synchronized either to the highly stable signal of the on-board H-maser or the up- linked signal of a ground H-maser. The RadioAstron mis- sion is served by two tracking stations capable of receiving the scienti c data from the satellite and transmitting an uplink reference signal to it: one at Pushchino (Moscow Region, Russia) and another at Green Bank (West Vir- ginia, USA). Additionally, these tracking stations perform Doppler measurements of the downlink signal during each scienti c observation conducted by the satellite. The qual- ity of these measurements is usually much higher than that of the standard radio tracking at the C-band (see Table 3 below). Since the scienti c observations are usually per- Figure 1: The approximation of the RadioAstron spacecraft surface formed several times a day and last for about an hour, used in the SRP model these measurements provide very valuable data for OD. A comparison of a posteriori estimated accuracy of ra- dio tracking observations is shown in Table 3. One-way OD to improve the estimate of the SRP coecients. The Doppler data provided by the Pushchino and Green Bank SRP coecients are assumed constant during the whole stations are much more accurate than the two-way Doppler orbit determination interval, which usually spans for 20{ data obtained by the regular tracking stations of Bear 30 days. Lakes and Ussuriysk. Relatively large two-way Doppler The complete speci cation of the dynamic model we noise at Bear Lakes and Ussuriysk is explained by the per- use to describe the orbital motion of the RadioAstron formance of the C-band telemetry, tracking and control spacecraft is given in Table 2. Between the events of re- system used in the project. However, these stations are action wheel desaturation (momentum dumps) the space- also capable of performing range measurements and their craft moves passively and every desaturation event is as- two-way Doppler data are free from the contribution of sociated with a certain Delta V vector. A priori values the a priori unknown frequency o set of the on-board H- of velocity change vectors are calculated using the atti- maser. tude data available via telemetry and the target values of Radio tracking of RadioAstron is supplemented by op- Delta V reported by the attitude control system for each tical tracking. Routine optical angular observations are thruster ring event. The model for the atmospheric drag conducted by a number of observatories, including the ones is signi cantly simpler than the SRP model, because most of the ISON collaboration, Roscosmos facilities, the MAS- of the time it has close to zero impact on the motion. For TER network and many others. More than 40 di erent the whole duration of the mission the perigee height was telescopes have observed the spacecraft so far and provided below 1500 km (the nominal height of the atmosphere in optical measurements of right ascension and declination of the density model used) only several times. the spacecraft. Laser ranging to RadioAstron is also possible and pro- 3. Tracking and on-board observations vides distance measurements with cm-level accuracy. How- ever, such observations are performed only occasionally for Several types of observational and telemetry data are two reasons. First, the retrore ector array is mounted on used to determine RadioAstron's orbit. Standard radio the spacecraft in such a way that obtaining laser echoes tracking consists of range and Doppler measurements at from it is possible only when the spacecraft is in a speci c the C-band. This tracking is performed in the two-way orientation (or oriented no more than approximately 10 mode with the 64-m antenna at Bear Lakes and the 70- degrees away from it). Because of this, it is almost impos- m antenna at Ussuriysk. Each tracking session is usually sible to perform laser ranging simultaneously with scien- several-hour long, and the observations are separated by ti c observations, with the exception of the gravity exper- 1{3 days. This time interval is much longer than the typ- iment. Second, the spacecraft spends most of the time at ical separation between successive desaturations of the re- distances exceeding 100,000 km, which makes it reachable action wheels (several hours), which makes it dicult to only to the most powerful laser ranging stations, usually achieve the required accuracy of orbit determination using those capable of ranging to the Moon. So far success- this type of data alone. ful laser ranging observations have been performed by the There is another highly valuable type of radio tracking following observatories: Grasse (France), Kavkaz (Russia), data, which can be obtained when the SRT is performing Yarragadee, Mt. Stromlo (both in Australia) and Wettzell a scienti c observation. The RadioAstron spacecraft does (Germany). not store any of the radio astronomy data collected by the A test of applying the Planetary Radio Interferome- SRT and transmits it in real time to Earth using its 1.5-m try and Doppler Experiment (PRIDE), which is an OD high-gain antenna during every experiment. technique based on observing spacecrafts using VLBI, was 4 Table 2: Description of the RadioAstron dynamic model. Perturbation Model Gravity Static EGM-96 up to 75th degree/order Third bodies JPL DE-405 Sun, Moon and planets Earth tides IERS-2003 conv. General relativity IERS-2010 conv. Surface forces Solar radiation 3 parameters ,  and 1 1 2 Earth radiation Applied (constant albedo coe .) GOST R 25645.166-2004 density model, Atmospheric drag cannonball force model with one solve-for parameter Desaturation of reaction wheels Delta V Direction Telemetry data from the on-board star sensors and attitude le Delta V magnitude Telemetry data for the duration of thruster ring and a priori thrust model scopes. This allows us to estimate the perturbing torque Table 3: A posteriori estimates of radio tracking errors based on OD as follows. Given that the spacecraft attitude with respect results from September 2016 to April 2017. Doppler observations at Bear Lakes and Ussuriysk are received over 1 second integration to an inertial frame is constant on the time interval be- interval, at Pushchino and Green Bank | normal points with 1 tween t and t , the change of the angular momentum can 1 2 minute averaging were employed. be represented in a body- xed frame as Range bias Doppler Tracking station (RMS), m noise, mm/s 2 I a ( (t ) (t )) = M(t)dt; (3) Bear Lakes, 64-m 2.9 1.21 i i i 2 i 1 i=1 Ussuriysk, 70-m 13.4 4.78 Pushchino, 22-m N/A 0.19 where I is the moment of inertia of the i-th wheel, a i i Green Bank, 140-ft N/A 0.04 is its axis of rotation, (t) is its angular velocity, and M(t) is the perturbing torque. SRP is usually the only major source of perturbing torque which is responsible for conducted during the in-orbit checkout of the RadioAstron the change of angular momentum of the reaction wheels. mission Duev et al. (2015). The technique showed the po- Therefore, Equation (3) allows one to obtain the observed tential for improving the accuracy of estimating the space- value of the SRP torque in a body- xed frame as long craft state vector. However, PRIDE operations for Ra- as the SRP variations on the time interval of (t ; t ) are 1 2 dioAstron would require massive involvement of ground- negligible based radio telescopes capable of observing at the X- and Ku-bands as well as signi cant data processing resources. M = I a ( (t ) (t )): (4) Such e orts were deemed unnecessary since other \tradi- SRP;o i i i 2 i 1 t t 2 1 i=1 tional" OD techniques proved sucient. Several types of data collected on board and available The corresponding computed value of the SRP torque is via telemetry proved helpful in modeling RadioAstron's given in Equation (2). The computed value depends on the dynamics and determining its orbit. The information on SRP coecients introduced in Section 2, thus the reaction the spacecraft attitude and rings of the attitude control wheel rotation rates contain valuable information, which thrusters is essential for modeling the perturbations due to otherwise may only be obtained from the orbital dynamics SRP and Delta V's that result from momentum dumping of RadioAstron. of the reaction wheels. Finally, also available via telemetry are the rotation rates of the reaction wheels, which implic- 4. Orbit determination technique itly contain information on the dynamics of the spacecraft center of mass. This information can be extracted in the According to the dynamic model described above, Ra- following way. dioAstron's orbital motion is determined by the following Dynamics of the center of mass and attitude dynamics vector of parameters Q = fX(t ); ;  ; ; v ; : : : ; v g, d 0 1 1 2 1 n are related to each other by the action of SRP. The atti- which are, respectively, the initial state vector of the space- tude control system of the spacecraft uses reaction wheels craft, the three SRP coecients and the n vectors of ve- instead of more commonly applied control moment gyro- locity changes due to desaturations of the reaction wheels, 5 0.00002 and where n is the number of desaturation events in the 0.00001 orbit determination time interval. 0.00000 The kinematic parameters that a ect only the com- 0.00001 puted values of observables are the session-wise range bi- 0.0004 ases and the frequency o set of the on-board H-maser, also assumed to be piece-wise constant in a way described in 0.0002 Section 6. The latter a ects the one-way Doppler observ- 0.0000 ables obtained by the Pushchino and Green Bank tracking stations. Standard orbit determination interval includes 0.00002 at least two full orbits of the spacecraft and usually does 0.00000 not exceed 30 days. Orbit determination is performed using the batch least squares estimator. The observations of the torque and a priori Delta V vectors are assumed to be independent Figure 2: Components of the SRP torque on a 48 hr interval in from each other and from the tracking data. Therefore, January 2015: observed (red), computed (blue). the functional to be minimized can be represented as =( ) P( )+ o c o c matrix,  is the error in the estimated magnitude of v, and  is the error in the estimate of v components in X d j j T SRP j j + (M M ) P (M M )+ the plane orthogonal to e. The typical value of magnitude SRP;o SRP j SRP;o SRP j=1 error  used in the OD is 5% of v, the value of the or- X thogonal component,  , corresponds to a 0.25 deg error 0 T 0 + (v v ) P (v v ): (5) i i i i i in direction or 4:36 10 of v. i=1 In reality the functional is not limited to the one in Equation (5). It accounts also for the a priori information The rst term in Equation (5) gives the contribution of the on range biases and the on-board H-maser frequency o set. tracking data with corresponding computed values o c and inverted covariance estimate of as weight matrix P. 5. ASC correlator overview The second term in Equation (5) contains the sum of squares of residual torques, the observed and computed Orbital solutions obtained as a result of the OD pro- values of which were described in the previous section cedure outlined above are used mostly for scienti c appli- (Equations (4) and (2)). cations, primarily for correlation of VLBI data collected The covariance matrices and their inverses, P , can by the mission. In this section we review the operations SRP be estimated using Equation (4), given the estimation er- of one of the correlators capable of processing RadioAs- rors of the rotation rates of the reaction wheels and the ori- tron data, the ASC Correlator. This FX correlator has entations of their axes are known. In practice, however, the been developed by the Astro Space Center of the Lebe- weights have to be adjusted to an accuracy level of about dev Physical Institute (ASC LPI) Likhachev et al. (2017) 2 10 Nm, since the perturbing torque model described speci cally to support the RadioAstron mission. It is used above is relatively simple and does not always match the to process the majority of observations performed by Ra- accuracy of observations. Fig. 2 shows the agreement be- dioAstron (95%) and thus provides the largest data set of tween the modeled torque and the observational data. It is the so-called residual delays and delay rates, which we will clear that for the X and Z components the degree of agree- use in the next section to evaluate the orbit determination ment, in relative terms, is signi cantly lower than that for accuracy. the Y component, but it is the latter that contains the Correlation of space-VLBI data is performed by the dominant part of the torque due to SRP. ASC Correlator in two steps, with the second step being The last term of Equation (5) takes into account the a necessary due to the orbit determination uncertainties. In priori information on Delta V's due to desaturations. For order to decrease the residual delay, delay rate and delay each desaturation event the value of net v is calculated acceleration, the rst step, or pass, is performed in the so- using information on every thruster ring that occurred called \wide" window mode. The window size along the during that event, including its duration, the fuel pressure delay is determined by the number of spectral channels, and the a priori thrust model. Each of the weight matrices, while along the fringe rate it is determined by the integra- P , in (5) is the inverse of the respective covariance matrix i tion time. Both parameters are set before the correlation given by process is started and can be easily adjusted. After a fringe was found, the second correlation pass is performed in a 2 T 2 T C =  (E e e ) +  e e : d m smaller window taking into account the residual delays ob- Here, e is the unit vector of v which has the same di- tained in the rst pass. rection as X-axis of the spacecraft, E is a 3  3 identity 02, 00:00 02, 06:00 02, 12:00 02, 18:00 03, 00:00 03, 06:00 03, 12:00 03, 18:00 04, 00:00 Mx N m M N m My N m RadioAstron's capability to simultaneously observe at observations performed since 2014 and a signi cant por- two di erent wavelengths is of great value for performing tion of earlier observations were correlated using orbits the fringe search. In case an observation was performed produced with this version of the algorithm. At the early at two di erent wavelengths, the residual values obtained stages of the mission, i.e. for experiments conducted be- as a result of the successful fringe search at the longer fore 2014, a previous, less sophisticated version of the OD wavelength can usually be used to signi cantly simplify algorithm was used. the fringe search at the shorter one. A delay model for space-VLBI is naturally more com- 1e 10 plicated than those used for ground-based VLBI. Apart ver. 2 ver. 1 from issues related to the orbit, it has to take into account the fact that the delay rate depends not only on relative ve- locities of the telescopes but also on the frequency o set of the uplink signal, which is used on board as a reference for the on-board scienti c equipment (for the two-way phase- 1 locked loop mode of synchronization), or on the frequency o set of the on-board H-maser (for the one-way mode). The latter, as it turned out, could result in delay rates as high as 30 35 ps/s. A distinctive feature of the ASC Correlator is its de- lay model ORBITA2012. This delay model was developed speci cally for the ASC Correlator and is capable of calcu- Date lating the delay up to O(c ) terms Vlasov et al. (2012), which provides for using smaller correlation window sizes at the rst pass of the correlation process. Moreover, this Figure 3: Residual delay rates obtained by the ASC Correlator using delay model is capable of taking into account the delay two versions of the orbit: generic algorithm, version 1 (red dots); new acceleration. algorithm, version 2 (blue dots). The post-correlation data reduction for all experiments considered in this paper was performed using the PIMA A statistically signi cant correlation detected between package Petrov et al. (2011). This included baseline fringe signals recorded by the RadioAstron space radio telescope tting, bandpass and amplitude calibration, and averag- and ground-based radio telescopes yields a set of three ing the data over time and frequency. The primary goal of residual values, that of delay, delay rate and delay acceler- the post-correlation processing is to allow for studying the ation. In this paper we focus exclusively on the delay rates physical properties of the observed objects. The residual for the following reasons. First, data collected by the space values of group delay, delay rate and delay acceleration radio telescope are not time-tagged on board, i.e. synchro- are a byproduct of this processing. The total delay and its nization of the on-board H-maser time scale with that of derivatives can be further used for studying the observed UTC is possible only indirectly, by means of time-tagging objects, e.g. in imaging experiments. This requires accu- the data stream owing from the satellite by the receiving rate calibration of a number of uncertainties, e.g. the rates tracking station. Second, there is a non-zero probability of the telescope clocks, their trajectories (including that of of inconsistent synchronization of ground radio telescopes the SRT), coordinates of the observed sources, propagation with UTC. And nally, as shown below, there are a num- media parameters, etc. ber of signi cant orbit-unrelated factors, which can con- The residual delays and their derivatives, which we tribute to the residual delay rate also a ecting the residual used in this analysis, were obtained as a result of correla- delay. All this makes it dicult to separate the contribu- tion of the observations of the RadioAstron AGN survey tion to residual delay due to errors in the orbit from those science program and were not used in the OD procedure. due to the involved clocks. On the other hand, possible The sources observed in this program are compact enough, contributions to the residual delay rate are much more so that the contribution to residual values due to non-point predictable. Time synchronization by means of a tracking structure of the sources is negligible. station does not a ect the delay rate, since any change in the delay it may introduce remains constant during every \scan" ( le) of the recorded data. Drifts of ground-based 6. Orbit determination results clocks are usually much smaller than the residual delay rates we observe and, moreover, when an experiment in- The orbit determination algorithm outlined in this pa- volves more than one ground radio telescope, relative drifts per has been implemented and used by the Keldysh In- of the ground-based clocks can be estimated by perform- stitute of Applied Mathematics (KIAM) to provide Ra- ing fringe searches on ground-only baselines. Finally, the dioAstron users with a posteriori orbits required to corre- clock drift of the spacecraft clock can be estimated inde- late VLBI data collected by the mission. The majority of pendently in the process of orbit determination. Residual delay rate, s/s Fig. 3 shows the residual delay rates obtained by the 1e 11 ASC Correlator as a result of processing the observations ARECIBO EVPTRIYA KALYAZIN TIANMA65 performed by the RadioAstron mission up to early 2017. ATCA-104 GBT-VLBA MEDICINA TORUN BADARY HARTRAO MOPRA WSTRB-07 Each data point in Fig. 3 corresponds to a value of the CEDUNA HOBART26 NOTO YEBES40M DSS63 IRBENE PARKES ZELENCHK residual delay rate determined by the correlator as a result EFLSBERG JODRELL2 SVETLOE of correlating two data scans, i.e. units of data typically of 10 min duration, one of which was recorded by the Ra- 0 dioAstron space radio telescope and the other by a ground radio telescope that participated in the experiment. Blue dots denote the residual delay rates obtained with version 4 2 orbits, i.e. those determined with the algorithm outlined in Section 4. Red dots denote the values obtained with verion 1 orbits, i.e. determined with a generic OD algo- rithm and a simpler dynamic model of the spacecraft. The Date earlier version of the algorithm uses a simple cannonball model to estimate the SRP and thus ignores its attitude dependence. Also, the velocity changes due to momentum Figure 4: The residual delay rates obtained using version 2 orbits. dumps are not estimated in the generic algorithm but set Color indicates the ground radio telescope of the baseline. to their a priori values, v . For each data point of Fig. 3 it is assumed that RadioAstron is the rst antenna while accurate to at least 10 : the ground radio telescope is the second one, i.e. the de- d 1 picted residual delay rates are relative to the space-based = k v (t ); (6) 0 1 2 antenna. dt c All the observations used to obtain the residual delay where k is a unit vector pointing in the direction of the rates of Fig. 3 were conducted in the so-called \one-way" source (the aberration factor can be omitted provided our mode, which is characterized by the following two condi- accuracy requirements). tions: a) the on-board scienti c equipment is synchronized The delay understood by the correlator is a di erence to the reference signal of the on-board H-maser; b) the car- of proper time intervals counted by the station clocks from rier of the downlink signal used to transfer the science data the moment of synchronization to the moment of signal re- from the spacecraft to a tracking station is also synchro- ception. Thus the estimate for the TCG delay rate given nized to the reference H-maser signal. by Equation (6) may be insuciently accurate, especially It is clear from Fig. 3 that the variation of the residual in the case when one of the stations is orbiting the Earth. delay rates obtained with version 2 of the OD algorithm is Assuming the clocks were synchronized at t the delay substantially smaller than that of the delay rates obtained should be as follows with the generic algorithm. The RMS of the residual de- Z   Z t 2 t 2 2 1 1 v 1 v 11 10 2 1 lay rates are 2:463 10 for version 2 and 1:097 10 =  + + U (r ) dt + U (r ) 0 2 1 2 2 c 2 c 2 for version 1. It is also clear that the residual delay rates t t s s obtained with version 2 orbits are o set from zero, with @V (R ) +V (R ) V (R ) r dt; 1  1 the mean value equal to 1:663  10 and also reveal a @R long-term linear trend (Fig. 4). The latter is clearly an in- (7) dication of a systematic e ect which cannot be attributed where  is the TCG delay, t and t are, respectively, the 0 1 2 to an OD error since these residual values were obtained moments of signal reception by station 1 and station 2, using a large number of independent orbit solutions and an U (r) is the Newtonian potential of the Earth at position even greater number of observations of radio sources scat- r, V is the sum of the Newtonian potentials of all bodies of tered all over the sky and observed at various projected the Solar System excluding the Earth, R is the barycen- baselines. tric position of the spacecraft, and R is the barycentric Let us consider the impact of the orbit errors on the position of the geocenter. The tidal potential is neglected calculated value of the group delay reported by the cor- for the ground-based station (i.e. station 2), following the relator. Denote the spacecraft state vector in a geocen- recommendations of Petit and Luzum (2010). tric inertial reference frame according to the orbital so- T T T The change of the delay given by Equation (7) due to lution by X(t) = (r (t) ; v (t) ) , the true state vector 1 1 T T T errors in the spacecraft position, r , and velocity, v , 1 1 by X(t) = (r (t); v (t)) and the di erence between the 1 1 T T T can be represented as two by X(t) = X(t) X(t) = (r (t) ; v (t) ) . Pro- 1 1 vided that the requirements to OD accuracy stated in Sec- (r ; v ) = (r ; v )) 1 1 0 1 1 tion 1 are met, the error in the Geocentric Coordinate 1 r  r 1 1 Time (TCG) delay rate Vlasov et al. (2012) due to OD v  v U dt; (8) 1 1 2 2 c r errors can be given by the following equation, which is s 1 Residual delay rate, s/s P P 3 3 1 @ V where the tidal term, which can be approximated by r r , i j 2 i=1 j=1 @R @R i j 1e 11 has been neglected since for RadioAstron's orbit and the expected magnitude of OD errors, the contribution of this term to the delay rate is less than 10 . Therefore the correction to the delay rate can be written as d d v  v r  r 1 1 1 1 =  U : (9) 2 2 dt dt c c r Provided that OD errors do not exceed the values stated in Section 1, the last two terms on the right-hand side of Equation (9) contribute at most 10 . Therefore, if we aim at evaluating the correction to the delay rate with Date an accuracy of 10 , these two terms can be neglected. Adding to this the correction due to the on-board H-maser Figure 5: Evolution of the RadioAstron on-board H-maser fractional clock drift, h (t) and that of the ground-based station, frequency bias. Each value is obtained as part of an orbital solu- h (t), we arrive at the following equation tion. The density of points before and after October 2014 is di erent because of a change of the frequency bias estimation strategy. d 1 ((t t ) (t t ) ) = k v (t ) h (t ) + h (t ); 2 s p 1 s p 1 2 1 1 2 2 dt c (10) i.e. when both the on-board science equipment and the downlink carrier are locked to the on-board H-maser ref- which can be used to investigate the spacecraft velocity erence signal. Therefore, if the on-board H-maser output error, provided it is larger than 10 but within the frequency is biased, i.e. there is a non-zero clock drift, both limit given in Section 1, i.e. less than 10 . The two the science data, which eventually results in the residual additional assumptions needed to justify the use of Equa- delay rates used in our analysis, and the downlink carrier tion (10) for this purpose are: a) the position and velocity frequency measurements, which are used in OD, will be of the ground-based station (#2) are known much better a ected by this same frequency bias. than those of the spacecraft and b) the error of model- In order to properly t one-way Doppler measurements ing the contribution of the propagation media to the delay in the OD process, frequency bias of the tracking station rate is less than 10 . While the rst assumption is un- doubtedly true, the second may be violated in rare cases H-maser and that of the on-board H-maser have to be of extreme weather conditions. taken into account. For each of the two tracking stations we used a priori values of the frequency biases, which we The delay model implemented by the ASC Correlator derived from long-term series of measured di erences be- assumes that the spacecraft and station clocks are ideal, tween the station's local time, synchronized to its H-maser, i.e. h (t)  0 and h (t)  0. Therefore, any instrumental 1 2 and the GPS time. We assume the on-board H-maser frac- e ect a ecting the clock drift, if present, would manifest tional frequency bias, f=f , to be identical to that of the itself in an extra residual delay rate reported by the cor- downlink carrier, and estimate the latter in the OD pro- relator. cess. If we assume that no other systematic e ects are present, The on-board H-maser frequency biases obtained this each data point shown in Fig. 3 corresponds to the right- way are shown in Fig. 5. The bias evolution is almost linear hand side of Equation (10) evaluated at a certain time over more than 4 years of the analyzed data. In late Octo- using a certain orbital solution plus an error of the delay ber 2014 an accident occurred on board after which the on- rate determination inherent to the fringe search procedure. board H-maser started experiencing a power surge every Therefore, if we are able to evaluate the clock drifts of the time one of the telemetry and command (T&C) transpon- space and ground-based station clocks, and apply the cor- der was turned on. Before that accident our approach responding corrections to each data point of Fig. 3, the to estimating the on-board H-maser frequency bias was result will be the spacecraft velocity errors projected onto to treat it as constant over the whole OD interval. Af- the directions to the observed celestial sources. ter the accident we changed our strategy and estimated We assume that the ground-based station clock drift, the bias independently for each interval between the T&C h (t), is always well below the level of 10 and thus link transponder on/o switches, which usually coincided can be neglected, which is justi ed by the results of fringe with radio tracking from Bear Lakes or Ussuriysk stations. searches at ground-only baselines. Under this assumption, This change of strategy explains the di erent density of any clock drift of the ground station, if present, will be attributed to the error in the spacecraft velocity. The drift data points in Fig. 5 before and after October 2014. Using the above assumption that the downlink carrier of the on-board clock, h (t), is estimated as part of the OD fractional frequency bias, f=f , is equal to that of the process in the way outlined below. on-board H-maser, i.e to the on-board clock drift, h , we All the observations used for the present analysis were can substract it from the residual delay rates using Equa- performed, as noted above, in the so-called one-way mode, Fractional frequency bias tion (10). The residual delay rates obtained by applying We are working on developing a full understanding of this correction are shown in Fig. 6 (cf. Fig. 4). With the why the correction of Equation (11) is required. How- correction applied the RMS of the residual delay rates re- ever, its origin is likely due to either an inconsistency be- duces to 1:239  10 and its mean value reduces from tween the models of orbital motion or the delay model used 11 12 1:663 10 to 3:735 10 . by the correlator, or both. This can be concluded from the following reasoning. Orbital solutions are provided in the geocentric coordinate system and the Terrestrial Time 1e 11 (TT) scale, as the correlator delay model expects. In that ARECIBO EVPTRIYA KALYAZIN TIANMA65 case the e ect of the solar gravitational potential on the ATCA-104 GBT-VLBA MEDICINA TORUN BADARY HARTRAO MOPRA USUDA64 delay is described by the last three terms in the second CEDUNA HOBART26 NOTO WSTRB-07 DSS63 IRBENE PARKES YEBES40M integral on the right-hand side of Equation (7), with the EFLSBERG JODRELL2 SVETLOE ZELENCHK potential V replaced by U . For the delay rate it di ers from the correction in Equation (11) by the third term, @U (R )=@R r =c , which is approximately equal to the expression in Equation (11) and makes the tidal e ect of the solar gravitational potential almost negligible in the case of RadioAstron. Therefore, the correction in the form of Equation (11) cannot be due to an unaccounted e ect of the solar gravitational potential but can only arise due to an inconsistency in the computation of delays using the Date provided orbit solutions. The residual delay rates after application of the cor- rection of Equation (11) are shown in Fig. 9. In addition Figure 6: The residual delay rates obtained using version 2 orbits and to overall reduction of delay rate variation, it is now clear corrected for the on-board clock drift. Color indicates the ground radio telescope of the baseline. that part of the data points, i.e. those that correspond to RadioAstron{Torun baselines, have an appreciable bias of According to Equation (10), we expect each value of about 1:9  10 , which can reasonably be attributed the residual delay rate corrected for the spacecraft clock to the clock drift of this ground radio telescope. drift to be equal to the corresponding spacecraft velocity Excluding the Torun data, which constitute 8.2% of the dataset, we arrive at the mean value of the residual delay error projected onto the direction of the observed radio 13 12 rates of just 4:387 10 and its RMS of 9:573 10 . source. Therefore, we expect them to be largely uncorre- The distribution of the data also becomes more symmetric lated and symmetrically distributed with respect to zero. (Fig. 8). However, the values shown in Fig. 6 still reveal signi cant In order to investigate the statistical properties of the systematic e ects, i.e. the average value is 1.12 mm/s in residual delay rates shown in Fig. 9 further, we exclude terms of velocity and there are noticeable variations with the data of experiments conducted in the summer peri- a period of about 1 year. The histogram of the data is ods, i.e. between the 1st of June and the 1st of Septem- shown on the left-hand side of Fig. 8. ber. These experiments constitute 10.4% of data, but their A more detailed analysis of the data shown in Fig. 6 RMS, equal to 2:307 10 , is more than 5 times larger using the Lomb-Scargle periodogram method reveals that, than that of the rest of the data. Degrade orbit deter- in addition to the annual variation, the residual delay rates are modulated with a period of about 8.8 day (Fig. 7), mination accuracy is de nitely one of the main reasons which is close to the average orbital period. why the summer experiments exhibit larger residual delay The periodic variations visible in Fig. 6 and con rmed rates. The number of scienti c observations drops signif- on the left panel of Fig. 7 are present in the original data icantly during summer, mainly due to constrains on the and not due to the applied clock correction h (t). spacecraft attitude with respect to the Sun, which reduce In an attempt to understand the origin of these pe- the visibility of radio sources of interest. The lack of ex- riodicities we tried several hypotheses. We were able to periments reduces the amount of Doppler data from the determine that the periodic patterns in the residual delay tracking stations which, in turn, a ects the orbit accuracy. rates can be noticeably reduced by applying the following After excluding the summer experiments and ltering correction out outliers, which, for the 5 threshold, constitute 1.15% of the original data, we obtain a dataset which contains _ = (U (R) U (R )) : (11) 88.4% of the initial number of residual delay rates. For this subset of the data the mean residual delay rate is 4:428 13 12 Here, U is the Newtonian potential of the Sun, R is the 10 and the RMS is 4:255 10 . This corresponds to barycentric position of the RadioAstron spacecraft during the standard deviation of the projected spacecraft velocity the experiment and R is the barycentric position of the error, k v , of 1.275 mm/s. geocenter. To summarize, we obtained a dataset of the spacecraft Residual delay rate, s/s Figure 7: A Lomb-Scargle periodogram of residual delay rates corrected for the on-board H-maser drift before (left) and after (right) application of the correction of Equation (11). Signi cant spikes are annotated with the periods (in days) corresponding to their frequencies. 1e 11 0.5 0.175 ARECIBO EVPTRIYA KALYAZIN TIANMA65 ATCA-104 GBT-VLBA MEDICINA TORUN BADARY HARTRAO MOPRA USUDA64 CEDUNA HOBART26 NOTO WSTRB-07 0.150 0.4 DSS63 IRBENE PARKES YEBES40M EFLSBERG JODRELL2 SVETLOE ZELENCHK 0.125 0.3 0.100 0.075 0.2 0.050 0.1 0.025 0.000 0.0 4 2 0 2 4 6 8 4 2 0 2 4 Date k· δv mm/s k· δv mm/s 1 1 Figure 8: Histograms of the residual delay rates after application Figure 9: The residual delay rates obtained using version 2 orbits of the correction for the on-board clock drift. Left: the additional corrected for the on-board clock drift and for the correction of Equa- correction of Equation (11) is not applied. Right: the correction of tion (11). Color indicates the ground radio telescope of the baseline. Equation (11) is applied. Experiments with the Torun ground radio telescope are excluded. to the tracking station that was receiving data from the spacecraft. velocity error estimates projected onto the directions of the Since every science observation is accompanied by si- observed sources, k v . We assumed that clock drifts of multaneous Doppler measurements performed by the track- ground-based VLBI stations that provided the ground legs ing station used to receive the science data, the largest of the baselines are negligible, the on-board clock drift is errors are to be expected in the velocity components that equal to the estimated value of the downlink carrier frac- belong to the plane which is normal to the tracking station tional frequency o set, and we applied the correction of direction. However, contrary to this expectation, Fig. 10 Equation (11) to remove systematic patterns in the resid- does not reveal any signi cant dependence of the projected ual delay rates due to suspected inconsistencies between velocity error, k v , on . A possible reason for this is the orbit and correlator delay models. Now, in order to that the corrected residual delay rate, and thus kv , con- obtain information on the spacecraft velocity error itself, tains errors due to inaccurate estimation of the on-board v , and not only on its projection, kv , we represent the 1 1 clock drift, h (t), and the assumption on the ground sta- residual delay rates as a function of the angle, , between tion clocks that h (t)  0, both of which do not depend the direction to the observed source, k, and the direction Residual delay rate, s/s on the source direction, k, and may provide signi cant vided by the discontinuity analysis and those obtained us- contributions to the residual delay rates. ing the residual delay rates are similar. However, since the former contain, as noted above, additional prediction errors, we expected them to be higher. This may be an indication of the fact that even after the corrections to the residual delay rates have been applied, clock-related errors still give a signi cant contribution to these data, resulting in overestimated values of velocity errors. 7. Discussion We obtained two independent estimates of the veloc- 2 ity errors, the one using the comparison of adjacent inde- pendent orbits and the other via the residual delay rate analysis. The obtained results agree very well. However, in order to obtain the velocity error estimates using the residual delay rate data, we had to apply two non-trivial 0 25 50 75 100 125 150 175 corrections to the data. Angle, α degrees The rst correction is to take into account an instru- mental e ect of the spacecraft clock drift. The signi cance Figure 10: Dependence of the residual delay rates on the angle be- of the on-board H-maser clock drift is evident from the tween the direction to observed source and the direction from the combined analysis of two-way range and Doppler track- spacecraft to the tracking station which was used to receive the data of that observation ing data together with the one-way Doppler measurements of the downlink carrier synchronized to the on-board H- We can estimate the velocity error along the \worst" maser. Since the ASC Correlator assumes the clocks to directions using the dependence of the projected veloc- be ideal, i.e. any non-zero clock drift goes directly into ity error, k  v , on . For the experiments, in which the residual delay rate, the necessity to correct for the 75   105 , the RMS of k  v is equal to 1.359 on-board clock drift is well justi ed. This correction, how- mm/s. Assuming the spacecraft velocity is determined ever, introduces additional errors into the residual delay not worse along other directions the error ellipsoid of v rate data due to the way it is estimated in the process of is con ned by the one corresponding to the covariance ma- OD. 2 2 2 2 2 2 trix diag(1:85 mm =s ; 1:85 mm =s ; 1:85 mm =s ). The necessity of the second correction, that of Equa- In order to obtain an independent estimate of the ve- tion (11), is also clear as it signi cantly improves the resid- locity error, we compared 55 independent adjacent orbit ual delay rates by cleaning them from arti cial periodic solutions from 2014 to 2018 for discontinuity in velocity. If patterns. However, its nature is currently not as well un- the rst solution was obtained for, e.g., the time interval derstood. We suspect the necessity of this correction is of (t ; t ) and the second one for (t ; t ), the spacecraft 1 2 2 3 due to an inconsistency between the models of orbital mo- state vector is compared at t . The time of comparison, tion or the delay model used by the correlator, or both. t , is thus either the beginning or the end of an OD inter- Investigation of the physical cause behind this corrections val. Unlike the case of the residual delay rates, these times is beyond the scope of the present work and would require do not correspond to any tracking activity and, therefore, a deeper insight in the experimental data accumulated by the state vectors evaluated at these moments contain ad- the mission. ditional prediction errors. The residual delay rate data provide a unique means The 3D RMS of the 54 velocity di erences computed of estimating velocity errors of a space-VLBI spacecraft. this way is 4.38 mm/s. Just like the residual delay rates, If the mission performs observations of radio sources dis- these velocity di erences exhibit signi cant increase in mag- tributed all over the sky, the method e ectively provides nitude in the summer months due to the decrease of the estimates of all components of the spacecraft velocity er- amount of tracking data. If we exclude the summer data, ror, as shown in Fig. 10. we arrive at a dataset of 42 points and 3D RMS of the The OD accuracy evaluation of the only previous space- velocity di erences of 3.52 mm/s. VLBI mission of VSOP/HALCA, given by You et al. (1998), However, we should take into account that such dif- did not include an analysis of the residual delay rate data. ferences contain errors of two independent orbit solutions, However, the overall OD accuracy of the HALCA space- therefore, their variance is as sum of variances of two orbit craft, obtained with traditional methods, was better than solutions. Thus, we can conclude that the 3D RMS of the that of RadioAstron. This can be easily understood as velocity error of a single solution is 3.10 mm/s for the full the HALCA spacecraft was tracked by a larger number of dataset and 2.53 mm/s for its summer-free subset. tracking stations, each possibly with better instrumental The RMS values of the spacecraft velocity error pro- stability than those involved in RadioAstron. k· δv , mm/s 1 8. Conclusion Bank Observatory is a facility of the National Science Foundation operated under cooperative agreement by As- We outlined our approach to orbit determination of sociated Universities, Inc. The Arecibo Observatory is op- the RadioAstron spacecraft. The method includes an in- erated by SRI International under a cooperative agreement house developed SRP model and an algorithm to take into with the National Science Foundation (AST-1100968), and account the accumulated angular momentum of reaction in alliance with Ana G. Mendez-Universidad Metropoli- wheels to improve our knowledge of the dynamics of the tana, and the Universities Space Research Association. spacecraft center of mass. We tested the performance of The Australia Telescope Compact Array (Parkes radio tele- this OD method using the unique \tracking data" available scope / Mopra radio telescope / Long Baseline Array) is only for a space-VLBI spacecraft, i.e. the residual delay part of the Australia Telescope National Facility which rates, which in our case were obtained by the ASC Corre- is funded by the Commonwealth of Australia for oper- lator of the RadioAstron mission as a result of processing ation as a National Facility managed by CSIRO. This the data of VLBI observations of celestial radio sources work is based in part on observations carried out using performed by the RadioAstron spacecraft together with the 32-meter radio telescope operated by Torun Centre ground radio telescopes. This analysis allowed us to con- for Astronomy of Nicolaus Copernicus University in Torun clude that the OD method we developed provides up to (Poland) and supported by the Polish Ministry of Science 11 times more accurate orbital solutions in terms of veloc- and Higher Education SpUB grant. Results of optical ity and residual delay rates as compared to a generic OD positioning measurements of the Spektr-R spacecraft by algorithm. the global MASTER Robotic Net Lipunov et al. (2010), Using the residual delay rate data we have obtained ISON collaboration, and Kourovka observatory were used an estimate of the standard deviation for every compo- for spacecraft orbit determination in addition to mission nent of the spacecraft velocity error. The residual delay facilities. rates obtained in non-summer observations, which consti- tute 88.4% of all data, exhibit a standard deviation of References the spacecraft velocity error of 1.4 mm/s or less for each velocity component. This result is consistent with an in- Duev, D.A., Zakhvatkin, M.V., Stepanyants, V.A., et al., 2015. Ra- dependent a posteriori estimate of the spacecraft velocity dioAstron as a target and as an instrument: Enhancing the Space VLBI mission's scienti c output. Astronomy & Astrophysics 573, error obtained using the analysis of velocity di erences A99. doi:10.1051/0004-6361/201424940, arXiv:1411.4576. computed at the boundaries of adjacent orbital solutions. Kardashev, N.S., Khartov, V.V., Abramov, V.V., et al., 2013. \RadioAstron"-A telescope with a size of 300 000 km: Main pa- rameters and rst observational results. Astronomy Reports 57, 9. Acknowledgements 153{194. doi:10.1134/S1063772913030025. Likhachev, S.F., Kostenko, V.I., Girin, I.A., et al., 2017. Soft- The RadioAstron project is led by the Astro Space ware Correlator for Radioastron Mission. Journal of Astro- Center of the Lebedev Physical Institute of the Russian nomical Instrumentation 6, 17. doi:10.1142/S2251171717500040, arXiv:1706.06320. Academy of Sciences and the Lavochkin Scienti c and Pro- Lipunov, V., Kornilov, V., Gorbovskoy, E., et al., 2010. Master duction Association under a contract with the Russian Robotic Net. Advances in Astronomy 2010, 349171. doi:10.1155/ Federal Space Agency, in collaboration with partner orga- 2010/349171, arXiv:0907.0827. Litvinov, D., Rudenko, V., Alakoz, A., et al., 2018. Probing the nizations in Russia and other countries including Keldysh gravitational redshift with an Earth-orbiting satellite. Physics Institute of Applied Mathematics of the Russian Academy Letters A 382, 2192 { 2198. doi:10.1016/j.physleta.2017.09. of Sciences. Partly based on observations performed with 014. Special Issue in memory of Professor V.B. Braginsky. radio telescopes of IAA RAS (Federal State Budget Scien- Nunes, N., Bartel, N., Bietenholz, M., et al., 2019. The gravita- tional redshift monitored with RadioAstron from near Earth up ti c Organization Institue of Applied Astronomy of Rus- to 350,000 km. Advances in Space Research doi:10.1016/j.asr. sian Academy of Sciences). Partly based on the Evpato- 2019.03.012, arXiv:1904.01060. ria RT-70 radio telescope (Ukraine) observations carried Petit, G., Luzum, B., 2010. IERS Conventions (2010). Technical Re- out by the Institute of Radio Astronomy of the National port. Verlag des Bundesamts fr Kartographie und Geodsie. IERS Technical Note ; 36. Academy of Sciences of Ukraine under a contract with Petrov, L., Kovalev, Y.Y., Fomalont, E.B., et al., 2011. The Very the State Space Agency of Ukraine and by the National Long Baseline Array Galactic Plane Survey { VGaPS. The As- Space Facilities Control and Test Center with technical tronomical Journal 142, 35. doi:10.1088/0004-6256/142/2/35, support by Astro Space Center of Lebedev Physical In- arXiv:1101.1460. Vlasov, I.Y., Zharov, V.E., Sazhin, M.V., 2012. Signal delay in the stitute, Russian Academy of Sciences. Partly based on RadioAstron ground-space interferometer. Astronomy Reports 56, observations with the 100-m telescope of the MPIfR (Max- 984{987. doi:10.1134/S1063772912120086. Planck-Institute for Radio Astronomy) at E elsberg. Partly You, T.H., Ellis, J., Mottinger, N., 1998. Navigation of the space based on observations with the Medicina (Noto) telescope VLBI mission-HALCA, in: Stengle, T. (Ed.), AAS/GSFC 13th International Symposium on Space Flight Dynamics, pp. 841{856. operated by INAF - Istituto di Radioastronomia. The Na- Zakhvatkin, M.V., Ponomarev, Y.N., Stepanyants, V.A., et al., 2014. tional Radio Astronomy Observatory is a facility of the Navigation support for the RadioAstron mission. Cosmic Research National Science Foundation operated under cooperative 52, 342{352. doi:10.1134/S0010952514050128. agreement by Associated Universities, Inc. The Green

Journal

AstrophysicsarXiv (Cornell University)

Published: Dec 4, 2018

References