TY - JOUR
AU - Holali, Apevienyeku Kwami
AB - Abstract Traffic accidents on highways are attributed mostly to the "invisibility" of oncoming traffic and road signs. "Speeding" also causes drivers to reduce the effective radius of the vehicle path in the curve, thus trespassing into the lane of the oncoming traffic. The main aim of this paper was to develop a multisensory obstacle-detection device that is affordable, easy to implement and easy to maintain to reduce the risk of road accidents at blind corners. An ultrasonic sensor module with a maximum measuring angle of 15° was used to ensure that a significant portion of the lane was detected at the blind corner. The sensor covered a minimum effective area of 0.5 m2 of the road for obstacle detection. Yellow light was employed to signify caution while negotiating the blind corner. Two photoresistors (PRs) were used as sensors because of the limited number of pins on the microcontroller (Arduino Uno). However, the device developed for this project achieved obstacle detection at blind corners at relatively low cost and can be accessed by all road users. In real-world applications, the use of piezoelectric accelerometers (vibration sensors) instead of PR sensors would be more desirable in order to detect not only cars but also two-wheelers. 1. Introduction The increasing rate of accidents on highways is generally attributed to late spotting, invisible oncoming traffic and the inability of vehicles to maintain their lanes. "Invisible" road signs are also a causative factor in road accidents, especially at night or during low-visibility weather conditions. To stem the rise in fatal crashes, a number of studies have recommended the adoption of integrated interventions spanning education, engineering measures, enforcement and trauma care [1, 2]. One can also argue that an automated highway system (AHS), as shown in Fig. 1, will be safer, since statistics suggest that human error accounts for 90% of accidents [3]. Fig. 1. Open in new tabDownload slide AHS technology concept [4]. Fig. 1. Open in new tabDownload slide AHS technology concept [4]. Land resources are limited in many countries. It is thus difficult to build new infrastructure such as highways and freeways [5]. Recent developments in automating highways in other parts of the world and the rise of smart vehicles have made it a necessity for Ghana to start developing automated devices that will pioneer the development of automated highways. This initiative has, in other parts of the world, met with considerable success, ensuring more efficient, safe and predictable road experiences. In this project, a multisensory obstacle-detection device was developed using the Arduino Uno microcontroller to reduce the risk of road accidents at blind corners. This microcontroller was chosen because it has an open-source programming environment. 1.1 Review of blind corners, signage and technologies Blind corners exist because of obstructions such as hills, mountains, buildings, curve which makes it impossible to have a clear line of sight of incoming traffic. A blind spot on a road is one that is generated by vehicles in front obstructing the view of the driver behind, who therefore has no view of the oncoming traffic [6]. A blind-corner sign gives drivers a "heads-up" on the potential danger of collision with the oncoming traffic when it suddenly appears. A noticeable yellow diamond and other traffic signs alert oncoming traffic that dangerous and hidden intersections or curves are ahead. In some countries, especially more developed countries, the use of a convex mirror to enhance traffic visibility is prominent. These convex mirrors have the limitation of providing a poor image if the surface of the mirror is dusty or when there is fog or low visibility. Since dusty conditions are prevalent in Ghana, convex mirrors may not be appropriate. To limit the frequency of accidents in blind-corner environments, Chelli et al. demonstrated that a geometry-based approach to deriving a narrowband single-input, single-output (SISO) vehicle-to-infrastructure (V2I) channel model can be implemented [7]. In Chelli et al.’s work, a transmitter MS tries to communicate with a fixed RSU deployed on one side of the road, as shown in Fig. 2. However, the implementation of the V2I channel model has financial and technical challenges. Fig. 2. Open in new tabDownload slide Geometrical blind-corner model for V2I communications [7]. Fig. 2. Open in new tabDownload slide Geometrical blind-corner model for V2I communications [7]. According to Elefteriadou et al. [8], Blind Corner Monitoring technology relates to the visual functions (the ability to spot incoming traffic) in approaching crossings and has a potential impact on gap acceptance properties. This technology makes use of frontal cameras with built-in prisms mounted in the middle of the grille, which send pictures to the display screen inside the car. When the car approaches a crossing where visibility is reduced to the left and/or to the right, the image on the display shows a view of approximately 20 m in both directions, at an angle of 25°. Shimizu et al. [9] emphasized that to improve the driver’s visual field, cameras at the front end of the vehicle must take enhanced images ahead of the vehicle when the driver enters an intersection with poor visibility. A similar report was made by Corbin et al. [10]. On the other hand, the use of these cameras has the limitation of providing bad visuals when there is fog or low visibility. Additionally, this technology for obstacle detection at blind corners is fixed on individual vehicles. This means that drivers who cannot afford such technologies are likely to be the most vulnerable at blind corners. In a similar vein to other technologies employed at blind corners, Khairnar et al. [11] argued that applications of automotive wireless communication—also known as vehicle-to-vehicle (V2V) communication—along with various automotive applications relying on wireless networks can improve the driver’s "sixth sense" at blind corners. An experiment conducted by Chirupphapa et al. [12] also shows the characteristics of IEEE 802.11p (the Institute of Electrical and Electronics Engineers Local Area Network protocols for vehicular communication systems) communication at blind corners. Chirupphapa et al. carried out preliminary experiments with IEEE 802.11p with real-time devices at a blind corner. Their tests revealed that the nodes cannot communicate if their distance from a blind corner is more than 20 m. They also found that within 15 m of a corner, the nodes send at least five packets of data to ensure successful transmission. These tests paved the way for efficient V2V and V2I interactions via wireless communication at blind corners. The late transmission (short distance to blind corners) of the data packets makes this unsuitable to apply; since the decision-making of the driver takes time, communication at a greater distance from the blind corner is required. Connolly [13] stated that "the Honda ASV-3 uses cameras and millimetre-wave radar to detect obstacles and approaching vehicles and assists in steering and braking, and communicates positional information with other vehicles". The ASV-3’s V2V communication system utilizes 5.8 GHz two-way radio signals to help detect obstacles at blind corners. Furthermore, pedestrians are asked to carry portable transmitters to alert drivers to their presence. This would be difficult to implement on Ghanaian roads because the cost of implementing this technology and the burden on pedestrians would be prohibitive. Finally, it is worth mentioning that the "traditional" methods (road signs for blind corners; convex mirrors) and existing technologies for signalling road users at blind corners have their own setbacks (as depicted in Table 1), such as the low visibility of the signals during unfavourable weather conditions (like fog); encroachment of the road by surrounding vegetation; and the late spotting of the signals before the driver reaches a blind corner, which can be attributed to the dimness of the environment and the characteristics of the blind corner. This project seeks to address the existing setbacks and make it possible for every individual road user to benefit. The use of "simple technologies" to resolve the issue of road accidents at blind corners is meagre, which makes this work relevant. Table 1. Comparison of the differences between existing technologies and the proposed prototype Existing technology . Limitation . Strength of the proposed lab-scaled prototype . Road traffic signs Vegetation encroaches the signs, making them ‘invisible’ to road users. The use of LEDs and the strategic position of the device makes it ‘immune’ to vegetation encroachment. Convex mirrors Provide a poor image when the surface of the mirror is dusty and when there is fog or low visibility. The use of LEDs makes the device more visible under low-visibility conditions. Geometrical blind-corner model for V2I communications with SISO V2I channel model [7] Financial and technical challenges Set-up costs are relatively low and technical expertise required for setting up and maintaining the device is minimal. Frontal cameras with built-in prisms mounted in the middle of the grille [8–10] Bad visuals when there is fog or low visibility The device functions well in spite of low-visibility conditions. Fixed on individual vehicles The device is fixed on the road, making it available to all road users. IEEE 802.11p to demonstrate V2V and V2I at blind corners [12] The nodes cannot communicate if their distance from a blind corner is more than 20 m. Signals indicating blind corners and obstacles at blind corners can be seen several metres away from the blind corners. Late transmission of data packets (especially at greater distance from blind corner) Cameras and millimetre-wave radar in V2V [13] Pedestrians are asked to carry portable transmitters to alert drivers to their presence Every individual road user can benefit without the need to carry transmitters. The cost of implementing this technology and the burden on pedestrians would be prohibitive The cost of implementing this technology is reasonable. Existing technology . Limitation . Strength of the proposed lab-scaled prototype . Road traffic signs Vegetation encroaches the signs, making them ‘invisible’ to road users. The use of LEDs and the strategic position of the device makes it ‘immune’ to vegetation encroachment. Convex mirrors Provide a poor image when the surface of the mirror is dusty and when there is fog or low visibility. The use of LEDs makes the device more visible under low-visibility conditions. Geometrical blind-corner model for V2I communications with SISO V2I channel model [7] Financial and technical challenges Set-up costs are relatively low and technical expertise required for setting up and maintaining the device is minimal. Frontal cameras with built-in prisms mounted in the middle of the grille [8–10] Bad visuals when there is fog or low visibility The device functions well in spite of low-visibility conditions. Fixed on individual vehicles The device is fixed on the road, making it available to all road users. IEEE 802.11p to demonstrate V2V and V2I at blind corners [12] The nodes cannot communicate if their distance from a blind corner is more than 20 m. Signals indicating blind corners and obstacles at blind corners can be seen several metres away from the blind corners. Late transmission of data packets (especially at greater distance from blind corner) Cameras and millimetre-wave radar in V2V [13] Pedestrians are asked to carry portable transmitters to alert drivers to their presence Every individual road user can benefit without the need to carry transmitters. The cost of implementing this technology and the burden on pedestrians would be prohibitive The cost of implementing this technology is reasonable. Open in new tab Table 1. Comparison of the differences between existing technologies and the proposed prototype Existing technology . Limitation . Strength of the proposed lab-scaled prototype . Road traffic signs Vegetation encroaches the signs, making them ‘invisible’ to road users. The use of LEDs and the strategic position of the device makes it ‘immune’ to vegetation encroachment. Convex mirrors Provide a poor image when the surface of the mirror is dusty and when there is fog or low visibility. The use of LEDs makes the device more visible under low-visibility conditions. Geometrical blind-corner model for V2I communications with SISO V2I channel model [7] Financial and technical challenges Set-up costs are relatively low and technical expertise required for setting up and maintaining the device is minimal. Frontal cameras with built-in prisms mounted in the middle of the grille [8–10] Bad visuals when there is fog or low visibility The device functions well in spite of low-visibility conditions. Fixed on individual vehicles The device is fixed on the road, making it available to all road users. IEEE 802.11p to demonstrate V2V and V2I at blind corners [12] The nodes cannot communicate if their distance from a blind corner is more than 20 m. Signals indicating blind corners and obstacles at blind corners can be seen several metres away from the blind corners. Late transmission of data packets (especially at greater distance from blind corner) Cameras and millimetre-wave radar in V2V [13] Pedestrians are asked to carry portable transmitters to alert drivers to their presence Every individual road user can benefit without the need to carry transmitters. The cost of implementing this technology and the burden on pedestrians would be prohibitive The cost of implementing this technology is reasonable. Existing technology . Limitation . Strength of the proposed lab-scaled prototype . Road traffic signs Vegetation encroaches the signs, making them ‘invisible’ to road users. The use of LEDs and the strategic position of the device makes it ‘immune’ to vegetation encroachment. Convex mirrors Provide a poor image when the surface of the mirror is dusty and when there is fog or low visibility. The use of LEDs makes the device more visible under low-visibility conditions. Geometrical blind-corner model for V2I communications with SISO V2I channel model [7] Financial and technical challenges Set-up costs are relatively low and technical expertise required for setting up and maintaining the device is minimal. Frontal cameras with built-in prisms mounted in the middle of the grille [8–10] Bad visuals when there is fog or low visibility The device functions well in spite of low-visibility conditions. Fixed on individual vehicles The device is fixed on the road, making it available to all road users. IEEE 802.11p to demonstrate V2V and V2I at blind corners [12] The nodes cannot communicate if their distance from a blind corner is more than 20 m. Signals indicating blind corners and obstacles at blind corners can be seen several metres away from the blind corners. Late transmission of data packets (especially at greater distance from blind corner) Cameras and millimetre-wave radar in V2V [13] Pedestrians are asked to carry portable transmitters to alert drivers to their presence Every individual road user can benefit without the need to carry transmitters. The cost of implementing this technology and the burden on pedestrians would be prohibitive The cost of implementing this technology is reasonable. Open in new tab 2. Methodology To generate preliminary concepts, AutoCAD drawings of different concept designs were made. The designs were then reviewed and corrected, and recommendations were made to improve the designs. In total, four concepts were created, and the final design was selected through the use of Pugh’s [14] concept-screening matrix method. 2.1 Concept summary For the purpose of illustration, two similar displays were used: light-emitting diodes (LEDs) for Design Concepts 1 and 3, as shown in Figs. 3 and 5, respectively; and liquid crystal displays (LCDs) for Design Concepts 2 and 4, as shown in Figs. 4 and 6, respectively. The same sensor (ultrasonic sensor HC-SR04) was employed for all four design concepts. Fig. 3. Open in new tabDownload slide Design Concept 1. Fig. 3. Open in new tabDownload slide Design Concept 1. Fig. 4. Open in new tabDownload slide Design Concept 2. Fig. 4. Open in new tabDownload slide Design Concept 2. Fig. 5. Open in new tabDownload slide Design Concept 3. Fig. 5. Open in new tabDownload slide Design Concept 3. Fig. 6. Open in new tabDownload slide Design Concept 4. Fig. 6. Open in new tabDownload slide Design Concept 4. In Design Concept 1 (DC 1), shown in Fig. 3, the sensing elements employed were two HC-SR04 ultrasonic sensors and two piezoelectric accelerometers; red and yellow LEDs were used for display, and an Arduino microcontroller (Arduino Uno R3) was used for signal processing. When the piezoelectric accelerometers picked up the road vibrations caused by another vehicle and the ultrasonic sensor detected the distance of the vehicle from the blind corner, signals were generated and sent to the microcontroller for processing. The distance detected by the ultrasonic sensor and the piezoelectric accelerometer signals corresponded to the distance and vibration ranges programmed into the microcontroller; the microcontroller then emitted signals to switch on either the red LEDs (to indicate the presence of obstacles on the road) or the yellow LEDs (to indicate that the road was free of obstacles). For example, when a vehicle driving outbound cycle (in CCW direction) was sensed by both piezoelectric accelerometer 1 and HC-SR04 1, the Arduino microcontroller would switch red LED 2 on to indicate the presence of obstacles at the blind corner on the other side of the road. In Design Concept 2 (DC 2), shown in Fig. 4, the sensing elements employed were strips of photoresistors (PRs) and HC-SR04 ultrasonic sensors; LCDs were used for display, and an Arduino Uno R3 microcontroller was used for signal processing. The PRs sent high signals when light rays fell upon them and low signals when the light intensity was reduced. The PRs sent high signals when there was no vehicle above them (hence no shadow over the PRs) and low signals when a vehicle was above them (that is, the vehicle overshadowed the PRs, reducing the light intensity). When the PRs sensed low light intensity but the ultrasonic sensor did not detect any obstacle on the road, the LCD would display a "No Danger!" indicator on screen. But if both the PR and the ultrasonic sensor detected obstacles on the road, the LCD would display a "Danger!" indicator on screen. For instance, when a vehicle driving outbound cycle (in CCW direction) was sensed by both PR 1 and HC-SR04 1, the Arduino microcontroller would display "Danger!" on LCD 2 to indicate the presence of obstacles at the blind corner on the other side of the road. In Design Concept 3 (DC 3), shown in Fig. 5, the same sensing elements as in DC 2 were employed, while LEDs were used for display. When the sensing elements detected an obstacle, a signal was generated and sent to the microcontroller, which in turn switched on the red LEDs. The yellow LEDs were switched on when there was no obstacle present. In Design Concept 4 (DC 4), shown in Fig. 6, the same sensing elements as those used in DC 1 were employed, while LCDs were used for display. When the sensing elements detected a road obstacle, the LCDs would display a "Danger!" indicator, and the LCDs would display "No Danger!" when there was no road obstacle. It should be noted that in all four design concepts, the sensing elements for one side of the road (Lane 1) were connected to displays in the other lane (Lane 2) and vice versa. 2.2 Concept screening and selection In this project, Pugh’s [14] method for concept screening and selection was employed. For the concept screening, scoring matrices for evaluating concept alternatives were generated (see Table 2). A reference design was selected as a datum. Selection criteria were established based on design requirements, which included the designers’ intentions. Table 2. Concept-screening matrix . Concept . Selection criteria . DC 1 (reference) . DC 2 . DC 3 . DC 4 . Ease of use 0 − + − Day reliability 0 + + 0 Night reliability 0 − − 0 Cost 0 + + − Ease of application 0 + + − Durability 0 + + − Sum +’s 0 4 5 0 Sum 0’s 6 0 0 2 Sum −’s 0 2 1 4 Net score 0 2 4 −4 Rank 3 2 1 4 Continue? Combine Revise Yes No . Concept . Selection criteria . DC 1 (reference) . DC 2 . DC 3 . DC 4 . Ease of use 0 − + − Day reliability 0 + + 0 Night reliability 0 − − 0 Cost 0 + + − Ease of application 0 + + − Durability 0 + + − Sum +’s 0 4 5 0 Sum 0’s 6 0 0 2 Sum −’s 0 2 1 4 Net score 0 2 4 −4 Rank 3 2 1 4 Continue? Combine Revise Yes No Open in new tab Table 2. Concept-screening matrix . Concept . Selection criteria . DC 1 (reference) . DC 2 . DC 3 . DC 4 . Ease of use 0 − + − Day reliability 0 + + 0 Night reliability 0 − − 0 Cost 0 + + − Ease of application 0 + + − Durability 0 + + − Sum +’s 0 4 5 0 Sum 0’s 6 0 0 2 Sum −’s 0 2 1 4 Net score 0 2 4 −4 Rank 3 2 1 4 Continue? Combine Revise Yes No . Concept . Selection criteria . DC 1 (reference) . DC 2 . DC 3 . DC 4 . Ease of use 0 − + − Day reliability 0 + + 0 Night reliability 0 − − 0 Cost 0 + + − Ease of application 0 + + − Durability 0 + + − Sum +’s 0 4 5 0 Sum 0’s 6 0 0 2 Sum −’s 0 2 1 4 Net score 0 2 4 −4 Rank 3 2 1 4 Continue? Combine Revise Yes No Open in new tab DC 1 was the reference design, so the 0 value (qualitatively) was given to it. This served as the ideal model. DC 2 and DC 4 were awarded "–" for the ease of use criterion, since using LCD characters would not be preferable to using simple LED lights that showed a colour (red/yellow) to signify the condition of the roadway. This preference of LEDs over LCDs was to ensure that all drivers’ concentration was on their driving manoeuvres [1]. Due to the use of PRs in DC 2 and DC 3, their night reliability was reduced compared to the other design concepts. When cost was considered as a criterion, DC 2 and DC 3 were the most favourable options. DC 4 cost more due to the combined cost of the LCDs and piezoelectric accelerometers. Unlike DC 4, DC 2 and DC 3 were the most preferable when the ease of application and durability criteria were considered. This was because of the choice of sensing components employed. That is, PRs would make the building of the lab-scaled prototype easier. For concept selection, the weight of each criterion was generated through team consensus. Design concepts were rated through a comparison to the datum design. As indicated in Table 3, representative ratings were given from 1 to 5, indicating how much worse or better each design was than the reference. Once all the concepts were rated, a total score for each design was calculated using the following formula: $$\begin{equation*} {\bf Total\ Score} = \sum\nolimits_{{\boldsymbol{i}} = {\bf 1}}^{\boldsymbol{n}} {{\boldsymbol{rw}}} \quad \quad [{\bf 2}] \end{equation*}$$ where r is the raw rating of the concept, w is the weighting of the criterion and n is the number of criteria. Table 3. Concept scoring . . Concept . Selection criteria . . DC 1 (reference) . DC 2 . DC 3 . DC 4 . . Weight (%) . Rating . Weighted score . Rating . Weighted score . Rating . Weighted score . Rating . Weighted score . Ease of use 10 3 0.3 1 0.1 4 0.4 2 0.2 Day reliability 15 3 0.45 4 0.6 4 0.6 3 0.45 Night reliability 15 3 0.45 1 0.15 4 0.6 3 0.45 Cost 10 3 0.3 5 0.5 5 0.5 1 0.1 Ease of application 25 3 0.75 5 1.25 4 1 1 0.25 Durability 25 3 0.75 5 1.25 4 1 2 0.5 Total score 3 3.85 4.1 1.95 Rank 3 2 1 4 . . Concept . Selection criteria . . DC 1 (reference) . DC 2 . DC 3 . DC 4 . . Weight (%) . Rating . Weighted score . Rating . Weighted score . Rating . Weighted score . Rating . Weighted score . Ease of use 10 3 0.3 1 0.1 4 0.4 2 0.2 Day reliability 15 3 0.45 4 0.6 4 0.6 3 0.45 Night reliability 15 3 0.45 1 0.15 4 0.6 3 0.45 Cost 10 3 0.3 5 0.5 5 0.5 1 0.1 Ease of application 25 3 0.75 5 1.25 4 1 1 0.25 Durability 25 3 0.75 5 1.25 4 1 2 0.5 Total score 3 3.85 4.1 1.95 Rank 3 2 1 4 Open in new tab Table 3. Concept scoring . . Concept . Selection criteria . . DC 1 (reference) . DC 2 . DC 3 . DC 4 . . Weight (%) . Rating . Weighted score . Rating . Weighted score . Rating . Weighted score . Rating . Weighted score . Ease of use 10 3 0.3 1 0.1 4 0.4 2 0.2 Day reliability 15 3 0.45 4 0.6 4 0.6 3 0.45 Night reliability 15 3 0.45 1 0.15 4 0.6 3 0.45 Cost 10 3 0.3 5 0.5 5 0.5 1 0.1 Ease of application 25 3 0.75 5 1.25 4 1 1 0.25 Durability 25 3 0.75 5 1.25 4 1 2 0.5 Total score 3 3.85 4.1 1.95 Rank 3 2 1 4 . . Concept . Selection criteria . . DC 1 (reference) . DC 2 . DC 3 . DC 4 . . Weight (%) . Rating . Weighted score . Rating . Weighted score . Rating . Weighted score . Rating . Weighted score . Ease of use 10 3 0.3 1 0.1 4 0.4 2 0.2 Day reliability 15 3 0.45 4 0.6 4 0.6 3 0.45 Night reliability 15 3 0.45 1 0.15 4 0.6 3 0.45 Cost 10 3 0.3 5 0.5 5 0.5 1 0.1 Ease of application 25 3 0.75 5 1.25 4 1 1 0.25 Durability 25 3 0.75 5 1.25 4 1 2 0.5 Total score 3 3.85 4.1 1.95 Rank 3 2 1 4 Open in new tab 2.3 Design parameter analysis The prime parameter considered in this project was the stopping distance. It has been observed that, at blind corners, it is the "speeding" level of vehicles that is a cause of the accident [15]. To counter this, the stopping distance for the road vehicle was calculated and used to signal drivers to slow down enough to maintain lane and negotiate the corner successfully. Other parameters considered were the voltage and current requirements of the microcontroller, sensors and displays. The operating specifications for the microcontroller (Table 4), the various sensors (Tables 5 and 6) and the displays were also considered. Table 4. Arduino Uno specifications Description . Values . Microcontroller ATmega328P Operating voltage 5 V DC Input voltage (recommended) 7–12 V DC Input voltage (limit) 6–20 V DC Digital I/O pins 14 (of which 6 provide PWM output) PWM digital I/O pins 6 Analogue input pins 6 DC current per I/O pin 20 mA DC current for 3.3 V pin 50 mA Flash memory 32 KB (ATmega328P) of which 0.5 KB is used by the bootloader SRAM 2 KB (ATmega328P) EEPROM 1 KB (ATmega328P) Clock speed 16 MHz Description . Values . Microcontroller ATmega328P Operating voltage 5 V DC Input voltage (recommended) 7–12 V DC Input voltage (limit) 6–20 V DC Digital I/O pins 14 (of which 6 provide PWM output) PWM digital I/O pins 6 Analogue input pins 6 DC current per I/O pin 20 mA DC current for 3.3 V pin 50 mA Flash memory 32 KB (ATmega328P) of which 0.5 KB is used by the bootloader SRAM 2 KB (ATmega328P) EEPROM 1 KB (ATmega328P) Clock speed 16 MHz Open in new tab Table 4. Arduino Uno specifications Description . Values . Microcontroller ATmega328P Operating voltage 5 V DC Input voltage (recommended) 7–12 V DC Input voltage (limit) 6–20 V DC Digital I/O pins 14 (of which 6 provide PWM output) PWM digital I/O pins 6 Analogue input pins 6 DC current per I/O pin 20 mA DC current for 3.3 V pin 50 mA Flash memory 32 KB (ATmega328P) of which 0.5 KB is used by the bootloader SRAM 2 KB (ATmega328P) EEPROM 1 KB (ATmega328P) Clock speed 16 MHz Description . Values . Microcontroller ATmega328P Operating voltage 5 V DC Input voltage (recommended) 7–12 V DC Input voltage (limit) 6–20 V DC Digital I/O pins 14 (of which 6 provide PWM output) PWM digital I/O pins 6 Analogue input pins 6 DC current per I/O pin 20 mA DC current for 3.3 V pin 50 mA Flash memory 32 KB (ATmega328P) of which 0.5 KB is used by the bootloader SRAM 2 KB (ATmega328P) EEPROM 1 KB (ATmega328P) Clock speed 16 MHz Open in new tab Table 5. HC-SR04 module electrical parameters Description . Values . Working voltage 5 V DC Working current 15 mA Working frequency 40 kHz Maximum range 400 cm Minimum range 2 cm Measuring angle 15° Description . Values . Working voltage 5 V DC Working current 15 mA Working frequency 40 kHz Maximum range 400 cm Minimum range 2 cm Measuring angle 15° Open in new tab Table 5. HC-SR04 module electrical parameters Description . Values . Working voltage 5 V DC Working current 15 mA Working frequency 40 kHz Maximum range 400 cm Minimum range 2 cm Measuring angle 15° Description . Values . Working voltage 5 V DC Working current 15 mA Working frequency 40 kHz Maximum range 400 cm Minimum range 2 cm Measuring angle 15° Open in new tab Table 6. Specifications for light-dependent resistor Description . Values . Operating temperature −30°C to +70°C Resistance (dark) 200 kΩ Resistance (light) 3 kΩ Maximum voltage rating @ 25°C 150 V DC Maximum allowable power dissipation @ 25°C 100 mW Description . Values . Operating temperature −30°C to +70°C Resistance (dark) 200 kΩ Resistance (light) 3 kΩ Maximum voltage rating @ 25°C 150 V DC Maximum allowable power dissipation @ 25°C 100 mW Open in new tab Table 6. Specifications for light-dependent resistor Description . Values . Operating temperature −30°C to +70°C Resistance (dark) 200 kΩ Resistance (light) 3 kΩ Maximum voltage rating @ 25°C 150 V DC Maximum allowable power dissipation @ 25°C 100 mW Description . Values . Operating temperature −30°C to +70°C Resistance (dark) 200 kΩ Resistance (light) 3 kΩ Maximum voltage rating @ 25°C 150 V DC Maximum allowable power dissipation @ 25°C 100 mW Open in new tab With the appropriate equations of motion, the stopping distance for an average family vehicle was generated with the following assumptions based on the findings of the Queensland Government Department of Transport and Main Roads [16]: In an emergency, the average driver takes approximately 1.5 seconds to react. A modern vehicle with good brakes and tyres, after braking, is capable of stopping at approximately 7 ms2. A dry road that is sealed and level enables good friction between the tyres and the road to help stop the vehicle sooner. Scientifically, it has a coefficient of friction of approximately 1. A wet road that is sealed and level has less friction between the tyres and the road, which increases the stopping distance of the vehicle. Scientifically, the coefficient of friction is approximately 0.7. For an average family car speeding at 110 km/h, the stopping distance will be 113 m on a dry road and 143 m on a wet road. 2.4 Final design $$\begin{eqnarray*} {\boldsymbol{S_{ D}}} = {\boldsymbol{T_{ D}}} + {\boldsymbol{B_{ D}}} \end{eqnarray*}$$(1) where SD is stopping distance, TD is thinking/reaction distance and BD is braking distance. $$\begin{eqnarray*} {\boldsymbol{S_{ D}\ }} = {\boldsymbol{\ vt}} + {\boldsymbol{\ }}\frac{{{{\boldsymbol{v}}^2}}}{{2{\boldsymbol{\mu a}}}} \end{eqnarray*}$$(2) where v is initial velocity (m/s), t is reaction time (s), a is deceleration (m/s2) and μ is coefficient of friction. The final design dimensions were generated from the above information. The stopping distance used was 145 m. This was to compensate for any other errors in the calculation, as well as to cater for both dry and wet road conditions. The final design concept as shown in Fig. 7 had the merits of providing good signalling even during poor visibility conditions and vegetation encroachment. Financial and technical constraints to this design were comparatively minor, and hence it was preferable to all other designs for blind-corner signalling and obstacle detection. Fig. 7. Open in new tabDownload slide Final design concept: (a) final design concept; (b) aerial view on position of sensory devices; (c) and (d) blind corner aerial view Fig. 7. Open in new tabDownload slide Final design concept: (a) final design concept; (b) aerial view on position of sensory devices; (c) and (d) blind corner aerial view 2.5 Electrical set-up for the final design The schematic for the electrical set-up (Figs. 8 and 9) was designed using Fritzing Beta (Version 0.9.3, 2016, Fritzing). The design calculations in terms of voltage and current requirements for each component were taken into consideration. Fig. 8. Open in new tabDownload slide Electrical Set Up for Final Design. This is a pictorial representation of the schematic shown in Fig. 9. The components included the HC-SR04, Arduino Uno R3, 9V battery, NPN transistors, LEDs (red and yellow), light dependent resistors, some required resistors and wiring harness Fig. 8. Open in new tabDownload slide Electrical Set Up for Final Design. This is a pictorial representation of the schematic shown in Fig. 9. The components included the HC-SR04, Arduino Uno R3, 9V battery, NPN transistors, LEDs (red and yellow), light dependent resistors, some required resistors and wiring harness Fig. 9. Open in new tabDownload slide Electrical circuit for the final design Fig. 9. Open in new tabDownload slide Electrical circuit for the final design 3. Results and discussion After the lab-scaled prototype (shown in Fig. 10) was built and tested, it was observed that the connections were not properly insulated, especially the soldered joints and points. This was corrected by using rubber glue to fill the points of soldering. The leads that were loosely connected were ruggedly fitted. Fig. 10. Open in new tabDownload slide Pattern marking and architectural model design of the final design concept Fig. 10. Open in new tabDownload slide Pattern marking and architectural model design of the final design concept The ability of the HC-SR04 module to determine the presence of an obstacle on the road model was within the range given in the data sheet (3 mm). The current and voltage applied to it were specified in the data sheet (15 mA and 5 V DC, respectively). When the module was used to detect the dummy car (shown in Fig. 11), there was a flint disturbance in the reading and detection. This was a result of the small effective frontal surface area of the dummy car. To compensate for this, the HC-SR04 was tilted forward to ensure that the ultrasonic waves were able to hit and reflect on the frontal surface area of the dummy car. The HC-SR04 module was able to sense the dummy car and sent signals to the Arduino microcontroller to switch on the red LEDs when the car was 30 cm away from the beginning of the blind corner. The measuring angle of 15° by the HC-SR04 module ensured that a significant portion of the lane was covered in the obstacle detection. This 15° radius also made it necessary to ensure that the effective area in which to carry out the obstacle detection was a minimum of 0.5 m2. Fig. 11. Open in new tabDownload slide Testing the HC-SR04 module Fig. 11. Open in new tabDownload slide Testing the HC-SR04 module The continuity tests on selected components, the wiring harness and the circuitry proved positive. The connection was not broken. Short circuits were also tested for and none was recorded. The DC voltage drop across one LED when calculated was 2 V, but in practice, this value varied by ± 0.2 V. The current passing through one LED when calculated was 20 mA, but this also varied in practice by ± 1 mA. Initially, the LEDs arranged in series did not favour the other LEDs at the end of the chain (that is, the first LED drained the current from the other LEDs). This was because, in a series circuit, the current through each component is the same in the circuit [17]. To counter this, the LEDs were rearranged in parallel and an NPN transistor was used to switch the LEDs on and off. This approach ensured that all the LEDs lit up well, since their current requirements were equally met, and made the LEDs more modular (as one defective LED could easily be replaced in the circuit). The PR indicated that it was within the working range. When the PR was covered with the dummy car, it had a resistance of 200 kΩ, compared to a 3 kΩ resistance when uncovered. The delays in the programming affected the sensitivity of the PR, so the delays were further reduced in the coding to ensure a quick response. This was because the sampling rate of the microcontroller was increased as the delay was reduced. 4. Conclusions The main aim of this project was to develop a multisensory obstacle-detection device that is affordable, easy to implement and easy to maintain to reduce the risk of road accidents at blind corners. The proposed technology has addressed the setbacks of previous solutions in terms of cost of implantation, clear visibility at great distance from the blind corner and "immunity" to vegetation encroachment, as indicated in Table 1. Out of the four design concepts generated, one design concept was chosen through the use of Pugh’s method for concept selection. The final concept was further developed and then manufactured and tested. Design parameters (stopping distance, required voltage and required current) were used in the final design, and a sketch of the final design was made using Autodesk AutoCAD (Version 2007). The electrical connection schematic was made using Fritzing Beta. The microcontroller (Arduino Uno R3) was programmed with a sketch using the Arduino IDE. An environment (architectural model) was built, sensors (HC-SR04 and PR) and displays (red and yellow LEDs) were mounted on it to demonstrate obstacle detection at blind corners. Before the sensors and other electronic components were mounted on the environment (architectural model), reliability tests were done to ensure they met the parameters set. The whole set-up was later tested and changes were made continuously to attain the desired outcome: obstacle detection at blind corners and signalling oncoming traffic about the presence of road obstacles. The initial detection at the blind corner was done in a suitable position before the corner negotiation began. In this project, only two PRs were provided for detection because of the limited number of pins on the microcontroller. If the obstacles at the corner were not detected, the system would be dangerous. In the real world, the PRs will be installed at regular intervals in each lane to ensure that obstacles are continually detected until vehicles or obstacles leave the danger zone. This will make the design more efficient throughout the perimeter of the blind corner. The components employed in the manufacturing of the obstacle-detection device are modular. The device developed in this project serves as a starting point for the development of smart roads (an AHS) in Ghana and the African diaspora. This device is still in its early stages, and more work can be done to make it smarter (enabling it not only to detect obstacles, but also to identify the kind of obstacle). Specifically, the design, manufacture and testing of the obstacle-detection device at blind corners were achieved. Following the successful completion of this project, it is recommended that the following should be considered if this work is to continue. The real-world applications of this design are limited by its use of the same sensors. A tweak in the design to have it use piezoelectric accelerometers (vibration sensors) instead of a PR sensor would allow it to sense not only cars but two-wheelers as well. The PRs should be installed at regular intervals in each lane at the corner to ensure that the obstacles at the corner can be continuously monitored as they passing through and out of the corner. In the real environment, a multi-sourced power supply should be implemented, employing solar-energy technology, battery power and grid energy. In future studies, the microcontroller should be programmed with a more sophisticated algorithm that would enhance the intelligence of the obstacle-detection device. This is because, although the device developed for this project successfully detected road obstacles to some degree, it needs to be smarter in identifying the kind of obstacle present so that logical steps can be taken. In addition, the parameter for this design was the speed of the vehicle to determine the stopping distance. In future studies, other parameters, like the radius of the blind corner or the skidding and overturning velocities of the vehicle, should be taken into consideration. ACKNOWLEDGEMENTS The authors would like to acknowledge the support of Ho Technical University, Ghana. This work was supported by the Ghana Government Book and Research Allowance for tertiary institutions. Conflict of interest statement None declared. References 1. Ackaah W , Adonteng DO. Analysis of fatal road traffic crashes in Ghana . Int J Inj Control SA . 2011 ; 18 : 21 – 7 . Google Scholar Crossref Search ADS WorldCat 2. Teye-Kwadjo E . Risk factors for road transport – related injury among pedestrians in rural Ghana: implications for road safety education . Health Educ J . 2017 ; 76 : 880 – 90 . Google Scholar Crossref Search ADS WorldCat 3. Mends-Brew E , Dadzie J, Dadson BA et al. Modelling the trend of road traffic accidents in Accra . Math Modelling Appl . 2018 ; 3 : 1 – 8 . Google Scholar OpenURL Placeholder Text WorldCat 4. Sokolovskij E . Automobile braking and traction characteristics on the different road surfaces . Transport . 2007 ; 22 : 275 – 8 . Google Scholar Crossref Search ADS WorldCat 5. Zhang J , Wang F-Y, Wang K et al. Data-driven intelligent transportation systems: a survey . IEEE Trans Intell Transp Syst . 2011 ; 12 : 1624 – 39 . Google Scholar Crossref Search ADS WorldCat 6. Kazuki I , Jun S. Image synthesis for blind corners from uncalibrated multiple vehicle cameras . In: 2008 IEEE Intelligent Vehicles Symposium , Eindhoven, Netherlands , 2008 ; 956 – 61 . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC 7. Chelli A , Hamdi R, Alouini M-S. A vehicle-to-infrastructure channel model for blind corner scattering environments . In: IEEE 78th Vehicular Technology Conference (VTC Fall) , Las Vegas, NV, USA , 2013 ; 1 – 6 . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC 8. Elefteriadou L , Martin B, Simmerman T et al. Using microsimulation to evaluate the effects of advanced vehicle technologies on congestion . Technical report . University of Florida Center for Multimodal Solutions for Congestion Mitigation 2011 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 9. Shimizu S , Taniguchi S, Kidena T et al. Multi angle vision system to supplement driver's visual field . Fujitsu Ten Tech J . 2011 ; 36 : 12 – 18 . Google Scholar OpenURL Placeholder Text WorldCat 10. Corbin S , Fischer T, Manente V et al. Decreasing the number and severity of pedestrian accidents by means of new technical systems in the vehicle . 2001 , http://webfiles.ita.chalmers.se/~mys/ActiveSafety05/ProjectReports05/Group4.pdf 21 (20 May 2020, date last accessed) . OpenURL Placeholder Text WorldCat 11. Khairnar VD , Pradhan SN. V2V communication survey wireless technology . arXiv:1403.3993 [cs.NI] , 2014 . OpenURL Placeholder Text WorldCat 12. Chirupphapa P , Jaktheerangkoon S, Nakorn KN et al. Preliminary performance evaluation of IEEE 802.11p in blind corner scenario . In: 13th International Conference on Electrical Engineering/Electronics, Computer, Telecommunications and Information Technology (ECTI-CON) , Chiang Mai, Thailand , 2016 ; 1 – 6 . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC 13. Connolly C . Collision avoidance technology: from parking sensors to unmanned aircraft . Sensor Rev . 2007 ; 27 : 182 – 8 . Google Scholar Crossref Search ADS WorldCat 14. Pugh S . Total Design: Integrated Methods for Successful Product Engineering , Pearson Education , Berkeley, CA , 1991 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 15. Aarts L , Van Schagen I. Driving speed and the risk of road crashes: a review . Accident Anal Prev . 2006 ; 38 : 215 – 24 . Google Scholar Crossref Search ADS WorldCat 16. Queensland Government Department of Transport and Main Roads . Sharing the Road With Bicycle Riders . 2016 . https://www.qld.gov.au/transport/safety/rules/other/cyclists/ (9 November 2020, date last accessed) . 17. Halliday D , Resnick R, Walker J. Fundamentals of Physics Extended 10E , Hoboken, NJ , Wiley , 2013 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC © The Author(s) 2020. Published by Oxford University Press on behalf of Central South University Press. This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact journals.permissions@oup.com
TI - Design, manufacture and testing of a prototype obstacle-detection device at blind corners: an automated highway system technology
JF - Transportation Safety and Open Environment
DO - 10.1093/tse/tdaa029
DA - 2020-12-28
UR - https://www.deepdyve.com/lp/oxford-university-press/design-manufacture-and-testing-of-a-prototype-obstacle-detection-7PFGSzq95u
SP - 1
EP - 1
VL - Advance Article
IS - 
DP - DeepDyve
ER -