Advanced Robotics

Ma, Shugen; Low, Kin Huat; Wang, Zhidong

doi: 10.1163/156855309X442602pmid: N/A

Kato, Naomi; Shigetomi, Toshihide

doi: 10.1163/156855309X443016pmid: N/A

This paper evaluates methods for the long-range navigation of autonomous underwater vehicles (AUVs) at a middle depth of deep water from the viewpoint of spatial fields-aided navigation. Motivated by experimental works on the geomagnetic field-aided navigation of aquatic animals, we focus on the application of inertia systems with reference to maps of geomagnetic and bathymetric fields of the Earth to the navigation of AUVs. The Inertia Navigation System (INS) in combination with acoustic transponders, the Doppler velocity log, normally used for navigation systems of AUVs near the sea floor, cannot be used for long-range navigation of an AUV at a middle depth of deep water. We first carried out experiments to obtain geomagnetic and bathymetric data in the Harima Sea with a gentle slope of the sea floor and geomagnetic field. At Suruga Bay, which has a steep seabed slope, however, we used a bathymetric map and digital geomagnetic map published by the Japan Coast Guard. We performed a simulation to evaluate the performance of these navigation methods, and found that the precision of navigation of an AUV in water current by the method using the geomagnetic map and bathymetric map without INS is better than by the method using INS alone or by the method using INS with a positional correction by geomagnetic and bathymetric information.

Low, K. H.; Zhou, Chunlin; Zhong, Yu

doi: 10.1163/156855309X443124pmid: N/A

This paper is dedicated to the implementation of biological fish swimming motion onto biomimetic fish robots. By learning from different species of fish, the mechanism design and the motor control of swimming machines could be shaped in different forms. In general, they can be grouped into two major forms, from an engineering viewpoint: serial open-chain design and parallel mechanism design. The gait planning on both forms is then performed based on the well-established theory of fish swimming. By using the associated kinematics equations, the generic solution of the gait planning for multi-link fish robots in the two respective forms of mechanisms is derived. The solution is taken as the gait control input for the swimming testing of a six-link body and/or caudal fin fish prototype and an eight-link media and/or paired fin fish prototype. The experiments show that smooth steady swimming and forward/backward swimming are achievable by making use of the gait planning and control.

Shibuya, Koji; Kawai, Kota

doi: 10.1163/156855309X443007pmid: N/A

This paper proposes a new buoyancy control device inspired by a hypothesis about the buoyancy control mechanism of sperm whales that claims that sperm whales control their buoyancy by melting or coagulating spermaceti oil in their heads to induce volume change. We have previously proposed a buoyancy control mechanism using a syringe and a piston inspired by this hypothesis. Since that system had problems, we built a new buoyancy control device with a mechanism that utilizes the elasticity of rubber to handle the volume change. In the new device, peltier elements and a nichrome wire were installed to heat and cool the paraffin wax and induce volume change. We conducted experiments to assess the buoyancy change and confirmed that the device works well. Then, we built a prototype of an underwater vehicle with four of the devices and confirmed that this robot benefited from 1.3-times higher buoyancy change per unit mass of paraffin wax than with the previous mechanism. Based on the above results, we discuss the feasibility of this method.

Jia, Wenchuan; Chen, Xuedong; Huang, Qingjiu; Sun, Yi; Luo, Xin; Xu, Jiaqiang

doi: 10.1163/156855309X443025pmid: N/A

In this paper, a biomimetic control architecture suitable for multi-legged robots is presented, which is inspired by decision and control behaviors of reptiles. The architecture consists of four functional blocks: information sensing, communication, behavior generation and joint driving. The behavior models are classified into five types according to intelligence level, and in-between two successive layers, a learning and evolution mechanism is designed so that a deliberate decision for a reflex event that is generated in the higher layer can be memorized in the lower layer and be re-invoked by the lower layer directly without rethinking when the same event is triggered. The presented biomimetic control has not only the deliberateness of acquired intelligence, but also the innate flexibility. Furthermore, a physical quadruped robotic embodiment and its virtual prototype are constructed for experimental tests and simulations of the proposed control architecture. The effectiveness of the proposed control method is validated by simulations and experiments.

Ye, Changlong; Ma, Shugen; Li, Bin; Wang, Yuechao

doi: 10.1163/156855309X443061pmid: N/A

A modular universal unit (MUU) is developed for snake-like robots, which has 3 d.o.f. with a series of passive rollers around its cylindrical shell. Among those d.o.f., pitching and yawing are actuated by means of differential gears to accomplish a large ratio of propulsion to mass. The series of passive rollers around the cylindrical aluminum shell of the MUU form another large wheel that can be used as a driving wheel of mobile robots. The snake-like robot composed of those MUUs has more powerful propulsion and higher mobility. By connecting MUUs in different forms, we can also realize a connected mobile platform or a manipulator in addition to a snake-like robot. Owing to it having 3 d.o.f., two or more MUUs can be connected to make up many mobile robots or form a manipulator that exhibits high mobility and agility. Some typical reconfigurable robots composed of these MUUs are analyzed for locomotion control. Finally, the locomotion experiments and simulations are given to show the characteristics of this MUU.

Zhang, Houxiang; Wang, Wei; González-Gómez, Juan; Zhang, Jianwei

doi: 10.1163/156855309X442990pmid: N/A

This paper presents a novel inspired modular climbing caterpillar. First, related issues such as the attachment principles and locomotion kinematics of climbing robots are summarized systematically. Based on the investigation of the movement mechanism of natural caterpillars, we combine climbing techniques with a modular approach to realize a novel prototype as a flexible wall-climbing robotic platform featuring an easy-to-build mechanical structure, a low-frequency vibrating passive attachment principle and various locomotion capabilities. The robot consists of cross-connected modules for moving. There are only two kinds of modules in the system: (i) the head and tail module, and (ii) the body module. Active joints actuated by RC servos endow the connecting modules with the ability to change shapes in two dimensions. After explaining these principles, the discussion focuses on the robot's various locomotion capabilities. Linear movements, turning movements, lateral movements, and rotating and rolling movements are achieved by an inspired control model to produce rhythmic motion. An easy-to-build modular caterpillar is designed and manufactured as an experimental prototype to confirm the feasibility of our design principle and the robot's capabilities. Finally, a conclusion is given and future work is outlined.

Watanabe, Kouki; Iwase, Masami; Hatakeyama, Shoshiro; Maruyama, Takehiko

doi: 10.1163/156855309X443115pmid: N/A

This paper describes a locomotion method of a snake-like robot with passive wheels based on a stabilizing control theory for nonlinear systems. The advantage of this method is that the locomotion can be realized only by a state feedback control law, while many reported methods require some reference trajectory such as the serpenoid curve or are realized not only by a locomotion control, but also by a posture control to avoid singular postures. The proposed method evaluates the friction force of passive wheels in the quadratic-like cost function, so that the efficiency with respect to the input can be improved. The state-dependent Riccati equation technique is utilized to realize this method and it permits us to tune the system performance like an optimal control case. The projection method, one of the nice modeling techniques, is utilized to derive both the plant and the friction model. The effectiveness of the proposed method is verified through numerical simulations and experiments.

Benvenuto, Antonella; Sergi, Fabrizio; Di Pino, Giovanni; Seidl, Tobias; Campolo, Domenico; Accoto, Dino; Guglielmelli, Eugenio

doi: 10.1163/156855309X443034pmid: N/A

Robots for space applications require a level of autonomy while operating in highly unstructured environments that current control architectures cannot manage successfully. We investigated including pre-developed insect brain tissue into the control architecture of space exploratory vehicles. A doubly hybrid controller is proposed that hinges around the 'insect-in-a-cockpit' concept towards an evolution of the classical deliberative/reactive paradigm, featuring a biological (insect brain) high-level deliberative module coupled with low-level reactive behaviors embedded in a robot. The proposed concept, its design methodology and functional description of the submodules are presented, along with a preliminary feasibility assessment mainly derived from an in-depth review of the state of the art.

Campolo, Domenico; Schenato, Luca; Pi, Lijuan; Deng, Xinyan; Guglielmelli, Eugenio

doi: 10.1163/156855309X443052pmid: N/A

In this paper, we address sensor fusion for the attitude estimation of micromechanical aerial vehicles (MAVs), in particular a biologically inspired robotic housefly. First, a dynamic observer is proposed that estimates attitude based on kinematic data available from different and redundant bio-inspired sensors such as halteres, ocelli, gravitometers, magnetic compasses and light polarization compasses. In particular, following a geometric approach, the traditional structure of complementary filters, suitable for multiple sensor fusion, is specialized to the Lie group of rigid-body rotations SO(3) and almost-global asymptotic stability is proved. Then, the filter performance is experimentally tested via a 3-d.o.f. robotic flapper and a custom-made set of inertial/magnetic sensors. Experimental results show good agreement, upon proper tuning of the filter, between the actual kinematics of the robotic flapper and the kinematics reconstructed from the inertial/magnetic sensors via the proposed filter.