Sensor signal processing and omnidirectional locomotion control of a bio-inspired hexapod robot

Control method based on a central pattern generator (CPG) was proposed to deal with the issue of intelligent locomotion control of bio-inspired hexapod robots. The traditional CPG-based control was unable to control the foot trajectory of robots. For the proposed CPG control method, trajectory generators were added to the control scheme to control foot trajectory. Thus, omnidirectional walking control could be realized by simply adjusting the value of control parameters. Two artificial neural networks for sensor signal processing were developed to overcome complexity in sensory feedback and parameter tuning in CPG control. The neural networks accomplished the fusion of multi-sensory signals, generating the values of the control parameters for robot behavior control. In this way, the robot realized autonomous obstacle-avoiding. A hexapod robot prototype was designed to conduct two real robot experiments in different situations. The experimental results proved that the effectiveness of proposed omnidirectional locomotion control algorithm and obstacle-avoiding algorithm.