Trajectory Planning and Control for a Human-like Robot Leg with Coupled Neural-oscillators

The control of legged robots ean be inspired from how biological systems (living creatures) control movements. In this paper, an oscillator-based gait rhythm generator is designed, and a computed torque controller implements the movement of a four-link robot leg. The coupled oscillator network generates trajectories of the hip, knee, and ankle joint angles resulting in a gait similar to that of human. Oscillator parameters and coefficients are tuned to achieve human-like @ajeetories. The desired torque applied to each joint is generated by the computed torque controller according to reference signals tlom the gait rhythm generator and fwdback signals tlom the joints. Oscillator based contxol strategies are implemented and tested in simulation on a PC system. The equation of motion (EOM) of a multi-degree of fi-eedom, four-link robot leg has been derived by Newton-Euler method to validate and evaluate the performance of gait rhythm generator and torque controllers.