Walking Pattern Generation of Biped Robot Using Trajectory Planning of Gravity Center

Zhao Jianghai,Ye Xiaodong,He Feng,Zhao Ziyi
DOI: https://doi.org/10.1109/icma.2014.6885815
2014-01-01
Abstract:A practical method of the gait generation for the biped robot based on the trajectory planning of centre of gravity(COG) is proposed. The control system of the biped robot is designed as the three layers structure: the intelligent decision layer, the motion control layer, and the motor driving layer. The walk process and the action sequence of biped robot are divided into three stages: the starting walk, cyclic walk and stopping walk. in order to improve the real-time performance and the walk stability of the biped robot, The gait data of the walk period is reasonably allocated to the single-stance phase and the double-stance phase, and the final velocity of planned trajectory for the single-support phase and the double-support phase is set to zero. So the landing velocity of the swing leg is efficiently decreased, and the swing leg strikes the ground gently. The three-order polynomial is used to generate the trajectory of the stance-leg and the swing leg, and then the trajectory is discretized with an interpolation period of 10ms. The position and velocity of the discrete point of trajectory is calculated and saved in the hard disk of control computer as the gait data of off line. The test results show that the robot can walk at a steady pace with the single support phase of 6s, the double-support phase of 4s, and the step-length of 400mm.
What problem does this paper attempt to address?