Active Balance Control of Humanoid Locomotion Based on Foot Position Compensation
- PDF / 2,237,897 Bytes
- 14 Pages / 595.22 x 842 pts (A4) Page_size
- 30 Downloads / 174 Views
Journal of Bionic Engineering https://www.springer.com/journal/42235
Active Balance Control of Humanoid Locomotion Based on Foot Position Compensation Chengju Liu1, Tong Zhang1, Ming Liu2, Qijun Chen1* 1. School of Electronics and Information Engineering, Tongji University, Shanghai 201804, China 2. School of Electronic and Computer Engineering, Hong Kong University of Science and Technology, Hong Kong 999077, China
Abstract A foot positioning compensator is developed in this paper for a full-body humanoid to retrieve its balance during continuous walking. An online Foot Position Compensator (FPC) is designed to improve the robustness of biped walking, which can modify predefined step position and step duration online with sensory feedback. Foot placement parameters are learned by the FPC based on the Policy Gradient Reinforcement Learning (PGRL) method. Moreover, the FPC assists the humanoid robot in rejecting external disturbances and recovering the walking position by re-planning the trajectories of walking pattern and the Center of Mass (CoM). An upper body pose control strategy is also presented to further enhance the performance of humanoid robots to overcome strong external disturbances. The advantages of this proposed method are that it neither requires prior information about the walking terrain conditions, nor relies on range sensor information for surface topology measurement. The effectiveness of the proposed method is verified via Webots simulation and real experiments on a full-body humanoid NAO robot. Keywords: humanoid walking, active balance, Foot Positioning Compensation (FPC), Policy Gradient Reinforcement Learning (PGRL) Copyright © Jilin University 2020.
1 Introduction Legged robots have an anthropomorphic appearance and more flexible mobility compared with wheeled robots. When exposed to unknown external disturbances, such as irregular terrain or pushes, the ability to perform robust and adaptive walking is the most basic and important precondition to fulfill various tasks[1–3]. Biped locomotion is generally achieved by designing the trajectory of the Center of Mass (CoM) so that the actual Zero Moment Point (ZMP) follows the desired one[4,5], which is decided by a pre-defined foot position. However, this planning approach is executed offline typically, and the walking gait can’t be adjusted flexibly. In online implementations, the robot is typically represented by simplified dynamics models, such as Cart-table model and Linear Inverted Pendulum Model (LIPM)[6–11], resulting in larger approximation errors than that of encountered in the actual humanoid robot dynamics. To guarantee the adaptive walking in unknown conditions in the presence of external disturbances, such as pushes or terrain irregularities, a great number of
control strategies have been proposed to regulate the CoM or tracks ZMP reference trajectories through robust feedback schemes[12–17]. Castano et al.[15] proposed a locomotion control strategy using feedback information of the actual CoM state to compensate the foot placem
Data Loading...
