Gait Planning And Terminal Sliding Mode Stability Of Biped Robot On Stairs

In the field of mobile robotics,due to the characteristics of bipedal robots such as human-like appearance,complex freedoms,and stronger environmental adaptability,they have a wider prospect and potential in the market.Research on the stability of bipedal robots,especially in complex environments,has always been a focus and difficulty in the field.When designing a stable walking scheme for the robot in the scene of climbing stairs,the overall effectiveness of the scheme is mainly determined by three aspects: whether the trajectory planning is appropriate,whether the system modeling is accurate,and whether the controller is reliable.The trajectory planning needs to reasonably design the zero torque point(ZMP),the center of mass(Co G),and the motion trajectory of the leg joints,which is a prerequisite for stable walking of bipedal robots.The system modeling needs to obtain detailed and accurate physical parameters of the robot body,and consider uncertain interference and inevitable modeling errors in the actual motion environment,which is the foundation for the stability of the robot.From the perspective of trajectory planning,the expected motion trajectory of each joint of the robot is obtained,and an efficient and reliable controller is designed to achieve real-time and accurate tracking,which is the focus of achieving stable walking.This article has done the following work in response to the above three issues:1.Unlike walking on a flat surface,the support region of the robot during motion on stairs is discontinuous,resulting in the inapplicability of the ZMP stability criterion.To address this issue,a virtual slope method was improved,where steps are regarded as flat slopes with a fixed slope rate.The virtual ZMP and virtual support region were obtained.At the same time,considering the possible collision problem during the motion of the lower limb,constraints were set to achieve a trajectory planning based on the ZMP stability criterion.2.In response to the inevitable existence of modeling errors and uncertain disturbances in the actual working state of the experimental model,an external disturbance observer was designed to accurately estimate the external disturbance,reducing the requirements for parameter accuracy and eliminating the impact of modeling errors on the system’s robustness and controller control accuracy.This improved the system’s stability and control precision.3.To address the issue of joint trajectory tracking,a non-singular terminal sliding mode control scheme based on a disturbance observer was designed,which avoided singularities and achieved system convergence in a finite time.At the same time,the use of saturation functions instead of sign functions reduced the occurrence of fluttering,and Lyapunov’s stability theorem was used to prove that the system state could converge to within a compact set near zero in a finite time under the control of the sliding mode controller.4.The overall scheme’s feasibility and effectiveness were verified through MATLAB simulation experiments.