Research On Push Recovery And Upper Limb Fault-tolerant Control Of Humanoid Robot

The walking stability of the robot and the instability of the system caused by joint failure are the key issues that limit the large-scale application of humanoid robots.To address the key technical challenges of humanoid robots,such as the walking instability caused by disturbance and the fault tolerance of the upper limb joint actuator,the corresponding walking control methods and fault tolerance control methods are proposed in this thesis.The main research contents are as follows.1)Aiming at the problems of the traditional gait planning method,such as high gait order and easy to appear oscillation.A gait planning method based on DCM(Divergent Component of Motion)is proposed.This method consists of two steps.The first step involves decoupling the DCM from the inverted pendulum model to plan a smooth and continuous trajectory of the robot’s centroid.The second step employs quintic interpolation to plan the motion path of the swing leg based on the requirements of robot walking.The simulation results show that the trajectory of DCM,centroid and swing leg planned by this method is smooth and monotonous in each time period.2)On the basis of the classical two-dimensional linear inverted pendulum model,a new simplified biped walking model is established by adding the ankle joint model.Building on this,we propose a new online gait adjustment method for humanoid robots that is based on the simplified model.Through decoupling the new model,the DCM expression of the new model is derived.By converting the optimization problem of DCM,walking period,and step length into a quadratic optimization problem,the method generates the optimized landing point and walking time period of the swing leg within the feasible region of gait parameter constraints.3)An ankle joint torque control method based on extended inverted pendulum DCM is proposed.This method abstracts the impact of sudden disturbances on stability during walking as DCM mutations,and designs the ankle joint torque control law that aims to eliminate DCM mutation at the next moment.The main feature of this method is that it has little impact on the original walking posture,and only needs to control the ankle joint to achieve the purpose of disturbed recovery.The simulation results show that this method can effectively eliminate the speed mutation of the robot caused by interference and improve the walking stability.4)A new adaptive sliding mode control method is proposed to compensate the effect of actuator failure on the control performance of robot upper limb joints.The controller ensures the finite-time convergence of the system error in the case of system failure by fusing the integral fast terminal non-singular sliding mode controller and robust adaptive technology.At the same time,in view of the complete failure of the robot actuator,a redundant control scheme is proposed,which can overcome the impact caused by the actuator failure online by designing an adaptive algorithm and improve the fault tolerance performance of the robot upper limb joints.