Research On Robot Trajectory Tracking And Compliance Control In Human-robot Physical Interaction

With continuous advancement of technology,the research direction of robotics has gradually shifted from "robotic substitution for humans" towards human-robot interaction.Human-robot physical interaction has gained significant attention from researchers as it is considered the most direct way of interaction between humans and robots.During human-robot interaction,robots need to ensure smooth and compliant movements to adapt to flexibility of humans,thereby enhancing comfort and safety of interaction.To achieve this goal,trajectory tracking control and compliance control have received much attention in human-robot physical interaction control research.Trajectory tracking control can achieve robot’s desired trajectory by controlling joint torque,allowing motion to exhibit desired behavior.Compliance control can adjust the relationship between interaction force and interaction position of robot,achieving compliant adaptation to humans,and enhancing physical interaction performance between humans and robots.Therefore,this thesis has been mainly focus on research of trajectory tracking control and compliance control for robots.The main work and contributions can be summarized in the following three aspects:1.Two-link robot kinematics model and dynamics model are developed.The principles of sliding mode control,optimal control and iterative optimal control have been outlined.The core ideas of impedance control in soft control methods are summarized.The mechanism of optimal impedance control is analyzed.The effects of different impedance parameters on performance of human-computer physical interaction tasks have been investigated.These provide basis for subsequent research in this thesis.2.A sliding mode trajectory tracking control method with fixed time convergence is proposed to solve the problem of unknown upper limit of tracking error convergence time.This method takes advantage of the bounded property of arc-tangent function and combines it with fixed-time stability theory.Firstly,sliding mode surface was designed based on arc-tangent function,and then control law was designed based on the fixedtime stability theory,so that motion time of trajectory tracking error in sliding phase and reaching phase was only related to parameters set.Meanwhile,singularity problem was avoided by designing different control law in singular region.Finally,through the simulation experiment of a two-link robot,it is verified that the proposed method can make robot track desired trajectory within known time limit.3.An optimal impedance human-robot physical interaction control method is proposed to solve the problem that iterative learning-based compliance control method needs to repeat same task many times.The proposed method draws on the mechanism by which iterative optimal control can optimize cost function to determine optimal control input to system without information of system matrix.It adopts a two-loop control structure.In the task-oriented outer loop,an iterative optimal impedance controller is designed to describe the problem of finding optimal impedance parameters as a Linear Quadratic Regulator problem,which uses iterative optimal control to find optimal feedback gain to minimize cost function including tracking error and interaction force.At the same time,a soft auxiliary function is introduced to avoid robot jitter caused by parameter mutation.In the robot-oriented inner loop,nonsingular terminal sliding mode control is used to make actual trajectory of robot track impedance trajectory.Continuous control law is obtained by eliminating chattering by saturation function.Simulation results show that proposed method can obtain optimal impedance parameters by using only interaction information at initial stage of a task,and minimize tracking error and force consumed by operator during task.