Research On Trajectory Tracking Control Of Modular Robot System Based On Adaptive Estimation Of Joint Torque

With the progress of science and technology,the rapid development of robot technology,now robots have gradually appeared in our social production and life,and even into people’s families.Robots can free human from heavy manual labor,provide entertainment,learning and company for human beings,in short,robots are playing an increasingly important role in all walks of life.Taking into account the traditional fixed configuration of the robot will not change because of the change of the task,which can only deal with a single repetitive work,when the robot’s workplace and task environment changes,this kind of industrial robot cannot change its configuration,in contrast to the modular robot has the characteristics of reconfigurability,self-repair,self-adaptability,strong environmental adaptability.Modular robot is composed of several basic module units with the same structure,interchangeable mechanical interface and electrical interface,sensor and motion ability.Through the connection between modules,a variety of configurations are formed to adapt to the requirements of work and environment.It has a good application prospect in the fields of deep-sea operation,deep space exploration,emergency rescue,rehabilitation and medical treatment.Because the joint torque value obtained by the modular robot during the execution of tasks will be affected by noise,the estimated joint torque value cannot be directly used in the control of the joint,so it is necessary to use the filtering algorithm to ensure the accuracy of the joint torque,and adopt the appropriate control strategy to ensure the accuracy and stability of the robot system.This thesis makes the following research on modular robot.(1)Estimation method of joint torque of modular robot based on harmonic drive deviceBecause the structure of the robot joint is complicated when the torque sensor is installed at the joint of the modular robot,the flexibility and accuracy of the robot joint module are affected.Therefore,the estimation method of joint torque is a promising technique to obtain joint torque.In this paper,the estimation method of joint torque is adopted,and the joint position information is used to estimate the value of joint torque under the condition of considering the flexibility of flexspline and wave generator,which is of great significance for improving the control accuracy of modular robot.In this paper,the effectiveness of the torque estimation method is verified by several experiments.(2)Establishment of modular robot dynamics model based on joint torque estimationThe harmonic drive device is an important part of the modular robot joint module.According to the kinematic relation and the torque transmission principle,the mathematical model of harmonic drive flexibility was established considering the flexibility of the wave generator and the flexibility of the flexspline.At the same time,a dynamic model of modular robot based on joint torque estimation is established by adopting torque feedback technique.At the same time,the friction term and the coupling term of the robot dynamic model are analyzed.Finally,the system state space description of the modular robot is given.(3)Decentralized robust control of harmonic drive robot based on redundant adaptive robust filter joint torque estimationIn this chapter,considering that the joint torque estimated by the harmonic drive may be affected by the change of joint load,a redundant adaptive robust extended Kalman filter algorithm is designed.In this algorithm,the optimization and robustness of joint torque adapt to load changes by self-adjusting filtering gain and self-switching filtering mode,at the same time,considering the influence of modeling error,redundant factors which can tolerate modeling errors are designed.The proposed method can effectively balance the estimation accuracy and response speed,which make the switching function work stably in a certain redundancy range,and switch between the optimal filtering mode and the robust filtering mode necessary.A decentralized robust control is proposed to compensate the friction and cross-linked coupling terms of the robot to improve the control accuracy of the robot.(4)Impedance control of harmonic drive robot based on adaptive filter joint torque estimationBecause the estimated joint torque of harmonic drive device is affected by noise,the estimated result cannot be directly used in the joint control law,so that the filtering algorithm is needed.Considering that the noise model is not accurate and the filtering error is large,an adaptive Kalman filtering algorithm based on Sage-Husa is designed.The process noise and measurement noise were updated iteratively,and the model parameters and noise characteristics were corrected in real time to reduce the error with the actual situation.At the same time,the deviation will occur when the robot moves in contact with the operating object,resulting in an interaction force,which will have a damaging effect on the object.Therefore,impedance control technology is used to keep the end contact force of the robot within a reasonable range,so that the robot can better track the predetermined curve.In order to reduce chattering in the robot system and improve the tracking accuracy of the robot position,an adaptive filtering impedance control method was proposed to track the robot trajectory.