Research On Control Of 7-DOF Manipulator Based On ROS

With the development of the economy and society,the demand for human application of robots has become larger and larger.Low-degree-of-freedom robots developed based on traditional robot control platforms are only able for non-contact work tasks with relatively simple operating environments.In the face of current modular development trends and dealing with various complex and unknown operating environments,such robots' limitations are becoming more and more obvious.This paper takes the redundant degree of freedom robot arm researched and developed by our laboratory as the research object.Based on the open source robot operation platform ROS,beyond smoothly and precisely controlling of position,the force and position of the robot arm in the unknown environment are simultaneously controlled.Firstly,the kinematics analysis of the manipulator is carried out.Based on the D-H parameter method,the forward kinematics equation is established and the workspace is solved.The variable step search and the minimum damped square method are combined to solve the inverse degree of the redundant degree of manipulator.This algorithm ensures that the iteration error converges quickly.Based on the constraints of position,velocity and acceleration,the trajectory planning algorithm is designed in the free workspace to achieve smoothly and precisely controlling of position.Secondly,based on the open source robot operating system ROS,the modular joint driver and the human-computer interaction program are developed.The upper and lower machines communicate based on the CANopen protocol,and the control program developed by QT and the URDF model are integrated into the rviz plug-in respectively,so that the manipulator can be controlled according to different operation modes and displayed on the PC.Then,we establish a dynamic model and implement control model,after which we get the conclusions and methods for improve the performance of algorithm by adjusting the parameters.An off-line estimation method of BP neural network environment equivalent stiffness based on joint current is proposed.At the same time,an adaptive algorithm is introduced to adjust the impedance parameters online,and finally a force/position control algorithm that exhibits good adaptability in an uncertain environment is designed.Finally,based on the experimental platform,we designed experiment and analyzed the results.Firstly,the kinematics experiments and the accuracy of the trajectory planning algorithm are verified by kinematics experiments.The controller PID parameter adjustment experiment was designed.The actual output torque and the ideal output torque of the seven joints were compared by design experiments to verify the effectiveness of the inner loop control of the impedance system.Then the BP neural network environment equivalent stiffness data acquisition experiment based on joint current is designed.Finally,through the design force/position tracking comparison experiment,the performance of the algorithm under uncertain environment is improved...