Control System Of Robot Joint Based On Sliding Mode Variable Structure

Robot joint servo drive is the core of the robot control system, it is directly related to the whole movement performance of robot whether the design is reasonable or not. Therefore, a reasonable control strategy, an effective joint controller and driver, and increasing the robot's position accuracy, rapidity and anti-jamming capability have a great significance.A sliding mode is adopted to control single-joint of robot. The simulations show that traditional proportion sliding mode has a problem of chattering, then a sliding mode based on index reaching law and a fuzzy adaptive sliding mode control are proposed to apply to control robot joint. The simulations show that the system not only reduces the chattering, but also has a fast response, a good tracking performance, and a certain anti-jamming capability. The major work and results show as follows:1. Mechanical structure of robot single joint composed of a DC torque and a harmonic reducer is introduced, mechanical properties are simulated to find that the joint system has a rapid response and a good regulation characteristic.2. It is established control model of robot joint based on traditional proportion sliding mode, sliding mode based on index reaching law, and fuzzy adaptive sliding mode variable structure, analyze their advantages and disadvantages through MATLAB simulation. Fuzzy adaptive sliding mode control can adjust itself parameter of sliding mode according to system status. It not only reduces chattering problem of the former two kinds of sliding mode control, but also enhances the system's anti-jamming capability. Compared with PID control, sliding mode control has the advantage of fast tracking and robustness.3. Design a robot joint control system: Hardwares such as drive control of joint, feedback of position angle and CAN communications are designed; Software are divide into upper machine software based on PC and lower machine software based on control panel whose main chip is Dspic30f/4012. Upper machine is developed by LabVIEW, and lower machine is developed by C language, including PWM speed control, reading of position signal, CAN communications, PID and fuzzy discrete sliding mode control algorithm. The control system can achieve function such as the selection of control mode and control strategy, the setup of PID and sliding mode parameters, the sending of target position, and the acquisition, display and storage of postion data.4. Build experimental platform, the static accuracy, dynamic experiment and robust experiment are tested. The results show that the steady-state accuracy of joint controlled by sliding mode can reach up to 0.09°which is the highest resolution of AS5045. Rise time of joint system controlled by sliding mode is 0.1s litter than that of joint controlled by PID, the maximum overshoot is 0.4%, and there is a smooth movement, a good dynamic performance. When joint is stable, 8.6N.m external load is put on joint, angle change of joint controlled by sliding mode is 0.18°litter than that of joint controlled by PID. The robot joint controlled by sliding mode has a high precision, a fast response, a good robustness, which is validated in the experiment.Model of robot joint controlled by sliding mode are established, simulations and experiments show that robot control system has a good static, a good dynamic characteristics and a certain anti-interference ability. Variable structure control combined with the neural network intelligent control can continue to be studied, and further improve the joint performance. The coupling between joints may also be considered, sliding mode control can be applied to control more joints of the robot.