Dynamic Modeling And Control Strategy For Flexible Joint Manipulator

The traditional robots dynamic model is hypothetical designed and controlled in rigid body dynamics. However, the accuracy and efficiency of these robots are limited in the real condition. Due to the features of high load-to-weight-ratio, low power consumption and lightweight, flexible joint manipulators have been widely used in some areas, such as:space exploration, high-precision robots and severe environment application. However, joint flexibility will cause vibration, which increases the difficulty to control. A controller was designed in this paper to improve the performance of flexible joint manipulators.Firstly, the kinematics model of six-DOF manipulator was established. Then, the Lagrange dynamics is used in the analysis to construct an integrated dynamic model of six degrees flexible joints manipulators. Secondly, we designed several controllers on the basis of the above dynamic model. A flexible joint manipulator was analyzed with traditional PID control and PID control based on neural networks, and simulated by Simulink. The simulation results illustrate that PID control based on neural network has more advantages than traditional PID control. Using PID control based on neural network has improved respond speed and control accuracy and parameter adjustments easier. It's also reduced tracking error of 50% and achieved good control effect. Then, the control law is created by using the neural network backstepping method in the case of SDOF manipulator with flexible joint. The simulation results show that the control method of model parameter variations are less affected.Finally, in order to verify the control performance of the flexible joint controller, a single degree of freedom flexible manipulator experimental platform was set in preparation for the follow-up study.