원문정보
초록
영어
In this paper, a parameters self-learning PID controller algorithm based on modified BP neural network is proposed to eliminate the influence of time delay on the stability and maneuverability of tele-operation manipulators. This control algorithm adjusts the three parameters of PID controller on line through BP neural network. Conjugate gradient method is used for real-time adjustment of weighted coefficient of BP neural network so as to adjust the output parameter of PID controller. The model of three-joint manipulator with three degrees of freedom (3-DOF) was established. The simulation results show that force tracking performance of master and slave manipulators is good, the maximum error is 0.15. The position tracking performance of slave manipulator is stable, the amplitude decay can be ignored, the maximum error is 3.9 and time delay is 0.3s. This control algorithm has fine self-learning capability and robustness. It had better time delay control effect and could improve the operability of internet-based tele-operation manipulators.
목차
1. Introduction
2. Bilateral Servo Control Architecture with Force Deviation Feedback
3. SimMechanics Model of 3-DOF Manipulator
3.1. Set the Parameters of Environmental Module
3.2. Set the Parameters of Revolute Module
3.3. Set the Parameters of Three Body Modules
3.4. Set the Parameters of Joint Actuator Modules
3.5. Set the Parameters of Joint Sensor Modules
3.6. Set the Parameters of XYGraph Modules
3.7. Run the Model and Visualization
4. Setup the Time Delay Simulation Platform Based On TrueTime Software
5. The Design of Controller
5.1. The Structure of BP Neural Network
5.2. Modified BP Learning Algorithm
5.3. The Structure and Algorithm of Controller
6. The Simulation Results and Analysis
7. Conclusions
Acknowledgements
References