Robot Control Algorithm And Simulation System

The controlled object of robot is the manipulator having some joints. Its mathematics modeling is very complicated and it is a nonlinear and strong coupled system having close connection with kinetics and the dynamics of robot. Theoretically speaking, though the position and force of robot is controlled rapidly and accurately by some traditional algorithms, it is very difficult to realize. The real time control is executed by high-speed computer in addition to much deducing. In recent years, the ANN appearing again supplies a new path to solve the problems of robot. Because the ANN has the characteristics of fuzziness, fault tolerance, adaptability and self-learning, it is superior to the method of deducing the mathematical modeling and searching the excellent parameters. A lot of scholars have solved the complicated problems of robot with the ANN.This article's task is as follows: Firstly, The traditional impedance control and adaptive impedance control are researched by simulation and compares of their performance under much condition are done. How to adjust impedance parameters is presented preliminarily. It is important to adjust impedance parameters during the force control and it effectively reduces the force between the robot and the environment. Secondly, Based on ANN inverse system method can overcome the difficulties of nonlinearities and strong coupling in multi-degrees of freedom robot system; with characters of simple structure and without knowing the controlled system's precise model, this method can be generalized to a class of MIMO nonlinear system in real field. This article discusses the algorithm: firstly the method of ANN inverse system makes the robot system linear, secondly the decoupling robot system is controlled by control strategy; the new algorithm is simulated and its result is very well. Finally, the neural network inverse system method is implemented on the PA—10 robots and the robot system are uncoupled. The experiment shows that thisalgorithm can make the robot track the trace very well.