Control And Simulation Of Internet Based Teleoperation System

Robot was one of the greatest inventions in last century. It extricates human form heavy physical work and increase efficiency and precision of industrial production. Robots can do the task for people in the environments where are inaccessible for human as well, such as deep ocean, aerospace and nuclear area. Moreover, robots are widely used in military affairs, medical treatment, entertainment and science research. Teleoperator based on telepresence technology is an important research of robot. With the rapid improvements of computer network, the communication of teleoperator system is replaced by Internet. But the random time-delay of Internet will affect the performance of teleoperator system, and even make the system unstable. To solve the problem, firstly, the time-delay of Internet will be analyzed and tested, and then the research on teleoperator system will be discussed. Main researches are below:The characteristic of time-delay were analyzed at first, and the primary factors which affect time-delay were summarized. Based on the analysis, a program for testing the network time-delay was designed and applied in testing the campus network time-delay, and the test result was analyzed. Some suggestions on choice of transmission protocols in teleoperation system were provided.A passive control algorithm based on two-network theory was discussed and analyzed. The teleoperator system can be equivalent as a passive two-network. With this control algorithm, the effect of stability of teleoperator system can be reduced and good operation performance will be achieved. And then the control of Internet with time-varying time delay and data loss problem was discussed. At last the effectiveness of the passive control algorithm was showed through simulation.As the environment of actual teleoperator system is usually uncertain, to take good control on system under uncertain environment, a predictive algorithm of neural network for teleoperator system was studied and analyzed. The model of slave and environment was created through neural network, and the force of slave was predicted, thus slave can follow the action of master without time delay. The effectiveness of this algorithm was show through simulation.