Robot Intelligent Control Strategy

The thesis commences the research by conducting a survey of the current technical status of industrial robot control at the execution level. The research involves the development of theoretical and applicable intelligent control strategies for robot motion control, mainly by virtue of fuzzy control, variable structure control, artificial neural network control and their integrativity techniques, striving for practical application to multi-DOF rigid manipulator trajectory tracking problems.The main achievement of the research is the development of a class of fuzzy variable structure control system which has the advantage of making full use of both the mathematical model of multi-DOF rigid manipulator in terms of differential equations and the behavioral model in terms of natural languages or alike. The system is designed to store the fuzzy rules distributively to strengthen the ability of its fuzzy rule base to resist against damage / the ability of self-recovery; to learn the rule and control parameters selectively / adaptively to strengthen its real-time operativeness.In Chapters 2 through 4, the robot motion control techniques related respectively to fuzzy control, artificial neural network control and variable structure control are investigated sequentially and separately, followed by addressing the alleviation of the amplitude and frequency of chattering as a breakthrough, which always appears in both fuzzy control and variable structure control systems, leading to an overall summarization of the similarities between a typical variable structure control scheme and a class of fuzzy control scheme. These work facilitates the revealing of the fuzzy essence of the boundary layer method frequently employed to alleviate chattering effectively in variable structure control in the second half of Chapter 4. Then along this theoretic line, the subject of integrativity of various intelligent control methodologies discussed independently thereinbefore is fully probed in Chapter 5.Fuzzy control is the kernel in the proposed robot intelligent control scheme.