Research On 6-DOF Stewart Platform's Decentralized Intelligent Control

6 DOF Stewart platform has many degrees of freedom, high rigidity, high precision, modular production can be advantages. With domestic and foreign research further, application range has also been expanding its main application area such as flight simulators, parallel machine, aviation docking space, medical and many other areas, and requires the platform to the movement control system with high accuracy and dynamic response, the defense, which has significant strategic business value and significance. Compared with the actual object, it is controllable, non-destructive, economic, and reliability advantages, so its prospects are very bright. Six degrees of freedom in this thesis is to improve the movement and control of Stewart platform performance, carried out on the Stewart platform kinematics, dynamics and intelligent control in the relevant research.Firstly, six degrees of freedom kinematics of Stewart platform has been studied, using neural network approximation of exact solutions of complex non-linear global search algorithm and the global search characteristics of genetic algorithm to combine the two organic, based on genetic-neural network six degrees of freedom kinematics of Stewart platform method was proposed. Genetic algorithm global search off the original sample, get a good initial vector neural networks, as the first phase of the output. Combining genetic algorithm and neural network advantages, the first stage of learning to return a sub-global optimum, to be used as a neural network to learn the second phase of the initial vector position of positive solutions to improve the learning speed and accuracy.Secondly, use the application of singularity theory to analyze bifurcation characteristics of Stewart platform. Stewart platform built bifurcation equation, according to the static bifurcation singularity theory to determine the body condition of the bifurcation point. Stewart platform based around the initial point of the non-singular space, bifurcation theory is proposed based on Stewart platform workspace calculation. Of the Stewart platform, the singular point and drive bar of the rod length between the corresponding diagrams.Then dynamics of the Stewart platform, in-depth study of Stewart platform using Lagrange method for dynamic modeling, given the full dynamic model, and further given the characteristics of the kinetic model description. On this basis, analysis of the inertia matrix of the block diagonally dominant characteristics of Stewart platform for joint space dynamic model of the coupling, the Stewart platform, six-input six-output system is divided into three two-input two-output subsystem, established subsystem dynamics model, weakening the strong coupling Stewart platform. To further reduce the interference subsystem coupling is proposed based on inverse dynamic compensation control, improved system coupled interference suppression.Finally, support vector machine inverse system control is studied, six-DOF Stewart platform on the decoupling of the three sub-system dynamics model DIDO decentralized control, a combined strategy of intelligent control was proposed, fuzzy control and support vector machine combination. This control strategy is to fully support vector machine approximation ability of nonlinear equations, but also played a support vector machine inverse control and fuzzy control of complementary advantages. Kinematics proposed by the compensation method, and the method and control strategy is applied to reverse Stewart platform DIDO force compensation control subsystem, and the SISO with the traditional PID control are compared to verify the control strategy and the dynamics model is correct.