Motion Control Implementation Of Two-wheeled Robot Based On Human-simulated Intelligent Control

Two-Wheel Mobile Robot(TWMR) is a integrated system which involves such disciplines as sensor technology, information processing, electrical engineering, computer engineering, automation control engineering and artificial intelligence, and also is one of the most active research. Therefore, the research of TWMR motion control is always a difficult problem at home and abroad.The research of TWMR motion control often mainly focused on two aspects: point stabilization control and trajectory tracking control. The design for these control methods is usually carried out in the offline situation, which is through the precise controlled object model to design control algorithm, and then use offline optimization algorithm to optimize the parameters of the controller and make the motion trajectory of TWMR get good control effect. But optimization of controller parameters in the offline situation will waste a lot of time, especially applied in physical systems which lead to the change of the controlled object model following the change of external environment and is more trouble for parameters adjustment.Compared with the off-line control method, online controller has good performance that control parameters are dynamically adjustable. It not only can save a lot of time and energy for adjusting the controller parameters, but also can strengthen the adaptive ability of the mobile robot. It has great significance to extend its application field.Through the research of controller parameters dynamically adjustable methods and parameters non-adjustable methods, it put forward a kind of controller that parameters were dynamic adjustable in this paper. The main work was as follows:① Presented a kind of controller that parameters were dynamic adjustable, which was based on Human-Simulated Intelligent control.The controller combines human-simulated intelligent control with neural network algorithm to get a three-layer structure of the controller. The bottom layer is executive layer. It mainly uses multiplex Human-Simulated Intelligent control as basic control unit. The middle layer is that dynamic parameters can adaptive adjust. It adjusts each controller parameter of the bottom layer online through neural network. The top layer is the task coordination. It carries out planning control to each basic control unit according to the state of the system.② Set up a simulation experiment platform and physical experiment platform for TWMR and designed the off-line parameter adjusted controller based on Back-stepping method. And compared the intelligent controller which was proposed in this paper with Back-stepping method.Established simulation platform based on the "equivalent" state space model and carried out the point stabilization experiment and trajectory tracking experiment of the two control methods. Observed control effect in the two cases of with disturbance and without disturbance.Set up the physical experiment based on DC double closed loop, and carried out the point stabilization experiment and trajectory tracking experiment using the TWMR human-simulated intelligent controller which was presented in this paper. The experimental results achieved good control effect.Compared with Back-stepping controller, comparative experiments show that human-simulated intelligent controller which is proposed in this paper has a smaller error, controls more quickly, and has smoother trajectory.