|
Wheeled mobile robot, the two self-balancing robot is an important biomimetic systems, it has a simple structure, small size, light weight, flexible movement and many other advantages, has great prospects for development in the social and industrial applications, as a means of transport, the two self-balance robot has a broad space for development, has a high commercial value and research value of two from the value for robotics research, this paper features two self-balance robot absolutely unstable, its balance control were studied. Analyze and learn from at home and abroad on the basis of two self-balancing robot, the hardware structure of the object of study and its major parts are described in detail, the working principle of the electrical system, software system and status detection system. System, the two methods of analysis and Lagrange equations of Newtonian mechanics mathematical modeling, and system stability, controllability and observability, and the relative degree of control to do a qualitative analysis. Control method, for the two self-balancing robot equilibrium point linear model, the use of modern control theory classic method of pole placement design state feedback controller for the simulation analysis, the initial inclination and initial displacement control, followed by the least square method identification of state feedback linear controller parameters, that is, first of two self-balancing robot nonlinear model of multi-point linearization, then the set of linear model using the pole placement method to obtain a set of control input and output data, the use of these data by least- state feedback matrix multiplication identification parameters, the least square method state feedback control and pole placement control. Finally, further design intelligent controller: Blur - proportional and differential (Fuzzy-PD) controller design, fuzzy proportional derivative control displacement combination of fuzzy control and inclination proportional and differential control, given an estimate on the basis of linear control fuzzy controller input and output scale factor and input output scale factor change of control effect were analyzed; fuzzy neural controller design, using adaptive fuzzy neural control sample data control using adaptive neuro-fuzzy inference system ANFIS training fuzzy membership function for fuzzy control rules, the training data input and output data from the pole placement of linear controller, fuzzy neural controller design for control of three degrees of freedom model, simulation and control effect analysis.
|