Abstract：Aiming at the problem of difficulty to determine the weighting matrix for a conventional linear quadratic regulator(LQR) optimal controller and the slow response caused by this difficulty,a two-wheel self-balancing vehicle with multivariable,strongly coupling,non-linear characteristics is used as a controlled object,puts forward a parameter optimization method of LQR controller based on genetic algorithm.Choosing the linear quadratic performance index as a objective function,the optimal solution of weighting matrix Q is obtained by using global optimization search ability of genetic algorithm.Thus,the state feedback control rate K can be designed,and a dynamical model for the simulation experiment of a system will be built.The results indicate that the GA-LQR controller has a better control effect than a traditional pole placement and an LQR controller,and that the system has more quick response speed and smaller overshoot.

Keyword：two-wheel self-balancing vehicle;linear quadratic regulator;genetic algorithm;optimal control;dynamical model;