|
This thesis addresses the collision avoidance control problem of a 4WSground vehicle. A linear quadratic regulator (LQR) which is a usefulcontrol method for linear systems is used to find the state feedbackgain matrix (K) of linear vehicle systems. Then the values of K computedby LQR set up the search space of genetic algorithm (GA) in which theoptimal state feedback gain of nonlinear vehicle system can be achieved.Because the GA-based control law needs much time to search for the optimalsolution of the optimization problem, it can't be a real time controllerfor the vehicle systems. The feedback gain matrix for some vehicle' svelocities searched by GA generates the continuous feedback gain for adymanic range of the speed between 10 m/s and 50 m/s by the interpolationmethod. At the end of this thesis, a Driver- Preview-Model-based GA controlleris derived for nonlinear vehicle systems to reduce the effects of the lateralacceleration response's delay, and make the controlled vehicle system hasa faster response and more stable performances.
|