Research On The Path Planning For Local Obstacle Avoidance And Lateral Motion Control Of Intelligent Vehicle

With the increasing quantity of vehicles, the social problems brought by the automobile are gradually coming out. In our country, the number of casualties caused by traffic accidents is high every year, and the reason for the accidents basically are caused by misoperation of the driver; at the same time urban traffic congestion, energy and environment problems are urgent to be solved. Research on intelligent vehicle is not only one of the effective ways to solve the above problems, but also has a broad space for development and application prospects in the military, industrial and scientific fields.The main content of this thesis take intelligent vehicle as the research object and is to do research on the two key technologies: local path planning and obstacle avoidance lateral movement control is studied based on the automotive system dynamics、swarm intelligence algorithms and control theory. At first the intelligent vehicle dynamics model is established after introducing the key technology research status of intelligent vehicle at home and abroad. The establishment of a two degree of freedom linear model vehicle model and control model equation of state of seven degrees of freedom and Dugoff tire model used in the compositi on.In view of the disadvantages of the traditional local path planning algorithm, such as path inaccessible, easy to fall into local optimum, obstacle avoidance model complex lead to large amount of calculation problems, this paper adopts a local obstacle avoidance path planning method based on bacterial foraging algorithm optimization. In order to avoid the traditional path planning for large-scale optimization calculation and other defects, we adopted continuous spline interpolation for obstacle avoidance model, the routing problem of discrete continuous treatment. In order to verify the advantages of the proposed algorithm, the bacterial foraging algorithm, particle swarm optimization algorithm and genetic algorithm three optimization algorithms are simulated and compared. The test results show that the bacterial foraging method has the characteristics of strong searching ability and fast convergence, can quickly search for the optimal path in the environment.The RBF neural network sliding mode control m ethod is adopted to solve the highly nonlinear and parametric uncertain properties of the intelligent vehicle model as well as the low control accuracy and poor robustness caused by the massive external interference during driving. A new equation of state is derived and a sliding mode controller is designed based on the state equation of the traditional two degree of freedom vehicle control model. By using RBF neural network to model imprecise part has carried on the adaptive compensation, finally through t he lyapunov stability theory deduce the weights of neural network and the stability of the control system is proved.The modeling and Simulation of the controller are researched in the light of the previous design in the Matlab/Simulink environment. The effectiveness and robustness of the algorithm are analyzed in detail. The neural network sliding mode control method can decrease the lateral deviation and yaw velocity deviation between the current location and the expected path rapidly comparing with the single sliding mode controller. It is also proved that the algorithm has a better robustness in many conditions such as different speed, different loads and external disturbance.