Study On Simulation Of Lane Keeping Control Based On BP Neural Network

Lane keeping control is one of the main points in vehicle safety assistant technology study. How can we detect the emergency and take measures in time to rectify the driving direction of the car, and then the car will move back to the current lane before the circumstance of lane departure occurs, so the traffic accident will be avoid eventually.According to the statistical data, the accidents of lane departure on highway make up one-third of the total traffic accidents. People have worked at the study on the method of avoiding the lane departure accidents for years, it can be seen clearly that the unconscious lane departure could be corrected timely and effectively to prevent the occurrence of the traffic accident if the vehicles were furnished with lane keeping system. More and more researchers think highly of the vehicle safety driving assistant technology—the purpose is to improve the active safety of the vehicle. For the moment, the best way to realize the safe and effective driving is to improve the self-safety of the car, also it can laid the foundations for the intelligent traffic.From the point of control algorithm, along with the development of computer technology and the perfection of intelligent control algorithm, the intelligent vehicle control more and more close to intelligentize, each control algorithm is both used in the intelligent vehicle control, and got a good effect.Based on the aboard achievements in this field, this paper mainly do some researches on the lane keeping control. The following issues are discussed: modeling of the dynamic car, analysis of the lane departure of the high speed moving car on the highways, the design of the control model. The following research has been completed:Firstly, the dynamic model of the car is the base of the study of the lane keeping control algorithm. A vehicle model with ninety six degree of freedom was created in this paper, the input of the model is steering wheel angle, and the output is the velocity, yaw angle, and so on. This model is made up of the steering model, tire model, suspension model, engine model, break model, etc. It ensures the accuracy in the simulation and it is suitable for the co-simulation of the lane keeping control.Secondly, this paper studies the math theory of BP neural network, to deal with the defects of the steepest descent in slowly converging and easily immerging in partial minimum frequently, the improved conjugate gradient algorithm is brought forward to solve the problem. This paper analyzes the algorithm deeply in theory, introduces the idea and process. Then the neural network trained by the linear hunting method developed by Fletcher and Reeves algorithm is applied into function approximation. This algorithm improves the convergence of training process and achieves excellent identification effect.Applying the BP neural network in PID control can efficiently overcome the limitations of badness of parameter adjusting and poor performance when the plant has nonlinearity, time-varying uncertainty and difficulty in setting up the accurate model. This paper studies the structure and algorithm of PID controller based on BP neural network, applies the improved conjugate gradient to neural network PID controller. This improved algorithm not only increases the convergence speed in the training process, but also adjusts the PID controller parameters on line, which has rather strong capabilities of adaptive and self-study. So it has better performance.Co-simulation technology is also used to link the control model established in MATLAB to the mechanical model established in ADAMS. Steering wheel angle is the input of the vehicle model, and it is also used as the output of control model. The outputs of vehicle model, velocity, yaw angle, and so on, are used as the inputs of control model. The co-simulation results prove the accuracy and robustness of the lane keeping control system.Finally, a summary is given to the current research work. Although the research in the paper is developed and comes to some achievements, the system is still have some faults and needed for the further development when it is applied into practice and there are much works left to the further research. Although the co-simulation based on virtual prototype was an emerging and challenging technology, the application of co-simulation in vehicle virtual experiments is very wide and innovative from the viewpoint of current research work.