The Adaptive Cruise Control System Research Based On Neural Network-PID Controller

Adaptive Cruise Control(ACC)is an advanced control system on intelligent vehicles and it has been more and more popular these days.The demand for car safety has also been higher and higher with the development of modern technologies.That makes the operating performance the main focus among the researches of the cruising situation of automatic driving cars.The vehicle's incapability of performing a brake automatically when facing a situation is one of the main factors causing a car accident.In order to ensure the safety of the driver and passengers,an effective control system is needed for the vehicle with ACC.The exsisting conditional PID controller is barely sophisticated enough for the ACC system.So in the paper,the Neural Network is introduced in the PID controller and a fourth layer is added to the BP Neural Network to make a more valid algorithm.This algorithm could control the output and adjust its own weights to perform more accurately.Also,the controller with this algorithm is used in the braking system of the vehicle and the ECU could collect the data about the road condition and predict the brake distance and then control the car to avoid crash.The paper also divides the cruising situation into three categories and discusses the threshold of the crash mode respectively.The Neural Network-PID controller is also used in those modes.At the end of the paper,the CarSim simulating software is used to test whether the improved Neural Network-PID controller could control the vehicle to perform a valid brake automatically and the results show that it is better than the exsisting BP Neural Network/PID controller in terms of the accuracy and efficiency.