Research On Control Strategy And Algorithm For Light Vehicle Electromechanical Braking System

Along with the improvement of the demand to the vehicle brake safety capability,the EMB(Electromechanical brake system) has been the direction of vehicle brake system and several famous motor companies of the world all agree with that. EMB comes from X-by-wire of the plane industry. EHB(Electrohydraulic Brake) and EMB are called brake-by-wire by a joint name,it is a filiation of X-by-wire. It is considered to be the first step to EMB.Several companies have made the prototype of EMB.According to some information,EMB will be popular in 2015.Researches on EMB are in commencement phase in China.At the same time, how to match up the existing vehicle control algorithm to EMB , such as ABS, TCS, ESP, etc.become one of the issues to resolve for those automobile engineers. Researches on EMB are in commencement phase in China. This paper mainly did some exploratory research and attempted to advance a ABS+EBD control algorithm , the following researches have been completed.First of all, the paper brought forward a ABS+EBD overall control programme based on EMB, Establishing the control logic which use the sliding rate as the control objectives, and giving its controlled circulation flow chart.And then, referencing the relevant literature,it ultimately selected neural network algorithm tuning PI control as ABS control mode,and logic threshold algorithm as EBD control mode. As the conventional BP neural network control algorithm slow convergence, network is easily in a local minima, the learning process often shocks and other shortcomings, this paper advanced a improved BP neural network algorithm by introducing a momentum factor to enhance neural network learning and computing speed, stabilize the learning process, and analysis its advantages through the relevant theoretics.Then, it confirmed various simulation module based on the input and output requirements of ABS + EBD algorithm goal and build corresponding simulation software in Matlab / Simulink,which included: driver intent module, EMB implementation module, load module, brake module, control logic module, vehicles module and tyres module etc.It built a reasonable driver intent module according to EMB braking requirement,to make automobile brake both with the intention to the driver and rapid eliminating the braking space.The paper also brought forward the requirement of brake module's system response according to the characteristics of EMB implementation and achieve its goal by using motor tricyclic control.After that,it started to simulate according to the typical working condition and finally got the best response parameters.In order to fully comprehensive comparing ABS + EBD control with only ABS control ,the paper simulate under related typical working conditions based on the selected vehicle braking I curve. The simulation results show that: in a variety of braking conditions, ABS + EBD control have obviously enhanced over only ABS control at the aspect of driving comfort, braking distance and braking time.At the same time the paper also briefly analyse the"Following attribution"principle of EBD control . On the other hand, in order to test the PI tuning controled ABS braking neural network's adaptability on different road adhesion coefficient, this paper selected three different step situation to observe.Simulation results fully demonstrated the neural network algorithm's self-learning and adaptability.