The Friction Coefficient Evaluation Between Tire And Road On The Process Of Running

With the volume of motor vehicles increasing, the number of traffic accidents is taking on the increasing trend, and the number of deaths, the economic losses caused by traffic accidents is rising in recent years. According to the investigation analysis showed that the incidence of traffic accidents has a great relation with friction coefficient. The lower friction coefficient, the longer the braking distance, the more easily occur to traffic accidents. The main purpose of this paper is forecast the friction coefficient between tire and road, let the driver can make accurate real-time control of the tire and road surface conditions, ensure the traffic safety.This paper summarized the research status of friction coefficient between tire and the road at home and abroad. Analyzed the influence factors of friction coefficient, i.e. tire, the status of road surface, slippage rate and speed and so on. Analyzed the impact of friction coefficient on traffic safety in different driven conditions. Compared and analyzed several familiar tire models, such as brush tire model, Fialia tire model, Magic Formula tire model, UniTire tire model and neural network tire model. At last, analyzed and evaluated friction coefficient with BP neural network model. Based on the influence factors of friction coefficient to design the experimentation of tire force measurement; Adopted MATLAB software to set up BP neural network model, and took longitudinal force, side force, slippage rate, cornering angle and load as the evaluation guideline. Finally adopted experience data, separated experience data into training samples and test samples, after selected the number of the hidden layer neurons, the training function and all need parameters of training, then programmed the BP neural network training program in MATLAB software, and training by training samples, after that, gained a steady neural network model, used the test samples to simulate in the previous training network. The simulation results are similar to the original data, the error is very small. Also can carry through visual simulation by GUI, eliminating write the process, using them more convenient.