Fault Diagnosis For Vehicle ESP System Based On Support Vector Machine

With the rapid development of science technology, in order to satisfy people's traffic safety requirements, many new electronic system have continually emerged, such as ABS, TCS, EPS, and so on. Increasing electronic system, so that traffic safety has been significantly improved, but it also brings new problems. The higher degree of automation of, it is more difficult to ensure the stability of vehicles. Therefore, while people design more stable automotive electronic control system, research on fault diagnosis of automotive electronic control system has become a hot topic in the community. In order to reduce the losses caused by accident, it is hoped that when the electronic control system fails, the driver can get timely warning. ESP a new generation of automotive active safety technology, which was developed from ABS and TCS. It not only had anti-lock braking and driver anti-skid function, but also in times of emergency imposed different braking force on each wheel to modified car oversteer or understeer, thus ensured the safety of car driving. With its superior performance, ESP system will become standard equipment of modern cars. We can say that, from the above analysis, study, Research in ESP system fault diagnosis has undoubtedly of great practical significance and application value.After decades of study, the fault diagnosis technology has made tremendous progress. Fault diagnosis method can be divided into three kinds:quantitative model-based approach, qualitative model-based methods and methods based on data-driven. The traditional model-based diagnostic methods have improved the theory, when have a precise system model, the result of diagnosis is satisfactory. But when the system is very complex,it is very difficult to create a system model. At this time, this results of this method will be poor. Vehicle is a very complex nonlinear dynamic system, inn this case, I used the method of support vector machines which is based on the data to fault diagnose the vehicle ESP system. There are six chapters in this paper.In the chapter one, mainly introduced the research background and significance. The history and present situation of automotive fault diagnosis technology is given. Model-based fault diagnosis method and fault diagnosis method based on the data was introduced, the basis for using data-driven method and the history of support vector machines was given. In the chapter two, the detailed description of the ESP system was given. This chapter begained with the introduce of the composition of ESP. ECU, sensors and actuators,the three parts was described in detail. Described the working process of hydraulic modulator and solenoid valve. Gave detailed description the principle of ESP from two example.In the chapter three,the model of vehicle that bult with AMESim was given, detailed description of software functions and features of AMESim were introduced And model of each part of vehicle were given.In the chapter four, detailed description of the theory of support vector machines was given. Support vector machine had developed based on the statistical theory. In this chapter, the theory of statistical and Support vector machine was introduced.In the chapter five, detailed description of fault diagnosis for vehicle ESP systems based on support vector machine was given. The fault system was built up with software MATLAB. The analysis of the ESP system failure mode was introduced, the specific practices of data collection, data processing, system training, prediction diagnosis, and the analysis of results was descript.In the chapter six, a summary and prospect was given. Introduced conclusion of this paper, and give the prospect of further study.