A Study On Financial Escort Vehicles’ Fault Warning Based On Ant Colony Optimization Fuzzy Neural Network

In recent years, China’s financial industry develops steadily. At the same time more and more people realize the significance of the financial security. Therefore, there are some professional financial escort companies being created and it’s forming a kind of new industry. The financial escort vehicles which these companies choose differ from common vehicles, especially in stability and security. And once the malfunction happens, the escort vehicle only must be sent back for repair instead of being repaired by common4S store because of its features such as not easy to disassemble. Supposing that some vehicles are sent back for repair, the corresponding escorting scheduling has to be made adjustments. It’s unfavorable not only in security but also in terms of financial aspect.The engine is the core component of a financial escort vehicle. And it’s also the system that frequent malfunctions occur in. With the development of vehicles’technology, vehicle electrification is increasingly high and functions are stronger and stronger. On the other hand, it also brings more escort vehicles’ fault points. Traditional fault diagnosis expert system that is mostly based on knowledge has some limitation. So it’s necessary to improve traditional method, in order to adapt to the escort vehicles’ requirements.The main research achievements include:(1) From the existing problems at present in the financial escort vehicles’ fault diagnosis, this paper summarizes existing mainstream fault diagnosis equipments and methods at home and abroad.(2) This paper comes up with a new process, dealing with the status signal by a fuzzy way before artificial neural network (ANN) operates, which can also be identified as a new system connected in series by fuzzy system and neural network. The reasons why paper raises a new fault warning method are that fault signal of a financial escort vehicle’s engine has fuzziness and nonlinear. And also because the data sensors collect isn’t in real time. Different sensor has different period for collecting.(3) This paper designs the neural network’s structure in detail, responding the speed accuracy and reliability of escort vehicles’ fault warning. And then paper uses the ant colony optimization (ACO) to optimize the ANN. In order to solve the problem that ANN is easily trapped into local minima value. At the same time ACO helps to improve convergence speed.(4) Because of the complexity of engine’s electronic controlled system, this paper puts forward a new segmentation fault method. An engine’s electronic controlled system is divided into several subsystems and then for example, for the electronic fuel injection system (EFI) case, the paper accomplishes the whole fault warming system’s structure based on ant colony optimization fuzzy neural network. It primarily includes fault information collections, design of BP network and the optimization process of ACO.The results can provide basis for further blending system development. And both the models and optimized algorithm in this system can be referenced to other systems.