Research On Key Fault Information Of Buses Based On CAN-BUS Data And Rough Set

With the increasing number and operating mileage of the capital buses,the high load operation of buses has resulted in frequent failures.After failure,the normal traffic order will be affected.Therefore,it is very necessary to analyze the key fault information of buses and scientifically monitor the health condition of the buses.However,the data of buses operation are characterized by diversity,large capacity and complex form.The traditional data analysis method can not effectively solve the problem.When the bus fails,the parameters state of the CAN-BUS(Chinese name of controller area network bus technology)system will change accordingly.In order to improve the efficiency of buses maintenance,the key information of CAN-BUS system are found out,which can improve the level of public transport operation.Firstly,the data in CAN-BUS system is processed in this paper,,including data cleaning,data import and fault time point judgment.Secondly,analysing a variety of common discretization algorithms and combining with the advantages and disadvantages of each discretization algorithm and the characteristics of the data used in this paper,the discretization algorithm based on the boolean logic algorithm is used to discretize the continuous attribute values of the buses.Thirdly,aiming at the high computational complexity of heuristic search and so on reduction algorithm,an attribute reduction algorithm based on genetic algorithm is proposed to reduce attributes.The algorithm is based on the basic principle of rough set,and sets the dependence between the conditional attribute and the decision attribute as the optimization target,sets the encoding method,and uses the attribute reduction algorithm based on genetic algorithm to filter the condition attributes.Finally,the attribute reduction algorithm based on cellular genetic algorithm is proposed for attribute reduction based on the problem that the calculation results of attribute reduction algorithm based on genetic algorithmresult is unstable and may fall into the local optimum.The designed algorithm is used to reduce the actual buses data in the CAN-BUS system,and the two algorithms are compared from the efficiency,stability,convergence quality and rule matching degree of the algorithm.Through attribute reduction of the algorithm in this paper,the key fault information of the buses are the instrument speed,oil pressure,torque percentage,instantaneous engine speed,coolant temperature.Further research shows that the attribute reduction algorithm based on cellular genetic algorithm is more applicable than the attribute reduction algorithm based on genetic algorithm in the efficiency,stability and the quality of convergence.Although the attribute reduction algorithm based on genetic algorithm is slightly better than the attribute reduction algorithm based on cellular genetic algorithm in the rule matching degree,the attribute reduction algorithm based on cellular genetic algorithm is more capable of mining information within the acceptable range of matching.Finally,the key fault information of buses is displayed in real time.The average value of the key fault information of the buses at different periods is calculated and displayed.Based on the buses month,buses type and route as the query condition,the average value of the key fault information of the buses is calculated and displayed in real time using Java language and Baidu Echarts plug-in in the paper.In the operation of the buses,we can quickly detect and solve the problem of buses failure by monitoring the real-time changes of these parameters.