Vehicle Engine Calibration Based On CAN Communication Software Research And Parameter Optimization

With the aggravation of automobile pollution,the increase of automobile ownership and the rapid development of electronic control technology,the calibration software based on automobile communication network becomes increasingly important.However,with the rapid development of electronic technology and artificial intelligence in China,the research on artificial neural network algorithm provides an opportunity to bridge the gap between domestic and foreign level.The artificial neural network learning nonlinear fitting ability and generalization ability provide a guarantee for the prediction of calibration data,and improve the prediction accuracy and accuracy under the condition of reducing the calibration workload,which is of great significance for the development of automobile and engine industry.Based on the analysis of vehicle CAN bus and CCP protocol based on working principle,based on the mechanism of automatic code generation and Simulink underlying driven design completed CAN drive development,implement integration of automotive engine calibration software CCP protocol,complete software and ECU calibration function,this paper proposes a calibration parameter optimization algorithm based on BP neural network.(1)Determined the overall architecture and technical route of the calibration software design by analyzing the functional requirements of the calibration software.Using CAN bus communication and CCP calibration protocol to design a set of small calibration system software,divided into lower computer(ECU end)calibration program and upper computer(PC end)calibration software,respectively responsible for ECU control and calibration process control.(2)Realized ECU communication support of calibration software by customizing the integration of CAN underlying driver file and CCP protocol based on MCU MPC5634.C-MEX S function file,TLC file and C-MEX S function are designed respectively to send and receive CAN,and the module encapsulation function is realized.Design interface files of CAN driver and CCP protocol driver and integrate them into codewarrior project file package to realize CCP protocol driver.(3)Designed the communication control module,A2 L file management module and BP neural network optimization algorithm module of the upper computer of the calibration software in the VB6.0 development environment.Prepare controlcan-dll file call module and interface function call module to complete CAN communication,define CCP protocol command function to realize upper CCP protocol driver;Preparation of A2 L file key field reading function,the realization of A2 L file parsing;A neural network is constructed to determine its network topology and learning algorithm.(4)Realized the function verification of the lower computer calibration program and the upper computer calibration software.Simulink CAN drive test model was built to verify the functions of lower computer CAN drive module.Through off-line calibration,CCP protocol function is verified by taking ECU of a general gasoline engine as the research object.The curve fitting ability and prediction ability of BP neural network are verified by selecting the MAP data of injection of part of the gasoline engine for optimization algorithm training.The test results show that the calibration software works stably,the communication data is complete and the operation is reliable.The optimization algorithm has reasonable design and excellent performance.