Optimization And Implementation Of Ee Architecture Of L4 Smart Electric Vehicle Based On Pareto

In recent years,with the development of driverless cars,the four electrifications of automobiles,namely electrification,intelligence,networking,and sharing,have gradually become the future development direction of the automotive industry.Automotive electronic and electrical architecture must support the future automotive "four modernizations" and vehicle With the rapid development of software,the basic requirements that must be met for this include high computing performance,high communication bandwidth,high functional security,high information security,and continuous software updates.However,the current electronic and electrical architecture has the problems of improving computing performance,increasing communication bandwidth,and gradually upgrading software,which cannot meet the needs of future smart car development.The development of smart electric vehicles has further increased the requirements for on-board electronic and electrical architectures,and more reasonable and complete electronic and electrical architectures are needed to support smart electric vehicles.Therefore,it is necessary to further improve and optimize the EE architecture of the current traditional automobile.In this paper,the principle of Pareto multi-objective optimization is firstly studied,and the fast and elitist non-dominated sorting genetic algorithm(NSGA-?)of elite control strategy is applied to the solution of the optimization of the later architecture model.Then analyze the electronic and electrical architecture network topology diagrams of the three bidding models BMW 7 Series,Audi A8 and Tesla,summarize the controllers,driver assistance system controllers,bus protocols of the three models and combine the L4 level standards to determine the next generation of intelligence The functional requirements of electric vehicles,and then worked out an architecture solution,which includes a gateway solution,network topology,power distribution solution,and ground point design.Then,in the PREEvision,the EE architecture model based on the functional domain and the central gateway is built and evaluated.PREEvision is a model-based secondary development tool for automotive electronic and electrical architecture,which covers the entire stage of V-model development.The software can model the electrical and electronic architecture at different levels,and each layer is mapped to each other.The model building mainly includes a requirements analysis layer,a logical architecture layer,a component network layer,a communication layer,an electrical principle layer,a wiring harness layer,and a topology layer.The built-in model uses JAVA programming to implement an evaluation algorithm for bus weight,cost,and load.Rate.Finally,the NSGA-? algorithm was used to optimize the model's architecture bus cost,busweight,and bus load rate.Based on the same bus cost or bus weight,the bus load rate can be reduced from 35.67% to 27.85%.Use a certain model of Dongfeng Liuqi for verification,make changes on the original model,add sensors such as high-definition camera,lidar,millimeter wave radar,GPS,night vision camera,and change some of the original CAN communication to Ethernet communication.Newly added Ethernet and LVDS communication.Under the condition that the weight and cost of the architecture bus similar to the previous evaluation and optimization are met,the VN5640 is used to collect bus data and analyze the architecture bus load rate.After 12 tests,the architecture bus load rate is between 24% and 28%.In line with the results of the optimization,the load rate is reduced,laying the foundation for the realization of advanced unmanned driving.