State Estimation And Traction Control For X-by-wire Electric Vehicle

In order to save energy and protect environment, people move the key of vehicleresearch to the BEV (battery electric vehicle). Because of the unique structure of the BEV,control methods become more flexible and more complex control algorithm can be realizedeasily. Therefore, many research designers have been interested in the BEV. Nowadays thecontrol technologies on electric vehicle are developed by innovating and updating traditionalICEV’s control technologies. Whatever active safety technologies are, vehicle state and roadcondition information should be acquired more or less, whereby control logic can be made tocontrol vehicles. Then, how to obtain these information exactly by using existing or low costsensors and not to increase the use of high cost sensors becomes the primary problem need tobe solved for vehicle active safety control system.This dissertation is based on the project named “Research on Control methods andKey Technology for X-By-Wire Vehicle” by National Natural Science Foundation ofChina. Based on the summary of the domestic and foreign research achievements on vehiclestate estimation and road condition identification and traction control, vehicle state and roadconditions would be obtained respectively by using UKF and fuzzy control theory. On thisbasis, traction control system is designed. The main research works are summarized asfollows：First, three degrees of freedom vehicle model using HSRI tire models is built. In orderto accord with X-by-wire characteristics, the vehicle model is four wheel independent steering and driving comparing with traditional vehicles. After this model built, UKF theoryis chosen to estimate vehicle state and Matlab/Simulink is used to build simulation model ofthe estimation algorithm. And then the algorithm is analyzed and verified by co-simulationbetween Matlab/Simulink and CarSim. In order to ensure that simulation can be close toactual situation, the vehicle model in CarSim is modified and noise is merged into the outputsignals from CarSim. Then, some special conditions based on X-by-wire EV are selected.In order to make the tire model obtain more accurate adhesion coefficient and toprovide the optimal slip rates to traction control system, the fuzzy control theory is used toidentify road after summarizing the existing researches on road identification methods, thenmembership functions are chosen and fuzzy logic rules are made. During making rules, it isdifficult to identify the actual road condition when the slide of wheels has not occurred, sothe value of road identification result will become higher after the slip of wheels occurs. Thefuzzy control logic is realized through fuzzy block in the Matlab/Simulink, and theidentification algorithm proposed in this dissertation is similarly verified by using the resultsof the simulation of modified vehicle model running in different conditions in the CarSimsoftware. One of the identification conditions is that the vehicle is accelerating until anywheel slips, then stop accelerating to brake until any wheel locks. Another condition is thevehicle model is accelerated on docking road to test the response of the vehicle.Using speed estimate information and the optimal slip rate information and in order toverify the algorithm in real electric vehicle platform in the future, this dissertation selects thePID control algorithm which is practical for vehicle traction control. This method is based onthe optimal slip rate to realize. In order to ensure that the slip rate of drive wheels will bebetter controlled in the optimal sliding rate, this dissertation proposed a reasonable tractionstrategy including the opportunities of involved and quit. This strategy can adjust the tractioncontrol period through the threshold selection. It is similar to the threshold control, but dueto introduction of PID control module, the effect of traction control is better. In order toreflect more intuitively the advantage of the control algorithm when the algorithm isanalyzed and verified by co-simulation between Matlab/Simulink and CarSim, Vehicle without the traction control is contrasted with vehicle with the traction control. Drivingstraightly on low adhesion road and split road，and moving along slanting trajectory on lowadhesion are used to verify the algorithms.