On The State Estimation And Control Strategy Of Collision Avoidance System For Electric Vehicle

The serious situation of modern road traffic safety makes automobile consumersand government managers pursue the increasingly high demand for automobile safety.The traditional passive safety systems have been unable to meet the requirements ofmodern transport. Therefore, it is significant to develop the advanced vehicle activesafety system, which is the urgent request for modern transportation. The complexityof automobile structure and the variability of external environment make the studyon vehicle collision avoidance system are becoming more and more complex. Thestudy on vehicle collision avoidance system has been developing rapidly and got someachievements, however, there are still many problems need to be further researchedand discussed. The existing research mostly focused on the research and design oflongitudinal collision avoidance systems, and yet the research and design of thelateral collision avoidance systems are few. A four-in-wheel-motor-drive electricvehicle is taken into account in this paper. The state estimation and the controlstrategy of collision avoidance system are further researched on safety distance model,vehicle dynamics modeling, and control, respectively, to improve the safety of vehiclewith functions of braking and steering collision avoidance. The main researchcontents and the innovation work in this paper are shown as follows:First, longitudinal safety distance model based on adhesion coefficient anddriving intention parameter, and lateral safety distance model based on vehicle edgeturning trajectory are proposed. In order to ensure vehicle safety and improve roadutilization, adhesion coefficient, which can represent the information on road, is usedin longitudinal safety distance model. Besides that, driving intention parameter,which can represent the information on drivers’ driving characteristics, is also used inlongitudinal safety distance model. Therefore, longitudinal safety distance model is proposed with adhesion coefficient and driving intention parameter. It can not onlyadapt to different road conditions, but also determine safety distance according to thedrivers’ driving intentions. The adaptability and the security of longitudinal vehiclesafety system can be improved effectively. Combining with minimum keep distance,lateral safety distance model based on vehicle edge turning trajectory is proposed. Itcan be able to adapt to different road conditions, but also determine safety distanceaccording to the distances between the vehicle edge turning trajectories. It is simplein structure, easy to be implemented. The adaptability and the security of lateralvehicle safety system can be also improved effectively.Second, the simplified method of tire cornering stiffness estimation and thenonlinear observer of vehicle sideslip angle are proposed. Combining the informationon tire lateral force, lateral tire dynamics models can be simplified. The simplifiedlateral tire dynamics models can result in simplifying recursive matrix in recursiveleast squares (RLS) algorithm. The secondary diagonal elements are all zero inrecursive matrix, in other words, the cornering stiffness of front and rear tires can becompletely decoupled to be estimated, respectively. Avoiding the matrix calculationin estimation, the calculation speed of estimator can be improved. Combining withthe information on cornering stiffness of front and rear tire, the nonlinear observer ofvehicle sideslip angle is designed. With the first-order Stirling’s interpolation filterand a low-pass filter, vehicle sideslip angle can be obtained accurately. Furthermore,the estimation process is stable, and the error is small.Third, the stable single input single output (SISO) system model of vehiclelateral dynamics is proposed. Combining with steering stability constraint and tirecornering stiffness information, vehicle lateral dynamic model can be simplified into alinear SISO system. Utilizing mixed sensitivity method, the performance criterion oflateral collision avoidance system can be converted into a standard problem in thesense of-norm. Therefore, H robust controller can be designed to restrain theeffect of uncertainties resulting from parameter perturbation and lateral winddisturbance for lateral collision avoidance system. Thereby, vehicle handling andstability can be improved effectively.Fourth, a braking force distribution strategy based on constrained regenerative braking strength continuity is proposed. The first step is safety brakes rangelinearization. Combining with the nonlinear characteristic of the safety brakes rangeand the distribution characteristics of hydraulic proportional-adjustable valve, thebraking force distribution curve of hydraulic proportional-adjustable valve can beoptimized to approach ideal braking force distribution curve (Curve I). With theoptimization curve of hydraulic proportional-adjustable valve and the tangent line ofminimum rear-wheel braking force (Curve M), safety brakes range can be linearized.It can not only ensure the safety of vehicle braking process, but also simplify theexpresstions of safety brakes range. The second step is the distribution betweenregenerative braking force and frictional braking force. Combining with linearizedsafety brakes range, the expresstions of regenerative braking force and frictionalbraking force can be calculated under different braking strength condition. Byutilizing the characteristic of regenerative braking strength continuity, the allocationparameters in the the expresstions of regenerative braking force and frictionalbraking force can be determined, and the expresstions of regenerative braking forceand frictional braking force can be obtained. This proposed distribution strategy cannot only be suitable for electric vehicle with two-wheel-drive and four-wheel-drivestructure, but also provide a theoretical basis for braking force distribution. It hasgenerality in practice.Finally, the main content of this dissertation is summarized, and further worksare discussed.