Estimation Of Vehicle Side-Slip Angle Under Limit Condition

Feedback control systems for active safety in automotive applications(VehicleStability Control) have entered production cars. Vehicle Stability Control (VSC)is an active safety device, which lets the vehicle run following the driver's intensionto prevent vehicle from losing stability. Now, VSC is the international researchfocus in the vehicle active safety field. With development of control techniques anddecrease of control system cost, VSC will replace ABS to be the standard deviceof modern vehicles.Many Vehicle Stability Control(for instance yaw stabilization systems such asBOSCH's ESP) have in common that the control action depends on informationabout vehicle status, for instance side-slip angle of vehicle. However, the side-slip angle of vehicle is rarely measured directly and must therefore be estimatedfrom other measurements, such as wheel speed, yaw rate, and vehicle accelerationmeasurements. Hence, it is the focus of Vehicle Stability Control designing toresearch the technique of vehicle side-side angle nonlinear estimation.The main goal of this thesis is to develop a nonlinear observer with a cascadedstructure in order to take nonlinear dynamics under tire adhesion limit into account.Stability of the observer is guaranteed in the form of input-to-state stability of theobserver error dynamics.An 8-DOF vehicle model under tire adhesion limit is given. A vehicle modelthat can describe the vehicle non-linear characteristics under limit condition isneeded for researching the estimation technique of vehicle side-side angle. The keypoint of the vehicle model is to build a tire model which can describe the tire'sproperties under tire adhesion limit. Based on Prof.Guo Konghui's tire theory, thispart gives a non-steady non-linear tire model in tire's ISO coordinate, in whichthe tire's longitudinal slip property, the lateral slip property are considered. Based on this tire model, an 8-DOF vehicle model is built for estimating vehicle statusparameters. We introduce the process of building the 8-DOF vehicle model usingMATLAB/Simulink particularly. At last, the vehicle model is verified using theexperimental data witch is get form the vehicle classic conditional test. It provesthat the vehicle model we built can describe the vehicle's dynamics property underlimit condition. It provides us a good vehicle model for estimating vehicle side-slipangle.A nonlinear observer based on a 7-DOF vehicle model are proposed to resolvethe problem of side-slip angle estimation under limit condition when the vehicleturns。The observers are based on a sensor suite that is standard in many newcars, consisting of acceleration and yaw rate measurements in addition to wheelspeed and steering angle measurements. A significant advantage of the proposedapproach over extended kalman filter is the real-time solution of the Riccati di?er-ential equations is avoided, such that the nonlinear observer can be implementedmore e?ciently in a low-cast electronic unit. It is necessary to know the forcesacting on the vehicle generated by friction between the tires and the road whenwe estimate the side-slip angle of vehicle using a nonlinear observer. Many papersmake a illogical assumptions that the longitudinal wheel forces are known whenthey are in the face of this problem, though such information is not always avail-able. A nonlinear tire-road friction model is used to calculate the force generatedby friction, in order to avoid making this assumption. A 7-DOF vehicle modeland a nonlinear tire-road friction model are given in this part. At last, many limitconditions under di?erent tire-road friction coe?cient and di?erent vehicle speedare simulated using 8-DOF vehicle model built in part Two. Then the nonlinearproposed in this part are applied to experimental data from the 8-DOF vehiclemodel. The experimental results prove that the nonlinear observer performs wellwhen it is used to estimate the vehicle side-slip angle.We analyze stability of the nonlinear observer which is proposed in Part Three.The nonlinear observer has a classical cascaded structure, which consists of theobserver for longitudinal velocity and the observer for lateral velocity. In thispart, two sub-observers are found input-to-state stability, hence we can establishinput-to-state stability of the whole nonlinear observer.E?ciency of the nonlinear observer is validated using well-known CarSim soft-ware witch can show how vehicles respond dynamically to inputs from the driver and the immediate environment(road and wind). Vehicle models in CarSim soft-ware are more precise than the 8-DOF vehicle model built in Part Two. So val-idation in this part using CarSim software can prove e?ciency of the nonlinearobserver. The result of validation show that the observer performs well.Deeper research work need to be done since some problems is still remain tobe solved, such as only vehicle turning condition is considered, excluding vehiclebraking condition. Yet turning is contemporary with braking in fact.