| VSC (Vehicle Stability Control) is designed to prevent vehicle lossof stability during traction or braking at corners. Recently, significantprogress has been made in VSC, however, its further extension andapplication are limited by two major respects: on one hand, more sensorsare required to improve the precision of the controller, which increasesthe cost of the control system; on the other hand, it’s hard for thecontroller to obtain the information of tire-road interaction, as a result ofwhich the control strategy can’t be adjusted when the road conditionvaries suddenly.Model based estimation is a solution for the design of a low costcontroller. Utilizing the existing measurement signals, this method is ableto obtain vehicle states and parameters which are currently hard tomeasure. However, the complexity and the accuracy of the estimationalgorithm depend on those of the vehicle model. Therefore, a full vehiclemodel is established which includes: the tire model, the steering systemmodel, the braking system model and the single wheel model. Moreover,in order to provide the estimation algorithm with a wide choice, threedifferent kinds of tire models are proposed based on the analysis of tirelateral characteristics.Based on the linear tire model, the Kalman filter algorithm is used toestimate four important vehicle states with the help of the measurementsignals of ABS (Anti-lock Braking System) wheel-speed sensors.Meanwhile, the linear relationship between the maximum total aligningmoment of the steering system and tire-road friction coefficient is derivedbased on Fiala tire model. A MAMM (Maximum Aligning MomentMethod) is developed for the estimation of tire-road friction coefficientutilizing only the measurement signals of EPS (Electric Power Steering).Furthermore, a RLS (Recursive Least Squares) algorithm is designed to identify the parameter online. Simulation and experimental results showthat the accuracies of Kalman filter algorithm and MAMM aresatisfactory under the condition of small steering angle and pure side slip;however the results become unsatisfactory with the increase of tire sideslip angles or tire longitudinal forces.Based on the Brush tire model, two nonlinear observers are designedto improve the accuracy of the linear algorithm under two differentarrangements of sensors. The high accuracy nonlinear observer performsvery well but requires two additional sensors to measure yaw rate andlateral acceleration of the vehicle. Therefore, it can be an option for thecontroller equipped in luxury cars. On the contrary, the low cost nonlinearobserver, which performs much better than linear algorithm, only requiresthe measurement signals of ABS and EPS. Furthermore, a combinedmethod (CM) is proposed to improve the accuracy of tire-road frictioncoefficient by taking advantages of both the algorithm of MAMM andnonlinear observer.Finally, an Enhanced Stability Control System (ESCS) is introduced,which borrows the algorithm of envelope control system applied inaircrafts and the current stability control system in production vehicles.With the help of the estimated parameters of the nonlinear observer, thecontroller of ESCS is able to obtain the real-time information of tire-roadinteraction and intervene at a better time. As a low cost and effectivesolution to extend the function of ABS to lateral dynamics in stability, theESCS control strategy is proposed based on only the existing EPS andABS sensors and the similar slip control algorithm in ABS. Meanwhile,utilizing the estimated tire-road friction coefficient and theone-dimensional search-gradient method, the so called “optimal slipratio” is calculated to make the performance of ESCS even better. |
PDF Full Text Request