Theoretical Research On Vehicle Rollover Mechanism And Active Steering Rollover Prevention Control

Currently,since the number of automobiles increases and the traffic environment becomes more complicated in China,the traffic safety draws many attentions.Compared to common traffic accidents,vehicle rollover always leads to a larger proportion of seriously injured occupants and fatalities,and often brings to a secondary accident,which cause incurable loss to people's lives and properties.Therefore,it has important social significance that to design an effective active vehicle rollover prevention system so as to reduce or even to avoid vehicle rollover.On the basis of the previous researches,in this thesis,the dynamic mechanism of vehicle rollover is systematically analysed,and the robust control theory for vehicle rollover prevention based on active steering is studied.The main research contents are as follows: 1.A general nonlinear dynamic vehicle is firstly modelled.Under reasonable assumptions,a simplified linear model is obtained for the control design.To detect rollover risk,a general load transfer ratio(LTR)model is built as the rollover index.Based on the LTR model,the impact factors of rollover and the corresponding methods to avoid rollover are investigated.The contribution of roll angle to LTR is also studied and for large roll motion vehicles,a novel perspective is proposed to understand the rollover mechanism in the sprung mass dynamic space.2.To estimate the side slip angle of vehicle online and overcome the weak robustness and low convergence speed of commonly used linear observers,nonlinear sliding mode observer is designed based on sliding mode theory.Furthermore,a robust kinematic-dynamic robust observer is proposed which combined with the excellent characteristics of low noise of a dynamic observer and strong robustness of a kinematic observer.3.Active steering rollover prevention of heavy vehicle and common vehicle are separately researched.It is proposed that,for a heavy vehicle,roll motion should be considered since a large gravity centre transition may be produced by the suspension during cornering.On the contrary,for a common small car,since the roll angle is pretty small,the contribution of roll motion to LTR is absolutely ignorable for simplification.Sliding mode control is utilized for the rollover prevention of both types of vehicles.Simulation verifies that compared to the open-loop vehicle,the controlled vehicle can efficiently avoid rollover by limiting its LTR under the chosen threshold value.4.To solve the problem of weak robustness due to the constant speed assumption to simplify the vehicle as a linear parameter invariant(LPI)system in common rollover prevention control,Udwadia-Kalaba servo control theory is introduced where the control based on the fundamental equation of constrained motion(FECM)can deal with linear parameter variant(LPV)systems.The control based on the FECM of the vehicle produces the nominal steering input and generates a reference state command in which the rollover prevention condition has been predesigned.To consider system uncertainties,a generalized sliding mode controller is proposed to guarantee the vehicle states to track the reference ones.Simulation results show that under speed variations,the proposed controller possess better robustness and less energy consumption compared to a traditional sliding mode controller.5.Adaptive robust active control for vehicle rollover prevention is introduced where system parameter uncertainties are discussed.In the constraint design,the LTR contributed by both lateral displacement and yaw rate of the vehicle are artificially distributed by a weighting coefficient.The controller consists of three parts where the nominal control based on the FECM of the vehicle provides the main steering input and the initial condition error is compensated by feedback control.In the adaptive control,a leakage-type adaptation law is designed and utilized to adjust the adaptation parameter to automatically cover the system uncertainties and disturbances.Simulation verifies that the proposed controller has an excellent performance to guarantee the lateral acceleration to converge to the desired value fast with less control input volatility.