Research On Vehicle Yaw Stability Control Approach Based On Uncertainty Of Drivers

With the increasing demand of vehicle safety performance,as well as the more and more complex traffic situation,automobile-manufacturers show growing interest on vehicle active control.As one of the major project in the active safety research filed,Uncertainties are not taken into account at the control design stage in traditional vehicle yaw stability control approach,which has weak robustness under some extreme conditions.Aiming at the uncertainties,such as driver behavior,vehicle yaw stability control approach applying robust model predictive control(RMPC)is proposed based on set invariance theory in this paper.In order to reduce complexity of control design and computational burden,a piecewiseaffine(PWA)hybrid vehicle model,whose state and input variables are constrained,is derived from nonlinear dynamical model of the vehicle with two degrees of freedom reflecting yaw and lateral motions through approximate linearization method.The driver's steering input is treated as a bounded disturbance in the model,whose uncertainty is modeled as a set-valued function of the vehicle states.Model mismatch is equivalent to uncertainty of input commands.Driver-in-the-loop vehicle model containing uncertainty is built as predictive model used for designing controller.According to set invariance theory and robust reachability framework,computing algorithm of robust invariant set is applied to reachability analysis of the constrained driverin-the-loop uncertain vehicle model.Firstly,the maximal robust invariant set of linear mode is computed as terminal set,and constraints of the original nominal system are tightened while bounding the error between the nominal and the uncertain system states.Meanwhile,three-step robust controllable sets is also computed to analyze controllability.And then,the control objective is put forward to keep cornering force of front and rear tires in linear region for all admissible disturbances.A RMPC problem of vehicle yaw stability,with reference tracking,is formulated to provide front corrections and yaw moment commands through solving constrained optimal control problems at each sampling time.And yaw moment is converted to four individual breaking torques using the algorithm of torque distribution.Finally,through simulations combining Simulink and CarSim on low fraction road,effectiveness of the controller and feasibility of the design approach are demonstrated.