Abstract: According to the human-machine coordination of steering control for lane departure assistance system(LDAS), a human-machine shared decision-making and control method is proposed. Accounting for the driver torque, road curvature and longitudinal speed, the virtual strip width is determined using a fuzzy controller and assistance control is based on whether the wheels are beyond the virtual strip. Then, according to the position relation between vehicle and lane, the dynamic model of vehicle lateral deviation at look-ahead point is established. The LPV/H∞ controller is designed to tracking center-line of the lane and determining the desired steering angle. A sliding mode observer with unknown inputs is designed to estimating the road reaction torque. A second-order sliding mode controller which take into account driver torque and road reaction torque is used to generate assist torque. The human-machine control authority is shifted by authority distribution module to achieve shared steering control between assistance system and driver. The proposed assistance system is evaluated in veDYNA/Simulink, simulation results showed that the proposed method can effectively avoid lane departure at different speeds, minimize controller intervention and human–machine conflict. A test platform for LDAS is established, which is based on veDYNA/LabVIEW. The proposed method is experimentally validated with driver and hardware in the loop, the result showed that this method can effectively assist the driver through the complicated lane and good results were obtained using a criterion for human-machine cooperation.

Key words: control, hierarchical control, lane departure assistance system, LPV/H&infin, shared control, sliding mode observer