Active Safety Measures for Vehicles Involved in Light Vehicle-to-Vehicle Impacts.

Abstract

Road traffic statistics have shown multi-event crashes typically result in higher fatalities and injuries than single-event crashes do, especially when the initial harmful event leads to a loss of vehicle directional control and causes secondary collisions. In this work, the topic of stabilization control for vehicles involved in light vehicle-to-vehicle impacts is addressed. A post-impact stability control (PISC) system is developed to attenuate undesired vehicle motions (spin-out, skid, rollover) induced by the initial impacts, so that subsequent crashes can be avoided or mitigated.

First a vehicle collision model is developed to characterize vehicle motions due to the light impact, which is based on an assumption of substantial changes of kinematic states but minor structural deformations. Colliding vehicles are modeled as rigid bodies with four degrees of freedom, and the influences of tire forces are taken into consideration to improve the prediction accuracy of collision consequences. Then a collision sensing/validation scheme is developed to detect impulsive disturbances and trigger the activation of PISC. The vehicle responses to the impulse are predicted and used to compare with subsequent measurements for collision confirmation.

The stabilization controller, which is derived from the multiple sliding surface control approach, regulates the disturbed vehicle motions via differential braking/active steering. The system effectiveness is verified through CarSim/Simulink simulations for angled rear-ends collisions. When compared with the performance of existing electronic stability control (ESC) systems and four-wheel braking approach, PISC demonstrates improved capability to reject the collision disturbances and to assist the driver to regain control.

For more integrated control of longitudinal/lateral/yaw/roll motions, a hierarchical control architecture for vehicle handling is proposed. It consists of three coordinated stages: the generation of virtual control commands through model predictive control, the generation of actual commands through constrained optimal allocation, and the tracking of wheel slips at the actuator level. This cascade modular design allows for better trade-off among various control objectives and explicit consideration of control input constraints at handling limits.

This proposed active safety feature can be deemed as a functional extension to current ESC systems, and constitutes a complementary module towards a comprehensive vehicle safety system.