Mike K. Rao

Obstacle State Estimation For Imminent Crash Prediction & Countermeasure Deployment Decision-Making

This paper describes how predictive crash sensing and deployment control of safety systems require reliable and accurate kinematic information about potential obstacles in the host vehicle environment. The projected trajectories of obstacles in the path of the vehicle assist in activation of safety systems either before, or just after collision, for improved occupant protection. An analysis of filtering and estimation techniques applied to imminent crash conditions are presented in this paper. Optimization of design criteria used to achieve required response performance, and noise minimization, are evaluated based on the safety system to be activated. The predicted target information is applied in the coordinated deployment of injury mitigation safety systems.

Supervisory vehicle impact anticipation and control of safety systems

Occupant safety systems are incorporated in vehicles to meet the requirements of occupant protection. For optimum performance safety devices require tailored activation. This paper presents a supervisory control approach using predictive collision sensor information to augment the performance of safety systems. The supervisory approach determines the potential for a collision to occur, and assists in deployment decision-making. Decision-making is based on the obstacle range, and closing velocity information obtained from the anticipatory sensor, and a reference signal indicative of the host-vehicle deceleration. The multi-input supervisory control system consists of a fuzzy rule-based system, which determines the potential for a collision to occur, and deployment command generation for activation of respective safety devices.