Collision prediction system for external airbag using an interacting multiple model estimation with multirate Kalman filter

Abstract

External airbag installed in vehicle can protect pedestrians and reduce injuries in the event of a crash. In this paper, a collision prediction system for external airbag using an interacting multiple model is proposed to detect a collision and inflates quickly enough to cushion the impact. The system includes two different models, constant velocity and constant acceleration, in order to estimate a distance between ego vehicle and front object. The integration of two models gives optimized object state estimation with low cost laser scanner and reduces the root mean squares error.