Design and Research of Occupant Type Feature Recognition Based on BP Neural Network

Abstract

Accurate detection of vehicle occupant classification is one of the prerequisites for the realization of intelligent airbag system. In order to reduce the damage caused to the passenger by accidental airbag in the collision accident, this paper uses the characteristics of the passenger's body pressure distribution on the seat to identify different occupant types. The traditional high-density array pressure sensors can better detect the pressure distribution characteristics of different occupant types on the seat, but the method is costly and is not suitable for commercial applications. In this paper, based on the theory of pressure sensitive points, the flexible array pressure sensors based on pressure sensitive points are used in this hardware and the occupant identification algorithm based on BP neural network is established. The occupant identification algorithm is composed of occupant type recognition algorithm and occupant sitting recognition algorithm. Real-time detection of changes in occupants type and sitting postures is achieved.