Research On Vehicle Occupant Type Recognition Algorithm Based On Millimeter Wave Rada

Improper deployment of automobile airbags is likely to cause secondary injuries to small occupants such as children.The accurate identification of automobile occupant types is of great significance to the research of smart airbags.Aiming at the problems of low accuracy and poor stability of traditional car occupant type recognition methods,this paper has carried out research on car occupant type recognition algorithms based on millimeter wave radar.This article introduces the basic theory of occupant type recognition based on millimeter wave radar,focusing on the composition of FMCW millimeter wave radar system,the estimation method of target distance,speed,angle parameter and the principle of constant false alarm detection algorithm,and establishes the occupant type based on AWR1642 radar module.Test experiment platform and environment.Researched the occupant type recognition algorithm based on machine learning.On the basis of the FMCW millimeter wave radar occupant echo signal processing,the multi-dimensional parameter characteristics are extracted;the occupant type recognition algorithm using the SVM model is designed,and the performance of the algorithm is evaluated through experiments.The results show that it is based on Cubic SVM The model’s occupant type recognition algorithm has an accuracy rate of 92.9%.Researched the occupant type recognition algorithm based on Faster R-CNN.On the basis of processing the echo signal of the occupants in the FMCW millimeter wave radar,the distance-angle heat map is extracted to construct the data set;the occupant type recognition algorithm based on Faster R-CNN is designed,and the performance of the algorithm is evaluated through experiments;At the same time,the occupant type recognition algorithm based on machine learning is compared,and the results show that different from the manual extraction of features in machine learning,the occupant type recognition algorithm based on Faster R-CNN that automatically extracts features has a higher accuracy rate of 98.0%.The machine-learning occupant type recognition algorithm has a faster recognition speed.Finally,the pros and cons of the two algorithms are analyzed.