Pedestrian Detection System with Edge Computing Integration on Embedded Vehicle

The article proposes pedestrian detection system with edge computing with multi-network integration on embedded vehicle. When camera of lens design in machine learning, the proposal design uses AdaBoost, support vector machine (SVM) and convolutional neural network (CNN). The disadvantage is that a large number of samples are needed for training, and the amount of operation and the large number of parameters cannot be used in the embedded system for vehicles. This article proposes to reduce the amount of computation and the number of parameters required by the network by integrating different optimization operations between networks of different architectures, so as to achieve prediction by using the Renesas R-car H3 of the embedded system on vehicle. The proposed design can maintain above a certain accuracy and cost lower than camera of lens with sensors of radar and lidar.