Research on pedestrian collision test method based on depth learning algorithm

The research on automatic driving test scenarios and collision warning methods is of great significance to the training and test verification of vehicle active safety and accelerated automatic driving algorithm. This paper deeply studies the in-depth learning algorithm, incorporates the common active safety technology in the autonomous driving scene, uses the principle of back propagation to carry out pedestrian collision detection, and analyzes different types of vehicle and pedestrian head collision test results, and verifies the effectiveness of the proposed pedestrian collision simulation algorithm, provides technical reference for the real-time collision warning model and collision avoidance strategy of intelligent connected vehicles.