Pedestrian safety system with crash prediction

Abstract

Every year nearly 1.5 million people are dying in traffic collisions around the world, due to the unexpected behavior of pedestrians while crossing the road. To address this problem an augmentation function for predicting the crash risk of the active pedestrian is proposed. The augmentation function has several functions like pre-crash scenario, vehicle trajectory, and pedestrian trajectory. In a Pre-Crash scenario, pedestrian movement such as entering the road boundary or not is detected. The input comes from the sensor which is located at the head of the car. After the pre-crash scenario vehicle trajectory is used to control the speed of the vehicle. Then the Markov IRW model-based pedestrian trajectory finding is used to predict the state of each pedestrian. The states are categorized into three types: Running, Walking and Standing. In this model, the pedestrian types whether the pedestrian is a child or young or old aged people is predicted. And the crash risk is evaluated based on the Monte Carlo algorithm that calculates the minimum detection range for active pedestrians. If the crash risk is exceeded the threshold limit then an augmentation function is activating the evasive action.