Early Detection of Occluded or Hidden Pedestrian by Using Mobile Base Station

The detection and prediction of occluded or hidden pedestrians are challenging for autonomous driving, especially in unusual scenarios such as adverse weather or busy urban areas. We observed that almost every teenager or adult carries a mobile phone when they go out. We propose a method to indirectly detect the occluded or hidden pedestrians by a using mobile base station. So that we can overcome the weakness of the direction detection algorithms. We predict the movement of pedestrians by taking the advantage of the Kalman filter. We also propose an algorithm to minimize the risk of vehicle-pedestrian collision during driving. The experiment and simulation results showed that the hidden pedestrian could be detected 90 meters away or more; by using the Kalman filter, the accuracy of location estimation could be improved by more than 55%.