Development of a vision-based collision avoidance system

Abstract

Vision based collision avoidance system is vital in the development of autonomous driving system. In recent years, the trend leans toward autonomous driving to alleviate accidents caused by human error and broaden the prospect of driving amongst the handicapped. The introduction of such autonomous vehicles into commercial markets is increasing. However, the more proficient prototypes are relying on expensive sets of sensors. Applications such as video surveillance, vehicle detection make use of continuous image acquisition and processing in real time. This requires high amount of bandwidth data and can incur high cost. This project aims to introduce a more affordable vision based collision system to ease the production of autonomous vehicles.