Object/Obstacles detection system for self-driving cars

Object detection systems are vital technology used in Autonomous Vehicle (AV) for operating safely. This paper focuses on reviewing the existing literature related to object detection systems in Autonomous Vehicle. The paper is also suggesting a framework based on the reviewed existing systems for future implementation. The proposed object detection system for AV can detect, classify, and localize each object in its appropriate class by using a set of cameras, radars, Lidar, GPS/INS sensors, Artificial Neural Networks (ANN) and advanced Computer Vision. The object detection for AV is achieved in five steps that include perception, data process, objects behavior and understanding, sensor fusion, and acting. The proposed system is also compared with the current object detection systems for AV for its reliability and robustness. The proposed object detection system is expected to overcome the current object detection systems for AV because of its ability to classify and detect the type and behavior of the objects on the road, allowing the AV to navigate safely in a known and unknown environment by localizing itself even in the most challenging scenarios.