Design, implementation and comparison ofsensor fusion methods for object detection and trackingbased on multiple 3D lidar sensors

Abstract

The automotive industry is a significant social and financial power around the world. This industry is constantly changing with the technological developments to meet customer needs such as better performance, safety driving etc. Today, many automotive companies are bringing their vehicles to a certain level of autonomy with Advanced Driver Assistant Systems (ADAS) to provide safe and comfortable vehicles. The increasing demand in this area and the investments of automotive companies make the subject of autonomous driving more attractive. In order to increase the autonomy of the vehicle, human intervention from the system should be removed and the vehicle must do everything which the driver does. Thus, researchers and automotive companies focus on completely eliminating human intervention from these autonomous systems. Autonomous systems require a complete perception of the surroundings and understanding the behaviour of objects around the vehicle. Accordingly, it is extremely important to detecting and tracking the objects on the environment in order to decide the motion of the vehicle. The reliable, verified environmental perception is provided by the use of multiple sensors with the correct combination.