Simultaneous Localization and Mapping (SLAM) for Autonomous Driving

Abstract

Simultaneous Localization and Mapping (SLAM) techniques have achieved astonishing evolution over the last few decades and are of growing interest to the autonomous driving community. SLAM has advantages over some traditional vehicle positioning and localization techniques since SLAM can support more reliable and robust localization, planning and controlling to meet some key criteria of autonomous driving. However, there are still some issues that adversely affect the behaviour of the classical SLAM techniques for autonomous driving applications. The fundamental properties of SLAM still need to better understood and appropriate quality analysis methods are required so as to improve the performance of SLAM. This study will review SLAM techniques in the context of autonomous driving. First, we give an overview of the different SLAM techniques and then discuss the possible applications of SLAM for autonomous deriving with respect to different driving scenarios, vehicle system parts and the characteristics of the SLAM techniques. We then focus on some challenging issues and potential solutions for the application of SLAM for autonomous driving. We also summarise some quality analysis algorithms that can be used to evaluate the characteristics and performance of SLAM system. Finally, we conclude with remarks on further challenges and future orientation of research.