LiDAR-Based Obstacle Avoidance With Autonomous Vehicles: A Comprehensive Review

Research on autonomous vehicles (AV) - self-navigating machines that transport both man and cargo has proliferated lately. While once limited to industrial or military uses, more attention is now given to their potential applications in broader society, especially in taking over the mundane, risky and taxing jobs from humans. From self-driving cars, to autonomous mobile robots, to unmanned air, surface and underwater vehicles, one challenge common to all of them is the need to navigate autonomously without human intervention as they operate in their intended environment. The ability to accurately detect and safely avoid obstacles is thus imperative to achieving greater autonomy in vehicles. In recent years, light detection and ranging (LiDAR) sensors - known for their accuracy and reliability in measuring distances- have been widely used for obstacle avoidance. However, as AVs are expected to function under a multitude of conditions, the usage of a single sensor is insufficient. Sensor fusion becomes the next logical step to allow the vehicle to detect and respond to a wider variety of situations. In this paper, we investigate the ways sensor fusion can be applied to improve the obstacle avoidance capability of various indoor and outdoor LiDAR-based AVs by reviewing recent publications over the past decade. The core contribution includes examining the types of secondary sensor used, the motivation behind their selection, as well as the obstacle avoidance algorithm used. Finally, the obstacle avoidance research trends driving indoor and outdoor AVs are discussed and future research directions are presented.