Path-Follower-2: A Narrow Environment Navigation System for Large Mobile Inspection Robot

Abstract

In industrial settings, mobile inspection robots, often large and heavy, are specifically designed for operation in confined environments. However, their considerable size often limits their ability to perform autonomous global path planning and self-directed obstacle avoidance. Typically, these robots adhere to pre-defined routes, executing high-precision maneuvers, halting for obstacles, and stopping at designated points for data collection. Maneuvering safely and efficiently in these restricted spaces is vital for their detection tasks and largely depends on their strict adherence to the pre-set global path. This paper introduces Path-Follower-2, a narrow environment navigation system for such robots, overcoming the limitations of existing methods that focus mainly on single functionalities like obstacle parking and avoidance. It integrates a novel obstacle descriptors, a global path preprocessing method for early narrow space detection, and adaptive collision detection for precise trajectory tracking in tight spaces. Tested in the Gazebo simulation against static and dynamic obstacles, Path-Follower-2 exhibits exceptional accuracy and smoothness in path tracking within narrow spaces, significantly surpassing most existing solutions.