Graph-based Path Planning and ABC-optimized IT2FLS for Autonomous Mobile Robot Exploration Within Unknown Environments

Abstract

Path planning and autonomous exploration in an unknown environment are primary tasks for realizing the environment perception of mobile robots. Aiming at closed structure scenes, such as large tunnels or pipelines, this paper introduces the graph-based path planning method to implement autonomous explorations of mobile robots within unknown environments. According to candidate vertices and edges composed of vertices, the graph-based path planning method determines a tracking target based on current planning gains. Based on continuous tracking targets, this paper implements an artificial bee colony (ABC)-based interval type-2 fuzzy logic system (IT2FLS) for trajectory tracking. The parameters of IT2FLS are obtained according to ABC optimization instead of subjective rule-based inferences. Linear and angular velocities of the robot are computed according to the difference between the pose of the robot and the current tracking target. This paper adopts a differential robot as a platform and builds a simulation environment for system training and evaluations. Under the overall longest path of 477.530 meters, the maximum root mean square error (RMSE) of 0.186 (m) and the maximum error of 1.507 (m) can be achieved.