A new method using knowledge reasoning techniques for improving robot performance in coverage path planning

Abstract

Robots are a rapidly evolving development field encompassing variable domains ranging from industrial robots to empathetic robots for human companions. Future robots will be highly dependent on the ability to understand, interpret, and generate a representation of the environment in which they are operating, ideally in both a human and machine-readable formalism. An important element in this process lies in Path Planning (PP) with obstacle avoidance in dynamic environments (including cleaning and monitoring in robotics) to identify optimal coverage paths. The study in this paper presents a new approach which combines knowledge reasoning techniques with Breadth First Search to find the optimal path for a cleaning robot in a dynamic environment. This approach is used to apply knowledge inference with conventional coverage PP algorithms to enable robot control and avoid obstacles with optimal coverage PP. The experimental results show that using the proposed approach a robot avoids fixed and mobile obstacles, optimal PP reducing both computational cost and time. When compared to other current approaches, the proposed approach with high-coverage rate and low-repetition rate in coverage performs better than the conventional robot algorithms.