An Integrated Approach to Goal Selection in MobileRobot Exploration.

Area of Science:

• Robotics
• Artificial Intelligence
• Computer Science

Background:

• Autonomous mobile robots require efficient exploration strategies for unknown environments.
• Existing methods face challenges in optimizing pathfinding and goal generation for complete coverage.

Purpose of the Study:

• To develop an efficient autonomous navigation framework for mobile robot exploration.
• To address the d-Watchman Route Problem using a novel evolutionary algorithm approach.

Main Methods:

• Formulating exploration as the d-Watchman Route Problem with coupled goal generation and pathfinding tasks.
• Solving the Generalized Traveling Salesman Problem variant using an evolutionary algorithm with indirect representation.
• Employing a nearest neighbor constructive procedure for feasible solution generation.

Main Results:

• The proposed method outperforms state-of-the-art exploration techniques by up to 12.5% in low-obstacle environments.
• Performance is slightly lower by 4.5% in office-like environments compared to other methods.
• Successful deployment on real hardware demonstrated practical applicability.

Conclusions:

• The evolutionary algorithm-based framework offers an efficient solution for autonomous robot exploration.
• The indirect representation effectively handles constraints in optimization problems.
• The method shows promise for real-world robotic applications requiring complete environmental mapping.