Enhancing autonomous exploration for robotics via real time map optimization and improved frontier costs

Area of Science:

• Robotics
• Artificial Intelligence
• Computer Vision

Background:

• Autonomous exploration and mapping are critical in robotics.
• Current methods face challenges with inefficient strategies and incomplete map coverage.
• Ground mobile robots require enhanced exploration performance.

Purpose of the Study:

• To propose an efficient autonomous exploration method for ground mobile robots.
• To improve mapping quality and overall exploration efficiency.
• To address limitations of existing exploration strategies.

Main Methods:

• Utilizing a frontier-based strategy with real-time grid map optimization (bilateral filtering and expansion).
• Developing a novel frontier cost function considering path length, sensor range, and information gain.
• Combining an autonomous exploration decision model with the Minimum Ratio Travelling Salesman Problem (MRTSP).

Main Results:

• Demonstrated 10-30% improvement in exploration efficiency compared to classic methods.
• Enhanced mapping quality through optimized frontiers.
• Reduced inefficiencies in autonomous exploration strategies.

Conclusions:

• The proposed method significantly enhances autonomous exploration and mapping efficiency for ground mobile robots.
• The integration of map optimization and advanced decision-making models is effective.
• This approach offers a more robust solution for navigating and mapping unknown environments.