Two-Dimensional Frontier-Based Viewpoint Generation for Exploring and Mapping Underwater Environments.

Area of Science:

• Robotics
• Autonomous Systems
• Oceanography

Background:

• Autonomous underwater vehicles (AUVs) require efficient exploration strategies for unknown environments.
• Traditional methods often rely on single-sensor data, necessitating multiple missions for comprehensive coverage.
• Prior information is often unavailable in complex underwater settings.

Purpose of the Study:

• To develop a novel robotic exploration algorithm for AUVs capable of operating without prior environmental information.
• To enable simultaneous data acquisition from multiple sensors for enhanced environmental mapping.
• To reduce the time and cost associated with underwater exploration missions.

Main Methods:

• The algorithm integrates view planning (VP) and frontier-based (FB) strategies.
• It utilizes a labeled quadtree occupancy map for environment representation and viewpoint generation.
• The system processes occupancy data from a profiling sonar and optical data from a camera concurrently.

Main Results:

• The algorithm achieves full environmental coverage using data from two distinct sensors in a single mission.
• It demonstrates computational efficiency suitable for online execution on an AUV's onboard computer.
• Successful sea trials were conducted using the Sparus II AUV, validating the algorithm's performance.

Conclusions:

• The developed algorithm offers an efficient and cost-effective solution for autonomous underwater exploration.
• Simultaneous multi-sensor data fusion enhances the completeness and detail of environmental mapping.
• The approach is robust and adaptable for real-world AUV operations in complex underwater environments.