Incremental Autonomous Exploration of Flooded Underground Mines Using Slice-Based Gap Detection and Cage-Deployed AUV
Year: 2026
Type: conference-paper
Abstract: Autonomous exploration of flooded underground environments such as mines and caves remains a major challenge due to restricted visibility, complex three-dimensional geometry, and limited communication. This paper presents an incremental autonomous exploration framework for the tether-free UX1-Neo underwater robot deployed through a tethered cage system. The approach combines multi-modal perception, three-dimensional occupancy mapping, and a slice-based gap detection algorithm for identifying traversable openings in complex underground structures. Exploration is organized incrementally, allowing the robot to periodically return to a deployment cage for data transfer and mission supervision. The proposed framework was validated through high-fidelity simulation using the Stonefish underwater simulator and hardware-in-the-loop experiments with a real robot operating in a laboratory water tank. Results demonstrate reliable detection of navigable passages, incremental map construction, and robust autonomous exploration in confined underwater environments.
Institution: INESC TEC
Language: English
Subfield: Ocean Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7201848167