Self-Reconfiguring Modular Robots for Deep-Sea Exploration

Abstract

Self‑reconfiguring modular robots (SRMRs) represent a paradigm shift in underwater robotics, offering unprecedented adaptability, resilience, and efficiency for deep‑sea exploration missions. This manuscript provides an in‑depth analysis of an SRMR system engineered to operate under extreme hydrostatic pressures (up to 150 MPa), subzero temperatures (down to –2 °C), and complex, unstructured seabed terrains. We detail a novel modular architecture comprising hermetically sealed aluminum alloy modules filled with dielectric oil, interconnected via magnetic latches and spring‑loaded electrical contacts. Each module houses a pressure‑tolerant brushless DC actuator and onboard microcontroller, enabling fully distributed decision‑making. A hybrid control framework integrates a centralized mission planner—responsible for global pathing and morphology directives—with module‑level behavior controllers that handle local reconfiguration, collision avoidance, and energy management. High‑fidelity simulations in Gazebo, augmented with hydrodynamic plugins to emulate abyssal currents (up to 1 m/s) and terrain types (sand, silt, and rocky outcrops), validate the system’s capabilities. Key performance metrics include reconfiguration time (12.3 ± 2.1 s between chain and wheel morphologies), locomotion efficiency (0.85 m/J in wheel mode, a 35% improvement over chain mode at 0.63 m/J), fault tolerance (92% waypoint completion under a 10% random module failure rate), and long‑duration uptime (>90% operational over 48 h). Results demonstrate that SRMRs can dynamically optimize their morphology to balance speed, stability, and energy consumption, negotiating obstacles via legged configurations when needed. We discuss the practical implications for seabed mapping, sample collection, and infrastructure inspection, highlighting areas for future work such as real‑world prototype testing, advanced energy storage, and enhanced acoustic communication strategies.