Electrosensing-Based Three-Dimensional Localization and Obstacle Avoidance for Underwater Robots with a Mobile...

Accurate localization and reliable obstacle avoidance remain challenging for underwater robots in shallow and unstructured environments, where optical sensing can be degraded by light attenuation and turbidity, while acoustic sensing may be affected by multipath propagation and environmental noise. Weakly electric fish exploit externally generated electric-field signals for passive electrosensing and electrocommunication, and actively emit and receive self-generated electric signals to perceive nearby objects in turbid and confined aquatic environments. Inspired by these capabilities, this study proposes a surface-beacon-based electrosensing system that supports passive electrosensing, active electrosensing, and electrocommunication for 3D localization and obstacle avoidance. Using the surface-beacon pose and passive electric-field measurements, a particle filter corrects the motion prior from robot dynamics, inertial measurements, and pressure sensing to estimate the 3D position of the underwater robot. For local perception, a multi-electrode active electrosensing module detects nearby obstacles and infers their spatial distribution. A rule-based 3D avoidance strategy was developed to support adaptive navigation behaviors, including vertical avoidance, obstacle bypassing, and pipe tracking. A key contribution of this study is the frequency-separated implementation of electric-field signals for passive localization, active perception, and surface-underwater communication, thereby reducing mutual interference among these functions. Experimental validation was conducted through two separate water-tank experiments using custom-built robots. The results demonstrate that the proposed methods improve localization accuracy compared with the motion-prior estimate and enable reliable obstacle avoidance in confined underwater scenarios. These results highlight the potential of bio-inspired electrosensing as a complementary near-field sensing modality to conventional optical and acoustic approaches, particularly in confined, visually degraded, or acoustically complex shallow-water environments, and provide a practical foundation for future deployment in real-world marine applications.