Bio-inspired offset array design for enhanced range in underwater active electrosensing with neural network-based localization

Addressing the critical detection range limitation in active electrosensing (AES) for underwater sensing, this study proposes an enhanced AES system via novel array optimization. While AES offers advantages like interference immunity, acoustic stealth detection, and low cost, its short range restricts applicability. A target perturbation model under differential signal acquisition reveals that signal strength increases with local electric field intensity, target size, differential channel spacing, and conductivity contrast, but decreases with target-electrode distance. To extend detection, novel array configurations were explored. Simulations demonstrate that both rectangular and offset arrays significantly outperform the traditional collinear layout. Specifically, an offset array (with 8 m transmitting–receiving spacing) achieved an effective detection range enhancement exceeding 83% under the same distortion threshold while maintaining simplified electrode structure. Experimental validation confirmed a 100% increase in maximum detection distance to 5 m under identical noise thresholds compared to the collinear array. Furthermore, a fully connected neural network-based localization model achieved a mean positioning error of 14.12 cm at 3.15 m in static scenarios. In dynamic scenarios within 1-3 m, mean errors were controlled between 13.19 cm and 27.56 cm. Mechanistic analysis indicates that increasing the array baseline enhances the signal-to-noise ratio by simultaneously suppressing near-field environmental noise and amplifying far-field signal reception. Structural innovations in array design enabled this study to significantly expand the detection range of AES systems without compromising cost efficiency. These advancements directly promote the engineering application of AES technology, offering critical technical support for underwater defense security monitoring, long-range early warning systems, and maritime rights protection. • Novel Array Design for Extended Detection Range: A bio-inspired offset array optimization scheme is proposed, effectively overcoming the limitations of traditional collinear layouts for underwater active electrosensing (AES). Experimental validation confirms this design doubles the maximum detection distance from 2.5 m to 5 m under identical noise thresholds, a 100% improvement over the collinear array. • Quantitative Signal Perturbation Model: This study establishes a theoretical model of target perturbation under differential signal acquisition, derived from 3D quasi-static field equations. This model reveals that signal strength is positively correlated with local electric field intensity, target size, differential channel spacing, and conductivity contrast, while it is negatively correlated with the target-electrode distance. This provides a clear physical basis for the array design. • Enhanced Noise Suppression Mechanism: Mechanical analysis indicates that extending the array baseline synchronously suppresses near-field environmental noise and enhances far-field signal reception. This dual effect optimizes the SNR for long-distance target detection. • High-Precision Neural Network-Based Localization: A fully connected neural network (FCNN) model was developed for target localization. In static scenarios, a mean positioning error of 14.12 cm was achieved at a distance of 3.15 m. In dynamic scenarios within a 1-3 m range, mean errors were consistently controlled within the 13.19 cm–27.56 cm interval.