2026 International Conference on Sensing, Measurement & Data Analytics in the era of Artificial Intelligence (ICSMD 2026)

During subsea cable magnetic localization, magnetic measurements are susceptible to environmental interference and sensor attitude variations, making cable position inversion sensitive to measurement errors and unstable across consecutive observations. To address these issues, this paper proposes an Electromagnetic Localization Algorithm-guided Grey Wolf Optimization method for subsea cable magnetic localization, integrating frequency domain magnetic feature extraction with sequential estimation. First, attitude correction is applied to the magnetic field data acquired by the dual triaxial magnetometer array, and the magnetic features at the target frequency are extracted using band pass filtering and a sliding window fast Fourier transform (FFT). The analytical estimate obtained by the electromagnetic localization algorithm (EMLA) is then introduced as a prior to guide the Grey Wolf Optimizer (GWO) in estimating the lateral offset and vertical distance of the subsea cable, while an unscented Kalman filter (UKF) is further introduced to exploit the temporal information of consecutive observations for sequential localization. Experimental results show that the median absolute error (MedAE) and root mean square error (RMSE) between the estimated and reference cable routes are 0.247 m and 0.331 m, reducing the two errors by approximately 41.1% and 43.0%, respectively, compared with standalone GWO, while also achieves the lowest error standard deviation and 95th-percentile (P95) error among all the compared methods.
 

subsea cable localization,sliding window FFT,grey wolf optimizer,Unscented Kalman Filter