Joint HFM-CW Active Sonar Tracking via Velocity-Domain Bayesian Fusion With Implicit HFM Range-Bias Compensation

Abstract

This letter addresses target tracking for combined hyperbolic frequency modulation (HFM) and continuous-wave (CW) active sonar in clutter-dominated environments. HFM provides range-azimuth evidence, whereas CW provides Doppler-derived radial velocity. Radial motion biases the HFM range coordinate, while clutter produces false spatial peaks. Our velocity-domain Bayesian fusion framework maps HFM candidates into previous-state-conditioned velocity hypotheses for joint evaluation of HFM evidence, CW consistency, kinematic feasibility, and an inertial prior. The method embeds HFM range-bias compensation into likelihood evaluation and suppresses false associations before maximum a posteriori state update. Semi-synthetic experiments on sea-trial backgrounds show 72.9–85.0% lower position root-mean-square error (RMSE) than cascaded normalized cross-correlation plus CW (NCC+CW) across nominal HFM target-to-background ratios (TBRs) of -25 to -5 dB. The 95th-percentile (P95) tracking-stage latency is 0.69 s.