Shipborne Multisource Fusion Framework for Ocean Wave Inversion: Integrating X-Band Radar and Motion Data

Accurate perception of near-field waves is critical for the safety of ship navigation. Existing X-band radar inversion methods face dual constraints: spectral-analysis methods require buoy-assisted calibration, and data-driven approaches fail to achieve comprehensive wave parameter inversion. To overcome these limitations, this paper proposes RaMoWave, a novel shipborne multi-source fusion framework for ocean wave inversion. The key innovations include: 1) A generalized ship motion intelligent inversion model that extracts significant wave height (SWH) and spectral peak period under arbitrary wave encounter angles; 2) A feature fusion strategy that optimizes radar-derived wave spectra and reconstructs wave fields without relying on buoy data. Extensive validation via numerical simulations and sea trials (SWH: 0.5-5 m (simulation) and 0.5-2.4 m (sea trials) ) demonstrates threefold superiority: 1) It enables comprehensive inversion of SWH, spectral peak period, characteristic period, and wave field—capabilities unattainable by existing intelligent approaches; 2) It achieves a 30% improvement in parameter inversion accuracy over spectral-analysis methods, with error reductions up to 50% in buoy-free scenarios; 3) It maintains relative inversion errors below 7% across all scenarios, exhibiting superior applicability and robustness.