Statistical modeling of sea clutter in USV-based maritime SISO and MIMO radar configurations

Sea clutter remains a major limitation in maritime radar systems, often obscuring small or low Radar Cross Section (RCS) targets and degrading detection performance. Most existing studies focus on shore-to-sea propagation using Single Input Single Output (SISO) radar configurations, while limited attention has been given to sea-to-sea propagation scenarios involving Unmanned Surface Vehicles (USVs) and Multiple Input Multiple Output (MIMO) radar systems. This research proposes a statistical sea clutter modelling framework for MIMO-based maritime radar operating under sea-to-sea propagation conditions. Experimental data were acquired in a controlled wave tank environment with varying wave heights, and radar returns were processed using Hilbert Transform-based envelope extraction. Sea clutter characteristics were modelled using multiple probability distribution functions, namely Log-Normal, Log-Logistic, Gamma, and Weibull distributions, with parameter estimation performed via Maximum Likelihood Estimation. Model validation was conducted using Goodness-of-Fit analysis, Root Mean Square Error, percentage error, and log-likelihood evaluation between measured and simulated data. Results demonstrate that MIMO radar provides enhanced spatial diversity but introduces increased clutter variability compared to SISO configurations. The proposed modelling framework achieves high agreement between measured and simulated clutter signals, with percentage errors below 0.4% for SISO and 0.1% for MIMO cases. Among the evaluated models, the Weibull distribution consistently provides the best representation across varying sea states. This study contributes a validated sea clutter model tailored for USV-based MIMO maritime radar systems, offering practical insights for optimizing radar deployment and improving small-target detection in complex maritime environments.