Novel Multi-Target Tracking Based on Poisson Multi-Bernoulli Mixture Filter for High-Clutter Maritime Communications

Effective multi-target tracking is vital for maritime communications. However, the maritime environment poses significant challenges, such as variable clutter and dynamic wave conditions, which often compromise the accuracy and reliability of conventional tracking methods. The filters based on Random Finite Set (RFS) are widely utilized to deal with the above challenges. In this paper, Poisson Multi-Bernoulli Mixture (PMBM) filter is involved in tracking multi-targets in a high clutter rate environment, where wind speed and wave height are considered as two key components to model maritime clutter. The Weibull distribution and an empirical model are used to represent maritime conditions. Simulation results confirm the PMBM filter’s superior performance, showing a significant improvement in Generalized Optimal SubPattern Assignment (GOSPA) and root mean square Generalized Optimal SubPattern Assignment (RMS GOSPA) metrics compared to the Multi-Bernoulli Mixture (MBM) and Probability Hypothesis Density (PHD) filters. These findings not only confirm the proposed filter’s efficacy but also provide efficient tracking schemes for maritime communications.