The Poisson multi-Bernoulli (PMB) filter offers an elegant Bayesian formulation for multi-target tracking by approximating the complex PMBM density as a PMB density. This approximation allows for more efficient computation and simpler implementation while maintaining reasonable accuracy. However, approximating the posterior density involves a certain degree of tracking accuracy loss, and different approximation methods may result in different types of tracking accuracy losses. To address this problem, this paper proposes a partitioned PMB (PPMB) filter that decomposes the multi-target tracking problem to smaller sub-problems, which can synthesize the advantages of different approximations to avoid tracking accuracy loss. Furthermore, to address the issue of tracking lost for close-spaced and undetected targets, an improved measurement oriented multi-Bernoulli Poisson (IMOMB/P) filter is also proposed, which utilizes more accurate marginal association probabilities. Simulation results demonstrate that the proposed PPMB filter outperforms existing PMB filters, such as the MOMB/P and track oriented multi-Bernoulli (TOMB) filters, thus showing promising application prospects.
Robust target tracking with non-stationary heavy-tailed measurement noise caused by fluctuating radar cross-section (RCS) remains a significant challenge. This letter proposes a novel robust Kalman filter that jointly estimates the target state and its RCS by fusing kinematic measurements and amplitude information. First, we propose a novel Normal-inverse-Wishart-Chi-square-inverse-Gamma distribution to model the non-stationary heavy-tailed measurement noise that arises from RCS fluctuations. Second, the Rician amplitude likelihood is simplified via a Gaussian approximation. Third, the tracking problem is formulated within a variational Bayesian framework to approximate the analytically intractable joint posterior distribution. Simulation results demonstrate that the proposed method significantly enhances tracking accuracy and robustness compared to conventional robust filters.
10 December 2014 in Journal of Information and Computational Science