An Adaptive δ-GLMB Filter Under Noise Statistics Mismatch

Abstract

Aiming at the problem of multi-target tracking under noise statistics mismatch, an adaptive δ-GLMB filter based on variational Bayesian (VB) approach is proposed. The joint distribution of predicted state and corresponding predicted error covariance matrix, and the joint distribution of measurement noise mean vector and covariance matrix are modeled as the Normal-inverse Wishart (NIW) distributions, in which the latent variables are described as the Gamma distributions. In this paper, the single-target filtering density is expressed as the mixture of Normal inverse Wishart inverse Wishart Gamma Gamma (NNIWNIWGG), and an NNIWNIWGG mixture implementation of δ-generalized labeled multi-Bernoulli (δ-GLMB) filter for linear Gaussian is given. According to the minimization of Kullback-Leibler divergence, the approximate solution of predicted likelihood is obtained. Simulation results show that the proposed adaptive δ-GLMB filter has high tracking accuracy in the case of noises statistics mismatch.