Tracking of non-ellipsoidal extended target using negative measurements

Abstract

Abstract. This paper presents an algorithm for tracking of non-ellipsoidal extended target using negative measurements, where the extension of the target is modeled as multiple ellipses rather than a simple geometric shape. When positive measurements stemming from the target are scarce due to occlusion, negative measurements stemming from elsewhere can be exploited to get a more reliable estimate. In this work, we introduce a recursive Bayesian filter that incorporates multiple random matrices framework with negative information, where Variational Bayes Expectation Maximization is used to solve the measurement-to-source association issue. The merits of the proposed method are verified via a gamma Gaussian inverse Wishart mixture implementation.