Gaussian Mixture PHD Filter with State-Dependent Jump Markov System Models
Dohyeung Kim, Inseok Hwang · 2019
The Gaussian mixture probability hypothesis density (GM-PHD) filter with jump Markov system (JMS) models assume that mode transition probabilities are constant, irrespective of the target state. However, in some applications (e.g., air traffic control), the mode transition of the target is dependent on its state and thus the mode transition probabilities are a function of the target state. This paper proposes a multiple model GM-PHD filter with state-dependent mode transition probabilities which are represented as Gaussian probability density functions. The performance of the proposed algorithm is illustrated with an example in an air traffic control application.