Maneuvering target tracking by using particle filter method with model switching structure

Tracking problem of maneuvering target is treated with assumption that the maneuver is unknown and its acceleration has abrupt changes sometimes. To cope with unknown maneuver, Bayesian switching structure model, which includes a set of possible models and switches among them, is used. It can be formalized into general (nonlinear, non-Gaussian) state space model where system model describes the target dynamics and observation model represents a process to observe the target position. Heavy-tailed uni-modal distribution, e.g. Cauchy distribution, is used for the system noise to accomplish good performance of tracking both for constant period and abrupt changing time point of acceleration. Monte Carlo filter, which is a kind of particle filter that approximates state distribution by many particles in state space, is used for the state estimation of the model. A simulation study shows the efficiency of the proposed model by comparing with Gaussian case of Bayesian switching structure model.