Combining the interacting multiple model method with particle filters for manoeuvring target tracking

Target tracking is an element of systems that performs tasks such as surveillance, navigation, aviation and obstacle avoidance. It is generally difficult to represent different behavioural aspects of the motion of a manœuvring target with a single model. Therefore multiple model-based approaches are usually required when seeking solutions for manœuvring target tracking problems, which are generally non-linear. In the recent years, new strategies have been developed via the combination of the interacting multiple model (IMM) method and variants of particle filters (PFs). The former accounts for mode switching, while the latter account for non-linearity andœor non-Gaussianity in the dynamic system models for the posed problems. This paper considers an IMM algorithm for tracking three-dimensional (3D) target motion with manœuvres. The proposed algorithm comprises a constant velocity model, a constant acceleration model and a 3D turning rate (3DTR) model. A variety of combinations of extended Kalman filters (EKFs), unscented Kalman filters (UKFs) and PFs are used for the models. The proposed IMM algorithm variants are applied to a problem on the 3D manœuvring target tracking. Simulation test results show that by using a computationally economical PF in the 3DTR model, with EKFs andœor UKFs in the remaining models, superior performance in state estimation can be achieved at relatively modest computational costs.