Interacting multiple model tracking algorithm fusing input estimation and best linear unbiased estimation filter

In manoeuvring target tracking, a primary tradeoff is the robust tracking of manoeuvres against the accurate tracking of constant velocity (CV) motion. To achieve this goal, an interacting multiple model (IMM) algorithm fusing input estimation (IE) and best linear unbiased estimation (BLUE) filter is presented. First, the constant input assumption of IE is modified to track possible manoeuvre fluctuation. Then, an innovation sequence modification technique is proposed so manoeuvre can be detected and estimated sequentially. With improvements above, a modified input estimator (MIE) is designed to track varying manoeuvres. In view of the optimal tracking of BLUE filter for CV motion, MIE and BLUE filter are fused within IMM framework to generate an overall adaptive tracking algorithm. Simulation results reveal the proposed approach yields better accuracy for CV tracking and robustness for manoeuvre tracking.