An IMM tracking algorithm with input estimation

A new approach to track manoeuvring targets is presented. A target model that combines Singer's model with a deterministic step manoeuvre model is also proposed to account for the various realistic evasive manoeuvre strategies. The interacting multiple models (IMM) method incurs a mean tracking error in the presence of a pilot-commanded abrupt target manoeuvre. A recursive real-time least-squares algorithm to compute the magnitude of the input acceleration is devised to reduce the tracking error. The combined scheme of this input estimation filter and the IMM algorithm markedly improves the tracking accuracy. Simulation results show that the proposed algorithm is superior to that of the IMM algorithm, especially in velocity and acceleration estimations.