An adjustable parameter model for tracking a manoeuvring target

Abstract

An adjustable parameter model for tracking manoeuvring targets is presented using a modified variable variance intensity technique. This model is based on the Singer model but with modifications to its fixed manoeuvre statistical characteristics. Meanwhile, a ‘current model’ concept is adapted to describe the statistical distribution of the target's acceleration. Since the variance intensity of the proposed model is calculated using a variable manoeuvring probability density function with adequate covariance, it exhibits a strong adaptive ability for tracking a manoeuvring target and a reduced steady estimation bias in tracking a non-manoeuvring target. Computer simulation results show that a Kalman estimator based on this model can track both manoeuvring and non-manoeuvring targets with greater accuracy than either the Singer model or the interacting multiple model method and is suitable for online processing.