An Adaptive Tracking Algorithm for Bearings-only Maneuvering Target

To improve the tracking performance of bearings-only maneuvering target, a dynamic and adaptive multiple models (MM) algorithm is proposed in this paper, i.e., the acceleration value of each sub-model of MM structure is adjusted dynamically, unlike the traditional MM method with fixed acceleration levels, by an economic on-line self-constructing neural fuzzy inference network (SONFIN) according to the changes of extracted feature information, i.e., innovation variation, heading change and bearing variation Then a set of Unscented Kalman Filter (UKF) is utilized to estimate target state. Numerical simulation results show that the performance of the proposed algorithm is nearly identical to that of the interactive multiple models (IMM), and it is also free of any prior information of target motion. Furthermore, it can deal with the maneuvering target with time varying acceleration.