Enhanced variable dimension filter for maneuvering target tracking

Abstract

An enhancement of the variable dimension (VD) filter for maneuvering-target tracking is presented. The use of measurement concatenation, a procedure whereby fast sampled measurements are stacked while maintaining their proper relationships with the states, leads to significant reduction in estimation error by low processing rate algorithms. The use of double decision logic (DDL) for the maneuver onset and ending detection as well as appropriate procedures for reinitialization of the estimation filters results in improved maneuver detection and filter adaptation. Simulation results show the performance of the proposed enhanced variable dimension (EVD) filter.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>