A comparison of adaptive tracking filters of variable maneuverability

This paper is concerned with a tracking problem in which the target randomly changes its degree of maneuverability. An optimal filter cannot generally be mechanized for this problem because it has excessive computational requirements. Consequently, two suboptimal adaptive filtering techniques are investigated and compared against a non-adaptive filter design. One of the adaptive techniques is new. Performance and sensitivity studies are presented which show the advantages and disadvantages of each filter. The results demonstrate that, in some situations, adaptive designs can provide a large improvement in system tracking performance over an optimized non-adaptive design.