Image correlation in nonlinear estimation

Even the performance of image-enhanced trackers degrades when an agile target moves along an osculating path. It is possible to improve tracking accuracy by replacing the conventional acceleration model with one which more closely represents measured vehicle motions. This paper presents a study of the performance improvement attainable with algorithms derived on the basis of more exact acceleration models. It is shown that using the correlation between image and kinematic measurements, improvement in tracking performance is possible.<<ETX>>