Title :
Maneuvering target motion analysis using BSpline representation
Author_Institution :
Thomson-Sintra ASM, 525 route des dolines, BP 157 06903 Sophia-Antipolis cedex, France
Abstract :
Target motion analysis (TMA) for a rectilinear source movement (RSM) has been intensively studied in the last ten years. But difficulties still exist, especially when source heading or speed changes are within the same time as the conventional TMA convergence time. This paper is concerned with a new method of batch TMA for maneuvering sources using a non-linear least-squares fit between the whole set of measurements and a BSpline trajectory representation. It provides a good way to globally estimate both the instants of maneuvers and their number with an experimentally robust model order selection method. This work includes tests on actual data from at-sea recordings.
Keywords :
Data models; Hidden Markov models; Noise; Noise measurement; Polynomials; Robustness; Trajectory;
Conference_Titel :
European Signal Processing Conference, 1996. EUSIPCO 1996. 8th
Conference_Location :
Trieste, Italy
Print_ISBN :
978-888-6179-83-6