In many practical source tracking applications, the interval of source stat
ionarity may severely vary with time, so that array observations may contai
n both almost stationary data blocks and nonstationary data intervals with
rapidly moving sources. Moreover, typical situations may occur where some s
ources move rapidly within the window exploited, whereas the motion of the
other sources is weak. In such scenarios, the traditional fixed-window appr
oach appears to be nonoptimal because it may lead to a very poor tracking p
erformance. Below, we address the narrowband direction of arrival (DOA) tra
cking problem using a new adaptive-window approach. In our technique, a sep
arate data-driven window is used for each source of interest. The optimizat
ion of window lengths is based on the bias-to-variance tradeoff. The compar
ison of our approach with conventional fixed-window algorithms is presented
showing that the underlying idea has an evident potential in nonstationary
scenarios with rapidly moving sources. A natural price for the improved tr
acking performance is a higher computational cost and the restriction of ou
r approach by the scenarios with 'well-separated' sources. (C) 2000 Elsevie
r Science B.V. All rights reserved.