Waveform Design For Track-Before-Detect-Based Cognitive Radars

Detect-before-track-based cognitive radars in which
threshold detections are taken as the input of tracking, irre-
versibly result in high false alarm under the case of low signal-
to-noise ratio (SNR). To solve this problem, in this paper, we
propose a framework of cognitive radars based on track-before-
detect (TBD) technique. This framework includes the TBD mea-
surement model consisting of received ambiguity function without
threshold detection, cubature Kalman filter to estimate target
state, and the feedback mechanism and optimization criterion for
the next transmitted waveform. In particular, waveform design
problem in the TBD-based cognitive radars is emphasized. This
work opens the door to the cognitive radars based on TBD
technique, and reveals their potential in target tracking under
the case of low SNR. Numerical results demonstrate that better
target tracking performance can be achieved by the TBD-based
cognitive radars, as compared with conventional radars.