For the first time ever, this paper presents the design and implementation of a next-generation of tracking radar systems: the cognitive tracking radar (CTR). At the heart of the CTR, we have a cognitive waveform-selection (CWS) algorithm that can optimally pick the transmit waveform from a prescribed library, in response to information fed back from the receiver to the transmitter. In accordance with dynamic programming, the waveform-selection algorithm seeks to minimize the expected tracking error over a horizon of prescribed length. Approximation of the problem is also studied to mitigate the burden of computational load. To evaluate the system, we introduce computer experiments on a classical ballistic target tracking problem, the results of which confirm the superiority of the CTR over a conventional radar with fixed waveform.
展开▼
