High range resolution profile estimation via a cognitive stepped frequency technique

The problem of high range resolution profile (HRRP) estimation is considered in this paper. In particular, stepped frequency waveforms are devised to enhance the target range profile (RP) estimation accuracy. The basic idea relies on the dynamic optimization of the probing waveform accounting for some feedback information to minimize the profile estimation error. The transmitted frequency hopping pattern is selected so as to minimize the predicted Cramer-Rao lower bound (CRLB) associated with the amplitudes and locations of the scattering centers falling in the coarse range bin under test (i.e., the target profile). Specifically, it is assumed that the target imaging is initially performed via a conventional linear stepped frequency transmission, hence, based on the collected data, an initial prediction of the target profile is derived (perception). Then, the RP estimation is enhanced progressively according to the cognitive paradigm, via a specific frequency pattern selection at the next transmission (action). The results highlight the capabilities of the cognitive approach to provide interesting benefits with respect to the classic linear stepped frequency strategy.