Applications of Spectral Estimation Techniques to Radar Doppler Processing: Simulation and Analysis of HF (High-Frquency) Skywave Radar Data

摘要

This work is the second paper in a series of studies of the application of spectral estimations techniques to Doppler processing of coherent radar signals. In this work, simulated high-frequency (HF) radar sea scatter time series are generated and processed by use of three different spectral estimation algorithms and the fast Fourier transform (FFT). The sea clutter is simulated by narrowband filtering a wideband Gaussian noise spectrum in the frequency domain, with filter widths appropriate to describe first-order Bragg lines and second-order continuum. Targets are introduced as sinusoids, stepped by 5 dB for eight different echo power values, and stepped in Doppler frequency for four different values relative to the clutter. These simulations identify problems that appear unique to Doppler processing of coherent radar data in the presence of broadband clutter, and are in distinction to the application of spectral estimation to processing in the spatial domain. In the latter case, the spectral contributions are generally narrowly confined in the angular power spectral estimate, and the aim is to separate these contributions in the presence of noise. The HF radar application is concerned with separation of weak targets in the presence of stronger clutter returns, which are relatively broad compared to the target return. It appears that the Burg maximum entropy method allows the detection of targets in clutter under conditions which the FFT is incapable of detection with any degree of certainty. (Author)