Parametric spectrum analysis for target characterization

The most commonly applied technique to estimate the power density spectrum of a frequency modulated, continuous wave (FMCW) microwave radar system is the fast Fourier transform (FFT), which applies a deterministic signal approximation by a linear combination of a set of orthogonal functions (e.g. sine and cosine waves). In this paper the authors present an original approach of FMCW radar spectrum analysis with the autoregressive parametric estimation, that yields additional information due to the fact that with this proposed method a more detailed signal characterization is possible. The main target of interest that is treated in this article is the road surface, which enables one to conduct both absolute speed over ground measurements and road classification according to its roughness. A 38 GHz FMCW laboratory radar serves as a test platform to verify the feasibility of the authors' approach. The road classification procedure is based on "test" signals previously recorded on different road surfaces. First evaluations with different classification methods, to determine the road structure, show promising results.