Delay-doppler frequency power spectrum estimation for vehicular propagation channels

In this contribution, an algorithm is derived based on the Space-Alternating Generalized Expectation-maximization (SAGE) algorithm for estimating the power spectrum of time-variant propagation channel dispersive in delay and Doppler frequency. The proposed algorithm utilizes a signal model where each component is characterized by a parametric delay-Doppler frequency power spectral density function. The parameter estimates obtained are directly the small-scale parameters required for constructing stochastic channel models. Compared with conventional modeling approaches that compute the small-scale parameters based on parameter estimates of propagation paths, the proposed algorithm can avoid the errors in the parameter estimation of specular paths, and provide accurate estimates of statistical parameters in the cases where a small amount of observation samples are available.