Joint maximum likelihood estimation of specular paths and distributed diffuse scattering

One of the main drawbacks of high-resolution channel parameter estimators proposed so far is that the underlying data model is based on the assumption that the radio channel can be modeled as a superposition of a finite number of specular-like propagation paths. Since this model does not hold in practice, the high-resolution parameter estimators produce erroneous result when they are applied to channel sounding measurements. This is due to the fact that the radio channel contains not only concentrated propagation paths, but also a large number of multipath components resulting from distributed diffuse scattering. To resolve this problem, we propose an algorithm that allows the joint estimation of the parameters of the specular paths and of the contribution of the distributed diffuse scatterer to the channel impulse response. Depending on the available measurements, the algorithm estimates the four coefficients of the polarimetric path weight matrix, the azimuth and elevation at the Tx-site, azimuth and elevation at the Rx-site, the path delay, and the Doppler-shift of the specular components. Furthermore, the estimator determines the parameters of the distribution of the dense multipath components. We outline the main steps of the algorithm and present parameter estimation results from channel sounding measurements.