Rapid expansion of wireless communication services during the last decade is expected to continue at an even faster pace in the future. This growth creates a demand for novel modulation, coding, and detection methods suitable for the rapidly varying multipath fading channels encountered in mobile radio. We propose to investigate a new approach to the estimation and prediction of the fading channel which does not rely on statistical characterization. In this method, the physical nature of the flat fading signal is exploited to yield a deterministic model that predicts the channel for an entire block of data. Specifically, the mobile radio signal is decomposed into several scattered components distinguished by their Doppler shifts; the slowly varying parameters associated with these components are determined and tracked, and the composite fading signal is predicted. This approach may potentially result in the ability to anticipate and compensate for the "deep fades that severely limit the performance of wireless communication systems and will aid in the development of low-power mobile radio systems.
