A high-resolution, multi-template deconvolution algorithm for time-domain UWB channel characterization

Time-domain ultra-wideband (UWB) channel characterization with deconvolution has become increasingly popular. However, owing to the nature of the frequency-selective non-line-of-sight channel on a general wideband channel-sounding pulse, the a priori assumption of the CLEAN deconvolution algorithm does not necessarily hold. Moreover, the assumption of a fixed a priori can easily be violated since the radio-wave propagation mechanisms experienced by the sounding pulse can be frequency-selective. This paper first summarizes the use of the CLEAN algorithm in UWB channel characterization, including a discussion about the a priori assumption, enhancements made to the basic algorithm, and the effect of the a priori assumption on channel characteristics. A study of the a priori dependence is then presented along with real-world measurements to demonstrate the shortcomings of the single-template CLEAN algorithm. Finally, a high-resolution, multi-template deconvolution algorithm is proposed to enhance the channel estimation performance. This algorithm incorporates realistic frequency-dependent pulse distortions and is shown to supersede its predecessors by being computationally efficient, simple, accurate, and deployable for point-to-point measurements.