Time resolution-signal-to-noise ratio trade-off in deconvolved and Wiener filtered radio channel sounding

It is an often anticipated fact that the time resolution of direct sequence-type radio channel sounding is limited only partially by the sounding sequence chip duration. We have previously introduced a computationally efficient and noise resistant method to achieve a time resolution approaching sample time called matched filter deconvolution (MFD). This paper considers additional effects of Wiener filtering when used with the MFD method. Our simulations show how filtering trades SNR to time resolution and how this trading reacts on sampling rate/sequence rate ratio. Based on this analysis we suggest Wiener filtering to be considered when deconvolving radio channels having poor SNR and/or when channel measurement results should be used for playback channel simulations.