Location of subchip radio channel multipath components from direct sequence sounding data

This paper presents novel instantaneous, radio channel modeling algorithm. It is based on deconvolution and has four basic modules that are: (1) determination of initial estimate by matched filter deconvolution, (2) suppression of deconvolution noise, (3) selection of signal taps and (4) complex tap amplitude accuracy enhancement. The paper analyzes parameter selection and gives an outline of theoretical and simulated performance. As an example we note that the algorithm can achieve with 6 times oversampled 15 chip maximal length code and 5 Rayleigh channel taps an average amplitude accuracy enhancement of 8 dB to 19 dB compared to uncompensated optimum deconvolution when the channel SNR is 20 to 40 dB respectively.