Adaptive Nonlinear Narrow Band Interference Rejection in Ternary DS-UWB

It is known that the narrowband interference (NBI) rejection capabilities of a direct sequence spread spectrum system can be improved by employing an adaptive FIR rejection filter. However, due to the non-Gaussian noise input to the filter, this rejection filter is suboptimum. The optimum filter can be achieved by introducing a nonlinear function to estimate the value of the spread spectrum signal, and subtract this estimate from the input to the rejection filter. The efficacy of linear prediction in the UWB environment has not been tested. This paper derives the optimum nonlinear function for a ternary DS-UWB system, and applies it to a LMS rejection filter, and the introduction of an interleaver eliminates the need for the symbols received during the filters training period to be discarded. A considerable gain is achieved with the inclusion of this function and interleaver, and simulations demonstrate the robustness of the proposed filter with a DS-UWB system co-existing with multiple 802.11a interferering signals.