Non-linearly Induced Interference and its Mitigation in Cognitive Wideband Receivers

Abstract:
Software defined radios are widely used as multimode and multi-band RF transceivers for modern communication systems, especially in cognitive radio. However, the design of flexible, cheap, and reconfigurable front-ends is a challenging task in terms of the RF transceiver electronics. Many requirements to the analogue components can be met by today's technologies, however, this may not always be practical except for measurements and prototyping. This paper addresses RF impairments in wideband cognitive radios, especially non-linear distortions in the receiver. Receiver non-linearities affect the reliability of spectrum sensing, a key issue in spectrum awareness. Here, the focus is on mitigation of non-linear distortions using "Dirty RF" techniques. We present experimental work studying the effects of non-linear distortions and analysing the performance of a typical wideband softwaredefined receiver. A feed-forward mitigation algorithm with an adaptive filter has been implemented and applied to real measurement data for cleaning the baseband spectrum of distortion products. The results obtained show that a significant suppression of distortion products can be achieved.