How wideband receiver nonlinearities impact spectrum sensing

Wideband receivers are appealing for cognitive radio spectrum sensing as they allow sensing multiple subbands in parallel. However, wideband radio front-end circuits introduce various impairments into the received signal, which make detection of unoccupied bands difficult. In particular, the low noise amplifier and mixer are typically nonlinear components due to which the wideband received signal experiences internal mixing of the signal components, often referred to as intermodulation terms. The impact of these intermodulation terms on the wideband spectrum sensing might be significant in cases of both detection of occupied and unoccupied bands. In particular, energy detection based spectrum sensing might result in high false alarm probability due to increased energy in the band of interest due to intermodulation terms. In this paper we analyze the impact of LNA and mixer nonlinearities on energy detection based spectrum sensing, and show that although both components result in intermodulation terms in different subbands of the wideband channel, they both result in severe degradation of the spectrum sensing performance. We also discuss possible digital signal processing algorithms for compensation of nonlinearities.