Hardware Impairment Detection and Prewhitening on MIMO Precoder for Spectrum Sharing

Multiple input multiple out (MIMO) cognitive radio offer the spatial degree of freedom that can be used to share the spectrum with less interference via Precoding Technique. Many precoding techniques in the literature assume that the underlying hardware is ideal. In the practical case, hardware adds many impairments at both the transmitter and receiver side. Such hardware impairments degrade the performance of the MIMO system. This impairment introduces additional interference to the primary user in minimum interference precoder of MIMO cognitive radio and spoils the objective of the precoder. This paper analyzes the performance of the cognitive radio with minimum interference MIMO precoding in the presence of hardware impairment. Detecting the presence of hardware impairment and its exact variance is required to mitigate the impairment by prewhitening. In general all prewhitening work, it is assumed that the transmitter knows the presence of impairments and the exact variance of it. This work presents support vector machine (SVM) based hardware impairment detection technique on the real-time acquired data by using national instruments, vector signal generator (VSG) 5673. The VSG is configured in 2 × 2 MIMO setup to acquire the transmitter data without impairment and with injected impairment like DC offset, phase noise, I/Q gain imbalance, quadrature skew and frequency offset. After detecting the presence of hardware impairment by SVM, the variance of the data is calculated using Error Vector Magnitude (EVM) to enable prewhitening. The results show that the detection of hardware impairment can be possible with 100% of accuracy and the prewhitening scheme completely removes the impact of it.