Research And Verification Of Wideband Zero-IF Receiver's IQ Imbalance Calibration Technique

Zero-IF receivers have high integration and low power consumption,and have been widely studied and applied in recent years.Zero-IF receivers need to process I/Q signals separately.Analog devices such as amplifiers,mixers,and filters in the RF channel cause the amplitude and phase characteristics of the two branches to be different,that is,I/Q imbalance.Affected by I/Q imbalance,the wanted signal will be interfered by mirror image,which will directly lead to receiver performance degradation.Current I/Q imbalance compensation algorithms typically assume that I/Q imbalance is frequency independent and therefore can only be used for narrowband signal correction.RF channels of wideband receiver is frequency selective,resulting in I/Q imbalance being also frequency dependent.It is difficult to compensate the frequency-dependent I/Q imbalance using traditional technique.Therefore,this paper studies the I/Q imbalance correction algorithm for wideband zero-IF receivers.The details include:Firstly,the correction requirements for the I/Q imbalance correction of the wideband zero-IF receiver are analyzed.A non-ideal zero-IF receiver model is designed and the deterioration of the received signal influenced by frequency-dependent I/Q imbalance is analyzed.The test platform based on AD9361 chip is built,and the internal IQ correction performance of AD9361 chip is analyzed.The frequency dependence of I/Q imbalance is verified,and the requirements of correction algorithm are proposed accordingly.Secondly,the I/Q imbalanced frequency domain correction algorithm for wideband zero-IF receivers is studied.Based on the analysis of signal power spectral density and frequency domain correlation,a blind estimation algorithm is designed,which can separately estimate and correct frequency-dependent and frequency-independent I/Q imbalance.The simulation results show that the error vector magnitude(EVM)of the corrected 16-QAM signal with 14 MHz bandwidth is reduced from-18 dB to-51 dB.Thirdly,the time domain I/Q imbalance correction algorithm based on recursive least squares(RLS)algorithm is studied.Based on the circularity of wideband random signals,an adaptive blind estimation algorithm based on RLS is designed.According to the simulation results,the EVM of the corrected wideband signal is reduced from-19 dB to-47 dB.Fourthly,the Newton-like time domain correction algorithm is studied.By tracking the properness of the signal,an adaptive blind correction algorithm based on Newton iteration method is designed,and the filter and step size design of the algorithm are simulated and analyzed.After verifying on the software radio platform,the performance of the algorithm is better than that of the AD9361,and the EVM of the measured LTE signal with 20 MHz bandwidth can be less than-36 dB.In this paper,a frequency-dependent I/Q imbalance correction algorithm is designed for wideband zero-IF receivers,and the performance of the scheme is verified by tests.The paper has certain reference significance for the IQ imbalance compensation algorithm design of wideband zero-IF receiver.