Research On Baseband Compensation For The Radio-frequency Non-idealities

With the highly commercialization of wireless communications applications,it is highly desirable to design low-cost low-power wireless communication systems in many scenarios.However,the low-cost system faces various challenges: due to the cost constraints,complex circuits and expensive devices are no longer applicable;moreover,circuits that adopt single-device optimization are also not suitable for the mass produced terminals;all these factors lead to the fact that the low-cost transceivers may suffer from severe radio-frequency(RF)non-idealities,which degrades the system performance drastically.To this end,the transceivers usually apply compensation or calibration algorithms in the digital baseband to mitigate the performance loss.In this paper,some typical RF non-idealities are investigated.Moreover,the impacts of RF non-idealities are analyzed and the corresponding calibration or compensation algorithms are proposed for various system structures.The details and innovations of this paper are as follows:First,a time-domain preamble-based compensation algorithm for the frequency-dependent I\Q imbalance in single-carrier direct-conversion receivers is proposed.The receiver uses the Golay complementary sequences to estimate the frequency-dependent I\Q imbalance.Due to their excellent autocorrelation properties,the Golay complementary sequences significantly simplify the estimation process.Since the Golay sequences are used as the preamble sequence or unique words in some standards(e.g.the 60 GHz standards IEEE 802.15.3c and IEEE 802.11ad),the proposed algorithm needs no modification in the frame structure when applied to these standards,which reduces the system overhead.The simulation results show that the compensation scheme can eliminate the impact of I\Q imbalance in both line-of-sight and non-line-of-sight channels.Second,an adaptive pre-distortion algorithm and an iterative tracking algorithm are proposed to calibrate I\Q imbalance,carrier leakage and in-band distortion in the transmitter.The inverse filter of each branch is estimated directly,which avoids redundant computation.Moreover,an iterative algorithm is proposed to track the parameters variation due to the changing in temperature in an online manner.By complexity analysis,the proposed scheme is proved efficient.The effectiveness of the scheme is testified by simulation.Third,the impacts of the transmitter and receiver I\Q imbalance are theoretically quantified.Moreover,a decision-directed widely-linear(WL)compensation algorithm is proposed to jointly compensate the transmitter and receiver I\Q imbalance,which mitigates the performance degradation caused by the I\Q imbalance.In addition,a generalized parameter-based(PB)algorithm is proposed to enhance the convergence rate thus has a better performance in the fast-fading channel.It also separates the I\Q imbalance parameters.Both algorithms use the differential coding rules to estimate and compensate in an adaptive manner,which avoids the need for training sequences or known pilot signals.The compensation performances of the two proposed algorithms have been verified by simulations.Finally,the impacts of the Wiener phase noise in the DSTC-OFDM system are theoretically analyzed.Without the help from known pilots or training sequences,a decision-directed algorithm is proposed to estimate and suppress the common phase error(CPE)of phase noise in DSTC-OFDM system.In addition,another grouping-based decision-directed estimation algorithm is proposed in order to enhance the performance of phase noise suppression under high-level phase noise.The grouping rules is designed to divide the subcarriers of one OFDM symbol into two groups,and then the decision-directed estimation is carried out in each group.After that,the estimation results are combined and selected in a strategy that effectively avoids the influence of decision error.Both phase noise compensation algorithms proposed are simulated and verified in simulations.The results show that the decision-directed estimation can compensate for the moderate phase noise while the grouping-based algorithm has better compensation performance under severe phase noise.