Key Technologies Of Signal Metric And Optimization For High Efficiency RF Amplification

In modern communication systems, the transmit signals have wide instantaneous bandwidth and large dynamic envelopes, which makes RF power amplifiers(PAs) operate at the large back-off power with low power efficiency, resulting in energy waste and system cost increase.To address this problem, it is desired to find the optimal metric method of PA input signals, which can accurately represent the power efficiency of PA. In this dissertation, the main contributions are summarized as follows.Firstly, it proposes a new metric called distortion components metric(DCM) for the input signals of RF PA. Based on the generalized polynomial PA model, it decomposes the PA output into two uncorrelated components according to the Gram-Schmidt orthogonalisation procedure, and finds that the PA nonlineatities not only contribute to the signal distortion but also to the line amplification of input signals. The DCM is proposed to quantify the distortion power caused by the nonlinear PA. Experiments results show that, in comparison with peak-to-average power ratio(PAPR) and cubic metric(CM), the DCM provides much higher accuracy in determining the amount of back-off power needed to satisfy a given performance requirement. For instance, for the 20 MHz LTE signals with QPSK modulation, the prediction error of DCM is reduced by 86% and 58%, respectively, than those of PAPR and CM.Secondly, it proposes DCM reduction methods for PA input signals based on the distortionless optimization techniques. It analyzes the key factors that affects the system performance in nonlinear transmission, and finds that in the distortionless optimization, minimizing the DCM of signals would directly reduce the signal-to-noise ratio(SNR) loss of systems. And then it employs the typically distortionless optimization techniques, tone reservation(TR) and multiple signals(MS) techniques, to reduce the DCM of signals.(1) According to the property of DCM, it formulates the DCM minimization in TR technique as a D.C.(Difference of Convex functions) programming, and designs an iterative algorithm to solve this problem. The proposed algorithm is able to converge to the optimal value with only two D.C. iteratives. In OFDM system with 128 subcarriers and QPSK modulation, the algorithm has a 4.6d BDCM-reduction performance when reserved subcarrier number is 8. Moreover, the algorithm offers 3d B and 1.1d B SNR gain, respectively, than PAPR and CM optimization, at BER of 310-.(2) The DCM is used as the criterion to select the optimal transmit signal in selective mapping(SLM) and partial transmit sequence(PTS) techniques, and the computation complexities of DCM, PAPR and CM optimizations are compared in detail. Numerical and simulation results show that, in comparison with PAPR and CM optimizations, the DCM optimization offers better system performance. In OFDM system with 128 subcarriers, the computation complexity of DCM optimization increases 10% and 2%, respectively, when compared to the PAPR and CM optimizations.Finally, it proposes DCM reduction methods for PA input signals based on the distortion optimization techniques. It formulates the DCM optimization as a problem of distortion minimization subject to the PA efficiency constraint. The optimization problem is solved by developing a customized interior-point algorithm. Numerical and simulation results show that, in comparison with conventional CM-reduction based convex opimization, the proposed algorithm has the same level of computation complexity, while provides more DCM-reduction and system performance gain. For instance, in OFDM system with 64 subcarriers and QPSK modulation, the proposed algorithm can achieve an additional DCM-reduction gain of 4.54 d B and SNR gain of 1.94 d B.The research results in this dissertation can be applied to the signal design in wireless communication systems, which has theoretical and practical significance for power efficiency enhancement and performance improvement of systems.