Adaptive nonlinearity compensation technique for 4G wireless transmitters

Abstract

A new adaptive digital predistortion based on an iterative injections approach is experimentally verified using a 4G signal. An error vector magnitude and spectrum regrowth improvements are achieved, while at the same time not requiring large digital signal processing memory resources and computational complexity. Therefore, the method can be a strong candidate to become a part of modern predistorters for nonlinearity compensation in 4G wireless transmitters.Introduction: The power amplifier (PA) linearisation method based on injection was initially introduced in [1] where second harmonic and frequency deference injections are applied. In [2], the authors proposed using third-order (IM3) component injection. Digital predistortion (DPD) methods based on injections are used in [3][4][5]. In [3], the IM3 and IM5 component injections were utilised. A combined look-up tables (LUTs)-injection approach was used in [4]. The authors in [5] used the iterative injection method for PA linearisation purposes in order to overcome the distortion compensation limit [3]. In general, linearisation methods based on the injection approach have smaller complexity compared with other proposed techniques for nonlinearity compensation based on Volterra, LUTs or polynomials [6 -8]. However, they usually suffer from a compensation limit introduced by new components that appear after injections [3]. Moreover, the existing injection-based techniques were tested in open loop and, due to that fact, are not adaptive and have limited usage in real transmitters. In addition, there was no experimental proof of how these techniques behave with 4G signals which have very variable amplitude.In this Letter, we demonstrate the PA linearisation using the proposed DPD based on adaptive iterative injections of the in-band distortion components. This method has two important advantages. First, it uses real multiplications and real additions only and avoids complex Volterra, LUT or polynomials for nonlinearity compensation. Secondly, in contrast to existing injection-based DPD [3][4][5], the technique uses an adaptive architecture [7], and therefore is adaptive to the variation of the PA nonlinear transfer function in real environmental conditions. The proposed DPD was verified experimentally using a more than 10 dB peak-toaverage power ratio (PAPR) 5 MHz downlink long-term evolution (LTE) signal.

Performance of fixed‐gain amplify‐and‐forward nonlinear relaying with hardware impairments

“…Impairments have a very large impact especially in modern multicarrier communication systems utilizing orthogonal frequency division multiplexing (OFDM) signaling. In such systems, a high peak to average power ratio [15,16] is a major problem resulting in its high sensitivity to nonlinear distortion due to PA nonlinearity. E jx 1 j 2 D P i .…”

Improved baseband digital predistortion for linearising PAs with nonlinear memory effects using linearly interpolated LUT

“…Classical PA models include the look‐up table (LUT) model, Volterra series model, memory polynomial (MP) model, Hammerstein model, Wiener model and so on [6]. In [1], an adaptive closed‐loop DPD based on polynomial nonlinearity compensation is proposed. In [2], a dual band DPD based on the MP model and the LUT is proposed.…”

“…The most efficient and promising technique for linearising PAs is digital predistortion (DPD) [1][2][3][4][5]. In DPD, a predistorter is inserted before the PA so that the whole system has a linear behaviour.…”

Digital predistortion method combining memory polynomial and feed‐forward neural network

“…They can be classified into three categories: look-up table (LUT)-based DPDs [2,3], polynomial-based DPDs [4][5][6] and neural network (NN)-based DPDs [7,8]. The LUT-based DPD is noted for its simplicity, but its linearisation performance depends on the size of the LUT.…”