Compensation of nonlinear distortions with memory effects in digital transmitters

High-level linear modulation schemes used in modern digital communication systems exhibit large peak-to-average power ratios (PAPRs). The performance of transceivers is very sensitive to nonlinear distortions, which arise mainly from the high power amplifier (HPA). Also, because of the wideband signals, the nonlinear distortions are frequency-dependent. The paper proposes an algebraic solution to compensate at the transmitter for the HPA nonlinearity. The HPA is represented by a memoryless nonlinear block followed by a linear filter. We first estimate the parameters of the unknown nonlinearity, which is modelled through a polynomial expansion. The frequency response of the unknown filter is then calculated, in order to capture the memory effects in the system. Using the identified nonlinear system parameters, a cascade of the inverse filter and the inverse memoryless nonlinearity is constructed preceding the HPA, in order to predistort the input signals and achieve overall linear transmitter characteristics. The scheme is examined through computer simulations for quadrature amplitude modulation (QAM). Improvements in the bit error rate (BER) and out-of-band spectrum regrowth are demonstrated for the travelling wave tube (TWT) HPA model. The results show that the proposed method is effective in compensating for amplitude-to-amplitude (AM/AM) distortions with memory using a relatively-small number of data points at the identification stage.