A Digital Predistorion Method with Low-complexity Update Process for Scenarios of Power Variation

This paper proposes a lightweight digital distortion (DPD) coefficient amendment method based on an indirect learning (ILA) architecture, which is especially suitable for occasions where the signal power fluctuates frequently. Static DPD coefficients are usually difficult to adapt to the dynamic nonlinear characteristic change of the power amplifier (PA) caused by input signal power variation. Therefore, it is necessary to adjust the DPD model coefficients appropriately to improve the out-of-band (OOB) nonlinear suppression effect. To verify the feasibility of the proposed method, simulations are carried out to assess the adjacent channel leakage ratio (ACLR) of the output signal of the PA. This low-complexity coefficient amendment method could be potentially employed in scenarios such as 6G intelligent radio, where the average power of the input signal is fastly time-varied.