A novel peak-power control algorithm for digital predistortion with intuitive direct learning

A novel algorithm to control peak-power of the predistorted signal is proposed. The digital predistortion expands peak-power uncontrollably when power amplifers are operated at saturation. In order to control peak expansion, we introduce penalty of a barrier function into the cost function of linearization optimization. Our simulation demonstrates the effectiveness of peak-power controlling for the adaptation with new cost function. Experimental tests show the peak-power controlled DPD makes peak-power under the barrier with only 1 dB linearization performance loss.