The invention discloses a power
amplifier pre-
distortion method for a complex-valued pipeline
recurrent neural network model. The power
amplifier pre-
distortion method comprises the steps: carrying out the modeling of a power
amplifier behavior model through the complex-valued pipeline
recurrent neural network model, solving a pre-distorter module, and carrying out the pre-
distortion operation ofa power amplifier input
signal, and specifically includes the steps: firstly, part of input and output signals of a power amplifier are used as test signals, and forward modeling is conducted on the power amplifier, and the model weight is optimized through an enhanced complex value real-time recursive learning
algorithm, and the optimal model weight is obtained, and the nonlinear and memory capacity of the power amplifier represented by the model is checked; secondly, inversion is performed on the model so as to perform
reverse modeling on the power amplifier to obtain a predistorter structure; and finally, the input
signal of the power amplifier obtains a
signal subjected to pre-distortion compensation through the predistorter structure, and sends the signal into the power amplifier, sothat the
adjacent channel power ratio of the output signal of the power amplifier obtained at the moment can be remarkably improved.