Power amplifier pre-distortion method of complex-valued pipeline recurrent neural network model

Abstract

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.

Description

technical field [0001] The invention relates to the field of digital pre-distortion in digital signal processing, in particular to a power amplifier pre-distortion method for a complex-valued pipeline recursive neural network model. Background technique [0002] With the rapid development of wireless communication technology, especially when 4G and 5G mobile communications are widely popularized, the demand for spectrum resources in modern communication systems continues to increase. In order to meet this requirement, high-order modulation techniques are widely used in communication systems, but such modulation techniques will make it more difficult to design RF power amplifiers, which play a crucial role in RF front-end components, and will also cause spectrum regrowth . In order to meet the needs of modern wireless communication technology, a series of power amplifier linearization and efficiency enhancement technologies have been proposed, such as feedforward technology ...

AI Technical Summary

Problems solved by technology

[0005] In view of the fact that the existing neural network model cannot well represent the memory effect of the power amplifier, and when the input signal is complex-valued, a cumbersome training algorithm will be introduced to cause lengthy training time, and it is proposed that the memory effect of the power amplifier must be represented , and can process complex-valued signals, and can quickly update the parameters of the model, a power amplifier pre-distortion method for a complex-valued pipeline recursive neural network model. The specific technical scheme is as follows:

Images