The power amplifier is one of the core components in the communication transmission system. Its main function is to amplify the modulated signal and transmit it to the antenna for transmission. The power amplifier will produce nonlinear distortion when amplifying the signal, and the parasitic spectrum generated by the nonlinear distortion will cause great interference to the adjacent channel and its own working channel. Nowadays, communication has developed vigorously. On the basis of satisfying basic voice communication, it also has the function of data transmission. Therefore, more complex tasks such as building a communication network put forward higher requirements on the linearity of the power amplifier.

Digital predistortion technology is a common method to improve the linearity of a power amplifier. The digital predistortion technology extracts the input and output characteristics of the power amplifier and constructs a power amplifier model of the power amplifier, and then constructs a suitable linearization strategy. The digital predistortion technology has no stability problems, has a wide applicable bandwidth, high precision, and low implementation cost. The many advantages of digital predistortion technology are very promising and have very good prospects.

The traditional digital predistortion technology models the power amplifier and performs reverse compensation. In recent years, deep learning has made great progress in academic and application.