Digital pre-distortion linearization parameter extraction method based on cloud platform

Abstract

The invention discloses a digital pre-distortion linearization parameter extraction method based on a cloud platform. A digital pre-distortion hardware platform comprises a vector signal generator, apower amplifier, a spectrum analyzer, an attenuator, a coupler and a load. The vector signal generator is connected with the power amplifier. The attenuator is connected with the coupler. The coupleris connected with the spectrum analyzer and the load. The cloud platform comprises measurement servers and a plurality of databases. Each database is connected with a plurality of measurement servers.The plurality of measurement servers are connected with an application server. The application server is connected with client terminals through the Internet. The measurement servers are connected with the vector signal generator and the spectrum analyzer. The method has the advantages that the hardware use cost can be reduced, the cost is relatively low, the digital pre-distortion linearizationefficiency is effectively improved, and a multi-power amplifier concurrent request digital pre-distortion linearization processing demand can be satisfied.

Description

technical field [0001] The invention relates to a method for extracting digital predistortion linearization parameters, in particular to a method for extracting digital predistortion linearization parameters based on a cloud platform. Background technique [0002] The main function of the power amplifier is to amplify the modulated signal to the required power. It is an indispensable key component in modern wireless communication systems. However, the nonlinear characteristics of the power amplifier itself, on the one hand, cause the signal to cause spectrum regeneration or expansion outside the band and interfere with adjacent channels; Especially in broadband communication, the memory effect is obvious, seriously affecting the normal transmission of the communication system. [0003] At present, RF power amplifiers mainly use digital pre-distortion technology to avoid interference to adjacent channels and distortion generated by signals passing through nonlinear RF power ...

AI Technical Summary

Problems solved by technology

The traditional digital pre-distortion system is based on hardware chip technology, but hardware implementation of digital pre-distortion will cause loop delay problems in the circuit, and requires expensive equipment support, high system cost, high power consumption, poor stability, and does not support multiple Unified coordination, centralized management and parallel processing of power amplifier predistortion parameter extraction

[0004] The effectiveness of the digital predistortion system largely depends on the accuracy of the nonlinear modeling of the power amplifier PA. There are many modeling types proposed at present, including Saleh model, memory polynomial model, Wiener model, Hammerstein model, Volterra model and Its tailoring model, neural network model, etc. Many models have problems such as complex algorithms and difficult convergence, and these models have higher and higher requirements for training and computing servers.

Among them, the pre-distortion method based on neural network has a good linearization effect, and is a more efficient solution for training nonlinear models, but the actual application situation is that due to the long data training time, it is easy to fall into local convergence, and the massive data cannot adapt to the actual situation. Issues such as requirements cannot be effectively implemented

At the same time, for the traditional predistortion system, the IQ two-channel signal fed back and collected back has higher and higher requirements on the processor speed, and the hardware overhead is large. For example, the sampling bandwidth of the current digital predistortion needs to be at least 3 to 5 times the signal bandwidth.

With the widening of communication bandwidth, such as LTE-Advanced five-carrier signal, which already has a signal bandwidth of 100MHz, according to the five times the sampling rate required by the pre-distortion feedback channel, a sampling rate of 500MSPS is required, and if intermediate frequency signal processing is used, 1GSPS is required Sampling rate, such A / D is very expensive, usually limited or no original device, thus greatly increasing the difficulty of pre-distortion implementation

On the other hand, with the expansion of applications, the concurrent pre-distortion processing requirements of multiple power amplifier systems are also increasing, and the professional experimental environment cannot meet the needs of users for remote use.

Images