Channel equalization method based on SHAP feature optimization

Abstract

The invention relates to a channel equalization method based on SHAP feature optimization, and the method comprises the steps: (1) enabling a transmitting end to transmit data containing a training sequence to a receiving end through an optical fiber, and enabling the receiving end to input the sampled data containing the training sequence to an equalizer module based on machine learning; (2) extracting a received training sequence by an equalizer module based on machine learning, and constructing a feature vector for the training sequence; (3) obtaining the SHAP value of each feature of each sample through calculation, and measuring the feature weight importance of each feature and the influence on the prediction result based on the index; (4) selecting features based on the feature weights, and constructing feature vectors for the training set and the test set according to a feature selection result; and (5) inputting the constructed feature vector into an equalizer based on a DSP algorithm to equalize subsequent effective data. According to the invention, the system performance is maintained, the influence of clock jitter on the system performance is reduced, and the calculation complexity of channel equalization is reduced.

Description

technical field [0001] The invention relates to a channel equalization method based on SHAP feature optimization applied in a high-speed optical fiber access network transmission system. Background technique [0002] In recent years, with the rise of cloud computing, unmanned driving, virtual reality and other new services, industry and system practicability have put forward higher requirements for data center interconnection based on optical fiber communication. However, short-distance optical transmission systems are extremely cost-sensitive. Therefore, the pursuit of increasing capacity while maintaining low cost is of great significance and is the focus of current data center interconnect system design. One solution to the need for low cost is to use mature and low-cost optoelectronic devices with low bandwidth. However, for high-speed systems based on band-limited transceivers, bandwidth-induced inter-system interference, fiber optic limitations, and cumulative disper...

AI Technical Summary

Problems solved by technology

However, the current equalization scheme based on machine learning has certain defects in terms of system implementation cost, algorithm performance, computational complexity, and robustness.

In addition, the current machine learning algorithms all limit their application in practical fiber optic communication system deployment due to their high complexity and long running time.

Images