Research on clustering analysis of eye diagram point set of digital signal based on equivalent time sampling

Jingkai Ni

DOI: https://doi.org/10.54097/e890q450

2024-06-30

Abstract:With the continuous development and improvement of communication technology, the research and application of high-frequency signals are becoming more and more important, especially in important fields such as electronic communication, aerospace and aviation. The analysis of high-frequency signals is the most basic and most important. High-frequency digital signals are mainly obtained by equivalent time sampling and sequential sampling. This paper first analyzes the basic principles of equivalent time sampling and real-time sampling, and compares the advantages and disadvantages of the two and the limitations of each sampling method through various indicators. After that, the eye diagram point set based on equivalent time sampling is clustered and analyzed. By comparing the contour coefficients, the sum of squared errors, and the time complexity of each clustering algorithm, the clustering method of the eye diagram point set is further optimized, and the most efficient and accurate clustering algorithm is selected. The clustering algorithm is optimized by increasing the multi-selective convergence threshold based on each cluster center.After many experiments and simulations, from the perspective of various clustering indicators, there is no significant difference between the K-Means clustering algorithm and the K-Mediods clustering algorithm in the case of fewer data points. However, in the case of relatively large data points, the K-Mediods clustering algorithm is more accurate and efficient than the K-Means clustering algorithm. Moreover, compared with the original K-Mediods clustering algorithm, the clustering effect of the K-Mediods clustering algorithm after multi-selective optimization in terms of convergence threshold is reflected in both the sum of squared errors and the contour coefficient. Both have better accuracy.