A Clustering Algorithm in Ultra-Dense Cellular Networks

Abstract

The invention discloses a clustering algorithm in an ultra-dense cellular network. According to the method, a set division process based on interference weight among users comprises two phases: an interference weight generation phase and a user clustering phase. In the interference weight generation phase, the users generate interference to other users and also bear the interference from the otherusers, so a weight balancing policy is employed, namely, through adjustment of useful signal weight design of the users and through adjustment of interference coefficients of the users to the other users, power utilization dimensions and space occupation dimensions of the users are obtained as gain when cooperative transmission is carried out. In the clustering phase, an improved KNN clustering algorithm is provided. According to the algorithm, the users are clustered according to a criterion of minimizing the interference among the users in the clusters. The users in the same cluster share spectrum resources. A utilization rate of the spectrum resources and system edge throughput are improved.

Description

technical field [0001] The invention belongs to the technical field of mobile communication and relates to a clustering algorithm, in particular to a clustering algorithm in an ultra-dense cellular network. Background technique [0002] With the popularization of networking and digitalization, breakthroughs are needed in the following aspects: wide-area coverage, high-capacity hotspots, large connections with low power, and high reliability with low latency. These developments naturally have higher requirements for segmenting user traffic and allocating communication traffic. [0003] Ultra-dense networks consisting of many low-power small cells represent the most innovative and transformative means of addressing this challenge. However, the interference problem caused by dense cell deployment will reduce network capacity and user experience, resulting in reduced spectrum utilization and cell edge throughput. [0004] In order to solve these problems, it is necessary to de...

AI Technical Summary

Problems solved by technology

However, coordinating among all cells in the network is a very complex task, because precise synchronization is required in the coordinating cells, which incurs additional navigation and signal processing overheads, in order to avoid these overheads, the It is necessary to carry out small-scale cooperative clustering on the cell, that is, the users of the cell can be clustered first, and then the cooperative transmission can be carried out within and between clusters

Most of the clustering algorithms proposed now have their own shortcomings. Most of the existing technologies are difficult to implement and difficult to understand, requiring a large number of estimated parameters and experimental verification; the existing technologies are easy to analyze and process common events, and lack the classification of rare events. The ability to deal with it; the existing technology has poor ability to deal with multi-classification problems

Images