Spectrum Allocation Scheme for Intelligent Partition Based on Machine Learning for Inter-WBAN Interference

摘要

A wireless body area network (WBAN) is a physical network system with functions such as information collection and data transmission, and provides a key means for real-time health detection and medical services. However, if WBANs deployed in a dense environment lack internetwork coordination, it can cause high interference. In addition, WBANs have the characteristics of social mobility, and regular frequency changes will increase energy consumption. In this article, the co-frequency interference between coexisting WBANs in dense environments is analyzed, and the real-time frequency of mobile WBANs is studied emphatically to alleviate these problems. A spectrum allocation scheme for intelligent partition based on machine learning is proposed as a mixed scheduling scheme that combines graph coloring and partitioning ideas. WBANs' self-organized dynamic clustering forms multiple cells according to location, and the cells use a vertex coloring algorithm of self-generating topological graphs to coordinate their frequency bands and assign different colors to coexisting cells. When a new user joins the area, it will self-organize and cluster to a divided cell, and the cell will divide frequency bands for the new user. No other bands need to be changed, and the experiments conducted in this research demonstrate that the proposed arithmetic can adapt well to the fast-changing WBANs of the topology, and the system anti-interference performance, frequency resource utilization, and system stability are favorable.