Research On Heterogeneous Network Switching Technology For Smart-home

The concept of smart-home travels to China at the late 1980s.In recent years,with the acceleration of 5G commercialized operation and Internet of things,everything can be and will be connected,smart-home network gets more and more attention by house-owners and enterprises.This work bases on the cooperation project "Smart-home oriented heterogeneous adaptive wireless network research" with Hisense Co.Ltd,while this paper focuses on the designation of the heterogeneous network switching technology.The smart-home heterogeneous network will be a LTE,D2D and WIFI co-existence scenario,with the purpose of realizing the smart-home network all-coverage and seamless switching,this paper is going to choose decision factors and design the switching process.This paper takes the received signal strength,network available bandwidth,network fees,signal stability,user preferences and user's location into consideration during the selection of decision factors.Take these six factors as the input of the decision system.We design the"preliminary decision aided fuzzy judgment" algorithm as the fundamental basis of selection of heterogeneous network.The algorithm firstly conduct some candidate networks by preliminary decision,which can greatly reduce the unnecessary switching called "ping-pong effect";and then by the selection of fuzzy judgment to determine the optimal network,which can improve the balance of the network.Concerning obtaining user's location,this paper adapted the location fingerprint method of WIFI indoor positioning technology by combining KFCM clustering and machine learning algorithms together,and proposed some optimizations upon clustering amount and kernel parameters,which greatly improved the processing speed and the accuracy of the indoor positioning and provided a more powerful safeguard for switching decisions.