Research On Real-time Data Processing In Perception Layer Of The IoT

With the development of science and technology, the Internet of Things(IoT)apply in the social production and human life more and more widely. Many practicalnetworks have appeared in smart home, intelligent power and industrial control field.Most of these networks require strong instantaneity, the real-time standard ofnetworks determines the overall performance of the system.However, in many networks of the Internet of things, the huge amounts ofterminals produce GB even TB perception data minutely. If we transfer and use thesedata without a certain processing, the network will produce a large number ofcollisions, seriously affect the real-time performance.At this stage，the network data processing mainly include data fusion, datacompression, data filtering and cloud processing several branches, domestic andforeign experts put forward many methods，and obtain a certain effect. But thesemethods can’t improve the real time standard according to the characteristics of theInternet of things, can’t reduce the total amount of data in the network effectively. Theexisting "Decentralized Perception-Data， Transmission-Decision， Process"architecture has been unable to meet the demand of practical application.Under the environment of Internet of things, the networks have differentstructures and different functions, so improve the real-time performance can’t treat allnetworks the same, must carry out a reasonable classification. Some networks are setup in order to achieve the established goals continuously, some networks are set up inorder to keep a certain environment under a fixed condition. This thesis named thesetwo kinds of networks Enterprise Networks and Maintain Networks. Compare to theEnterprise Networks, Maintain Networks mostly simple, easy to establish a unifiedmodel for research, and also have high demands on real-time. Therefore, in this thesis,in order to improve the real-time of the Maintain Networks, move key point of thedata processing to perception layer. The main works are as follows.(1)Introduce the generation, development, application field and the core technology of the Internet of Things, summarizes the characteristics of the data in theInternet of things. Study the existing data processing scheme of the Internet of things,analysis the advantages and disadvantages.(2)The existing data processing method of the Internet of Things put moreemphasis on the network layer and application layer, unable to effectively solve theproblem of large amount of data, transmission delay, data collision, and poorreal-time in the Maintain Networks. In this thesis, we based on the real-time demandof the data, put forward a data processing model in perception layer. First of all,classify the Internet of Things isomerous terminals, set up a layer-cluster model.Then, analysis of the node error distribution regularity of sensory data, setting theerror threshold, Define events and response rules, implement the simple eventreal-time response, and the complex events quick upload.(3)In order to meet the demand of real-time in data processing model, based onthe no beacon mode of the IEEE802.15.4MAC layer, proposed an improvedLN-MAC protocol. By classifying data, introduce channel unfair competitionmethods, improve the real-time data’s ability of capturing channel. This protocol don’tchange the original way of channel competition, only collocate the real-time data andnon-real-time data with different transmission parameters. This protocol do not addany communication burden, at the same time, improve the quality of real-time data’stransmission, control of the redundancy, reduce the amount of data in the network. Byestablishing the protocol’s markov chain of the nodes’ buffer queue, analyzes twokinds of data’s channel competition ability.(4) Instructions the data processing model under the condition of smart home,discuss the response of different types of events. To ensure the real-time datatransmission of LN-MAC protocol under the NS2simulation environment. Comparethe transmission performance of the real-time data and non-real-time data. Throughthe comparison with other similar MAC protocols, confirmed the validity of theprotocol.