Study On Environmental Change Events Aware Method For Internet Of Things Application

The Internet of things (IOT) is the third world information industry tide following the Internet. According to the agreed protocols, it can connect any items sensed and controlled to the Internet for accomplishing intelligent identification, location, tracking, monitoring and management through various information sensing devices. As an important part of the IOT, Wireless sensor network is largely responsible for obtaining and acquiring information and contributes to improving the sensing ability of the IOT system to the surrounding environment. Recently many researchers convert the monitoring data obtained by sensors to context as the environment information which will be provided to Internet of Things application. But the running of the Internet of Things application system depends not only on the state of the environment, but also the change of the environment state, that is. environmental change event. Therefore, how to process raw monitoring data obtained by sensor to form reliable and effective environmental change event and accomplish efficient information awareness is the important problem to be solved in IOT.To address this issue, this thesis established an approach of environmental change events aware for IOT application. Based on the idea of information fusion, this approach can aware predefined environmental change events according to monitoring data obtained by Wireless Sensor Networks. Some environmental change events are related to some certain parameters of environment. But in fact many environmental change events are more complex. Their occurrence is also related to other events, that is temporal or causality relationship between events. Analyzing of the correlation between environmental change events contributes to improving the efficiency of awareness. In this thesis, environmental change events are divided into three categories according to their characteristic. And then different approach is adopt for different category. Firstly, to the environmental change events which lead to changes in a single environment parameter, we adopt the approach based on rules discrimination which can discriminate environmental change events according to real-time environmental state data sequence and rule curve fitting degree. Secondly, to the environmental change events which leads to changes in several environment parameters, we adopt the approach based on BP neural network which can be trained by history data set to get corresponding aware model based on BP neural network and then environmental change events can be aware in accordance with real-time environmental state data sequence. Thirdly, to the composite environmental change events which lead to changes in several environment parameters, we adopt the approach of Dynamic Bayesian Network whose structure is built according to the expert experience and parameter is learned through history data set, thus used for real-time environmental change events aware. Finally, this thesis designs and realizes these three approaches of environmental change events aware separately, analyzes and evaluates the approach from the accuracy and data size to verify the effectiveness of the proposed approach.