Wireless Sensor Networks’ Energy Saving Research Based On Data Fusion In Building Environment

With the the rapid development of sensor detection technology,wireless communication technology, data processing technology and automation control technology,Wireless Sensor networs(WSNs) came into being and are playing a more and more important role. While used in building environmental monitoring,the network will generate plenty of redundant data that will deplete the network’s energy and decrease the lifetime of the whole network. Therefore,it is necessary to study on data fusion algorithm for wireless sensor networks(WSNs) used in the scenario of building to reduce the amount of data transmission in WSNs, cut down the system energy consumption and operating cost with reflecting the environment more accuracily.Using a laboratory in the school as the research object,the thesis works on the data of temperature and humidity fusion. The network is clustered based on geographic location and cluster head nodes act as routers,relaying fused data to the network coordinator for energy saving in the intelligent building’s query and monitoring. Using two levels of data fusion can provide high performance in monitoring accuracy and decrease energy consumption at the same time. PC monitors nodes collecting data based on IOTV2 hardware platform and IAR, Visual Studio software development platform, while doing research and development of communications between Cortex A8 boards and PC to achieve monitoring environmental parameters.For an assigning range of WSNs with uniform nodes collecting data, non-uniform clustering area is optimized by Particle Swarm Optimanation algorithm whose global optimazation objective is to prolong survival life cycle.The WSNs is divided into clusters based on the nodes’ geographic information. Matlab shows the method can effectively increase the life cycle of the network with less communication energy consumption as while as keeping the coverage and connectivity of the network.The data fusion algorithm in single cluster is studied based on clustering results. In order to build a better data fusion system,BP neural network is used to optimize fuzzy prediction and train the membership degree of collecting data which are training sample got from the monitoring platform. The data fusion based on BP neural network can guarantee high quality while decreasing the quantity of sensoring data.Two kinds of different parameters used to train BP neural network are proposed and compared with precision of the fuzzy prediction which are implemented for better precision and less energy consumption.The input training sample of the mentioned data fusion algorithm is collected based on the temperature and humidity monitoring platform for wireless sensor network. In the paper, the WSNs for building environment is integrated into OURS platform and Collect environmental parameters.Environment temperature and humidity monitoring platform of WSNs is built based IOTV2 Things chamber and IAR, Visual Studio software development platform. Data is collected and transfered to host computer for environment parameters monitoring, control, preservation, analysis and other functions. The input training sample of the mentioned data fusion algorithm is collected based on the platform and simplified fusion algorithm is embedded into node. Cortex A8 development board and the host computer can be designed to communicate in LAN.