The Research On The Localization Scheme And The Routing Of WSN

As a new technology of sensing the physical word, wireless sensor networks (WSN) narrow down the distance between the people and the physical word based on the information collected by many sensor nodes, who are the main component of the WSN. Recently, the WSN is widely used in many areas, such as:military, industry, agriculture, environment, transportation and smart home etc.Since sensor nodes in WSN are deployed randomly, to sense the physical word by a large scale of WSN is to obtain the positions of each senor firstly. Based on the scenario of sensors deployed, the kernel of the localization mechanisms of the WSN is to propose a method to compute the localizations of each sensor simply accurately and stably. Meanwhile, the WSN may become paralysis since many sensors whose power are exhausted are died. To design an efficient, energy-efficient and stable routing is the key point to extend the WSN’s life and to improve the performance of the WSN.Many improved localization methods are proposed in this paper based on the classical schemes, such as:weighted center localization (WCL), approximate point in triangle test (APIT) etc. To track the moving sensors accurately, the statistical arguments of the measuring noise and the system noise in the tracking model are estimated on line. After that, a layer based multi-hop dynamic clustering routing algorithm (LMDC) is proposed based on the sensors’ localizations estimated by the schemes above. The simulation results show that the methods proposed in this paper are perfect in theory and suitable to be used in actual word. The main composite contents of this paper are shown as follows:1. A dynamic weighted center localization based on a Gaussian mixed modelA Gaussian Mixed Model (GMM) is established based on the off-line received signal strength indicator measured by the receive sensor nodes and heard from the transmitting sensor. Based on GMM, a new measurement of RSSI is proposed and a dynamic weighted argument of WCL is established. The simulation results show that the new measurement method is better than the log model of RSSI, and the localization accuracy is better than the WCL with a fixed weighted argument. Meanwhile, the average error and the maximum error of the WCL is decreased with the dynamic weighted argument based on the GMM.2. An improved APIT algorithm is proposedAn improved APIT algorithm is proposed in this paper, meanwhile, it is combined with a fuzzy system to adapt to the degree of the irregular (DOI) of the RSSI. The defects of the original APIT are analyzed in two cases in this paper. In one case, the neighbor sensor selected by the unknown sensor is closer to three vertex anchor nodes of a triangle than the unknown. In this case, the original APIT is re-proved and is perfect. In the contrary case, the neighbor is further away from there vertexes than the unknown. In this case, a proper threshold of the distance between the neighbor and the unknown is provide to decrease the errors of the original APIT. The simulation results show that the improved APIT is better than the original APIT in many testing scenarios, and the new scheme combined with improved APIT and the fuzzy system has good performance with the different DOI.3. An improved time-lag unscented Kalman filter (UKF)An improved time-lag unscented Kalman filter (UKF) is proposed to track the moving sensors. The statistical arguments of the measurement noise in UKF are estimated by wavelet transformation (WT) on line, and after that, the statistical arguments of the system noise in UKF are estimated by the Sage-Husa filter. Meanwhile, a criterion of the divergence trend of the new UKF is proposed and the correction method of the new UKF is proposed. 4. A layer based multi-hop dynamic clustering routing algorithmA layer based multi-hop dynamic clustering routing algorithm is proposed based on the localization information of each sensor. First, a multi-layer structure of the WSN is formed based on the distance between the source sensor node and the Sink node. Second, the cluster area of each layer is formed by the K means. Third, the cluster head in the first layer and the other layer are selected by the different methods. The simulation results show that the routing proposed in this paper delay the whole network’s life and improve the performance of the network.5. A person localization systemA person localization system is established base on the WSN in this paper. A schema of the localization system is designed based on the special terrain of the tunnel of a mine. The hardware of sensor nodes is composed by ZigBee CC2430, and the software of the localization system is deployed by C# based on the geography information system (GIS). The localization schemes of the system are realized by the methods discussed in chapter 2 and chapter 4. In the actual localization system, the localization schemes proposed in this paper are proved good.