Research On Localization And Tracking Based On Wireless Sensor Networks

As a new hot field, wireless sensor network (WSN), which integrates thetechnologies of sensor, microelectronics, wireless communication and distributedcomputing, is a highly concerned technology in the academia and industry as itsadvanced idea and broad application prospect. And it has become a hot research topic inthe current information technology field. WSN has the advantages of low-cost, smallvolume, flexible networking and flexible deployment, it can be used widely in military,target tracking for battlefield, anti-terrorist, rescue, environmental monitoring, medicalcare, space exploration, traffic management, agriculture, and so on.Sensor node localization and target tracking are two vital problems in the researcharea and application area of WSN. Based on the existing localization mechanisms andtracking mechanisms, the sensor node localization and target tracking applicationtechnology in WSN were analyzed from the views of improving precision, reducingenergy consumption, prolonging the network life based on the Rang-free theory,coordination theory, particle filter and other computing methods. This dissertation seeksbreakthroughs on research methods and ideas. The major contribution of thisdissertation is specifically stated as follows:①Localization is the basis and prerequisite for target tracking. In order toaccurately locate and track the target, the sensor node which involves in localization andtracking must be aware of its own position. For the location problem of sensor node inWSN, we proposed a RSS-assisted Range-free Mobile Beacon Localization algorithm(RRML). In the network, the mobile beacon node moves around the sensing field, andperiodically broadcasts its own position. In stead of the traditional position fixed beaconlocation method, sensor nodes compute their positions by using the virtual beaconpoints. RRML algorithm allows a sensor node to locally compute its position by usingonly three received mobile beacons. First, the sensor node chooses the first and lastmobile beacons of the visitor list as beacon points. According to the geometricreasoning and RSS between sensor node and the two beacon points, the positionalrelationship between the beacon points and the sensor node’s communication circle isdetermined, and the radius parameters for computing the sensor node are adaptivelyadjust. Then, the two candidate positions of the sensor node are calculated based onsimple geometric constraints. At last, the third beacon point is chosen to determine the accurate position of the sensor node. Meanwhile, the upper limit of the localizationerrors is carried out, and also we analyze the main impact factors of the localizationerror. The simulation results show that the RRML algorithm has a good localizationaccuracy, scalability, and robustness.②The essence of target tracking based on WSN is a collaboration process of thesensor nodes. Through mutual cooperation and coordination between the nodes, aneffective node management mechanism can be established to achieve target tracking.On the target detection stage, due to the uncertainty and unpredictability of real-worldtarget’s motion, we proposed an energy efficient sensor activation protocol based onpredicted region technique. The predicted region sensor activation algorithm (PRSA)predicts the moving region of target in the next time interval instead of predicting theaccurate position, and decreases the missing rate caused by the traditional predictionschemes. Then an activated strategy is established to activate the fewest essentialnumber of sensor nodes within the communication intersection region, to monitor thetarget’s location in the next time by analyzing the position relation between the targetand the nodes. The proposed algorithm can reduce the number of nodes which involvedin tracking the target, to prolong the network’s operational lifetime. Simulation studiesshow that the proposed algorithm provides significant energy savings and prolongs thenetwork’s operational lifetime. On the target tracking stage, we proposed a dynamictracking cluster construction strategy based on node cooperation mechanism. Thedistance and the residual energy of sensor node are taken as constraint conditions, tobuild tracking cluster to guarantee network efficient clustering. It can improve theefficiency of target tracking task of the network.③Combining sensor activation algorithm and dynamic clustering algorithm, aParallel Extended Kalman Particle filter (PEPF) algorithm was proposed for targettracking. Based on the dynamic cluster model, particles are divided into multiplesubsets by the cluster head (CH), which run in parallel on the member node. Thecomputed result of each member node will be fused on the CH, so we can get theoptimal state estimation of the target. There is no information exchange among themember nodes, only the member nodes and the CH have information exchange, whichcan reduce the energy consumption of communication. Due to the particle filters areparallel executed on each sensor node, the proposed algorithm can improve the particlefiltering efficiency and avoid energy excessive consumption on one sensor node, andbalance the energy consumption of the network. Meanwhile, in order to avoid the traditional methods select the transition probability density as the importance densityfunction, which make the particles seriously depend on the system state transition model,Extended Kalman filter algorithm is used to generate the importance density function ofparticle filter, which can make the sampling of the important density function moreclose to the sample of the posterior probability density. The proposed algorithmimproves the efficiency of the particle filter, and avoids particle degradation problems.The simulation results show that the proposed algorithm which has high trackingaccuracy can predict and track the moving target well, and effectively balance energyconsumption of the network.④Considering battlefield target tracking and intrusion detection of specific regionsapplication environments, a localization and tracking prototype system based on WSNwas built. For the localization problem in WSN, the localization algorithm based onmobile beacon was designed and implemented. Then, combining with nodecollaborative algorithm, target tracking prototype system was built, which can lay thefoundation for further application of localization and tracking technology in WSN.