Study on the Quality of Localization Guarantee Strategy in Wireless Sensor Networks
|School||Wuhan University of Technology|
|Course||Traffic Information Engineering \u0026 Control|
|Keywords||The inconsistencies degree (DoI) Particle filter Node Localization Wireless sensor networks|
With modern sensor technology, intelligent technology, MEMS technology, embedded computing technology, wireless communication technology, the rapid development of wireless sensor networks, new information acquisition and processing technology in the real world become more realistic , so the technology has been in academia, industry, and government departments attach great importance. One of the main functions of the wireless sensor network sends the collected environmental information to the user, or for processing the collected information, make decisions whether there is interest in the event invasion and send the results to the user. The lack of position information for the vast majority of applications of the wireless sensor network, the data is meaningless, such as in the event detection applications require not only able to detect the event, but also need to be able to obtain the exact position of the event; in target tracking applications , you must first know the location of the node to target real-time location updates. Therefore, the need to solve the basic problems of the wireless sensor network node localization problems become one. Low on standard static network node density, iterative positioning (ie unknown nodes can own position estimation) can effectively improve the network node localization rate neighbors already positioning node, but if iterative positioning error propagation problem has not been well solve a single node positioning errors will soon affect the entire network, to cause positioning accuracy deteriorated sharply, leading to the node localization quality can not meet your application needs. To address this problem, propose an iterative positioning algorithm based on particle filter and control strategies based on inconsistent degree of error. First, to propose a particle filter positioning algorithm, the unknown node uses a series of sample particles and their corresponding weights to achieve position estimation. These particles are updated according to the posterior probability density of the node position gradually. In addition, according to the uniform distribution of the characteristics of wireless sensor network node randomly proposed called inconsistent degree (Degree of Inconsistency, DoI) the concept of positioning error is larger node will be gradually identified inconsistencies degrees and their location updating, in order to improve the positioning accuracy, and inhibit the accumulation of errors. Mobile wireless sensor networks, a new mobile Marke Gauss Markov walk model under the premise of Free Ranging node localization algorithm. Estimate algorithm based on the centroid of a polygon formed by the intersection of the boundary line Marke enter and leave the communication range of unknown nodes forming two or more sets in the node position, so that computational complexity is reduced, and the algorithm can also be in the accuracy does not satisfy requirements by adding the positioning method of a boundary line, the location updating, to achieve the improvement of the positioning accuracy. The simulation results show that the increase in the number of boundary line group formed as Marke mobile positioning accuracy is also rising.