Font Size: a A A

Research On Mobile Node Localization Algorithm In Wireless Sensor Network

Posted on:2012-07-19Degree:MasterType:Thesis
Country:ChinaCandidate:L K ZhaoFull Text:PDF
GTID:2218330338968156Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
Localization technology for unknown nodes in Wireless Sensor Network is one of the key technologies. Localization is determining the position coordinates for unknown nodes, it could be divided into static localization algorithm and dynamic localization algorithm according to whether nodes mobile. As the development of technology, static localization algorithm in WSN has been improved increasingly, but because of the great complexity of the dynamic localization algorithm, the development of localization for mobile nodes in WSN are lagged behind, and there are still many deficiencies.This paper focuses on the localization technology for mobile nodes in WSN. For now, the algorithms for mobile nodes mainly are MCL (Monte Carlo Localization) and MCB (Monte Carlo Localization Boxed) which is the classical improved algorithm. MCB is due to the significant increase efficiency of the algorithm and improved localization accuracy in localization for mobile nodes, it has become one of the main algorithms in dynamic localization algorithms.Through further study and research MCL algorithm and MCB algorithm, this paper found two defects that exist in the MCB algorithm:1. When unknown nodes detect the number of anchor nodes is insufficient, the positioning error too large to accept;2. When mobile nodes detect the number of anchor nodes and hop size always keep consistent, the positioning errors will accumulation due to the algorithmic characteristic.In order to solve these two defects, this paper proposed corresponding improved algorithms: MCBPA (Monte Carlo localization Boxed use provisional anchor node) algorithm and MCBMPF (Monte Carlo Localization Boxed use multi-Point average forecast) algorithm.MCBPA algorithm used unknown nodes that they have high positioning accuracy upgrade to provisional anchor nodes. When unknown nodes detect numbers of anchor nodes are low, and receive the broadcast informations come from the provisional anchor nodes, MCBPA algorithm can use provisional anchor nodes to assist unknown nodes localization, and finally achieve the purpose of improve positioning accuracy. Simulation shows that localization accuracy of unknown nodes which they use MCBPA algorithm have improved, and the positioning accuracy increase along with numbers of incremental mobile nodes in the network.MCBPAF algorithm firstly predict coordinate of unknown nodes at the next moment, secondly reconstruct the bounding box center on the predicted location, and further reconstruct the sampling box. MCBPAF reduced the positioning error effectively, because of the more accuracy sampling box. Mobile nodes produce mobile trajectories in CRW (Continuous Random Walk Mobility Model), this model simulate movement of mobile nodes according to the Newtonian mechanics principle. The trajectory that was produced by CRW and the real trajectory are similar, both of them have certain regularity and greatly randomness, so CRW could simulate movement of nodes successfully. Due to unknown nodes can only estimate own location through the localization algorithm, but there are errors between estimated position and actual position, it reduced the forecasting accuracy drastically. MCBMPF algorithm, in order to ensure the accuracy of predict, use multiple estimate positions before next localization. MCBMPF algorithm use mathematical expectation of the multiple estimate positions simulate mathematical expectation of the real positions, and use the standard deviation of multiple estimate positions simulate the standard deviation of real positions. Through calculate the multipoint standard deviation of unknown nodes, MCBMPF algorithm estimate the speed of mobile nodes at the next moment. The simulation results show that the method improved the accuracy of nodes localization and sampling efficiency greatly.Finally, MCBPAF (Monte Carlo algorithm: use Provisional node localization Boxed percentage Forecast) was put forward on the basis of MCBPA algorithm and MCBMPF algorithm. MCBPAF algorithm construct anchor box based on MCBPA algorithmic rules and construct boundary box based on MCBMPF algorithmic rules. In the end, create higher precision sampling box. Simulation shows that algorithmic accuracy for mobile node localization is improved significantly, and the algorithm has better adaptability in most simulation environment.
Keywords/Search Tags:WSN(wireless sensor network), mobile node localization, MCB(Monte Carlo localization boxed), provisional anchor node, multi-point average forecast, continuous random walk mobility model
PDF Full Text Request
Related items