Research On Mobile Nodes Localization Algorithm In WSNs

After monitoring an event, an important problem is the location where it occurred. Location information is an indispensable part of datum collected by sensor nodes, which means that monitoring information is usually meaningless without the location information. Although localization algorithm used in WSNs has been researched in many previous works, there are only few of them that focus on mobile nodes.Firstly, this dissertation introduces the structure, key technology, characteristic and applications of the wireless sensor networks, and the concept, classification and evaluation of the localization of nodes in the wireless sensor networks. Then this paper describes the existing localization algorithms for mobile nodes in the WSNs, finally, focuses on the RSSI-based localization algorithm for indoor environments.Localization algorithm for indoor environments is a challenging issue concerning WSNs based location systems. This thesis present a novel method for indoor-location estimation using a mixture strategy. The motive is to obtain accurate coordinates without incurring too much human calibration. In addition, based on manifold learning, this thesis improves the probabilistic method and presents a semi-supervised algorithm to reduce data collection and processing workload in fingerprinting method. The experiments show this algorithm is robust in indoor environments and can get relatively accurate coordinates with little labeled RSSI vector. In other words this algorithm is a low-cost and efficient algorithm. At the end of the thesis, this algorithm is used to tracking a mobile toy car in indoor environments, the result shows, with calibration, the performances of the probabilistic semi-supervised algorithm is relatively accurate.