Research On Matrix Based Localizationin Mobile Wireless Sensor Networks

Wireless sensor networks(WSNs) consist of static or mobile sensor nodes. These nodes can communicate with each other to perform sensing and data processing cooperatively in ad-hoc manner. Wireless sensor networks have many applications such as environmental monitoring, tracking, traffic control, human health monitoring, and disaster relief. Since the many applications of wireless sensor networks depend on the location information of sensor nodes, how to get the accurate coordinates has become the focus of attention. GPS or manual deployment for all nodes will be network issues such as cost, position power, network scalabilit y limitations, even impossibility in some cases. Therefore, we must adopt certain algorithms and mechanisms to locate the sensor itself. Esisting node localization algorithms of wireless sensor networks still subject to many problems, such as greate impacted by circumstances, large positioning errors, high algorithm complexity and huge energy-costs, unsuitablity for mobile sensor networks. With the development of wireless sensor network technology and the expansion of applications, node localization technology has great theoretical and practical significance.In this thesis, we present an innovative localization scheme based on matrix completion, MALL, that utilizes the collected connectivity and distance information to achieve high-precision localization. Since MALL only involves convex optimization and low-complexity non-convex optimization, it can localize mobile nodes at a very fast pace. MALL only makes use of the information from anchor nodes within two hops and normal nodes within one hop. Consequently, MALL leads to satisfactory localization precision with low communication cost. An in-depth analysis of the time complexity and communication cost of MALL is also included. Through intensive simulation, we found that MALL outperforms the state-of-the-art localization schemes.