Study Of Coverage And Localization Methods Based On Fuzzy Information Processing In Sensor Networks

Sensor network is an innovative information acquisition system in an unmanned surveillance region, which plays an important role in critical infrastructure security, counter-terrorism and explosion-proof in large-scale social activities, security detection in high-risk industries and anti-aircraft defense, etc. Deploying sensor networks is a very important method for regional information acquisition, which can be developed to achieve regional close surveillance, reconnaissance, target localization, identification and tracking, which isIn a practical large scale sensor network, the resources of a sensor node such as the energy, processing, communication, and sensing capabilities are constrained. Moreover, there are limited cognitive abilities for human; the network topology is changing dynamically, inexact values of reference standards or parameters are always used in the model. As a result, there would be some fuzziness and randomness for information acquisition and network topology in sensor networks, where the acquisition of information is not perfect. In detail, there become some uncertainty in sensor location and sensing ability. Therefore, the network information cannot be obtained in a precise manner. There exist significant limitations of handling these uncertainties based on traditional geometric analytical methods in sensor networks.In this context, this paper focuses on the uncertainty of coverage and localization for target detection, localization and track oriented sensor networks. A new analysis approach of fuzzy information processing has been proposed. The fuzzy metric for uncertain information is proposed in ad-hoc sensor networks, and its fuzzy geometric property is also studied. The purpose of the paper is to do some theoretical discussions to handle the uncertainty issues for the construction of practical sensor networks. This paper could be divided into six chapters, which are organized as follows:In Chapter 1, the research background and significance of this dissertation, sensor network and its characteristics are briefly described. The current research of sensor networks, coverage problem, and localization issues is summarized. Finally, the main achievement and arrangement of this dissertation are concluded.In Chapter 2, based on fuzzy plane geometry theory, an analysis approach of fuzzy high-dimensional geometry is proposed, including fuzzy point, the distance between two fuzzy points, fuzzy vector, fuzzy line and fuzzy hyper-conicoid in Rn (n≧1) space, which properties are also discussed. Based on the approach above, a fuzzy geometric bearing-only target localization algorithm for fuzzy line observer trajectory, fuzzy triangle observer structure and regular fuzzy ellipse observer trajectory in passive sensor networks is proposed. A fuzzy geometric bearing-only target localization algorithm for a moving target in R3 space is also proposed, which can realize the four-dimensional localization including fuzzy estimate coordinate and speed of the target by measuring the fuzzy azimuth angle and the fuzzy elevation angle at intervals of fixed time. In Chapter 3, the basic concepts and properties of Sugeno measure are described.Based on the theory, an information coverage analysis approach based on Sugeno measure in sensor networks is also proposed. A perimeter coverage analysis approach by the use of the interaction between two nodes is provided. The analysis of information coverage for cellular-model deployment in sensor networks is also provided. Finally, an analysis approach for best-case fuzzy information coverage path in sensor networks based on fuzzy measure is proposed.In Chapter 4, the definitions and properties of gλrandom variable on Sugeno measure space are described. Based on the theory, the differential geometry of statistical model proposed by Amari is generalized from probability space to Sugeno measure space, and a new notion of fuzzy manifold based on Sugeno measure is proposed, including the fundamental differential-geometrical structures of statistical models, the tangent space, and the Riemannian metric in a fuzzy manifold. The algorithm of the signal-strength matching localization based on Sugeno model in sensor networks is proposed.In Chapter 5, the model of fuzzy information coverage in sensor networks is proposed, the membership function and fusion operator for coverage intensity are defined. Based on this model, the analysis of fuzzy information coverage for deterministic and random deployment in sensor networks is provided. The analysis approach of higher order fuzzy information coverage for sensor networks is proposed. The analysis of best-case and worst-case fuzzy information coverage in sensor networks is also provided.In Chapter 6, a summary of the work of the whole paper is given and a prospect of the issues concerned in this field is also presented.