Subspace Localization Algorithm Based On Optical Waveguide Sensor

Sensor has played a significant role in many fields of current society, especially in wireless communication, military reconnaissance and electro-magnetic compatibility. Traditional sensors of electric field have some disadvantages in applications. Thus, integrated optical waveguide electric field sensor using advance integrated optics technology can be adopted, which makes the whole measurement system super-small, steady and reliable, is characteristic of broad bandwidth, immunity to electromagnetics and tiny perturbation. They can be interconnected by optical fiber so that remote control and monitoring can be realized. At present, the electric field sensor is the one of research focus using integrated optics technology based on LiNbO₃ substrate. Therefore, integrated optical waveguide electric field sensors are developed in this paper, by analyzing the structure of the electric field sensor with segmented electrodes and establishing the approximate mathematical model of sensor signal.The implement of sensors in localization area has alrealy been a hot research direction in recent years. In this paper, two main branches were discussed: sensor array localization and sensor networks localization. There were some novel algorithms proposed for each branch.First, this paper focuses on a seriel of problems in the near field localization. Current research situation of sensor array localization was present, and some classical localization approaches were also given. Based on the symmetric property of sensor array, afar field approximated localization algorithm was proposed; for further improving the accuracy, we introduce the alternative projection scheme between DOA and range parameters into traditional weighted subspace fitting algorithm.Wireless sensor networks, which integrate technologies of MEMS, sensing, embedded computing, distributed information processing and wireless communication, lead to a new pattern of information gathering and processing. The fundamental problems of sensor networks are deployment and coverage, localization, and networking protocols. And localization, which has been focused as a basic problem, is the precondition of many applications of sensor networks, such as target identification, surveillance and tracking. In this thesis, we discuss the self-localization technology in detail. Since the nodes are restricted by size, cost, energy, and work under the environment that should not be accessed, the basic approach is to use beacons that aware of their positions as reference, and the other nodes compute their positions through the relationship of distance, angel, or network connectivity to the beacons.This dissertation analyzed the subspace algorithm based on multiangulation informations, and gives simulations to examine the localization performance with the parameters varying; then a improved normalization subspace method was proposed based on range measurement subspace method, meanwhile, the reason of localization improvement was also introduced, finally, the stability problem of this novel algorithm was solved by introducing a protection range strategy.