Design,Optimization And Acceleration Of Indoor Positioning System On Mobile Computing Platform

In recent year,Indoor Positioning System has become a popular topic in research as Location Based Services are playing a more and more important role in people's daily life while Global Positioning System is unable to provide accurate position information indoors.The topic of this dissertation is Indoor Positioning System.Study and explorations are made on drawbacks of existing research based on a comprehensive literature review.This research chooses smart mobile devices as platforms because of its popularity and designs a hybrid Indoor Positioning System with features of multi-mode,low-cost and high accuracy.The most significant feature is multi-mode where the mode of how smart devices are carried when using the Indoor Positioning System can change freely in all supported modes.Moreover,the low computational complexity and deployment cost of the designed Indoor Positioning System make it more suitable for easier deployment on mobile flatforms with limited computation and energy resources.In an experiment with length over one kilometer,the position error of the user's last location is only 1.36m.In addition,new optimization and acceleration algorithms are proposed in this dissertation for Particle Filter,which a commonly used algorithm in Indoor Positioning Systems with significant hardware overhead and computational complexity,to cutdown the demand for memory resources and computation time,as well.The high hardware efficiency and real-time capability of the proposed algorithms are proved by theoretical analysis,simulation and prototype design on Field Programmable Gate Array.In case study of the proposed algorithm,the algorithm with low memory consumption can reduce up to 49.69%of memory without sacrificing filter accuracy and the algorithm with high real-time capability is able to accelerate filtering process by 38.62%when compared to linear speedup of generic distributed particle filters.