Font Size: a A A

Research On High-Accuracy And Robust Localization Techniques Based On Magnetic Field And Multmodal Fusion Algorithm

Posted on:2015-04-29Degree:MasterType:Thesis
Country:ChinaCandidate:R TaoFull Text:PDF
GTID:2298330467463305Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
With the development of mobile Internet, location based service is changing our lives. Precise positioning, as one of the core LBS services, gradually become the research focus in the field.In this paper, the research focused on indoor positioning technology and theory, based on existing research. This paper proposes a multimodal indoor positioning algorithm, and on the basis of the algorithm, implements a positioning system so called ICTIPS and a LBS application system. The algorithm make full use of multiple sensors on the smartphones, including WiFi signal strength measurements, magnetometer readings, acceleration readings, gyroscope readings, compass, etc. And built a pedometer module, a dead reckoning and corner detection module, a step length adapting module, a magnetic matching module, a confidence module, and finally the particle filter framework is used to achieve the integration. The main focuses of this paper is geomagnetic signal positioning method. The algorithm includes two stages of off-line training and on-line positioning. In the offline training phase, WiFi fingerprints and magenetic fingerprints are collected, which are being used in the online positioning phase. In online positioning phase, the WiFi positioning result is used to initialize our multimodal algorithm, and then the tracking part of the algorithm is driven by the pedometer module, using the magnetic characteristics to get particle weights updated, using map information to control particle movements and using confidence analysis to monitor state of system, and updating the estimated position of the user in each round. The contributions of this dessertation are listed as follows. Firstly, studies on geomagnetic information characteristics of the indoor environment is carried out, a standard of the fingerprinting is proposed, magnetic fingerprint matching algorithms are compared and analyzed. Secondly, this paper proposes a corner detection, step length adapting method. Thirdly, this paper puts forward a method of confidence analysis and location correction strategy. Finally, the multimodal algorithm uses all available sensors, and provides a robust and reliable positioning services.
Keywords/Search Tags:Particle Filter, Magnetic Localization, Dead Reckoning, Multimodal
PDF Full Text Request
Related items