Trajektooride taastamine ja inimeste liikumise mustrite analüüs mobiiltelefoni andmete põhjal
Files
Date
2017
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Tehnoloogiad, mis kasutavad geograafilisi andmeid, on muutunud meie igapäevaelu tähtsaks osaks. Tänu sellele on kasvanud asukoha andmetemassiliine salvestamine ja kaevandamine. Seni on GPS tehnoloogiad olnud põhiliseks geograafiliste andmete kogumismeetodiks. Sellega paralleelselt on populaarsust kogunud mobiiliandmete kasutamine positsiooni tuvastamiseks ja liikumismustrite analüüsimiseks. Mobiiliandmete (CDR) põhjal trajektooride taastamiseks on vajalik meetodite kohendamine selleks, et tulemused oleksid korrektsed. Tänu sellele, et telekommunikatsiooni ettevõtted on alustanud suuremat koostööd ja hakanud CDR-andmeid järjest rohkem avalikustama, on mobiiliandmete kasutamine mitmetel aladel suurenenud. Töödeldud mobiiliandmed aitavad anda ülevaadet rahvastiku liikumisest erinevates ulatustes. Samal ajal on trajektooride taastamine CDR-andmetest kohati raskendatud võrreldes GPS-andmetega. Suurimaks probleemiks on algus- ja lõpp-positsioonide asukoha määramine, mis on veelgi enam raskendatud juhul kui objekt liigub.Selle lõputöö eesmärgiks on trajektooride taastamine anonüümsete kasutajatepoolt genereeritud CDR-andmete põhjal. Tulemuste valideerimine GPS-andmetega, mis on loodud paralleelselt mobiiliandmetega ning on vajalik selleks, et määrata saadud trajektooride täpsust. Loodud trajektoore saab kasutada objektide, sealhulgas ka inimeste, liikumismustrite analüüsimiseks ja rahvastiku paiknemise tuvastamiseks, mis aitab linnade planeerimisel ja infrastruktuuride optimeerimisel. Lõputöö väljunditeks on trajektooride taastamine ja täpsuse analüüsimine, lisaks sellele inimese liikumismudelite tuvastamine ja tihedamini külastatavate asukohtade identifitseerimine nagu näiteks kodu, töökoht ja poed.
Up until now, GPS data has been greatly used for collecting highlyprecise locational data from moving objects including humans. In contrast, mobile phone data is becoming more and more popular in the last few years. The usage of mobile phone data, that is also known as CDR data, has many benefits over the widely used GPS. This means that the methods used for example in GPS trajectory reconstruction, need to have modifications made be compatible with CDR data.The fact that telecommunication companies have started to cooperate moreand share the CDR data with the public is also a boost to the usage of CDRdata. The processed and analyzed CDR data can be used to get an overview ofcrowd movement in different scales, for example traveling inside a city as opposed to between countries. Extracting trajectories from CDR data has numerous complications.This is due to the fact that the data might not be continuous anddiscovering of the starting point of the object in motion is complicated.The goal of this thesis is to use CDR data in the reconstruction of trajectoriesmade by an anonymous user and to validate the results with GPS data generated in parallel to the CDR data. Reconstructed trajectories can be used for movement analysis and population displacement and would help city planning by optimizing the infrastructures.Outcomes of this thesis are the reconstructed trajectories based on CDR dataand the precisions of final paths. Also, the frequency of CDR events is analyzedin addition to distance distribution. After that the areas that the user visits most frequently are extracted, such as home and work locations.
Up until now, GPS data has been greatly used for collecting highlyprecise locational data from moving objects including humans. In contrast, mobile phone data is becoming more and more popular in the last few years. The usage of mobile phone data, that is also known as CDR data, has many benefits over the widely used GPS. This means that the methods used for example in GPS trajectory reconstruction, need to have modifications made be compatible with CDR data.The fact that telecommunication companies have started to cooperate moreand share the CDR data with the public is also a boost to the usage of CDRdata. The processed and analyzed CDR data can be used to get an overview ofcrowd movement in different scales, for example traveling inside a city as opposed to between countries. Extracting trajectories from CDR data has numerous complications.This is due to the fact that the data might not be continuous anddiscovering of the starting point of the object in motion is complicated.The goal of this thesis is to use CDR data in the reconstruction of trajectoriesmade by an anonymous user and to validate the results with GPS data generated in parallel to the CDR data. Reconstructed trajectories can be used for movement analysis and population displacement and would help city planning by optimizing the infrastructures.Outcomes of this thesis are the reconstructed trajectories based on CDR dataand the precisions of final paths. Also, the frequency of CDR events is analyzedin addition to distance distribution. After that the areas that the user visits most frequently are extracted, such as home and work locations.