Transportation mode detection based on mobile sensor data

Urbančič, Jasna (author), Pejović, Veljko (mentor), Mladenić, Dunja (comentor)

This thesis addresses transportation mode detection based primarily on mobile phone data using machine learning methods. Our approach uses short samples of accelerometer readings taken while traveling in a vehicle to distinguish between three modalities --- car, bus, and train. We use gravity estimation to pre-process the samples. We extract features from statistical, frequency-based, and peak-based domain. With statistical analysis of the features we gain an introspective into the data. To additionally analyze the features we construct several feature sets for classification. As a classifier we use random forest, support vector machine, and neural network. Our approach correctly classifies 65% cars, 63% buses, and 18% trains using neural network.

• Language: English
• Keywords: machine learning, mobile sensing, data mining, pattern recognition, intelligent transportation systems
• Work type: Master's thesis/paper
• Organization: FRI - Faculty of Computer and Information Science
• Year: 2018
• PID: 20.500.12556/RUL-106015
• Publication date in RUL: 14.01.2019

URBANČIČ, Jasna, 2018, Transportation mode detection based on mobile sensor data [online]. Master’s thesis. [Accessed 19 August 2025]. Retrieved from: https://repozitorij.uni-lj.si/IzpisGradiva.php?lang=eng&id=106015

Secondary language

• Language: Slovenian
• Title: Detekcija prevoznega sredstva z mobilnimi senzorji

Abstract

V delu obravnavamo detekcijo prevoznega sredstva z mobilnimi senzorji in metodami strojnega učenja. Pri tem uporabljamo kratke vzorce podatkov iz pospeškometra, ki jih zajamemo med uporabnikovim potovanjem v vozilu. Razločujemo med tremi prevoznimi sredstvi --- avtom, avtobusom in vlakom. Vzorce predobdelamo tako, da iz pospeškov izločimo gravitacijsko komponento. Iz vzorcev izločimo statistične in frekvenčne značilke ter značilke vrhov. S statistično analizo značilk dobimo vpogled v podatke. Dodatno analiziramo značilke preko različnih množic značilk, ki jih uporabljamo za klasifikacijo. Kot klasifikatorje uporabljamo naključne gozdove, metodo podpornih vektorjev in nevronske mreže. Z uporabo nevronskih mrež smo pravilno razpoznali 65% avtomobilov, 63% avtobusov in 18% vlakov.

• Keywords: strojno učenje, mobilno zaznavanje, podatkovno rudarjenje, razpoznava vzorcev, inteligentni transportni sistemi