Improvement of algorithms to identify transportation modes for MobilitApp, an Android Application to anonymously track citizens in Barcelona

TL;DR

This paper presents improved algorithms for transport mode detection in MobilitApp, an Android app that collects mobility data in Barcelona using accelerometry, aiming to enhance accuracy and independence from external services.

Contribution

The study develops a new transport mode detection algorithm leveraging accelerometer and gyroscope data, improving accuracy and autonomy of MobilitApp.

Findings

Parameters for different transport modes were identified.

A model for transport mode detection was designed.

The algorithm operates in background without interfering with app use.

Abstract

MobilitApp is an Android application whose objective is to obtain mobility data from the citizens of the metropolitan area of Barcelona. The current project is based on the research of more trustful and stronger transport decision algorithms using advantages of accelerometry techniques. The developed algorithm reads data from the mobile's accelerometer and gyroscope and writes it in a file that is afterwards sent to the server. This process is executed in background without interfering in the main application activity. Collected data has been processed and used to analyze the behaviour of the mobility pattern of the distinct transport modalities. The obtained result has been parameters which allow us to configure a model for each mean activity and designing a transport mode detection algorithm which would use information obtained from the mobile's own sensors. MobilitApp is still executing its main functionality, monitoring mobility data from the citizens. In future versions the solution proposed to detect the transport systems will be integrated into the application, in this way the app will work with information obtained from the device and will not depend on any other services.