Developing a passive GPS tracking system to study long-term travel behavior

Abstract

This paper describes development and testing of a passive GPS tracking smartphone application and corresponding data analysis methodology designed to increase the quality of travel behavior information collected in long-term travel surveys. The new approach is intended to replace the pencil-and-paper travel diaries and prompted recall methods that require more user involvement due to requirements for manual data entry and/or high battery usage. Reducing the burden placed on users enables researchers to collect data over longer periods, thus improving the quality of travel behavior research. To reduce battery use the smartphone-based application collects GPS data less frequently than other methods. Therefore, new algorithms were developed to identify trips and activities, transport mode, and even the specific vehicle used by the traveler. An important finding was the significant advantage of using users past data to improve mode detection results. The system was tested successfully in Zürich and Basel (Switzerland).

On the requirements on spatial accuracy and sampling rate for transport mode detection in view of a shift to passive signalling data

“…Many classifiers have been used for the task of classification in mode detection. The ones that are most closely tied to the specificities of the problem at hand are the rule-based classifiers that typically have relatively rigorous boundaries on a relatively small number of features (Bohte & Maat, 2009;Chen et al, 2010;Gong et al, 2012;Sauerländer-Biebl et al, 2017;Schuessler & Axhausen, 2009;Stopher et al, 2008;Marra et al, 2019). In situations where there are more features affecting the classification, support vector machines (Bolbol et al, 2012;Pereira et al, 2013), decision trees (Reddy et al, 2010), and random forests (Ellis et al, 2014;Mäenpää et al, 2017) are the most popular choices.…”

WITHDRAWN: A Survey on Local Transport Mode Detection on The Edge

“…In fact, the smartphone mobility introduces noises when it moves close to the human body or when it is placed at different positions. In order to provide an accurate classification it is pivotal to use some techniques such as data filtering (i.e., removing the data that does not represent user real positions) and smoothing (to help reducing random noise present in the data) to ensure better accuracy [34].…”

“…In Marra's work [34] a passive GPS tracking application is proposed; it is claimed that it consumes very little battery power by reducing the GPS sampling rate. However, there is no evaluation of such a claim in Marra's article.…”

“…Position User Location independency independency independency Reddy [10] Yes Yes No Hemminki [12] Yes Yes Yes Sonderen [13] No No NA Martin [11] Yes No No Guvensan [4] Yes Yes Yes Marra [34] NA Yes No Soares [39] NA No No Liang [24] Yes No No Zhao [23] No NA NA…”

“…Variance along the Discrete Fourier Transform (DFT) energy coefficients between 1-3 Hz from the accelerometer and the speed from the GPS receiver were selected as the feature set. [a] Data filtering removes the data that do not represent the correct information [34].…”