Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2014 Apr 22:2:36.
doi: 10.3389/fpubh.2014.00036. eCollection 2014.

Identifying Active Travel Behaviors in Challenging Environments Using GPS, Accelerometers, and Machine Learning Algorithms

Katherine Ellis  1 , Suneeta Godbole  2 , Simon Marshall  2 , Gert Lanckriet  1 , John Staudenmayer  3 , Jacqueline Kerr  2
Affiliations

Identifying Active Travel Behaviors in Challenging Environments Using GPS, Accelerometers, and Machine Learning Algorithms

Katherine Ellis et al. Front Public Health. .

Abstract

Background: Active travel is an important area in physical activity research, but objective measurement of active travel is still difficult. Automated methods to measure travel behaviors will improve research in this area. In this paper, we present a supervised machine learning method for transportation mode prediction from global positioning system (GPS) and accelerometer data.

Methods: We collected a dataset of about 150 h of GPS and accelerometer data from two research assistants following a protocol of prescribed trips consisting of five activities: bicycling, riding in a vehicle, walking, sitting, and standing. We extracted 49 features from 1-min windows of this data. We compared the performance of several machine learning algorithms and chose a random forest algorithm to classify the transportation mode. We used a moving average output filter to smooth the output predictions over time.

Results: The random forest algorithm achieved 89.8% cross-validated accuracy on this dataset. Adding the moving average filter to smooth output predictions increased the cross-validated accuracy to 91.9%.

Conclusion: Machine learning methods are a viable approach for automating measurement of active travel, particularly for measuring travel activities that traditional accelerometer data processing methods misclassify, such as bicycling and vehicle travel.

Keywords: physical activity; random forest.

PubMed Disclaimer

Figures

Figure 1
Figure 1
The classification pipeline. (1) We started from raw sensor data, which was split into 1-min windows. (2) Features were extracted from each window of data. (3) Then the features from each window were classified into transportation modes.
Figure 2
Figure 2
Average roll, pitch, and yaw angles.
Figure 3
Figure 3
Example output for 1 day of data. We plot the activity mode vs. time for 1 day of data. In the top plot, we show the activity mode predicted by the random forest algorithm. In the middle plot, we show the smoothed predictions output by the moving average filter. In the bottom plot, we show the ground truth annotations for this day. Minutes in black were correctly classified, minutes in red were misclassified, and minutes in blue had no ground truth annotation with which to compare.
See this image and copyright information in PMC

References

    1. Martin WA, McGuckin NA. Travel Estimation Techniques for Urban Planning (No. 365). Washington, DC: National Academy Press; (1998).
    1. Merom D, van der Ploeg HP, Corpuz G, Bauman AE. Public health perspectives on household travel surveys: active travel between 1997 and 2007. Am J Prev Med (2010) 39(2):113–21 10.1016/j.amepre.2010.04.007 - DOI - PubMed
    1. Sahlqvist S, Song Y, Ogilvie D. Is active travel associated with greater physical activity? The contribution of commuting and non-commuting active travel to total physical activity in adults. Prev Med (2012) 55(3):206–11 10.1016/j.ypmed.2012.06.028 - DOI - PMC - PubMed
    1. Tudor-Locke C, Ainsworth BE, Popkin BM. Active commuting to school. Sports Med (2001) 31(5):309–13 10.2165/00007256-200131050-00001 - DOI - PubMed
    1. Bassett DR, Jr, Pucher J, Buehler R, Thompson DL, Crouter SE. Walking, cycling, and obesity rates in Europe, North America, and Australia. J Phys Act Health (2008) 5(6):795–814 - PubMed