Traveler Mobility and Activity Pattern Inference UsingPersonal Smartphone Applications and ArtificialIntelligence Methods
Author | : Ali Yazdizadeh |
Publisher | : |
Total Pages | : 0 |
Release | : 2022 |
ISBN-10 | : OCLC:1393188734 |
ISBN-13 | : |
Rating | : 4/5 ( Downloads) |
Download or read book Traveler Mobility and Activity Pattern Inference UsingPersonal Smartphone Applications and ArtificialIntelligence Methods written by Ali Yazdizadeh and published by . This book was released on 2022 with total page 0 pages. Available in PDF, EPUB and Kindle. Book excerpt: Recent advances in communication technologies have enabled researchers to collect travel data from location-aware smartphones. These advances hold out the promise of allowing the automatic detection of the critical aspects (mode, purpose, etc.) of people's travel. This thesis investigates the application of artificial intelligence methods to infer mode of transport, trip purpose and transit itinerary from traveler trajectories gathered by smartphones. Supervised, Random Forest models are used to detect mode, purpose and transit itinerary of trips. Deep learning models, in particular, Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN), are also employed to infer mode of transport and trip purpose. The research also explores the use of Generative Adversarial Networks (GANs), as a semi-supervised learning approach, to classify trip mode. Moreover, we investigate the application of multi-task learning to simultaneously infer mode and purpose. The research uses several different data sources. Trip trajectory data was collected by the MTL Trajet smartphone Travel Survey App, in 2016. Also, other complementary datasets, such as locational data from social media, land-use, General Transit Feed Specification (GTFS), and elevation data are exploited to infer trip information. Mode of transport can be inferred with Random Forest models, ensemble CNN models, and RNN approaches with an accuracy of 87%, 91%, and 86%, respectively. The Random Forest and multi-task RNN models to infer trip purpose achieve an accuracy of 71% and 78%, respectively. Also, the Random Forest transit itinerary inference model can predict used transit itineraries with an accuracy of 81%. While further improvement is required to enhance the performance of the developed artificial intelligence models on smartphone data, the results of the research indicate the capability of smartphone-based travel surveys as a complementary (and potentially replacement) surveying tool to household travel surveys.