Transport mode detection (TMD) is an essential cornerstone for advancing the field of mobility research, providing vital insights into human travel behavior, urban mobility, and transportation infrastructure usage. It plays a fundamental role in devising efficient transportation strategies, optimizing public transit systems, and informing sustainable mobility initiatives. However, challenges posed by GPS data, including poor reception and varying accuracy, demand innovative solutions. This paper presents a novel and comprehensive method that integrates multi-source data, including GPS, smartphone sensors, and road network information, to enhance TMD accuracy for real-world trips. By consolidating neural networks, map-matching techniques, and a trajectory similarity function the solution proposed tackle the complexities of TMD effectively. Through extensive validation using the Geco Air application, we demonstrate the robustness of the novel method over existing approaches, showcasing high performance even in challenging GPS scenarios. Leveraging data from smartphones' embedded sensors and road networks further enhances the precision of mode detection, contributing to the development of more reliable transportation systems. This solution offers a holistic and practical approach to automatic transport mode detection. By leveraging multi-source data, we present a promising direction for future research, shaping the landscape of mobility analysis and decision-making in transportation domains.
A Robust Approach for Transportation Mode Detection Using Smartphone-Based GPS Sensors and Road Network Information
2023-09-24
8171065 byte
Conference paper
Electronic Resource
English
German Aerospace Center (DLR) | 2022
|