Improving GPS-Based Mode of Transport Detection in Multi-Modal Trips using Stop Analysis

Jens Klinker, Mariana Avezum-Mercer, Stephan M. Jonas

Abstract: This paper presents an extension to existing GPS-based approaches for tracking modes of transportation in multimodal trips. The extension focuses on analyzing stops and mapping them to surrounding public transport stations in order to improve the accuracy of the mode of transport detection. The proposed method is evaluated using data from the city of Munich, resulting in a 17% improvement of the F1-Score, from 73% to 90%. It is applicable to any GPS-based mode of transport detection system to potentially improve their accuracy.