Exploring Hybrid-Multimodal Routing to Improve User Experience in Urban Trips

• Published in: Applied sciences 2021-05, Vol.11 (10), p.4523
• Main Authors: Rodrigues, Diego O., Maia, Guilherme, Braun, Torsten, Loureiro, Antonio A. F., Peixoto, Maycon L. M., Villas, Leandro A.
• Format: Article
• Language: English
• Subjects: Algorithms; Bicycling; big data; Cities; Clustering; Datasets; flow clustering; Global positioning systems; GPS; intelligent transportation systems; Mobility; Motivation; multi-source data analyses; Route choice; Route selection; spatio-temporal data analyses; Taxicabs; Traffic congestion; Traffic jams; Transportation; Trends; Urban transportation; User experience
• ISSN: 2076-3417; 2076-3417
• DOI: 10.3390/app11104523

Millions of individuals rely on urban transportation every day to travel inside cities. However, it is not clear how route parameters (e.g., traffic conditions, waiting times) influence users when selecting a particular route option for their trips. These parameters play an important role in route recommendation systems, and most of the currently available applications omit them. This work introduces a new hybrid-multimodal routing algorithm that evaluates different routes that combine different transportation modes. Hybrid-multimodal routes are route options that might consist of more than one transportation mode. The motivation to use different transportation modes is to avoid unpleasant trip segments (e.g., traffic jams, long walks) by switching to another mode. We show that the possibility of planning a trip with different transportation modes can lead to improvement of cost, duration, and quality of experience urban trips. We outline the main research contributions of this work, as (i) an user experience model that considers time, price, active transportation (i.e., non-motorized transport) acceptability, and traffic conditions to evaluate the hybrid routes; and, (ii) a flow clustering technique to identify relevant mobility flows in low-sampled datasets for reducing the data volume and allow the execution of the analytical evaluation. (i) uses a Discrete Choice Analyses framework to model different variables and estimate a value for user experience in the trip. (ii) is a methodology to aggregate mobility flows by using Spatio-temporal Clustering and identify the most relevant of these flows using Curvature Analysis. We evaluate the proposed hybrid-multimodal routing algorithm with data from the Green and Yellow Taxis of New York, Citi Bike NYC data, and other publicly available datasets; and, different APIs, such as Uber and Google Directions. The results reveal that selecting hybrid routes can benefit passengers by saving time or reducing costs, and sometimes both, when compared to routes using a single transportation mode.