Recurrent flow networks: A recurrent latent variable model for density estimation of urban mobility. (September 2022)

• Record Type: Journal Article
• Authors: Gammelli, Daniele; Rodrigues, Filipe
• Is Part Of: Pattern recognition. Volume 129(2022)
• Publication Date: 2022-09
• Subjects: Urban mobility -- Latent variable models -- Normalizing flows -- Variational inference; Pattern perception -- Periodicals; Perception des structures -- Périodiques; Patroonherkenning; 006.4
• DOI: 10.1016/j.patcog.2022.108752
• Languages: English
• ISSNs: 0031-3203

Abstract

Highlights: Mobility demand characterized by spatial and temporal variability. Recurrent Flow Networks (RFN) formulated for spatio-temporal density estimation. RFNs exhibit long-term predictions and fine-grained distributions on urban topologies. Experiments with synthetic and real-world data demonstrate solution approach. Abstract: Mobility-on-demand (MoD) systems represent a rapidly developing mode of transportation wherein travel requests are dynamically handled by a coordinated fleet of vehicles. Crucially, the efficiency of an MoD system highly depends on how well supply and demand distributions are aligned in spatio-temporal space (i.e., to satisfy user demand, cars have to be available in the correct place and at the desired time). To do so, we argue that predictive models should aim to explicitly disentangle between temporal and spatial variability in the evolution of urban mobility demand. However, current approaches typically ignore this distinction by either treating both sources of variability jointly, or completely ignoring their presence in the first place. In this paper, we propose recurrent flow networks 1 (RFN), where we explore the inclusion of (i) latent random variables in the hidden state of recurrent neural networks to model temporal variability, and (ii) normalizing flows to model the spatial distribution of mobility demand. We demonstrate how predictive models explicitly disentangling between spatial and temporal variability exhibit several desirable properties, and empirically show how this enables the generation of distributions matching potentially complex urban topologies.