Revealing the day-to-day regularity of urban congestion patterns with 3D speed maps

Abstract

In this paper, we investigate the day-to-day regularity of urban congestion patterns. We first partition link speed data every 10 min into 3D clusters that propose a parsimonious sketch of the congestion pulse. We then gather days with similar patterns and use consensus clustering methods to produce a unique global pattern that fits multiple days, uncovering the day-to-day regularity. We show that the network of Amsterdam over 35 days can be synthesized into only 4 consensual 3D speed maps with 9 clusters. This paves the way for a cutting-edge systematic method for travel time predictions in cities. By matching the current observation to historical consensual 3D speed maps, we design an efficient real-time method that successfully predicts 84% trips travel times with an error margin below 25%. The new concept of consensual 3D speed maps allows us to extract the essence out of large amounts of link speed observations and as a result reveals a global and previously mostly hidden picture of traffic dynamics at the whole city scale, which may be more regular and predictable than expected.Studying human mobility in large cities is critical for multiple applications from transportation engineering to urban planning and economic forecasting. In recent years, the availability of new data sources, e.g. mobile-phone records and global-positioning-system data, has generated new empirically driven insights on this topic. A central question at large spatial and temporal scales is which (dynamic) components of human mobility can be considered as predictable and thus suitable for explanatory and predictively valid mathematical models, and which part is unpredictable. Earlier studies of human trips shows that traveled distance can be described by random walks and more precisely as Lévy-flights 1 . Latter studies partly amend this theory by recognizing some regularity features in peoples' trips. Individuals obviously frequently move between specific locations, such as home or work 2 . Such patterns are also regular in time 3,4 meaning that the most frequent locations are likely to be correlated with daily hours and dates. Regularity can also come from decomposition by transportation modes 5 . Human mobility can be studied at the microscopic level, i.e. through person trajectories, but also at the macroscopic level, for example by estimating commuting flows between different regions (origins to destinations) or on the different links of a transportation network 6,7 . Such collective mobility patterns can be explained for example by distances between regions 8,9 , trip purposes 10 and road attractiveness related to road types, e.g. freeways, or locations, e.g. in major business districts 11 . Predicting commuting flows often requires local data for calibration 12 meaning that results cannot easily be transferable to other regions or cities. Recent findings 13 , however, show that a scale-free approach corresponding to an extension of the radiation model can successfully be applied to commuting flow estimation. This means that some...

Unraveling reaction-diffusion-like dynamics in urban congestion propagation: Insights from a large-scale road network

“…Recent findings from empirical and simulated data 26,27 have identified the spatial distribution of vehicle density in the network as one of the key components that influence the shape and the scatter of an MFD. These findings are of great importance because the concept of an MFD can be applied for heterogeneously loaded cities with multiple centers of congestion if these cities can be partitioned into a number of homogeneous spatially connected clusters 19,[28][29][30] .…”

Estimation of the shared mobility demand based on the daily regularity of the urban mobility and the similarity of individual trips

“…A significant regularity appears at a larger scale, making it possible to identify recurrent trends in urban mobility even if variability can be observed in the trips at the scale of a particular individual [ 3 ] and to derive behavioral laws [ 4 – 6 ]. It appears that patterns repeating from one day to the next can be identified either by observing the traffic conditions [ 7 ], the most used roads [ 8 , 9 ], or in the choice of modes of transport [ 10 ]. Despite this conformity in individual human behavior, private cars and taxis (and related services of transportation network companies) remain largely inefficient modes of transport, carrying less than 1.5 passengers on average, leading to significant traffic congestion and generating irreversible economic and environmental impacts [ 11 ].…”

Recognizing Spatio-Temporal Traffic Patterns at Intersections Using Self-Organizing Maps

“…The aim of analyzing traffic states is not new and has been tackled by different authors in the past [4][5][6][7][8][9][10][11][12][13][14][15]. The focus on temporal factors, which might influence traffic states, e.g.…”

“…Many previous approaches focused on network-wide travel time prediction. Lopez et al [9] proposed an urban-wide prediction model based on clustering. The approach identified day-to-day regularities in urban congestions.…”

“…As spatio-temporal analysis method for detecting traffic patterns, the application of clustering algorithms is common. In [9] the authors used k-means and DBSCAN. Other authors applied statistical models [14,8,7], fuzzy fitting [4] and piece-wise linear curves [5].…”