Constructing Time-Dependent Origin-Destination Matrices With Adaptive Zoning Scheme and Measuring Their Similarities With Taxi Trajectory Data

Abstract

There has been a recent push towards using opportunistic sensing data collected from sources like automatic vehicle location (AVL) systems, mobile phone networks, and global positioning system (GPS) tracking to construct origin-destination (O-D) matrices, which are an effective alternative to expensive and time-consuming traditional travel surveys. These data have numerous drawbacks: they may have inadequate detail about the journey, may lack spatial and temporal granularity, or may be limited due to privacy regulations. Taxi trajectory data is an opportunistic sensing data type that can be effectively used for O-D matrix construction because it addresses the issues that plague other data sources. This paper presents a new approach for using taxi trajectory data to construct a taxi O-D matrix that is dynamic in both space and time. The model's origin and destination zone sizes and locations are not fixed, allowing the dimensions to vary from one matrix to another. Comparisons between these spatiotemporal-varying O-D matrices cannot be made using a traditional method like matrix subtraction. Therefore, this paper introduces a new measure of similarity. Our proposed approaches are applied to the taxi trajectory data collected from Lisbon, Portugal as a case study. The results reveal the periods in which taxi travel demand is the highest and lowest, as well as the periods in which the highest and lowest regular taxi travel demand patterns take shape. This information about taxi travel demand patterns is essential for informed taxi service operations management. INDEX TERMS Dynamic origin-destination matrix, adaptive zoning scheme, origin-destination matrix similarity measure, taxi trajectory data, taxi travel demand.

Mobility Signatures: A Tool for Characterizing Cities Using Intercity Mobility Flows

“…These matrices contain the number of trips between two geographical points or areas. There are several choices that need to be made when constructing OD matrices, including choosing spatial scales and period of observations (Friedrich et al, 2010 ; Mungthanya et al, 2019 ). The spatial scale represented by the elements of the OD matrix can be, for example, a city, county, municipality, or country.…”

Modeling Location Choice of Taxi Drivers for Passenger Pickup Using GPS Data

“…There has been a number of studies aimed at modeling occupied taxi trips. For instance, Liu et al [57], Werabhat et al [58], and Zhang et al [59] estimated the O-D trips of occupied taxis using GPS data. Further improvements on the aforementioned studies were achieved through the development of trip distribution models [20] [19].…”

Inference of dynamic origin–destination matrices with trip and transfer status from individual smart card data

“…First approaches for the estimation of origin-destination (OD) matrices were based on statistical inference from interviews or/and surveys. However, with the monitoring of individual movements in the network, it has been possible to model dynamic and more accurate matrices of the state of urban traffic through sensory data sources such as phone mobile records [23], global position system trajectories [19], and smart card records [24]. In fact, most studies in the scope of public urban transport with AFC systems are dependent on smart card information.…”

“…Firstly, OD matrices can be classified as either static or dynamic. A static OD matrix considers time-independent flows over the space [19]. For this typology, methodologies have been developed to capture average flows between OD pairs within a geographic area, in a single matrix, such as gravity models, entropy maximization, information minimization [20].…”