Methodology for Validating Dynamic Origin–Destination Matrix Estimation Models with Implications for Advanced Traveler Information Systems

This paper describes a methodology for validating online dynamic O–D matrix estimation models using loop detector data in large-scale transportation networks. The simulation procedure focuses on travel aspects related to the collective trip structure of users, including the amount and duration of trips between O–D pairs, trip departure rates, average travel time from each origin and combinations of them. The analysis identifies emerging systematic patterns between these factors and issues related to the model performance, including network scale effects. This procedure aims to enhance the usage of prior O–D information based on, e.g. travel surveys, that are typically used in the estimation process. Moreover, it seeks to integrate the validation of dynamic O–D matrix estimation models with strategies for identifying target population groups for online planning and assessment of real-time travel information services within the context of Advanced Traveler Information Systems (ATIS).     

Limits and perspectives of effective O–D matrix correction using traffic counts

Correction of the O–D matrix from traffic counts is a classical procedure usually adopted in transport engineering by practitioners for improving the overall reliability of transport models. Recently, Papola and Marzano [Papola, A., Marzano, V., 2006. How can we trust in the O–D matrix correction procedure using traffic counts? In: Proceedings of the 2006 ETC Conference, Strasbourg] showed through laboratory experiments that this procedure is generally unable to provide for effective correction of the O–D matrix. From a theoretical standpoint, this result can be justified by the lower number of (stochastic) equations (independent observed link flows) with respect to the unknowns (O–D flows). This paper first confirms that this represents the main reason for the failure of this procedure, showing that satisfactory correction is generally obtained when the number of equations is greater than the number of unknowns. Then, since this circumstance does not occur in practice, where the number of O–D pairs usually far exceeds the number of link counts, we explore alternative assumptions and contexts, allowing for a proper balance between unknowns and equations. This can be achieved by moving to within-day dynamic contexts, where a much larger number of equations are generally available. In order to bound the corresponding increase in the number of unknowns, specific reasonable hypotheses on O–D flow variation across time slices must be introduced. In this respect, we analyze the effectiveness of the O–D matrix correction procedure in the usually adopted linear hypothesis on the dynamic process evolution of O–D flows and under the assumption of constant distribution shares. In the second case it is shown that satisfactory corrections can be performed using a small number of time slices of up to 3 min in length, leading to a time horizon in which the hypothesis of constant distribution shares can be regarded as trustworthy and realistic.     

Using repeated cross-sectional travel surveys to enhance forecasting robustness: Accounting for changing mode preferences

Transportation system capacity and performance, urban form and socio-demographics define the influences and constraints conditioning the preferences of urban residents for different transport modes. Changes in characteristics of urban areas are likely to lead to changes in preferences for alternative modes of transport over time; as a consequence, statistical models to forecast mode choice need to be sensitive to both purposeful changes to urban systems as well as exogenous shocks. We make use of the 1996, 2001 and 2006 household surveys conducted in the Greater Toronto and Hamilton Area to study mode preference evolution and model forecasting performance. These repeated cross-sectional household surveys provide an opportunity to investigate aggregate structural changes in commuting mode preferences over time, in a manner sensitive to changes in the urban area. We focus on commuting mode choices because these trips are prime determinants of peak period congestion and peak spreading. We then address how to combine the three cross-sections econometrically in a robust way that allows for use of a single mode choice model across the entire period. Using independent data from 2012, we are able to compare the individual year and combined models in terms of forecasting performance to demonstrate the combined model’s more robust forecasting performance into the future.     

Patterns of car usage and restraint modelling

Transport demand forecasting procedures have traditionally employed household based modal split models implicitly assuming a selection of mode for each trip based on relative generalised cost. A detailed examination of the trip patterns of a sample of household in West Yorkshire shows that in fact there is little discretionary choice of public transport; public transport trips in car owning households generally being explained in terms of the specific unavailability of the car for such trips. Two versions of a category analysis model for modal split are based on this observation and applied to household data for Glamorgan and Monmouthshire to show that such a procedure is workable and produces results comparing favourably with traditional approaches. The likely implications of three types of restraint policy are examined and it is concluded that the existing interdependence in trip patterns and modal choice within the household is of great significance in determining their effects. In particular it appears that positive attempts to increase vehicle occupancy at the peak are likely to be more favourable to public transport finances than the more negative policies to restrain use of the car for journey to work, or second car ownership.     

Assessing the accuracy of the Sydney Household Travel Survey with GPS

Over the past few years, GPS has been used in a number of surveys in the US to assess the accuracy of household travel surveys. The results have been somewhat alarming in that most of these exercises have shown that the standard trip-based CATI survey conducted in the US under-reports travel by about 20–25%. It was decided to use GPS to assess the accuracy of the Sydney Household Travel Survey, a continuous survey conducted by face-to-face interviewing. The procedure used was for the interviewers to recruit households for the household travel survey in the normal manner, and then, if the household met certain criteria, to endeavour to recruit the household to also undertake a GPS survey. A small sample of about 50 households was obtained, and GPS devices successfully retrieved that measured data on the same day as the travel diary was completed. In addition, participants in the GPS survey completed a prompted recall survey a week or two later, using maps and tabulations of travel obtained from the GPS devices, to identify mode, purpose and occupancy for trips measured by the GPS, and also to check for accuracy in defining trip ends and total number of trips. Based on the analysis of the GPS compared to the diary results, it was found that respondents under-reported their travel by about 7%, which is much less than in the US CATI results. Respondents were also found to under-report travel distances and over-report travel times. There was also a high incidence of non-reporting for VKT.

Combining web and face-to-face in travel surveys: comparability challenges?

Response rates for household travel surveys are tending to fall, and it seems unlikely that this trend will be reversed in the future. In recent years, travel data collection methods have evolved in order to obtain reliable data that are sufficiently detailed to feed increasingly complex models, and in order to integrate new technologies into survey protocols (Internet, GPS??). Combining different media is an obvious low-cost way of improving data quality as it increases the overall response rate. But the question of the comparability of data over time and between different survey modes remains unresolved. This paper makes a comparative analysis between the travel behaviours of web-based survey respondents and respondents to a face-to-face interview. The data were obtained from the 2006 Lyon conurbation household travel survey. Our analysis shows that the Internet respondents reported fewer trips per day than the face-to-face respondents (3.00 vs. 4.04 daily trips), and that the differences between the two groups varied according to the travel mode and trip purpose. While part of this difference can be explained by socioeconomic disparities (the Internet respondents had a specific profile) we cannot exclude the possibility of under-reporting due to the web medium.     

Impact of interviewing by proxy in travel survey conducted by telephone

Telephone‐interview surveys are a very efficient way of conducting large‐scale travel surveys. Recent advancements in computer technology have made it possible to improve upon the quality of data collected by telephone surveys through computerization of the entire sample‐control process, and through the direct recording of the collected data into a computer. Notwithstanding these technological advancements, potential sources of bias still exist, including the reliance on an adult member of the household to report the travel information of other household members. Travel data collected in a recent telephone interview survey in the Toronto region is used to examine this issue. The statistical tool used in the research was the Analysis of Variance (ANOVA) technique as implemented within the general linear model framework in SAS. The study‐results indicate that reliance on informants to provide travel information for non‐informant members of their respective households led to the underreporting of some categories of trips. These underreported trip categories were primarily segments of home‐based discretionary trips, and non home‐based trips. Since these latter two categories of trips are made primarily outside the morning peak period, estimated factors to adjust for their underreporting were time‐period sensitive. Further, the number of vehicles available to the household, gender, and driver license status respectively were also found to be strongly associated with the underreporting of trips and thus were important considerations in the determination of adjustment factors. Work and school trips were found not to be underreported, a not surprising result giving the almost daily repetitiveness of trips made for these purposes and hence the ability of the informant to provide relatively more precise information on them.     

Identifying travel mode with GPS data

Travel mode identification is an essential step in travel information detection with global positioning system (GPS) survey data. This paper presents a hybrid procedure for mode identification using large-scale GPS survey data collected in Beijing in 2010. In a first step, subway trips were detected by applying a GPS/geographic information system (GIS) algorithm and a multinomial logit model. A comparison of the identification results reveals that the GPS/GIS method provides higher accuracy. Then, the modes of walking, bicycle, car and bus were determined using a nested logit model. The combined success rate of the hybrid procedure was 86%. These findings can be used to identify travel modes based on GPS survey data, which will significantly improve the efficiency and accuracy of travel surveys and data analysis. By providing crucial travel information, the results also contribute to modeling and analyzing travel behaviors and are readily applicable to a wide range of transportation practices.     

Trend analysis of activity generation attributes over time

Activity generation models are relatively poorly developed in activity-based travel demand modelling frameworks. This research investigates whether observed distributions of activity attributes (activity frequency, start time and duration) used as inputs in the activity generation component of an activity-based travel demand model have changed over time. This research empirically examines changes in the distributions of activity generation attributes over time in the Greater Montreal area (GMA), Quebec, Canada. It also focuses on how these attributes vary with peoples’ socio-demographic characteristics. This research relies on the 1998, 2003 and 2008 origin–destination (O–D) household travel surveys of the GMA. The comparative analysis at three time points in a 10-year period clearly reveals that distributions of activity attributes are significantly changing over time. Work and school activities show similar trends; frequency “1” has increased and frequency “2+” has decreased over time. The occurrence of shopping activity on weekdays is decreasing over time. Start time and duration distributions for each activity have also changed significantly over time. The research allows preparing activity attributes for the application of an activity-based model, TASHA, such that they reflect temporal changes in travel behaviour of the GMA.     

Some comments on origin–destination matrix estimation

Estimation of origin–destination (O–D) matrices from link count data is considered. This problem is challenging because the number of parameters to be estimated is typically larger than the number of network links. As a result, it is (usually) impossible to identify a unique optimal estimate of the O–D matrix from mean link traffic counts. However, information from the covariance matrix of link count data collected over a sequence of days can relieve this problem of indeterminacy. This fact is illustrated through a simple example. The use of second-order statistical properties of the data in O–D matrix estimation is then explored, and a class of estimators proposed. Practical problems of model mis-specification are discussed and some avenues for future research outlined.     

Tour-based mode choice of joint household travel patterns on weekend and weekday

Joint household travel, with or without joint participation in an activity, constitutes a fundamental aspect in modelling activity-based travel behaviour. This paper examines joint household travel arrangements and mode choices using a utility maximising approach. An individual tour-based mode choice model is formulated contingent on the choice of joint tour patterns where joint household activities and shared ride arrangements are recognised as part of the joint household decision-making that influences the travel modes of each household member. Two models, one for weekend and one for weekday, are estimated using empirical data from the Sydney Household Travel Survey. The results show that weekend travel is characterised by a high joint household activity participation rate while weekday travel is distinguished by more intra-household shared ride arrangements. The arrangements of joint household travel are highly associated with travel purpose, social and mobility constraints and household resources. On weekends, public transport is mainly used by captive users (i.e., no-car households and students) and its share is about half of that on weekdays. Also, the value of travel time savings (VOTs) are found to be higher on weekends than on weekdays, running entirely counter to the common belief that weekend VOTs are lower than weekday VOTs. This paper highlights the importance of studying joint household travel and using different transport management measures for alleviating traffic congestion on weekdays and weekends.     

Transit to eternal youth: lifecycle and generational trends in Greater Montreal public transport mode share

Young people appear to be using public transit more than their predecessors, reversing twentieth century trends, but the importance of such findings depends on whether high transit use persists as these riders age. This paper examines whether transit mode share for commuting trips is increasing; socio-economic and geographic trends are also explored to attempt to determine whether these trends are likely to continue. The study uses repeated cross-sectional origin–destination surveys of Greater Montreal (1998, 2003 and 2008). Over 45,000 home-to-work and home-to-school trips are studied for each survey year. A general lifecycle pattern of decreasing transit share with age is apparent within cohorts until individuals reach their early 30s, followed by decades of stability. This pattern appears to hold in recent years, but with higher youth use rates, and it is argued that the higher use will continue as current younger cohorts mature. Suburbanization by those in their early 30s is evident and, along with household composition changes, appears to explain much of the final within-cohort mode share declines before equilibrium. Transit providers might see lasting ridership gains, as those currently in their early 30s and younger replace lower-use cohorts in the workforce, provided service provision keeps pace. Addressing the needs of young people, whose mode choices are comparatively unsettled, should be a priority for transit agencies to ensure higher transit usage in the future.     

The ambivalence of ridesharing

Ridesharing is quite a popular topic of discussion among transport authority personnel. It is perceived to be a viable alternative to classical modes of transportation, and receives a great deal of political support from transport planners. However, not much objective information is available on ridesharing behaviors. We use travel survey data to study the evolution of the ridesharing market in an urban area. Our study is based on data from four large-scale OD surveys conducted in the Greater Montreal Area (1987, 1993, 1998 and 2003). In the latest survey conducted in Montreal, car passengers were asked to identify the driver who gave them the opportunity to travel in this way. Their answers were classified according to the type of driver; for instance, a member of their household, a neighbor or a co-worker. We use this information to calibrate a model matching car passengers and car drivers belonging to the same household. This will be referred to as IHHR (intra-household ridesharing). Preliminary results reveal that approximately 70% of all trips made by car passengers are the result of IHHR. Furthermore, around 15% of those trips are questionable, in that they were exclusively generated for another individual’s purposes, consequently generating an additional trip for the journey back home. Moreover, this percentage increased over time. Objective data regarding ridesharing and its evolution in an urban area will undoubtedly help decision makers gain a clearer profile of this means of travel and help to realign attitudes on the issue.

Integrated imputation of activity-travel diaries incorporating the measurement of uncertainty

Procedures to transform GPS tracks into activity-travel diaries have been increasingly addressed due to their potential benefit to replace traditional methods used in travel surveys. Existing approaches for data annotation however are not sufficiently accurate, which normally involves a prompted recall survey for data validation. Imputation algorithms for transportation mode detection seem to be largely dependent on speed-related features, which may blur the quality of classification results, especially with transportation modes having similar speeds. Therefore, in this paper we propose an enhanced integrated imputation approach by incorporating the critical indicators related to trip patterns, reflecting the effects of uncertain travel environments, including bus stops and speed percentiles. A two-step procedure which embeds a segmentation model and a transportation mode inference model is designed and examined based on purified prompted recall data collected in a large-scale travel survey. Results show the superior performance of the proposed approach, where the overall accuracy at trip level reaches 93.2% and 88.1% for training and surveyed data, respectively.     

Space and time related determinants of public transport use in trip chains

This research aims at gaining a better understanding about time and space related determinants, which are generally acknowledged to be important factors in the choice of transport mode. The effect of trip chaining is taken into account to improve the insight in the relation between the choice of transport mode and time factors. The data source is the first large scale Belgian mobility survey, carried out in 1998–1999, complemented with a newly created database, containing for each trip a calculated public transport trip. This allows comparing for each trip the actual travel time with the calculated travel time by public transport. Using elasticities and regression techniques the relation between travel time components and public transport use is quantified. On trip level, a clear relation is found between waiting and walking time and public transport use. On trip chain level, travel time variables for the whole trip chain such as the maximum and the range in the travel time ratio provide a significant improvement to the explanatory power of the regression model. The results contain parameters for model input and recommendations to public transport companies on information provision, intermodality and supply.     

The use of a hazard-based duration model for imputation of missing personal income data

This paper suggests using a proportional hazard model to predict personal income, for the purpose of imputing missing income data in household travel surveys. The model has a hazard function that comprises two multiplicative components: (1) a non-parametric baseline hazard function that is dependent only on the income level and (2) a function that is dependent only on the other personal attributes of the survey respondents (excluding income). To estimate and validate the model, data is drawn from a travel characteristics survey conducted in Hong Kong in year 2001. The model is found to have a much higher accuracy when compared with a conventional ordered probit model based on the assumption that the logarithm of income is normally distributed.

A decomposition scheme for estimating dynamic origin–destination flows on actuation-controlled signalized arterials

This paper presents a decomposition framework for estimating dynamic origin–destination (O–D) flows on actuation-controlled signalized arterials from link counts, mainly addressing the issues of incomplete information and the large number of O–D pairs. The framework decomposes the original high-dimensional problem into much smaller sub-problems at the intersection and corridor levels. At the intersection level, turning movements are inferred with incomplete information; at the corridor level, the final estimates of O–D flows are constructed as weighted averages of the estimates from the column and row decompositions. Numerical examples are presented to demonstrate the effectiveness and the computational efficiency of the decomposition framework.     

Quasi-dynamic estimation of o–d flows from traffic counts: Formulation,statistical validation and performance analysis on real data

The paper proposes a “quasi-dynamic” framework for estimation of origin–destination (o–d) flow from traffic counts, under the assumption that o–d shares are constant across a reference period, whilst total flows leaving each origin vary for each sub-period within the reference period. The advantage of this approach over conventional within-day dynamic estimators is that of reducing drastically the number of unknowns given the same set of observed time-varying traffic counts. Obviously, the gain in accuracy depends on how realistic is the underlying assumption that total demand levels vary more rapidly over time than o–d shares. Firstly, the paper proposes a theoretical specification of the quasi-dynamic estimator. Subsequently, it proposes empirical and statistical tests to check the quasi-dynamic assumption and then compares the performances of the quasi-dynamic estimator of o–d flows with both classical off-line simultaneous dynamic estimators and on-line recursive Kalman filter-based estimators. Experiments are carried out on the real test site of A4–A23 motorways in North-Eastern Italy. Results confirm the acceptability of the assumption of quasi-dynamic o–d flows, even under the hypothesis of constant distribution shares for the whole day and show that the quasi-dynamic estimator outperforms significantly the simultaneous estimator. Data also suggest that using the quasi-dynamic estimates instead of the simultaneous estimates as historical o–d flows improves significantly the performances of the Kalman filter, which strongly depends of the quality of the seed o–d flows. In addition, it is shown that the aggregation of quasi-dynamic o–d estimates across subsequent time slices represents also the most effective way to obtain o–d estimates for larger time horizons (e.g. hourly estimates). Finally, a validation based on an hold-out sample of link flows (i.e. counts not used as inputs in the o–d estimation/updating process) revealed the quasi-dynamic estimator to be overall more robust and effective with respect to the other tested estimators.     

Linking land use with household vehicle emissions in the central puget sound: methodological framework and findings

A leading cause of air pollution in many urban regions is mobile source emissions that are largely attributable to household vehicle travel. While household travel patterns have been previously related with land use in the literature (Crane, R., 1996. Journal of the American Planning Association 62 (1, Winter); Cervero, R. and Kockelman, C., 1997. Transportation Research Part D 2 (3), 199–219), little work has been conducted that effectively extends this relationship to vehicle emissions. This paper describes a methodology for quantifying relationships between land use, travel choices, and vehicle emissions within the Seattle, Washington region. Our analysis incorporates land use measures of density and mix which affect the proximity of trip origins to destinations; a measure of connectivity which impacts the directness and completeness of pedestrian and motorized linkages; vehicle trip generation by operating mode; vehicle miles/h of travel and speed; and estimated household vehicle emissions of nitrogen oxides, volatile organic compounds, and carbon monoxide. The data used for this project consists of the Puget Sound Transportation Panel Travel Survey, the 1990 US Census, employment density data from the Washington State Employment Security Office, and information on Seattle’s vehicle fleet mix and climatological attributes provided by the Washington State Department of Ecology. Analyses are based on a cross-sectional research design in which comparisons are made of variations in household travel demand and emissions across alternative urban form typologies. Base emission rates from MOBILE5a and separate engine start rates are used to calculate total vehicle emissions in grams accounting for fleet characteristics and other inputs reflecting adopted transportation control measures. Emissions per trip are based on the network distance of each trip, average travel speed, and a multi-stage engine operating mode (cold start, hot start, and stabilized) function.