Identifying trip ends from raw GPS data with a hybrid spatio-temporal clustering algorithm and random forest model: a case study in Shanghai

ABSTRACT

Smartphones have been advocated as the preferred devices for travel behavior studies over conventional surveys. But the primary challenges are candidate stops extraction from GPS data and trip ends distinction from noise. This paper develops a Resident Travel Survey System (RTSS) for GPS data collection and travel diary verification, and then uses a two-step method to identify trip ends. In the first step, a density-based spatio-temporal clustering algorithm is proposed to extract candidate stops from trajectories. In the second step, a random forest model is applied to distinguish trip ends from mode transfer points. Results show that the clustering algorithm achieves a precision of 96.2%, a recall of 99.6%, mean absolute error of time within 3?min, and average offset distance within 30 meters. The comprehensive accuracy of trip ends identification is 99.2%. The two-step method performs well in trip ends identification and promotes the efficiency of travel survey systems.     

Review of GPS Travel Survey and GPS Data-Processing Methods

Abstract

Global positioning system (GPS) devices have been utilised in travel surveys since the late 1990s. Because GPS devices are very accurate at recording time and positional characteristics of travel, they can correct the trip-misreporting issue resulting from self-reports of travel and improve the accuracy of travel data. Although the initial idea of using GPS surveys in transport data collection was just to replace paper-based travel diaries, GPS surveys currently are being applied in a number of transport fields. Several general reviews have been done about GPS surveys in the literature review sections in some papers, but a detailed systematic review from GPS data collection to the whole procedure of GPS data processing has not been undertaken. This paper comprehensively reviews the development of GPS surveys and their applications, and GPS data processing. Different from most reviews in GPS research, this paper provides a detailed and systematic comparison between different methods from trip identification to mode and purpose detection, introduces the methods that researchers and planners are currently using, and discusses the pros and cons of those methods. Based on this review, researchers can choose appropriate methods and endeavour to improve them.     

Detecting trip purposes from smartphone-based travel surveys with artificial neural networks and particle swarm optimization

Travel surveys based on global positioning system (GPS) data have exponentially increased over the past decades. Trip characteristics, including trip ends, travel modes, and trip purposes need to be detected from GPS data. Compared with other trip characteristics, studies on trip purpose detection are limited. These studies struggle with three types of limitations, namely, data validation, classification approach-related issues, and result comparison under multiple scenarios. Therefore, we attempt to obtain full understanding and improve these three aspects when detecting trip purposes in the current study. First, a smartphone-based travel survey is employed to collect GPS data, and a surveyor-intervened prompted recall survey is used to validate trip characteristics automatically detected from the GPS data. Second, artificial neural networks combined with particle swarm optimization are used to detect trip purposes from the GPS data. Third, four scenarios are constructed by employing two methods for land-use type coding, i.e., polygon-based information and point of interest, and two methods for selecting training dataset, i.e., equal proportion selection and equal number selection. The accuracy of trip purpose detection is then compared under these scenarios. The highest accuracies of 97.22% for the training dataset and 96.53% for the test dataset are achieved under the scenario of polygon-based information and equal proportion selection by comparing the detected and validated trip purposes. Promising results indicate that a smartphone-based travel survey can complement conventional travel surveys.     

A modeling approach for matching ridesharing trips within macroscopic travel demand models

State of the art travel demand models for urban areas typically distinguish four or five main modes: walking, cycling, public transport and car. The mode car can be further split into car-driver and car-passenger. As the importance of ridesharing may increase in the coming years, ridesharing should be addressed as an additional sub or main mode in travel demand modeling. This requires an algorithm for matching the trips of suppliers (typically car drivers) and demanders (travelers of non-car modes). The paper presents a matching algorithm, which can be integrated in existing travel demand models. The algorithm works likewise with integer demand, which is typical for agent-based microscopic models, and with non-integer demand occurring in travel demand matrices of a macroscopic model. The algorithm compares two path sets of suppliers and demanders. The representation of a path in the road network is reduced from a sequence of links to a sequence of zones. The zones act as a buffer along the path, where demanders can be picked up. The travel demand model of the Stuttgart Region serves as an application example. The study estimates that the entire travel demand of all motorized modes in the Stuttgart Region could be transported by 7% of the current car fleet with 65% of the current vehicle distance traveled, if all travelers were willing to either use ridesharing vehicles with 6 seats or traditional rail.     

An analysis of the factors influencing differences in survey-reported and GPS-recorded trips

Recent advances in global positioning systems (GPS) technology have resulted in a transition in household travel survey methods to test the use of GPS units to record travel details, followed by the application of an algorithm to both identify trips and impute trip purpose, typically supplemented with some level of respondent confirmation via prompted-recall surveys. As the research community evaluates this new approach to potentially replace the traditional survey-reported collection method, it is important to consider how well the GPS-recorded and algorithm-imputed details capture trip details and whether the traditional survey-reported collection method may be preferred with regards to some types of travel. This paper considers two measures of travel intensity (survey-reported and GPS-recorded) for two trip purposes (work and non-work) as dependent variables in a joint ordered response model. The empirical analysis uses a sample from the full-study of the 2009 Indianapolis regional household travel survey. Individuals in this sample provided diary details about their travel survey day as well as carried wearable GPS units for the same 24-h period. The empirical results provide important insights regarding differences in measures of travel intensities related to the two different data collection modes (diary and GPS). The results suggest that more research is needed in the development of workplace identification algorithms, that GPS should continue to be used alongside rather than in lieu of the traditional diary approach, and that assignment of individuals to the GPS or diary survey approach should consider demographics and other characteristics.     

Travel time estimation for urban road networks using low frequency probe vehicle data

The paper presents a statistical model for urban road network travel time estimation using vehicle trajectories obtained from low frequency GPS probes as observations, where the vehicles typically cover multiple network links between reports. The network model separates trip travel times into link travel times and intersection delays and allows correlation between travel times on different network links based on a spatial moving average (SMA) structure. The observation model presents a way to estimate the parameters of the network model, including the correlation structure, through low frequency sampling of vehicle traces. Link-specific effects are combined with link attributes (speed limit, functional class, etc.) and trip conditions (day of week, season, weather, etc.) as explanatory variables. The approach captures the underlying factors behind spatial and temporal variations in speeds, which is useful for traffic management, planning and forecasting. The model is estimated using maximum likelihood. The model is applied in a case study for the network of Stockholm, Sweden. Link attributes and trip conditions (including recent snowfall) have significant effects on travel times and there is significant positive correlation between segments. The case study highlights the potential of using sparse probe vehicle data for monitoring the performance of the urban transport system.     

Investigation of the use of smartphone applications for trip planning and travel outcomes

This paper explores the use of smartphone applications for trip planning and travel outcomes using data derived from a survey conducted in Halifax, Nova Scotia, in 2015. The study provides empirical evidence of relationships of smartphone use for trip planning (e.g. departure time, destination, mode choice, coordinating trips and performing tasks online) and resulting travel outcomes (e.g. vehicle kilometers traveled, social gathering, new place visits, and group trips) and associated factors. Several sets of factors such as socio-economic characteristics and travel characteristics are tested and interpreted. Results suggest that smartphone applications mostly influence younger individuals’ trip planning decisions. Transit pass owners are the frequent users of smartphone applications for trip planning. Findings suggest that transit pass owners commonly use smartphone applications for deciding departure times and mode choices. The study also identifies the limited impact of smartphone application use on reducing travel outcomes, such as vehicle kilometers traveled. The highest impact is in visiting new places (a 48.8% increase). The study essentially offers an original in-depth understanding of how smartphone applications are affecting everyday travel.     

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.

Evaluating the feasibility of a passive travel survey collection in a complex urban environment: Lessons learned from the New York City case study

The combination of increasing challenges in administering household travel surveys and advances in global positioning systems (GPS)/geographic information systems (GIS) technologies motivated this project. It tests the feasibility of using a passive travel data collection methodology in a complex urban environment, by developing GIS algorithms to automatically detect travel modes and trip purposes. The study was conducted in New York City where the multi-dimensional challenges include urban canyon effects, an extreme dense and diverse set of land use patterns, and a complex transit network. Our study uses a multi-modal transportation network, a set of rules to achieve both complexity and flexibility for travel mode detection, and develops procedures and models for trip end clustering and trip purpose prediction. The study results are promising, reporting success rates ranging from 60% to 95%, suggesting that in the future, conventional self-reported travel surveys may be supplemented, or even replaced, by passive data collection methods.     

Trip chain complexity: a comparison among latent classes of daily mobility patterns

This paper studies the relationship between trip chain complexity and daily travel behaviour of travellers. While trip chain complexity is conventionally investigated between travel modes, our scope is the more aggregated level of a person’s activity-travel pattern. Using data from the Netherlands Mobility Panel, a latent class cluster analysis was performed to group people with similar mode choice behaviour in distinct mobility pattern classes. All trip chains were assigned to both a travel mode and the mobility pattern class of the traveller. Subsequently, differences in trip chain complexity distributions were analysed between travel modes and between mobility pattern classes. Results indicate considerable differences between travel modes, particularly between multimodal and unimodal trip chains, but also between the unimodal travel modes car, bicycle, walking and public transport trip chains. No substantial differences in trip chain complexity were found between mobility pattern classes. Independently of the included travel modes, the distributions of trip chain complexity degrees were similar across mobility pattern classes. This means that personal circumstances such as the number of working hours or household members are not systematically translated into specific mobility patterns.

     

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.     

Link travel time inference using entry/exit information of trips on a network

This paper studies link travel time estimation using entry/exit time stamps of trips on a steady-state transportation network. We propose two inference methods based on the likelihood principle, assuming each link associates with a random travel time. The first method considers independent and Gaussian distributed link travel times, using the additive property that trip time has a closed-form distribution as the summation of link travel times. We particularly analyze the mean estimates when the variances of trip time estimates are known with a high degree of precision and examine the uniqueness of solutions. Two cases are discussed in detail: one with known paths of all trips and the other with unknown paths of some trips. We apply the Gaussian mixture model and the Expectation–Maximization (EM) algorithm to deal with the latter. The second method splits trip time proportionally among links traversed to deal with more general link travel time distributions such as log-normal. This approach builds upon an expected log-likelihood function which naturally leads to an iterative procedure analogous to the EM algorithm for solutions. Simulation tests on a simple nine-link network and on the Sioux Falls network respectively indicate that the two methods both perform well. The second method (i.e., trip splitting approximation) generally runs faster but with larger errors of estimated standard deviations of link travel times.     

Tracking daily travel; Assessing discrepancies between GPS-derived and self-reported travel patterns

Global Positioning Systems (GPS) technologies have been used in conjunction with traditional one- or two-day travel diaries to audit respondent reporting patterns, but we used GPS-based monitoring to conduct the first assessment to our knowledge of travel reporting patterns using a seven-day travel log instrument, which could reduce response burden and provide multiple-day, policy-relevant information for evaluation studies. We found substantial agreement between participant-reported daily travel patterns and GPS-derived patterns among 116 adult residents of a largely low-income and non-white transportation corridor in urbanized Los Angeles in 2011–2013. For all modes, the average difference between daily GPS- and log-derived trip counts was only about 0.39 trips and the average difference between daily GPS- and log-derived walking duration was about −11.8 min. We found that the probability that a day would be associated with agreement or discrepancies between these measurement tools varied by travel mode and participant socio-demographic characteristics. Future research is needed to investigate the potential and limitations of this and other self-report instruments for a larger sample and a wider range of population groups and travel patterns.     

Deriving and validating trip purposes and travel modes for multi-day GPS-based travel surveys: A large-scale application in the Netherlands

In the past few decades, travel patterns have become more complex and policy makers demand more detailed information. As a result, conventional data collection methods seem no longer adequate to satisfy all data needs. Travel researchers around the world are currently experimenting with different Global Positioning System (GPS)-based data collection methods. An overview of the literature shows the potential of these methods, especially when algorithms that include spatial data are used to derive trip characteristics from the GPS logs. This article presents an innovative method that combines GPS logs, Geographic Information System (GIS) technology and an interactive web-based validation application. In particular, this approach concentrates on the issue of deriving and validating trip purposes and travel modes, as well as allowing for reliable multi-day data collection. In 2007, this method was used in practice in a large-scale study conducted in the Netherlands. In total, 1104 respondents successfully participated in the one-week survey. The project demonstrated that GPS-based methods now provide reliable multi-day data. In comparison with data from the Dutch Travel Survey, travel mode and trip purpose shares were almost equal while more trips per tour were recorded, which indicates the ability of collecting trips that are missed by paper diary methods.     

Tracking a system of shared autonomous vehicles across the Austin,Texas network using agent-based simulation

This study provides a large-scale micro-simulation of transportation patterns in a metropolitan area when relying on a system of shared autonomous vehicles (SAVs). The six-county region of Austin, Texas is used for its land development patterns, demographics, networks, and trip tables. The agent-based MATSim toolkit allows modelers to track individual travelers and individual vehicles, with great temporal and spatial detail. MATSim’s algorithms help improve individual travel plans (by changing tour and trip start times, destinations, modes, and routes). Here, the SAV mode requests were simulated through a stochastic process for four possible fare levels: $0.50, $0.75, $1, and $1.25 per trip-mile. These fares resulted in mode splits of 50.9, 12.9, 10.5, and 9.2% of the region’s person-trips, respectively. Mode choice results show longer-distance travelers preferring SAVs to private, human-driven vehicles (HVs)—thanks to the reduced burden of SAV travel (since one does not have to drive the vehicle). For travelers whose households do not own an HV, SAVs (rather than transit, walking and biking) appear preferable for trips under 10 miles, which is the majority of those travelers’ trip-making. It may be difficult for traditional transit services and operators to survive once SAVs become available in regions like Austin, where dedicated rail lines and bus lanes are few. Simulation of SAV fleet operations suggest that higher fare rates allow for greater vehicle replacement (ranging from 5.6 to 7.7 HVs per SAV, assuming that the average SAV serves 17–20 person-trips per day); when fares rise, travel demands shift away from longer trip distances. Empty vehicle miles traveled by the fleet of SAVs ranged from 7.8 to 14.2%, across the scenarios in this study. Implications of mobility and sustainability benefits of SAVs are also discussed in the paper.     

Exploring key determinants of travel satisfaction for multi-modal trips by different traveler groups

Using a primary dataset from an experimental survey in eight European cities, this study identified the key determinants of satisfaction with individual trip stages as well as overall journey experience for different travel modes and traveler groups. Multivariate statistical analyses were used to examine the relationships between overall satisfaction and travel experience variables, trip complexity, subjective well-being indices, travel-related attitudes as well as individual- and trip-specific attributes. The results indicate that for certain traveler groups, such as women, young and low-income or unemployed travelers, there are distinctive determinants of satisfaction with trip stages for various travel modes. The results also indicate that satisfaction with the primary trip stage is strongly linked to overall trip satisfaction, while satisfaction levels with access and egress trip stages are strongly related to satisfaction with the primary trip stage. Past experience, traveler expectations and attitudes, and the emotional state of travelers are also significant explanatory variables for travel satisfaction. The results indicate that when an individual consciously chooses a particular travel mode, they will report a higher level of satisfaction with that chosen mode. Notwithstanding, while past experience highly influences an individual’s current travel satisfaction, the more they travel with the current mode, the less satisfied they are with their choice. The results of this study highlight the importance of gaining a better understanding of the interaction between instrumental variables and non-instrumental variables at different trip stages and the influence on user preferences, satisfaction and decision-making processes.     

A theoretical and empirical model of trip chaining behavior

This paper addresses the theoretical and empirical issues involved in modeling complex travel patterns. Existing models have the shortcoming of not representing the interdependencies among trip links in trip chains with multiple non-home stops. A theoretical model based on utility theory and explicitly accounting for the trade-offs involved in the choice of multiple-stop chains is developed. Using this theoretical model, utility maximizing conditions for a household's choice of a daily travel pattern are derived. The optimum travel pattern is described in terms of the number of chairs (tours) traveled on a given day and in terms of the number of stops (sojourns) made on each of those chains. For a given household, the form of the optimum pattern is a function of the transportation expenditures (time, cost) required to reach potential destinations. Constraints on the conditions of optimality due to the limited and discrete nature of travel pattern alternatives are also considered. Parameters of the general utility function were estimated empirically using actual travel data derived from a home interview survey taken in Washington, D.C. The multinomial logit model is used to relate utility scores for the alternative travel patterns to choice probabilities. The resulting parameter estimates agree with theoretical expectations and with empirical results obtained in other studies. In order to demonstrate the empirical and theoretical implications of the model, forecasts for various transportation policies (e.g., gasoline price increases, transit fare reductions), as made by this model and by other less complex models, are compared. The results of these comparisons indicate the need for expanding the scope of existing travel forecasting models to explicit considerations of trip chaining behavior.     

Door-to-door travel times in RP departure time choice models: An approximation method using GPS data

A common way to determine values of travel time and schedule delay is to estimate departure time choice models, using stated preference (SP) or revealed preference (RP) data. The latter are used less frequently, mainly because of the difficulties to collect the data required for the model estimation. One main requirement is knowledge of the (expected) travel times for both chosen and unchosen departure time alternatives. As the availability of such data is limited, most RP-based scheduling models only take into account travel times on trip segments rather than door-to-door travel times, or use very rough measures of door-to-door travel times. We show that ignoring the temporal and spatial variation of travel times, and, in particular, the correlation of travel times across links may lead to biased estimates of the value of time (VOT). To approximate door-to-door travel times for which no complete measurement is possible, we develop a method that relates travel times on links with continuous speed measurements to travel times on links where relatively infrequent GPS-based speed measurements are available. We use geographically weighted regression to estimate the location-specific relation between the speeds on these two types of links, which is then used for travel time prediction at different locations, days, and times of the day. This method is not only useful for the approximation of door-to-door travel times in departure time choice models, but is generally relevant for predicting travel times in situations where continuous speed measurements can be enriched with GPS data.     

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.     

Impacts of highly automated vehicles on travel demand: macroscopic modeling methods and some results

Automated vehicles (AV) will change transport supply and influence travel demand. To evaluate those changes, existing travel demand models need to be extended. This paper presents ways of integrating characteristics of AV into traditional macroscopic travel demand models based on the four-step algorithm. It discusses two model extensions. The first extension allows incorporating impacts of AV on traffic flow performance by assigning specific passenger car unit factors that depend on roadway type and the capabilities of the vehicles. The second extension enables travel demand models to calculate demand changes caused by a different perception of travel time as the active driving time is reduced. The presented methods are applied to a use case of a regional macroscopic travel demand model. The basic assumption is that AV are considered highly but not fully automated and still require a driver for parts of the trip. Model results indicate that first-generation AV, probably being rather cautious, may decrease traffic performance. Further developed AV will improve performance on some parts of the network. Together with a reduction in active driving time, cars will become even more attractive, resulting in a modal shift towards car. Both circumstances lead to an increase in time spent and distance traveled.