Using non-real-time Automatic Vehicle Location data to improve bus services

Although real-time Automatic Vehicle Location (AVL) data is being utilised successfully in the UK, little notice has been given to the benefits of historical (non-real-time) AVL data. This paper illustrates how historical AVL data can be used to identify segments of a bus route which would benefit most from bus priority measures and to improve scheduling by highlighting locations at which the greatest deviation from schedule occurs. A new methodology which uses historical AVL data and on-bus passenger counts to calculate the passenger arrival rate at stops along a bus route has been used to estimate annual patronage and the speed of buses as they move between stops. Estimating the patronage at stops using AVL data is more cost-effective than conventional methods (such as surveys at stops which require much more manpower) but retains the benefits of accuracy and stop-specific estimates of annual patronage. The passenger arrival rate can then be used to calculate how long buses spend at stops. If the time buses spend at stops is removed from the total time it takes the bus to traverse a link, the remaining amount of time can be assumed to be the time the bus spends moving and hence the moving speed of the bus can be obtained. It was found that estimation of patronage and the speed of buses as they move between stops using AVL data produced results which were comparable with those obtained by other methods. However the main point to note is that this new method of estimating patronage has the potential to provide a larger and superior data set than is otherwise available, at very low cost.     

Passenger waiting time and information acquisition using automatic vehicle location for verification

Bus riders utilize a variety of information media to learn how to travel to their destinations and to learn when they should arrive at bus stops. As part of the OCTA (Orange County Transit Authority) Transit Probe evaluation, 1199 passengers were surveyed to measure relationships between information acquisition and waiting time. A unique aspect of the survey was that some of the data could be correlated with automatic‐vehicle‐location (AVL) measurements of bus lateness at stops. Insights are provided as to the types of information riders acquire based on the nature of the trip and demographic characteristics. Insights are also provided as to factors affecting perceived waiting time. We found age group, whether a person needs to arrive at a destination by a specific time, primary language, and whether the person is a first‐time user of the bus line to be significant causal factors.     

A congested and dwell time dependent transit corridor assignment model

This research proposes an equilibrium assignment model for congested public transport corridors in urban areas. In this model, journey times incorporate the effect of bus queuing on travel times and boarding and alighting passengers on dwell times at stops. The model also considers limited bus capacity leading to longer waiting times and more uncomfortable journeys. The proposed model is applied to an example network, and the results are compared with those obtained in a recent study. This is followed by the analysis and discussion of a real case application in Santiago de Chile. Finally, different boarding and alighting times and different vehicle types are evaluated. In all cases, demand on express services tends to be underestimated by using constant dwell time assignment models, leading to potential planning errors for these lines. The results demonstrate the importance of considering demand dependent dwell times in the assignment process, especially at high demand levels when the capacity constraint should also be considered. Copyright © 2017 John Wiley & Sons, Ltd.     

Estimation of train dwell time at short stops based on track occupation event data: A study at a Dutch railway station

Train dwell time is one of the most unpredictable components of railway operations, mainly because of the varying volumes of alighting and boarding passengers. However, for reliable estimations of train running times and route conflicts on main lines, it is necessary to obtain accurate estimations of dwell times at the intermediate stops on the main line, the so‐called short stops. This is a great challenge for a more reliable, efficient and robust train operation. Previous research has shown that the dwell time is highly dependent on the number of boarding and alighting passengers. However, these numbers are usually not available in real time. This paper discusses the possibility of a dwell time estimation model at short stops without passenger demand information by means of a statistical analysis of track occupation data from the Netherlands. The analysis showed that the dwell times are best estimated for peak and off‐peak hours separately. The peak‐hour dwell times are estimated using a linear regression model of train length, dwell times at previous stops and dwell times of the preceding trains. The off‐peak‐hour dwell times are estimated using a non‐parametric regression model, in particular, the k‐nearest neighbor model. There are two major advantages of the proposed estimation models. First, the models do not need passenger flow data, which is usually impossible to obtain in real time in practice. Second, detailed parameters of rolling stock configuration and platform layout are not required, which makes the model more generic and eases implementation. A case study at Dutch railway stations shows that the estimation accuracy is 85.8%–88.5% during peak hours and 80.1% during off‐peak hours, which is relatively high. We conclude that the estimation of dwell times at short stop stations without passenger data is possible. Copyright © 2016 John Wiley & Sons, Ltd.     

An application-oriented model of passenger waiting time based on bus departure time intervals

This paper develops an application-oriented model to estimate waiting times as a function of bus departure time intervals. Bus stops are classified into Type A and B depending on whether they are connected with urban rail transit systems. Distributions of passenger arrival rates are analyzed based on field data for Beijing. The results indicate that the best fits for the distribution of passenger arrival rates for Type A and B bus stops are the lognormal distribution and gamma distribution, respectively. By analyzing relationships between passenger arrival rates and bus departure time intervals, it is demonstrated that parameters of the passenger arrival rate distribution can be expressed by the average and coefficient of variation of bus departure time intervals in functional relationships. The validation shows that the model provides a reliable estimation of the average passenger waiting time based on readily available bus departure time intervals.     

An improved headway-based holding strategy for bus transit

Abstract

This paper presents an improved headway-based holding strategy integrating bus transit travel and dwelling time prediction. A support vector machine-based (SVM) model is developed to predict the baseline travel and dwell times of buses based on recent data. In order to reduce prediction errors, an adaptive algorithm is used together with real-time bus operational information and estimated baseline times from SVM models. The objective of the improved holding strategy is to minimize the total waiting times of passengers at the current stop and at successive stops. Considering the time-varying features of bus running, a ‘forgetting factor’ is introduced to weight the most recent data and reduce the disturbance from unexpected incidents. Finally, the improved holding strategy proposed in this study is illustrated using the microscopic simulation model Paramics and some conclusions are drawn.     

Rethinking bus punctuality by integrating Automatic Vehicle Location data and passenger patterns

This paper investigates punctuality at bus stops. Although it is typically evaluated from the point of view of bus operators, it must also account for users, as required in recent service quality norms. Therefore, evaluating punctuality at bus stops is highly important, but may also be a complex task, because data on both bus arrivals (or departures) and users must be taken into account and processed. Data on buses can be collected by Automatic Vehicle Location (AVL) systems, but several challenges must be addressed in order to use them effectively. Passengers data at bus stops cannot be derived from AVL, but they can be used to derive passenger patterns and need to be integrated into processed AVL data. This paper proposes a new punctuality measure defined as the fraction of passengers who will be served within an acceptably short interval after they arrive. A method is proposed to determine this measure: it provides (i) several rules to handle AVL collected data, (ii) a procedure integrating processed AVL data and potential passengers’ patterns and (iii) a hierarchical process to perform the punctuality measure on each bus route direction of a transit network, as well as for every bus stop and time period. The paper illustrates the experimentation of this method on more than 4,000,000 data of a real bus operator and represents outcomes by easy-to-read control dashboards.     

Passengers' perceptions and effects of bus‐holding strategy using automatic vehicle location technology

This study reports bus passengers' behavior and perceptions related to the use of potential features of an automatic vehicle location (AVL) system in bus transit through conducting an attitudinal on‐board survey in Bangkok. A passenger waiting‐time survey conducted as part of this study revealed that passengers perceive waiting‐time at bus stops to be greater than actually experienced. The other aim of this study is to examine the potential benefits of bus‐holding using an AVL technology, in terms of waiting‐time, through minimizing bus bunching under different congestion levels. The results are obtained using PARAMICS, and reveal a significant reduction in average waiting‐time.     

Using survival models to estimate bus travel times and associated uncertainties

Transit agencies often provide travelers with point estimates of bus travel times to downstream stops to improve the perceived reliability of bus transit systems. Prediction models that can estimate both point estimates and the level of uncertainty associated with these estimates (e.g., travel time variance) might help to further improve reliability by tempering user expectations. In this paper, accelerated failure time survival models are proposed to provide such simultaneous predictions. Data from a headway-based bus route serving the Pennsylvania State University-University Park campus were used to estimate bus travel times using the proposed survival model and traditional linear regression frameworks for comparison. Overall, the accuracy of point estimates from the two approaches, measured using the root-mean-squared errors (RMSEs) and mean absolute errors (MAEs), was similar. This suggests that both methods predict travel times equally well. However, the survival models were found to more accurately describe the uncertainty associated with the predictions. Furthermore, survival model estimates were found to have smaller uncertainties on average, especially when predicted travel times were small. Tests for transferability over time suggested that the models did not over-fit the dataset and validated the predictive ability of models established with historical data. Overall, the survival model approach appears to be a promising method to predict both expected bus travel times and the uncertainty associated with these travel times.     

Time of day intervals partition for bus schedule using GPS data

Time of day partition of bus operating hours is a prerequisite of bus schedule design. Reasonable partition plan is essential to improve the punctuality and level of service. In most mega cities, bus vehicles have been equipped with global positioning system (GPS) devices, which is convenient for transit agency to monitor bus operations. In this paper, a new algorithm is developed based on GPS data to partition bus operating hours into time of day intervals. Firstly, the impacts of passenger demand and network traffic state on bus operational performance are analyzed. Then bus dwell time at stops and inter-stop travel time, which can be attained based on GPS data, are selected as partition indexes. For buses clustered in the same time-of-day interval, threshold values of differences in dwell time at stops and inter-stop travel time are determined. The buses in the same time-of-day interval should have adjacent dispatching numbers, which is set as a constraint. Consequently, a partition algorithm with three steps is developed. Finally, a bus route in Suzhou China is taken as an example to validate the algorithm. Three partition schemes are given by setting different threshold values for the two partition indexes. The present scheme in practice is compared with the three proposed schemes. To balance the number of ToD intervals and partition precision, a Benefit Evaluation Index is proposed, for a better time-of-day interval plan.     

Integrated optimization of bus priority operations in connected vehicle environment

Most previous works associated with transit signal priority merely focus on the optimization of signal timings, ignoring both bus speed and dwell time at bus stops. This paper presents a novel approach to optimize the holding time at bus stops, signal timings, and bus speed to provide priority to buses at isolated intersections. The objective of the proposed model is to minimize the weighted average vehicle delays of the intersection, which includes both bus delay and impact on nearby intersection traffic, ensuring that buses clear these intersections without being stopped by a red light. A set of formulations are developed to explicitly capture the interaction between bus speed, bus holding time, and transit priority signal timings. Experimental analysis is used to show that the proposed model has minimal negative impacts on general traffic and outperforms the no priority, signal priority only, and signal priority with holding control strategies (no bus speed adjustment) in terms of reducing average bus delays and stops. A sensitivity analysis further demonstrates the potential of the proposed approach to be applied to bus priority control systems in real‐time under different traffic demands, bus stop locations, and maximum speed limits. Copyright © 2016 John Wiley & Sons, Ltd.     

Bus arrival time prediction at bus stop with multiple routes

Provision of accurate bus arrival information is vital to passengers for reducing their anxieties and waiting times at bus stop. This paper proposes models to predict bus arrival times at the same bus stop but with different routes. In the proposed models, bus running times of multiple routes are used for predicting the bus arrival time of each of these bus routes. Several methods, which include support vector machine (SVM), artificial neural network (ANN), k nearest neighbours algorithm (k-NN) and linear regression (LR), are adopted for the bus arrival time prediction. Observation surveys are conducted to collect bus running and arrival time data for validation of the proposed models. The results show that the proposed models are more accurate than the models based on the bus running times of single route. Moreover, it is found that the SVM model performs the best among the four proposed models for predicting the bus arrival times at bus stop with multiple routes.     

Crowding in public transport systems: Effects on users,operation and implications for the estimation of demand

The effects of high passenger density at bus stops, at rail stations, inside buses and trains are diverse. This paper examines the multiple dimensions of passenger crowding related to public transport demand, supply and operations, including effects on operating speed, waiting time, travel time reliability, passengers’ wellbeing, valuation of waiting and in-vehicle time savings, route and bus choice, and optimal levels of frequency, vehicle size and fare. Secondly, crowding externalities are estimated for rail and bus services in Sydney, in order to show the impact of crowding on the estimated value of in-vehicle time savings and demand prediction. Using Multinomial Logit (MNL) and Error Components (EC) models, we show that alternative assumptions concerning the threshold load factor that triggers a crowding externality effect do have an influence on the value of travel time (VTTS) for low occupancy levels (all passengers sitting); however, for high occupancy levels, alternative crowding models estimate similar VTTS. Importantly, if demand for a public transport service is estimated without explicit consideration of crowding as a source of disutility for passengers, demand will be overestimated if the service is designed to have a number of standees beyond a threshold, as analytically shown using a MNL choice model. More research is needed to explore if these findings hold with more complex choice models and in other contexts.     

The effects of AVL accuracy upon public service bus system performance

Through an examination of the dependence of several key performance parameters of a public bus system upon Automatic Vehicle Location (AVL) estimation accuracy, it becomes possible to place in perspective the value of AVL technology to improved public conveyance performance. Important bus transportation system performance measures dependent upon AVL estimation accuracy are: (1) Headway Control Error; (2) Time-Of-Passage Error, and (3) Required Layover Reserve. An analytical model of the dependence of these bus system performance measures upon AVL estimation error has been constructed. In addition, error models of three basic types of AVL systems, i.e., dead reckoning, proximity, and radio location have been developed and validated by experimental comparisons. By employing both sets of models, i.e. for the bus transport and AVL systems it becomes possible to recommend appropriate AVL technologies that best meet the performance requirements of a public bus service. The accompanying text synopsizes the noted models and provides an example of their use.     

Estimating bus run times for new limited-stop service using archived AVL and APC data

In recent years, several transit agencies have been trying to be more competitive with the automobile to attract choice riders. Transit agencies can only be competitive if they can provide services that are reliable, have a short access and egress time, and have run times that are comparable to automobiles. Several transit agencies try to be competitive through offering faster service, such as limited-stop (express) bus service. This study uses AVL and APC data, in addition to a disaggregate data obtained from a travel behavior survey, to select stops and estimate run times for a new limited-stop service that will run parallel to a heavily used bus route (67 Saint-Michel) in Montréal, Canada. Three different scenarios are developed based on theory and practice to select stops to be incorporated in the new limited service. The time savings for each scenario are then evaluated as a range and a fourth scenario is developed. A limited-stop service is recommended based on selecting stops serving both directions of the route, major activity points and stop spacing. This study shows that implementing a limited-stop service would yield substantial time savings for both, the new limited service and the existing regular service running in parallel.     

Evaluate bus emissions generated near far-side and near-side stops and potential reductions by ITS: An empirical study

In this paper, we use second-by-second automatic vehicle location data to estimate bus emissions near far-side and near-side stops. We classify the bus running state near a stop into approach, dwell, and departure. A vehicle specific power approach is used to estimate bus emissions for each state. We show that bus emissions generated near stops can be significantly reduced by using certain intelligent transportation systems techniques.     

A model of bus bunching under reliability-based passenger arrival patterns

If bus service departure times are not completely unknown to the passengers, non-uniform passenger arrival patterns can be expected. We propose that passengers decide their arrival time at stops based on a continuous logit model that considers the risk of missing services. Expected passenger waiting times are derived in a bus system that allows also for overtaking between bus services. We then propose an algorithm to derive the dwell time of subsequent buses serving a stop in order to illustrate when bus bunching might occur. We show that non-uniform arrival patterns can significantly influence the bus bunching process. With case studies we find that, even without exogenous delay, bunching can arise when the boarding rate is insufficient given the level of overall demand. Further, in case of exogenous delay, non-uniform arrivals can either worsen or improve the bunching conditions, depending on the level of delay. We conclude that therefore such effects should be considered when service control measures are discussed.     

Prediction of ambient carbon monoxide concentration using nonlinear time series analysis technique

This study evaluates the potential of nonlinear time series analysis based methods in predicting the carbon monoxide concentration in an urban area. To establish the functional relationship between current and future observations, two models based on local approximations and neural network approximations are used. To compare the performance of the models, an autoregressive integrated moving average model is also applied. The multi-step forecasting capabilities of the models are evaluated.     

Evaluating the effects of bus design on passenger flow: Is agent-based simulation a feasible approach?

In this paper, we propose an agent-based simulation approach that is capable of simulating the flow of passengers on board buses and at bus stops. The intention is that it will be applied during vehicle development to analyze how vehicle design affects passenger flow, and thus also how it affects system performance such as dwell time. In turn, this could aid the developers in making design decisions early in the development process. Besides introducing the simulation tool itself, the paper explores the realism of the data generated by the tool. A number of passenger flow experiments featuring a full-scale bus mockup and 50 participants were carried out. The setup of these experiments mirrored a number of ‘bus journeys’ (regarding vehicle design, number of passengers boarding/alighting at each stop and so on) that had previously been simulated using the simulation tool. When the data from the simulations were compared with the data from the passenger flow experiments, it could be concluded that the tool is indeed able to generate realistic passenger flows, although with some errors when a large number of passengers board/alight. The simulated dwell times were rationally affected by the tested bus layout aspects. It was concluded that the tool makes it possible to evaluate how variations in bus layouts affect passenger flow, providing data of sufficiently high quality to be useful in early phases of vehicle design.     

Predicting short-term bus passenger demand using a pattern hybrid approach

This paper proposes an Interactive Multiple Model-based Pattern Hybrid (IMMPH) approach to predict short-term passenger demand. The approach maximizes the effective information content by assembling the knowledge from pattern models using historical data and optimizing the interaction between them using real-time observations. It can dynamically estimate the priori pattern models combination in advance for the next time interval. The source demand data were collected by Smart Card system along one bus service route over one year. After correlation analysis, three temporal relevant pattern time series are generated, namely, the weekly, daily and hourly pattern time series. Then statistical pattern models are developed to capture different time series patterns. Finally, an amended IMM algorithm is applied to dynamically combine the pattern models estimations to output the final demand prediction. The proposed IMMPH model is validated by comparing with statistical methods and an artificial neural network based hybrid model. The results suggest that the IMMPH model provides a better forecast performance than its alternatives, including prediction accuracy, robustness, explanatory power and model complexity. The proposed approach can be potentially extended to other short-term time series forecast applications as well, such as traffic flow forecast.