Abstract: In this paper, the authors summarize work concerned with the simulation of large road networks using a cellular automaton (CA) model. A discussion of road traffic simulations is presented, and the cellular automaton model is introduced. Some results are presented, showing that this approach is sufficient to reproduce characteristics of real traffic. The paper concludes with a discussion of how the simulation code can be parallelized.
Abstract: SISTM (Simulation of Strategies for Traffic on Motorways) is a microscopic motorway simulation programme for investigating alternative traffic management measures. It has been enhanced to model variable speed limit systems such as the system currently in operation on part of the M25. Version 4.1 of the software can model different parmeter values for the flow thresholds, the flow measurement periods and smoothing factors. Traffic data from this section has been used to predict the effect of 50 and 60mph speed limits on the target speed of drivers in SISTM. This paper describes SISTM's speed control modelling and how it is calibrated for speed control operations.
Abstract: This article describes microscopic traffic simulation with fuzzy logic.
Abstract: This paper presents an application of microscopic traffic modeling to perform real time highway traffic simulation and prediction. The simulation uses inductance loop data as a model input and predicts the downstream traffic based on a microscopic model. The preliminary results show high cross-correlation between the prediction and the real data.
Abstract: This paper discusses issues associated with accurate modeling of microscopic driver behavior. It also examines the degree of shortfall present incurrent simulation and analysis techniques, and potential guidelines for increasing the validity of microscopic simulation models are suggested.
Abstract: In this paper, the authors analyze weaving patterns and large changing behavior of vehicles on a highway interchange using a microscopic simulation model employing a car following theory. The model is described by three basic equations consisting of the simple forward movement, the following movement, and the stopping movement.
Abstract: The objective of this paper is to demonstrate the extent to which the results produced by the INTEGRATION traffic-simulation model, an automated tool capable of capturing travel-time impacts of Intelligent Vehicle Highway Systems (IVHS) technologies, match those estimated by the standards of the Environmental Protection Agency (EPA). The intent is to establish the level of credibility of the INTEGRATION fuel-consumption and emission estimates prior to evaluating IVHS technologies for their environmental impact using the model.
Abstract: As the initial step in formalizing the TRANSIMS design, this paper translates the interpretation of the analytical needs that are explicitly or implicitly derived from recent federal legislation into high-level simulation design. This document serves as a basis for feedback from the user community regarding the overall policy environment. The document is directed at metropolitan planning office personnel, federal and state departments of transportation and air quality personnel, municipal and county planners and other stakeholders in the air quality and transportation planning communities.
Abstract: Computer simulation has become a very powerful decision aid for varied facets of traffic engineering. Simulation experiments are often used to fit a metamodel of interest between the mean response and a selected set of input factors. This is done by carefully designing statistical experiments under alternative system designs, which are referred to as multipopulation simulation experiments. Validation and statistical analysis procedures are presented on linear metamodels from multipopulation traffic simulation networks under the common random number (CRN) strategy on three sample networks using the TRAF-NETSIM model. Under the CRN strategy, positive correlations are induced among the observations, and hence the usual statistical analysis cannot be applied to obtain point estimates and confidence intervals; therefore it must be modified. Before the statistical analysis is conducted, certain assumptions of the CRN strategy should be validated--those that, if violated, render the modified statistical analysis invalid.
Abstract: This paper discusses the advantages of object-oriented rule-based modelling in traffic simulation based on experience gained in the development of the HUTSIM traffic signal simulation system in Helsinki. The basic features of HUTSIM are object-oriented rule-based simulation of vehicle and control interactions and precise presentation of the signal control, the latter being realised by connecting a real adaptive signal controller to the micro-computer-based simulation system. Principles behind object-oriented modelling and elements of the model are discussed. Rule-based and state-based vehicle dynamics are described. The interactive role of graphics is considered. It is suggested that object-oriented programming with rule-based presentation is particularly suitable for traffic simulation.
Abstract: A day-to-day dynamic framework, in which the DYNASMART simulation assignment model was applied to evaluate the performance of traffic networks, was developed to study network dynamics under different information systems. Two levels of tripmaker decision-making processes are identified: (a) day-to-day dynamics and (b) real-time dynamics. Day-to-day dynamics consider the choices of departure time and route according to indifference bands of tolerable "schedule delay" defined as the difference between the user's actual and preferred arrival times. Real-time dynamics consider en route switching decisions. Numerical experiments were conducted to investigate the day-to-day evolution of network flows under real-time information and assess the effectiveness of such information in a proper dynamic perspective.
Abstract: A new type of probabilistic cellular automaton for the physical description of single and multilane traffic is presented. In this model space, time and the velocity of the cars are represented by integer numbers (as usual in cellular automata) with local update rules for the velocity. The model is very efficient for both numerical simulations and analytical investigations. The numerical results from extensive simulations reproduce very well data taken from real traffic (e.g. fundamental diagrams). Several analytical results for the model are presented as well as new approximation schemes for stationary traffic. In addition the relation to continuum hydrodynamic theory and the follow-the-leader models is discussed. The model is part of an interdisciplinary research program in Northrhine-Westfalia for the construction of a large scale microsimulation model for network traffic, supported by the government of NRW.
Abstract: Several recent strategic studies have used linear relationships between system speed (veh-km/veh-h) and veh-km to represent the demand-supply interaction. These "area speed-flow relationships" are based on empirical work in the 1970s, but there has been no detailed investigation of the way in which such relationships are generated, or how they are affected by the nature of the network or the pattern of demand. This paper discusses the background to the problem, and presents results of an EPSRC funded study which is investigating the interaction between system speed and vehicle kilometres in different types of network. Four relationships are being obtained, between (i) speed (veh-km/veh-h) and actual flow (veh-km/h), (ii) speed (veh-km/veh-h) and demanded flow (veh-km/h), (iii) time/km (veh-h/veh-km). The data is collected directly from the micro-simulation package NEMIS for an origin-destination matrix which is factored to represent different levels of demand. The modelling and data collection methodology is described and results for various networks presented. The implications of these results are discussed in terms of definition of network capacity and in implications for strategic modelling and evaluation.
Abstract: This paper presents the use of an extensively modified version of the micro-simulation model TRAF-NETSIM, to represent the performance of vehicles at road junctions in Bangkok, Thailand. In particular, it considers the tendency of vehicles to pass through gaps, and to swing around the backs of traffic queues in cross-streets. Tests of the modified TRAF-NETSIM on a hypothetical network showed that it generally worked as expected. Both versions of TRAF-NETSIM were applied to a model of the Bangkapi area of Bangkok. The street network there has large signalised junctions, with up to 20 entry lanes, and cycle times up to 7-10min during peak periods. One of the main traffic problems is blocked upstream junctions. The simulation outputs were investigated for average speeds and average and maximum queues. It was shown that: (1) the modified TRAF-NETSIM provides results closer to observed data than those from the standard TRAF-NETSIM; and (2) simulation of the actually occurring unequal cycle times and irregular stages requires further research. The study showed the need to adapt 'standard' packages to allow for local road network characteristics and driver behaviour, if meaningful results are to be obtained.
Abstract: This seminar includes papers on traffic speeds and flows, data collection and errors, traffic assignment, activity models, models incorporating search and learning, microsimulation, advanced discrete choice models, evaluation (of bus priority schemes), and modelling responses to ATT.
Abstract: This paper presents a method to manage an urban environment complex database, making use of hierarchical elision techniques over a spatial connectivity network. It also shows how to adapt and refine traffic microsimulation algorithms to provide visual quality in a local area.
Abstract: In this paper, the authors examine a number of problems surrounding the use of microscopic highway simulation models. They then identify the next steps required to ensure a greated degree of validity in the models.
Abstract: An exploratory procedure for analyzing the permitted left-turn capacity with exclusive lanes is presented, including several empirical models for opposing queue length prediction, permitted saturation flow of mixed traffic, and effect of bay length on the left-turn capacity. Some critical factors, such as the number of opposing lanes and the interactions between upstream and downstream green time-cycle length ratios, have been incorporated in the proposed procedures for capacity estimation. A discrete choice modeling methodology has been applied to predict the fraction of time in a cycle during which the through queue length may be over a certain distance. Such a model enables traffic engineers to determine the left-turn bay length from a cost-benefit perspective. It should be noted that all proposed empirical models are grounded on the simulation experiments with TRAF-NETSIM. Hence, adjustments or modifications may be necessary after extensive field observations are conducted to calibrate TRAF-NETSIM.
Abstract: The effectiveness of control strategies applied to alleviate traffic congestion depends heavily on the accuracy and credibility of the data sources used. Among data sources now available, loop detector systems can provide large quantities of high-quality data. The feasibility of using detector data to improve the performance of advanced traveler information systems functions, such as arterial travel time estimation, is examined. The relationships between travel times and flow/occupancy information are assessed using simulation techniques and field data. A common study area is used for both types of experiments. The NETSIM model was selected as the best simulation tool available. Recent enhancements of the model, expected to be incorporated in its next release allow for the simulation of surveillance detector information such as vehicle counts, percentage occupancy values, and average spot speed. Several experiments are performed to incorporate variations in entry flows and turning movement percentages, as well as the randomness of traffic phenomena. Besides the simulation experiments, field studies are carried out as part of a validation effort. These studies include on-site travel time data collection and concurrent detector output consideration for detectorized links in the study area. The explanatory analysis presented indicates that both approaches support the following conclusions: (a) travel time is independent from both flow and occupancy under conditions of low traffic demand, and (b) generalized regression equations can be fitted for certain ranges of occupancies to properly model the relationships between travel time and detector data.
Abstract: Stochastic traffic simulation models, such as TRAF-NETSIM, use random number seeds to generate variables to describe driver, roadway, and traffic characteristics. In analyzing outputs from these models, one should consider the variability of the responses. The variability of NETSIM's output using the methods of replication and batch means was explored. For the batch means method, it is proposed to compute the measures of effectiveness (MOEs) for intermediate time intervals using a proposed interval calculation (PIC) procedure. The MOEs were evaluated at the network, intersection, and link levels of aggregation. Depending on the MOE and level of aggregation, the two methods yielded significantly different results. Hence, depending on the study objective, outputs may need to be examined at different levels of aggregation to obtain meaningful results. The practical implications of the variability are also discussed, and statistical approaches are proposed to deal with output variability. Auto- and cross-correlations must be examined explicitly, particularly when dealing with link MOEs resulting from very short simulation time. Ignoring positive cross-correlation is not detrimental but leads to more conservative confidence intervals. Either the batch means with PIC method or replication method must be used to build confidence intervals. NETSIM's direct output for intermediate time intervals should not be used to build a confidence interval unless an autocorrelation analysis is done. Not using proper statistical procedures can lead to erroneous and misleading conclusions.
Abstract: An integrated real-time ramp metering model for nonrecurrent freeway congestion among link flows has been developed and tested in this study. The core concept of the proposed algorithm is to capture the dynamic traffic state evolution with a two-segment linear flow-density model. To be implemented in real time, an effective solution algorithm has been proposed for determining the time-varying metering rates. The entire algorithm has also been integrated with INTRAS, the most well-known freeway simulation model, for conducting simulation experiments. Preliminary research results indicate that the proposed integrated control model is promising because its effectiveness increases with the severity of accidents and the level of congestion. The model execution time is also sufficiently short for potential real-time operations.
Abstract: Over the past three decades, considerable ramp metering algorithms of both time-of-day and traffic-responsive types have been proposed and tested. The time-of-day algorithms determine the pre-timed metering rates based on regular demand and roadway traffic pattern, hence they are not effective in dealing with non-recurrent congestions. An appropriate trade-off must be made between the model complexity and the availability of solution algorithms. With this attempt, an integrated real-time ramp metering model for non-recurrent freeway congestion among link flows has been developed and tested in this study. The core concept of the proposed algorithm is to capture the dynamic traffic state evolution with a two-segment linear flow-density model. The entire algorithm has also been integrated with INTRAS, the most well-known freeway simulation model, for conducting simulation experiments.
Abstract: CORFLO is a strong tool for evaluating coordinated freeway corridor traffic management strategies because of its capability of explicitly simulating freeways and surface streets within a single environment. The lack of a fuel consumption estimation capability in FREFLO, the freeway simulation component of CORFLO, however, limits its application for evaluating corridor management strategies directed at the much needed conservation of scarce petroleum resources. The objective was to develop and implement a fuel consumption model based on the most currently available fuel consumption data. A review of previous work on fuel consumption modeling indicated that a model based on the energy consumed to overcome the forces resisting a vehicle's motion is suitable for implementation in FREFLO. The relation between acceleration noise and density was used to estimate the acceleration owing to vehicle interaction. The model was calibrated by using fuel consumption data obtained from FHWA; these data are representative for 64% of the passenger vehicle fleet from 1980 to 1992. The model explains 99.5% of the variation in the constant-speed fuel consumption data and 86.1% of the variation in fuel consumption due to acceleration. Air drag and rolling friction coefficients computed from the model parameters were near the lower bounds of the range of theoretically possible values. The fuel consumption estimates obtained from FREFLO and INTRAS for a small segment of I-35 near Austin, Texas, were comparable under most traffic conditions.
Abstract: Since its inception in the early 1980s, the Integrated Traffic Simulation (INTRAS) model has been used in many studies involving freeway corridor traffic simulation. The model was originally calibrated with data collected in the 1970s in Los Angeles. In view of changing traffic conditions during the past decade, the validity of the parameter values as calibrated in the original setting is questionable. In several recent studies that used INTRAS as an evaluation tool, the model has been recalibrated with recent data. However, because of the different applications of the INTRAS model in these studies, the calibrations were made with output averaging at longer time intervals and for different output variables. To simulate traffic operation on Southern California freeways consistent with surveillance data currently being collected by the California Department of Transportation in traffic operations centers, INTRAS has been calibrated with respect to loop detector data at 30-sec intervals. The calibration process involved traffic during conditions with and without incidents, based on data collected along a 5-mi section of a major freeway in Orange County. Key parameters calibrated in this study include car-following sensitivity constants, minimum car-following distance, vehicle lengths, effective detector lengths, and the INTRAS "rubbernecking factor". The calibrated model has been used to simulate detector data for evaluating incident detection algorithms and for training artificial neural network models to detect freeway incidents.
Abstract: Systematic scientific research on highway capacity started in Finland during the 1960s. From the beginning, vehicle speeds have been actively studied. In the 1970s a wide before-and-after analysis of general road section speed limits was done followed by an analysis of lowered wintertime speed limits during the late 1980s and early 1990s. Highway capacity, on two-lane roads and on motorways, has been actively researched especially in the 1980s and 1990s. At the moment three-lane highways (with the centre lane given alternately for one direction of flow only) and exceptionally widelanes (5.25 m) on two-lane highways are being demonstrated and studied. A special project of wintertime effects on traffic is going on and a traffic monitoring system of 180 automatic speed and flow measuring points is in use for census and research purposes. New research on signalised and non-signalised intersections is starting, partly to update old capacity values, and partly to calibrate and validate the Finnish HUTSIM microscopic simulation program. Interest in simulation is growing, and among others, the Australian TRARR program has been used in traffic flow research.
Abstract: The operation of the HUTSIM-traffic signal simulator is based on the basic rules of dynamics. The behaviour of the vehicles is regulated by the microlevel parameters. The calibration and validation study was divided into the definition of the parameters and the comparison of the macrolevel results.
Abstract: Several optimization models have been proposed for the dynamic traffic assignment problem. However, little computational experience is available. In this paper, initial work to fill this gap by formulating and solving a class of discrete time, nonlinear models is presented. An exit function is proposed, and its effects on traffic flow are analyzed using small traffic networks. The performance of three widely available optimizers in solving these models is described. Solution features of both nonlinear and piecewise linear versions of the model are presented and compared with simulation results generated using the Dynasmart simulator.
Abstract: This paper summarizes a report on the testing, calibration and validation of FRESIM and NETSIM traffic simulation models. The process for each model is described separately although extensive similarities exist between these processes for both models.
Abstract: This paper describes the use of microscopic simulation models to solve complex design problems at freeway interchanges. The application of micro-simulation is illustrated by examples of three design problems: (1) Complex cross weave; (2) Complex multi-lane merge; and (3) High volume off-ramp in a weaving section. These examples are all taken from the design of the systems interchange between I-94 and I-894/USH 45, the "Zoo Interchange" in Milwaukee, Wisconsin.
Abstract: This paper summarises the findings of an initial study in which neural networks were developed for several kinds of transport problem. The first problem addressed was that of recognising whether the road system is in a particular state (e.g. that certain links are "congested"). The second problem was similar, but concerned with short term forecasting from real time data; some comparisons with Box-Jenkins methods are reported. Because certain traffic states or parameters (e.g. queue lengths) are not readily measurable on-street, an experiment is also reported where a neural network is used to infer relationships between these parameters and more readily measurable quantities. Data for this experiment were provided by micro-simulation. Such modelling capabilities were investigated further in the fourth area, by applying neural networks to complex multivariate data (here a computer-based survey of drivers' route choice); it was found that neural networks provided an effective and quicker alternative to logit models of individual choice. However, there remained the worry that the relationships derived are not made explicit by the neural network. This shortcoming was circumvented in the fifth study, in which it was shown that by dithering the data inputs, useful insights were obtained into the underlying structure of the relationships.
Abstract: This paper reports on a simulation-assignment framework for the analysis of time-dependent vehicular trips in a congested urban road network. The core is a simulator (DYNASMART) that allows consideration of individual tripmaker decisions at the origin as well as at each node along the way to the destination, in addition to describing the dynamics of vehicular traffic flow. The simulator is incorporated in an algorithmic framework for solving both system-optimal and user-equilibrium time-dependent assignment of trips to the network. Alternatively, assignment may be determined entirely by the specified user choice rules. Particularly useful is a formulation and solution algorithm for a network with multiple user classes possibly following different assignment principles and/or receiving different types of information (or none at all). Results of computer implementation of the procedures on test networks are included to illustrate their applicability to evaluate transportation system improvements and information supply strategy.
Abstract: To comply with requirements of the 1990 Clean Air Act, the Maine Department
of Transportation (MDOT) released a two phase Request For Proposal in early
June of 1992. Phase I of the project was to establish 1990 baseline levels of
vehicular congestion as measured by total vehicle delay (in vehicle-hours) at
372 signalized intersections. These intersections are located in the
Environmental Protection Agency (EPA) designated nine county clean air
non-attainment area within the State. Phase 2 of the project was to recommend
and quantify traffic type improvements designed to reduce congestion and obtain
a 15 percent decrease in total vehicle delay.
Abstract: This paper chronicles the calibration of NETSIM, a traffic network simulation model, in the CBD of San Antonio, Texas, with the use of the two-fluid model; a procedure, which, until now, had been performed only with a hypothetical network. Turning movement counts were used for input to NETSIM, and field studies were conducted to calibrate the two-fluid model for the San Antonio CBD. NETSIM was considered calibrated when a series of runs over all volume conditions resulted in the same (or similar) two-fluid parameters. Several features of NETSIM were used in its calibration: creation of new vehicle classes based upon differing maximum accelerations, maximum speeds, and queue discharge characteristics; parking intensity and short-term events. By far, the most effective tool in calibrating NETSIM in this network was the addition of new vehicle classes.
Abstract: Traffic safety is an important concept in the evaluation of a transportation system and its impact on public health. Using reported crash rates as an indicator of safety of a freeway facility has many drawbacks, such as errors in the reporting and recording of crashes, inaccuracies in the way in which the exposure measure is derived, and the wait involved for a sufficient sample size to materialize. Conflict rates provide an alternative to crash rates as an indicator of safety. Benefits of their use include the ease and accuracy with which conflict rates at ramp weaves can be obtained and the high frequency at which they occur with no physical harm to the public. Computer subroutines were added to the Integrated Traffic Simulation (INTRAS) to count conflicts, and freeway traffic was simulated at 10 modeled ramp weaves on Interstate 294 (Interstate 294 serves as a quasi-beltway for the Chicago metropolitan area). The resulting conflict rates were then noted. Volume, geometric, and crash data for the 10 sites were provided by the Illinois State Toll Highway Authority and on-site visits. The conflict rates at the sites were applied in a test of their ability to identify the known hazardous ramp weaves, and the relationship between conflict and crash rates was examined.
Abstract: The current procedures for analyzing freeway weaving sections with four or more directional freeway lanes contain approximations and estimates based on observed freeways with fewer lanes. The use of larger cross-section freeways to cope with congestion reinforces the need for updated ramp merge/diverge and weaving area analysis procedures. This study documents the inappropriate results provided by the 1985 Highway Capacity Manual method of analysis. The new FRESIM microscopic freeway simulation model under development by the Federal Highway Administration provides reasonable estimates of merge/diverge and weaving area operations. This study substantiates the validity of the FRESIM model to predict the general performance of four and five lane freeways in Texas. Finally, recommendations for use of auxiliary lanes in Texas, based on FRESIM simulations, are presented in a tabular "operational matrix".
Abstract: This paper investigates the performance of a traffic network under real-time information, under three distinct assignment rules reflecting different information supply strategies and associated user behaviour principles: (1) descriptive real-time information with boundedly-rational path switching rules; (2) time-dependent system optimal assignment, corresponding to normative route guidance information; and (3) time-dependent user equilibrium assignment, reflecting possibly the long term evolution of a system under descriptive information. The investigation uses a modelling framework (DYNASMART) for the assignment of time-varying trip desires to the network and the simulation of the resulting traffic patterns in the network. A principal feature is that vehicle paths in the network are modelled explicitly as the outcome of individual path selection decisions at each mode in the network. Traffic flow is represented using a hybrid approach where vehicles are tracked individually and/or in macroparticles, and moved consistently with macroscopic traffic flow relations between speed and density. Junction delays are explicitly modelled. Vehicles are routed in the network according to individual decisions made at decision points under real-time information availability for equipped vehicles. The simulator is integrated in a new algorithmic procedure for the solution of the time-dependent system optimal and user equilibrium assignment problems. Results of a set of experiments are used to investigate the effectiveness of real-time information systems and the key parameters that affect network performance, such as the overall loading level, the temporal loading pattern, and market penetration. It is shown that system optimal assignment offers meaningful improvement in total system cost over the other strategies, especially as congestion increases in the network.
Abstract: Presents a study into traffic safety and efficiency in adverse weather conditions. A model is derived for car-following using optimal control theory. The results can be used for predictive studies on microscopic traffic behavior as well as for in-vehicle control. A traffic simulation model is built, based on literature sources and model analysis. This model incorporates the influences of reduced friction and visibility. Simulation of a sudden visibility reduction shows that road capacity and traffic safety are both decreased.
Abstract: Computer simulation is now used routinely as a decision support tool in many facets of transportation engineering. Simulation models are often used to compare system alternatives, such as alternative system designs and/or operating policies. The presence of high variance of the simulation output variables is a critical problem in such comparative analyses because the models must either be run longer or executed several times to achieve reasonably accurate point and interval estimates of the parameters of interest. The high variability in output measures can also lead to concern about the validity of the model when they are used by the practitioners who are not intimately familiar with the stochastic nature of the model processes. This paper describes and illustrates the effectiveness of variance reduction based on the concept of common random numbers (CRN) for the TRAF-NETSIM simulation model.
Abstract: NETSIM, a very detailed computer model for traffic simulation developed in the United States, is validated and calibrated for use in urban street networks in New Zealand conditions. It is particularly valuable for dense and complex networks where other methods are seldom appropriate. The development, logic and structure of NETSIM is documented. Each submodel is examined, and its associated requirements for sensitivity, validation and calibration, as well as its embedded parameters are identified. Although more field calibrations and analyses of other than the four submodels considered for this report are required, most of the submodels tested were found to be valid. It is concluded that NETSIM can be usefully applied with its embedded parameters to New Zealand conditions. Recommendations for further research and development of NETSIM for use by local authority practitioners and consultants are supplied.
Abstract: Research for the project, "Urban road traffic models for economic appraisal", identified and developed a range of traffic analysis software suitable for New Zealand use to evaluate urban road improvement schemes. Outputs of the models, in the form of traffic performance characteristics, are used to estimate expected economic benefits from alternative options for a road improvement scheme. Criteria were defined for evaluating existing traffic analysis software. Overseas and New Zealand software were then appraised, and detailed analysis of the selected software was conducted. Research results summarise the analysis and evaluation of existing computer-based packages for intersections, arterial networks, motorways, small area networks and city-wide networks. They also summarise the in-depth investigations of existing models such as NETSIM and MULATM. A summary of theoretical work and accompanying references is given, together with an outline of the program MULDEL. This program was developed to estimate delay at a priority intersection based on the user optimal principle for lane choice. Economic evaluation obtained from traffic models, including intersection, small area network and city-wide network models, as well as directions for future research required for intermeshing economic evaluation and traffic models are summarised.
Abstract: The effectiveness of variance reduction techniques that users can apply to improve the efficiency and reliability of simulation experiments with the TRAF-NETSIM simulation model is described and illustrated. The two variance reduction techniques, antithetic variates and common random numbers, reduce the variance of simulation output by replacing the original sampling procedure by a new procedure that yields the same parameter estimate but with a smaller variance. Thus, the users can obtain greater statistical accuracy for the same number of simulation runs. A recent modification of the stochastic sampling process has made the TRAF-NETSIM model amenable to these variance reduction techniques and allows the users to apply these techniques with minimal additional effort. The effectiveness of these techniques is evaluated through an analysis of simulation output data from a TRAF-NETSIM case study. The estimated values and variances are computed for some representative measures of effectiveness after 10, 20, and 30 replications. The results indicate that both techniques are effective in reducing variance of the model output. By using the variance reduction techniques, the variance of parameter estimates is reduced on the average by 65% in the 24 comparisons that are made. The common random numbers strategy is more effective than the antithetic variates procedure. Over 50% reduction in variance is obtained using the common random numbers strategy in all comparisons and 80% or more in 6 of the 12 comparisons. In all cases studied, better statistical precision is obtained by making the two-thirds fewer simulations than under conventional multiple replications-based experimentation.
Abstract: In this paper, the mobility model of the integrated simulation tool "SIMCO2" (SImulation of Mobile COmmunication), for the performance evaluation and verification of short-range vehicle-beacon and intervehicle communication protocols and new Road Transport Telematics (RTT) application (Automatic Fee Collection, Route Guidance, Cooperative Driving, etc) is presented and validated with motorway measurements of various traffic scenarios and intensities on motorway A2 between Utrecht and Amsterdam performed by the Dutch Ministry of Transportation (Rijkswaterstaat). A comparison of the motorway measurements and the traffic scenarios simulated by SIMCO2 show a very good correspondence in important aspects like following distances between vehicles, average speed of vehicles, distribution of vehicle classes over the lanes, depending on traffic intensity. These results are discussed in this paper.
Abstract: Major weaving sections are a major source of congestion on freeways. The conflicts between vehicles making lane changes can create significant problems both at and upstream of the weaving location. This report implements a method to calculate and analyze point flows at various locations within a weaving section. The method also estimates the amount of lane changing as a measure of the likelihood of successfully negotiating a weaving section. Included in the report is information on the operation of the program and the data embedded in the program. The data are based on empirical data from nine California freeway sites with origin-destination flows and ramp spacings extrapolates using the microscopic freeway simulation model, INTRAS.
Abstract: Many recurring operational problems on freeways occur because of the turbulence caused by traffic entering the freeway and the presegregation of exiting traffic in the vicinity of ramps. This report proposes a method to estimate point flows at various locations immediately downstream of on-ramps and upstream of off-ramps in the right-most two freeway lanes. The point flows are computed from multi-variate linear regression equations using each origin-destination flow and the length between ramp pairs as independent variables. Models have been developed for isolated on-ramps, isolated off-ramps, on-off ramps, on-on ramps, and an on-off-off ramp multiple weave site all with four freeway lanes and an on-off ramp combination with three freeway lanes. All ramps are single lane with no auxiliary lanes. The equations are based on empirical data from seven California freeway sites with origin-destination flows and ramp spacings extrapolated using the microscopic freeway simulation model, INTRAS.
Abstract: This paper documents a case study on freeway improvements evaluation through the use of FREFLO, a macroscopic simulation model, and FRESIM, a microscopic simulation model. Both models are part of the Federal Highway Administration's TRAF simulation system. The case study involved a section of the I-70 freeway in Columbus, Ohio. Geometric and access control improvements were proposed to alleviate congestion on the freeway. The existing conditions, and the proposed improvements were modelled under current, as well as future, traffic scenarios. Results from the two models are compared in terms of validity (against field observations), completeness, and the suitability and usefulness of output measures for specific and/or day-to-day traffic operations analysis. The paper concludes with a list of recommendations for future freeway modelling efforts.
Abstract: This report is one of eight documents prepared under a U.S. Department of Transportation, Federal Highway Administration (FHWA) contract entitled "Analysis of Complex Congested Corridor Locations". The purpose of the project was to evaluate the effectiveness and ease of use of three freeway simulation models and the freeway analysis procedures described in the 1985 Highway Capacity Manual (HCM) by applying them to a series of real world situations. The three simulation models tested were: FREFLO - a macroscopic model developed and supported by the FHWA as part of the TRAF modeling system; FREQ - a second macroscopic model developed and supported by the Institute of Transportation Studies, University of California at Berkeley; and FRESIM - a microscopic model developed and supported by the FHWA, representing an extension of an earlier FHWA model known as INTRAS. Each of these models, together with the computerized version of the 1985 HCM procedures, known as the Highway Capacity Software (HCS), was applied to each of five case studies. Each case study location was experiencing severe operating problems. The models were used to evaluate these problems, and to test a range of possible remedial actions. This report is an executive summary of the project. The other volumes in the series are: FHWA-RD-92-103 Volume I: Final Report; FHWA-RD-92-104 Volume II: Technical Report; FHWA-RD-92-105 Volume III: Alternative Analysis Report, I-94 Eastbound between U.S.H. 45 (Zoo Interchange) and I-43 (CBD Terminus) in Milwaukee, Wisconsin; FHWA-RD-92-106 Volume IV: Incident Management Alternative Analysis Report for the George Washington Bridge and Cross Bronx Expressway Eastbound, New York/New Jersey; FHWA-RD-92-107 Volume V: Alternative Analysis Report, I-494 Eastbound between Prairie Center Drive and T.H. 5 in Minneapolis, Minnesota; FHWA-RD-92-108 Volume VI: Alternative Analysis Report, I-70 and I-71 Northbound Immediately to the South and East of Downtown Columbus, Ohio; and FHWA-RD-92-109 Volume VII: Alternative Analysis Report, I-5 Southbound, between S.R. 900 and S.R. 18 in Seattle, Washington.
Abstract: This research report documents the development models for control of signalized diamond interchanges during oversaturated traffic conditions. Oversaturated traffic conditions occur when the average traffic demand exceeds the capacity of the signal system. The dynamic optimization model proposed is the principal product of this research. The control objective of the dynamic model is to provide maximum system productivity as well as minimum delay for a selected roadway system. A special feature predetermined upper limits. The dynamic model was developed for conventional diamond interchanges and three-level diamond interchanges. The model takes the form of mixed integer linear programming. The effectiveness of the control strategies generated by the dynamic model was compared to those derived from conventional signal timing models, using the TRAF-NETSIM microscopic simulation model. It was found that the dynamic models produced optimal signal timing plans for the oversaturated signalized interchanges. The dynamic model consistently outperformed conventional models with respect to system productivity. This conclusion was drawn from the TRAF-NETSIM simulation. The dynamic model solutions significantly reduced total system delay for most test cases, while slightly increasing the delay for a few test cases.
Abstract: Describes a theoretical model which describes the influence of information technologies on driving behavior and traffic flow. The model has been integrated into a microscopic traffic simulation system. Differences in traffic flow between not supported, partially supported, and completely supported driving are presented.
Abstract: The need to develop and implement quickly new bus priority measures is probably now more urgent than ever. Increased congestion and the effects of deregulation make a very good case for implementing bus priority measures to reduce delays, operating costs and environmental impacts. The paper considers the different types of measures including: Priority at junctions, selective vehicle detection, with and contra-flow bus lanes, priority access to streets and facilities and parking/loading restrictions and more comprehensive schemes. There is a need to identify very quickly areas and bottlenecks where these types of measures would be more advantageous. Experience in the company is used to illustrate the quick response methods developed to achieve this identification. The paper then considers the need to provide quick analysis of these sites and produce design for the priority measures; software is expected to speed up this task. The use of programs for junction design (ARCADY, OSCADY, SIGSIGN), bus lane design (BLISS) and network software (TRANSYST, CONTRAM, SATURN and TRAFFICQ) in assisting the design of these measures is then discussed and their advantages and limitations put in context. The appropriateness of each of these programs is different in different areas of application and this is highlighted in the paper. The paper then considers other areas which could provide some additional assistance. One of these is the design of median bus lanes, following the extensive experience in Brazil. Another is the physical design of bus stop areas thus reducing interference among buses and between buses and the rest of the traffic. The program IRENE and its role in assisting these tasks is also discussed. The approaches are illustrated using examples from application by the company in different locations in the UK and Overseas.
Abstract: The utility of the simulation package TRAF-NETSIM to the traffic engineer is assessed and demonstrated by means of a case study. The methodology employed in performing the analysis is presented in a way that will aid future users of TRAF-NETSIM. The advantages and disadvantages of TRAF-NETSIM are documented along with the human resource requirements for a first-time application of the program. TRAF-NETSIM permits the engineer to compare alternative control and design strategies for a traffic intersection, corridor or network and allows the user to design and test within the office environment the simulation of many traffic options. TRAF-NETSIM attempts to be as realistic as possible. Lanes can be channelized for turns only or designated for carpool or bus activity. Pedestrian activity, long- and short-term events, and bus routes can be simulated as well. Creativity permits the engineer to evaluate unusual networks when required. The output of TRAF-NETSIM provides the user with a host of measures of effectiveness to compare traffic options. Delay time/vehicle, number of phase failures, speed, vehicle miles, stops/vehicle trip are some of the measures of effectiveness that can be used to evaluate networks.
Abstract: This study addresses the problem of queueing on highway facilities, wherein a large number of computerized methods for the analysis of different queueing situations are available. A three-tier classification system of the methodologies was used with the following categories: dedicated techniques, classical queueing theory, and simulation. A knowledge base for selecting an appropriate technique for a specific facility and problem is provided. The utilization of the video camera to capture queueing data in the field is described and applied to evaluate alternative methods to analyze queueing at signalized intersections. This evaluation revealed three distinct approaches from the respective categories for the evaluation of queueing at signalized intersections: the HCM method, the vacation-server queueing model, and TRAF-NETSIM. It was found that the queueing model and simulation methods offer flexibility over the more structured, dedicated HCM method and should be considered in the analysis of other situations as well as of signalized intersections.
Abstract: We explore the prospect of using TRAF-NETSIM, a microscopic simulation model, to estimate capacity and level of service through a case study. In the case study, we collected stopped delay, saturation flow, bus dwell time, double parking duration, vehicle and pedestrian volumes, etc. These data served as the bases to run and to calibrate the model, and to check the model results. Although TRAF-NETSIM does not provide capacity and level of service directly, we showed how to make use of its detail simulation capabilities and graphics to obtain capacity and level of service. We also showed how to calibrate the model to represent local traffic conditions. The simulated capacity, stopped delay and level of service were very close to the field results. Since TRAF-NETSIM is a stochastic model, there is concern that its results may vary. We examined its variability by inputting different random number seeds with different simulation times. We find that the variation of capacity was insignificant while that of stopped delay was mixed. We also examined the required number of runs and length of simulation times to obtain 95% level of confidence. TRAF-NETSIM has many advantages. Its animated and static graphics can show what is going on or how the result is derived. Its numerous calibrating parameters enable it to be applicable to many traffic conditions. It produces many statistics which are useful for other analyses. It considers individual factors as well as the interaction of different factors which may affect capacity and level of service. One can analyze the impacts of these factors on one intersection or on the network as a whole. The prospect of using a simulation model such as TRAF-NETSIM for capacity and level of service appears to be promising.
Abstract: No abstract provided.
Abstract: The papers in this session include: Technology Evaluation of Variable Message Signs: A Comparative Analysis of LED versus Fiberoptic Technology; State-of-the-Art of AVI Technology: Potential Applications to Florida's Turnpike; Emerging Technological Tools for Detailed Freeway Design and Operational Analysis: TRAF-FRESIM, FSMTUTOR, and FEDIT; A New Technique for Incident Detection in Light Traffic.
Abstract: In the report traffic operational aspects of road work on freeways have been investigated. The results show that road work strongly effects highway capacity. For instance, the closure of one lane on a two-lane highway due to maintenance (so called 2,1 work zone) causes the capacity to drop by more than 50%. The major issues in the report are determining the capacity and explaining the influence factors at (2,1) and (4,0) work zones. In a (4,0) work zone the two lanes of both directions are led on one carriageway while the other is available for road work. The number of lanes is not reduced, but the lanes are narrower than in the undisturbed situation. The research is carried out by means of an extensive literature search and by using a microsimulation model.
Abstract: Model verification and validation are two important tasks in developing a traffic simulation model. Traffic simulation models have unique characteristics because of the interaction among the drivers, vehicles, and roadway. The effects of the interaction on traffic flow should be considered in verification and validation of the models. If these two tasks are not properly performed, a traffic simulation model may not provide accurate results. A procedure for verification and validation of microscopic traffic simulation models is developed, and its application to a car-following simulation model, CARSIM, is demonstrated. The validation part of the procedure is emphasized. The validation efforts are performed at the microscopic and macroscopic levels. For validation at the microscopic level, the speed change patterns and trajectory plots obtained from simulation models are compared with those from field data. For validation at the macroscopic level, the average speed, density, and volume for simulated platoons are compared with those of field data. Also, variation of these parameters when the platoons go through a disturbance and interrelationships between these variables computed from the simulation models and the field data are examined. Regression analysis and analysis of variance of the simulation results versus the field data are discussed. The procedure may be considered as a step toward development of a comprehensive systematic approach for verification and validation of traffic simulation models.
Abstract: This article describes an enhancement to the TRAF-NETSIM simulation model, a Fortran-based program modelling the operational performance of vehicles, travelling in a network of surface streets. In this model, each vehicle is categorised and its performance determined uniquely every second. The enhancement provides an ability to model, with traffic streams exhibiting identical routeing patterns, driver-vehicle and other characteristics through a series of runs. The user can thus make a series of simulation runs by retaining the traffic stream of an initial run and varying traffic control, volume or other operational conditions during subsequent runs. The generation of stochastic behaviour and the need to generate identical traffic streams is discussed. Two approaches are considered and program modifications are described. The selected approach is based on using the vehicle-specific random number seeds to generate identical traffic streams from one run to the next. These number seeds are generated from a 'base' random number seed. Some representative simulation results are presented and the applications and limitations of the identical traffic streams feature are examined.
Abstract: Statistical issues involved in the use of stochastic traffic simulation are briefly discussed, and simulation experiments using TRAF-NETSIM are described, with emphasis on effective procedures for output analyses. The exploratory results obtained from simulation experiments are further investigated, along with the development of guidelines for output variability assessment. It is noted that the variability of variables generated by a traffic simulation model can be assessed by variables' confidence intervals computed via the two most convenient approaches, the batch means, and replication methods.
Abstract: Papers in this session include: The 1985 Highway Capacity Manual Delay Equation, CJ Messer; SIDRA for the Highway Capacity Manual, R Akcelik; Interpretation of Traffic Density as Measure of Level of Service, HS Joubert; Comparing Capacities and Delays Estimated by Highway Capacity Software and TRAF-NETSIM to Field Results, SY Wong.
Abstract: Chapter 4 of the 1985 Highway Capacity Manual uses weaving and nonweaving speeds as measures of effectiveness (MOEs) to evaluate the quality of service in freeway weaving sections. However, recent research suggests that speed may not be a reliable indicator of traffic performance. Speed and conflict rates (in particular, lane change (LC) and rear-end (RE) conflicts) are tested in terms of their sensitivity to geometric and flow variables. The testing environment is a microscopic simulation model developed for FHWA named Integrated Transportation Simulation (INTRAS), which has been extensively validated on freeway segments throughout the country. For simple one-sided freeway weaving sections, proposed conflict rates were found to be potentially more effective than speeds as an MOE. This finding is demonstrated by a higher sensitivity of the conflict rates MOE compared with the speed MOE to several geometric and flow variables at the weaving section. LC and RE conflict rates were sensitive to changes in the volume-to-capacity ratio (VC), reaching their maximum level for VC in the range 0.9 to 1.0. LC and RE conflict rates were also sensitive to changes in the volume ratio (VR), reaching their maximum level for VR in the range 0.3 to 0.5.
Abstract: The results are given of recent computational experience with the NETSIM simulation model on a CRAY X-MP/24 supercomputer. The application of NETSIM to a large urban network is demonstrated, its computational performance on the supercomputer relative to conventional mainframes is compared, and the types of modifications to the NETSIM program that would be necessary to take better advantage of the parallelism present in the supercomputer architecture are identified. The paper reviews relevant background on supercomputing and CRAY, and the simulation experiments are described. Future modifications to NETSIM to enhance its performance in vector-processing environments are discussed, and comments are made on substantive questions in traffic theory and practical traffic problems that can benefit from enhanced computational capabilities.
Abstract: This paper reports a series of empirical comparisons of several well-known incident-detection algorithms that have been performed by the authors during recent years. All these algorithms are designed to use traffic data, measured on-line at the site, and detect incidents by recognising unexpected changes in the traffic data. The suitability of detection algorithms depends on the given standard of the data collection facilities and on the traffic volumes, which both greatly influence the detection rate. The traffic data used originate from incidents observed on various motorways. In addition, microsimulation data provide a basis for comparing the algorithms in ideal situations. The following guidelines are given for using the examined detection algorithms: (1) for closely spaced adjacent sites, use the correlation algorithm for low traffic volumes, otherwise use exponential smoothing or short-period forecasting; (2) for widely spaced adjacent sites, use a filtering technique. As no existing method is superior in all traffic situations, the authors propose combining the most adaptable algorithm, the Kalman filter, with an on-line parameter optimisation procedure, the correlation approach for low traffic flows, and a robust local forecast algorithm that detects significant shock waves very clearly.
Abstract: This paper describes the results of field observations designed to measure the interactions of entryflow with the circulating flow which enters the roundabout at the previous entry, and that which leaves the roundabout at the previous exit. Data collected during peak periods at normal roundabouts and at a mini roundabout have been used to demonstrate the effect of central island size on the entry mechanism. It was found that at mini roundabouts the traffic which entered at the previous entry reduced any entry opportunity while at large roundabouts vehicles were able to enter by accepting a relatively small gap in this stream. The difference in entry mechanism and behaviour of traffic as a result of choice of entry lane is also considered. The effect of introducing the interactions described above on the entry performance as measured by average delay has been evaluated using a simulation model. Results obtained from the simulation model have been compared with those of the models ARCADY2, and TRAFFICQ runs as a simulation model of traffic movement at roundabout based on gap-acceptance criteria. It is shown that by modelling the entry interactions greater accuracy is obtained under congested conditions.
Abstract: This paper describes an enhancement to the TRAF-NETSIM simulation model. This enhancement provides users the ability to simulate with identical traffic streams through a series of runs. Thus, the user can make a series of simulation runs by retaining the traffic stream of an initial run while simulating under different traffic control or operational conditions in subsequent runs. The representation of stochastic behavior in the TRAF-NETSIM simulation model is described and the need for the generation of identical traffic streams is discussed. The two approaches considered in implementing this feature and the program modifications necessary to implement the selected approach are described. Some representative simulation results are presented and the applications and limitations of the identical traffic streams feature are discussed.
Abstract: This study relates to the evaluation of bus lanes. It has concentrated on deriving methods for predicting the principal effect of bus lanes, which is on the journey times of priority and non-priority vehicles, although all impacts are considered. This was undertaken by initially collecting data at 22 with-flow and 3 contra-flow bus lanes throughout the UK, covering parameters such as traffic flows and journey times for priority and non-priority vehicles. Such measurements allowed estimates to be made of the maximum journey time (and associated cost) savings for priority vehicles due to the bus lane. However, the prediction of the effects of the bus lanes on non-priority traffic required an assessment to be made of the suitability of three computer-based traffic models, TRAFFICQ, CONTRAM and BLAMP, and their application to all or a sample of the study sites. This application involved initial 'calibration' of the models to reflect existing traffic conditions at each site followed by further modelling with the bus lane(s) removed to obtain predictions of 'without' bus lane journey times. These results have enabled predicted economic benefits/disbenefits of each bus lane to be determined. A step-by-step procedure for evaluating bus lanes, using either modelling or non-modelling techniques as appropriate, has also been developed and is described in the report. This procedure is illustrated using data from two bus lanes recently introduced in London, where 'before-and-after' surveys were carried out which have also allowed elements of the evaluation procedure to be validated. For other studies undertaken as part of the review of bus priority measures see Contractor Reports nos 88 and 89.
Abstract: This article describes a new computer model, BLISS, (bus lane interactive simulation system), used to design with-flow bus lanes, and compares it with other such models (CONTRAM, TRAFFICQ, SATURN and BLAMP). Brief descriptions of these other programs which list their advantages and limitations are also provided, as is a section on the design of an ideal with-flow bus lane and the general methodology used in design production and evaluation. The BLISS program is written in Turbo Pascal with extended graphics facilities. Details of its methods of queue handling, platoon dispersion, delay calculation, lane choice, user interface, and data entry, and its simulation program are provided. The advantages of BLISS over BLAMP (Bus Lane Algorithm Modelling Program) are listed, as are its limitations. Because BLISS is a link-based model no reassignment is considered. Traffic diversion is not currently modelled, and neither is turning traffic. However since the model uses saturation flow as an element of discharge per second, satisfactory representation of turning traffic can be obtained by careful treatment of input levels. BLISS can be used on IBM PC microcomputers with 512 k bytes RAM and even without a floating point co-processor will run much faster than real time.
Abstract: Delay is an important measure of effectiveness in traffic studies; it represents the direct cost of fuel consumption and indirect cost of time loss to motorists. The 1985 Highway Capacity Manual (HCM) used delay as the principal measure of level-of-service at signalized intersections. The manual introduced a delay model that is based partially on the traditional Webster model. The goals of this study were: (1) to review the delay models currently being used in the SOAP84 computer model for an isolated intersection and by the TRANSYT-7F computer model for an arterial; (2) compare them with the original model developed by Webster and the new model proposed by the HCM; and (3) assess the HCM model in both isolated intersection and urban arterial environments. The results show that the new HCM delay model is sensitive to arrival type; thus it is important to know which arrival type to use. The delay values calculated by the new HCM model, SOAP84, and TRANSYT are close to the values resulting from NETSIM.
Abstract: This is a summary of a study to determine if computer simulation programs could be used to evaluate the operation of new signal systems. The study showed that both NETSIM and TRANSYT 7F models accurately estimate travel time, delay and fuel consumption, and can be used in establishing priorities among candidate signal systems. A more complete description of the project is contained in the Final Report.
Abstract: A study of three recently installed signal systems in Maryland included travel time, fuel consumption and traffic volumes for all movements at all signalized intersections. A field after study, a field before study (system modified to represent before - uncoordinated - conditions), and computer simulation runs were compared, using travel time, delay and fuel consumption MOE's. Both NETSIM and TRANSYT 7F models were used. It was shown that computer simulation models can be used to predict changes in these parameters which can then be used to make decisions on which of several candidate new signal systems will likely be the most cost effective to meet budget constraints.