Traffic breakdowns are frequently observed phenomena on roads in larger cities, especially during peak hours. Locations along a road stretch with frequently observed breakdowns are known as recurrent bottlenecks. Knowledge about bottleneck locations are important for improvement of traffic conditions at these locations. Bottleneck locations can be identified through manual inspection of data. However, due to the comprehensive amount of data that are available today, it becomes impractical to manually identify breakdowns and instead, an automated process can be used. We propose such an automated method. The proposed method is applied to a use case south of Stockholm in Sweden. One month of data collected at densely spaced detectors is used to investigate the sensitivity of the parameter settings. After calibration of the threshold values, 100% of the larger breakdowns and 40% of the medium size breakdowns are identified. Smaller breakdowns, not giving significant impact on the traffic conditions, are only detected in 10%-20% of the cases. Thereafter, the method is applied to 1 year of data to illustrate the applicability of the method on a larger data set. The results show that the method is promising to use for identification of recurrent bottleneck locations.

The gradual deployment of Connected and Automated Vehicles (CAV) in traffic will result in a transition period in which vehicles with various levels of automation and connectivity will have to co-exist with non-connected and non-automated road users for quite some time. Consequently, new types of interactions will emerge (and old types of interactions are likely to become more complicated) between vehicles at different levels of automation and other road users which could have significant implications on traffic safety and efficiency. Understanding the nature of these interactions, how humans might adapt their behavior, how connectivity can be utilized to proactively enhance drivers char63 driving performance, and how automated vehicles can be programmed to behave in different driving situations to guarantee safety and efficiency remain among the key knowledge gaps that require scientific research. This knowledge is crucial for the development of adequate integration policies of connected and automated vehicles in mixed traffic environment, for updating and improving automated vehicles char63 algorithms and software, for designing the physical and digital road infrastructure, and for operating and managing traffic on the road network.

The gradual deployment of automated vehicles on the existing road network will lead to a long transition period in which vehicles at different driving automation levels and capabilities will share the road with human driven vehicles, resulting into what is known as mixed traffic. Whether our road infrastructure is ready to safely and efficiently accommodate this mixed traffic remains a knowledge gap. Microscopic traffic simulation provides a proactive approach for assessing these implications. However, differences in assumptions regarding modeling automated driving in current simulation studies, and the use of different terminology make it difficult to compare the results of these studies. Therefore, the aim of this study is to specify the aspects to consider for modeling automated driving in microscopic traffic simulations using harmonized concepts, to investigate how both empirical studies and microscopic traffic simulation studies on automated driving have considered the proposed aspects, and to identify the state of the practice and the research needs to further improve the modeling of automated driving. Six important aspects were identified: the role of authorities, the role of users, the vehicle system, the perception of surroundings based on the vehicle’s sensors, the vehicle connectivity features, and the role of the infrastructure both physical and digital. The research gaps and research directions in relation to these aspects are identified and proposed, these might bring great benefits for the development of more accurate and realistic modeling of automated driving in microscopic traffic simulations.

Microscopic emission models estimate second-by-second emissions and fuel consumption for individual vehicles based on vehicle trajectories. A vehicle trajectory describes how the position, speed and acceleration of a vehicle evolves over time. In practice, collecting a complete trajectory data set on a road stretch is not always feasible due to economic and privacy constraints. Therefore, several researchers suggest approaches for generating Virtual Vehicle Trajectories (VVT) given some partially observed traffic data. However, the traditional VVT generation approaches, being originally developed for travel time estimations, usually consider a simplified description of vehicle kinematics, hindering their applicability in emission modelling. In this paper, we suggest a novel approach for generating VVT, which facilitates their use in emission modelling. We empirically evaluate our method by comparing it to the traditional approaches. The results are promising, showing that, under certain experimental settings, our method can enhance the accuracy of emission estimations.

Microscopic traffic simulation is a useful tool for the planning of motorized traffic, yet bicycle traffic still lacks this type of modeling support. Nonetheless, certain microscopic traffic simulators, such as Vissim, model bicycle traffic by applying models originally designed for car traffic. The gradient of a bicycle path has a significant impact on the speed of cyclists; therefore, this impact should be captured in microscopic traffic simulation. We investigate two calibration approaches to reproduce the effect of gradient on the speed of cyclists using the default driver behavioral model in Vissim. The first approach is to modify the simulated gradient to represent different values of the gradient-acceleration parameter: a fixed value that represents a decrease in the maximum acceleration that cyclists can apply on an uphill. The second approach is to adjust the maximum-acceleration function. We evaluate both approaches by applying a Vissim model of a bidirectional bicycle path with a 3% gradient in Stockholm. The results show that the current default implementation in the Vissim model underestimates the effect of gradient on speed. Moreover, the gradient-acceleration parameter does not directly reduce the maximum acceleration of all cyclists, but only of those cyclists riding above a certain speed. We conclude that by using a higher gradient-acceleration value than the default, we accurately estimate the observed mean speed on the uphill. However, neither of the investigated calibration approaches provides accurate estimates of the speed distributions. We emphasize the need for developing more accurate behavioral models designed for cyclists.

Dynamic Network Loading (DNL) models are typically formulated as a system of differential equations where travel times, densities or any other variable that indicates congestion is endogenous. However, such endogeneities increase the complexity of the Dynamic Traffic Assignment (DTA) problem due to the interdependence of DNL, route choice and demand. In this paper, attempting to exploit the growing accessibility of traffic-related data, we suggest that congestion can be instead captured by exogenous variables, such as travel time observations. We propagate the traffic flow based on an exogenous travel time function, which has a piece-wise linear form. Given piece-wise stationary route flows, the piece-wise linear form of the travel time function allows us to use an efficient event-based modelling structure. Our Data-Driven Network Loading (DDNL) approach is developed in accordance with the theoretical DNL framework ensuring vehicle conservation and FIFO. The first simulation experiment-based results are encouraging, indicating that the DDNL can contribute to improving the efficiency of applications where the monitoring of historical network-wide flows is required. Abbreviations: DDNL - Data Driven Network Loading; DNL - Dynamic Network Loading; DTA - Dynamic Traffic Assignment; ITS - Intelligent Transportation Systems; OD - Origin Destination; TTF - Travel Time Function; LTT - Linear Travel Time; DL - Demand level

The introduction of automated vehicles (AVs) is commonly expected to improve different aspects of transportation. A long transition period is expected until AVs become prevalent on roads. During this period, different types of AVs with different driving logics will coexist along human driven vehicles. Using microscopic traffic simulation, this study investigates the range of potential impacts on traffic performance in terms of throughput and travel delays for different types of AVs and human driven vehicles on motorways. The simulation experiment includes scenarios with combinations of three different driving logics for AVs together with human driven vehicles at increasing penetration rates. The utilized AV driving logics represent the evolution of AVs, they were defined in the microscopic simulation tool Vissim and were created by modifying and extending the human driver behaviour models. The results of the simulation experiment show a decrease in vehicle throughput and significant effects on delay times when AVs with a more cautious driving logic are predominant. Overall, results show higher vehicle throughput and lower travel delays as AVs evolve to more advanced driving logics.

Dynamiska trafikmodeller förväntas i större utsträckning komplettera och på sikt helt eller delvis ersätta statiska trafikmodeller. Samtidigt blir trafikmätningar mer och mer tillgängliga, och utifrån dessa kan tidsvarierade estimeringar av flöden och hastighet över större områden tas fram. I Sverige används idag HBEFA för estimering av emissioner. HBEFA är en EU-gemensam modell som används i flera europeiska länder för beräkning av luftföroreningar och bränsleförbrukning. Detta projekt syftar till att ta fram en metod för tillämpning av HBEFA på trafikdata från mätningar och dynamiska trafikmodeller på ett sätt som är konsistent med den tillämpning av HBEFA som idag sker med statisk trafikdata. Projektet fokuserar främst på tillämpningar där lokala emissioner från vägtrafik är av intresse. Exempel på sådana tillämpningar är uppföljning, estimering och prediktering av luftkvalitet. Men även analys av förändring av utsläpp från vägtrafiken över tid, samt före/efter studier vid förändringar av trafikinfrastrukturen. Detta kräver att den tillämpade beräkningsmetodiken för utsläpp anpassas för att kunna dra nytta av den rikare information som dynamisk trafikdata innehåller.I HBEFA är skattning av trafiktillstånd centralt för estimering av emissioner. I denna studie har vi studerat och jämfört tre olika sätt att definiera trafiktillståndet i HBEFAstrafiksituationer nämligen baserat på: flöde (det som används i nuläget), hastighet och densitet. Flödesbaserade definitioner har en inbyggd begränsning i att de inte entydigt kan beskriva trafiksituationen, då lågt flöde kan innebära både låg trängsel och hög trängsel, beroende på hastighet. Detta är inget problem i nuvarande tillämpning av HBEFA med statisk trafikdata, då en förenklad beskrivning av det överbelastade fallet görs, som tillåter flöden över kapaciteten, men som då representerar en efterfrågan snarare än flöde. Undersökningarna i denna studieindikerar att en hastighetsbaserad definition av trafiktillstånd inte är att föredra då hastigheten kan variera relativt mycket mellan olika fordon och över en sträcka vid liknande flödes- och densitetsförhållanden. Hastighetsbaserade gränser mellan trafiktillstånd blir främst problematiskt vid tillämpning på mätdata. Vid hastighetsbaserade gränser blir varje fordonspassage med låg hastighet automatiskt klassade i ett trafiktillstånd som innebär trängsel. Densitet är ett mer stabilt mått vid liknande trafikförhållanden och kan sägas bättre definiera graden av trängsel i trafikteoretisk mening. En densitetsbaserad definition kan entydigt bestämma samtliga trafiktillstånd, till skillnad från en flödesbaserad definition som endast kan bestämma icke överbelastade trafiktillstånd entydigt. För de icke överbelastade trafiktillstånden (dvs. 1 och 2) har vi också kunnat visa att en densitetsbaserad definition ger ungefär samma fördelning av trafikarbete mellan trafiktillstånden.För totala emissioner är skillnaden mellan skattade emissioner med nuvarande flödesbaserade definitioner av trafiktillstånd och utvärderad densitetsbaserade definitioner av trafiktillstånd oftast liten. Dock är den densitetsbaserade versionen bättre på att fånga hur emissionerna varierar i tid och rum.

The introduction of automated vehicles is expected to affect traffic performance. Microscopic traffic simulation offers good possibilities to investigate the potential effects of the introduction of automated vehicles. However, current microscopic traffic simulation models are designed for modelling human-driven vehicles. Thus, modelling the behaviour of automated vehicles requires further development. There are several possible ways to extend the models, but independent of approach a large problem is that the information available on how automated vehicles will behave is limited to todays partly automated vehicles. How future generations of automated vehicles will behave will be unknown for some time. There are also large uncertainties related to what automation functions are technically feasible, allowed, and actually activated by the users, for different road environments and at different stages of the transition from 0 to 100% of automated vehicles. This article presents an approach for handling several of these uncertainties by introducing conceptual descriptions of four different types of driving behaviour of automated vehicles (Rail-safe, Cautious, Normal, and All-knowing) and presents how these driving logics can be implemented in a commonly used traffic simulation program. The driving logics are also linked to assumptions on which logic that could operate in which environment at which part of the transition period. Simulation results for four different types of road facilities are also presented to illustrate potential effects on traffic performance of the driving logics. The simulation results show large variations in throughput, from large decreases to large increases, depending on driving logic and penetration rate.

In large urban areas, the estimation of vehicular traffic emissions is commonly based on the outputs of transport planning models, such as Static Traffic Assignment (STA) models. However, such models, being used in a strategic context, imply some important simplifications regarding the variation of traffic conditions, and their outputs are heavily aggregated in time. In addition, dynamic traffic flow phenomena, such as queue spillback, cannot be captured, leading to inaccurate modelling of congestion. As congestion is strongly correlated with increased emission rates, using STA may lead to unreliable emission estimations. The first objective of this paper is to identify the errors that STA models introduce into an emission estimation. Then, considering the type and the nature of the errors, our aim is to suggest potential solutions. According to our findings, the main errors are related to STA inability of accurately modelling the level and the location of congestion. For this reason, we suggest and evaluate the postprocessing of STA outputs through quasidynamic network loading. Then, we evaluate our suggested approach using the HBEFA emission factors and a 19 km long motorway segment in Stockholm as a case study. Although, in terms of total emissions, the differences compared to the simple static case are not so vital, the postprocessor performs better regarding the spatial distribution of emissions. Considering the location-specific effects of traffic emissions, the latter may lead to substantial improvements in applications of emission modelling such as dispersion, air quality, and exposure modelling.

Microscopic traffic simulation is an ideal tool for investigating the network level impacts of eco-driving in different networks and traffic conditions, under varying penetration rates and driver compliance rates. The reliability of the traffic simulation results however rely on the accurate representation of the simulation of the driver support system and the response of the driver to the eco-driving advice, as well as on a realistic modelling and calibration of the drivers behaviour. The state-of-the-art microscopic traffic simulation models however exclude detailed modelling of the driver response to eco-driver support systems. This paper fills in this research gap by presenting a framework for extending state-of-the-art traffic simulation models with sub models for drivers compliance to advice from an advisory eco-driving support systems. The developed simulation framework includes among others a model of drivers compliance with the advice given by the system, a gear shifting model and a simplified model for estimating vehicles maximum possible acceleration. Data from field operational tests with a full advisory eco-driving system developed within the ecoDriver project was used to calibrate the developed compliance models. A set of verification simulations used to illustrate the effect of the combination of the ecoDriver system and drivers compliance to the advices are also presented.

The mission of the H2020 CoEXist project is to enable mobility stakeholders to get “Automation-ready” – which CoEXist currently defines as conducting transport and infrastructure planning for connected and automated vehicles (CAVs) in the same comprehensive manner as for existing modes such as conventional vehicles, public transport, pedestrians, and cyclists, while ensuring continued support for existing modes on the same network. This definition will be fine-tuned through stakeholder engagement processes. The H2020 CoEXist project started in May 2017 and will run until April 2020. This paper introduces this project and covers its progress until January 2018, with a focus on the methodology of the “Automation-ready framework” that provides a planning framework for urban road authorities to prepare for the introduction of CAVs on the road network. The framework includes elements about strategic urban mobility planning for CAVs and a clear guide for urban transport planners with a list of concrete actions that cities can do now to plan for CAVs on their road network

In recent years, various field operational tests (FOTs) have been carried out in the EU to measure the real-world impacts of Intelligent Transport Systems (ITS). A challenge arising from these FOTs is to scale up from the very localised effects measured in the tests to a much wider set of socio-economic impacts, for the purposes of policy evaluation. This can involve: projecting future take-up of the systems; scaling up to a wider geographical area – in some cases the whole EU; and estimating a range of economic, social and environmental impacts into the future. This article describes the evaluation conducted in the European project ‘ecoDriver’, which developed and tested a range of driver support systems for cars and commercial vehicles. The systems aimed to reduce CO2 emissions and energy consumption by encouraging the adoption of green driving behaviour. A novel approach to evaluation was adopted, which used scenario-building and micro-simulation to help scale up the results from field tests to the EU-28 level over a 20 year period, leading to a cost-benefit analysis (CBA) from both a societal and a stakeholder perspective. This article describes the method developed and used for the evaluation, and the main results for eco-driving systems, focusing on novel aspects, lessons learned and implications for policy and research. © 2018 World Conference on Transport Research Society

The traffic performance at oncoming lane separated highways with alternating dedicated overtaking lanes (so called 2+1 roads), is dependent on the share of two lane segments (also known as the share of overtaking length). In order to maximize utilization and traffic performance, the configuration of the overtaking lanes should be designed to avoid congestion and delays. Short overtaking lanes implies limited time of queue discharge, but gives frequently recurring possibilities to overtake. Increased lengths of overtaking lanes imply the possibility to overtake several vehicles per overtaking lane, but increases the risk of catching up slower vehicles since the configuration also results in increased lengths of one lane segments.

This report presents a traffic simulation study of how different configurations affects the throughput at 2+1 roads. The results indicate that overtaking lanes between 1 050 and 1 400 meters result in shortest travel time. However, the differences are small (~0.4 seconds/km) and not statistically significant. Thus, the benefit of optimizing the configuration in terms of throughput could be questioned. Based on the results, it becomes reasonable to question the concept of designing 2+1 roads with long overtaking lengths (which corresponds to the recommendations from the Swedish Transport Administration (Trafikverket)). The major risk of catching up a slower vehicle at the one lane segments obviously affects the travel time.

The continuous traffic growth has led to highly congested cities, with negative environmental effects, both related to air quality and climate change. According to the European Environment Agency, transportation remains a significant contributor to the total emissions of the main air pollutants, (EEA, 2016). Specifically, Nitrogen Oxides (NOx), Carbon Oxide (CO) and fine particulate matter (PM2.5) make up 32%, 23% and 8% of the total emissions, respectively. This vigorous impact of vehicular emissions to the urban environmental air quality, raises concerns over the impact of traffic on human health. Therefore, the effective implementation of emission reducing policies, such as traffic control measures or congestion pricing, becomes crucial for many European cities in order to meet the air quality standards and mitigate the human exposure to pollution. To quantify the environmental effects of these measures and demonstrate their effectiveness, a reliable estimation of pollutants concentrations through emission and dispersion modelling is needed....

Speed-flow relationships are an important part of the Swedish Transport Administration (Trafikverket) model for evaluation of effects of road facilities (the EVA model). This report present suggestions for new speed-flow relationships for motorways (MV), low standard motorways (4F), oncoming lane separated highways with grade separated intersections (MML), oncoming separated highways with at grade intersections (MLV), and two-lane highways. The suggestions are based on data from measurements using the Swedish Transport Administration’s traffic measurement system TMS in combination with model calculations. The TMS data have, for each road category, been quality checked, processed and analysed. The data material is presented as speed-flow diagrams for passenger cars, buses and trucks without trailer, and trucks with trailers. A comparison of the current speed-flow relationships and the TMS-measurements was then conducted for each road category, and if needed a suggestion for a revision was presented. The most significant changes from last revision from 2013 are: average free flow speed for trucks without trailer have in general increased for all road types except two lane highways for which the speed has decreased; average free flow speed for trucks with trailers have in general decreased; and the average speed on two lane highways have in general decreased

Route choice calculations in static traffic assignment models (as Emme, Visum, TransCad) are based on travel time estimations using volume delay functions. The volume delay function (also denoted travel time functions) describe how the travel time depend on the traffic volume for different types of roads. The volume delay functions are one of the base elements in travel prognosis models as the Swedish Sampers model system. This report presents a pre-study with the aim to investigate how volume delay functions should be designed and calibrated, including which road classification to use, which type of volume delay function that should be used, how the functions should be calibrated and which data that is needed for the calibration. These questions were investigated by a literature review on state-of-practice, workshops with experienced Sampers users to collect information and experiences of the current volume delay functions in Sampers, workshops with research experts on data collection of travel times, and project internal discussions on calibration methodologies.

The literature review showed that there are few guidelines on how volume delay functions can or should be calibrated. The calibration is commonly conducted by fitting the volume delay function curve to cross-sectional measurements of flow and mean speed. There are some examples of calibration based on travel time measurements based on floating car measurements or number plate recognition. These calibration approaches focus on describing travel time for a given link based on the flow at the link. However, based on the literature review and experience from earlier research in Sweden it is concluded that volume delay functions that represent the traffic process on a road link in a good way do not necessary give a good fit of the static assignment calculated and observed link and route flows and travel times. There are several attempts described in the literature of calibration approaches that aim to minimize the difference between model calculated and observed flows and travel times using optimization techniques. The suggestion from the pre-study is that such an approach should be investigated for calibration of the Sampers volume delay functions. To avoid overfitting and unrealistic parameters values the optimization should include lower and upper limits of the parameters.

The calibration requires both link flow and travel time observations. Link counts are regularly measured for other purposes and can be collected from the Swedish Transport Administration and municipality regular traffic measurement programs. The suggestion for travel time data is to use the travel time data that currently is commissioned by the Swedish Transport Administration and Stockholm and Göteborg municipality. Our recommendation is also that the Swedish Transport Administration investigate the possibility to buy travel time data for the Swedish main road network.

Purpose Road traffic incidents cause delay, affect public safetyand the environment. The CEDR PRIMA project aims toextend practical guidance for traffic managers in pro-activeTraffic Incident Management (TIM) techniques to reduce theimpacts and associated costs of incidents.Methods The paper describes modelling methods used in theproject for assessing the effect of different management techniqueson incident duration and travel delay under various scenarios,including collision, adverseweather, heavy vehicle breakdownand other obstruction, assuming various management strategiesand generic impacts of novel technologies. Macroscopicsimulations of 178 variations of 13 basic scenarios have beenperformed using a flexible and computationally efficientmacroscopic queue model, results being verified by simulationusing a velocity-based Cell Transmission Model (CTM-v).Results The results of the two modelling methods are broadlyconsistent. While delays estimated by the two methods candiffer by up to 20%, this is small compared to the factor of30 range of modelled delays caused by incidents, dependingon their nature and circumstances, and is not sufficient toaffect general conclusions. Under the peak traffic conditionsassumed, the most important factor affecting delay is whetherrunning lanes can be kept open, but quick clearance of carriagewayis not always feasible.Conclusions Comparison of two very different modellingmethods confirms their consistency within the context of highlyscenario-dependent results, giving confidence in the results.Future research and data needs include further validation ofthe models, potential application to traffic flow and conflictprediction and incident prevention, and more complete andconsistent recording of incident timelines and impacts.

The rapid growth of traffic congestion has led to an increased level of emissions and energy consumption in urban areas. Well designed infrastructure and traffic controllers along with more efficient vehicles and policy measures are required to mitigate congestion and thus reduce transport emissions. In order to evaluate how changes in the traffic system affect energy use and emissions, traffic analysis tools are used together with emission models. In large urban areas emission models mainly rely on aggregated outputs from traffic models, such as the average link speed and flow. Static traffic models are commonly used to generate inputs for emission models, since they can efficiently be applied to larger areas with relatively low computational cost. However, in some cases their underlying assumptions can lead to inaccurate predictions of the traffic conditions and hence to unreliable emission estimates. The aim of this paper is to investigate and quantify the errors that static modeling introduces in emission estimation and subsequently considering the source of those errors, to suggest and evaluate possible solutions. The long analysis periods that are commonly used in static models, as well as the static models’ inability to describe dynamic traffic flow phenomena can lead up to 40 % underestimation of the estimated emissions. In order to better estimate the total emissions, we propose the development of a post processing technique based on a quasi-dynamic approach, attempting to capture more of the excess emissions created by the temporal and spatial variations of traffic conditions.

The first Swedish 2+1 median barrier road was opened in 1998. The concept was to retrofit the standard existing two-lane 13 m paved width cross-section at 90 and 110 kph posted speed limit without widening. This design has one continuous lane in each direction, a middle lane changing direction every one to three kilometres with a median barrier separating the two traffic directions. Today over 2 700 km 2+1 median barrier roads are opened for traffic. AADTs vary from some 3 000 to 20 000 with an average just below 10 000 nowadays normally with 100 kph. The concept has lately been enhanced also to cover the existing 9 m paved width cross-section. The design concept is the same from a drivers viewpoint, one continuous lane in each direction with a middle lane changing direction and a separating median barrier. This is created by introducing a continuous median barrier and adding overtaking lanes within an overtaking strategy. The differences are the existence of 1+1-sections, less overtaking opportunities and a slightly more narrow cross-section. Some 15 projects are opened. The purpose of this paper is to summarize present knowledge on level-of-service issues as they are presented in Swedish design and assessment guidelines and to give an overview of field measurements and theoretical analytical and simulation studies supporting the recommendations.

This paper presents an investigation of capacity reduction in connection with roadwork zones. The paper presents a state-of-the-art description on roadwork effects on capacity. Based on the literature on this topic the most important parameters that should be incorporated in a Swedish capacity manual for the operation and maintenance roadwork are: the proportion of heavy traffic; lane width; number of closed lanes; closed road shoulder; proportion of commuter traffic; and length of roadwork zone. The paper presents a comparison of a composite model of correction factors from Germany, USA and Denmark and the Dutch model for computation of capacity reduction. The comparison show that the two models essentially gives the same results. Based on these results a model was developed. The model developed was validated using empirical data from a full scale test at the motorway network in Gothenburg. The throughput was measured in two cases during the morning and afternoon peak hour. The capacity for the normal site conditions was estimated based on traffic flow and speed data from the same site. The result shows that the empirically estimated capacity reduction is consistent with the reduction calculated with the new model for the different road work designs evaluated. The conclusion is that the model developed seems to be valid for capacity reduction estimations of roadworks on Swedish motorways but that more empirics are needed to ensure general validity.

Vägtrafiken är en stor källa till utsläpp av både lokalt, regionalt och globalt förorenande ämnen. Lokalt och regionalt förorenande ämnen så som partiklar, kolmonoxid, kolväten och kväveoxider har påverkan på både natur och människa, och hälsoskadliga effekter i stadsmiljö är välkända. Globalt förorenande ämnen från biltrafiken, främst koldioxid, är ett stort problem ur ett klimatperspektiv. Hälsovådliga utsläpp är främst intressanta ur ett luftkvalitetsperspektiv, det är då inte de faktiska utsläppen från fordon (emissioner) som är mest intressanta utan halten förorenande ämnen i luften. Halten av förorenande ämnen beror naturligtvis på emissioner, men också på utsläpp från andra källor och väder. Genom en förenklad värdering av denna typ av utsläpp, där man tar hänsyn till lokala variationer i spridning av emissioner i luften, värderas luftkvalitet i samhällsekonomiska kalkyler i Sverige direkt baserat på fordonsemissioner. Mer detaljerat beräknas spridning av emissioner när kommuner gör luftkvalitetsberäkningar i verktyget SIMAIR. Koldioxid värderas direkt utifrån sina emissioner i samhällsekonomiska kalkyler.

Oavsett om syftet är att göra samhällsekonomiska kalkyler eller mer detaljerade luftkvalitetsberäkningar i SIMAIR används i Sverige vanligen utdata från statiska trafikmodeller, baserade på användarjämvikt, som indata till emissionsberäkningarna. En statisk nätutläggning tar inte hänsyn till variationer över tid, eller utbredning av köer i trafiknätverket. Även om mer detaljerade trafikanalyser ofta görs med dynamiska trafikmodeller i senare planeringsskeenden, är inte resultatet från dessa analyser direkt tillämpbara i de emissionsberäkningar som görs i Samkalk (det verktyg som vanligen används för samhällsekonomiska kalkyler i Sverige) eller i SIMAIR.

I Sverige används främst emissionsmodellen HBEFA som är en databas med emissionsfaktorer. Emissionsfaktorerna beskriver utsläpp i gram per fordonskilometer för en given fordonstyp, vägklass och trafikförhållande, som antingen definieras av flödesgränser eller hastighetsgränser. Emissionsfaktorer för fyra olika trafiksituationer finns definierade; ”free flow”, ”heavy”, ”saturated” och ”stop-and-go”. För att erhålla emissioner per timme på ett vägavsnitt multipliceras emissionsfaktorn med trafikflöde (fordon per timme) och vägavsnittets längd.

När utdata från statiska trafikmodeller används som indata till emissionsberäkningar uppstår främst två problem. Statiska trafikmodeller medelvärdesbildar efterfrågan, och därmed trafikförhållanden, över en längre tid. Dessutom modellerar statiska modeller fördröjning till följd av köer på ett enskilt vägavsnitt, och köer kan inte propagera i ett vägnät. Potentiellt kan dessa två faktorer leda till att emissioner under- eller överskattas.

För att undersöka omfattning och karaktären av dessa fel har trafikmätningar från delar av E4 genom Stockholm använts. Flödes- och hastighetsmätningar har aggregerats till 15 minutersperioder och använts som indata till emissionsberäkningar av CO₂ (och därmed också energianvändning), CO, HC och NO_x.

Statiska trafikmodeller använder reseefterfrågan över en längre period (exempelvis en maxtimma) som indata. I statiska modeller förväntas lika många resor som påbörjas också avslutas, vilket kräver att tidsperioden är tillräckligt lång (i vårt fall med morgontrafiken i Stockholm innebär det aggregering över en tretimmarsperiod). Detta i sig gör att flöden i trafiknätverket är medelvärden för den studerade perioden, och topparna smetas därför ut. Det är dock flödestopparna som i verkligheten kraftigast bidrar till utsläppen. Jämförelsen mellan emissioner från aggregerade flöden och 15 minuters flöden ger en inblick i hur stort fel medelvärdesbildningen medför. Underskattningen av emissioner för samtliga utsläpp, utom HC, ligger mellan 10% och 15%. Utsläppen av HC underskattas med ca 20%.

Vidare så analyserades effekten av att statiska modeller inte kan modellera uppbyggnad och avveckling av köer. HBEFA tillämpades få enligt den specifikation som Trafikverket tagit fram till Samkalk och SIMAIR. Sedan studerades, länk för länk, hur emissionsberäkningarna avviker mot när emissionsfaktorerna appliceras direkt på trafikmätningarna. På enskilda vägavsnitt kan underskattningen av utsläpp av HC och CO nå 40%, och totalt över hela den studerade vägsträckan underskattas utsläppen med 15% till 25% beroende på vilket ämne som avses.

Slutligen har en metod för efterbearbetning av trafikdata tagits fram, baserad på den quasi-dynamiska nätutläggningsmetoden STAQ. Denna metod kan ses som en diskritiserad version av länktransmissionsmodellen, där förändringar i trafikförhållanden beskrivs som diskreta händelser i en simulering. STAQ använder fundamentaldiagram för att beskriva relationen mellan densitet, flöde och hastighet på varje vägsträcka, och kan delvis beskriva köers uppbyggnad och avveckling. En händelse i simuleringen motsvaras av att densiteten förändras på vägavsnittet, och därmed hastigheten. Däremot hanteras, genom de diskreta händelserna, tid på ett förenklat sätt, där trafikförhållanden på vägavsnitt endast registreras i samband med de diskreta händelserna. STAQ är framför allt utvecklad för att bättre bestämma fördröjning till följd av köer. I detta arbete har vi istället utvecklat ett angreppssätt som utnyttjar information som finns tillgänglig vid de diskreta händelserna för att förbättra beräkningen av emissioner. Med denna efterbearbetningsmetod underskattas emissionerna istället med cirka 10%, förutom för HC som underskattas med 20%.

Vi har identifierat brister som uppstår när utdata från statiska trafikmodeller används som indata till emissionsberäkningar, och föreslagit en efterbearbetningsmetod för att försöka minska de fel som uppstår. Fortsatt arbete behöver fokusera på hur denna metodik fungerar i större nätverk och hur man bättre kan ta hänsyn till hur efterfrågan varierar inom en maxperiod.

This report presents method and result for the development of new travel time functions for the Swedish national transport planning modelling system Sampers. Travel time functions include one part that describes the travel time delay on road links and one part that describes the delay at intersections. It is difficult and expensive to conduct synchronized measurements of traffic flow and travel times. An alternative approach has therefore been applied in which the travel time functions were calibrated based on calculations of intersection delay for different intersection designs using the intersection delay and capacity model Capcal. The travel time functions developed were tested and validated are now implemented in the Sampers system.

This paper presents a micro-simulation model that takes flared design of rural intersections into consideration. The intersection model is designed with input parameters that describe the geometric conditions of the flare. The behavior model includes both a traditional gap-acceptance sub-model and a passage model for modelling of vehicles’ possibility to pass other vehicles using the flare. The intersection model developed has been implemented in the traffic micro simulation model RuTSim. The gap-acceptance part of the model has been calibrated using data for stop and yield 3-way intersections. The validation was performed by using video recordings to calculate delay for the yield regulated intersection and time in queue and service time for the stop regulated intersection. The results from the validation simulations correspond well with the empirical validation data. The effect of the flare on delay has been studied by using 3 different intersection lay-outs and different levels of minor and major flow. The result shows that the delay is decreasing with increasing intersection radius.

Dedicated bus lanes and bus streets have, in recent years, become common measures for prioritisation of public transport. By ensuring free path along routes, they increase average speed and travel time reliability of buses. However, a major drawback is that the total traffic capacities of the roads decrease. Hence, these measures are only suitable when the total traffic flow is low enough to allow for a reduction of lanes; if it is possible to reroute adjacent traffic; or if it is possible to extend the road with additional lanes. A supplementary priority measure could be to utilize dynamic bus lanes (also called intermittent bus lanes and bus lanes with intermittent priority). Dynamic bus lanes are only dedicated for buses when and where the buses need them, and otherwise open for all vehicles to use. At any given point, adjacent traffic is only permitted from using the dynamic bus lanes at the stretches where buses are in the vicinity. This report presents the results from a pre-study, investigating the potential that dynamic bus lanes could have as a priority measure for public transport in a Swedish context.

Knowledge of situations in which dynamic bus lanes have the highest potential, and their implementation requirements is scarce. It is moreover uncertain how they would affect traffic safety, level of service and user experience. Two real world field tests have been conducted; one in Lisbon and one in Melbourne. The installation in Melbourne is now permanently applied for trams on one street. The field test in Lisbon was on the contrary not made permanent, although the results showed large benefits for buses and limited adverse effects on other vehicles. Dynamic bus lanes have also been investigated by means of traffic analysis and traffic simulation experiments. In general, these studies show that the effects on travel time for buses are in general positive and delays for other vehicles are limited. Results from example calculations in this pre-study show that this also could be true for a Swedish context. It has also been identified that the effects on travel times are highly dependent on factors such as: the total traffic flow; the bus flow, the capacity of roads and junctions; the distance between junctions and bus stops; the type of bus stops and the yielding rules at bus stops. The effects on travel time variations are unclear and need to be further investigated.

Few rigorous research studies have in general been undertaken to measure the user experiences or road safety implications of bus priority schemes, and evidence from those that do exist are mixed. Anyhow, the experiences from Lisbon and Melbourne suggest that drivers in adjacent lanes in general understand and accept that they are deprived of the right to use the lane when the buses need it, and that they will behave appropriately. Neither of the field tests has observed any negative impact on road safety. A workshop was conducted within this pre-study in order to further investigate plausible user experiences. The results indicate that bus drivers’ stress levels could be reduced; the relative attractiveness of travelling by bus might rise; and that motorists probably would experience the introduction of dynamic bus lanes as neither good nor bad, as long as the system is fairly intuitive.

Technical solutions for implementing dynamic bus lanes exist. A dynamic bus lane system would require development of a system control unit and integration with bus sensors, sensors for traffic flow measurement, variable message signs (to inform road users of the current status of the dynamic bus lane) and traffic signals. It is moreover, in Sweden, possible to develop a local traffic rule that regulates dynamic bus lanes. However, the rule needs to be properly specified, designed, communicated, signed and marked on the road.

The overall conclusion form the pre-study is that dynamic bus lanes could be a useful complementary priority measure for public transport vehicles in Sweden, especially when dedicated bus lanes are not feasible or desirable. However, a real world installation in Sweden, including pre implementation traffic analysis, is needed, in order to further investigate the potential and consequences. Thus, the next step is to plan for an implementation on a specific road stretch. That would include both estimation of costs, and generate input to further studies of effect on level of service and user experience. Driving simulators and traffic simulation experiments are applicable methods for investigating these issues.

To improve road safety on parts of the road network carrying low traffic volumes, road designs are proposed including single-lane road segments and periodic overtaking lanes. These roads have been proven to contribute to substantial benefits in terms of road safety. However, overtaking of slower vehicles is only possible on segments including an overtaking lane and not on the single-lane road segments. Driver and vehicle heterogeneity resulting in differences in desired speeds are consequently decisive for the traffic performance. Sufficient quality of service is relying on an appropriate design and distribution of single-lane segments and overtaking lanes. In this paper, we study the effect of the desired speed distribution on traffic performance on single-lane road segments. Expressions are derived for the travel time, delay and percent time spent following. The derived expressions link the desired speed distribution, the single-lane segment length and the traffic flow to the resulting traffic performance. The results are verified through comparison with measures based on microscopic traffic simulation. The conclusion is that there is a good agreement between derived measures and simulation results. The derived measures should therefore not only be of theoretical interest, but also of practical use to estimate traffic performance on single-lane road segments.

In driving simulation, a scenario includes definitions of the road environment, the traffic situation, simulated vehicles’ interactions with the participant’s vehicle and measurements that need to be collected. The scenarios need to be designed in such a way that the research questions to be studied can be answered, which commonly imply exposing the participant for a couple of predefined specific situations that has to be both realistic and repeatable. This article presents an integrated algorithm based on Dynamic Actor Preparation and Automated Action Planning to control autonomous simulated vehicles in the simulation in order to generate predefined situations. This algorithm is thus able to plan driving actions for autonomous vehicles based on specific tasks with relevant contextual information as well as handling longitudinal transportation of simulated vehicles based on the contextual information in an automated manner. The conducted experiment shows that the algorithm is able to guarantee repeatability under autonomous traffic flow. The presented algorithm can benefit not only the driving simulation community, but also relevant areas, such as autonomous vehicle and in-vehicle device design by providing them with an algorithm for target pursue and driving task accomplishment, which can be used to design a human-vehicle cooperation system in the coming era of autonomous driving.

Detta kapitel behandlar beräkning av kapacitet, fördröjning, andel stopp och kölängd för:

• Cirkulationsplatser i 3- och 4-vägs korsningar med ett eller två cirkulerande körfält. Varierande antal cirkulerande körfält behandlasinte.
• Metoderna kan dock relativt enkelt utökas för att behandla cirkulationsplatser med fler än fyra ben.

Metoden behandlar också överbelastning enligt metodik i Trafikverkets Effektkatalog Bygga om och Bygga nytt (version april 2014). Förutsättning för överbelastning är att överbelastningen varar en timme med trafikflöde 0 efter denna timme. Metoden är implementerad i Capcal 4.0, (se Capcal 4.0 Användarhandledning Trivector2013:87).

För varje delavsnitt finns kommentarer på vänster sida och beräkningsstegen på högersida. Dokumentet bör således läsas och skrivas ut dubbelsidigt för bästa läsbarhet.

Definitioner i form av allmänna termer och beteckningar är dokumenterade i kapitel 1 avsnitt 1.7. och litteraturreferenser i avsnitt 1.8.

The global aim of the ecoDriver project is to increase the fuel efficiency by 20% by optimising the driver-powertrain-environment feedback loop and delivering effective advice to drivers. In the course of the project, field experiments will take place with a wide range of vehicles — e.g. cars, light trucks and vans, medium and heavy trucks and buses — covering both individual and collective transport. The last step of the project (Sub Project 5; SP5) is to scale up the results from these tests and analyse costs and benefits for a number of futurescenarios.

The aim of SP5 is to predict the impact of the ecoDriver systems and solutions in the future, drawing on all the evaluations carried out in the project. With the results of SP5 it will be possible to make estimates about the costs and benefits of the suggested green driving support systems on a global (EU-27) level, both for society as a whole and for sub-groups like manufacturers and consumers. SP5 will construct a set of possible scenarios for the future depending on various road maps envisioned today. The predictions for future years will be made based on available data from within and outside of the project, and on advanced microscopic traffic modelling. SP5 takes the following steps to meet the objectives:

• Collect data needed for scaling up and developing scenarios
• Create a range of scenarios
• Assess the network implications of green driving support systems for future networks
• Predict the global impacts for a range of systems and scenarios
• Carry out a cost benefit analysis for a range of systems and scenarios

This deliverable describes the data needs for each step. It also contains a description of the approaches proposed for the scenario building, the microscopic traffic simulations, the scaling up and the cost-benefit analysis.

Detta kapitel behandlar beräkning av kapacitet, fördröjning, andel stopp och kölängd för:

• 3- och 4-vägskorsningar med stopp- eller väjningsplikt förunderordnade tillfarter
• Det finns även korrigeringar för beräkning av korsningar medhögerregel.

Metoden behandlar också överbelastning enligt metodik i Trafikverkets Effektkatalog Bygga om och Bygga nytt (version april 2014). Förutsättning för överbelastning är att överbelastningen varar en timme med trafikflöde 0 efter denna timme. Metoden är implementerad i Capcal 4.0, (se Capcal 4.0 Användarhandledning Trivector2013:87).

För varje delavsnitt finns kommentarer på vänster sida och beräkningsstegen på högersida. Dokumentet bör således läsas och skrivas ut dubbelsidigt för bästa läsbarhet.

Defnitioner i form av allmänna termer och beteckningar är dokumenterade i kapitel 1 avsnitt 1.7. och litteraturreferenser i avsnitt 1.8.

Syftet med föreliggande handbok är att beskriva hur trafiksimulering kananvändas som en alternativ metod eller komplement till analytiska metoderför att bestämma kapacitet och framkomlighet. Liksom metodbeskrivningarnai TRV2013/64343 är beskrivningarna avsedda att kunna användas för att medhjälp av trafiksimulering uppskatta effekterna av en given utformning isamband med planering, konsekvensanalys, projektering och drift avvägtrafikanläggningar. Simulering kan användas som ett komplement till deanalytiska metoderna, eller som ersättning i fall som inte täcks av dessametoder. Härigenom minskas risken för onödiga kostnader förorsakade avsåväl över- som underkapacitet.

Syftet med föreliggande handbok är att beskriva hur trafiksimulering kan användas som en alternativ metod eller komplement till analytiska metoder för att bestämma kapacitet och framkomlighet. Liksom metodbeskrivningarna i TRV2013/64343 är beskrivningarna avsedda att kunna användas för att med hjälp av trafiksimulering uppskatta effekterna av en given utformning i samband med planering, konsekvensanalys, projektering och drift av vägtrafikanläggningar. Simulering kan användas som ett komplement till de analytiska metoderna, eller som ersättning i fall som inte täcks av dessa metoder. Härigenom minskas risken för onödiga kostnader förorsakade av såväl över- som underkapacitet.

Kapitlet omfattar beräkning av kapacitet, belastningsgrad och reshastighet för personbilar (inkl. släp), lastbilar utan släp och lastbilar med släp för:

• tvåfältig landsväg (tvåfält),
• mötesfri motortrafikled (MML) med varierande andel omkörningsfältmellan 40 – 85 %,
• mötesfri landsväg (MLV) med varierande andel omkörningsfält mellan 15 – 85 %,

med hastighetsgräns 70 km/h eller högre.

Med ömkörbar längd avses andel sträcka i en riktning som är två körfält inkluderat inledningssträckor och exkluderat avslutningssträckor.

För MML med 100 % tvåfält, som innebär två genomgående körfält genom trafikplats, behandlas vägen som en fyrfältsväg (4 F) med motsvarande hastighetsgräns, hastighet och kapacitet, se Kapitel 2. Vid enbart ett genomgående körfält i trafikplats erhålls max 85 % omkörningsfält och övergången två till ett ger en väsentlig lägre kapacitet.

Reshastighetsberäkningen avser förhållanden utan överbelastning. För överbelastade situationer anges samma schablonmetod som i Trafikverkets effektsamband (version april 2014).

De geometri- och regleringsdata som ingår är andel omkörbarlängd (för mötesseparerad väg), hastighetsbegränsning och siktklass.

Trafikflödesberoende data som behövs är trafikflöde, riktningsfördelning och andel tunga fordon. För tvåfältsvägar ingår vägbredd som indata medan för MML och MLV antas att:

• MLV 15-30 % har en standardvägbredd på 8-12 m
• MML/MLV 40 % har en standardvägbredd på 13-14 m
• MML/MLV 60 % har en standardvägbredd på 13-17 m
• MML/MLV 85 % har en standardvägbredd på 16-18 m,

Beräkningsmetoden är en uppdatering av beräkningsmetoden i Effektsamband för transportsystemet Bygg om eller bygg nytt (Trafikverket, 2014 April; Carlsson, 2007). För varje delmoment finns kommentarer på vänster sida och beräkningsstegen på högersida. Dokumentet bör således läsas och skrivas ut dubbelsidigt för bästa läsbarhet.

Definitioner omfattande allmänna termer och deras beteckningar är dokumenterade i kapitel 1 avsnitt 1.7.

Litteraturreferenser inkluderande de som avser landsvägar är dokumenterade i kapitel 1 avsnitt 1.8

The purpose of this study is to take stock of Swedish data and studies that could form the basis of the estimation of marginal costs for congestion on roads and scarcity of railway capacity. Furthermore, the development of methods to estimate and evaluate the congestion in public transport is discussed. The Transport Administration’s investigation from 2013 indicates that there are persistent congestion problems in Stockholm that would be affected by adjustments to the toll cordon, congestion tax levels and differentiation with respect to time and place. A second part quantifies congestion in the road network outside urban areas by the use of extensive flow and velocity measurements from the E4 south of Stockholm as an example. This choice is justified as an example of a highly trafficked road link where the speed reductions occur regularly. The results show that the flows during May to December 2013 regularly were so high that speed dropped below 60 kilometer per hour for long periods. The marginal cost of congestion can be expressed as the change in the cost of a change in density and is highest at densities close to the road’s capacity. When the density is greatest, the marginal cost of one further car is about 10 SEK per kilometer. Compared to the taxes on petrol which is about 0,34 SEK per kilometer, which is considered to cover the marginal costs of all other externalities including carbon emissions. Thus, congestion costs may therefore be considered significant. For train slots, this study has used the Transport Administration’s electronic record of the operators’ requests for train slots from the National train plan for 2013 and the corresponding documentation of the determined slots. We find that the total of allocated slots for the largest operator of passenger trains in Sweden, SJ, received 99 percent of the slots it had applied for and Green Cargo 97 percent. This is not a strong indication of scarcity. For crowding in public transport there are a number of British studies of the valuation of travel in crowded conditions in passenger trains. These valuations have mostly been calculated as a multiplicative factor on time values when the passengers travel without crowding. The report presents studies that show that the willingness to pay may be substantially higher for a shorter travel time if the trip takes place in crowded conditions.

This report presents suggestions for new speed-flow relationships for motorways (MV), low standard motorways (4F), oncoming lane separated highways with grade separated intersections (MML), oncoming lane separated highways with at grade intersections (MLV), and two-lane highways. The suggestions are based on measurements from Trafikverket’s traffic count measuring system TMS in combination with model calculations. The TMS data have, for each road category, been quality checked, processed and analyzed. The data material is presented as speed-flow graphs for personal cars, trucks/buses without trailers and trucks with trailers. A comparison of the current speed-flow relationships and the TMS measurements was then conducted for each road category. The revised set of relationships then constituted the base for the 2012 revision of Trafikverket's publication "Effect calculations for road facilities". The analysis conducted resulted in suggestions to decrease the free flow speed and the travel speed at higher flows for most of the road categories. For motorways also a decrease in capacity is suggested. For oncoming separated highways (both MML and MLV) are minor changes of the capacity suggested. The suggested capacity value is for MML and MLV 1550 vehicles/h independently of speed limit and lane/road width.

This VTI report presents a method for calculating expected queue length and travel timedelay on one lane road sections without overtaking possibilities. The method wasdeveloped 2001 and presented in a working paper. The background for the modeldevelopment was that the Swedish Road Administration (now the Swedish TransportAdministration) planned to build so called 1+1 roads, i.e. roads with longer sectionswithout overtaking possibilities. The method developed has later on also shown to bevaluable for level of service calculations of 2+1 roads with varying share of two lanesections and for developing speed-flow relationships for the Administration’s ”Effectcalculations for road facilities”.The method uses section length, traffic flow, average speed and standard deviation asinput. The method is divided with respect to calculation of effects due to single slowrunningvehicles and effects at “normal” speed distribution. Since no data wereavailable when the model was developed, the model results were instead compared totraffic simulations with the microscopic traffic simulation model AIMSUN. The resultsshow a good correlation but the analytical model gives in general approximately 1.2 percent lower travel time delay. The differences can probably partly be explained by thestochastic parts of the simulation model. One should also remember that neither theanalytical model nor the simulation model has been calibrated and validated with realdata for this type of roads. Thus, the differences between the models do not necessaryimply that the analytical model is the one deviating from reality.

Microscopic traffic simulation is often used as an alternative or complementary tool to analytical methods and procedures for level-of-service analyses of road traffic facilities. The increased usage of traffic simulation for level-of-service analysis has raised a need for guidelines on how to apply and use traffic simulation models. Many countries have developed or are currently developing traffic simulation guidelines. This is also the case in Sweden, were the new Swedish highway capacity manual will include a chapter on traffic simulation. This paper presents a survey of the current traffic simulation guidelines in USA, Germany, UK, Denmark and Sweden. The guidelines have been analysed with respect to the aspects covered: when to apply simulation; the workflow of a simulation study; data collection needs; calibration and validation; experimental design; statistical analysis; and calculation of level-of-service measures. The guidelines analysed are focused on different aspects and none of them covers all of the topics listed above. Some of the guidelines are connected to specific simulation software packages and some are written in a more general manner. Most of the aspects covered are general and applicable in any country. The main reason for developing country specific guidelines is often a need for guidelines in the local language. Experimental design and statistical analysis are not treated extensively in the guidelines; neither do the guidelines discuss how to deal with calibration based on limited real world measurements. Calculation of level-of-service measures are quite extensively treated in some of the guidelines and to a little extent in others. All of the guidelines contain important contributions for the simulation chapter of the new Swedish highway capacity manual

Autonomous vehicles can be used to create realistic simulations of surrounding vehicles in driving simulators. However, the use of autonomous vehicles makes it difficult to ensure reproducibility between subjects. In this paper, an effort is made to solve the problem by combining autonomous vehicles and controlled events, denoted plays. The aim is to achieve the same initial play conditions for each subject, since the traffic situation around the subject will be dependant upon each subject's actions while driving in the autonomous traffic. This paper presents an algorithm that achieves the transition from autonomous traffic to a predefined start condition for a play. The algorithm has been tested in the VTI driving simulator III with promising results. In most of the cases the algorithm could reconstruct the specified start condition and conduct the transition from autonomous to controlled mode in a non-conspicuous way. Some problems were observed regarding moving unwanted vehicles away from the closest area around the simulator vehicle, and this part of the algorithm has to be enhanced. The experiment also showed that the controlled every-day life traffic normally used in the VTI driving simulator makes subjects drive faster than in autonomous traffic.

Roundabouts have become a more common type of intersection in Sweden over the last 30 years. In order to evaluate the roundabout level-of-service both analytical models and simulation models are being used. Analytical traffic models for intersections, such as the Swedish capacity model Capcal, has difficulties estimating the level-of-service of a roundabout if there are pedestrians and cyclists at crossings located close to the roundabout. It is well known that a crossing located after a roundabout exit can cause an up-stream blocking effect that affects the performance of the roundabout. But how the upstream blocking effect depends on the different flows of vehicles and pedestrians is not known. In this paper an existing analytical model by Rodegerdts and Blackwelder has been investigated and compared to simulations in VISSIM and measurements from Swedish roundabouts. The purpose of this investigation is to examine if the model by Rodegerdts and Blackwelder is suitable for implementing into existing analytical models such as Capcal. The results show that the model by Rodegerdts and Blackwelder can estimate if a capacity loss will occur, but the magnitude of this loss is more difficult to evaluate. The conclusion and recommendation is that the model by Rodegerdts and Blackwelder should be implemented into the Swedish capacity model Capcal. The model by Rodegerdts and Blackwelder is to be used as a warning system if the results in Capcal are too uncertain to use for analysis of the roundabout performance.

This paper presents a design methodology for driving simulator scenarios in which autonomous and controlled surrounding vehicles are combined. The main motives are to achieve both a high realism and high reproducibility. The methodology is introduced using a theater metaphor in which a driving simulator scenario is defined as a constellation of: everyday life driving, preparations for plays, and plays. Advantages, disadvantages and difficulties with the proposed methodology are discussed.

This paper presents a modified version of the Intelligent Driver Model (IDM) [M. Treiber, A. Hennecke, and D. Helbing, Phys. Rev. E. 62, 2 (2000)]. The IDM is a car-following model. A car-following model controls the accelerations of individual vehicles in a microscopic traffic simulation model. The original IDM has been observed to result in negative vehicle accelerations in situations where the distance to the preceding vehicle is much larger than the estimated desired safety distance. In this paper, we propose a modified function for the interaction with preceding vehicles which do not include this model property. A comparison of the results of simulations with the original and the modified IDM shows that the modified IDM results in higher average speed for a specific flow, a less steep speed-flow relationship and higher capacity. The speed-flow relationships of simulations with the modified IDM are also shown to better match the speed-flow relationships in real traffic on Swedish freeways. The differences between the results for the original and the modified IDM increase if the models are extended to include drivers' anticipation of the downstream traffic condition.

Driving simulators and microscopic traffic simulation are important tools for making evaluations of driving and traffic. A driving simulator is de-signed to imitate real driving and is used to conduct experiments on driver behavior. Traffic simulation is commonly used to evaluate the quality of service of different infrastructure designs. This thesis considers a different application of traffic simulation, namely the simulation of surrounding vehicles in driving simulators.

The surrounding traffic is one of several factors that influence a driver's mental load and ability to drive a vehicle. The representation of the surrounding vehicles in a driving simulator plays an important role in the striving to create an illusion of real driving. If the illusion of real driving is not good enough, there is an risk that drivers will behave differently than in real world driving, implying that the results and conclusions reached from simulations may not be transferable to real driving.

This thesis has two main objectives. The first objective is to develop a model for generating and simulating autonomous surrounding vehicles in a driving simulator. The approach used by the model developed is to only simulate the closest area of the driving simulator vehicle. This area is divided into one inner region and two outer regions. Vehicles in the inner region are simulated according to a microscopic model which includes sub-models for driving behavior, while vehicles in the outer regions are updated according to a less time-consuming mesoscopic model.

The second objective is to develop an algorithm for combining autonomous vehicles and controlled events. Driving simulators are often used to study situations that rarely occur in the real traffic system. In order to create the same situations for each subject, the behavior of the surrounding vehicles has traditionally been strictly controlled. This often leads to less realistic surrounding traffic. The algorithm developed makes it possible to use autonomous traffic between the predefined controlled situations, and thereby get both realistic traffc and controlled events. The model and the algorithm developed have been implemented and tested in the VTI driving simulator with promising results.

  In recent years, the transport simulation of large road networks has become far more rapid and detailed, and many exciting developments in this field have emerged.  Within this volume, the authors describe the simulation of automobile, pedestrian, and rail traffic coupled to new applications, such as the embedding of traffic simulation into driving simulators, to give a more realistic environment of driver behavior surrounding the subject vehicle. New approaches to traffic simulation are described, including the hybrid mesoscopic-microscopic model and floor-field agent-based simulation. Written by an invited panel of experts, this book addresses students, engineers, and scholars, as well as anyone who needs a state-of-the-art overview of transport simulation today.