Department of Science and Technology
Institutionen för teknik och naturvetenskap
Linköping University
Linköpings universitet
g
n
i
p
ö
k
r
r
o
N
4
7
1
0
6
n
e
d
e
w
S
,
g
n
i
p
ö
k
r
r
o
N
4
7
1
0
6
-E
S
LiU-ITN-TEK-A--17/030--SE
Evaluation of bus terminals
using microscopic traffic
simulation
Caroline Askerud
Sara Wall
LiU-ITN-TEK-A--17/030--SE
Evaluation of bus terminals
using microscopic traffic
simulation
Examensarbete utfört i Transportsystem
vid Tekniska högskolan vid
Linköpings universitet
Caroline Askerud
Sara Wall
Handledare Therese Lindberg
Examinator Anders Peterson
Upphovsrätt
Detta dokument hålls tillgängligt på Internet – eller dess framtida ersättare –
under en längre tid från publiceringsdatum under förutsättning att inga
extra-ordinära omständigheter uppstår.
Tillgång till dokumentet innebär tillstånd för var och en att läsa, ladda ner,
skriva ut enstaka kopior för enskilt bruk och att använda det oförändrat för
ickekommersiell forskning och för undervisning. Överföring av upphovsrätten
vid en senare tidpunkt kan inte upphäva detta tillstånd. All annan användning av
dokumentet kräver upphovsmannens medgivande. För att garantera äktheten,
säkerheten och tillgängligheten finns det lösningar av teknisk och administrativ
art.
Upphovsmannens ideella rätt innefattar rätt att bli nämnd som upphovsman i
den omfattning som god sed kräver vid användning av dokumentet på ovan
beskrivna sätt samt skydd mot att dokumentet ändras eller presenteras i sådan
form eller i sådant sammanhang som är kränkande för upphovsmannens litterära
eller konstnärliga anseende eller egenart.
För ytterligare information om Linköping University Electronic Press se
förlagets hemsida
http://www.ep.liu.se/
Copyright
The publishers will keep this document online on the Internet - or its possible
replacement - for a considerable time from the date of publication barring
exceptional circumstances.
The online availability of the document implies a permanent permission for
anyone to read, to download, to print out single copies for your own use and to
use it unchanged for any non-commercial research and educational purpose.
Subsequent transfers of copyright cannot revoke this permission. All other uses
of the document are conditional on the consent of the copyright owner. The
publisher has taken technical and administrative measures to assure authenticity,
security and accessibility.
According to intellectual property law the author has the right to be
mentioned when his/her work is accessed as described above and to be protected
against infringement.
For additional information about the Linköping University Electronic Press
and its procedures for publication and for assurance of document integrity,
please refer to its WWW home page:
http://www.ep.liu.se/
Abstract
Traffic simulation is a safe and efficient tool to investigate infrastructural changes as well as traffic conditions. This master thesis aims to analyse a microscopic traffic simulation method for evalua-tion of bus terminal capacity. The evaluaevalua-tion is performed by investigating a case study of the bus terminal at Norrk¨oping travel centre. The analysed method, referred to as terminal logic in the thesis, uses a combination of time based and event based simulation. Through the combination of time and event, it is possible to capture all movements within the terminal for individual vehicles. The simulation model is built in the software Vissim.
A new travel centre for Norrk¨oping is under development. Among the reasons for a new travel centre is the railway project Ostl¨anken in the eastern part of Sweden. An evaluation of the bus terminal is interesting due to a suspicion of overcapacity and the opportunity of redesigning. To investigate both the terminal capacity and the terminal logic, three scenarios were implemented.
• Scenario 1: Current design and frequency • Scenario 2: Current design with higher frequency
• Scenario 3: Decreased number of bus stops with current frequency
The results from the scenarios confirm the assumption of overcapacity. The capacity was evaluated based on several different measures, all indicating a low utilization. Even so, the utilization was uneven over time and congestion could still occur when several buses departed at the same time. This was also seen when studying the simulation, which showed congestions when several buses departed at the same time. The case study established the terminal logic to be useful when evaluating capacity at bus terminals. It provides a good understanding of how the terminal operates and captures the movements. However, it was time-consuming to adjust the logic to the studied terminal. This is a disadvantage when investigating more than one alternative. The thesis resulted in two main conclusions. Firstly, a more optimised planning of the buses at Norrk¨oping bus terminal would probably be achievable and lead to less congestions at the exits. Secondly, the terminal logic is a good method to use when evaluating bus terminals but it is not straight forward to implement.
Keywords: Microscopic traffic simulation, Vissim, VisVap, Bus terminal, Capacity, time based simulation, event based simulation.
Sammanfattning
Trafiksimulering ¨ar ett s¨akert och effektivt verktyg f¨or att unders¨oka b˚ade infrastrukturf¨or¨andringar andra trafiksituationer. Syftet med detta examensarbete ¨ar att analysera en mikroskopisk trafik-simuleringsmetod f¨or utv¨ardering av kapaciteten hos bussterminaler. Norrk¨opings resecentrum anv¨ands som ett praktikfall f¨or att genomf¨ora utredningen. Den analyserade metoden, h¨anvisad som terminallogik i examensarbetet, best˚ar av en kombination av tidsbaserad och h¨andelsebaserad simulering. Kombinationen av tid och h¨andelse m¨ojligg¨or att f˚anga r¨orelser inom terminalen f¨or individuella fordon. Simuleringsmodellen ¨ar byggd i simuleringsverktyget Vissim.
Ett nytt resecentrym f¨or Norrk¨oping ¨ar under utveckling. En av de bakomliggande orsakerna till det nya resecentrumet ¨ar j¨arnv¨agsprojektet Ostl¨anken som ska g˚a igenom ¨Ostra Sverige. En utv¨ardering av den nuvarande bussterminalen ¨ar intressant p˚a grund av att det finns en misstanke att terminalen har ¨overkapacitet samt att det finns m¨ojlighet att f¨or¨andra terminelen i och med nya resecentrum. F¨or att unders¨oka b˚ade kapaciteten hos Norrk¨opings bussterminal och terminal-logiken formulerades tre olika scenarion.
• Scenario 1: Nuvarande utformning och frekvens • Scenario 2: Nuvarande utformning men h¨ogre frekvens • Scenario 3: Minskat antal h˚allplatser men nuvarande frekvens
Resultaten fr˚an scenariona bekr¨aftar antaget om ¨overkapacitet vid terminalen. Kapaciteten utv¨arder-ades baserad p˚a flera olika m¨atv¨arden som alla indikerade l˚ag utnyttjandegrad. Utnyttjandegraden var dock oj¨amn ¨over tid, vilket ledde till att tr¨angsel kunde uppst˚a n¨ar flera bussar avgick samtidigt fr˚an terminalen. Detta kunde ocks˚a ses genom att studera simuleringen som visade att det blev tr¨angsel n¨ar flera bussar avgick samtidigt. Praktikfallet p˚avisade att terminallogiken ¨ar anv¨ and-bar f¨or att utv¨ardera kapaciteten hos bussterminaler. Terminallogiken tillhandah˚aller f¨orst˚aelse f¨or hur terminaler fungerar och f˚angar bussarnas r¨orelser. Dock var det tidskr¨avande att anpassa logiken till den studerande terminalen. Det ¨ar en nackdel om flera alternativ ska unders¨okas. Ex-amensarbetet resulterade i tv˚a huvudslutsatser. F¨or det f¨orsta borde det vara m¨ojligt att skapa en mer optimerad planering f¨or bussarna som trafikerar terminalen, vilket ocks˚a borde leda till mindre tr¨angsel vid utfarterna. F¨or det andra ¨ar terminallogiken en bra metod att anv¨anda f¨or utv¨ardering av bussterminaler, men den ¨ar inte helt okomplicerad att implementera.
Keywords: Mikroskopisk trafiksimulering, Vissim, Bussterminal, Kapacitet, tidsbaserad simuler-ing, h¨andelsebaserad simulering.
Acknowledgements
Firstly we would like to thank our supervisor Therese Lindberg and examiner Anders Peterson at Link¨oping University for their support and feedback during this thesis. We would also like to thank Sweco Society in Norrk¨oping for the opportunity to perform this thesis, the employees also deserves a thank for being welcoming and helpful during the thesis. A special thank at Sweco is to our supervisor, Johan Ericsson, and technical supervisor, Magnus Fransson, for their guidance and support throughout the thesis. ¨Ostg¨otatrafiken, Weidermans buss och Nobina Sverige AB deserves a special thanks for providing us with data and taking the time to answer our questions about the bus terminal
Additionally we would like to thank PTV for letting us use an academic license of the microscopic simulation software Vissim, which was necessary in order to perform this thesis.
Norrk¨oping, June 2017
Contents
1 Introduction 1
1.1 Background . . . 1
1.2 Purpose and research questions . . . 2
1.3 Delimitations . . . 2
1.4 Outline . . . 2
2 Planning and designing bus terminals 3 2.1 Planning guidelines . . . 3
2.2 Requirements for bus terminals . . . 4
2.3 Designing bus terminals . . . 4
2.4 Design alternatives . . . 5
2.4.1 Saw-tooth design . . . 5
2.4.2 Drive-through . . . 6
2.4.3 Centre platform . . . 6
2.4.4 Angle berth . . . 7
3 Bus terminal capacity 9 3.1 Capacity definition . . . 9
3.2 Determination of the capacity . . . 9
3.2.1 Using analytical methods . . . 10
3.2.2 Using simulation . . . 12
4 Microscopic traffic simulation 13 4.1 Time vs. event based simulation . . . 13
4.2 Behaviour models . . . 14
4.2.1 Car-following model . . . 14
4.2.2 Gap-acceptance model . . . 14
4.3 Required data . . . 15
4.4 The simulation software Vissim . . . 15
4.5 Microsimulation of stops and terminals . . . 16
4.6 Verification, calibration and validation . . . 16
5 Framework for terminal logic 19 5.1 Preliminaries . . . 19
5.2 Operations in Vissim . . . 20
5.2.1 Generating buses . . . 21
5.2.2 Buses entry the terminal . . . 21
5.2.3 Bus stops and layover area . . . 21
5.3 Operations in VisVap . . . 22
5.3.1 Receive . . . 23
5.3.2 Serve . . . 23
5.3.3 Release . . . 23
5.3.4 Reset . . . 23
5.4 Implementation of terminal logic operations in VisVap . . . 24
6 Case study of Norrk¨oping bus terminal 29 6.1 The current terminal . . . 29
6.2 Technical details for the implementation . . . 30
6.3 Data collection . . . 31
6.3.1 Bus movements within the terminal . . . 31
6.3.2 Peak hour . . . 32
6.3.3 Field observation . . . 34
6.4 Implementation of Norrk¨oping bus terminal . . . 35
6.5 Modifications of the general terminal logic . . . 36
6.6 Model analysis . . . 38
6.6.1 Verification and validation . . . 38
6.6.2 Scenario 1: Current design and frequency . . . 39
6.6.3 Scenario 2: Current design with higher frequency . . . 39
6.6.4 Scenario 3: Decreased number of bus stops with current frequency . . . 40
6.7 Sources of error . . . 40
7 Result and analysis 41 7.1 Case study of Norrk¨oping bus terminal . . . 41
7.1.1 Scenario 1: Current design and frequency . . . 41
7.1.2 Scenario 2: Current design with higher frequency . . . 44
7.1.3 Scenario 3: Decreased number of bus stops with current frequency . . . 47
7.1.4 Capacity evaluation of the numerical results for the case study . . . 49
8 Discussion 51 8.1 Evaluation of the thesis . . . 51
8.2 Evaluation of the terminal logic . . . 52
9 Conclusion and further research 55 9.1 Conclusion . . . 55
9.2 Further research . . . 55
References 57
List of Figures
1 The saw-tooth design. . . 5
2 The drive-through design. . . 6
3 The centre platform design. . . 7
4 The angle berth design. . . 7
5 Flowchart for simulation. . . 13
6 Illustration of how VisVap and VisVap are connected, with the different inputs and outputs. . . 19
7 The queue construction. . . 20
8 The lap queue construction . . . 22
9 Flowchart of the terminal logic. . . 22
10 Flowchart for VisVap. . . 24
11 The different windows in VisVap, one for the flowchart and several for inputs and subroutines (PTV GROUP, 2014). . . 25
12 Location of Norrk¨oping bus terminal. . . 29
13 Design of Norrk¨oping bus terminal. . . 30
14 Number of arrivals and departures for each minute. . . 32
15 Number of arrivals and departures per 10 minutes intervals. . . 33
16 Number of arrivals and departures per hour. . . 33
17 The created network of Norrk¨oping bus terminal in Vissim. . . 35
18 The number of occupied bus stops for the current situation, bus stops at the layover area is not included. The total number of bus stops is 12. . . 42
19 The number of occupied bus stops at the layover area. The total number of bus stops is 13. . . 43
20 The number of occupied regular bus stops over time with increased frequency. The total number of bus stops is 12. . . 45
21 The number of occupied bus stops at the layover area with increased frequency. The total number of bus stops is 13. . . 45 22 The number of occupied bus stops with current frequency but a third of the bus stops. 47 23 The number of occupied bus stops at the layover area with a third of the bus stops. 48
List of Tables
1 The receive function . . . 25
2 The serve function . . . 26
3 The release function . . . 27
4 The reset function . . . 27
5 Data necessary for the simulations . . . 31
6 Type of the gathered data. . . 34
7 Validation of average waiting times at exits. . . 39
8 Utilization of bus stops . . . 43
9 Number of delays within each time interval. . . 43
10 Delay at the terminal. . . 44
11 Utilization of bus stops with higher frequency . . . 46
12 Number of delays within each time interval for the increased frequency. . . 46
13 Delay at the terminal for the increased frequency. . . 46
14 Utilization of bus stops with a third of the bus stops. . . 48
15 Number of delays within each time interval with a third of the bus stops. . . 48
16 Delay at the terminal for the current frequency and a third of the bus stops. . . 49
17 A summary of the numerical results for the three scenarios. . . 50
Abbreviations
HCM2000 Highway Capacity Manual 2000
ITS Intelligent Transport System
OD-matrix Origin-Destination Matrix
Macro Macroscopic
Meso Mesoscopic
Micro Microscopic
PASSION PArallel Stop SimulatION
PT Public Transport
TRAST TRafik f¨or en Attraktiv STad (Transport for an attractive city)
USA United States of America
Terminology
Bus stop A designated place where buses stop for passengers to board or alight from the bus.
Bus terminal A terminal for buses, often placed within or connected to a travel centre. The bus terminal consists of several individual bus stops and can have different designs.
Critical-gap The minimum major-stream headway during which a minor-street vehicle can make a maneuver.
Dwell time The time a vehicle spends at a scheduled stop without moving. Interchange Change between different public transport modes.
TransLink A division of the Department of Transport and Main Roads in Queensland, Australia.
Travel centre Railway station or railroad station and a junction for different means of travel.
1 INTRODUCTION
1
Introduction
Travel centres, and consequently also bus terminals, have an important role when developing cities and infrastructure. The railway has worked as an engine for cities for a long time and the public transport becomes more and more important. Hence, travel centres and bus terminals needs to be well-functioning in order to achieve sufficient public transport with the station as a link to the city (Trafikverket, 2013).
1.1
Background
In Sweden, there is an infrastructural project towards a new generation of railways, called Ostl¨anken (East Link project). Ostl¨anken will go through the eastern part of Sweden, making several au-thorities and municipalities involved in the project. The authority responsible for Ostl¨anken is Trafikverket (Swedish Transport Administration). Information about the Ostl¨anken project can be found at for example Trafikverket (2014), Ostl¨anken (n.d.) and Nyk¨opings kommun (n.d.). One of the cities Ostl¨anken will pass through is Norrk¨oping, which makes the municipality of Norrk¨oping a part of the project. Trafikverket and Norrk¨oping municipality have different interests regard-ing the outcome of the project. Trafikverket’s main interest is to achieve a good accessibility for Ostl¨anken through Norrk¨oping, while the municipality focuses on good connections and urban environments for the city. Ostl¨anken is expected to result in shorter travel times leading to more travellers using railways for their trips. The reduced travel time will likely lead to more commuters and work opportunities in cities at reach. Norrk¨oping municipality has plans to build a new travel centre associated to the Ostl¨anken project. Travel centre refers to the whole station area, both trains and buses. Information about Ostl¨anken and the new travel centre in Norrk¨oping can be found at Next:Norrk¨oping (n.d.).
The current travel centre consists of three different parts. One part is for trains, another is for trams and city buses and then there is a separate bus terminal for buses going outside of the city. The bus terminal handles public transport in form of express buses between different towns in the region and also regional buses to rural areas around Norrk¨oping. The terminal also handles long-distance traffic to other parts of Sweden. The regional public transport provider ¨Ostg¨otatrafiken is responsible for all traffic at the terminal except the long-distance traffic. Several private bus companies use the terminal to provide long-distance traffic.
When planning the new travel centre in Norrk¨oping, there is an opportunity to change the design of the bus terminal. The capacity of the existing bus terminal is unknown, but overcapacity is suspected. Since the travel centre is located near the city centre, it is placed on valuable land. For city planners, it can be interesting to use this land for properties, parks or other urban purposes. The possible traffic increase is another reason to aim for a space-efficient terminal. The urban environment in combination with increased demand restricts the possibilities for expanding the terminal in order to fulfil the demand. Therefore, it is of high importance to have a space-efficient bus terminal. In order to decide whether a change of the terminal is desirable, the current terminal and its capacity needs to be evaluated.
To enable an evaluation of the bus terminal, capacity, which is a measure of the utilization, is of high importance and needs to be properly defined. This can be done in several different ways and thus needs to be investigated. This will require a good understanding of terminal functionalities. Microscopic traffic simulation is a tool that can be used to evaluate bus terminals and evaluate capacity. It provides a high level of detail and can capture details and movements of individual vehicles at bus terminals. Sweco, a Swedish consultant firm, has performed several evaluations of bus terminals using the microscopic traffic simulation tool Vissim. In excess of being used by Sweco, Vissim has an additional program called VisVap that can be used to manage vehicle movements within terminals. These two factors make Vissim an appropriate tool to use to evaluate bus terminals and have been chosen for this thesis.
1.2 Purpose and research questions
1.2
Purpose and research questions
Bus terminals can be simulated and evaluated in different microscopic simulation softwares. Sweco has performed evaluations of bus terminals in Vissim in the past. The method that Sweco has used can be adjusted depending on the design of the terminal that is being investigated. Therefore, it is interesting to investigate their general terminal logic and how it can be adjusted to the current bus terminal in Norrk¨oping. When adjusted to Norrk¨oping bus terminal, the terminal logic is analysed based on its possibilities to enable capacity evaluations of bus terminals.
The purpose of this thesis is to investigate how the microscopic traffic simulation tool Vissim can be used to evaluate bus terminals and furthermore, to investigate how the method used by Sweco can be adjusted to Norrk¨oping bus terminal and thereby also evaluate the capacity.
The contribution of this master thesis is to analyse and investigate a method for using micro sim-ulation to evaluate the capacity of bus terminals regarding utilization and delay. More specifically the contribution will be an evaluation of using the software Vissim in combination with VisVap to achieve both time and event based simulation. Additionally, this thesis will investigate the term capacity and how the capacity can be determined for bus terminals.
The following research questions has been formulated to answer the purpose.
• How can Vissim in combination with VisVap, terminal logic, be used to evaluate bus terminals with respect to capacity?
– How can the model be implemented and verified?
• Which estimations are possible to make about the capacity at Norrk¨oping bus terminal? – Furthermore, are there any improvements to be made regarding space or efficiency?
1.3
Delimitations
The simulations in this thesis focuses on bus traffic within a bus terminal. Thus, no car traffic or pedestrians are included in the simulations. The study area is delimited to only include the bus terminal at the travel centre in Norrk¨oping. The simulations are performed in Vissim and no other microsimulation tool is used.
1.4
Outline
This thesis is structured as follows. Chapter 2. Planning and designing bus terminals and 3. Bus terminal capacity contains a theoretical framework for the thesis. Chapter 2 is about planning, design and localisation of bus terminals. The chapter covers information about where travel centres should be located in a city, where bus terminals should be located in relation to the rest of the travel centre and different layout alternatives for bus terminals. Chapter 3. Bus terminal capacity, is about bus terminal capacity. This chapter contains information about how capacity can be defined and determined for bus terminals.
Chapter 4. Microscopic traffic simulation and 5. Framework for terminal logic contains a de-scription of the method used in the thesis. The fourth chapter contains information about traffic simulation with focus on microscopic simulation and Vissim. The fifth chapter describes the ter-minal logic, previously used by Sweco.
The terminal logic is adjusted and used in a case study on Norrk¨oping bus terminal. The case study is presented in Chapter 6. Case study of Norrk¨oping bus terminal. This chapter also includes changes and additions to Sweco’s logic.
Chapter 7. Result and analysis presents the results from the case study together with an analysis of the results and the terminal logic. Chapter eight and nine contains discussion and conclusions from the thesis.
2 PLANNING AND DESIGNING BUS TERMINALS
2
Planning and designing bus terminals
Planning of bus terminals and travel centres in general, is a complex matter within the public transport area. The discipline is complex since it requires a combination of aspects from both transport engineering and urban planning.
2.1
Planning guidelines
There are authorities and public transport companies from several countries that provides guide-lines and handbooks for planning of bus terminals and travel centres. Mostly Swedish guideguide-lines are covered in this thesis. A handbook from Australia is mentioned as an example of an external guideline.
The Australian guideline is the Public Transport Infrastructure Manual by TransLink (2016), which contains best practises and design principles for public transport infrastructure in Queensland, Australia. The manual clearly states TransLink’s expectations for both new and upgraded public transport infrastructure within the TransLink network. Several aspects need to be taken into consideration when planning for public transport. One aspect mentioned in the manual is the urban design. The infrastructure must work with the existing physical and social context, be sustainable, feel safe etcetera. Another handbook about travel centres and station areas is Stationshandbok, provided by Trafikverket (2013). The aim of this handbook is to create better travel centres in Sweden; better designed, more functional and more effective for travellers. The handbook is divided into several parts, covering the whole station area. Different guidelines and design principles are presented for each part of the travel centre. For example, the requirements are not the same for the platform and for the area inside a station building. The handbook has a specific subsection about buses and interchanges between different public transport modes. Fast and safe transfers should be given priority at travel centres. (Trafikverket, 2013)
Sveriges kommuner och landsting (Swedish Association of Local Authorities and Regions), Trafikver-ket and BoverTrafikver-ket (National board of housing, building and planning) has created a handbook called TRAST (Sveriges kommuner och landsting, 2015). The aim with TRAST is to guide planners in their work towards creating attractive cities, the main focus is on traffic. Kol-TRAST is a com-plementary handbook to TRAST and is immersed within the area of public transport planning (Sveriges kommuner och landsting, 2012). The handbook has a specific section about bus stops and transfer points, but the focus is more on bus stops in the route network than on bus terminals. As mentioned above, there are numerous handbooks for public transport planning available. Un-fortunately, many of them focuses more on travel centres in general than on bus terminals. When evaluating bus terminal capacity, it is the bus terminal area that is interesting and not the travel centre or the station area. Nevertheless, information regarding travel centres and single bus stops might be useful for terminal capacity due to limited existence of terminal specific information. There are a couple of handbooks with terminal focus available. One example of such a handbook is RITERM -09, developed by SL (public transport provider in Stockholm, Sweden) (SL, 2009). RITERM -09 is a set of guidelines with main focus on bus terminals and the handbook brings up fundamental conditions for bus terminals. For example, conditions for localisation, design and function of bus terminals. The handbook also describes how bus terminals should be designed within SL’s network. This means that some of the conditions in the guidelines are specific for SL, while other information is universal for bus terminals overall.
2.2 Requirements for bus terminals
2.2
Requirements for bus terminals
There are several different factors affecting whether a bus terminal is considered adequate. The factors is for example information, design or functionality. According to Sveriges kommuner och landsting (2012), bus terminals should have both high traffic safety for travellers and good ac-cessibility for public transport vehicles. Conflicts between different means of transport should be minimized and short driving distances should be the aim within the terminal. It is also desired that people with disabilities are taken into consideration when designing walkways and waiting areas within the terminal. Other things to take into consideration are possibilities for service, cleaning and management. It is also important that the terminal can stay effective during for example bad weather conditions.
It is important to consider required space, dimensioning and functionality when planning a bus terminal. The required space for the terminal can depend on for example the traffic load, the route network and the purpose of the bus lines using the terminal. These aspects can be difficult to know in advance. Sveriges kommuner och landsting (2012) brings up that terminals can be dimensioned considering articulated buses when designing the bus stops and bogie buses when determining the geometry within the terminal. This is a technique to determine the required space without knowing the exact demand and usage in advance. When it comes to the functionality, there are several different parts that is necessary for the terminal to be as effective as possible. Functionalities that always should be present and accommodated are: drop-off, layover time and pick-up. Other functionalities can vary depending on the size of the terminal. (Sveriges kommuner och landsting, 2012)
2.3
Designing bus terminals
The localisation and design of bus terminals and travel centres in general has an important role for the transport system and for the urban development. The localisation of bus terminals and travel centres can be seen as one, since bus terminals normally are a part of travel centres. Sveriges kommuner och landsting (2015) discuss where within the travel centre to place the bus terminal. The recommended placement depends on which kind of bus traffic that is going to use the bus terminal. For local traffic, it is recommended to use kerbside bus stop directly outside the entrance to the station or the platform. Therefore, the local buses do not always belong to the bus terminal but rather have their own bus stops. The regional traffic can be combined with the local traffic and located at the same place if no layover time is needed. Layover time is the time between departures, when the bus has no passengers aboard but not enough time to drive to a garage. This time is often spent at a layover area, which can be located close to the bus terminal. If layover time is necessary, it might be better to place the bus stops for the regular traffic at a specific terminal where a layover area is accessible.
Other factors important for the localisation of bus terminals are available space, required space and environmental requirements. RITERM -09 presents four questions that could be asked when building a new terminal or improving an existing one (SL, 2009).
• Which bus lines should use the terminal?
• Which functionalities are required for the terminal and how large area is necessary (need of space)?
• How should the terminal area be disposed in order to meet all requests regarding closeness, safety and connections between areas within the travel centre?
• How can the traffic be managed to minimize conflicts?
The answers to these questions are essential in order to obtain an overview over the bus terminals’ flows and functionalities, which gives an indication of the location and design needs for the terminal. The flows through the terminal can be determined by investigating the number of arriving and departing bus lines at the terminal. The flow affects the required space in several ways. Since bus lines have different functionalities, the required space can differ. Turning bus lines often has a common drop-off location and an individual pick-up location. A throughgoing bus line requires
2 PLANNING AND DESIGNING BUS TERMINALS
two separate bus stops, one in each direction. Turning bus lines may need a space near the bus terminal for layover time between arrivals and departures. (SL, 2009)
As mentioned previously, the terminal with bus stops for regional traffic can be placed together with the bus stops for local traffic at the travel centre. The problem can be to allocate space for layover time for the regional buses. Sveriges kommuner och landsting (2015) points out that bus lines does not necessarily need to have the terminal as end destination and thus no layover time there. If the terminal is just as any stop along the route, there is no need for space for time regulation. When the terminal is the end of the line and time regulation is required, more space is needed and the terminal cannot be placed right outside the entrance. In Trafikverket (2013), docking is proposed as a suitable solution. Trafikverket (2013) emphasis that the travellers and their interchanges should have focus when planning bus terminals and travel centres. Furthermore, the design and placement of the bus stops within the terminal should enable interchanges to be made with a normal walking pace. This means that the bus stops, and the bus terminal, need to be located near the main walkways from the train platforms. The location should be near the train platforms but without compromising the travellers’ safety. It is not recommended that travellers need to cross a busy street when changing travel mode. It is also important that the bus stops or the bus terminal is designed in a way that makes it possible for travellers to change between buses without crossing bus traffic in an unsafe way.
2.4
Design alternatives
The bus terminal at a travel centre can be designed in several different ways. The different design types vary in how the buses should be located within the terminal area, for example where they should have their pick-up and drop-off location. Several summaries of design alternatives exist. One is a diploma thesis written by N¨atterlund and Thomasson (2011), parts of the thesis is based on SL (2009). The following text about design alternatives for the buses’ loading area (berth) at bus terminals is based on Sveriges kommuner och landsting (2015), SL (2009), N¨atterlund and Thomasson (2011) and Brinckerhoff (2002).
2.4.1 Saw-tooth design
The loading area is designed so that the kerbside gets the shape of a ”saw-tooth”. The buses are placed with an angle against the street, see Figure 1. Saw-tooth design requires less space than placing the buses along the kerbside, but it can still require a lot of space depending on the number of buses. A benefit with this design is that it is easy for the passengers to see the destination signs on the buses. Another benefit is that it is possible to park the bus close to the kerbside, which enables comfortable boarding and disembarking for passengers. A disadvantage with this design is that it can be considered as ugly and not urbane. When using the saw-tooth design, the buses can leave their place at the terminal without reversing. This is an important advantage for pedestrian safety and simplifies for the drivers. The saw-tooth design is often used in combination with another design, in order to reduce the required space.
2.4 Design alternatives
2.4.2 Drive-through
The loading area of a drive-through design is designed so that the buses can drop-off and pick-up passengers and then ”drive through” the terminal on its way out. There are two different types of drive-through designs, a straight one and an angled one. The straight drive-through design is illustrated in Figure 2. The main advantage with the design is that it is space efficient. Overall, there are more disadvantages than advantages with this design. The drive-through design can be dangerous for passengers since they need to cross the buses path on their way to the right bus stop. The design can also lead to poor overview over the terminal. Despite all the disadvantages, it can be situations when a drive-through design is the only option due to shortage of space.
Figure 2: The drive-through design.
2.4.3 Centre platform
The loading area is located around a centre platform, where all pick-up and drop-off locations are located at the platform, see Figure 3. The traffic around the centre platform should be unidirec-tional and the travellers should be able to reach the platform without crossing the buses’ path. This can be ensured by using different levels for the centre platform and for the surrounding traffic. Then, the travellers can reach the platform by for example a tunnel or a bridge. An advantage with the centre platform design is that it can be a very safe alternative for travellers, since no crossing of the roadway is required. The safety perspective assumes that different levels are used. Another advantage is that it is easy to transfer since all the buses are located around the platform. A disadvantage with the design is that it can require a lot of space, depending on the size of the centre platform. If there is a lack of space at the terminal, the outside of the roadway can be used for more bus stops. This solution counteracts the safety benefit with this design since it could induce more travellers crossing the roadway.
2 PLANNING AND DESIGNING BUS TERMINALS
Figure 3: The centre platform design.
2.4.4 Angle berth
The design of the loading area makes the buses park with the front end facing the travellers waiting area when arriving to the bus stop at the terminal. The angle berth design, docking, is illustrated in Figure 4. A benefit with this design is that the travellers get a clear overview of the terminal and can wait inside for their bus. The design is also very safe, when used right. The buses need to gear out from their berth so it is of high importance that travellers do not cross the roadway behind the buses. The angle berth design is most suitable for bus lines having the terminal as their end destination or have layover time at the terminal. Otherwise, it might not be time well spent to drive in and out of the berth. The design is for this reason not recommended for throughgoing lines.
3 BUS TERMINAL CAPACITY
3
Bus terminal capacity
Capacity evaluation of bus terminals could easily be presumed as a well-developed research area. The interest and value in the field should not be a problem. Municipalities are eager to provide good public transport, but are often restrictive with usage of urban ground. Nevertheless, the available research and studies indicates a lack of knowledge in this problem area. There is much research about capacity for bus stops, but not for bus terminals. No specific recommended method for determining and evaluating the capacity for bus terminals seem to exist. There are, in general, two types of models that can be used for traffic analysis: analytical models and simulation models. These two types will be presented in more detail in Chapter 4. Microscopic traffic simulation. Both these types can also be used for bus terminal capacity.
3.1
Capacity definition
A first step when discussing how to determine the capacity of a bus terminal is to decide how to define and measure capacity. The term capacity is quite vague and can be defined in several different ways. Bus stop and bus terminal capacity (Al-Mudhaffar et al., 2016) is an article that focuses on how the capacity can be defined and determined both for bus stops and for bus terminals. For a single bus stop, the capacity can be defined as the maximum number of buses per berth per hour (buses/h). This definition is from a model presented in HCM2000 (The highway capacity manual) provided by the Transportation Research Board National Research Council (2000) in USA. HCM2000 presents three definitions of capacity: a general one, a definition for vehicle capacity and a definition for person capacity. Broadly speaking, the general definition is that the capacity of a facility is the maximum rate (per hour) which persons or vehicles can be expected to cross a point or a section of a roadway during a given time period. The definition for vehicle capacity is most interesting for this thesis however. Vehicle capacity is defined as:
Vehicle capacity is the maximum number of vehicles that can pass a given point during a specified period under prevailing roadway, traffic, and control conditions. This assumes that there is no influence from downstream traffic operation, such as the backing up of traffic into the analysis point. (Transportation Research Board National Research Council, 2000: page 2-2)
The definition above defines the capacity for a single bus stop (Al-Mudhaffar et al., 2016; Trans-portation Research Board National Research Council, 2000). When evaluating the capacity for bus terminals, the capacity of all the stops within the terminal need to be taken into consideration. One researcher that has contributed significantly to the research about bus stop capacity is Rodrigo Fern´andez. The article On the capacity of bus transit systems brings up that capacity need to be handled differently for transit stations or terminals than for isolated bus stops (Fernandez and Planzer, 2002). The authors argue that the capacity for a terminal can be defined as the numbers of vehicles that can be served, or the number or passengers that can be transferred. Again, the capacity can be defined either for the vehicles or for the passengers. Al-Mudhaffar et al. (2016) provides a definition for bus terminal capacity with focus on vehicles:
Bus terminal capacity can be defined as the total number of buses that can be served by the terminal per time unit (e.g. hour) at a given frequency ratio for each bus line. (Al-Mudhaffar et al., 2016: page 1770)
3.2
Determination of the capacity
As mentioned in the beginning of this chapter, no recommended method for determining bus terminal capacity seem to exist. Several different methods are discussed in research articles, but the focus are often more on capacity for single bus stops than on capacity for terminals. Mostly, the methods for determining bus stop capacity or bus terminal capacity are divided into analytical methods or simulation methods. This classification will therefore be used throughout this chapter. Analytical models for determining the capacity of bus stops are mentioned in several articles and simulation, the other method, is also brought up. One article that has some focus on capacity
3.2 Determination of the capacity
for bus terminals is Al-Mudhaffar et al. (2016). Two different methods for estimation of bus terminal capacity are presented. The methods are empirical analysis and simulation of bus terminal operations. The empirical analysis is performed by calculating the capacity for the individual bus stops and then adding those capacities together. Simulation of bus terminals can determine the terminal capacity in form of the highest bus flow that the terminal can handle before breakdown. 3.2.1 Using analytical methods
Fernandez and Planzer (2002) presents a way to determine the terminal capacity analytically. Using this method, the capacity can conceptually be expressed as shown in Equation (1). The transfer capacity is expressed in vehicles per time unit (e.g. buses/h), assuming each loading position only accepts one vehicle at a time.
Qterminal= α · N
t0
(1) where:
Qterminal = Transfer capacity of the terminal
N = Number of loading positions or berths α= Availability of the loading positions t0 = Occupancy time of each loading position.
The problem with this method is that the equation only describes the terminal capacity conceptu-ally. Both the availability and the occupancy of the loading position need to be calculated before they can be used in the formula. The availability of the loading positions can be expressed as the share of the time that the loading position is free. This share depends on several conditions, for example how the loading positions are allocated to vehicles and which queuing method that is used for entering and exiting a loading position. The occupancy time of the loading positions can be expressed as a function of the types of vehicles and passengers. Some types of vehicles require longer time to be accommodated and loaded than other vehicle types and passengers paying with cash need more time than passengers with bus passes. (Fernandez and Planzer, 2002)
Fernandez and Planzer (2002) discuss the problems with Equation (1) and comes with another way to classify the factors affecting terminal capacity. The different types of factors are presented below.
• Physical
Number and layout of the loading positions, facilities for loading and unloading and type of vehicles.
• Operational
Arrival of vehicles and passengers and use of loading positions. • Behavioural
Types of drivers and passengers.
3 BUS TERMINAL CAPACITY
After establishing that the analytical method in Equation (1) is not fully satisfying, Fernandez and Planzer (2002) mention the HCM2000 model for calculating capacity as an option, see Equation (2). The HCM2000 model is also presented both in Adhvaryu (2006) and Al-Mudhaffar et al. (2016). The model is an analytical model adjusted for calculating the capacity of single bus stops.
Qbus stop = 3600 ·g C tc+ g C · td+ Za· Cv· td (2) where:
Qbus stop = Maximum number of buses per bus stop per hour (buses/h)
g
C = Effective green time per signal cycle (1.0 for no signal at exit) tc = Clearance time between successive buses (s)
td = Average dwell time (s)
Cv = Coefficient of variation of dwell times = standard deviation/mean for td
Za = One tail normal variation corresponding to probability that queues will form behind a bus
stop
The probability that queues will be formed behind a bus stop can also be called failure rate and is derived using statistics. Za represents the area under one tail of the normal curve beyond the
acceptable levels of probability that a queue will form. A table with values for Za can be found in
the Highway Capacity Manual (Transportation Research Board National Research Council, 2000: page 27-12).
As for the model presented in Equation (1), there are several factors not covered in the HCM2000 model presented in Equation (2). For example, Equation (2) does not consider the time it might take for the bus to enter the terminal or the loading area. This can for example be time due to deceleration or turning movements. Another uncertainty is that the risk of queues behind the bus, Za, is calculated assuming normal distributed probabilities. (Al-Mudhaffar et al., 2016)
A main difference between the capacity calculation presented in Equation (1) and the one presented in Equation (2) is that the first equation is for bus terminals and the second for bus stops. The available articles in this field contains discussions whether the capacity of a bus terminal can be calculated based on the bus stops within the terminal or not. As mentioned in the beginning of Chapter 3.2. Determination of the capacity, the capacity of a bus terminal can be calculated as the sum of the capacity for the individual bus stops (Al-Mudhaffar et al., 2016). This approach would make the HCM2000 model in Equation (2) valid and useful for both bus stop capacity and bus terminal capacity.
Both Al-Mudhaffar et al. (2016) and Adhvaryu (2006) mentions that a summation of the individual capacities is not always a satisfying way to determine the capacity of the whole terminal. According to Al-Mudhaffar et al. (2016), the terminal capacity can be reduced at higher traffic loads due to factors such as queueing buses, blocked entrances or passengers moving across the terminal. When this is the case, the terminal capacity can be calculated as the sum of the capacity of the individual bus stops multiplied with a factor. The factor is defined as [1 - a reduction rate], where the reduction rate is based on the elements affecting the terminal capacity. Adhvaryu (2006) mentions two main reasons why estimation of bus terminal capacity differs from estimation of bus stop capacity. One reason is that the time needed to manoeuvring the bus within the terminal is not covered in methods developed for bus stops. Another reason is the delay that can emerge due to jaywalking passengers in the terminal area. This delay can be of various length depending on the terminal design. Adhvaryu (2006) ends up with two conclusions: either bus stop capacity models need to be modified in order to be able to handle these two factors or a simulation model for estimating bus terminal capacity is a better option. Al-Mudhaffar et al. (2016) suggests that the use of a factor can be a modification of the existing analytical capacity models.
3.2 Determination of the capacity
Analytical models are easy to use and are not so time-consuming compared to simulation models. However, analytical models are not preferable when analysing complex traffic situations, for exam-ple intersections with more advanced design (Trafikanalysforum, n.d.). The dynamic and stochastic parts of the traffic system can be captured better with simulation models while analytical models are better at making rough estimations of for example ”standard” intersections. According to Allstr¨om et al. (2008), traffic simulation is better than analytical models when it is desired to take a larger part of the traffic system into consideration. Traffic simulation is also better at vehicle actuated signal controls than analytical models.
3.2.2 Using simulation
Simulation can be a better option than analytical models for complex traffic situations. Since no recommended analytical method for calculating bus terminal capacity seem to exist, simulation might be a suitable solution. There is not so much research available about using simulation to determine bus terminal capacity. One article that mentions microsimulation models for bus terminal capacity is Adhvaryu (2006). Adhvaryu (2006) means that a problem with analytical models can be that the results can vary a lot due to gross values used in the formulas. Analytical models often include constants that can be difficult to estimate. When using simulation, observed individual values could be used instead of gross values which makes it easier to adapt the model to the specific context. In the article, the microsimulation model PASSION was used to calculate the bus terminal capacity. The output from the simulation model was used to determine the capacity. The capacity was determined using the bus flow (buses/h), the berth capacity (buses/h) and the saturation (%). Adhvaryu (2006) does not explain how the simulation output is used in order to calculate the capacity.
Al-Mudhaffar et al. (2016) also mentions microsimulation as a method to determine bus termi-nal capacity. One reason is that atermi-nalytical equations, for example Equation (2), does not always consider variance in arrival times. Not considering arrival distributions can cause capacity overes-timation since utilization over time is not included. Equation (1) considers the occupancy time of each loading position, t0, but has other drawbacks compared to simulation.
4 MICROSCOPIC TRAFFIC SIMULATION
4
Microscopic traffic simulation
When conducting traffic analysis there are two common models to use, analytical or simulation models. Analytical models can consist of queueing theory, optimization theory or differential equations (Olstam, 2005). Simulation models uses several sub-models for describing traffic states (Olstam and Tapani, 2004). Three classes of simulation models exist, microscopic, mesoscopic and macroscopic. The classification depends on the level of detail describing the traffic state. With its high level of details regarding traffic state, micro (microscopic) simulation provide the possibility to simulate individual vehicles and how they interact with other vehicles. To be able to describe and simulate individual vehicles micro simulation uses sub-models in the form of behaviour models. Micro simulation is often carried out on a limited area, for example a signalized intersection (Barcel´o et al., 2010). Traditionally it has been used for conducting evaluations of capacity, level-of-service or different design alternatives. Micro simulation can be used both in urban and rural areas. Besides the traditional area of use, Olstam (2005) brings up ITS (Intelligent Transport Systems) as a new age area at that time. Example of use are Intelligent Speed Adaptation and Adaptive Cruise Control systems.
4.1
Time vs. event based simulation
A simulation model can either be time based or event based. The differences between the two are how the simulation is triggered to move forward. The time based simulation evolves as the time progresses while event based simulation are modelled as a series of events. The type used in simulation depends to some extend on the chosen software for the simulation (Robinson, 2014). For time based simulation, the time is divided into time steps of 0.1-1 second. For each time step the model is computed and updated, if there are no changes an unnecessary computation has been done. The computation is based on the specified behaviour of the model, for each computation the animation of the simulation model is also updated. After each time step the simulation clock is increased with the time step and the models enter the next time step (Robinson, 2014; Olstam, 2005). A flowchart of the time based simulation is presented in Figure 5. Robinson (2014) brings up the difficulties to determining the time step. There is also possible to determine the time step based on how long time the activities in the model require. However, activities require different amount of time. The duration of each time step is suggested in Olstam (2005), who also says that micro simulation normally is time-discrete but that event based exist.
4.2 Behaviour models
According to Robinson (2014), in event-discrete simulation it is only the events that changes the system that are simulated. The time between events are not simulated and therefore the events happen directly after each other. An event can both be triggered by other events or by the simulation clock. For event based simulation different approaches exists for example, three-phase approach where events are classified either as conditional events or as bound or booked events.
4.2
Behaviour models
As mentioned before, micro traffic simulation enables simulation of individual behaviour and in-teraction between different traffic users and environments. The behaviour of drivers decides the interaction between vehicles (Barcel´o et al., 2010). For example, car-following, lane-changing and gap-acceptance models are used as behaviour models. These are considered the most important ones, but other models exist as well. Which models that actually are used in the simulation de-pends on its’ purpose. For simulation of bus terminals, car-following and gap-acceptance models are the primary used behaviour models. The different classes of behaviour models are used in different sub-models in the simulation. All sub-models handle specific tasks and behaviours in the simulation model (Olstam and Tapani, 2004). The simulation software also matters for which sub-models and behaviour models that are used.
4.2.1 Car-following model
Car-following models describe a vehicle’s interaction to a preceding vehicle in the same lane. Definition of following a vehicle is if the interaction, together with the desired speed, would lead to a collision (Olstam, 2005). The aim of car-following models is the same for all, describe and control the interaction of a following vehicle. However, several classes of car-following models exist with different ways to control the following vehicle. Even though car-following models have existed since the 1950s there is still ongoing research in the area according to Olstam and Tapani (2004). The authors also bring up that the perfect car-following model may not exist or have not been discovered yet.
4.2.2 Gap-acceptance model
The aim of gap-acceptance models is to determine if the gap is adequate for the driver to fit (Olstam, 2005). In some cases, gap-acceptance models can be seen as a module to lane-changing models since it can determine if a lane change is possible. Olstam (2005) brings up the safety aspect, where gap-acceptance models are used to make sure the lane change can be done safely. Apart from being used in lane-changing models, gap-acceptance models are used when vehicles need to yield for a larger traffic stream. The needed gap for a driver can be modelled to be individual.
According to Archer (2005) micro simulation uses a fixed value for critical-gap, which describes the minimum gap needed to enter another traffic stream. The critical-gap can either be in time or distance. The value can for example depend on the speed of the road or if a connecting traffic stream must yield or stop before entering. The critical value can be determined by observation or literature studies.
4 MICROSCOPIC TRAFFIC SIMULATION
4.3
Required data
Micro simulation is an individual simulation of movements being dynamic and stochastic (Dowling et al., 2002). Due to the details in micro simulation, a high level of data is needed in order to create a good model. The level of data may also depend on how accurate the model needs to be and what is being simulated. Obviously geometric data and geographic data is needed for building the model. California Department of Transportation in USA (United States of America) offers a guideline of required input data for traffic micro simulation. The required input data are presented in the list below. (Dowling et al., 2002)
• Geometry (lengths, lanes, curvature) • Controls (signal timing, signs)
• Existing Demands (turn volumes, OD matrix) • Calibration Data (performance data: speeds, queues) • Future Demands (turn volumes, OD matrix)
The geometry data can often be collected from building drawings of the studied area. The demand data for the current situation are best received from a counting station, manual of automatic. In the guideline from California Department of Transportation, a method called license plate survey is stated as the best way to obtain traffic volumes over the selected area. Check-points for all possible routes in the area are then established and all passing vehicles license number are registered. The method can consume a lot of resources but the traffic volumes will be accurate. For the calibration data to be useful, it should be collected at the same time as the traffic count is performed. The calibration data can consist of several different measures. The measures can be for example a measure of capacity or other measures of the system such as speed, travel time, queues and delays. Dowling et al. (2002) points out that a field observation can be useful for the calibration in order to detect issues that have not been considered. Estimations of future demands and travel patterns are best received from local authorities that handles traffic planning. (Dowling et al., 2002)
4.4
The simulation software Vissim
Vissim is a time-discrete and behaviour based software simulation tool for micro simulation. The software can be used for both rural and urban areas; it can also handle multimodal transport operations (PTV GROUP, 2014). It is often used to analyse and optimise traffic flows. (Barcel´o et al., 2010)
There are two options for how traffic is assigned to the network in Vissim, one is static assignment and the other one is dynamic assignment (PTV GROUP, 2014). A main difference between them is how the routes are allocated in the network. In the static assignment, the vehicles follow manually defined routes and the drivers has no choice which path to follow from origin to destination. The user must create both routes and vehicle inputs to use static assignment in Vissim.
In the user manual for Vissim 7 (PTV GROUP, 2014), it is stated that Vissim is based on a traffic flow model and light signal control. The traffic flow model includes car-following models and a lane-change model. There are two car-following models included in Vissim which are developed by Wiedemann. In 1974 Wiedemann developed a model suitable to use for models of urban traffic, the model is called Wiedemann ’74. There is also a model for freeway traffic without merging areas, it is named Wiedemann ’99 (Axelsson and Wilson, 2016; Barcel´o et al., 2010). The car-following model uses threshold values to determine when to change the behaviour of the driver. The behaviour can only be changed when such value is reached, the change can for example be the velocity or distance to the preceding vehicle (Olstam and Tapani, 2004). In Barcel´o et al. (2010) it is brought up that Vissim performs a lane selection which determines the desired lane to change to. Before changing lane, the gap to other vehicles need to be enough. The gap size depends on the speed of the vehicle wanting to change lane and the speed of the approaching vehicle in the lane to change to. Car-following models and gap-acceptance models were discussed in Chapter 4.2. Behaviour models.
4.5 Microsimulation of stops and terminals
Vissim is a time based software, but has an additional program called VisVap that can be used to regulate the light signal controls used in Vissim. By using signal controls, it is possible to create events in the network, one event at each signal head. Signal controls consists of a traffic-dependent control logic unit, which can be modelled by external programs like VisVap. The control unit can determine the signalling status for all signals in the next time step, thereafter sending the status back to the Vissim simulation (PTV GROUP, 2014). The signalling status can be determined by using detectors in the Vissim network. VisVap uses the programming language VAP (Vehicle-Actuated Programming), where it is possible to let signal controls be vehicle actuated. By placing out detectors before a signal head in Vissim, vehicles at the signal head can be detected and thereafter VisVAP can decide when to change the signal light.
4.5
Microsimulation of stops and terminals
Most of the existing micro traffic simulation models are developed focusing on car traffic (Fernandez et al., 2010). These models, oriented mainly around movements of cars, does not provide sufficient level of detail for behaviour between public transport vehicles and the surrounding traffic. Hence, the models have limitations when used to simulate public transport. Most commercial traffic simulation softwares like Vissim and Aimsun (TSS-Transport Simulation Systems, 2017) provides possibilities to simulate public transport by using for example embedded PT stops and PT lines. Simulation of public transport is not the main focus of these softwares and it often fails to simulate buses in a realistic way (Kramer, 2013).
Many time based models focuses on car-following models and car traffic. To capture behaviour at the terminal, event based simulation can contribute (Lindberg et al., 2017). Both Kramer (2013), (Lindberg et al., 2017) and (Adhvaryu, 2006) mentions that bus stop operations happen in parallel and that event based simulation might be a solution. In Fern´andez (2010), it is mentioned that simulation models developed for bus stops in the car network exists. One example of such model is PASSION, mentioned above. PASSION was developed with the aim to overcome the drawbacks with models focusing on car traffic. The model considers interactions between buses, passengers and traffic at bus stops.
Vissim is a time based software but can capture parallel operations by using VisVap. A combination of time based and event based simulation is necessary in this thesis due to the bus movements being triggered by time but controlled by previous and future events. By using both Vissim and VisVap, it is possible to make the simulation both time based and event based.
4.6
Verification, calibration and validation
It is almost impossible to make an exact imitation of a real system on a computer. Therefore, the simulation model needs to be verified, calibrated and validated to make sure that the results are accurate. Sometimes these parts can be time-consuming and must be done more than once. According to Olstam (2005), using traffic simulation for traffic analysis can be cost-efficient since it offers a safe way to experiment with traffic systems, both existing ones and systems that are under development. The experiments can for example regard design or different system alternatives. Since simulation models are imitations of real systems, it is of great importance to check the reliability of the model. Rakha et al. (1996) explains the difference between verification, calibration and validation. Validation checks that the model output correspond to the real system. Calibration is the process of adjusting the model until it produces output close enough to the desired values. Verification on the other hand, is not related to the real system. The aim of verifying a model is to ensure that the model logic works as desired and as specified by the user. Barcel´o et al. (2010) brings up the simulation time when calibrating and validating simulation models. A simulation runs over a predetermined time and therefore, it is assumed that the model and the modelled system evolves similar during the simulated time period. This assumption is analysed in the validation; the output of the model is compared to observed values of the real system during the same time period. The validation is carried out on a calibrated model with the purpose of checking the logic and behaviour of the model. Also, to make sure the model imitations are correct. If not, the calibration needs to be redone. (Barcel´o et al., 2010)
4 MICROSCOPIC TRAFFIC SIMULATION
As mentioned, before a validation can take place the model needs to be calibrated. The aim of the calibration is to fit the model to the data. In Treiber and Kesting (2013) validation is defined as following:
Validation is the process of determining the reliability of a model, i.e., the degree to which it is an accurate representation of the real world from the perspective of the intended uses. (Treiber and Kesting, 2013: page 333)
There exists more than one validation technique. The common factors with the techniques are: simulation of a model and its prediction and enable comparison with already available data. It is common to split the available data into training data for the calibration and validation data for the model prediction. To be able to perform a good validation, it is ideal that the two datasets represent identical situations. (Treiber and Kesting, 2013) Since the aim of validation is to compare results of the simulation with the real system, it is almost impossible if the system does not exist. (Kelton et al., 2015) B˚ang et al. (2014) also brings up that a validation only is possible if more than one dataset exists.
According to Rakha et al. (1996), there are two objectives when verifying a traffic model: ensure that the model logic provides expected outputs and verify that the outputs is consistent for a range of typical input values. The verification process can be seen as a five step procedure, where the different steps is performed in sequence. The steps regards choice of input parameters, checking consistency of input and output parameters and evaluate the output data and model logic.
5 FRAMEWORK FOR TERMINAL LOGIC
5
Framework for terminal logic
In this thesis, a method for evaluating bus terminals using micro simulation has been used. This method is presented in this chapter and is developed and provided by Sweco. The chapter explains how the terminal logic is applied. Operations for the logic, events, needs to be implemented in both Vissim and VisVap. The first subchapter presents an introduction to the terminal logic. The next two subchapters presents the operations for each program. The last subchapter focuses on implementation of the terminal logic in VisVap.
5.1
Preliminaries
Applying the terminal logic requires implementation both in Vissim and VisVap. The buses are considered passive in the terminal logic and are only sent between different queues in the network. The different queues and operations for the buses’ movements within the terminal are built in Vissim. Conditions for sending buses between the queues are checked and performed in VisVap. Hence, a network needs to be built in Vissim that uses the output from VisVap as input. The output file from Vissim is used as input file in the signal controller in Vissim. That is how Vissim and VisVap are connected, see Figure 6. The concept behind the terminal logic is general but needs to be adjusted to the specific terminal that is being studied, since all terminals are unique.
Figure 6: Illustration of how VisVap and VisVap are connected, with the different inputs and outputs.
Figure 6 shows the input and output data to both VisVap and Vissim. VisVap contains several files with the extension *.vv, one for each subroutine. Three files are input to Vissim, *.vap, *.pua and *.dll. The *.vap file is created when compiling VisVap, the other two are standard files.
5.2 Operations in Vissim
5.2
Operations in Vissim
To apply the terminal logic in Vissim, the desired operations for the terminal need to be built in the model. Since the Vissim model uses the output from VisVap as input, the model needs to enable this kind of input. In order to use the output as input the operations need to exist in both programs. The operations require the same functionalities but need to be implemented differently. Development of the Vissim model, for the terminal logic, is based on how bus stops can be created in Vissim besides the embedded function. All the driving spaces, including the bus stops, is represented by links. The links represents roadways or traffic lanes. Creating bus stops with links instead of the embedded bus stops is possible due to signal control. Using signal heads to force buses to stop at links representing bus stops creates the same behaviour as normal bus stops. This way, each bus stop represents a queue where a bus is held by a signal head. Detectors are used both before and after the signal heads in order to detect when to change signal state and let the buses leave the queue. All signal heads are set to red in the beginning of the simulation, leading to buses always stopping when arriving to a queue. Figure 7 illustrates how a queue is constructed.
Figure 7: The queue construction. The two blue squares are detectors. The first line, turquoise, ends routes while the last line, pink, assigns new routes. The middle line, red, is a signal head.
Buses are sent between the queues to represent the movements within the terminal. All the parts of a queue have a purpose for sending and receiving the buses. The first detector is used to detect when a bus has stopped. The detection of a vehicle starts checks in VisVap regarding the purpose of the stop and the release time from the stop. When the release time is reached, the signal head turns green and the bus departs. The detector after the signal head is used to detect when the bus has departed and the signal state is reset to red. In the figure, the first line is connected to the buses’ route and illustrates the end point of the previous route. The last line is a decision point where the bus is assigned to a new route. The queues have routes between them, meaning that when the bus is allocated a new queue it follows a predetermined route to that queue. Thus, most queues have two points connected to the routes. First one line where the previous route ends and then another line where the new routes starts.
Queues are used everywhere in the network where a decision is required (i.e. where the bus’ route can change). Hence, the buses can be described as packages being sent between the queues. As mentioned, the buses are considered passive throughout the simulation. Depending on the studied terminal, different operations are needed to represent the system. Operations used to represent the terminal are presented below.
• Generating buses
• Buses entering the terminal
• Holding buses at bus stops and layover area • Letting the buses exit the terminal
• Taking a lap within the terminal
Different operations are handled differently in VisVap. Each operation is presented below except buses exiting the terminal. That operation only ensure the buses can leave the terminal.
