A BRT Corridor Through Stockholm’s Inner-city: Assessing the Operational Impacts of a BRT Corridor Along Bus Line 4 Using Microscopic Simulation

STOCKHOLM SWEDEN 2020 ,

A BRT Corridor Through Stockholm’s Inner-city

Assessing the Operational Impacts of a BRT Corridor Along Bus Line 4 Using Microscopic Simulation

FRANCISCO CARON MALUCELLI

A BRT Corridor Through Stockholm’s Inner-city

Assessing the Operational Impacts of a BRT Corridor Along Bus Line 4 Using Microscopic Simulation

Francisco Caron Malucelli

Degree Project in Transport Science Master in Transport and Geoinformation Technology School of Architecture and the Built Environment KTH Royal Institute of Technology

Supervisor at KTH: Erik Jenelius Host Company: Sweco Society Supervisor at Host Company: Anton Holgersson Examiner: Albania Nissan, KTH

August 2020

Abstract

Bus Rapid Transit (BRT) corridors and systems have emerged in the past three decades as affordable solutions of medium capacity public transport services to highly urbanized areas, especially in Latin America and Asia. In Stockholm, trunk bus lines have gained priority over mixed traffic over the years through exclusive bus lanes, signal priority, and reliability control, for example, but no complete BRT solution has been implemented yet. Among the inner-city trunk lines, Line 4 is the most demanded with around 70,000 passengers boarding the service daily. This thesis proposes, then, to assess the operational impacts that BRT solutions as segregate median lanes, stations with off-board fare collection and platform level boarding and alighting through all bus doors, full signal priority and headway control strategy, would have in bus Line 4, using a microscopic simulation approach. Two scenarios were simulated, and the results compared to the existing conditions (Base Scenario).

Scenario 1 considered a 5-minute headway service and Scenario 2, 3-minute headways. Overall, the proposed scenarios reduce travel times by 37.6-49.1%, increase average operational speeds (including dwell times) by 60.4-96.6%, decrease dwell times by 57.9-65.6%, decrease delays by 18.4-36%, decrease vehicle occupancy rates by 3.5-44.9% and improve the Coefficient of Variation of the headways from 0.83-0.85 in the Base Scenario to 0.1 in Scenario 1 and 0.2 in Scenario 2. As a result of the reduction in travel times, a BRT service would need 13 buses to operate a 5-minute headway and 21 buses for a 3-minute headway, against 27 vehicles that are used currently for a 4 to 6-minute headway during peak hour.

Keywords: BRT, Bus Rapid Transit, microscopic simulation, Vissim, trunk buses, Line 4

Sammanfattning

Bus Rapid Transit (BRT) korridorer och system har dykt upp under de senaste tre decennierna som prismässigt överkomliga lösningar för kollektivtrafiktjänster med medelkapacitet till mycket urbaniserade områden, särskilt i Latinamerika och Asien. I Stockholm har busslinjer tagit prioritet över blandad trafik under åren, men ingen komplett BRT-lösning har ännu implementerats. Bland inner- stadens stombusslinjer har linje 4 högst efterfrågan med cirka 70 000 passagerare per dag. Den här avhandlingen syftar till att utvärdera de operativa effekterna BRT-lösningar som segregerade mittkörfält, stationer med off-board-biljettuppsamling och plattformsnivå påstigning och avstigning genom alla bussdörrar, full signalprioritet och headway-kontrollstrategi, skulle ha i busslinje 4, med hjälp av mikroskopisk simulering. Två scenarier simulerades, och resultaten jämfördes med de befintliga förhållandena (Basscenario). Scenario 1 bestod av ett headway på fem minuter medans scenario 2 studerade ett headway på 3 minuter. Sammantaget de föreslagna scenarierna minskar restider med 37,6-49,1%, den genomsnittliga drifthastigheten ökar (inklusive uppehållstider) med 60,4-96,6%, uppehållstiderna minskar med 57,9-65,6%, förseningarna minskar med 18,4-36%, fordonsbeläggningsgraden minskar med 3,5-44,9% och variationskoefficienten förbättra för huvudvägarna från 0,83-0,85 i Basscenario till 0,1 i Scenario 1 och 0,2 i Scenario 2. Som ett resultat av restidsminskningen skulle en BRT-tjänst behöva 13 bussar för att bedriva trafik med fem minuters headway och 21 bussar med tre minuters headway. Detta kan jämföras med de 27 fordon som krävs idag med ett headway 4–6 minuter under högtrafik.

Nyckelord: BRT, Bus Rapid Transit, mikroskopisk simulering, Vissim, stombussar, Linje 4

Acknowledgements

Quitting a stable job, selling your house, your car and getting rid of most of your belongings at age 39 to move to a new country after new experiences, knowledge and excitement was not an easy decision for our family. But it definitely was the right one. And this would not have been possible if it was not for, first of all, my beloved wife Priscila, my high school sweetheart, my lifetime partner and my better half, and our lovely daughter Sofia (my best, best friend!), both who make my days always happier and agreed to embark in this great journey into the unknown! I love you forever and ever!

I am extremely grateful to my parents Paulo and Sueli for been supportive of our decision and for always showing me the right way in life. I would like to thank my brother Andre and the rest of our family that always stood behind us. You all live in our hearts!

I am grateful for all the help I got from Professors Semida Silveira, Keiko Fonseca Ono and Ricardo Luders, that believed in me and invited me for a summer course at KTH in 2017. You planted the seed for this thesis!

I would like to thank my KTH supervisor, Professor Erik Jenelius for helping and guiding me through the process and Professor Albania Nissan for being a friend and for always incentivize me to keep going. Thanks, Bibbi!

When we moved to Stockholm I would never have imagined that a job opportunity in an important company as Sweco would appear so soon. And it was made possible because of my good friend Michael De Lange introduced me to Martin Holmstedt. I will always appreciate what you did for me Michael! Martin, getting to know you, being a part of your team at Sweco and having you as my mentor makes this transition feel easier than it is. I am very grateful for the opportunity you gave me. You are an amazing person!

I would like to thank my Sweco supervisor and colleague, Anton Holgersson, for helping not only with the thesis, but adapting in a new place and working environment. Thanks Anton! I appreciate all the help and support from all of my Sweco colleagues. What an honor to be part of this great company!

I am very thankful for the precious information and time given by Triin Reimal, from Trafikkontoret and Astrid Adelsköld, from Keolis.

I would like to tell my colleagues from KTH what a great, fun ride it was for an old guy like me to be a part of a group of young, smart and wonderful people like you! Martina, Nana, Zandile, Adele, Mina, Robby, Arefin, Georgios, Nik, Bruno, Dominik, Memo, Tianshu, Zhai, Axel, Ruben, Tina, Signe, Elvira, Elin. You are the best! Special thanks to Michael Sederlin for always challenging me, for the great talks and for being an inspiration.

Finally, I would like to thank the friends that stayed in Brazil but that are, and always will be, part of our lives. Amigos do River forever!

Stockholm, 17 ^th of September 2020

Table of Contents

1. Introduction ... 1

Background and Motivation ... 1

Objective, Research Question, Scope and Limitations ... 5

1.2.1 Objective and Research Question ... 5

1.2.2 Scope ... 6

1.2.3 General Limitations ... 6

2. Literature Review... 7

BRT Overview ... 7

BRT Definition and Components ... 9

2.2.1 Right of Way and Busway Alignment ... 12

2.2.2 Stations ... 15

2.2.3 Intersection Treatments ... 18

2.2.4 Advanced Public Transport Systems - ATPS ... 22

Previous Work done for Bus Line 4 ... 23

3. The Bus Line 4 ... 26

General description ... 26

Line 4 Operation Diagnosis ... 29

3.2.1 Travel Times ... 31

3.2.2 Average Speed ... 33

3.2.3 Dwell times at bus stops ... 35

3.2.4 Standstill time/stopped time ... 39

3.2.5 Passenger Load – Boarding and Alighting Passengers ... 42

3.2.6 Reliability – Headway Distribution ... 47

4. Methodology ... 52

Studied Scenarios ... 52

Modeling Approach ... 52

Proposed BRT Solutions and Model Inputs ... 53

4.3.1 BRT Right of Way, Busway Alignment and Line Route ... 53

4.3.2 BRT Stops/Stations ... 67

4.3.3 BRT Vehicle and Modeled Bus Lines ... 71

4.3.4 Passenger Demand ... 73

4.3.5 Traffic Signals and Signal Priority ... 76

4.3.6 Holding Strategy ... 81

4.3.7 Number of Replications, Warmup Time and Simulation Resolution ... 85

5. Simulation Results – Analysis and Scenario Comparisons ... 87

Travel Times ... 87

Average Speeds ... 90

Dwell Times at Stations/Stops ... 95

Standstill time/stopped time (delays) ... 99

Passenger Load – Vehicle Occupancy... 104

Reliability – Headway Distribution ... 110

6. Conclusion, Discussion and Future Work ... 114

Conclusions and Discussion ... 114

Future Research ... 117

7. References ... 119

8. Appendix ... 124

Passenger Demand Tables – Boarding Passengers ... 124

Passenger Demand Tables – Alighting Passengers... 126

BRT Station Configuration in Vissim ... 127

Lists of Figures

Figure 1. Daily number of passengers boarding in the trunk buses for both Regional (light blue) and Inner City (dark blue) lines for the years 2014, 2017, 2018 and 2019. (Stockholms läns landsting/Trafikverket, 2020). ... 4 Figure 2. Daily average number of boarding passengers per trunk bus line for the years 2014, 2017, 2018 and 2019 (Stockholms läns landsting/Trafikverket, 2020). ... 4 Figure 3. Evolution of the constructed length of BRT infrastructure through the years (BRT+ Centre of

Excellence and EMBARQ, 2020)... 8 Figure 4. Malmö Expressen vehicle/platform BRT operation (left) and the blue BRT bus to be used in the Akalla- Barkarby BRT service (right). Sources: (BIL Swede; Sveriges Bussföretag, 2020) and (Järfälla kommun, 2020).8 Figure 5. BRT standards in Sweden (Ringqvist, et al., 2015). ... 11 Figure 6. Criteria set of components for assessing a BRT corridor (Odbacke, 2018). ... 12 Figure 7. Examples of a full segregated, median busway in Curitiba (top left) (Prefeitura de Curitiba, 2014), exclusive bus lanes in the middle, without physical segregation in Malmö (top right) (Google Inc., 2020) and a curbside bus lane in Stockholm (bottom) (Stockholms stad, 2020). ... 13 Figure 8. Median busway delimited with road marking or low curb (Ringqvist, et al., 2015). ... 14 Figure 9. Suggestion of minimal bus and passenger volumes do implement each type of ROW according to TCQSM Exhibit 6-34 (Ryus, et al., 2013). ... 14 Figure 10. Relationship between stop/station distance, maximum permitted speeds and operational speeds (Nielsen, et al., 2005). ... 16 Figure 11. Schematic view of a narrow, elongated BRT station (top) (ITDP, 2017), and examples from Quito (bottom left) and Lanzhou (bottom right) (ITDP, 2016). ... 17 Figure 12. Examples of narrow general traffic lanes behind BRT stations with shared bicycle lane in Curitiba (left) and Nantes (right). Source: Google Street View, 2020. ... 18 Figure 13. Schematic view of the detection systems (left) and the detection positioning within the intersection (Wahlstedt, 2014)... 20 Figure 14. Exhibit 6-42 from the TCQSM showing a summary of benefits and impacts of TSP solutions in different North American cities (Ryus, et al., 2013). ... 21 Figure 15. The blue buses are a characteristic of trunk bus lines in Stockholm. Source:

https://sl.ifokus.se/articles/51bfa0f18e0e7403c200070d-4 (2019). ... 26

Figure 16. Current Bus Line 4 route, bus stops and distance between stops. Source: made by the author based

on (Stockholm Stad, 2020). ... 27

Figure 17. Existing bus lanes in the middle of the road (blue) and on the side of the road (red) along Line 4's

route (Holgersson, et al., 2019). ... 28

Figure 18. Timetable for Line 4 during weekdays (SL, 2020). ... 29

Figure 19. Composition of the total travel time, in percentage, for the traveling direction towards Radiohuset

(Holgersson, et al., 2019). ... 32

Figure 20. Composition of the total travel time, in percentage, for the traveling direction towards Gullmarsplan

(Holgersson, et al., 2019). ... 32

Figure 21. Average speeds for Line 4 during weekday PM peak including (left) and excluding (right) the dwell

times at the stops (Holgersson, et al., 2019). ... 33

Figure 22. Percentages of average speed goal compliance for Line 4 during PM peak, for both directions and

including and excluding the dwell times. ... 34

Figure 23. Distance between stops and segment compliance to the distance between stops goal (left) and

percentages of the total Line 4 route length that comply or not to the 500m goal (Holgersson, et al., 2019)... 35

Figure 24. Average dwell times at the bus stops of Line 4 in the direction to Gullmarsplan, at the PM peak

(Holgersson, et al., 2019). ... 36

Figure 25. Average dwell times at the bus stops of Line 4 in the direction to Radiohuset, at the PM peak (Sweco,

2019). ... 36

Figure 26. Categorized average dwell times at the bus stops of Line 4 in the direction to Gullmarsplan (Holgersson, et al., 2019). ... 37

Figure 27. Categorized average dwell times at the bus stops of Line 4 in the direction to Radiohuset (Holgersson, et al., 2019). ... 37

Figure 28. Example of shared stop/station in Kungsträdgården and how different services create “congestion” at the stop which could lead to major delays. Source: taken by the author, 2019. ... 38

Figure 29. Heatmaps of the recorded standstill times for both directions of Line 4, for a regular Kernel density surface (left) and for a Kernel density surface weighted by the standstill times (right) (Holgersson, et al., 2019). ... 40

Figure 30. Average standstill time for each segment of Line 4 route during the PM peak, in the direction towards Gullmarsplan (Holgersson, et al., 2019)... 41

Figure 31. Average standstill time for each segment of Line 4 route during the PM peak, in the direction towards Radiohuset (Sweco, 2019). ... 41

Figure 32. Histograms of the standstill data for both directions of Line 4 during the PM peak, with the cumulative percentage frequencies. ... 42

Figure 33. PM peak average of total boarding, alighting and load of passengers towards Gullmarsplan (Holgersson, et al., 2019). ... 43

Figure 34. PM peak average of total boarding, alighting and load of passengers towards Radiohuset (Sweco, 2019). ... 43

Figure 35. Average vehicle occupancy rate at stops towards Radiohuset (Holgersson, et al., 2019). ... 44

Figure 36. Average vehicle occupancy rate at stops towards Gullmarsplan (Holgersson, et al., 2019)... 45

Figure 37. Average number of passengers (boarding + alighting) at each station, per vehicle, in the direction to Gullmarsplan, during PM peak, with the major transportation nodes highlighted in green. ... 46

Figure 38. Average number of passengers (boarding + alighting) at each station in the direction to Radiohuset, during PM peak, with the major transportation nodes highlighted in green. ... 46

Figure 39. Linear regression of passenger demand for PM peak and dwell times at each station, for both traffic directions ... 47

Figure 40. Boxplot of the headway data for Line for during PM peak. ... 48

Figure 41. Boxplot of the headway data for Line for during PM peak, without the extreme outliers. ... 49

Figure 42. Histograms of the headways for both directions of Line 4 during PM peak (no extreme outliers). ... 50

Figure 43. Existing road widths along Line 4 route. Source: made by the author based on (Stockholm Stad, 2020). ... 54

Figure 44. Example images of Rosenlundsgatan with Stockholms Södra on the left-hand side (left) and Ringvägen (right). Source: Google Street View, 2020. ... 55

Figure 45. Proposed design for Odenplan area including the BRT busway and the BRT Odenplan Station (top) and the Vissim network for the same area (bottom). ... 56

Figure 46. Proposed route for a full BRT operation of bus Line 4. Source: made by the author based on (Stockholm Stad, 2020) ... 57

Figure 47. Schematic cross section of a 24m wide street, no parking allowed. Source: done with (Streetmix, 2020). ... 58

Figure 48. Schematic cross section of a 24m wide street with a curbside loading bay in one of sides of the road. Source: done with (Streetmix, 2020)... 58

Figure 49. Schematic cross section of a 30m wide street, no parking allowed. Source: done with (Streetmix,

2020). ... 59

Figure 50. Schematic cross section of a 30m wide street with a curbside loading bay in one of sides of the road.

Source: done with (Streetmix, 2020)... 59 Figure 51. Schematic cross section of a 30m wide street with a BRT station in the middle and prohibited parking on both sides of the road. Source: done with (Streetmix, 2020). ... 59 Figure 52. Stockholm's road hierarchy according to the NVDB. Source: made by the author based on

(Stockholm Stad, 2020). ... 60 Figure 53. BRT Line 4 corridor as coded in Vissim (in green) with the proposed intersections, for the segment between Radiohuset and Odengatan, ... 61 Figure 54. BRT Line 4 corridor as coded in Vissim (in green) with the proposed intersections, for the segment of Odengatan between Valhallavägen and Sankt Eriksplan, ... 61 Figure 55. BRT Line 4 corridor as coded in Vissim (in green) with the proposed intersections, for the segment between Sankt Eriksplan and Västerbroplan, ... 62 Figure 56. BRT Line 4 corridor as coded in Vissim (in green) with the proposed intersections, for the segment of Hornsgatan between Västerbron/Långholmsgatan and Ringvägen, ... 62 Figure 57. BRT Line 4 corridor as coded in Vissim (in green) with the proposed intersections, for the segment of Ringvägen between Hornsgatan and Götgatan, ... 62 Figure 58. The existing bridge structure in Skanstullbron and its connection to Gullmarsplan, including the end of the bus lane (left) and the single lane that goes up to the roundabout (right). Source: Google Street View (2020). ... 63 Figure 59. BRT Line 4 corridor as coded in Vissim (in green) with the proposed intersections, for the segment between Hornsgatan and Gullmarsplan, ... 64 Figure 60. Example of a desired speed distribution coded in the model. ... 65 Figure 61. Speed limits set to the BRT corridor in the simulation model. Source: made by the author based on (Stockholm Stad, 2020). ... 66 Figure 62. Location of the existing Line 4 bus stops to be reallocated or removed, the proposed BRT stations and the average distances between stations. Source: made by the author based on (Stockholm Stad, 2020). 67 Figure 63. Walking distance from the proposed BRT station and Stockholms Södra Train Station. Source:

Goolgle Maps (2020). ... 68

Figure 64. Schematic design of the typical BRT station for a 30m wide street. ... 69

Figure 65. Examples of a center station with a split operation in Lanzhou (ITDP, 2016) and of elevated parallel

side streets around the BRT station in Nantes (Finn, et al., 2011). ... 69

Figure 66. A screenshot from the simulation showing two buses serving passengers at the same time in one of

the BRT stations. ... 70

Figure 67. Boarding delay function implemented in the model, based on the recommendation from TCQSM

(Ryus, et al., 2013)... 71

Figure 68. Characteristics of the articulated bus model used in Vissim (PTV, 2020). ... 72

Figure 69. Schematic map of all trunk bus lines that share the BRT corridor with Line 4. Source: made by the

author based on (Stockholm Stad, 2020)... 73

Figure 70. Redistribution of the existing passenger demand into the new proposed BRT stations. ... 74

Figure 71. Schematic view of VisWalk pedestrian simulation objects in a typical BRT station of the model... 75

Figure 72. Location and type of the traffic signal controllers along the BRT corridor in the model. Source: made

by the author based on (Stockholm Stad, 2020). ... 78

Figure 73. Schematic view of the detector position within the simulation model in Vissim... 79

Figure 74. Flowchart for the VAP logic of a 3-stage signal with BRT priority, where stages 1 and 2 are regular

signal stages and stage 3 is the exclusive BRT stage. Source: made by the author in Lucidchart (Lucid Software

Inc., 2020)... 80

Figure 75. Space-time diagram showing the simulation second against the average total traveled distance for each one of the buses in the simulation. ... 81 Figure 76. Detector and signal head scheme in Vissim for the holding strategy at a BRT station... 82 Figure 77. Flowchart of the VAP logic used for the holding strategy signals in the model. Source: made by the author in Lucidchart (Lucid Software Inc., 2020). ... 83 Figure 78. Simulation parameters used in Vissim for this study... 86 Figure 79. Comparison between the average travel time for the Base Scenario and for the proposed Scenarios 1 and 2, with average total travel times (left) and the relative difference of Scenarios 1 and 2 to the Base Scenario (right). ... 88 Figure 80. Composition of the total travel time, in percentage, for Scenario 1 (left) and Scenario 2 (right). ... 88 Figure 81. Comparison of the total travel time composition between the Base Scenario and Scenarios 1 and 2.

... 89 Figure 82. Average speed comparison between the Base Scenario and the simulated Scenarios 1 and 2 (left) and the relative difference of average travel times between Scenarios 1 and 2 and the Base Scenario (right). . 91 Figure 83. Average speeds for Line 4 including the dwell times at stations for the Base Scenario (top), Scenario 1 (bottom left) and Scenario 2 (bottom right)... 93 Figure 84. Segment of the BRT Line 4 between Fleminggatan and Fridhemsplan. ... 94 Figure 85. Percentages of average speed goal compliance for Line 4 during PM peak, for both directions and including and excluding the dwell times. ... 95 Figure 86. Screenshot of an instant of the simulation where it is possible to observe passengers boarding and alighting through all doors. ... 96 Figure 87. Average dwell times per BRT station for Scenario 1 and Scenario 2 in the traffic direction towards Radiohuset. ... 97 Figure 88. Average dwell times per BRT station for Scenario 1 and Scenario 2 in the traffic direction towards Gullmarsplan. ... 97 Figure 89. Relative decrease between the average dwell times at bus stops between the Base Scenario and Scenarios 1 and 2 for the direction towards Radiohuset. ... 98 Figure 90. Relative decrease between the average dwell times at bus stops between the Base Scenario and Scenarios 1 and 2 for the direction towards Gullmarsplan. ... 99 Figure 91. Heatmaps based on Kernel density surfaces showing the registered locations with higher

concentration of stops (speeds under 3km/h) for Scenario 1 (left) and Scenario 2 (right). ... 100 Figure 92. Relative delays measured in a link level for Scenario 1 (left) and Scenario 2 (right). ... 101 Figure 93. Detailed view of the relative delays around Västerbroplan Station for Scenario 1 (left) and for the segment between Fleminggatan and Fridhemsplan for Scenario 1 (right). ... 102 Figure 94. Average holding time per station for both traffic directions... 102 Figure 95. Average signal delays per segment, presented as the sum of both directions of Line 4. ... 103 Figure 96. Average vehicle occupancy rate variation during the analysis period for Scenario 1 and Scenario 2.

... 105

Figure 97. Average vehicle occupancy rate per station for the direction towards Radiohuset. ... 106

Figure 98. Average vehicle occupancy rate per station for the direction towards Gullmarsplan. ... 107

Figure 99. Average vehicle occupancy rate for each segment of the BRT network for Scenario 1 (left) and

Scenario 2 (right). ... 107

Figure 100. Vehicle occupancy rate histograms, with the respective cumulative frequency curves, for Scenario 1

(left) and Scenario 2 (right). ... 109

Figure 101. Boxplot for the measured headways in Scenario 1 and 2. ... 110

Figure 103. Vehicle trajectories for Scenario 1 (left) and Scenario 2 (right). ... 112

Figure 104. Example of BRT stations on the right-hand side of the busways... 117

Lists of Tables

Table 1. Data type, sources and collection period (Holgersson, et al., 2019). ... 30

Table 2. Average total travel times for Line 4, for both directions in different time periods (Holgersson, et al., 2019). ... 31

Table 3. Average speeds for Line 4 during the PM peak hour, both including and excluding the dwell times at bus stops (Holgersson, et al., 2019). ... 33

Table 4. Vehicle occupancy rates for Line 4 (Holgersson, et al., 2019). ... 44

Table 5. Summary statistics for the headway data of Line 4 during PM peak. ... 48

Table 6. Summary statistics for the headway data of Line 4 during PM peak, without the extreme outliers. ... 49

Table 7. Measure of reliability of the transit service based on the headway adherence (Ryus, et al., 2013). ... 51

Table 8. Travel time calibration results for the segment between Skanstull and Gullmarsplan. ... 65

Table 9. Social Force model parameters used in the simulations. ... 75

Table 10. Composition of pedestrian types and desired speeds used in the simulation. ... 76

Table 11. Descriptive statistics for the measured headways in the preliminary simulation. ... 81

Table 12. Summary of the sensitivity analysis results for Scenario 1. ... 84

Table 13. Summary of the sensitivity analysis results for Scenario 2. ... 84

Table 14. Calculation of the necessary number of simulation replications. ... 85

Table 15. Travel time results from the simulation for Scenarios 1 and 2. ... 87

Table 16. Average speeds for both directions of Line 4 and both simulated scenarios. ... 90

Table 17. Average speeds per direction, including the dwell times, with measures of dispersion. ... 91

Table 18. Average dwell times per station, for each traveling direction and simulated scenario. ... 96

Table 19. Summarized values of average standstill (delay) times for all studied scenarios, with the 99% confidence interval between parenthesis. ... 104

Table 20. Summary statistics for the vehicle occupancy rates... 106

Table 21. Summary of vehicle crowding compliance rates of Scenarios 1 and 2 to the standards defined by Trafikförvaltningen (2014). ... 109

Table 22. Descriptive statistics for the measured headways in Scenarios 1 and 2, in minutes. ... 110

Table 23. Measure of reliability of the transit service based on the headway adherence (Ryus, et al., 2013). . 113

Table 24. Summary and comparison of the aggregated values for the performance indicators from the Base

Scenario and the simulated Scenarios 1 and 2. ... 115

1. Introduction

Background and Motivation

Stockholm is a fast growing city which has a population of 975,904 in the inner city and 2,383,269 in the so-called Stockholm Region (SCB, 2020), and it is forecasted to grow 25% until 2030 (Firth, 2012).

Stockholm holds a work force of around 570,000 people of which 54% live within the limits of Stockholm City and 22% live in one of the ten adjacent municipalities (Järfälla, Sollentuna, Sundbyberg, Solna, Danderyd, Lidingö, Nacka, Tyresö, Huddinge and Ekerö). From those who must commute to work every day, 16% live in and commute from one of the county’s 15 municipalities (not Stockholm city), while only 8% commute daily from another county (Firth, 2012).

To accommodate this growth and intense daily movement, both urban development and transportation systems planning are well documented in a series of local and regional strategic plans and policies publications. In these documents, the public transit network plays a major role for the achievement of a sustainable, climate friendly society in the near future.

On a local level, the City of Stockholm published in 2007 the results of a project that was a long-term vision for a sustainable growth and development of the city, titled Vision Stockholm 2030 (Stockholms stad, 2007). By the time this thesis is being written, an updated document was released in June/2020, called Vision Stockholm 2040, but it was not used as reference for this work. As a part of this vision, the City of Stockholm’s traffic department (trafikkontoret) developed the Urban Mobility Strategy to help the city to achieve the goals set in Vision 2030. In the overall strategy of the Mobility Plan, one of the set points is the expansion of the public transport infrastructure including, amongst several other projects, the “conversion of Bus Rapid Transit line No. 4 to a tramway” (Firth, 2012). In the document, four main aims are set and each of these aims has specific objectives:

• Aim A refers to increasing transportation capacity,

• Aim B to promote transportation accessibility,

• Aim C focus on attractiveness and

• Aim D on sustainability.

Four of the fourteen total objectives in the aims regard specific improvements of the public transport system: transport capacity, public transport, journey time reliability and bus speeds (Firth, 2012).

Objectives B1 and B2 are the most interesting for this degree project as they focus on the importance of travel time reliability and transit operational speeds into attracting more users to the public transport system.

The main target set by objectives B1 and B2 is a minimum average speed, including dwell times at

stops, of 20km/h in the inner city by 2030. To reach this target, public transit needs dedicated space

and right-of-way (ROW) with traffic signal priority at intersections, a minimum of 500 meters between consecutive stops and shorter boarding/alighting times at these stops. However, giving this type of priority to faster public transit systems will have its impacts on the other, lower prioritized modes (Firth, 2012). And here, it is important to directly quote Stockholm’s Urban Mobility Strategy: “It must be possible to motivate the cost for other road users with the benefits generated for rapid transit passengers and the urban environment. If dedicated lanes were provided for the entire BRT and tram network in Stockholm City, this would correspond to approximately 3-4 per cent of the entire road network. A very preliminary assessment is that space could be dedicated to the rapid transit network in the inner city without any major impact on speeds in the regional road network, but with a higher capacity for movement in the local street network” (Firth, 2012).

On a regional level, the rapid transit services have a high degree of importance in the public transport network and are included with certain priority in the planning documents and policies. The rapid transit service networks are referred to as Stomnätet in Swedish, or Trunk Network in English, and include rail traffic and the blue buses traffic (Stjärnström, et al., 2014). It is of high importance to Region Stockholm, which is the organization responsible for planning and implementing public transportation infrastructure in the greater Stockholm, that the trunk buses service be characterized by high frequency, speed, clarity and reliability (Region Stockholm, 2020).

A cooperation work led by Region Stockholm and the Swedish Transport Administration (Trafikverket), and closely followed by the concerned municipalities, called “Green light to trunk buses” (Grönt Ljus stombuss) investigated how to improve accessibility for the blue buses in the trunk network by 2030.

“To give public transport priority over other vehicle traffic is not just about creating a fast and reliable bus service. It is also about increasing the capacity of streets and roads with heavy pedestrian traffic.

Accessibility is often limited when the buses have to share the space with the cars. Measures that prioritize public transport are necessary to make the trunk bus network attractive” (Stockholms läns landsting/Trafikverket, 2016). Among the investigated measures to improve trunk buses service, the most important for this work are the following (which can be combined for the best effect):

• Segregated bus lanes - provide faster, more reliable and shorter travel routes than car traffic.

Bus lanes are more efficient where there is constant congestion by providing faster movements through intersections;

• Signal prioritization – 70% to 80% of all delays in bus services are due to waiting times at traffic signals. Giving green light faster can reduce buses’ delay by 10 to 20%;

• Bus stop design has an effect on the speed - access to the stop should be easy, straight and, ideally, without a special bay or pocket. it should be simple, straight and without a special pocket;

• Better boarding and alighting – Dwell times at stops have a great impact on total bus travel time.

door, which takes time. Thus, improved ticketing and boarding/alighting processes will improve total travel times (Stockholms läns landsting/Trafikverket, 2016).

Since 2012, Trafikförvaltningen, which is the traffic management department within Region Stockholm (regional level), has been working in a partnership with the City of Stockholm in a joint action plan to improve the accessibility of trunk buses in Stockholm, called Trunk Network Plan (Stomnätsplan). The work has been done for trunk buses that serve the inner parts of the city and for the lines that serve the county area. The measures contained in the plan for the lines that serve the inner part of Stockholm, including bus Line 4, the subject of this thesis, aim to reduce the travel times, reduce the variability in the travel times thus increasing predictability, lower the passenger waiting times by improving regularity and increase the general public transport attractiveness (Region Stockholm, 2020). The plan also defines the three main principles for planning trunk services: good regional availability, attractive and competitive public transport and integrated planning for an attractive urban environment. The second principle includes the following goals:

• Average operational speeds, including dwell times, of at least 20km/h;

• Trunk buses are prioritized in some segments by receiving their own segregated spaces, at the expense of car traffic;

• Minimum distance between stops of 500 meters;

• Frequency between 2 and 7.5 minutes;

• A minimum of 500 passengers on the mostly demanded segment during peak hours (Stjärnström, et al., 2014).

It is quite clear from Stockholm’s urban development and public transit strategic documents that trunk buses have a much higher priority than private vehicles in the city’s transport network. According to the data published the Trunk Busses Annual Report (Stombussbokslut) from 2019, bus Line 4 has the highest ridership of all trunk lines, with a little more than 70,000 boarding passengers per day, which represents approximately 20% of the total daily trips from all trunk buses and 40% when only the inner city lines is considered (Stockholms läns landsting/Trafikverket, 2020).

Figure 1 shows the daily total number of passengers boarding both the Regional and the Inner-City

trunk bus lines for the years of 2014, 2017, 2018 and 2019. The figure shows that the number of

passengers has been rising through the years and that the split between Regional and Inner-City

passengers is also increasing in the latest years.

Figure 1. Daily number of passengers boarding in the trunk buses for both Regional (light blue) and Inner City (dark blue) lines for the years 2014, 2017, 2018 and 2019. (Stockholms läns landsting/Trafikverket, 2020).

Figure 2 shows the average number of passengers boarding each of the trunk lines per day, being lines 1 to 6 the inner-city trunk lines and the remainder the trunk lines that connect Stockholm with the metropolitan region. From Figure 2 it is possible to observe that Line 4 has the highest demand of all the trunk lines and that its demand is approximately the double of lines 1, 2 and 3, which are the other inner-city trunk bus lines. It can be also observed that the number of passengers using Line 4 increased considerably between 2018 and 2019, despite being stagnant in the previous period.

Figure 2. Daily average number of boarding passengers per trunk bus line for the years 2014, 2017, 2018 and 2019 (Stockholms läns landsting/Trafikverket, 2020).

Despite its clear importance to Stockholm’s public transport system, Line 4’s operational performance

is below the goals set by Trafikförvaltningen and Stockholms stad and, therefore, has been the subject

of many studies, tests and action plans regarding possible improvements. However, none of it

investigated the full quantitative impacts of a combination of the measures suggested/defined in the

strategic documents for the trunk services, hence the motivation for the present degree project.

Besides the strategical and technical aspects that motivates this work, there is also a personal motivation. The author was born and raised in Curitiba, Brazil, which was one of the berths of Bus Rapid Transit (BRT) systems as they are seen today and has a tendency of looking into public transportation bus services through those lenses. Therefore, testing similar solutions that exist in Curitiba since the 1980’s, and that were proven to be successful, into Stockholm’s most important bus line, and investigate how those solutions can improve the bus services seems natural to the author.

Objective, Research Question, Scope and Limitations

This section presents the objectives of the thesis and the respective research question, the spatial and temporal scopes, as well as the general limitations for the work.

1.2.1 Objective and Research Question

Considering the background and motivation presented in the previous chapter, the main objectives of this thesis are the following:

• Propose BRT solutions to Stockholm’s Bus Line 4, including:

o dedicated/segregated right of way – ROW;

o BRT stations that allow for off-board fare collection and level boarding/alighting;

o boarding and alighting through all bus doors and o full signal priority.

• Assess the impacts (positive and negative) of the proposed measures to the operation of Stockholm’s Bus Line 4 specifically, by comparing different indicators of service performance between the simulation and data collect from the field, including:

o Average travel time and speed;

o dwell times at bus stops;

o delay, or standstill time/stopped time;

o vehicle crowding (occupancy) and;

o Reliability.

Thus, the research question to be answered is: Does the implementation of a full BRT corridor

along bus Line 4 route improves the operational performance of the service compared to the

existing conditions and, in case it does, by how much?

1.2.2 Scope

In order to have a more comprehensive and detailed analysis of the operational impacts that full BRT measures can bring to an existing bus line that runs through the central area of a large city, the spatial focus of the study is the whole route of Stockholm’s bus Line 4, between Gullmarsplan and Radiohuset.

Field data collected from Line 4 (as it will be presented later) shows that the service has operational problems in both morning and afternoon peak periods, mostly due to traffic congestion. Ideally, the proposed measures should be tested for both peak periods and for the off peaks as well, so a more comprehensive understanding of the impacts would be achieved. However, due to the effort needed to model each time period and the time limitations of the degree project itself, it was decided to select only one of the peak periods to be analyzed. The temporal scope choice was based on the operational indicator that has the most impact on the service and that can be easiest interpreted: the travel time.

Data shows that, on average, Line 4 buses take around 107 minutes to travel a full round trip during the AM peak, 104 minutes during low traffic (off peak) periods and 114 minutes during the PM peak.

Therefore, it was defined that the time scope of the study will be restricted to the PM peak period, which is defined by the data to be between 15h00 and 18h00.

1.2.3 General Limitations

It is very important to highlight that the main objective of this thesis is to test a BRT corridor operation of bus Line 4 and to assess the impacts only on that specific bus line service, and not on general traffic or other bus services. Given the existing space availability on the streets where Line 4 route is located and the suggested implementation of an exclusive, segregated BRT corridor along it, the remaining space for general traffic lanes will be restricted and negative effects would be expected regarding congestion. However, considering the extensive scope of the simulation model, the time constrains to develop the research and the stronger interest in the specific benefits of the measures for trunk bus lines, the main limitation of this thesis is to not include the general traffic in the simulation model and, therefore, not analyze how it would be affected. Most importantly, this includes the regular bus services, or the red buses, that would not be able to use the BRT corridor infrastructure and would have to share the general traffic lanes with the other vehicles. In a nutshell, the analysis concerns only the methods and actions to have the best service as possible for Bus Line 4 in order to achieve the operational goals set by public policies and strategic documents.

There are other specific assumptions and limitations regarding the model and its components that are

described in the appropriate chapters, making it easier for the readers to follow.

2. Literature Review

BRT Overview

The concept of providing a better and faster public transport in growing cities by giving priority and improving the management and operation of bus services is present in plans and studies in North America since the 1930’s and in implementation in Europe in the 1970’s. Eventually, the differences in the city structures and development and in the general culture would lead to the two largely used acronyms in the transportation field: Bus Rapid Transit (BRT) and Buses with High Level of Service (BHLS). The former is a term first used in the United States of America in 1966 and was an evolution of improved bus systems that date from the early 1930’s, and the latter is the European version of high- quality bus services that started being implemented in the 1970’s through bus lanes, bus-only roads, traffic management, priority for buses at traffic signals (Finn, et al., 2011).

While the precursors of BRT emerged in the USA as bus lanes, or busways, on freeways that later were converted to High Occupancy Vehicles (HOV) lanes (Finn, et al., 2011), in South America, most specifically in the city of Curitiba, a combination of North American and European concepts gave birth to the first full-featured, well-known BRT system in 1982 (later improved in 1991) (Hidalgo, 2013). The easy, low-cost and fast implementation of a BRT corridor, together with its high performance in moving people through urban areas, brought the system a high degree of popularity among urban planners and it quickly spread to other location in Latin America and Asia, and later to the USA and Europe (Hidalgo, 2013).

Today, there is approximately 5,196km of BRT corridors implemented in 173 cities around the world, transporting close to 34,026,459 passengers per day (BRT+ Centre of Excellence and EMBARQ, 2020).

From the total implemented length, 35.2% is located in Latin America, 31.26% in Asia, 16.8% in Europe,

12.06% in North America, 2.52% in Africa and 2.09% in Oceania. Although Latin America and Asia

have similar BRT extensions, the systems in Latin America carry over 21 million passengers per day,

which represents 62% of the total, while in Asia 9.5 million passengers are served daily by BRT

systems. Nevertheless, it is a transit system that has had a significant growth in the past two decades

and became the backbone of many developing countries public transportation systems. Figure 3 was

extracted from a website called Global BRT Data which provides extensive information about BRT

system in the world. The graph shows the evolution through the years of the constructed length of BRT

networks around the globe between 1966 and 2020. It is possible to see that from the 1990’s to the

current date the growth of constructed BRT infrastructure is very prominent, specially between 2000

and 2010. Today, according to the data, there are over 5,000 kilometers of BRT infrastructure

implemented in the world.

Figure 3. Evolution of the constructed length of BRT infrastructure through the years (BRT+ Centre of Excellence and EMBARQ, 2020).

In Sweden, the trunk bus lines that operate in Stockholm, Malmö, Gothenburg, Karlstad and Jönköping, for example, have some components of a BRT system but are far from the large corridors and systems in South America and Asia. However, it is certain that BRT meets the same needs in Sweden as in many countries, indicating the urge to do feasibility studies. (Kottenhoff, et al., 2009).

Malmö has one bus line operating what is called the Malmö Express, using vehicles with BRT like identity and median bus lanes in parts of the line (BIL Swede; Sveriges Bussföretag, 2020). In the Stockholm Region, the first full BRT service is planned to start running in August of 2020, connecting the public transport terminal Akalla to Barkarby Staden with exclusive, segregated busways in the middle of the road, signal priority and state of the art vehicles (Järfälla kommun, 2020). Figure 4 shows examples of the Malmö Express and the full electrical bus that will operate the line between Akalla and Barkarby Stad.

Figure 4. Malmö Expressen vehicle/platform BRT operation (left) and the blue BRT bus to be used in the Akalla-Barkarby

BRT service (right). Sources: (BIL Swede; Sveriges Bussföretag, 2020) and (Järfälla kommun, 2020).

BRT Definition and Components

There is not a single, precise definition of BRT and several good descriptions can be found in the literature. “Bus Rapid Transit (BRT) a bus-based rapid transit system that can achieve high capacity and speed at relatively low cost by combining segregated bus lanes that are typically median aligned, off-board fare collection, level boarding, bus priority at intersections, and other quality-of-service elements (such as information technology and strong branding)” (ITDP, 2017). BRT has also been defined by the North American Federal Transit Administration (FTA) as a “rapid mode of transportation that can provide the quality of rail transit and the flexibility of buses.” The Transit Cooperative Research Program (TCRP) Report 90 expands this definition to “a rubber-tired form of rapid transit that combines stations, vehicles, services, running ways, and Intelligent Transportation Systems (ITS) elements into an integrated system with a strong image and identity” (Danaher, et al., 2007).

There are standards in the literature by which a bus service or bus corridor can be classified or not as a BRT and it is dependable, mainly, on the components that constitute the system. The Institute for Transportation and Development (ITDP) publicizes a method for classifying BRT corridors based on a punctuation system called “The BRT Standard”. A BRT corridor is defined as “a section of road or contiguous roads served by a bus route or multiple bus routes with a minimum length of 3 kilometers (1.9 miles) that has dedicated bus lanes” (ITDP, 2016). The classification is done according to five main categories and the points are attributed for several subcategories within each main category. The main categories are BRT Basics, Service Planning, Infrastructure, Stations and Access and Integration. BRT Basics sets the five essential elements that differentiates a BRT from a normal bus service, which are as follows (ITDP, 2016):

• Dedicated right-of-way (ROW);

• Busway alignment;

• Off-board fare collection;

• Intersection treatments and

• Platform-level boarding.

Another approach to BRT classification is proposed by a project called COST – European Cooperation

in Science and Technology in its report “Buses with High Level of Service – Fundamental

Characteristics and recommendations for decision-making and research” (Finn, et al., 2011). Although

COST is focused on the European BHLS, it presents a BRT classification method from (Gray, et al.,

2006) that categorizes into BRT-Lite, BRT-Heavy and Full-BRT. While BRT-Lite is characterized by

faster speeds than a regular bus service, usually achieved by greater spacing between stops, Full-BRT

encompasses systems with grade-separated ways, off-board fare collection, high frequency services

and modern vehicles, with performance similar to metro rail services. The BRT-Heavy is an in-between

concept that uses at grade, on-street busways (Finn, et al., 2011).

In the report called “Guidelines för attraktiv kollektivtrafik med fokus på BRT” (Guidelines for Attractive Public Transport with Focus on BRT), some standards for a Swedish BRT are defined with two main levels of operation: green and yellow levels (Ringqvist, et al., 2015). While green level refers to a full, adequate BRT that provides high attractiveness and efficiency, a yellow level service can be accepted as a partial BRT solution and represents a good level of service for trunk lines. The main characteristics of a BRT system, according to Ringqvist, et al. (2015) are:

• Easy to understand and use;

• High visibility in the urban environment, own identity, design and branding;

• Stops and stations, connecting public transport and the urban environment with high quality, in interaction with city life;

• High frequency and long period of operations during the day;

• Uninterrupted travel between stops, full priority at intersections and

• Short line segments with smooth alignment and high-quality road surface.

The Swedish standards used to classify BRT services into green or yellow levels, are presented in

Figure 5.

Figure 5. BRT standards in Sweden (Ringqvist, et al., 2015).

Odbacke (2018) proposes an important and relevant assessment tool for BRT based on 24 criteria divided into four main categories, which receive points in the case the criteria are fulfilled. A larger set of criteria was firstly defined based on “The BRT Standard” (ITDP, 2016) and on “The Guidelines for Attractive Public Transport with Focus on BRT” (Ringqvist, et al., 2015) and after the application of the methodology and a workshop important actors in the field, the criteria set was reduced to the 24 items showed in Figure 6.

Figure 6. Criteria set of components for assessing a BRT corridor (Odbacke, 2018).

Next, some of the most relevant components in a BRT corridor or system, according to the literature, including the ones that are applied in this thesis are presented in more detail.

2.2.1 Right of Way and Busway Alignment

The selection of a proper road infrastructure plays a fundamental role in the type and quality of the BRT service. There are two aspects that need to be considered when designing the road infrastructure for a BRT service: the dedicated right-of-way (ROW) and the alignment of the busway.

The ROW ensures that the buses are not moving in a mixed traffic environment, which is an important factor to ensure optimal operational speeds, reliability and a safe environment (Ringqvist, et al., 2015).

Ideally, BRT lanes should be physically segregated from other vehicle’s traffic, resulting in good compliance and easier enforcement (ITDP, 2016). Types of ROW include median busway, curbside bus lane, bus-only streets, grade separated busways, fixed guided busways, bi-directional one-lane configuration, virtual busway, contra-flow busway (ITDP, 2017).

The alignment of the busways corresponds to the location of the infrastructure within the street area.

aligned with the center of a two-way road. This reduces the number of conflicts with other traffic, including turning vehicles, parking, taxis and delivery vehicles, increasing the general performance of the public transport system (ITDP, 2016). A fully segregated, in the middle of the carriageway, with long extension and well-marked busway get a high punctuation according to Odbacke (2018) classification system. Ringqvist, et al. (2015) emphasizes that the infrastructure must be well adapted to the surrounding environment for safety reasons, especially in highly urbanized areas and city centers, where the number of pedestrians and bicyclists is elevated. Figure 7 shows examples of different types of ROW in different locations. In Curitiba (top left), the busways are in the middle of the road and fully segregated from mixed traffic while in Malmö (top right) the segregation and done by painted divisional lines on the pavement. In Stockholm (bottom) the majority of the ROWs is exclusive bus lanes on the right-hand side of the roads.

Figure 7. Examples of a full segregated, median busway in Curitiba (top left) (Prefeitura de Curitiba, 2014), exclusive bus lanes in the middle, without physical segregation in Malmö (top right) (Google Inc., 2020) and a curbside bus lane in

Stockholm (bottom) (Stockholms stad, 2020).

Regarding dimensions, which are important for the proposed measures in this work, Ringqvist, et al.

(2015) recommends the cross section showed in Figure 8 for busways in the middle of the road section, segregated by white road lines or narrow physical medians.

Figure 8. Median busway delimited with road marking or low curb (Ringqvist, et al., 2015).

The Transit Capacity and Quality of Service Manual (TCQSM) suggests minimum bus and passenger volume ranges for which to deploy each type of ROW, including complementary measures that have to be taken into account. It can be observed in Figure 9 that the requirement for a median bus lane is a bus volume of 60 to 90 buses per hour during the peak hour, which translates to one bus per minute to one bus per 40 seconds. Besides the high bus hourly volume, it is suggested that at least two lanes for general traffic are available in the same direction (Ryus, et al., 2013).

Figure 9. Exhibit 6-34 from TCQSM showing suggestion of minimal bus and passenger volumes to implement each type of

ROW (Ryus, et al., 2013).

The same publication presents some observed travel time savings and reliability improvements for bus services in the United States of America. In New York, the dual bus lanes implemented in Madison Avenue brought a 43% reduction in travel times for express buses while improving the reliability by 11%. In Los Angeles, a bus lane deployed in Wilshire Boulevard improved reliability by 12 to 27% with travel time gains of 0.3-0.5 minutes per kilometer (Ryus, et al., 2013).

Ben-Dor, et al.(2018) assessed the impacts and compared the effects of adding a dedicated bus lane (DBL) and converting one of the existing lanes into a DBL in the city of Sioux Falls using a multi-agent simulation approach. The study explored the effects of the two types of dedicated bus lanes have on travel times for each transport mode and for different population sizes. The results show that adding bus lanes makes the travel times for the buses during peak hour be the same as the travel times in mixed traffic during off-peak, even for a population twice as higher than the original size. It also showed that for the same population growth, the travel times with no bus lanes implemented would be twice as high as well. Finally, the improvement brought by the dedicated lanes promoted a 10% increase in the modal share for public transport.

Kim, et al. (2019) investigated the environmental benefits of BRT corridor in Seoul, South Korea, using a microscopic simulation methodology. Two scenarios were simulated and compared: one with median bus lanes implemented and other with no bus lanes and the buses running in mixed traffic. Results show an increase of 6.2% in the average speed for the buses and 50.2% for the general traffic when the BRT infrastructure is present, which directly impacts the vehicle hours traveled and, consequently, the fuel consumption and vehicle emissions. Fuel consumption is reduced by 18.5% and emissions by 19.3-31.4%, depending on the pollutant.

An interesting research was done by Jiang & Murga (2010) were they investigated the impacts on the whole area of the so-called Chicago Loop, if one of the traffic lanes of Michigan Avenue, in Chicago, was converted to an exclusive bus lane. What makes this study interesting was the unconventional large size (around 4.09km ² ) of the selected study area for a microscopic simulation. The research found that by implementing the dedicated bus lane increase bus average speeds by 21% while general traffic speeds would be reduced by 5.5%.

2.2.2 Stations

Stations are the component that connect the passengers to the vehicles and in BRT corridors, or

systems, they should provide, preferably, pre-boarding fare collection and at level boarding/alighting

through multiple doors (ITDP, 2017). Station characteristics and design, including fare payment, vehicle

floor configuration, door usage and platform configuration can directly influence the buses dwell time

which is one of the main causes of delay in bus services (Ryus, et al., 2013).

Another determinant of bus performance and system capacity is station spacing. A certain space between stops is needed in order for the buses to be able to reach the desired, maximum permitted speed and the higher the permitted speed, the higher the distance between stations needs to be.

However, longer distances between stations incur in longer walking distances and less accessibility, making the service less attractive to the users. Therefore, the balance point must be found (Bösch, et al., 2013). Figure 10 shows the relationship between station distance and vehicle speeds according to Nielsen, et al. (2005). It is possible to observe that for urban buses, distances in the range of 500 to 600 meters, with permitted speeds from 30-50km/h will result in operational speeds between 20- 30km/h.

Several sources recommend a station spacing in the order of 500m to optimize BRT services.

The BRT Standard suggests an optimal distance of 450m and rewards maximum points to corridors that have average distances in the range 300-800 meters (ITDP, 2016). The Swedish BRT guidelines consider green level operation services that have station spacings between 500 and 800 meters Ringqvist, et al.

(2015) and Odbacke (2018) suggest maximum points for corridors that have over 90% where the distances between stations are longer than 500m.

Buses incur in a minimum of 15 seconds delay at each bus stop to serve passengers and, thus, reducing the number of stops and increasing the distance between stops improves the service, even though the dwell times in the remaining stops would increase. Increasing distances from 135m to 400 meters would increase the buses average speed by 24% (Ryus, et al., 2013).

The station size should be dimensioned according to the forecasted passenger demand and with the local requirements for personal space. The configuration and design can change to better fit into constrained street spaces, although it is recommended that the width of the sidewalks should be kept constant around the stations as pedestrian volumes tend to increase (ITDP, 2017). The suggested

Figure 10. Relationship between stop/station distance,

maximum permitted speeds and operational speeds (Nielsen, et

al., 2005).

minimum station width in Sweden is 3.5 meters, including space for the weather protection and information signs, with a desired area of 1m ² per passenger (Ringqvist, et al., 2015).

Center stations that can serve both travel directions are desirable to promote easy transfers between lines and becomes more important as the BRT network expands (ITDP, 2016). In case of space restrictions, stations can be designed longer than the necessary and serve both directions in a split operation. Then, the stopping areas (docking bays) for buses in both directions are not directly in from of each other, reducing the passenger density when two buses stop at the same time. However, it is expected narrower stations to be less comfortable for the users (ITDP, 2017). Figure 11 shows a schematic view of the split operation in a narrow BRT station, according to The BRT Planning Guide (ITDP, 2017) and examples of the same case in Quito, Peru (bottom left) and Lanzhou, China (bottom right).

Figure 11. Schematic view of a narrow, elongated BRT station (top) (ITDP, 2017), and examples from Quito (bottom left)

and Lanzhou (bottom right) (ITDP, 2016).

In the case of center stations, it is desirable that the general traffic lane behind the station have the narrowest dimension possible, ideally 3 meters, to encourage lower speeds and promote a safer environment for passengers moving to and from the station (Ringqvist, et al., 2015). Examples of this type of design from Curitiba and Nantes can be seen in Figure 12.

Figure 12. Examples of narrow general traffic lanes behind BRT stations with shared bicycle lane in Curitiba (left) and Nantes (right). Source: Google Street View, 2020.

2.2.3 Intersection Treatments

Intersection treatments are measures applied to reduce bus delay at intersections and are related to increasing the green time that serves the bus lane, either by reducing the number of conflicting movements with the bus flow or by implementing signal priority (ITDP, 2016). Considering that BRT systems are generally built on corridors or locations where traffic congestion represents a problem to bus operations, implementing exclusive infrastructures of some kind directly eliminates a considerable amount of delay from the service. Intersections, as stations, represent a bottleneck in the system and can highly affect the performance. The main objectives, thus, of intersection design in a BRT corridor are threefold: “to provide safe and convenient crossings for pedestrians, to minimize delay for BRT vehicles and to minimize delay for mixed traffic” (ITDP, 2017).

In general, but specially on high frequency corridors, eliminating turning movements from mixed traffic

across the BRT corridor is more efficient than signal priority (ITDP, 2016). This reduces the number of

phases/stages needed in the traffic signal that controls the intersection, allowing more green time to

the remaining phases/stages. However, eliminating movements in one intersection will divert traffic to

adjacent intersections, considering that demand remains the same, reducing the general accessibility in the area and creating problems elsewhere (ITDP, 2017).

The BRT Planning Guide (ITDP, 2017) suggests the following tool set to deal with intersections in a BRT corridor.

• Simplify the BRT system’s routing structure to optimize turning movements into the corridor;

• Optimize the number of intersections along the corridor;

• Restrict as many mixed traffic turning movements on the BRT corridors as possible;

• Optimize the location of the station relative to adjacent intersections;

• Optimize the signal phasing and consider signal priority for public transport vehicles.

Transit Signal Priority (TSP) is another method of intersection treatment and can be described as the act of alter the signal timings at intersections in a way to give the bus corridor a time advantage over the other movements (Ryus, et al., 2013). Signal priority can be classified into Passive and Active.

Passive signal priority includes all types of measures that reduces delay in the intersection for the prioritized vehicle group or movement without any time of vehicle detection, including exclusive lanes, optimal location of bus stops, green waves and green time distribution optimized for the priority movement (Vägverket, 2004). Passive signal priority for BRT corridors typically requires cycle times between 60 and 120 seconds with 50% of the green times given to the corridor (30 seconds for BRT in a 60 second cycle, for example) (ITDP, 2017).

Active signal priority requires any type of vehicle detection that differentiates the prioritized vehicle from the rest of the demand at an intersection. In this type of TSP there are several actions that can be planned and programed into the traffic controllers:

• Green extension - Extension of the current green time;

• Red truncation - Shortening the conflicting phase/stage in case the prioritized movement has red light when detected;

• Rearrange the phasing/stage sequence to give priority and

• Extra phase/stage - Add an extra phase/stage only when detection occurs (Vägverket, 2004).

The effectiveness of TSP is highly dependable and can, therefore, be affected by long minimum green

times for pedestrians in the conflicting movement, conflicting requests for priority from buses in the main

and side streets and high frequency of buses in a corridor (Wahlstedt, 2014).

There are several types and technologies for vehicle detection in TSP, being the most common the combination of a magnetic detector loop and a vehicle transponder or radio based systems with Automatic Vehicle Location (AVL) systems in the buses (Wahlstedt, 2014). In both cases, there should be a priority request loop, in case of magnetic detectors, or a specific request position, in the case of radio systems, as well as cancelation loops/position. Figure 13 illustrates the concepts of both systems and the detection location within the intersection.

Figure 13. Schematic view of the detection systems (left) and the detection positioning within the intersection (Wahlstedt, 2014).

In Sweden , the active TSP system is called PRIBUSS, which stands for PRIoritering av BUssar i Samordnade Signalsystem, or Prioritization of Buses in Coordinated Signal Systems, and consists of a set of functions that can be customized and programed by the operator to give conditional priority in both coordinated and isolated control. The basic functions are as follows (Vägverket, 2004):

• Bus green extension (BF): the green time that serves the bus approach is extended;

• Early green (ÅTS): the bus approach receives immediate green unless the conflicting movements have just changed to green (min. green time respected);

• Extra stage/phase (EF): an extra signal stage (or phase) is initiated when the prioritized vehicle is detected. After the extra phase, the signal program goes back to subsequent stage/phase or to previous stage/phase;

• Dual extra phase (DEF);

• Abbreviation (AK): the bus signal groups that got Early Green or the Extra Stage can have their

green time canceled as soon as the vehicles have been served and the controller goes back to