traffic capacity models for mini-roundabouts in the …docs.trb.org/prp/13-0209.pdf · 1 traffic...

Traffic Capacity Models for Mini-roundabouts in the United 1

States: Calibration of Driver Performance in Simulation 2

3 4

Taylor W.P. Lochrane* 5 Turner-Fairbank Highway Research Center 6

Federal Highway Administration 7 6300 Georgetown Pike 8

McLean, VA 22101 9 [email protected] 10

11 Nopadon Kronprasert, Ph.D. 12

Virginia Tech, National Capital Region 13 Department of Civil and Environmental Engineering 14

7054 Haycock Road 15 Falls Church, VA 22043 16

[email protected] 17 18

Joe Bared, Ph.D., P.E. 19 Turner-Fairbank Highway Research Center 20



25 Daniel J. Dailey, Ph.D. 26

University of Washington 27 Department of Electrical Engineering 28

1410 NE Campus Parkway 29 Seattle, WA 98195 30

[email protected] 31 32

Wei Zhang Ph.D., P.E. 33 Turner-Fairbank Highway Research Center 34



39 40

*Corresponding Author 41 42

Revised Submission Date: November 15, 2012 43

Submission Date: March 30, 2012 44 Number of Tables and Figures 13*250 = 3,250 45

Number of Words: 4,113 46 Final Word Count: 7,363 47

48 49

TRB 2013 Annual Meeting Paper revised from original submittal.

Lochrane, T., Kronprasert, N., Bared, J., Dailey, D. and Zhang, W.

2

ABSTRACT 50 The design of mini-roundabouts has been around and practiced in Europe for decades. It has 51 been a successful and low-cost intersection configuration. Nevertheless, accessible traffic 52 capacity models for mini-roundabouts do not exist. 53 54 This study provides design recommendations and a simulation approach for capacity models of 55 mini-roundabouts from USA data. Two typical geometries are selected that have a 24’ or 36’ 56 approach widths typical of conventional intersections. The mini-roundabouts are best defined by 57 the Inscribed Circle Diameter of 50’ and 75’. Mini-roundabouts are low-cost treatments using 58 existing external boundaries of intersections. Field data are collected on Critical Gap and 59 Headway acceptance for a similar design located in Stevensville, MD in order to calibrate a 60 simulation. VISSIM Microsimulation software is used to model the selected prototype designs 61 for capacity estimations. The defining feature for mini-roundabouts is the traversable central and 62

splitter islands for large vehicles that make through or left turn movements. 63 64

The linear capacity models presented, estimate the capacity of the mini-roundabouts to be lower 65 than that of the single-lane roundabout. However the mini-roundabout has a higher capacity per 66 square foot of land which would be an innovative solution for urban areas for increasing capacity 67 at existing AWSC intersections at lower cost than single-lane roundabouts. 68

69



3

INTRODUCTION 70 Over the past decade in the USA, single–lane and dual-lane roundabouts are becoming the 71 preferred choice of intersection design for States, cities and towns. They have proven benefits 72 in safety and traffic flow improvements when properly designed. One of the designs that may 73 capture the attention of engineers in the US is the mini-roundabout. However this design is not 74 new to the roundabout community outside the US. The mini-roundabout is an innovative 75 intersection design in the US that can improve safety, reduce delays and facilitate slower speeds. 76 77

According to the NCHRP, of the Federal Highway Administration’s (FHWA) 78 “Roundabouts: An Informational Guide,” mini-roundabouts are small roundabouts with fully 79 traversable central islands used in low speed urban environments (1)(5). Mini roundabouts are 80 described by their traversable inscribed circle diameter (ICD) which is the distance between the 81 outer edges of the circular roadway. The design of a mini-roundabout should fit within existing 82

boundaries whether they are curb returns or edge of pavements with shoulder. 83 Much of the literature about mini-roundabouts focuses on the design criteria from other 84

countries and how they can be applied in the USA. One of the advantages in designing a mini-85 roundabout is that the diameter can range from 45 feet to 80 feet to stay within current 86 boundaries of a typical intersection (1). The mini-roundabout is designed such that trucks and 87 busses are allowed to traverse over the central island. Experience in Germany with trucks and 88 busses traversing over the center island developed a limit of the height for the central island to 89 not exceed 4.7 inches (2). Clive Sawers provides recommendations that in the USA the height of 90 the center island should not be lower than 2 inches, a slight cross slope from the center and a the 91 curbing to be 1:4 (3). This is to provide a safe height measure for trucks and busses to go over 92 the island, ensures proper drainage and guides passenger cars safely though the mini-roundabout. 93 One of the concerns with constructing a new design such as the mini-roundabout in an area 94 unfamiliar with the design is the impact on operational performance. This was investigated in 95

Germany with a study that looked at the impact of converting 13 unsignalized intersections to 96 mini-roundabouts. This study concluded no significant delay impacts with the mini-roundabout 97 at volumes of approximately 17,000 vehicles per day (2). 98

The only capacity model that is available for analyzing mini-roundabouts is a empirical 99 based model from a study by the United Kingdom in a program called ARCADY executed by 100 Transport Research Laboratory (4). One of the equations from ARCADY/3 used in Europe for 101 calculation of 4-way mini roundabouts is shown in equation 1 (7). 102

103

(1) 104 105

where 106

( ) ( )

107 The NCHRP Report 572, “Roundabouts in the United States,” provides capacity equations for 108 single and multilane roundabouts (5). The equation provided in NCHRP-572 for a single lane 109 roundabout, is show in equations 2. This is the same model used in the HCM 2010 (6). 110

111 112



4

( ) (2) 113 114 where 115

( )

( ) 116

METHODOLOGY 117 To model the capacity of a mini-roundabout, three steps are undertaken. First, data on driver 118 behaviors and travel characteristics at mini-roundabouts are observed from a similar mini-119 roundabout located in Stevensville, MD. Second, a microscopic traffic simulation model is 120 developed to emulate drivers’ behavior and to simulate for different traffic flow scenarios in over 121 saturated conditions. Finally, a regression model is developed and fitted to the simulated data to 122 estimate the capacity of mini-roundabouts. 123

124

PERFORMANCE OF SIMILAR DESIGNS IN THE UNITED STATES 125 There are very few mini-roundabouts constructed in the United States that have all the desirable 126 design recommendations. More importantly, no mini-roundabouts in the US operate at or near 127 capacity. One site constructed in Stevensville, Maryland conforms closely to the basic design of 128 a mini roundabout with an ICD of 80’. Nevertheless, the central and splitter islands are not raised 129 and have no passenger car deterrent except for flex-posts located around the central island. This 130 site was selected to evaluate the driver behavior with regard to gap and headway decisions. 131 Video recordings were collected using cameras that captured data from 3:45 pm to 5:45 pm. The 132 volume for this intersection is listed in Table 1. The cameras were set 30’ high on a telescopic 133 pole shown in Figure 1. 134

The video data were used to collect time gaps (both accepted and rejected gaps) and 135 follow-up time. An accepted gap is where a driver on the approach decides to move into the 136

circulating stream as the (time) gap between vehicles is perceived sufficiently long. Rejected 137 gaps are where a driver chooses not to move into the circulating stream as the gap is insufficient. 138 Follow-up time is the (time) gap between the second vehicle and lead vehicle when entering the 139 circulating stream. The driver behavior for cars and heavy vehicles were analyzed separately. 140

141 Figure 1: Data collection (left) and Google aerial photo (right) Stevensville, MD 142

143 144



5

TABLE 1: Traffic Characteristics Stevensville, MD 145 Measures Intersection Approaches Total

West Leg

(Left)

North Leg

(Off-ramp)

South Leg

(Bottom)

Entry flow, peak (veh/h) 290 320 270 880

Entry flow, off-peak (veh/h) 160 240 200 600

Peak-hour factor 0.6 0.65 0.65 -

Vehicle composition (%)

- Passenger cars

- Pickup trucks/SUV’s

- Trucks

47%

51%

2%

48%

49%

3%

44%

48%

4%

46%

49%

3%

Turning movement (%)

- Left-turn

- Right-turn

82%

18%

61%

39%

36%

64%

60%

40%

Note: * Calculate from the maximum 5-minute peak entry flow. 146

147

TABLE 2: Field Data Results Stevensville, MD 148 Approach Decision Critical Lag Critical Gap Follow-up Headway

Range Mean S.D. N Range Mean S.D. N Range Mean S.D. N

West Leg Accepted 2.6-4.9 3.6 0.8 9 3.5-4.3 3.9 0.6 2 2.2-3.8 2.7 0.5 13

Rejected < 4.3 2.0 1.2 28 1.9-4.2 3.1 0.8 22

North Leg Accepted 1.7-6.7 3.4 1.1 20 5.3-5.4 5.3 0.1 2 2.0-3.6 2.7 0.6 5

Rejected < 3.5 1.8 1.2 9 1.8-2.6 2.2 0.4 3

South Leg Accepted 2.1-4.7 3.5 0.8 10 5.5 5.5 - 1 2.1-3.6 2.8 0.5 10

Rejected < 4.3 2.4 1.1 11 1.7-4.5 3.1 0.9 12

Note: S.D. = standard deviation, N = sample size. 149 The results are based on passenger cars and SUVs. only. 150

151

One observation in Table 2 are that the vehicles on the west leg have shorter critical gap because 152

the distance between the entry of the north leg and west leg is shorter. As a result, vehicles on the 153 west leg entering the roundabout were more aggressive with the gap from the vehicles 154 approaching from the north leg. These critical gap values and follow up headway will be used in 155 the calibration of the microsimulation model. 156

157

DRIVER BEHAVIORS AT MINI-ROUNDABOUTS 158 By the design of a mini-roundabout, heavy vehicles (HV) need to climb over the traversable 159 circular island because their tracking surface requires a wider swept path Hence, heavy vehicles 160 and cars traverse along different paths through a mini-roundabout and the drivers of these two 161 vehicles behave in a different way. The cars’ drivers yield to the vehicles in the circulating 162 stream similar to those at a modern roundabout. On the other hand, trucks’ drivers at a mini-163 roundabout are more likely to behave as those at an all-way stop-controlled (AWSC) 164 intersection. Heavy vehicles not only yield to the circulating vehicles but also verify that there is 165

no other vehicle entering the intersection; in other words, when the trucks enter the mini-166 roundabout, vehicles from other approaches are stopping. 167

There are two factors influencing the drivers’ yield decision, drivers’ yielding behavior 168 and the other is the drivers’ gap acceptance. 169

170 171 172



6

Drivers’ Yielding Behavior 173 Based on field observations, when vehicles arrive at a mini-roundabout, they yield to other 174 vehicles. Yielding behaviors of drivers at a four-legged mini-roundabout can be classified into 175 four rules based on types of vehicle conflicts (Car vs. Car, Car vs. HV, HV vs. Car, and HV vs. 176 HV.) Consider vehicles from the eastbound (EB) approach. 177 178 Rule 1: Entering cars yield to other cars in the circulating stream, if the gap time and headway are 179

sufficiently safe. 180 181

Rule 2: Entering cars yield to (i) left-turn and through heavy vehicles from the southbound (SB) 182 approach; (ii) left-turn heavy vehicles from the westbound (WB) approach; and (iii) left-183 turn and through heavy vehicles from the northbound (NB) approach, if the gap time and 184 headway are sufficiently safe. 185

186 Rule 3: Entering heavy vehicles yield to other cars in the circulating flow or from the SB 187

approach, if the gap time and headway are sufficiently safe. 188 189

Rule 4: Entering heavy vehicles yield to (i) left-turn and through heavy vehicles from the SB 190 approach; (ii) left-turn and through heavy vehicles from the WB approach; and (iii) left-191 turn and through heavy vehicles from the NB approach, if the gap time and headway are 192 sufficiently safe. 193

194

Figure 2 shows that conflict areas that the drivers of cars and heavy vehicles approaching a mini-195 roundabout that are taken into account in simulation. When a car arrives, it yields to the vehicles 196 in the circulating stream and to those on his/her left approach. If four cars arrive (one on each 197 approach) at a mini-roundabout, then the total of four cars can be served at the same time. When 198

a heavy vehicle arrives, the mini-roundabout is operated as first-come-first-serve. When the 199 heavy vehicle enters, all other vehicles stop until the heavy vehicle leaves the intersection. The 200 only difference between the yielding rules between cars and heavy vehicles is that a car may 201 enter when there is a through heavy vehicle from the opposite direction. However, this is not the 202 case for a heavy vehicle. A heavy vehicle will stop and wait until no other vehicles occupy the 203 intersection box. 204 205

206 Figure 2: Clearance Zones for Cars and Heavy Vehicles 207

208 209



7

Setting the reasonable gap times and headways for each conflict in the simulation is challenging 210

and these values affect the capacity estimate of the mini-roundabout. 211 212

Drivers’ Gap Acceptance 213 Once the data were collected and collated, curves representing the frequency of rejected and 214 acceptance gaps were developed, and critical gap and lag are calculated. Gap acceptance 215 parameters form a crucial part of many analytical and simulation models to evaluate the capacity 216 of roundabouts. Critical gap is defined as the threshold by which drivers approach at a 217 roundabout judge whether to reject or accept the available gap (yield or not yield to vehicles 218 circulating). Critical lag happens when the number of accepted shorter lags is equal to the 219 number of rejected longer lags (14). 220 221

MIRCOSIMULATION 222 VISSIM simulation software is used to model the capacity of 50’ and 75’ ICD mini-roundabouts. 223 VISSIM is a microscopic, time-step and behavior-based traffic flow simulation model that 224 includes urban traffic operations. (10) The design templates of the 50’ and 75’ ICD mini-225 roundabouts designed from a CAD are imported as background to the VISSIM software. The 226 links in VISSIM are created for all vehicle movements and the vehicle characteristics are 227 specified. For both 50’ and 75’ ICD mini-roundabouts, desired speeds of vehicles are set 228 between 22 and 28 mph, and entry speeds of vehicles are 15 to 20 mph for both cars and heavy 229 vehicles. Circulating speeds of all vehicle types are 8 to 12 mph in a 50’ ICD mini-roundabout, 230 and they are between 13 to 17 mph in a 75’ ICD mini-roundabout. These speeds are based on 231 field measurements at the Stevensville, MD site 232

In the simulation, the eastbound (EB) approach of a four-legged mini-roundabout 233 intersection was selected to measure estimated capacity, or maximum throughput. The concept 234 for estimating capacity on the EB approach is to flood vehicles from this approach while varying 235

the combination of traffic volumes at the other approaches. The volumes are designed to 236 represent different traffic conditions with each approach simulating different intersection 237 demands. The design for the input volumes simulated 343 different traffic scenarios (7 traffic 238 volumes––0, 200, 400, 600, 800, 1000, 1200 veh/h on three approaches). The calculation is 239 repeated for every 2 percent increase in heavy vehicles (from 0 to 10%). The proportions of left-240 turn and right-turn movement on each approach are selected randomly between 0-20%. 241 242

Vehicle Paths in VISSIM 243 One of the main differences between the mini-roundabout and the single-lane roundabout is the 244 route for heavy vehicle (HV) movements. Given the complexity of heavy vehicle movements 245 and their permutations (Car vs. Car, Car vs. HV, HV vs. Car, HV vs. HV) heavy vehicles and 246 cars were designated to have separate VISSIM links/connectors and specific routing 247

assignments, as shown in Figure 3. The priority rules for heavy vehicles to traverse the island 248 require all other approaches to be clear as a heavy vehicle enters into the roundabout. 249

250 251 252 253 254



8

255 256

257 Figure 3: Routing for Car’s (above), Routing for HV’s (below) 258

259 Priority Rules in VISSIM 260 In a VISSIM simulation, priority rules are used to simulate vehicle movements at conflict 261 locations. The two parameters that are used to control priority rules are; (1) gap time and (2) 262 headway spacing. Gap times are defined as the minimum gap in time that a merging vehicle 263 needs to enter a circulating traffic stream. The headway is defined as a distance that identifies a 264 space that will need to be clear for the merging vehicle to enter the traffic stream under the 265 congested condition. The priority rules for the mini-roundabout design are more complicated 266

than a typical single-lane roundabout. This part of the design is most critical as a result of the 267 interaction between cars and heavy vehicles that requires different headways and gap time 268 calibrations between different layers. Figure 4 shows the clearance zones for single-lane, AWSC 269 and Mini-roundabout. The clearance zone shows the different priority rules for each of the three 270 designs and how the mini-roundabout is a combination of the single-lane and AWSC. 271

272 (a) Single-lane roundabout (b) AWSC intersection (c) Mini-roundabout 273

274 Figure 4: Clearance zones for different intersection controls 275



9

In VISSIM, priority rules are set by two markers. One is a red stop line marker, and the other is a 276

set of green conflict markers. Figure 4 shows examples of priority rules used for left-turn 277 entering car. The red stop marker on the entry link indicates entering vehicle boundary and the 278 green conflicting markers represent the circulating flow boundaries. The space between the 279 downstream green dashed line and the upstream green dashed line represents the headway 280 spacing. The priority rules for a single lane roundabout are much simpler than that of a mini-281 roundabout design with only the merge conflict being the only location for priority rules. For 282 example, for left-turn entering cars, nine priority rules are considered as shown in Figure 5. 283 Based on drivers’ yielding behaviors, four stop lines are used at the entry of each approach for 284 different types of vehicle conflicts. These four stop lines are all set at the same location on each 285 approach (just before entering the mini-roundabout.) For each stop line, the conflict markers are 286 defined according to different movements of entering vehicles and conflicting vehicles. 287 288

The critical gap values from field observations were used to create the clearance zones (as 289 demonstrated in Figure 5), although the observation site has different intersection configuration 290 from the proposed mini-roundabouts. These conflicting zones are then defined by the gap (time) 291 during uncongested condition and headway (distance) during congested condition in VISSIM. 292

293 These conflicting zones are then defined by the gap (time) during uncongested condition and 294 headway (distance) during congested condition in VISSIM. The conflicting zones for Car vs. 295 Car (M1-M2) in Table 3(c) are defined by both gap (time) and headway (distance) to ensure that 296 the entering cars will yield to the conflicting cars from the left approach and circulating cars. 297 The values were derived from the mean accepted gaps obtained from field data; however, the 298 values were refined in order to match with the typical 2-lane and 3-lane 4-legged intersections. 299 300

The conflicting zones for Car vs. HV, HV vs. Car, and HV vs. HV are defined by headway 301

(distance). These clearance zones are defined from the departure point to the 10-ft downstream 302 of the stop markers of the conflicting approaches. By locating these conflicting markers (from 303 the departure points to the 10-ft downstream), it will ensure that the vehicles will operate as first-304 come first-serve at the intersection and when a truck is in the mini, no other vehicles will enter 305 the intersection. The gap and headway values for each clearance zone are varied based on the 306 inscribed diameter of a mini-roundabout and intersection configurations. 307 308



10

309

Figure 5: Example of priority rules for a left-turn entering car 310

Table 3 compares the priority rules used in VISSIM for controlling vehicles at a single-lane 311 roundabout, an all-way stop-controlled (AWC) intersection, and a mini-roundabout. For a single-312

lane roundabout, only three priority rules are defined: (i) left-turn or through entry vehicles 313 conflicting with circulating vehicles from left or opposite approach; (ii) right-turn entry vehicles 314 with leading left-turn or through; and (iii) left-turn or through entry vehicles with leading right-315 turn for a total of 9 rules. For an AWSC intersection, 13 priority rules are defined, five rules for 316

entry vehicles conflicting with vehicles from left approach, four rules with vehicles from 317 opposing approach, and four rules with vehicles from right approach, for a total of 52 rules. For 318 a mini-roundabout, 30 priority rules are defined for different combinations of vehicle conflicts on 319

one approach (12 for entry cars and 18 for entry HVs.) These priority rules are set similarly for 320 all approaches; hence, there is a total of 120 rules for a four-legged mini-roundabout. The 321 priority rules for Car vs. Car at a mini-roundabout (M1-M3) are similar to those for a single-lane 322 roundabout (R1-R3), and the priority rules for HV vs. HV at a mini-roundabout (M18-M30) are 323

Circulating cars from

the left and opposite approach

Leading right-turn car

on the same approach

Left-turn heavy vehicles from

the left approach

Through heavy vehicles from

the left approach


the opposite approach


the right approach

Through heavy vehicles from

the right approach

Left-turn heavy vehicles


Leading right-turn heavy vehicle


M1 M3 M4

M5 M7 M9

M10 M11 M12



11

the same as those for an AWSC intersection (S1-S13). The added priority rules needed for a 324

mini-roundabout are those for Car vs. HV (M4-M12) and HV vs. Car (M13-M17). 325 326

TABLE 3: Comparison of Priority Rules for Different Intersection Controls in VISSIM 327 328

329 (a) Priority Rules for a Single-Lane Modern Roundabout 330

331

332 (b) Priority Rules for an All-Way Stop-Controlled Intersection 333

334

335 (c) Priority Rules for a Mini-Roundabout 336

337

CAPACITY MODELS 338 The capacity models of mini-roundabouts in this study are developed using a calibrated micro-339

simulation model, that use gap-acceptance modeling to emulate drivers’ yielding behavior to 340

circulating vehicles before entering a mini-roundabout. The parameters for the functional form of 341 the capacity of a mini-roundabout are estimated using a linear regression model. It is assumed 342 that the capacity (entry flow rate) is a function of circulating cars and conflicting truck 343 movements HVs, shown in Figure 6(a). 344 345

346



12

Determination of Entering and Circulating Flows 347 To model the capacity of the mini roundabout, multiple data evaluation points are created in a 348 VISSIM simulation model in the roundabout to record the entering and circulating flow rates, see 349 Figure 6(b). The first evaluation point is set on the eastbound approach and counts cars for 350 maximum throughput or estimated capacity. This approach is flooded in the simulation (with 351 2400 veh/hr) to make assure that there is always a queue. To calculate the entering volume 352 passenger car equivalence (PCE) for trucks, the AWSC PCE for level terrain is weighted by a 353

( ) default value of 1.7 (6). According to HCM 2010, a factor 2.0 is applied to trucks to 354 convert to passenger car equivalents for a single lane roundabout (11). Based on the turning 355 movements of heavy vehicles for the mini-roundabout which relate similar to AWSC, the default 356 value was set to 1.7, see Eq. (2). 357 358

Entry Flow Rate (Ve) 359

(2) 360

361

where: 362

( )

363 Only one evaluation point is needed to capture passenger cars that are circulating and are located 364 north of the entering volume and are in conflict with the eastbound approach. In Figure 6(a), 365

three evaluation points are needed to capture the heavy vehicle movements; NB Left ( ), 366

WB Left ( ) and SB Thru( ) . The circulating volume is calculated by the 367

weighted sum of circulating vehicles and heavy vehicles as shown in Eq. (3). 368

369 Circulating Flow Rate (VC) 370

( ) (3) 371

372

where: 373

( ) 374 For the circulating volumes conflicting with the eastbound approach, the west, north and south 375

bound approaches are set to provide combinations of volumes were the sum did not exceed 1400 376 veh/hr. 377

378



13

379 (a) Vehicle movements (b) Data collection points 380

Figure 6: Entering and Circulating Flow for a Mini-Roundabout 381 382

Estimation of Capacity Models 383 The equations developed for 50’ and 75’ ICD are shown in Equation 4 and 5. The approach 384 capacities are estimated using various ranges of circulating volumes (which is a result of 385 different combination of approach volumes from conflicting approaches, % HV’s, % Left turns 386 and % Right-turns) that are derived from hypothetical flows on adjacent approaches. The 387 simulation results indicated that the capacity of 50’ mini-roundabout is smaller than that of 75’ 388 mini-roundabout. An increase in conflicting vehicles causes a reduction in the mini-roundabout 389 capacity for the 50’ verses the 75’ mini-roundabout results (see coefficient values: -1.025 for the 390 50’ and -0.944 for the 75’). 391

392

(4) 393

394

(5) 395

396

where is the conflicting vehicles in passenger car equivalent per hour. 397

398 Figures 7(a) and 7(b) illustrate capacity estimates from VISSIM for mini-roundabouts with 50’ 399 and 75’ ICD, respectively. In these figures, the simulated data points are plotted and the best 400 fitting curve based on a linear regression is presented. 401 402



14

403 (a) 50-ft ICD (b) 75-ft ICD 404 405

Figure 7: Mini-Roundabout Entry Capacity as a Function of Conflicting Flow 406

407 Effect of Heavy Vehicles 408 Heavy vehicles used in this simulation are appropriate for light trucks, school busses, fire trucks, 409 delivery trucks, tow trucks etc. Trucks are shown to have an effect on capacity as shown in 410 Figure 8. Simulations where run with 50% and 100% trucks that show significant impacts on 411 capacity but these situations are not typical at roadway performance and were not included in the 412 model design. The vehicles with trailers such as WB-50 trucks were not modeled in this 413 simulation. The WB-50 truck will still be able to traverse the 50’ and 75’ ICD mini-roundabout 414 tested using AUTO turn. The study estimates the capacity model for 0-10% heavy vehicles only 415

which more realistically represents typical conditions in the field. 416

417 418

419 420 421

422 423 424 425

426

200

400

600

800

1000

1200

Conflicting Flow Rates (pcu/h)

Maxim

um

En

try F

low

Rate

s (

pcu

/h)

00 200 400 600 800 1000

50’ ICD Mini-roundabout

C50ICD = 1009 – 1.025Vc; R2 = 0.978

200

400

600

800

1000

1200

Conflicting Flow Rates (pcu/h)

Maxim

um

En

try F

low

Rate

s (

pcu

/h)

00 200 400 600 800 1000

75’ ICD Mini-roundabout

C75ICD = 1020 – 0.944Vc; R2 = 0.967



15

427

428

Figure 8: Effect of heavy vehicles on the mini-roundabout capacity 429

430 The capacity estimates of mini-roundabouts when compared to other intersection alternatives, 431 including an all-way stop-controlled (AWSC) intersection and a single-lane modern roundabout, 432 show that the capacity of a mini-roundabout is higher than that of an all-way stop-controlled 433 intersection, but lower than that of a single-lane modern roundabout. This is because there are 434 more complex interactions among vehicles at mini-roundabouts than modern roundabouts. For 435 example, the simulation suggests that the NCHRP-572 equation for single lane roundabout 436 predicts higher values for capacity than the mini roundabout model presented in this paper, see 437

Figure 9. 438 439

0

200

400

600

800

1000

1200

0 200 400 600 800 1000

Maxim

um

En

try F

low

Rate

s (

pcu

/hr)

Conflicting Flow Rates (pcu/hr)

No Heavy Vehicles (0% HV)

0

200

400

600

800

1000

1200

0 200 400 600 800 1000

Maxim

um

En

try F

low

Rate

s (

pcu

/hr)


10% HV

0

200

400

600

800

1000

1200

0 200 400 600 800 1000

Maxim

um

En

try F

low

Rate

s (

pcu

/hr)


50% HV

0

200

400

600

800

1000

1200

0 200 400 600 800 1000

Maxim

um

En

try F

low

Rate

s (

pcu

/hr)


100% HV



16

440

Figure 9: Comparison of Capacity Estimates of Different Intersection Designs 441

Capacity to Area 442 One of the advantages of the mini-roundabout is the amount of land required to build or convert 443 to this intersection. When building a single-lane roundabout additional right-of-way is usually 444 required based on the size of the ICD compared to an AWSC. In Figure 10, the entry capacity 445 per foot-print area is shown as a function of circulating flow. The 50’ and 75’ ICD mini-446

roundabouts have a larger entry capacity per area than that of the single lane roundabout. This 447 means that the mini-roundabout uses the area more efficiently based on the demand of the 448 entering capacity. This would be applicable for converting an AWSC intersection to a Mini-449 roundabout resulting in an increase in capacity. For example; if 200 vehicles are circulating, the 450 entry capacity for the 50’, 75’and single lane roundabout is about 400, 200 and 100 vph / 1,000sf 451 respectively. Thus, the 50’ mini-roundabout has higher capacity per square foot up to a 452 circulating volume greater than 800vph, where the single-lane roundabout surpasses the mini-453 roundabout. 454

0

200

400

600

800

1000

1200

0 200 400 600 800 1000

Maxim

um

En

try F

low

Rate

s (

pcu

/hr)

Conflicting Flow Rates (veh/hr)

50-ft ICD mini-roundabout


Single-lane roundabout (NCHRP)



17

455

Figure 10: Entry capacity per area 456

CONCLUSION 457

This research has presented a methodology to estimate parameters of an analytical capacity 458 model for the 50’ and 75’ ICD mini-roundabouts. The linear models presented are based on 459 VISSIM simulation results where VISSIM has been calibrated from field data. The models 460 presented include parameters for the 50’ and 75’ ICD mini-roundabout designs. The simulation 461

results suggest that linear models provide useful estimates of capacity for the two selected 462 prototype designs. Estimated capacity for the 50’ICD mini-roundabout is smaller than the 75’ 463 ICD where heavy vehicles are shown to affect the capacity at high heavy vehicle presence. When 464 comparing the land required to construct or convert from an AWSC to a mini-roundabout or 465 single-lane roundabout, the mini-roundabout has higher entering capacity per square foot. This 466 would be a useful design when optimizing existing land in urban areas to increase capacity at 467

lower cost. 468

469

0

100

200

300

400

500

600

0 200 400 600 800 1000

En

try C

ap

acit

y p

er

Are

a (

pcu

/hr/

1,0

00 ft2

)

Conflicting Flow Rates (veh/hr)



Single-lane roundabout (NCHRP)



18

REFERENCES 470

471 1. Robinson, B. W., Rodegerdts, L., Scarborough, W., Kittleson, W., Troutbeck, R., Brilon, 472

W., et al. (2000). “ROUNDABOUTS: An Informational Guide.” Federal Highway 473 Administration, McLean. 474

2. Brilon, W. (2005). “Roundabouts: A State of Art in Germany,” National Roundabout 475 Conference: 2005 Proceedings, p. 15, Transportation Research Board, Washington, D.C. 476

3. Sawers, C. (2009). “Mini-Roundabouts for the United States.” ITE Journal , 79(2), pp. 477 46-50. 478

4. “Assessment of Roundabout Capacity and Delay (ARCADY).” Retrieved July 25, 2011, 479 from <https://www.trlsoftware.co.uk/ProductContent.aspx?ID=1> 480

5. Rodegerdts et al. (2007). “NCHRP 572 Applying Roundabouts in the United States.” 481 Transportation Research Record of the National Academies. Washington, D.C. 482

6. Highway Capacity Manual 2010. Transportation Research Board, Washington, D.C.. 483 2010. Chapter 20. 484

7. Cicu, F., Illotta, P., Bared, J., and Isebrands, H. (2011). “VISSIM Calibration of 485 Roundabouts Traffic Performances,” Transportation Research Board 90

th Annual 486

Meeting, Washington, D.C. 487 8. Federal Highway Administration. (2010). “Mini-Roundabouts: Technical Summary.” 488

FHWA Office of Safety, Washington, D.C. 489

9. Waddell, E., & Albertson, J. (2005). “The Diamondale Mini: America's First Mini-490 Roundabout,” National Roundabout Conference: 2005 Proceedings, Transportation 491 Research Board, Washington, D.C. 492

10. PTV America (2009). VISSIM Users Manual. Version 5.10. 493 11. Highway Capacity Manual. Transportation Research Board, Washington, D.C.. 2000. 494

ISBN 0-309-06681-6 495

12. Kyte, M. and Marek, J. (1989). Estimating Capacity and Delay at a Single-Lane 496 Approach, All-way Stop-Controlled Intersection. Transportation Research Record: 497 Journal of Transportation Research Board No. 1225, Washington D.C., pp.73-82. 498

13. Sawers, C. (1996). “Mini-Roundabouts, getting them right!” Euro-Marketing 499 Communications, Canterbury. UK. 500

14. Raff, M. S. and Hart, J. W. (1950). A Volume Warrant for Urban Stop Signs. Eno 501 Foundation for Highway Traffic Control, Saugatuck, CN. 502


traffic capacity models for mini-roundabouts in the …docs.trb.org/prp/13-0209.pdf · 1 traffic...

Documents