Journal of Transport System Engineering 1 : 1 (2014), 67-73

JTSE

Roundabout or Traffic Signal: A Selection

Dilemma
a,* b
Akmal Abdelfatah, Anil Minhans,
a
Department of Civil Engineering, American University of Sharjah, Sharjah, UAE.
b
Faculty of Civil Engineering, Universiti Teknologi Malaysia, Skudai, Johor, Malaysia.

*Corresponding author: akmal@[Link]

Abstract

At-grade intersections are one of the vital elements in any road network. Traffic control for at-grade
intersections plays a significant role in the level of delay experienced by the vehicles on the road network.
Traffic signals and roundabout control are the most common types of at-grade intersection control. Many
transportation planners around the world debate the selection between these two types of control for a new at-
grade intersection. The aim of this research is to provide some guidelines for selecting one of the two types of
control for some at-grade intersections. The research utilized the well-known software packages of Synchro and
Sidra to evaluate the performance of traffic signals and roundabouts, respectively. The experimental work
considered a four leg at-grade intersection that was operated under different traffic congestion levels and
different traffic distribution percentages. The research results indicated that roundabouts are recommended for
small traffic volumes while traffic signals are proven to be more suitable for high traffic volumes.

Keywords: Traffic signals, roundabouts, intersection delays, planning applications

Usually, urban planners cannot recommend the use

1. INTRODUCTION of a roundabout or a traffic signal or replacing one
of them at any intersection without consulting
Traffic congestion is one of the major problems in traffic engineers. Traffic engineers usually use
many cities around the world. Local agencies try to computer software programs to analyze whether the
often propose some solutions to reduce traffic intersection should be implemented with a
delays at intersections. Some of the proposed signalized intersection or a roundabout intersection.
solutions include changing roundabout intersections Then, based on a comparison on the average delay
to signalized intersections. Nonetheless, many in each case, a decision can be made.
professionals believe that the solution to this daily
problem is much more complicated than just Even though there are plenty of commercial traffic
replacing roundabout intersections with signalized design software programs available, choosing
ones. This is a concern to both urban planners and between a roundabout or a signalized intersection is
traffic engineers on the spot, as the public expect not a straightforward decision. The reasons behind
urban planners to open new areas carefully to avoid that are: firstly, the software programs are not
future congestion and traffic engineers to propose always affordable or available. Secondly, urban
solutions to existing intersections. planners are usually not familiar with the software
programs. Therefore, the urban planners need

67
 Journal of Transport System Engineering 1 : 1 (2014), 67-73

traffic engineers to run the software analysis. life operation of roundabouts. The application of
Running the software analysis is time consuming, the proposed formulation to a case study indicated
while finding and hiring a traffic engineer might be that the proposed model can improve the
cost-intensive. Therefore, this study is aimed at roundabout performance and provide some
providing some guidelines on the suitability of guidelines for their use in real-life applications [10].
constructing a roundabout or a traffic signal based Different evaluation techniques for applying traffic
on the level of traffic congestion. signals to roundabouts have been discussed in
several papers [11-16].
In general, the objective of this study is to develop
a selection procedure that can assist urban planners 3. EXPERIMENTAL DESIGN
in selecting the type of intersection control (i.e.
roundabout or traffic signal) without conducting a In this study, we analyzed a 4-leg intersection with
detailed analysis of the considered intersection. the 4 approaches labeled as A, B, C, and D, as
shown in Figure 1. Each approach had 2 lanes. The
2. RESEARCH BACKGROUND intersection was evaluated for both roundabout and
traffic signal control for the following conditions.
The delay and level of service on at-grade
intersections have been considered by many
researchers in the past. The use of roundabouts has
been increasing over the past decade. A study
estimated the number of roundabouts in the US and
Canada to be more than two thousand in 2010 [1].
Delays at roundabouts have been considered in
several researches since the 1960’s and 1970’s [2-
5]. Most roundabout evaluation studies adopted
SIDRA software for the estimation of delay and
level-of-service characteristics [6, 7]. The use of a
roundabout has been recommended as it has been
proven to improve safety and provide better traffic
performance for high traffic volumes, than all-way-
stop controlled intersections or traffic signals [8].
One such research study had compared the Figure 1. Considered Intersection
roundabout performance under different control
types (i.e. yield control, two-way stop, all-way stop,  The total traffic volume on the intersection
or traffic signal on the roundabout). The analyses was varied from low traffic congestion to
concluded that roundabouts are recommended for very high traffic congestion. The total
high through or left turning traffic volumes [9]. traffic volume for the intersection on
However, this recommendation was based on two approaches A, B, C and D were 2500,
scenarios of traffic distribution only; equal split of 3000, 3500, 4000, 4500, and 5000 veh/hr.
traffic on the four approaches or a minor street with These values represent the total traffic
a very small traffic volume. Therefore, the volume on the four approaches.
generalization of the results cannot be simply
justified.  For the purposes of the study, the traffic
distribution on the approaches was
In recent studies, the use of signals on roundabouts considered for the three cases:
has been considered as a special case of traffic
signals to determine the optimum lane marking and I. Equal distribution of traffic volume on
signal timing. Using the problem solving all approaches (25% on each
formulation, these studies aimed at maximizing the approach)
capacity, minimizing the delay and the cycle length,
as well as much needed requirements to match real-

68
 Journal of Transport System Engineering 1 : 1 (2014), 67-73

II. Dominant traffic flow was purposely 1. The percentage of vehicles turning left
set to flow on one street (both on approach A would increase from 0
directions). The dominant traffic was to 50% (increment of 10%) of the
considered to be on approaches B and volume assigned to left turn
D. The percentage of traffic using movement on approach A. On
these two directions were 60%, 70% approach B, vehicles turning left
or 80%. The distribution of the would decrease from 50% to 0%
dominant flow over the two (decrement of 10%). Approaches C
approaches started with a 50/50 split and D would have a constant left
and then the percentage of traffic on turning percentage of 25%.
one of the approaches (B) increased
by an additional 10% or 20%. Table 1 2. The percentage of vehicles turning left
summarizes the considered cases for on approach C would increase from 0
this scenario. The traffic volume on to 50% of the volume assigned for the
approaches A and C was equal. left turn movement on approach C. On
approach B, vehicles turning left
Table 1. Traffic Distribution for Case II would decrease from 50% to 0%. Left
60 % of traffic a. 30% on B and 30% on D turn percentage on approaches A and
flow assigned to b. 40% on B and 20% on D D would be fixed at 25%.
approaches B&D c. 50% on B and 10% on D
70 % of traffic d. 35% on B and 35% on D 3. The percentage of vehicles turning left
flow assigned to e. 45% on B and 25% on D on approach D would increase from 0
approaches B&D f. 55% on B and 15% on D to 50%. On approach B, vehicles
80 % of traffic g. 40% on B and 40% on D
turning left would decrease from 50%
flow assigned to h. 50% on B and 30% on D
approaches B&D to 0%. A constant left turning
i. 60% on B and 20% on D
percentage of 25% would be fixed on
III. The dominant traffic flow was in two approaches A and C.
perpendicular directions i.e. on
approaches A and B. The traffic The analyses included the evaluation on the delay
distribution on approaches A and B per vehicle and level of service (LOS) for each
was similar to case II, as shown in traffic volume. Each case was individually analyzed
Table 2. The traffic on the other two for the traffic distribution and scenarios.
approaches had an equal split.
Three Cases (I,II, and III) were illustrated for the
Table 2. Traffic Distribution for Case III experimental design and different cases were
60 % of traffic a. 30% on A and 30% on B considered. For Cases II and II, there were 9
flow assigned to b. 40% on A and 20% on B possible traffic distribution percentages (i.e. a, b, c
approaches A&B c. 50% on A and 10% on B … etc.), as shown in Tables 1 and 2. Then, there
70 % of traffic d. 35% on A and 35% on B were three cases of turning percentages (1, 2, and
flow assigned to e. 45% on A and 25% on B 3). Accordingly, a case having a reference code of
approaches A&B f. 55% on A and 15% on B Case III.b.3, means that the dominant traffic was
80 % of traffic g. 40% on A and 40% on B according to Case III (dominant traffic on two
flow assigned to h. 50% on A and 30% on B perpendicular approaches). The distribution to the
approaches B&D i. 60% on A and 20% on B approaches was b, which indicated 70% on the
dominant approaches with 45% on A, 25% on B,
 The traffic movements on all approaches
while approaches C and D had 15% of the total
were assumed either to go straight or left.
traffic on each. Finally, the left turning percentages
The right turning vehicles were assumed to
followed Scenario 3, which showed an increase in
not use the intersection. Accordingly, the
the left turn percentage on approach D from 0 to
turning percentages on each approach were
50%, decrease in the left turn percentage on
assumed to have the following scenarios:
approach B from 50% to 0%, and the left turn

69
 Journal of Transport System Engineering 1 : 1 (2014), 67-73

movement on approaches A and C were fixed at Table 4. Summary Results for Case I.2 (Traffic
25%. Figure 2 provides a schematic representation Signal)
of the experimental design. Signal Case I.2
2500 3000 3500 4000 4500 5000
LT % C D L D L D L D L D L D L
0% 42 D 51 D 76 E 110 F 153 F 196 F
10% 39 D 42 D 59 E 85 F 116 F 153 F
20% 34 C 37 D 47 D 67 E 87 F 115 F
30% 34 C 37 D 47 D 67 E 87 F 115 F
40% 39 D 42 D 59 E 85 F 116 F 153 F
50% 42 D 51 D 76 F 110 F 153 F 196 F
D: Delay L: Level of Service

Figure 2. Experimental Design

The comparison between traffic signal and

roundabout performances was made based on the
delay per vehicle.

4. RESULTS AND ANALYSIS

Figure 3. Summary Results for Case I.2
As discussed in the previous section, the number of The sample analysis for Case II is presented with
evaluation cases was very large. A sample of the the summary of two cases. The first case is Case
evaluation results is presented in this section. The II.a.2, which represents dominant traffic on
sample had been selected to represent the general approaches B and D with a percentage of 30% on
trend of all cases. The following tables and charts each approach and the left turn percentage of
illustrate the results for Case I.2 (equal split of approach C increased from 0 to 50%. Tables 5 and
traffic on all approaches) and Scenario 2 for left 6 present the summary for the roundabout and
turn percentages (increase in the left turn traffic signal evaluations, respectively. Figure 4
percentage on approach C and reduction in the left shows the combined summary of the roundabout
turn percentage on approach B). Table 3 and traffic signal evaluation.
summarizes the results for roundabout evaluation,
while Table 4 shows the results of a traffic signal Table 5. Summary Results for Case II.a.2
evaluation. Figure 3 provides a combined (Roundabout)
representation of the results for roundabout and RA Case II.a.2
traffic signal evaluation for Case I.2. 2500 3000 3500 4000 4500 5000
%LT C D L D L D L D L D L D L
0% 11 B 18 B 57 E 133 F 192 F 263 F
Table 3. Summary Results for Case I.2 (Roundabout)
10% 10 B 16 B 42 D 109 F 171 F 246 F
RA Case I.2
20% 10 A 15 B 32 C 95 F 161 F 238 F
2500 3000 3500 4000 4500 5000
30% 10 A 14 B 20 C 88 F 155 F 214 F
LT %C D L D L D L D L D L D L
40% 10 A 14 B 28 C 92 F 144 F 202 F
0% 12 B 13 B 29 C 73 E 120 F 183 F
50% 9 A 14 B 30 C 86 F 146 F 254 F
10% 9 A 13 B 24 C 66 E 119 F 177 F
20% 9 A 13 B 23 C 62 E 111 F 165 F D: Delay L: Level of Service
30% 9 A 13 B 23 C 64 E 111 F 165 F
40% 9 A 13 B 24 C 66 E 126 F 180 F
50% 10 B 13 B 29 C 68 E 128 F 196 F
D: Delay L: Level of Service

70
 Journal of Transport System Engineering 1 : 1 (2014), 67-73

Table 6. Summary Results for Case II.a.2

(Traffic Signal)
Signal Case II.a.2
2500 3000 3500 4000 4500 5000
%LT C D L D L D L D L D L D L
0% 35 D 40 D 66 E 104 F 152 F 213 F
10% 33 C 35 D 48 D 80 F 120 F 179 F
20% 31 C 35 D 46 D 64 E 96 F 143 F
30% 34 C 36 D 48 D 65 E 95 F 134 F
40% 32 C 39 D 53 D 75 E 103 F 134 F
50% 37 D 44 D 63 E 91 F 125 F 162 F
D: Delay L: Level of Service

Figure 5. Summary Results for Case II.i.3

For Case III, which represents dominant traffic at

approaches B and D, it was represented by 2 cases,
namely Case III.a.3 and Case III.i.1.

Tables 9 and 10 and Figure 6 illustrate the results

for Case III.a.3, while Tables 11 and 12 and Figure
7 represent the results for Case III.i.1.

Table 9. Summary Results for Case III.a.3

(Roundabout)
Figure 4. Summary Results for Case II.a.2
RA Case III.a.3
2500 3000 3500 4000 4500 5000
The second case is Case II.i.3, which represents a %LT D D L D L D L D L D L D L
dominant traffic of 80% at approaches B and D, 0% 11 B 15 B 34 C 96 F 155 F 214 F
distributed as 60% at B and 20% at D, and the left 10% 11 B 11 B 39 D 100 F 158 F 217 F
20% 11 B 15 B 45 D 117 F 168 F 226 F
turn percentage at D increased from 0 to 50%. The
30% 11 B 16 B 54 D 141 F 178 F 242 F
results of this case are presented in Tables 7 and 8 40% 11 B 18 B 67 E 171 F 205 F 262 F
and Figure 5. 50% 11 B 22 C 83 F 210 F 245 F 295 F
D: Delay L: Level of Service
Table 7. Summary Results for Case II.i.3
(Roundabout) Table 10. Summary Results for Case III.a.3
(Traffic Signal)
RA Case II.i.3
2500 3000 3500 4000 4500 5000 Signal Case III.a.3
%LT D D L D L D L D L D L D L 2500 3000 3500 4000 4500 5000
0% 11 B 22 C 62 E 117 F 179 F 250 F %LT D D L D L D L D L D L D L
10% 10 B 17 B 50 D 109 F 177 F 251 F 0% 38 D 52 D 78 E 111 F 152 F 202 F
20% 10 B 17 B 58 E 122 F 196 F 280 E 10% 33 C 43 D 61 E 87 F 118 F 157 F
30% 11 B 8 A 71 E 142 F 225 F 300 F 20% 31 C 38 D 51 D 72 E 108 F 140 F
40% 11 B 9 A 87 F 166 F 258 F 368 F 30% 30 C 38 D 52 D 74 E 106 F 145 F
50% 12 B 9 A 105 F 192 F 296 F 421 F 40% 32 C 44 D 64 E 96 F 130 F 174 F
50% 36 D 52 D 81 F 119 F 164 F 218 F
D: Delay L: Level of Service
D: Delay L: Level of Service
Table 8. Summary Results for Case II.i.3
(Traffic Signal)
Signal Case II.i.3
2500 3000 3500 4000 4500 5000
%LT D D L D L D L D L D L D L
0% 49 D 56 D 82 F 116 F 111 F 145 F
10% 33 C 46 D 53 D 72 E 67 E 90 F
20% 32 C 38 D 45 D 60 E 54 D 71 E
30% 33 C 39 D 48 D 64 E 68 E 100 F
40% 37 D 54 D 64 E 101 F 106 F 153 F
50% 54 D 61 D 99 F 150 F 153 F 210 F
D: Delay L: Level of Service

71
 Journal of Transport System Engineering 1 : 1 (2014), 67-73

traffic signal performance under all traffic

conditions.

In the cases where there was a predominant traffic

in two opposite directions (i.e. One road with both
directions, Case II, it appears that the roundabout
showed lower delays and better level of service
when the total traffic volume on the intersection
was 3000 veh/hr or less. However, for traffic
volumes around 3500 veh/hr, the roundabout
Figure 6. Summary Results for Case III.a.3 presented a better performance in cases where the
left turn percentage was less than 20%. The traffic
Table 11. Summary Results for Case III.i.1 signal showed better performance for high traffic
(Roundabout) volumes (total traffic volume of 4000 veh/hr or
RA Case III.i.1 more on the intersection).
2500 3000 3500 4000 4500 5000
%LT A D L D L D L D L D L D L
0% 10 B 59 E 180 F 242 F 315 F 386 F
Finally, for the cases where there was high traffic
10% 10 B 60 E 178 F 242 F 315 F 386 F volume on two perpendicular approaches, Case III,
20% 11 B 60 E 178 F 243 F 316 F 387 F the results almost matched those of Case II, except
30% 11 B 60 E 178 F 244 F 316 F 388 F
for total traffic volume of 3500 veh/hr, where the
40% 11 B 61 E 179 F 245 F 317 F 389 F
50% 12 B 63 E 182 F 249 F 320 F 392 F roundabout showed better performance for left turn
D: Delay L: Level of Service
percentage of 30% or less. For low traffic volumes
(3000 veh/hr or less), the roundabout performed
Table 12. Summary Results for Case III.i.1 better while for high traffic volumes (4000 veh/hr
(Traffic Signal) or more), the traffic signal illustrated better
Signal Case III.i.1 performance.
2500 3000 3500 4000 4500 5000
%LT A D L D L D L D L D L D L
6. RECOMMENDATIONS
0% 55 E 65 E 107 F 158 F 156 F 265 F
10% 36 D 53 D 65 E 101 F 102 F 187 F
20% 36 D 41 D 52 D 66 E 66 E 142 F As general recommendations, the following criteria
30% 30 C 49 D 49 D 80 F 82 F 154 F can be used as guide for selecting the type of
40% 38 D 81 F 80 F 134 F 141 F 224 F intersection control are:
50% 59 E 133 F 136 F 212 F 232 F 347 F
1. For low traffic volume on the intersection
D: Delay L: Level of Service
(3000 veh/hr or less), a roundabout is
recommended, regardless of the traffic
distribution on the approaches.
2. If the traffic volume on the intersection is
distributed uniformly (i.e. Equal traffic
volume on all approaches), then a
roundabout is recommended regardless of
the traffic volumes and the turning
percentages. This is in agreement with
some previous studies [9][17]
3. For high traffic volume (4000 veh/hr), a
traffic signal is recommended for all traffic
Figure 7. Summary Results for Case III.c.3.1
distributions and turning percentages. The
5. DISCUSSION OF RESULTS conclusion in this case contradicts with
some previously reported results [9]. The
Considering the results for Case I (equal traffic main reason for the difference is the fact
volume on all approaches), it can be concluded that that the previous work did not consider
the roundabout performance was better than the unbalanced traffic on the approaches.

72
 Journal of Transport System Engineering 1 : 1 (2014), 67-73

4. In case of moderate traffic volume Controls, Delaware Center for Transportation, DCT
(between 3000 veh/hr to 4000 veh/hr), 199, University of Delaware, Newark, Delaware.
special attention should be paid to the left [9] Sisiopiku, V. & Oh, H. (2001). Evaluation of
Roundabout Performance Using SIDRA. Journal of
turn percentage before deciding on the
Transportation Engineering,127, 2, 143-150.
type of control for such intersection.
[10] Ma, W., Liu Y., Head, L., & Yang, X. (2013).
It is recommended to consider the following points Integrated Optimization of Lane Markings and
for future research: Timings for Signalized Roundabouts. Transportation
1. Consider having different layout of the Research Part C, 36, 307-323.
intersection (i.e. More than two lanes per [11] Azhar, A, & Svante, B. (2011). Signal Control of
approach). Roundabouts. Proceedings of the 6th International
2. Utilize micro simulation models to Symposium on Highway Capacity and Quality of
confirm the results obtained from this Service Stockholm, Sweden.
[12] Natalizio, E. (2005). Roundabouts with Metering
study.
Signals, Paper presented at the Institute of
3. Consider the case for signalized
Transportation Engineers Annual Meeting,
roundabouts as a third option for the Melbourne, Australia.
intersection control. [13] Stevens, C. (2005). Signals and Meters at
Roundabouts, Proceedings of the Mid-Continent
7. ACKNOWLEDGEMENT Transportation research Symposium, Ames, Iowa.
[14] Davies, P., Jamieson, B., Reid, B. (1980). Traffic
The authors would like to acknowledge the support Signal Control of Roundabouts. Traffic Engineering
from the Faculty Research Grants (FRG) program & Control, 21, 7, 354-357.
at the American University of Sharjah (AUS) for [15] Weihua, Z. & Ruijuan, D. (2009). Setting Conditions
providing the funding for this research project. of Roundabout under Different Signal Control
Also, the authors would like to acknowledge the Methods. Journal of Southeast University, 39, 2, 389-
participation of AUS students Ahmed Gamal and 393.
[16] Chaudhary, N. & Songchitruksa, P. (2008). Traffic
Yusri Al Yaarubi in conducting the software runs.
Adaptive Signal Control at Roundabouts.
Proceedings of the 4th International Gulf Conference
8. REFERENCES on Roads, 141-150.
[17] Ben-Edigbe, J & Minhans, A. (2011). Influence of
[1] Pochowski, A. (2010). An Analytical Review of
Statewide Roundabout Programs and Policies. Civil Composite Traffic Control Mechanisms on Four-Arm
and Environmental Engineering, Georgia Institute of Highway Intersection, Proceedings, Annual
Technology. Conference of the Canadian Society for Civil
[2] Tanner, J. (1962). A Theoretical Analysis of Delays Engineering, CSCE, Vol. 1 pp 195-203, Ottawa,
at an Uncontrolled Intersection. Biometrika, 49, 163– Ontario, Canada.
170.
[3] Blackmore, F. (1963). Priority at Roundabouts.
Traffic Engineering & Control, 5, 6, 104–106.
[4] McDonald, M. & Noon, C. (1978). Delays at
Roundabouts Caused by Geometric Design Factors.
Journal of the Institution of Highway Engineers, 25,
12, 9–13.
[5] Country Roads Board, CRB (1979). Guidelines for
the Design and Installation of Roundabouts.
Technical Bulletin No. 30, Sydney, Australia
[6] Akçelik, R. (1997). Lane-by-Lane Modelling of
Unequal Lane Use and Flares at Roundabouts and
Signalised Intersections: The SIDRA Solution. Traffic
Engineering & Control, London, 38, 7/8, 388-399.
[7] Akçelik, R., & Besley, M. (1996). SIDRA 5 user
guide, ARRB Transport Research, Ltd., Vermont
South, Sydney, Australia.
[8] Mensah, S. & Eshragh, S. (2010). Use of
Roundabouts as Alternatives to All-Way Stop

73