Research Article

Transportation Research Record

1–13
Ó National Academy of Sciences:
Vehicle and Pedestrian Level of Service Transportation Research Board 2020

in Street Designs with Elements of Article reuse guidelines:

sagepub.com/journals-permissions
Shared Space DOI: 10.1177/0361198120933627
journals.sagepub.com/home/trr

Ioannis Kaparias1 and Rui Wang1

Abstract
Inspired by developments in urban planning, the concept of ‘‘shared space’’ has recently emerged as a way of creating a better
public realm. This is achieved through a range of streetscape treatments aimed at asserting the function of streets as places
by facilitating pedestrian movement and lowering vehicle traffic volumes and speeds. The characteristics of streets with ele-
ments of shared space point to the conjecture that traffic conditions and road user perceptions may be different to those on
streets designed according to more conventional principles, and this is likely to have an impact on the quality of service. The
aim of this paper is, therefore, to perform an analysis in relation to level of service (LOS) and to investigate how this may
change as a result of the implementation of street layouts with elements of shared space. Using video data from the
Exhibition Road site in London during periods before and after its conversion from a conventional dual carriageway to a lay-
out featuring several elements of shared space, changes in relation to LOS for both vehicle traffic and pedestrians are investi-
gated, by applying the corresponding methods from the 2010 Highway Capacity Manual. The results suggest that streets with
elements of shared space provide a much improved pedestrian experience, as expressed by higher LOS ratings, but without
compromising the quality of vehicle traffic flow, which, in fact, also sees slight improvements.

The concept of ‘‘shared space’’ has gained global traction varying extents of shared space elements can be found
in recent years as a novel approach to the design of around the world and include: the concepts of ‘‘woonerf’’
urban streets, where both pedestrians and vehicles are and ‘‘home zone’’ in residential areas in the Netherlands
present (Figure 1). At its core, shared space aims at and UK respectively; the ‘‘Manual for Streets’’ approach
asserting the function of streets as places rather than as in the UK (6, 7); and the ‘‘Complete Streets’’ initiative in
arteries, and this entails designing layouts geared for eas- the U.S.A. (8).
ier pedestrian movement and lower vehicle speeds. As The characteristics of streets with elements of shared
such, shared space contrasts with the traditional car- space point to the conjecture that traffic conditions and
oriented approach, which is based on greater segregation road user perceptions on such streets may be different
of pedestrians and vehicles to ensure unobstructed traffic than those on streets designed according to more conven-
flow (1). tional principles. Indeed, it is aspired, through shared
Elaborating more on the term ‘‘shared space,’’ conver- space designs, to encourage high levels of street sharing,
sely to popular belief, this is not used to characterize which translates to enabling pedestrians to move more
entire streets as ‘‘shared’’ or ‘‘not shared,’’ particularly freely around the street and use parts of it that would
given that streetscape design cannot be standardized and otherwise be dedicated to vehicle traffic. Naturally,
needs to be context-sensitive. Instead, shared space is higher levels of sharing are likely to affect the quality of
used as an ‘‘umbrella’’ term to refer collectively to a service offered, as most recently defined in the 2016
range of streetscape treatments, aiming at creating a Highway Capacity Manual (HCM 2016) (9), and it can be
more pedestrian-friendly environment. These may range
from the removal of guardrails and the introduction of
1
‘‘informal’’ (uncontrolled) pedestrian crossing facilities in Transportation Research Group, Faculty of Engineering and Physical
Sciences, University of Southampton, Southampton, UK
a traditional ‘‘kerbed’’ street layout, through to layouts
with a single surface and little delineation between pedes- Corresponding Author:
trian and vehicle areas (2–5). Examples of streets with Ioannis Kaparias, [email protected]
2 Transportation Research Record 00(9)

Figure 1. Examples of shared space: (left) New Road, Brighton, UK and (right) The Wharf, Washington, DC, U.S.A.

hypothesized that significant increases in pedestrian qual- various steps involved, and elaborates on the practical
ity of service may be accompanied by minor decreases in steps of the analysis. The fifth section presents the results
vehicle traffic capacity. Nevertheless, to the best of the and discusses the trends in vehicle and pedestrian quality
authors’ knowledge, this hypothesis has yet to be verified, of service in relation to elements of shared space. The
as a comprehensive assessment of the level of service final section concludes the paper and identifies areas of
(LOS) of streets designed according to the principles of future research.
shared space has not been carried out.
The aim of the present study is, therefore, to evaluate
the quality of service for both vehicle traffic and pedes-
Background
trians, and specifically to investigate how the respective The LOS method, according to HCM 2010 (10) and
LOS ratings may change as a result of the implementa- HCM 2016 (9), is designed to translate complex road
tion of street layouts with elements of shared space. The performance and experience factors into a single indica-
analysis is carried out at the Exhibition Road site in tor that reflects the quality of service perceived by road
London’s South Kensington area, with data coming users. The indicator is calculated using mathematical
from video observations during periods before and after models based on various service measures, and consists
its conversion from a conventional divided roadway to a of a six-level evaluation system, with scores ranging from
single surface featuring several elements of shared space. A (best conditions) to F (worst conditions). The LOS
The relevant vehicle LOS (VLOS) and pedestrian LOS method is used by operational agencies to assess whether
(PLOS) methods from the Highway Capacity Manual the performance of road facilities is satisfactory and also
2010 (10) are used, as opposed to the most recent HCM whether changes in the future would tend to be sup-
2016 (9), as much of the analysis pre-dates the publica- ported by the general public. It also provides guidance
tion of the latter. The work complements previous for transport engineers and planners to predict facility
research on the topic, which assessed the behavior and operating conditions from a planning level. Given that
perceptions of road users in street designs with elements LOS is a simplified and straightforward indicator, it can
of shared space by looking at pedestrian–vehicle traffic then function as the basis of decision making for policy
conflicts (11, 12) and behavioral interactions (13), will- makers who may not necessarily have the relevant tech-
ingness of pedestrians and drivers to share (14), and nical background.
pedestrian gap acceptance behavior (15). The LOS concept was originally designed to assess the
The present paper is structured as follows. The second quality of service of highway and street facilities from
section presents the background of the study, focussing the viewpoint of motorized traffic. The respective VLOS
primarily on the LOS concept and on previous research method was initially presented in the 2nd edition of the
related to the capacity of shared space. The third section Highway Capacity Manual in 1965 (16) and has seen sev-
goes on to introduce the study site and to report on the eral revisions since then until the current 6th edition
data collection carried out. The fourth section then out- (HCM 2016) (9). It considers the speed, traffic density,
lines the VLOS and PLOS methods, describes the travel time, and delays for vehicle traffic to provide an
Kaparias and Wang 3

overall indicator of how drivers perceive the traffic con- brings about lower traffic flows and speeds, which may
dition and road environment. It has been extensively result in less congestion, and this may mean improved
applied in numerous case studies in the U.S.A. and inter- VLOS scores. Indeed, there is empirical evidence of
nationally, and has formed the basis of the respective shared space features having had a positive impact on
manuals of many other countries (e.g., Germany [17]). traffic efficiency, such as, for example, Laweiplein,
In the later versions of the HCM the LOS concept Drachten, in the Netherlands, and Svallertorget,
was extended to pedestrians, enabling the assessment of Norrköping, in Sweden (4, 26–28). More recent systema-
the quality of service provided on walkways, pedestrian tically obtained evidence from shared and conventionally
crossings, and other types of pedestrian facilities, reflect- designed junctions in five countries, including the
ing pedestrians’ perceptions of the surroundings and of U.S.A., has also suggested improvements in relation to
the traffic conditions. The relevant PLOS method, first both pedestrian and vehicle traffic junction delays (29).
introduced by Fruin (18) and refined thereafter, quanti- Therefore, improved VLOS performance is a potential
fies the performance on the basis of complex factors that outcome that cannot be dismissed.
may influence the pedestrian movements and walking The present study extends existing knowledge by
experience, and, just like VLOS, delivers a score ranging investigating VLOS and PLOS changes through video
from A to F. Being a more recent addition to the HCM observation of a street site before and after its redevelop-
methods, PLOS has seen fewer applications than VLOS, ment as a layout with elements of shared space, and the
with perhaps the most prominent example having been a next sections outline how this is done.
2006 case study in New York City (19). In that study, the
criteria of the average pedestrian space, flow rate, walk-
ing speed, and volume-to-capacity ratio were applied, Data Collection and Processing
and the outcome was a universal standard for the evalua- The data collection and processing steps are outlined in
tion of different types of pedestrian facilities, which could this section. This includes a description of the site and is
be applied in other cities with different characteristics followed by an account of the data collection methods
and required only limited input data. Nevertheless, the and tools employed. Then, the processing of the data col-
application focussed predominantly on the geometrical lected is explained, in preparation for the reporting of
features and pedestrian traffic conditions, and devoted the analysis methodology and results in the following
little attention to the interaction of pedestrians with other sections.
road users.
Refinements of the PLOS method (e.g., 20–22) have
been included in HCM 2010 and HCM 2016, and have Site Description
thus provided the ability to consider the role of pedes- Exhibition Road is an 800 m long road located in the
trians in multimodal traffic and to evaluate their percep- Royal Borough of Kensington and Chelsea (RBKC) in
tion of the presence of other road users. In a recent study London and is home to several London’s most popular
(23) the PLOS method was applied on a mixed lane envi- museums (Natural History, Science, V&A). The sur-
ronment and the PLOS index was related with several rounding area of South Kensington is well known as a
influencing factors by means of regression analysis. It cultural center, including other venues, such as the Royal
was found that, in addition to the conditions of pedes- Albert Hall, as well as many academic institutions, such
trian movement (flow, speed, etc.), PLOS was also influ- as Imperial College London. As the previous conven-
enced by the levels and patterns on vehicle traffic and tional divided roadway layout of Exhibition Road was
roadside parking. But while mixed lanes bear some simi- crowded (a problem exacerbated by numerous pedestrian
larity with shared space schemes, there are fundamental barriers) and dominated by high traffic flows and parked
differences, in that the former essentially ‘‘force’’ pedes- vehicles, the RBKC undertook an engineering scheme
trians to share some of the space with vehicle traffic, that included redevelopment featuring several elements
while the latter are designed to encourage them to do so. of shared space (Figure 2).
It can be, therefore, anticipated, in line with related find- The project was implemented over four years from
ings in the field so far (e.g., 24, 25) that shared space fea- mid-2008 to completion in late 2011. The following three
tures will have a positive contribution to PLOS ratings. main streetscape treatments were carried out:
The impact of shared space on VLOS, on the other
hand, is less clear. On the one hand, shared space aims at 1. Re-allocation of street space (Figure 2a): The pre-
improving the pedestrian environment while at the same vious layout of the 24 m wide Exhibition Road
time acting as a deterrent to vehicle traffic, and from that consisted of a 16 m wide divided roadway, accom-
viewpoint VLOS ratings would be expected to worsen. modating one lane of traffic in each direction as
On the other hand, however, shared space typically well as ample excess width allocated to parked
4 Transportation Research Record 00(9)

Figure 2. Exhibition Road (left) before and (right) after redevelopment.

vehicles, and two 4 m wide footpaths on either was shifted to the eastern side of the road to
side of the carriageway, accommodating pedes- occupy an 8 m wide two-lane road (termed the
trians. As a result of the redevelopment, traffic ‘‘traffic zone’’), with the former western side of
Kaparias and Wang 5

the divided roadway becoming a so-called ‘‘transi- with a wide (12 m) straight-across crossing, allow-
tion zone’’ accommodating primarily pedestrians, ing pedestrians to complete their crossing in a sin-
but also parking, cycles, and coaches stopping to gle phase. The scheme also included the removal
drop-off or pick-up passengers. The two 4 m foot- of pedestrian guardrails and other street clutter
paths remained in place and formed the so-called to further facilitate pedestrian movement.
‘‘pedestrian zone.’’ The space also saw the
removal of the kerbs and the implementation of
an end-to-end single surface, with 800 mm tactile Data Collection
paving strips delineating the pedestrian zone from Video footage has been collected through high-mast
the traffic and transition zones respectively. cameras for periods before and after the redevelopment
2. Unravelling of a one-way system (Figure 2, b and as part of recent studies analyzing traffic conflicts, beha-
c): In the original layout, a one-way system was vioral interactions, and gap acceptance in the area (11–
in place around the South Kensington Station 13, 15). This has also been complemented by vehicle traf-
area, whereby the southbound traffic was led fic and pedestrian counts. In this study, the data col-
along the southern tip of Exhibition Road and lected is used to assess the impact of the new design of
along Thurloe Street, while the northbound traf- Exhibition Road on the quality of service, as expressed
fic was guided along Thurloe Place. As a result of by VLOS and PLOS ratings. In the before case, the data
the redevelopment, Thurloe Place was converted refers to August 2008, before the start of the redevelop-
to a two-way street, accommodating both the ment works, and has been collected from several key
northbound and the southbound traffic, while locations around the site. For the after situation, the
Thurloe Street was converted to an access-only video footage comes from the same locations in October
street. 2011, following the completion of the scheme.
3. Re-design of pedestrian crossing facilities (Figure The locations are the following (Figure 3):
2d): At the intersection of Exhibition Road with
Cromwell Road, the original design included a  L1: Exhibition Road main body (Before: Cameras
staggered north–south pedestrian crossing on the A & B—After: Cameras 4, 5, 6 & 7):
western side of the site, which, however was not In the original layout, pedestrians were confined
following the desire-lines and required pedes- in the two 4 m wide sidewalks on either side of the
trians to cross in two stages, thus resulting in a divided roadway layout. These, however, were
high number of jaywalkers. The redevelopment often crowded by queuing visitors waiting to enter
removed the staggered crossing and replaced it the various museums, thus acting as a barrier to

Figure 3. Camera locations at the Exhibition Road site in the (left) before and (right) after monitoring.
6 Transportation Research Record 00(9)

pedestrian movement. At the same time, vehicle  12:00–13:00 (midday, when many tourists enter
traffic was often interrupted by buses stopping in and exit the museums); and
front of the museums. These issues have been  17:00–18:00 (evening peak hour, with tourists and
addressed in the new layout through the provision workers leaving the area, and locals returning).
of more pedestrian space, but also through the
establishment of the transition zone, in which For the before case, the analysis concentrates on
buses can now stop for passengers to board or Thursday, August 27, 2008, while for the after case it
alight. focuses on Thursday, October 27, 2011. Therefore, a
 L2: Cromwell Road junction (Before: Cameras C total of 18 h of video footage are processed by a single
& D—After: Cameras E, F, G & H): observer with respect to quality of service evaluation fol-
In the original layout, the facilities provided to lowing the HCM. For each of the locations, VLOS and
pedestrians wishing to cross Cromwell Road to PLOS levels are calculated for both the before and the
continue walking on either the eastern or the west- aftercase, so each hour of footage is analyzed twice,
ern sidewalks of Exhibition Road were two stag- bringing the total number of analysis hours to 36. As,
gered pelican crossings, which required a detour however, pedestrian flows are lower in the morning peak
and often long waiting times for a ‘‘green man’’ hour, PLOS is only analyzed for the midday and evening
signal. As a result, the vast majority of the pedes- periods. Also, as Thurloe Street (L3) has become an
trians used ‘‘shortcuts,’’ bypassing the staggered ‘‘access-only’’ street post-redevelopment and no longer
crossings and jaywalking, thus coming into con- accommodates through vehicle traffic, only PLOS is con-
flict with right-turning southbound traffic from sidered at that location and so the relevant footage is
Exhibition Road in the case of the western cross- only analyzed once. As such, the total number of hours
ing, or with left-turning southbound traffic in the analyzed is 24 (L1 VLOS and PLOS, L2 VLOS and
case of the eastern crossing. The western crossing PLOS, L3 PLOS only). The analysis follows the metho-
has been replaced by a wide straight-across cross- dology outlined in the next section.
ing in the new layout, while the eastern one has
been retained but redesigned.
 L3: Thurloe Street (Before: Camera F—After: Analysis Methodology
Cameras 1 & 2): The methodology adopted includes the processes
Pedestrians using this location in the original lay- involved in the assessment of the quality of service
out were faced with two problems: the lack of ade- offered by a street layout with respect to vehicle traffic
quate pedestrian crossing facilities, and the (VLOS) and pedestrians (PLOS). It is noted that the rele-
insufficient space for pedestrians on the southern vant methods from HCM 2010 (10) are used, as opposed
sidewalk of the road. In the new layout, this loca- to HCM 2016 (9), as large parts of the analysis were car-
tion has been redesigned as ‘‘access-only,’’ giving ried out before the publication of the latter.
more space to pedestrians.

Quality of Service for Vehicle Traffic (VLOS)

Data Processing
Following HCM 2010 (10), VLOS in the present study is
The performance of a road facility can be varying at dif- evaluated at the segment and the intersection levels. A
ferent days and times of the day, depending on the segment refers to the combination of a link and one or
changes in vehicle traffic and pedestrian volumes. For more of its boundary intersections, and segment-based
example, a street may perform satisfactorily (LOS D or VLOS is calculated separately for the different directions
greater) during low-volume periods, but the quality of of traffic. VLOS is evaluated according to the combined
service may drop to unsatisfactory levels (LOS E or F) performance of the link and of its downstream boundary
during times of high traffic volumes. Clearly, the latter is intersection. As such, based on the availability of the
a more critical condition, as it is these periods of under- videos from the entire site, one segment is considered,
performance that any design aims to address. and this is the southbound traffic on the main body of
In this study, periods of weekday peak vehicle traffic Exhibition Road (L1) leading into the junction with
and pedestrian flows are chosen for the analysis, and this Cromwell Road (L2). In addition, the VLOS evaluation
results in 3 h of analysis per location for the before and also covers the intersection of Exhibition Road and
after cases. These are: Cromwell Road itself (L2).
The segment and intersection analyzed are shown in
 08:00–09:00 (morning peak hour, with high vehicle Figure 4, and the process of the determination of the rel-
traffic flows); evant VLOS ratings is outlined in Table 1.
Kaparias and Wang 7

Table 1. Intersection and Segment VLOS Calculation Process, Adapted from HCM 2010 (10)

Intersection Segment

1. Lane group saturation flow (s), calculated by applying 1. Through movement volume-to-capacity ratio (V/C),
certain adjustment factors to a base value, according taken from Step 3 of the intersection VLOS calculation
to traffic and geometric characteristics measured at the site. for the relevant through movements. If there are
multiple lane groups, the weighted average V/C is calculated.
2. Lane group g/c ratio, that is, the ratio of the green time (g) to 2. Through movement control delay (dt), taken from Step 4 of the
the cycle length (c) of each lane group. As traffic signals may be intersection VLOS calculation for the relevant through
adaptive and the programs may vary, averages of the sequences movements. If there are multiple lane groups, the weighted
shown by the signals during the observation period are taken. sum of their control delays according to volume is calculated.
3. Lane group volume-to-capacity ratio (X = V/C), where V is the 3. Base free-flow speed (Sf0), as measured at
hourly demand traffic flow in the subject lane group, as the site at traffic conditions of low
measured at the site, and C is the lane group capacity, calculated volume and no impedances.
as , C ¼ N  s  gc , N = number of lanes in the group.
4. Lane group control delay (d), calculated as 4. Segment vehicle travel speed (ST,seg),
d = d1  PF + d2 + d3 , where 3600L
calculated as ST;seg = ,
0:5cð1gÞ 5280ðtR + dt Þ
 d1 ¼ 1minð1;Xc Þg is the uniform delay;
c where L is the length of the segment
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi and tR is the segment running (travel)
 d2 ¼ 900  T  ðX  1Þ þ ðX  1Þ2 þ 4X CT time, as measured on site.
is the incremental delay, with T being the
duration of the analysis period;
 d3 is the initial queue delay, where d3 = 0 unless
unmet demand is carried over between analysis periods; and
 PF is the so-called progression adjustment factor,
which is dependent on the g/c ratio.
5. Approach and intersection control delays (dA and dI), 5. Segment vehicle travel speed percentage
calculated as weighted averages of the control delays (SVTSP) of the base free-flow speed, calculated as
S
of the approach’s or intersection’s lane groups, SVTSP = T,Sf seg
0
 100.
weighted by the relevant traffic volumes.
6. VLOS rating, determined by comparing the calculated 6. VLOS rating, determined by comparing
dI value with the following bands: the calculated SVTSP value with the following bands:
dI (s/veh) VLOS SVTSP (%) VLOS
\10 A .85 A If V/C .1, then VLOS = F
10–20 B 67–85 B regardless of the SVTSP value.
20–35 C 50–67 C
35–55 D 40–50 D
55–80 E 30–40 E
.80 F \30 F

Note: HCM = Highway Capacity Manual; VLOS = vehicle level of service.

Starting from the intersection, this is signal controlled,

and according to HCM 2010, the VLOS rating is deter-
mined according to the control delay. This is calculated
as the average of the average delay per vehicle in each of
the lane groups in all of the intersection’s approaches,
weighted by the relevant traffic volumes. The resulting
value is then compared with the control delay threshold
bands provided in HCM 2010 to assign a VLOS rating.
Lane groups are, therefore, the basic unit of analysis of
the intersection VLOS evaluation. In the before case,
there are nine lane groups, of which three are exclusive
through traffic (one two-lane and two single-lane), five
are shared through traffic and left turn, and one is exclu-
Figure 4. (Left) The segment and (right) intersection considered sive right turn (all single-lane). In the after case, and fol-
in the VLOS evaluation of Exhibition Road. lowing several turn bans and other traffic management
Note: VLOS = vehicle level of service. changes (see Figure 2c), there are only four lane groups,
8 Transportation Research Record 00(9)

three of which are exclusive through traffic (two two-

lane and one single-lane) and one is shared through traf-
fic and left turn (single-lane). It should be noted that all
turning movements are operated in a ‘‘permitted’’ mode,
which means that the turning vehicles must yield to any
semi-compatible/conflicting traffic and pedestrian streams.
For the segment VLOS evaluation, two decision vari-
ables come into play: the volume-to-capacity ratio and the
segment vehicle travel speed percentage. The former is cal-
culated as part of the VLOS evaluation of the downstream
boundary intersection and reflects the ability of the seg-
ment to serve the traffic using it: if the ratio is larger than 1
(i.e., if demand exceeds capacity), then a VLOS rating of F Figure 5. The segments considered in the PLOS evaluation of
applies by default. The latter variable, on the other hand, Exhibition Road.
expresses the travel speed of vehicles on the segment, as Note: PLOS = pedestrian level of service.
measured on site, expressed as a percentage of the free-flow
speed. It is estimated on the basis of the observed travel
time and of the relevant control delay obtained from the vehicle traffic, PLOS is determined based on the average
downstream intersection VLOS calculation, and the result- pedestrian space measure only.
ing value is then compared with the respective threshold
bands provided to assign a VLOS rating.
Results
The video data collected are processed and analyzed
Quality of Service for Pedestrians (PLOS) using the method described in Section 4, and the results
The PLOS evaluation in the present study concentrates are presented in this section.
on four walkways around the Exhibition Road site, and
specifically both sidewalks of the main body of Exhibition
Road (L1), including the crossings to Cromwell Road VLOS Assessment
(L2), and both sidewalks of Thurloe Street (L3). The The results of the VLOS evaluation in the Exhibition
Exhibition Road sidewalks are assessed at the segment Road site before and after redevelopment to a design
level, where each segment consists of a link and a cross- with elements of shared space are shown in Table 3, in
ing, while the two Thurloe Street sidewalks are evaluated which part (a) reports the evaluation of the intersection
at the link level for the before case. In the after case, as and part (b) that of the segment.
Thurloe Street has been re-developed to an access-only Looking at part (a) of Table 3, it can be observed that
street, it is assessed as an off-street facility. The PLOS rat- the quality of service for vehicle traffic at the junction of
ing determination process is outlined in Table 2, and the Exhibition Road with Cromwell Road pre-redevelopment
walkways analyzed are shown in Figure 5. for each of the morning, midday, and evening observation
According to HCM 2010, the PLOS rating on a seg- periods had VLOS ratings of A, B, and B, respectively,
ment is determined on the basis of two decision variables: and that post-redevelopment this has improved to A in all
the average pedestrian space and the pedestrian perception three periods, with intersection control delays dropping to
score. With respect to the former, this reflects the average well below 10 s. With respect to the individual junction
amount of road space, as measured on site, that is available arms, VLOS on the two Cromwell Road approaches
per pedestrian in relation to their ability to continue to appears to be largely unaffected by the junction re-design
walk along their desired path without altering their pace or (and in the case of the eastern approach during the eve-
course. As concerns the latter, this is an aggregate measure ning peak, even slightly improved), with the measured
combining many different factors reflecting the pedestrian’s control delays remaining at the same low values as before
perception of the walking experience. It is estimated on the the redevelopment and affecting traffic flows of similar or
basis of the geometric characteristics, the various delays slightly higher levels. At the same time, vehicle traffic on
experienced by pedestrians and the difficulty of crossing the Exhibition Road approaches appears to experience
the roadway. The resulting values are then compared with slightly longer control delays overall, but these affect sig-
the average pedestrian space and perception score thresh- nificantly lower traffic flows.
old bands provided in HCM 2010 to assign a PLOS rating. Elaborating on the reasons behind these effects, a key
When it comes to the evaluation of off-street pedes- determining factor of the quality of service of the junc-
trian facilities, and given the absence of interaction with tion is the traffic signal program. Being adaptive, it has
Kaparias and Wang 9

Table 2. Segment, Link, and Off-Street Facility PLOS Calculation Process, Adapted from HCM 2010 (10)

Segment, link, and off-street facility

1. Average pedestrian space (Ap), calculated as Ap = 60Sp =Vp , where

 is the pedestrian flow per unit width, which depends on the hourly pedestrian flow on the walkway, Vped, and on the effective
sidewalk width WE, that is, the actual walkway width available to pedestrians,
 as measured
 on site; and
 Sp is the average pedestrian walking speed, which is estimated as Sp = 1  0:00078Vp 2  Spf on the basis of the pedestrian flow Vp
and on a free-flow pedestrian walking speed value of 4.4 ft/s (1.34 m/s).

Segment and link

2. Pedestrian perception score for link (Ip,link), calculated as Ip, link = 6:0468 + Fw, link + Fv, link + Fs, link , where Fw,link, Fv,link, and Fs,link are the
cross-section, vehicle traffic volume, and vehicle traffic speed adjustment factors, respectively, and whose values are determined on
the basis of the geometric features and surrounding traffic conditions of the link.

Segment
3. Pedestrian delays, namely:
 Pedestrian crossing delay (dpc), that is, the delay when waiting to cross a signalized junction, calculated as dpc = ðc  gwalk Þ2 =2c,
for cycle time c and pedestrian green time gwalk;
 Pedestrian diversion delay (dpd), that is, the extra time incurred because of having to divert from the desire-line to cross at a
signalized junction, with dpd = Dd =Sp + dpc , for total diversion distance Dd; and 
 Average pedestrian waiting delay (dpw), calculated as dpw = V1  eVtc  V  tc  1 , where V is the traffic flow along the road being
crossed and tc is the critical gap, estimated as tc = L=Scp + ts on the basis of the road width L, the average pedestrian crossing
speed Scp, taken as 3.5 ft/s (1.07 m/s) and the total pedestrian start and end clearance time, taken as 3 s.
4. Pedestrian perception score for intersection (Ip,int), with Ip, int = 0:5997 + Fw, int + Fv, int + Fs, int + Fd, int , where Fw,link, Fv,link, Fs,link, and
Fd,link are the cross-section, vehicle traffic volume, vehicle traffic speed, and pedestrian delay adjustment factors, respectively, and
whose values are determined on the basis of the geometric  features
 and surrounding traffic conditions of the intersection.
 
1
5. Roadway crossing difficulty factor (Fcd), Fcd = 1 + 7:5  0:1dpx  0:318Ip;link + 0:22Ip;int + 1:606 , where dpx = min dpd , dpw , 60 .
 
6. Pedestrian perception score for segment (Ip,seg), with Ip;seg = Fcd  0:318Ip;link + 0:22Ip;int + 1:606 .

Segment and link Off-street facility

7. PLOS rating, determined by comparing 7. PLOS rating, determined by comparing the
the calculated Ap and Ip,seg (or Ip,link) with the bands: calculated Ap value with the bands:

Ip,segorIp,link Ap(ft2/ped) Ap(ft2/ped) PLOS

.60 40–60 24–40 15–24 8–15 \8 .60 A

\2.00 A B C D E F 40–60 B
2.00–2.75 B B C D E F 24–40 C
2.75–3.50 C C C D E F 15–24 D
3.50–4.25 D D D D E F 8–15 E
4.25–5.00 E E E E E F \8 F
.5.00 F F F F F F

Note: HCM = Highway Capacity Manual; PLOS = pedestrian level of service.

responded to the changes in traffic volumes post- etc.). This results in slightly inferior, but still satisfactory
redevelopment through longer cycle times and green time (B or C), VLOS scores post-redevelopment for the
allocations to the Cromwell Road approaches than the Exhibition Road approaches.
Exhibition Road ones, and this results in overall longer Considering the VLOS segment evaluation results in
waiting times for drivers on the latter. However, the ban part (b) of Table 3, it can be observed that the quality of
of all right and most left ‘‘permitted’’ turns at the junc- service of vehicle traffic pre-redevelopment had a rating
tion (with relevant flows having been re-routed else- of D in the morning and midday periods, and that post-
where) appears to compensate for these effects and redevelopment this has improved to C; in the evening
ensure that individual cycle failures (which occurred period, the VLOS rating was C before, and has remained
occasionally for the southbound right-turning traffic the same after. This is despite the increases in both con-
before the redevelopment) no longer occur, despite the trol delay (resulting from the junction changes) and seg-
more pedestrian-oriented, and therefore more con- ment running time (as a result of the more pedestrian-
strained, road layout (fewer lanes, narrower roadway, friendly layout), which have resulted in lower calculated
10 Transportation Research Record 00(9)

Table 3. VLOS Evaluation Results in the Exhibition Road Site before and after Redevelopment

Morning Midday Evening

(a)
Before After Before After Before After

Northern approach (Exhibition Road)

Traffic volume (veh/h) 477 88 589 203 744 144
Approach delay (s/veh) 20.3 18.6 24.2 30.2 12.6 21.3
VLOS C B C C B C
Western approach (Cromwell Road)
Traffic volume (veh/h) 1008 1328 997 1008 988 1074
Approach delay (s/veh) 3.9 5.9 6.4 4.2 8.3 5.2
VLOS A A A A A A
Southern approach (Exhibition Road)
Traffic volume (veh/h) 332 64 308 118 338 136
Approach delay (s/veh) 14.5 18.1 12.0 20.9 8.5 19.6
VLOS B B B C A B
Eastern approach (Cromwell Road)
Traffic volume (veh/h) 880 916 929 945 1116 1060
Approach delay (s/veh) 5.2 4.0 8.2 3.8 14.4 4.9
VLOS A A A A B A
Intersection TOTAL
Intersection delay (s/veh) 8.6 5.9 11.3 7.2 11.5 6.8
VLOS A A B A B A

Morning Midday Evening

(b)
Before After Before After Before After

Traffic volume (veh/h) 477 88 589 203 744 144

Control delay (s/veh) 20.3 18.6 24.2 30.2 12.6 21.3
Running time (s) 25.7 35.4 25.9 35.9 26.4 35.7
Travel speed (mph) 14.1 12.0 12.9 9.8 16.6 11.4
Travel speed (km/h) 22.7 19.3 20.8 15.8 26.7 18.3
Travel speed % 49.8% 61.8% 45.7% 50.5% 58.7% 58.6%
Volume-to-capacity (V/C) ratio 0.37 0.18 0.34 0.61 0.57 0.37
VLOS D C D C C C

Note: Base free-flow speed = Before: 28.3 mph (45.5 km/h); After: 19.4 mph (31.2 km/h). veh/h = vehicles per hour; VLOS = vehicle level of service.

average travel speeds, as these have been accompanied borderline satisfactory quality of service and are likely to
with a reduction in the estimated base free-flow speed. have negative experiences about some elements when
As such, the travel speed percentages have remained the walking along the segment, such as long delays at the
same, or have even increased post-redevelopment. This intersection and lack of adequately long gaps for crossing
finding confirms the hypothesis that shared space may at mid-segment locations. This is reflected in the negative
bring about traffic efficiency benefits, as even though the (high) pedestrian perception scores for the intersection,
new layout may act as a deterrent to vehicle traffic and especially in the western sidewalk, as well as in the high
may result in lower volumes and speeds, the drivers that values of the crossing difficulty factor.
do use the segment benefit from less congestion and Following the redevelopment, however, it can be seen
therefore improved conditions. that the quality of service has improved significantly,
with the relevant PLOS rating rising to level B in both
walkways. In the western sidewalk this can be attributed
PLOS Assessment to the reduction in vehicle traffic volume and speed, to
The results of the PLOS evaluation in the Exhibition Road the re-design of the relevant Cromwell Road crossing
site pre- and post-redevelopment are shown in Table 4. from staggered to straight-across (resulting in improved
Looking at the western and eastern sidewalks of the link and intersection perception scores), but also to the
Exhibition Road main body, it can be observed that in greater ease in crossing provided by the new layout, as
both walkways and for the two observation periods ana- expressed by the reduction of the crossing difficulty fac-
lyzed the quality of service is rated as D pre-redevelop- tor. A similar trend is observed in the eastern sidewalk,
ment. This may indicate that pedestrians receive a even though the eastern Cromwell Road staggered
Kaparias and Wang 11

Table 4. PLOS Evaluation Results in the Exhibition Road Site before and after Redevelopment

Midday Evening
Before After Before After

Exhibition Road (western sidewalk)

Pedestrian volume (ped/h) 1732 2132 1835 2192
Average pedestrian space (ft2/ped) 67.8 122.1 63.9 118.7
Average pedestrian space (m2/ped) 6.3 11.3 5.9 11.0
Pedestrian perception score (link) 2.21 1.73 2.25 1.78
Pedestrian perception score (intersection) 3.13 2.58 3.05 2.58
Crossing difficulty factor 1.39 0.72 1.40 0.73
Pedestrian perception score (segment) 3.60 2.18 3.59 2.19
PLOS D B D B
Exhibition Road (eastern sidewalk)
Pedestrian volume (ped/h) 702 574 771 532
Average pedestrian space (ft2/ped) 146.3 179.1 133.1 208.2
Average pedestrian space (m2/ped) 13.6 16.6 12.4 19.3
Pedestrian perception score (link) 2.42 1.99 2.76 1.86
Pedestrian perception score (intersection) 2.63 2.61 2.67 2.67
Crossing difficulty factor 1.40 0.71 1.39 0.72
Pedestrian perception score (segment) 3.54 2.25 3.69 2.23
PLOS D B D B
Thurloe Street (northern sidewalk)
Pedestrian volume (ped/h) 743 na 824 na
Average pedestrian space (ft2/ped) 116.8 na 105.2 na
Average pedestrian space (m2/ped) 10.9 na 9.8 na
Pedestrian perception score (link) 2.08 na 2.11 na
PLOS B na B na
Thurloe Street (southern sidewalk)
Pedestrian volume (ped/h) 1237 na 1132 na
Average pedestrian space (ft2/ped) 11.8 na 13.6 na
Average pedestrian space (m2/ped) 1.1 na 1.3 na
Pedestrian perception score (link) 2.08 na 2.11 na
PLOS E na E na
Thurloe Street (access-only)
Pedestrian volume (ped/h) na 2608 na 3030
Average pedestrian space (ft2/ped) na 125.1 na 107.4
Average pedestrian space (m2/ped) na 11.6 na 10.0
PLOS na A na A

Note: na = not applicable; PLOS = pedestrian level of service.

crossing has been retained, resulting in roughly footpath, combined with the presence of bus shelters,
unchanged intersection perception scores. It should addi- rubbish bins, and other obstructions, as well as of high
tionally be noted that both walkways have also seen a sig- pedestrian volumes entering and exiting the adjacent
nificant increase in average pedestrian space, which, South Kensington Underground station. Consequently,
being above the threshold of 60 ft2 (5.6 m2) per pedestrian a PLOS rating of E is assigned to the southern sidewalk,
may not directly affect the PLOS ratings, but may still which means that the quality of service is unsatisfactory.
have an indirect influence through the perception scores. Post-redevelopment, on the other hand, Thurloe
Considering the evaluation results of Thurloe Street, Street has become an access-only street with much of the
whose two sidewalks are assessed at the link level in the vehicle traffic and most of the previously cluttering street
before case, it can be observed that the adequate pedes- furniture and other objects having been removed. This
trian space, together with a good link perception score, has considerably increased the average pedestrian space,
result in a PLOS rating of B for the northern sidewalk. and given that this is the only criterion in the evaluation
For the southern sidewalk, however, despite the link per- of off-street pedestrian facilities, PLOS has improved to
ception score (which depends on geometry and vehicle the maximum rating of A, which means that pedestrians
traffic characteristics) being the same, the average pedes- are able to move in their desired path without needing to
trian space is considerably smaller (as low as 11.8 ft2 or alter their pace or course to avoid conflicts with other
1.1 m2 per pedestrian) because of the relatively narrow road users.
12 Transportation Research Record 00(9)

All in all, the results of the PLOS evaluation of the corresponding safety analyses to explore potential safety
Exhibition Road site pre- and post-redevelopment impacts of the introduction of streetscape schemes with
appear to support the logical hypothesis that the imple- elements of shared space, not just at the site itself, but also
mentation of such features improves the quality of ser- in different (and perhaps not so obvious) network areas.
vice provided to pedestrians. Furthermore, a limitation of the present study has
been that, because of the timing of the analysis, the
HCM version before the latest one was used, that is, the
Conclusion 5th edition (HCM 2010) (10). This means that the analy-
In light of the shift in focus in urban street design, this sis could not be conducted using the improved LOS eva-
paper has examined the under-explored topic of how the luation methodologies included in the latest HCM 6th
quality of service changes as a result of the implementa- edition (HCM 2016) (9). It would be useful in future
research to perform an analysis using HCM 2016 instead
tion of street layouts with elements of shared space.
and to compare the results with the findings of this study.
Using video data from the Exhibition Road site in
Some discrepancy could be expected, in particular with
London during periods before and after its conversion
respect to the VLOS results, as the VLOS evaluation
from a conventional divided roadway to a single surface,
methodology for urban streets has changed from the pre-
featuring several elements of shared space, changes in
vious version.
relation to LOS for both vehicle traffic and pedestrians
Moreover, the research has been constrained to the
have been investigated.
case study in question, and the findings may therefore be
The results suggest that the redevelopment has consid-
biased by its specific characteristics. The further evalua-
erably improved the quality of service offered to pedes-
tion of other street sites with varying levels of implemen-
trians, with relevant PLOS ratings having increased from
tation of elements of shared space and in different cities
D to B and from E to A respectively at the two locations
and countries would be a very useful next step toward
evaluated. This is an expected finding, given that shared
obtaining more generic conclusions. Also, a further lim-
space features are intended to improve the pedestrian
itation of the present study has been that the video data
environment. What is also interesting, however, is that
were analyzed manually and by a single observer. It
vehicle traffic quality of service appears not to have been
would be of value to perform some additional analysis
compromised and has, in several locations, even
using multiple observers and technological tools (e.g.,
improved. In fact, VLOS ratings have overall increased
detectors and imaging software), complemented by rele-
from B to A and from D to C at the intersection and seg-
vant on-site observations, to further assert the validity of
ment examined, respectively.
the results and to improve on their accuracy.
Naturally, these results cannot be treated in isolation
In addition, future work could concentrate on evaluat-
from other related impacts of shared space design, and
ing the quality of service experienced by other road users
in particular safety-related ones. For instance, previous
in shared space environments, such as bicyclists (e.g.,
research on the Exhibition Road site found that there
using the corresponding LOS method in HCM), which
has been a change in vehicle–pedestrian traffic conflict
could deliver much different conclusions. Finally, the
patterns, with more slight and fewer severe occurrences
results obtained could be complemented by road user
post-redevelopment (11, 12). This suggests potentially
surveys, from which it will be possible to investigate more
safer conduct of vehicle drivers and pedestrians, but also
systematically the views and perceptions of pedestrians
implies fewer obstructions to the movement of vehicles
and drivers with respect to the quality of service of streets
and pedestrians, which would explain the quality of ser-
with elements of shared space.
vice improvements found by the present study. This is an
encouraging finding from the point of view of shared
space design. Acknowledgments
While this study has thrown some light on the topic of The authors would like to thank the Royal Borough of
quality of service in street layouts with elements of shared Kensington and Chelsea for supporting this work by providing
space, there are several future research directions that data and guidance.
remain to be explored next. For instance, a limitation of
the present study is that a decisive contributing factor of Author Contributions
the estimated VLOS and PLOS improvements has been
The authors confirm contribution to the paper as follows: study
the reduction in traffic flows at the case study site. Given conception and design: I. Kaparias, R. Wang; data collection:
that these flows have been diverted elsewhere, it would be R. Wang; analysis and interpretation of results: I. Kaparias, R.
important also to investigate potential VLOS and PLOS Wang; manuscript preparation: I. Kaparias. All authors
impacts at the network level, in addition to the site-level reviewed the results and approved the final version of the
analysis carried out. This could be complemented by manuscript.
Kaparias and Wang 13

Declaration of Conflicting Interests 15. Kaparias, I., J. Hirani, M. G. H. Bell, and B. Mount.
The author(s) declared no potential conflicts of interest with Pedestrian Gap Acceptance Behaviour in Street Designs
respect to the research, authorship, and/or publication of this with Elements of Shared Space. Transportation Research
article. Record: Journal of the Transportation Research Board,
2016. 2586: 17–27.
Funding 16. Highway Capacity Manual. Special Report 87, Highway
Research Board, Washington, D.C., 1965.
The author(s) received no financial support for the research, 17. Forschungsgesellschaft für Straßen- und Verkehrswesen.
authorship, and/or publication of this article. Handbuch fuär die Bemessung von Straßenverkehrsanlagen:
HBS 2015. FGSV Verlag, Köln, Germany, 2015.
References 18. Fruin, J. J. Designing for Pedestrians: A Level-of-Service
1. Buchanan, C., G. H. C. Cooper, A. MacEwen, D. H. Concept. Presented at 50th Annual Meeting of the Trans-
Crompton, G. Crow, and G. Michell. Traffic in Towns. portation Research Board, Washington, D.C., 1971.
Her Majesty’s Stationery Office, London, 1963, p. 224. 19. Bloomberg, M. R., and A. M. Burden. New York City
2. Hamilton-Baillie, B. Urban Design: Why Don’t We Do It Pedestrian Level of Service Study Phase I. NYC DCP,
in the Road. Journal of Urban Technology, Vol. 11, 2004, Transportation Division, New York, 2006.
pp. 43–62. 20. Landis, B. W., V. R. Vattikuti, R. M. Ottenberg, D. S.
3. Hamilton-Baillie, B., and P. Jones. Improving Traffic McLeod, and M. Guttenplan. Modeling the Roadside
Behaviour and Safety through Urban Design. Proceedings Walking Environment: Pedestrian Level of Service. Trans-
of the Institution of Civil Engineers - Civil Engineering, Vol. portation Research Record: Journal of the Transportation
158, 2005, pp. 39–47. Research Board, 2001. 1773: 82–88.
4. Hamilton-Baillie, B. Towards Shared Space. Urban Design 21. Dowling, R. G., D. B. Reinke, A. Flannery, P. Ryus, M.
International, Vol. 13, 2008, pp. 130–138. Vandehey, T. A. Petritsch, B. W. Landis, N. M. Rouphail,
5. Hamilton-Baillie, B. Shared Space: Reconciling People, Places and J. Bonneson. NCHRP Report 616: Multimodal Level of
and Traffic. Built Environment, Vol. 34, 2008, pp. 161–181. Service Analysis for Urban Streets. Final Report NCHRP
6. Department for Transport. Manual for Streets. Thomas Project 3-70. Transportation Research Board of the
Telford Publishing, London, 2007. National Academies, Washington, D.C., 2008.
7. Department for Transport. Manual for Streets 2 – Wider 22. Dowling, R., A. Flannery, B. Landis, T. Petritch, N. Rou-
Application of The Principles. Chartered Institute of High- phail, and P. Ryus. Multimodal Level of Service for Urban
ways and Transport, London, UK, 2010. Streets. Transportation Research Record: Journal of the
8. McCann, B. Completing Our Streets: The Transition to Transportation Research Board, 2008. 2071: 1–7.
Safe and Inclusive Transportation Networks. Island Press, 23. Raghuwanshi, A. K., and V. Tare. Assessment of Pedes-
Washington, D.C., 2013. trian Level of Service for Mixed Lane. Research Journal of
9. Highway Capacity Manual: A Guide for Multimodal Mobi- Engineering and Technology, Vol. 7, 2016, pp. 11–14.
lity Analysis, 6th ed. Transportation Research Board, 24. Karndacharuk, A., D. J. Wilson, and R. C. M. Dunn.
Washington, D.C., 2016. Analysis of Pedestrian Performance in Shared-Space Envir-
10. Highway Capacity Manual 2010. Transportation Research onments. Transportation Research Record: Journal of the
Board of the National Academies, Washington, D.C., 2010. Transportation Research Board, 2013. 2393: 1–11.
11. Kaparias, I., M. G. H. Bell, J. Greensted, S. Cheng, A. 25. Ruiz-Apilánez, B., K. Karimi, I. Garcı́a-Camacha, and R.
Miri, C. Taylor, and B. Mount. Development and Imple- Martı́n. Shared Space Streets: Design, User Perception and
mentation of a Vehicle-Pedestrian Conflicts Analysis Performance. Urban Design International, Vol. 22, 2017,
Method: Adaptation of a Vehicle-Vehicle Technique. pp. 267–284.
Transportation Research Record: Journal of the Transporta- 26. Euser, P. The Laweiplein: Evaluation of the Reconstruction
tion Research Board, 2010. 2198: 75–82. into a Square with Roundabout. Noordelijke Hogeschool
12. Kaparias, I., M. G. H. Bell, W. Dong, A. Sastrawinata, A. Leeuwaarden, Leeuwaarden, The Netherlands, 2006.
Singh, X. Wang, and B. Mount. Analysis of Pedestrian- 27. Edquist, J., and B. Corben. Potential Application of Shared
Vehicle Traffic Conflicts in Street Designs with Elements Space Principles in Urban Road Design: Effects on Safety
of Shared Space. Transportation Research Record: Journal and Amenity. Report to the NRMA-ACT Road Safety
of the Transportation Research Board, 2013. 2393: 21–30. Trust. Monash University, Accident Research Centre, Mel-
13. Kaparias, I., M. G. H. Bell, T. Biagioli, L. Bellezza, and B. bourne, Australia, 2012.
Mount. Behavioural Analysis of Interactions Between 28. Karndacharuk, A., D. J. Wilson, and R. C. M. Dunn. A
Pedestrians and Vehicles in Street Designs with Elements Review of the Evolution of Shared (Street) Space Concepts
of Shared Space. Transportation Research Part F: Traffic in Urban Environments. Transport Reviews, Vol. 34, 2014,
Psychology and Behaviour, Vol. 30, 2015, pp. 115–127. pp. 190–220.
14. Kaparias, I., M. G. H. Bell, A. Miri, C. Chan, and B. 29. Wargo, B. W., and N. W. Garrick. Shared Space: Could
Mount. Analysing the Perceptions of Pedestrians and Driv- Less Formal Streets Be Better for Both Pedestrians and
ers to Shared Space. Transportation Research Part F: Traf- Vehicles? Presented at 95th Annual Meeting of the Trans-
fic Psychology and Behaviour, Vol. 15, 2012, pp. 297–310. portation Research Board, Washington, D.C., 2016.