Content uploaded by Jiayu Pan
Author content
All content in this area was uploaded by Jiayu Pan on Jul 14, 2021
Content may be subject to copyright.
CIBSE Technical Symposium, UK 13-14 July 2021 2020
Page 1 of 16
Redesigning the working space for social
distancing: Modelling the movement in an open-plan
office
MISS JIAYU PAN BSC, MPHIL
Department of Architecture, University of Cambridge
jp844@cam.ac.uk
MR TZE YEUNG CHO BA, MENG
Department of Engineering, University of Cambridge
zc282@cam.ac.uk
DR RONITA BARDHAN BARCH, MCP, PHD
Department of Architecture, University of Cambridge
rb867@cam.ac.uk
Abstract
The global pandemic has reshaped the use of working space dramatically, mainly
due to the implementation of the working from home policy and indoor social
distancing requirements. This calls for the rethinking of the current office design
approaches and the proposal of specific design strategies to provide a safer and
healthier environment for employees to return to their offices. Also, the demand for
enhancing office design to accommodate special and rare events like the pandemic
is identified. More resilient and flexible office designs are needed for the post-
pandemic era. This study aims to provide an insight into the human movement in
the open-plan office setting by simulating the movement using agent-based
modelling. Normal scenario and special scenarios with social distancing standards
and reduced office capacity are simulated. The simulated scenarios contribute to
the development of the new adaptive office design approaches for a safe office
resume.
1. Introduction
The COVID-19 pandemic has significantly reshaped the work routine and office
experience. The containment measures including lockdown, travel restrictions and
social-distancing requirements have reduced the mobility of people, while
companies respond to the pandemic challenge with ‘work from home’ policy.
Although ‘work from home’ can lead to some improvements in the productivity (1),
office still has its irreplaceable role on providing face-to-face communication and
building up social connections. The plan for returning to offices is resumed with the
delivery of mass vaccination and the release of lockdown.
It is inevitable for people to gather, contact and interact when they go back to
offices, which may lead to an increase in the risk of transmission. The non-
pharmaceutical interventions like increasing air ventilation and social distancing are
proved to be effective (2), while the design of physical environment can help to
address the interventions. The effectiveness of the potential ventilation strategies in
indoor environment has been discussed in the previous studies (3–6). The questions
of how many people can safely access the office, what is the optimal social
CIBSE Technical Symposium, UK 13-14 July 2021 2020
Page 2 of 16
distance in office, how many people are allowed to use the facilities around the
office and how to alter the furniture layout to avoid overcrowding situation need to
be addressed. The corresponding office design solutions can contribute to reducing
the concentration of people thus reducing the potential of virus spreading. The
understanding of how people use and move around the office space is expected to
contribute to the decision making on the office reopening plans. The simulation
modelling approach can help to model the different scenarios with various
intervention strategies like layout changing, social distancing and population
limitations. An optimum strategy can be found through this process to ensure a safe
resumption.
This study is going to investigate the application of social distancing strategy in the
office. It models the movement of people in an office environment with agent-based
pedestrian simulation tool and examines the scenarios with various social
distancing requirements and limitations on the populations. This investigation is
essential for figuring out the appropriate strategy for safe reopening. Also, an
enhanced understanding on the movement and potential strategies can contribute
to future design. The effectiveness of different layouts and design strategies can be
tested with the simulation tools. Furthermore, the agent-based modelling in the
normal office setting has the potential to be developed as a comprehensive project,
from the observation and collection of behavioural data to establish a model
applying specifically to the workspace. This study provides an initial insight for the
potential of the investigation.
2. Literature review
The literature review part learns from the existing research on the movement in the
office environment and the simulation of pedestrian movement in the built
environment. The findings provide a background overview and pave the way for the
proposed study.
2.1 Movement in office
A large body of literature has discussed the observation of human activities in the
office environment. The behaviours of sitting, standing and walking are widely
measured and analysed in the domain of physical health research. The physical
activity monitor devices are used to trace the movement and record the sitting and
standing time and walking steps, while the questionnaire is applied to collect the
self-reported data (7–10). The sedentary behaviour, which indicates a prolonged
sitting time of more than 30 minutes (11,12), is considered a threatening to health
(13,14). The researchers test different approaches to reduce the sedentary time,
such as the use of sit-stand work station (15–17), the renovation of office layout
from traditional cells to activity-based type (18) and the seminar, workshop and
training sessions. The collected dataset show that the sitting time usually occupies
over 60% of the working time (16,19,20), while the largest proportion noted is 93%
(21). The recorded standing and walking time is much less than sitting, while the
figure varies in different studies, with a range of 10% to 20% for standing (15,22,23)
and 7% to 12% for stepping (15,23,24). Spinney et al. (25) report the frequency of
trips in an office. The trips from the desk per hour is 1.6, and the average number of
trips to the restroom and kitchen per hour is 0.38 and 0.96 separately.
CIBSE Technical Symposium, UK 13-14 July 2021 2020
Page 3 of 16
The correlation between the movement and spatial configuration is also a major
found in the previous research. Carter and Whitehead (26) study the pedestrian
circulation in Rail House and notes down 69 activities that can generate
movements, such as the visit to mailroom and the use of ground floor entrance.
Rassia (27) points out the informal communication for job interaction as a key
reason for movement in the office, while the movement is found along the corridors
and to landing and reception areas. The movement and interaction hotspots are the
individual desks, water coolers and print station (28). The indoor architectural
design can have an influence on the occupants’ movement pattern by considering
the layout carefully (27). The spatial system explains the movement flow in offices
effectively, and the results from two case studies show that the metric analysis
demonstrates the workplace movement best (29). The movement in offices is driven
by the geometric and non-geometric properties and the needs of travelling to
attractors. The geometric property refers to the pure spatial configuration
measurement, while the non-geometric feature takes the functional configuration
(the locations of furniture and functional spaces) into consideration. The travel
concentration measurement constructs paths from each seat to attractors of the
building entrance, the closest canteen, kitchen and restroom and transforms the
paths to the visible zones. This new spatial metric is tested against a large dataset
and proved to have some predictive power to the movement activity (30). The
correlational analysis of spatial configuration features and movement at micro, meso
and macro levels identifies the metrics of visual control, isovist min radial and travel
concentration are the most significant measurements to the movement activities.
The finding supports the argument of movement mainly taking place in the
circulation spaces (31).
2.2 Agent-based movement simulation
The movement simulation has been widely applied in built environment research,
because the understanding of how pedestrians move around urban spaces or
buildings can play a significant role in assisting the planning and design. The
microscopic models like cellular automata, social force and agent-based models are
created for the movement simulation (32,33). The agent-based model has a unique
feature of involving variability and randomness which leads the system to unknown
limits (34,35). Schelhorn et al. (36) propose the agent-based STREETS model to
look at the pedestrian activities in urban districts by integrating the socioeconomic
and behavioural characteristics of agents and the physical environment datasets in
the GIS system. The geometric agent-based pedestrian simulation PEDFLOW for
the microscope environment is addressed in the work of Kerridge, Hine and Wigan
(37). PEDFLOW is a complement work to the STREETS model with a higher level
network and decision structure.
The recent two decades have witnessed the emergence of agent-based crowd
simulations for different cases and scenarios. A classification framework
categorises the type of modelling by environment and behaviour. The type of
environment includes small-scale enclosed spaces, large-scale enclosed spaces,
mixed-mode, open space and hybrid, while the behaviour categories are purposeful
and familiar, purposeful and unfamiliar, purposeless, evacuation, forced-waiting and
temporal constraints (38). Shaaban and Abdelwarith (32) explore the pedestrian
behaviour when crossing a road, while different pedestrian attributes, like gender,
CIBSE Technical Symposium, UK 13-14 July 2021 2020
Page 4 of 16
age, clothing, are considered in the investigation. The human and social behaviour
in emergency evacuations is demonstrated in works (39–41). Rassia and Siettos (42)
introduce the molecular dynamics approach to model the evacuation in office
buildings. The simulations at a larger scale or mix-mode environment, like the
railway station (43), metro station (44) and outdoor recreation spaces (45).
For the post-pandemic world, simulation tools could assist the decision making by
performing possible scenarios of human activities. The importance of simulating the
risks and scenarios is also highlighted by the insight from Arup. The Space Explorer
tool is introduced as a service to understand people’s movement and risk for
returning to workspaces (46). The office behaviours and movements are different
from the normal crowd movements on streets or around public spaces, as the office
is a semi-public environment that involves more complex functions and activities
and a longer staying time. While sitting accounts for the majority of working time,
relatively less time is spent on standing and walking. Many of the movements
around office are generated by the visits to the attractors like printers, restroom and
meeting room. The employees have various levels of interactions like chatting,
visiting and meeting. According to the classification framework of agent-based
pedestrian models by Ronald, Sterling and Kirley (38), the model for the office
environment is classified as the small-scale enclosed spaces modelling with
purposeful and familiar behaviours. This combination remains a gap in the field.
Therefore, there is a need for an integrated tool to model the movement specifically
for the working environment as an aid to the post-pandemic office design.
3. Method
This section introduces the data and method used in the movement simulation.
Figure 1 illustrates the step-by-step flow chart of this study.
Figure 1 – Steps of the study
3.1 Data collection
The office plans applied in the study are extracted from the collection of plans in
Koutsolampros's work (31). The plan comes from the UK office dataset collected by
SpaceLab, while a set of predictive analysis on the spatial configuration and the
CIBSE Technical Symposium, UK 13-14 July 2021 2020
Page 5 of 16
activities of movement and interaction is performed in the previous works
(30,47,48). The first and second floors of Case 27 are selected for the analysis
considering the availability of relevant information (shown in Figure 2). Except for the
floor plan, the function distribution information is also extracted and illustrated
(shown in Figure 3). There are a total of 180 desks on the first floor and 174 desks
on the second floor with an open office plan. Therefore, the assumptions on the
maximum capacity of the offices are 180 and 174 respectively. Both of the floors
feature a dense seating plan, classified as dense workspace floors (31). The floor
entry and exit are the staircase and three lifts on the corridor. The first floor has four
meeting rooms and one alternative working space. The second floor has five
meeting rooms and two alternative working areas. There is a restroom and a tea
point with seats on each floor and facilities like printers and lockers locating around
the floor. The slight differences in the layouts of the two floors make their simulation
results comparable.
Figure 2 – Floor plan, Case 27 First floor (1F) and Second floor (2F) (adapted
from ref. (31))
Figure 3 – Floor plan with functions, Case 27 First floor (1F) and Second floor
(2F) (adapted from ref. (31))
CIBSE Technical Symposium, UK 13-14 July 2021 2020
Page 6 of 16
3.2 Pedestrian simulation setting
The pedestrian simulation plug-in PedSim Pro working under Grasshopper in Rhino
is applied to simulate the movement in the open-plan office (49). It is an agent-
based simulation tool based on a particle spring system with agents driven by
various forces like repulsion and attraction forces. This tool is selected because it is
easy to access and allows users to run and design their own scenarios. Its
predecessor PedSim is the most popular pedestrian simulation tool in Grasshopper,
while PedSim Pro has more features with greater modelling flexibility. The simulation
is run on a 3.1GHz Intel Core i7 computer with Rhino 6 and PedSim Pro 1.1.1.
The major components in the simulation modelling are stops, interests, obstacles,
person templates and the engine. Stops and interests are the location points on the
plan that agents can visit. Stops are the planned targets on the route, including a
start, a destination and multiple other visits. The movement route of the agent is
defined by the planned stops. Interests are the unplanned stops, as they are visible
and accessible. The interest and stop can be set as a room or a stand. The agents
who visit a room disappear from the map, while those who visit a stand still remain
on the plan and interact with the other agents. Obstacles can be either opaque or
transparent, and the setting of obstacles influences the decision of routes. The
person template defines the type of agents with different combination of stops and
interests. The engine is the essential component that integrates all the stops,
interests, obstacles and person templates.
In the simulation, a full circulation from the entry to the desk to the exit is built for
each agent. The entry and exit are the staircase and lifts. Each staircase or lift can
be either starting point or destination. The planned stop is the desks around the
office, and the desks are grouped into four sets according to their zone. The
restroom, meeting rooms and an alternative working space are set as interest with
the room feature, while the facilities, tea point and an open-plan alternative
workspace are assigned as the interest with the stand feature. Each agent can visit
a maximum of five interests on their designated route.
Object
Type of
program
Stand or
Room
Number of
objects
(on each floor)
Visiting
Time
(Unit: s)
Capacity
(For each
program)
Lift
Stop
Room
3
-
10
Staircase
Stop
Room
1
-
80
Desk
Stop
Stand
4 groups
180 (1F), 174 (2F)
30
1
Meeting room
Interest
Room
4 (1F), 5 (2F)
20
6-12
Restroom
Interest
Room
1
3
6
Alternative
workspace
Interest
Room and
stand
2 (1F), 3 (2F)
30
3-6
Teapoint
Interest
Stand
8
2-10
1
Facilities
Interest
Stand
31 (1F), 44 (2F)
2
1-2
CIBSE Technical Symposium, UK 13-14 July 2021 2020
Page 7 of 16
Table 1 – Model setting in PedSim Pro
Table 1 shows the setting of stops in the PedSim Pro model. The default fixed unit
setting is meter (unit: m) for distance and second (unit: s) for time. Due to the
limitation of the computing power, the time frame in simulation is scaled down. In
the model, 1 second corresponds to 1 minute of real-world time. To get a
representative result with the limited computational power, the simulation is run for
90 seconds for each scenario. The profile of the agent, including body radius, mass
and speed, follows the default setting in the plug-in, and the vision is set as
‘panorama’. There are 16 different combinations of ‘entry-desk-exit’ route (two entry
and exit options of lift or staircase and four groups of seats), thus a total of 16 agent
templates are created in modelling. An agent generates from the entry points in
every two seconds, and the maximum capacity of the lifts is 10 for each lift. The
visit time to working desks is about 30 seconds, which indicates a sedentary time of
30 minutes. Each desk can hold only 1 person. The visit time to the meeting room is
set at 20 seconds, and the capacity of the meeting room varies with the size, from 6
people to 12 people. For the tea point, restroom and facilities, the staying time
ranges from 3 seconds to 15 seconds and the capacity differs case by case.
3.3 Proposed scenarios
Total population
allowed in the office
Social
distancing (2m)
Interest
capacity
Seating
rule
Scenario 1
Full capacity
No
Normal
No
Scenario 2
3/4 (75%) of full capacity
Yes
Half
No
Scenario 3
1/2 (50%) of full capacity
Yes
Half
Yes
Table 2 – Scenario settings
Three basic scenarios with different virus containment strategies are proposed for
simulation (summarised in Table 2). Scenario 1 is the normal scenario with the full
capacity. The population presented in the office is 180 on the first floor and 174 on
the second floor. Scenario 2 simulates the situation with a reduced number of
people and a social distance of 2 meters. Only 75 per cent of the full population can
access the office, while the number of people allowed in the meeting rooms, lifts
and alternative working spaces is reduced to half of the original capacity. Only one
agent can use one facility every time. In scenario 3, half of the population is allowed
in the office. The social distancing requirement and the capacity limitations on the
facilities and functional rooms are the same as the scenario 2 settings. Additionally,
the gap between seats is placed. Only half of the desks are accessible (shown in
Figure 4). The snapshots for each scenario are taken, showing the position of
agents. All scenarios are run for the first floor, while only the normal scenario is
conducted for the second floor. Figure 5 presents the series of snapshots at 30, 60
and 90 seconds for scenario 1 on the first floor, while Figure 6 illustrates the
snapshots for all simulated scenarios at the time of 60 seconds.
CIBSE Technical Symposium, UK 13-14 July 2021 2020
Page 8 of 16
Figure 4 – Illustration of the seat gap
Figure 5 – Scenario snapshots for Case 27, First floor, Scenario 1 (t=30s, 60s,
90s)
Figure 6 – Scenario snapshots for Case 27, First floor, Scenario 2, Scenario 3;
Second floor Scenario 1 (t=60)
The data generated from the model is visualised in the format of heat maps, which
shows the movement density around the space. Following the previous analysis
(31), the lattice grid with a cell size of 45 by 45 centimetres is used. The number of
visits to each cell is counted and presented in the heatmap. The number of visits to
the interests is counted to represent their attractiveness.
4. Result and Discussion
4.1 Simulation results
Figure 7 presents the groups of movement density heat maps, and Figure 8 shows
the visit count to each interest. Generally, the highest density occurs around the
corridor. The visiting density decreases as the distance from the lift and staircases
increases. The facilities and desks located far from the corridor have the least visit
CIBSE Technical Symposium, UK 13-14 July 2021 2020
Page 9 of 16
counts. The tea point is less visited, which could be explained by its poorly visible
location. The meeting room and restroom along the corridor are the most visited
hotspots, while the visit count to the meeting rooms in the distant area is much less.
The facilities around the circulation area also cause some density concentration.
Figure 7 – Heat map
CIBSE Technical Symposium, UK 13-14 July 2021 2020
Page 10 of 16
Figure 8 – Visit count of interests
The reduction on the number of occupants in the office has effectively reduced the
movement density and the concentration around the interests. The movement area
contracts as the number of agents reduces, as the agents do not travel further to
reach the desks at the distant corners. In scenario 3, the overcrowding situation still
occurs at the entry and exit points and the restroom, while the movement density
along the corridor decreases significantly. The impact of social distancing and the
seat gap strategies is not apparent on the heatmap. The effectiveness of social
distancing measurement is demonstrated by the difference in the use of corridor. In
scenario 1 of the first floor, the flow concentrates in the middle of the corridor, while
the highest density occurs at the two sides of the corridor near the wall in scenario
2 and 3. The movement flows are separated to maintain the social distance. A
noticeable increasing in the distance between agents are shown on the snapshots
(Figure 6).
The normal scenario is run for both the first and second floor, while there are some
differences in the layout. The main distinctions in the layout occur on the right side
of the plan. The upper right quadrant on the second floor has a denser seating
arrangement comparing to the first floor. However, there is no significant variation
found in the movement around the area. The set of facilities locating along the
circulation area opposite to the tea point are visited 16 times, but those interest
points do not lead to extra concentration comparing to the same location on the
CIBSE Technical Symposium, UK 13-14 July 2021 2020
Page 11 of 16
first floor. As the entry and exit points, the restroom, the corridor and the circulation
remain in the same location, the general movement flow in both floors looks similar.
The results echo some of the findings from spatial configuration analysis (shown in
Figure 9). The three most important metrics are visual control and isovist min radial
based on the accessible areas and the in-floor travel concentration (31). Visual
control compares the space visible from a cell to the other directly visible cells.
Isovist min radial measures the distance from the isovist origin to the nearest
obstacle. Travel concentration evaluates the effect of attractors. It is not surprising
to see the corridor become the movement hotspot in the plan considering the
relatively high indoor travel concentration and isovist min radial value shown in the
previous analysis. The high values of the spatial structure metrics are also found
around the circulation area outside the tea point, while the area features a relatively
dense movement as well.
Figure 9 –The three most important spatial configuration metrics to movement,
Case 27 First Floor (extracted from ref. (31))
4.2 Potential strategies
In addition to the containment strategies of introducing social distancing, limiting
capacity and reducing crowding, the design strategies can be applied to the office
space including changing the layout, setting up the one-way system and adding
barriers between seats. Also, the ‘Working safely during coronavirus’ guideline
provides a series of suggestions, such as carrying out risk assessments, arranging
the workspace to keep people apart, using back-to-back or side-to-side seating,
increasing the frequency and restricting the access to different areas (50). The
simulation results support the application of social distancing and capacity
limitation, while the findings also lead to further reflections on the design of the
space.
In the floor plans examined in this study, a high concentration of movement is found
in the corridor area, mainly due to the location of staircase and lifts. At the same
time, employees need to pass the corridor to get access to the meeting rooms and
restroom, which adds to the crowding density. To reduce the high traffic around the
corridor, diverting the movement flow in the corridor is essential. The entry and exit
points could be separated, which indicates a greater distance between the staircase
CIBSE Technical Symposium, UK 13-14 July 2021 2020
Page 12 of 16
and lifts. The meeting rooms can be re-layout from their current location to the less
accessible areas. This measure could alleviate the travel concentration around the
corridor and discourage employees to visit the meeting rooms, as meeting rooms
are small and enclosed areas with a higher risk of spreading. Some of the facilities
can be relocated from the locations around the circulation area to less accessible
spaces.
5. Conclusion and limitations
In conclusion, this study provides an initial insight into the movement simulation in
the workspace. The literature review indicates that there is a need for more targeted
simulation tools for different scales, scenarios and environments. The study
demonstrates how the movement simulation in the office environment could
potentially support the decision making in the office reopening in the post-
pandemic era. While the government guideline sets the general approaches, the
movement simulation tool helps to provide more detailed analysis and solutions by
looking into a specific case and modelling case-specific scenarios. Comparing to
the existing agent-based modelling research which focuses more on the design of
models, this study explores the potential of applying the simulation results to real
cases. In this case, the results highlight that the corridor is the high movement
density area. There is a need for some specific considerations on reducing the
movement concentration around the corridor.
There are several limitations of this study, which influence the accuracy of the
simulation. Firstly, the model inputs are designed based on the secondary sources
and assumptions with a lack of primary data involved. The generalised and
simplified movement pattern used for simulation is relatively simple, and it could not
accurately reflect the complex situation in the real office environment. Secondly, the
tool used in this simulation is designed for the understanding of general movement
instead of for office-specific scenarios. This plug-in is unable to simulate the vertical
movement between different floors, though the two plans used in this study are
from the same case. Only the movement within floor is modelled. The personal
templates allow some degree of diversity and flexibility in the agent setting, the
agents still tend to be homogenised in the model. Thirdly, the agent-based
simulation results involve randomness, which indicates that the results for every
single simulation could have slight differences. The future investigation can be
coupled with environment modelling, especially the simulation of air ventilation in
the office space, to understand the behaviour and assess the risk of returning to the
offices.
Acknowledgements
The authors would like to thank Dr Petros Koutsolampros for his valuable
discussion in office space interactions and the office case study from his thesis.
Reference
1. Bloom N, Liang J, Roberts J, Ying ZJ. Does working from home work?
Evidence from a Chinese experiement. Q J Econ. 2014;130(1):165–218.
2. Sun C, Zhai Z. The efficacy of social distance and ventilation effectiveness in
CIBSE Technical Symposium, UK 13-14 July 2021 2020
Page 13 of 16
preventing COVID-19 transmission. Sustain Cities Soc [Internet].
2020;62(June):102390. Available from:
https://doi.org/10.1016/j.scs.2020.102390
3. Melikov AK. COVID-19: Reduction of airborne transmission needs paradigm
shift in ventilation. Build Environ. 2020;186:19–21.
4. Nembhard MD, Burton DJ, Cohen JM. Ventilation use in nonmedical settings
during COVID-19: Cleaning protocol, maintenance, and recommendations.
Toxicol Ind Health. 2020;36(9):644–53.
5. Barbosa BPP, de Carvalho Lobo Brum N. Ventilation mode performance
against airborne respiratory infections in small office spaces: limits and
rational improvements for Covid-19. J Brazilian Soc Mech Sci Eng [Internet].
2021;43(6). Available from: https://doi.org/10.1007/s40430-021-03029-x
6. Bhagat RK, Davies Wykes MS, Dalziel SB, Linden PF. Effects of ventilation on
the indoor spread of COVID-19. J Fluid Mech. 2020;903.
7. Thøgersen-Ntoumani C, Quested E, Smith BS, Nicholas J, McVeigh J, Fenton
SAM, et al. Feasibility and preliminary effects of a peer-led motivationally-
embellished workplace walking intervention: A pilot cluster randomized trial
(the START trial). Contemp Clin Trials [Internet]. 2020;91(February):105969.
Available from: https://doi.org/10.1016/j.cct.2020.105969
8. Smith L, Sawyer A, Gardner B, Seppala K, Ucci M, Marmot A, et al.
Occupational physical activity habits of UK office workers: Cross-sectional
data from the active buildings study. Int J Environ Res Public Health.
2018;15(6).
9. Fisher A, Ucci M, Smith L, Sawyer A, Spinney R, Konstantatou M, et al.
Associations between the objectively measured office environment and
workplace step count and sitting time: Cross-sectional analyses from the
active buildings study. Int J Environ Res Public Health. 2018;15(6).
10. Olsen HM, Brown WJ, Kolbe-Alexander T, Burton NW. A brief self-directed
intervention to reduce office employees’ sedentary behavior in a flexible
workplace. J Occup Environ Med. 2018;60(10):954–9.
11. Danquah IH, Kloster S, Holtermann A, Aadahl M, Bauman A, Ersbøll AK, et al.
Take a Stand!-A multi-component intervention aimed at reducing sitting time
among office workers-a cluster randomized trial. Int J Epidemiol.
2017;46(1):128–40.
12. Brakenridge CL, Healy GN, Winkler EAH, Fjeldsoe BS. What do workers do to
reduce their sitting time? the relationships of strategy use and workplace
support with desk-basedworkers’ behavior changes in aworkplace-delivered
sitting-reduction and activity-promoting intervention. J Occup Environ Med.
2018;60(11):1026–33.
13. Clemes SA, O’Connell SE, Edwardson CL. Office workers objectively
measured sedentary behaviour and physical activity during and outside
working hours. J Occup Environ Med. 2014;56(3):298–303.
14. Brombacher H, Arts D, Megens C, Vos S. Stimulight: Exploring social
interaction to reduce physical inactivity among office workers. Conf Hum
Factors Comput Syst - Proc. 2019;1–6.
15. Li I, Mackey MG, Foley B, Pappas E, Edwards K, Chau JY, et al. Reducing
Office Workers’ Sitting Time at Work Using Sit-Stand Protocols: Results From
a Pilot Randomized Controlled Trial. J Occup Environ Med. 2017;59(6):543–9.
CIBSE Technical Symposium, UK 13-14 July 2021 2020
Page 14 of 16
16. Huysmans MA, Srinivasan D, Mathiassen SE. Consistency of sedentary
behavior patterns among office workers with long-term access to sit-stand
workstations. Ann Work Expo Heal. 2019;63(5):583–91.
17. Sharma PP, Mehta RK, Pickens A, Han G, Benden M. Sit-Stand Desk
Software Can Now Monitor and Prompt Office Workers to Change Health
Behaviors. Hum Factors. 2019;61(5):816–24.
18. Hallman DM, Mathiassen SE, Jahncke H. Sitting patterns after relocation to
activity-based offices: A controlled study of a natural intervention. Prev Med
(Baltim) [Internet]. 2018;111(November 2017):384–90. Available from:
https://doi.org/10.1016/j.ypmed.2017.11.031
19. Gao Y, Cronin NJ, Nevala N, Finni T. Validity of long-term and short-term recall
of occupational sitting time in Finnish and Chinese office workers. J Sport
Heal Sci [Internet]. 2020;9(4):345–51. Available from:
https://doi.org/10.1016/j.jshs.2017.06.003
20. Jancey JM, McGann S, Creagh R, Blackford KD, Howat P, Tye M. Workplace
building design and office-based workers’ activity: A study of a natural
experiment. Aust N Z J Public Health. 2016;40(1):78–82.
21. Ezezue A, Ibem E, Odum C, Obiadi B. Architectural design interventions for
sedentary behaviour among workers in office buildings in Enugu, Nigeria. Civ
Eng Archit. 2020;8(6):1451–62.
22. Macdonald B, Gibson AM, Janssen X, Kirk A. A mixed methods evaluation of
a digital intervention to improve sedentary behaviour across multiple
workplace settings. Int J Environ Res Public Health. 2020;17(12):1–27.
23. Gorman E, Ashe MC, Dunstan DW, Hanson HM, Madden K, Winkler EAH, et
al. Does an “Activity-Permissive” Workplace Change Office Workers’ Sitting
and Activity Time? PLoS One. 2013;8(10):6–11.
24. Morris AS, Mackintosh KA, Dunstan D, Owen N, Dempsey P, Pennington T, et
al. Rise and recharge: Effects on activity outcomes of an e-health smartphone
intervention to reduce office workers’ sitting time. Int J Environ Res Public
Health. 2020;17(24):1–18.
25. Spinney R, Smith L, Ucci M, Fisher A, Konstantatou M, Sawyer A, et al. Indoor
tracking to understand physical activity and sedentary behaviour: Exploratory
study in UK office buildings. PLoS One. 2015;10(5):1–19.
26. Carter DJ, Whitehead B. A study of pedestrian movement in a multi-storey
office building. Build Environ. 1976;11(4):239–47.
27. Rassia S. The analysis of the role of office space architectural design on
occupant physical activity. PLEA 2008 - Towar Zero Energy Build 25th PLEA
Int Conf Passiv Low Energy Archit Conf Proc. 2008;(October):20–5.
28. Rassia ST, Hay S, Beresford A, Barker N V. Movement dynamics in office
environments. 3rd CIB Int Conf Smart Sustain Built Environ (SASBE 2009).
2009;
29. Sailer K. Movement in workplace environments – configurational or
programmed? 6th Int Sp Syntax Symp İstanbul 2007 [Internet]. 2007;68.
Available from: http://eprints.ucl.ac.uk/3497/
30. Koutsolampros P, Sailer K, Haslem R. Travel Concentration: The effects of
attractor-bound movement on workplace activity. Futur Work. 2020;1–13.
31. Koutsolampros P. Human behaviour in office environments. Finding patterns
of activity and spatial configuration in large workplace datasets. University
CIBSE Technical Symposium, UK 13-14 July 2021 2020
Page 15 of 16
College Londo; 2021.
32. Shaaban K, Abdelwarith K. Pedestrian attribute analysis using agent-based
modeling. Appl Sci. 2020;10(14).
33. Sun Q. A Generic Approach to Modelling Individual Behaviours in Crowd
Simulation [Internet]. PhD Thesis. 2013. Available from:
http://usir.salford.ac.uk/30771/
34. Batty M. Agent-Based Pedestrian Modeling. Environ Plan B Plan Des.
2001;28(3):321–6.
35. Bonabeau E. Agent-based modeling: Methods and techniques for simulating
human systems. Proc Natl Acad Sci U S A. 2002;99(SUPPL. 3):7280–7.
36. Schelhorn T, O’Sullivan D, Haklay M, Thurstain-Goodwin M. Streets: An
agent-based pedestrian model. Cent Adv Spat Anal Work Pap Ser. 1999;
37. Kerridge J, Hine J, Wigan M. Agent-based modelling of pedestrian
movements: The questions that need to be asked and answered. Environ Plan
B Plan Des. 2001;28(3):327–41.
38. Ronald N, Sterling L, Kirley M. An agent-based approach to modelling
pedestrian behaviour. Int J Simul Syst Sci Technol. 2007;8(1):25–38.
39. Pan X, Han CS, Dauber K, Law KH. A multi-agent based framework for the
simulation of human and social behaviors during emergency evacuations. AI
Soc. 2007;22(2):113–32.
40. Roan T-R. Developing an agent-based evacuation simulation model based on
the study of human behaviour in fire investigation reports [Internet]. 2013.
Available from:
http://oatd.org/oatd/record?record=%22oai:ethos.bl.uk:626669%22&title=De
veloping an agent-based evacuation simulation model based on the study of
human behaviour in fire investigation reports
41. Toyama MC, Bazzan ALC, Silva R Da. An agent-based simulation of
pedestrian dynamics: From lane formation to auditorium evacuation. Proc Int
Conf Auton Agents. 2006;2006:108–10.
42. Rassia ST, Siettos CI. Escape Dynamics in Office Buildings: Using Molecular
Dynamics to Quantify the Impact of Certain Aspects of Human Behavior
During Emergency Evacuation. Environ Model Assess. 2010;15(5):411–8.
43. Rindsfüser G, Klügl F. Agent-based pedestrian simulation A case study of the
bern railway station. Disp. 2007;170(3):9–18.
44. Wang WL, Lo SM, Liu SB, Ma J. On the use of a pedestrian simulation model
with natural behavior representation in metro stations. Procedia Comput Sci.
2015;52(1):137–44.
45. Morelle K, Buchecker M, Kienast F, Tobias S. Nearby outdoor recreation
modelling: An agent-based approach. Urban For Urban Green [Internet].
2019;40(January 2018):286–98. Available from:
https://doi.org/10.1016/j.ufug.2018.07.007
46. Arup. Arup Space Explorer [Internet]. Arup services. 2021 [cited 2021 Apr 20].
Available from: https://www.arup.com/expertise/services/digital/space-
explorer
47. Koutsolampros P, Sailer K, Pomeroy R, Austwick MZ, Hudson-Smith A,
Haslem R. Spatial databases: Generating new insights on office design and
human behaviours in the workplace. SSS 2015 - 10th Int Sp Syntax Symp.
2015;1–16.
CIBSE Technical Symposium, UK 13-14 July 2021 2020
Page 16 of 16
48. Koutsolampros P, Sailer K, Varoudis T, Haslem R. Dissecting Visibility Graph
Analysis: The metrics and their role in understanding workplace human
behaviour. Proc 12th Sp Syntax Symp. 2019;191.1-191.24.
49. Parametric Design Studio. PedSim Pro [Internet]. 2020 [cited 2021 Apr 18].
Available from: https://www.pedsim.net/
50. Department for Business Energy & Industrial Strategy, Department for Digital
Culture Media & Sport. Working safely during coronavirus (COVID-19).
Coronavirus (COVID-19) Rules, guidance and support. 2021.
