Content uploaded by Holger Schnädelbach
Author content
All content in this area was uploaded by Holger Schnädelbach on Jun 01, 2016
Content may be subject to copyright.
International Adaptive Architecture Conference, Building Centre, London, March 2011: H. Schnädelbach 1
Physiological Data in Adaptive Architecture
Holger Schnädelbach
Mixed Reality Laboratory
Computer Science, Jubilee Campus, University of Nottingham, Nottingham, NG81BB, UK
T: (+44) 0115 9514094 F: (+44) 0115 8466416
Biography
Dr Holger Schnädelbach is a Senior Research Fellow in the Mixed Reality Lab (MRL), Computer
Science, University of Nottingham. His PhD at the Bartlett School of Architecture was concerned with
the spatial aspects of the relationship between physical and virtual environments, leading to the
prototypical Mixed Reality Architecture. His work at the MRL has provided experience in designing,
implementing and evaluating interactive systems. His work has resulted in publications in leading
conferences and journals, such as ACM CHI, TOCHI, CSCW, Presence and Space Syntax
Symposium. During a recent Leverhulme Fellowship he was given the space to concentrate on
Adaptive Architecture, concerned with buildings that adapt to their environment, their inhabitants and
objects in their vicinity. This work has developed a particular focus on what it means to inhabit
adaptive buildings, considering the motivations, strategies and challenges that individuals, groups and
organisations face when the buildings around them adapt.
International Adaptive Architecture Conference, Building Centre, London, March 2011: H. Schnädelbach 2
Physiological Data in Adaptive Architecture
Holger Schnädelbach
Mixed Reality Laboratory
Computer Science, Jubilee Campus, University of Nottingham, Nottingham, NG81BB, UK
T: (+44) 0115 9514094 F: (+44) 0115 8466416
Abstract
This paper discusses physiological data as one stream of personal data available to drive adaptations
in architecture. Physiological data here refers to data such as heart rate, respiration and brain activity
that can be measured, manipulated and to some extent be interpreted to say something about the
inhabitants’ emotional state. An iteratively developed prototype, the ExoBuilding, allowed the
exploration of some of the salient issues in this space, and this exploration highlighted the possibility
for a biofeedback loop emerging between the inhabitant and their ‘physiologically’ adaptive
environment. An associated early consideration of the possible mappings between physiological data
and the building fabric gave rise to an initial model of physiological data in Adaptive Architecture,
presented here, which sets out the relationship between inhabitant, physiological data, actuation and
effect with biofeedback being a special case specific to the context of this work. An outline of a series
of hypothetical case studies explores the mappings between physiological data and the building
fabric, which in turn leads to the suggestion that adaptive architecture best be understood as socio-
technical system.
Introduction
Adaptive Architecture is concerned with buildings that are specifically designed to adapt to their
environments and their inhabitants, whether the latter are individuals, groups or organisations.
Adaptive Buildings have been and are currently of great interest in the Architecture, Design and
Computing communities [5, 10, 13, 19, 40]. Adaptations in such buildings can be directly inhabitant
controlled or automated while those two mechanisms are frequently combined in the same structure.
In either case, at least in most modern buildings and architectural prototypes in the field, some form of
computational ‘building management system’ mediates between inhabitant input and desired
adaptation, whether this is through networked light switches or activity sensors, just to mention two
examples. In this context, automated adaptation generally tends to draw on networks of sensors,
middleware software and actuators controlling for example the internal climate, access or security and
therefore directly impacting on inhabitants in a variety of ways. In doing so, such buildings draw on a
set of internal and external sources of data, for example measuring environmental temperature,
occupancy levels, and inhabitant activity. One currently underexplored source of data for adaptations
in buildings is that of personal data, where such data might range from data held by individuals
themselves, data published online (e.g. personal web sites, social networking sites, etc.) and data
held by commercial and governmental organisations. From a technological perspective, the personal
data of building inhabitants can be made widely available in an architectural context. Whether it is
social networking profiles or for example tracking data detailing position and/or activities, such
information can be used to adapt buildings in a variety of ways. The resulting technologically driven
Adaptive Architecture responds to people whether that is for environmental, practical/organisational or
artistic purposes, and it might over time learn how to best respond.
Out of a possible range of personally held data, this paper focuses on one particular type:
physiological data. Physiological data such as heart rate, respiration or brain activity has been
considered in Human Computer Interaction research for some time. Such data also opens up
interesting avenues for design exploration in Architecture as well as particular challenges. At the
Mixed Reality Laboratory we have recently explored the application of such data in Adaptive
Architecture, the result of which being the prototypical ExoBuilding. ExoBuilding immerses inhabitants
International Adaptive Architecture Conference, Building Centre, London, March 2011: H. Schnädelbach 3
in their own physiological data, breathing with them and sonifying their heart beat [31]. Roughly room
sized, it allows people to explore their own physiology through voluntarily changing their behaviour,
while also producing measurable unconscious changes in that behaviour in a direct feedback loop.
This paper will present ExoBuilding as novel Adaptive Architecture driven by inhabitants’ physiological
data as an example of the wider possibilities in this design space. ExoBuilding very directly
demonstrates the possible feedback loops between adaptations in the built environment and
inhabitants and this will be discussed through an initial model considering the relationship between
inhabitants, physiological data, actuation and effect. A number of future possible case studies
introduce hypothetical mappings between physiological data and building fabric, and those emphasise
the need to consider adaptive architecture as socio-technical systems.
ExoBuilding – A building prototype driven by physiological data
Out of previous work applying physiological data in the theme park environment [34, 39], emerged a
very broad and blue-sky research question: what applications are there for physiological data in
Adaptive Architecture. ExoBuilding, a prototypical adaptive building, begins to explore this relationship
and begins to address some of the salient issues in this area. The following provides a brief overview
of the design and implementation process with a full description available in a previous publication
[31].
Iterative design and prototyping
At the beginning of 2009, a variety of mappings between technically available physiological data and
the building fabric were explored as a starting point, for example linking levels of skin conductance to
room ambiance or an individual’s body temperature to localised heating and cooling. In early
discussions with colleagues, it emerged that the most interesting aspect of the investigation was the
mapping of physiological data to building extent and form. Most unusual and therefore promising
seemed approaches that link a person’s respiration to building size and form and their heart rate to
sound output. Dynamic architectural form is also a ‘relatively’, i.e. outside of Adaptive Architecture,
underexplored aspect of the built environment and we were not aware of any other direct mappings
between physiological data and the building fabric.
As a first step in an iterative design and prototyping process, a to-scale prototype was constructed. A
simple wooden framework holds standard model servos over a miniature tent structure, which is in
turn pinned to the wooden support floor. Initially, pre-recorded data was used to drive the servo
motors. This was combined with playing a heart beat sound and triggering embedded LEDs in the
same frequency, roughly at 60 beats per minute.
Figure 1 ExoBuilding to-scale prototype
As a next step, the to-scale prototype was augmented with live physiological data. For that purpose a
physiological data monitoring kit (see below for more detail), detecting heart beat and respiration was
interfaced and the prototype driven with live data. Concretely, the following mappings were explored:
Firstly, live respiration was linked to the prototype size, i.e. if a person breaths in, the prototype
increases in size and when the person breaths out, the prototype decreases in size. Secondly, a
person’s heart beat was sonified through a speaker system and displayed through a set of red LEDs
International Adaptive Architecture Conference, Building Centre, London, March 2011: H. Schnädelbach 4
embedded in the wooden base. This proof-of-concept implementation demonstrated that the technical
feasibility of linking physiological data to the building fabric. Although even this limited prototype
sparked a lively discussion, it did not provide any sense of what inhabiting such a structure might be
like, as it had a very limited scale.
An inhabitable prototype
To be able to investigate inhabitation further, a larger prototype needed to be constructed and for that
reason, it was scaled up to roughly room-size, the main constraint being the size of one experimental
bay in the Mixed Reality Laboratory, at roughly 6x6x2.7 metres. The fabric was measured off the to-
scale prototype and scaled up using readily available stretchable jersey. The full-scale prototype was
constructed as a tent structure. It was installed in the lab, suspended from a ceiling-mounted sub-
frame. With a floor footprint of roughly 4.5x3.5 m and a height ranging between 1.3 -1.6 m, it is large
enough to sit in on a reclined chair or to lie in on the floor or a low bench. The ‘spine’ of the structure
is re-enforced, using aluminium tubing. At two points on that spine, the fabric is pulled up towards the
ceiling. At five points, the fabric is pinned down with cast iron stage weights.
The ExoBuilding’s main adaptive feature is the physical upwards and downwards movement of the
structure. Counterbalanced drive arms pull the fabric up and then release it back down, with the
tension in the fabric providing the downward pull. The arms are driven by large but standard model
servo motors strengthened by additional gear boxes which were in turn controlled through a Phidgets
interface toolkit. Phidgets provide standardised, plug and play sensing and control connected to PCs
via USB and they are widely used in digital prototyping [25].
Figure 2 ExoBuilding contracted (top) and expanded (bottom)
Figure 2 shows the range of the physical movement of the prototype, from 1.3m to 1.6m shoulder
height. The seemingly subtle change in size belies the effect felt inside ExoBuilding, as the change in
overall volume is much larger than the photos could illustrate. In addition to the movement, a data
projector can be used to project information on its surface. LEDs embedded into the fabric can be
used to display further information. Finally, a sound system can be used to display sound through
audio and associated vibrations of the floor. All four adaptive features (movement, projection,
embedded LEDs and sound) of the prototype are controlled through an open-source middleware
platform [9]. This platform allows the connection of sensors, software components and actuators using
a graphical programming approach. In this instance it is used to read in physiological data live,
suitably convert the data for live display and then pass it on to the actuators.
Physiological data and mapping to ExoBuilding
Physiological data from a single participant is acquired to drive the adaptations in ExoBuilding. This
was done live using a Mind Media NeXus10 device. The Bluetooth enabled and battery powered
device is easily portable. It offers 10 hardware channels which allow measurement of physiological
data (e.g. EEG, ECG, respiration, etc.) [23].Via Bluetooth it is connected to a PC running the
associated Biotrace software [22]. Biotrace takes the 10 hardware channels and provides a series of
live data channels. As an example, one data channel derives heart rate and heart rate variability from
ECG. Taking this one step further, the combination of data from different hardware sensors then also
allows the analysis of derived channels such as HR/Respiration coherence within the same software.
For driving the prototype adaptation, three physiological data channels were being used. An ECG
signal, is gathered using three electrodes placed on the participant’s chest and torso. Biotrace makes
International Adaptive Architecture Conference, Building Centre, London, March 2011: H. Schnädelbach 5
the signal available as heart rate, which is then converted to heart beat events inside the
aforementioned middleware platform. The heart beat is played through the speaker system using a
pre-recorded heart beat sample, and it is displayed on the tent fabric via the embedded LED. Via a
subwoofer the floor is made to vibrate in sync with the sound output. Respiration data was gathered
using a respiration belt fitted around the participant’s torso, measuring the rising and falling extent of
their abdomen. Through Biotrace and the middleware platform, this was converted to the full range of
the servo motors to be able to change the extent of the fabric structure. The participant’s respiration
drives the shape and size of ExoBuilding, so that the spatial volume expands during inhalation and it
contracts during exhalation. Finally, using two finger electrodes, electro dermal activity (EDA
(frequently called GSR)) was measured. EDA drives the display of a graphic on the outside of the
fabric. When their EDA signal rises, the image fades in and when the signal declines, the image fades
out. The following table provides an overview of the three mappings of data to actuation.
Sensor Signal Actuation
ECG Heart beat Heart beat sound and LED
RSP belt Extent of abdominal breathing Extent of ExoBuilding
EDA EDA Visibility of graphic projected on ExoBuilding
Table 1Physiological data - Actuation mapping
The end-result is a somewhat surreal, very much embodied experience where one’s own ‘general’
physiological activity and the physiological responses to the embedded interactivity are exposed to
one-self and those observing the experience. Only exploring the prototype itself can get this
experience across while a video provides a good initial indication (video illustrating single inhabitant
interaction at: http://www.youtube.com/watch?v=NAZPQDqr8aQ).
In addition, the following is an attempt to capture the most important properties of the prototype. In
summary, in can be described as:
Multi-sensory The data display is multi-sensory, as information can be seen (e.g. the
projected graphics and movement of the fabric), heard (e.g. the sound system) and felt (e.g.
vibrations of the floor, air flow generated by the moving fabric and the fabric occasionally
touching people).
Immersive The data display is immersive in that it physically immerses the body of an
inhabitant into the data to be displayed, in this particular case their own physiology.
Visceral Taken together, this resulted in an almost visceral experience. Especially when the
sound was turned up and the floor started to vibrate, participants in the formative study
reported that they felt that their whole body was affected by the experience.
A formative study with three participants then highlighted some interesting potential effects that the
prototype might have on people [31]. People reported a sense of relaxation after 3 minutes of use.
The physiological data highlighted falling EDA traces in addition to slower than average respiration
rates and deeper respiration amplitudes. Both can be interpreted as a physiological sign of relaxation,
suggesting that the prototype might act as an effective immersive biofeedback device, for example for
respiration training, which in turn is used in the mitigation of a number of health conditions. This area
remains an active research interest which is currently being pursued. To be able to make any
definitive statements in this area, more scientific evidence for the effectiveness of ExoBuilding as
immersive biofeedback device will be needed. The analysis of the data of a controlled experimental
study is currently ongoing and would go beyond the scope of this paper.
Related work
The prototype has been developed in an environment of increased interest in the use of physiological
data in interactive experiences. Beyond interest in human computer interaction [27], there have been
a number of arts projects that probe in this direction and it is worth focusing on a few examples that
were developed around respiration specifically. The project ‘Breathe’ involved the recording of a
International Adaptive Architecture Conference, Building Centre, London, March 2011: H. Schnädelbach 6
participant’s respiration to be played back to the person following, exploring how one affected the
other [15]. Kuth’s Drawing Breath allows a single participant to explore their respiration in detail in the
form of a graphically animated projection [18]. The same artist has also worked on a series of projects
exploring heart beat. Related to both the above, combining collaboration and graphics, the project
‘Lungs The Breather’ brings together four participants ‘collaborating’ in the indirect control of a visual
display through their breathing [12].
In Architecture, the term ‘breathing’ is most conventionally associated with the way that a building is
ventilated, with a ‘breathing building’ being linked to façade designs that do not hermetically seal of
interior from exterior. There are also architectural installation projects and investigations that play with
the notion of breathing in a much less utilitarian sense. Hylozoic Soil [4] envelopes inhabitants in a
series of sensors and actuators, and the author describes the resulting overall breathing movement of
the piece in [2]. In related work, Thomsen’s Vivisection embeds sensors in an interactive fabric
installation separating a larger space into three responsive breathing chambers driven by large fans
[38]. Those are in turn triggered by a network of embedded proximity sensors. To the best of the
author’s knowledge however, no direct mappings of physiological data to actuated spatial enclosure
have previously been made either in arts or architecture explorations.
With the aim to expose some of the overall general possibilities in this space/ architectural design the
discussion now turns to types of physiological data that might be architecturally relevant, methods of
their technical acquisition and the possible mappings between that data and the architectural building
fabric, before considering the overall effects that can be created.
Physiological data in Adaptive Architecture – A model
Physiology ‘concerns the study of bodily function or how the parts of the body work’ [6, p.2]. One way
of studying physiology is by acquiring, recording and analysing physiological data that can be
recorded about a person. Such recordings are standard in medical practice and were historically done
manually, for example through a doctor taking a patient’s pulse, taking note of counts per time unit
and recording this in one form or another. Today there are many technical means to acquire
physiological data and it is those that are most relevant in this context. This is because the data
needs to interface with the building fabric, which is in turn driven by a technical infrastructure itself.
What types of data are then available in a building context then depends largely on how this data can
be captured and one might distinguish two overall forms.
Acquisition
In direct acquisition, data is acquired directly about inhabitants’ physiology, as exemplified by
ExoBuilding. There is a variety of wireless data acquisition devices similar to the one used for the
prototype that could make live physiological data available in a building context. Very frequently,
those devices provide the Electrocardiogram (ECG) (e.g. for heart rate, heart beat information, heart
rate variability), respiration information (e.g. for respiration rate, respiration – heart rate coherence),
the Electromyogram (EMG) (e.g. for the detection of facial muscle movements detecting smiles and
frowns) and Electroencephalography (EEG) (e.g. for the description of brain activity), among other
data streams. At present such data acquisition devices tend to be single-user, battery powered and
they need to be in range of a wireless networking infrastructure to provide real-time data. In addition,
there are other technologies to measure physiological data for example ultrasound and MRI scanners,
but they are (currently at least) too impractical to be considered for general use in the built
environment.
In indirect acquisition, data is acquired by measuring the effects that physiological activity of
inhabitants has on something else, typically the environment. A simple example is the operation of
modern HVAC systems that would measure internal environmental conditions to adapt their operation,
e.g. starting to cool when temperature has been pushed up in a particular room by an increase in
occupant numbers. Khan makes use of a similar strategy in his Open Columns project [17] for a much
more interesting purpose. In this project, ceiling mounted adaptive columns react to C02 levels in the
room under consideration. When C02 levels are pushed up in the room, the available space is
automatically decreased having the effect that people are being dispersed.
International Adaptive Architecture Conference, Building Centre, London, March 2011: H. Schnädelbach 7
In both acquisition scenarios, direct and indirect acquisition, a number of potential sensor placements
are conceivable. Sensors might be implanted and wirelessly received and transmit data, not
necessarily a far-fetched idea, considering the fact that artificial pacemakers have been fitted since
the 1960s. Sensors might be placed on the body, in a similar way to the device described in use with
the ExoBuilding. There are already new techniques available that make the sensing of some
physiological data in this way less intrusive to the person concerned. Finally, physiological data
monitoring might become embedded into the architectural infrastructure. The Open Columns project
introduced above provides an example for such an approach for indirect acquisition, in that case
detecting CO2 levels. For direct acquisition there are for example some approaches that detect
respiration through calibrated cameras [41].
Raw data – Data interpretation
Physiological data, however it was acquired, can then be understood as one component of personal
data that can be used in making buildings adaptive. Such data can be made available to an
architectural infrastructure in its raw form or be interpreted in some way. For example, in ExoBuilding
the raw respiration trace was mapped directly to building fabric movement, while heart beat, mapped
to sound and light output, was already one step removed from the ECG signal that was acquired (this
interpretation happened to occur in the associated software infrastructure). Data operations in this
context might also include averaging over time, the combination of data streams as in for example the
coherence between two sets of and very importantly in the context of building data the aggregation of
data emanating from multiple inhabitants, as explored in the project ‘A Sophisticated Soirée’ [3].
An entirely different and altogether more challenging approach is to attempt to connect psychological
state from physiological data, to make inferences from such data to understand moods, sentiments,
frustrations and elations [6]. In human computer interaction (HCI) this has found prominence in the
area of ‘Affective Computing’, proposed originally by Rosalind Piccard [26]. Her starting point was the
realisation that intelligent behaviour cannot be de-coupled from emotional competence, resulting in
the idea that for the creation of intelligent machines, the development of affective computational
behaviour is necessary. In parallel to this is the drive to provide computational devices with the ability
to display emotional behaviour or affect ‘themselves’ with the aim to improve our interaction with
them. In addition to physiological data, affective computing might then also draw on other data
streams to be interpreted, for example that of facial expression and body language that would be
acquired through camera recordings. In Adaptive Architecture, instead of the building fabric being
driven by respiration for example, it might be driven by an interpretation of the inhabitant’s mood, for
example providing spatial openness with a confident state of mind and more enclosure with a hesitant
state of mind. Affective computing has seen some sustained criticism however, because of the
difficulty of deriving emotional state from physiological data alone. At present, it is simply not clear
how emotion in its entirety, i.e. including intentions, appraisals and feelings in a particular context, can
be observable at all [8]. Just like ExoBuilding, other projects in HCI have therefore responded by
avoiding the representations of emotions altogether, for example [36] and [24]. Even if not practically
feasible in its full vision to date, the potential to infer emotional state from the interpretation of
physiological data opens up interesting possibilities in Adaptive Architecture.
Whether based on raw or interpreted data, the creation of an adaptive effect in a building needs to
consider the possible ways for actuation. Both effect and actuation will briefly be discussed below,
drawing on and expanding the author’s previous work on a categorised framework of Adaptive
Architecture [30].
Actuation
In an adaptive building, physiological data can be mapped to a variety of available actuators, where
‘actuator’ stands for a technical device that can create a desired effect in the environment enclosed by
the building. Some possible mappings have been considered as a first step in the ExoBuilding
prototyping process [31] and some more concrete examples will be considered in the case study
section below. In more general terms, physiological data might be used to drive a building’s lighting
system (e.g. switching, light levels, light temperature), it can be used to drive the sound infrastructure
(e.g. changing tune, volume), it might trigger changes in data flow and communication (e.g.
establishing a video link to a particular person), it might be used to control media displays (e.g.
bringing up additional information), it might drive architectural components and elements (e.g.
opening a roof or partition, sliding away seating in a theatre), it might trigger environmental controls
International Adaptive Architecture Conference, Building Centre, London, March 2011: H. Schnädelbach 8
(e.g. changing settings on the HVAC) and finally, physiological data could potentially be used to
control resource supply (e.g. controlling the mains water). As will be immediately clear, the actuators
briefly raised above are already present in buildings today and are typically controlled based on other
information. However, they are also principally available to be interfaced with physiological data.
Effect
Actuation in the sense it is described above is then used to create an effect in the environment
enclosed by a building. Such effects can be more or less directly mapped to what can be actuated
and therefore one might expect changes in the lighting and sound landscape, the density of
information that a space holds, its physical form and size, its permeability (e.g. to what extent an
inhabitant can traverse through the architectural configuration), its architectural topology and layout
and finally its environmental properties (e.g. its air quality, temperature). Such changes to an
architectural environment are ultimately designed to have an effect on its inhabitants and those might
be briefly summarised as levels of inhabitant comfort, convenience, usability of an environment,
spatial access, access to information, safety and security and organisational flexibility.
Of particular interest in the context here are the effects on inhabitants that are specific to physiological
data and therefore the argument is returning to some of the properties of the original ExoBuilding
prototype exploring this area. Buildings driven by physiological data will result in a feedback loop
between inhabitant and building fabric under certain conditions. These conditions are assumed to be
that the biofeedback is legible (i.e. an inhabitant can understand that an effect created in the
environment is linked to them) and that the data is not aggregated from multiple inhabitants, which is
very much connected to the first condition. Such feedback can be presented in the model illustrated in
Figure 3, pulling together the categories discussed above.
Figure 3 Model: Physiological Data in Adaptive Architecture
In this model of physiologically driven Adaptive Architecture, physiological data is acquired from
inhabitants; it is manipulated, sometimes combined with other personal and environmental data, and
sometimes interpreted to then feed into actuation of the spatial enclosure. Actuation results in
changes in the environment which in turn affect the inhabitants of the environment. As has been
highlighted previously, if those changes are indeed legible on some level, a biofeedback loop is
established, where the physiological data ‘displayed’ through architectural elements impacts on the
physiology of the person that the data is acquired from.
International Adaptive Architecture Conference, Building Centre, London, March 2011: H. Schnädelbach 9
Building types driven by physiological data
As the section on the acquisition of physiological data has demonstrated, the currently available
equipment means that such data is not available for general purpose use today. It is simply not
practical or desirable from a technical and from a social perspective for all inhabitants to be wearing
the necessary equipment. However, it is clearly conceivable that an infrastructure could exist in future
where the acquisition of inhabitants’ physiological data will be unobtrusive and possibly even
undetectable. Such a deployment would raise significant ethical issues, a detailed discussion of which
would go beyond the scope of the paper, while the paper will briefly return to the issue in the
discussion. In such a future scenario, physiological data would be technically available within a
building to be used in a variety of ways. In what follows, the paper steps through a number of future
possible case studies, making that assumption. Without worrying about how the data is acquired or
how it is manipulated, the cases are thought experiments addressing where work in this area might be
heading in future.
Architecture of Hope
Biofeedback as a strategy in the built environment might well present its own focussed application
area. There is good evidence that conventional Architecture can have a direct impact on recovery
from illness. Building on this evidence, hospital environments can be re-designed to better support
patients as well as staff supporting those patients. Measures can be as simple as providing better
privacy, quieter spaces and views of the outside environment [20]. Drawing on this background
evidence but also motivated through personal experience, Jencks founded the Maggie’s centre
initiative [16]. Maggie’s centres are attached to hospitals and provide for terminally ill patients in a
non-hospital, welcoming and ‘domestic’ environment. At the time of writing there were seven such
centres operating in the UK putting into practice the idea that the architectural environment is key to
patient recover.
What about Architecture taking a much more pro-active role in recovery, it becoming an active part in
the journey from illness to health? As the name ExoBuilding suggests, the idea has been from the
outset that the building envelope takes on a supporting function in some way, similar to the now
historical concept of a technical Exoskeleton. Such a supporting role might take on two forms. Firstly,
there is a more traditional medical use, in which environment-embedded physiological monitoring
could support carers and patients to monitor recovery progress and identify emergency situations,
something explored in Assisted Living HCI research. One can also imagine much more pro-active
support in learning biofeedback techniques that have been demonstrated to be effective in the
treatment of a variety of medical conditions, such as stress, anxiety disorders, asthma and insomnia.
For this purpose, actuated biofeedback environments would be integrated into dedicated ‘treatment’
rooms in a way that is unobtrusive and sympathetic to the architecture. There is also the possibility to
link to other such rooms (local and remote) for collaborative exercising.
The Millennium Dome
With the first case the discussion has not really strayed very far from the use of the original use of the
ExoBuilding prototype. Simply because it is the largest fabric structure that is around, the London
Millennium Dome might serve as an entirely different basis for departure and seemed the perfect case
example to consider respiration feedback on a massive scale [29]. In addition, the original science
centre and event space actually contained a ‘Rest Zone’, exhibiting constantly changing colours and
gentle music, to where visitors could retreat and relax [42]. With the original Millennium Experience
heavily criticized, the structure is now very successfully being used as the O2 arena, a concert and
event venue.
International Adaptive Architecture Conference, Building Centre, London, March 2011: H. Schnädelbach 10
Figure 4 Millennium Dome, Greenwich UK
Here one might consider a physically actuated fabric surface that is capable of moving in a given
rhythm. Given the overwhelming size of the dome and its current use as an entertainment venue, the
mapping of an individual’s physiological data to the structure would not make any sense. Instead, the
aggregation of data might need to be considered. Considering an occasion where it is used to full
capacity, the fabric movement might become linked to a self-selected (addressing the issue of
consent) but somehow representative sub-set of occupants taking part in a particular event. The
structure would move in synchrony with an averaged respiration taken from all measured participants,
with their heart beat being used to drive the beat of the music being mixed on stage. In this way, the
building and occupants become fundamentally embedded into a shared cultural experience which is
clearly visible to people outside and around the building, giving it an animate appearance and
reflecting what is going on inside. Unfortunately, this would probably never work technically, as
human respiration rates are too high to be successfully transmitted to a structure of this size, not even
considering the air movement caused by this design.
Institut du Monde Arabe
Switching context and to a more manageable scale, the following example considers the ‘Institut du
Monde Arabe’, a cultural centre celebrating Arab culture across the world in the centre of Paris. It’s
most striking and well-known feature is the mechanically adaptive main façade [35]. Across more than
100 square panels, light sensors were designed to dynamically open and close camera like
diaphragms to control thermal exposure, the interior lighting and views from inside to outside to some
extent as well. The centre represents an early example of a media façade, although the façade itself
has never operated reliably since the building was opened.
Figure 5 Iris Facade
International Adaptive Architecture Conference, Building Centre, London, March 2011: H. Schnädelbach 11
The iris metaphor provides the starting point for speculating about building elements that are linked to
their inhabitants’ iris but also eye lid movements (blinking and closing one’s eyes). For an individual, a
single façade element might recognise something about an inhabitant, such as their current mood to
provide the right level of illumination for specific activities. Linking the mechanical iris to the movement
of a human eye lid might seem unpractical because of the frequency and speed of the blink reflex but
it would offer interesting possibilities. The interior illumination from natural light of a building would dip
directly in response to someone shutting their eyes. This link would also give the building an
immediately recognisable animate appearance. If averaged over time, i.e. if the blink reflex is taken
out, the building façade would publicly proclaim when someone is asleep, while also providing them
with the right light conditions for sleeping, which might be particularly relevant in the bright summers
of northern hemisphere. Moving away from an individual to groups of inhabitants, a strategy in this
area would have to involve some form of data aggregation. It is also useful to consider other forms of
physiological data in this context. For example, a building might detect the state of alertness of groups
of people in different parts of building. The façade might adjust the level of daylight reaching the
interior in an attempt to influence that state of alertness in the desired way.
The Generator Project
Approaching a more hypothetical example, Price’s Generator project (1976-79) foreshadowed
ubiquitous computing technologies embedded into architectural infrastructure [28, 37]. His overall aim
was to create a re-configurable architectural layout that would be modified to suit a variety of functions
and purposes. It was designed to be responsive in that occupant preferences were taken into account
in deciding what layout to establish. The design revolved around 150 4x4m standardised architectural
units that could be arranged in different combinations, connected by walkways and ramps. Occupants
interactively and collaboratively with a ‘facilitator’ established layouts to be implemented and units
would be lifted in place by mobile crane controlled by a human operator. As a further extension to this
original idea, each unit was to have an embedded chip to allow an associated computer process to
keep track of units, support occupants to design new layouts and keep track of layout changes over
defined periods with the aim to suggest new layouts and identify inefficient use of layouts. This latter
intervention was designed to allow Generator to ‘surprise’ its inhabitants, with new layouts that the
system had identified.
The topological flexibility that forms the basis for this project idea lends itself to be explored in the
context of this paper. Spatial relationships very frequently express social relationships. The way these
social relationships become embedded into building topologies is a key focus of architecture as a
discipline and has for example been formally explored by Hillier [14]. The way that inhabitants feel
about spatial relationships in a given building, whether fixed or emerging, might just be detectable in
physiological data, or at least one could imagine that it could be by evaluating the relationship
between spatial topology and inhabitant emotional state. In a project like Price’s generator, the
building might then propose new and interesting layouts that are designed to have a positive effect on
people’s mood, with the overall topology constantly evolving. Where in a physical project the
automation of such a process remains a significant engineering challenge, this becomes a lot simpler
where the spatial relationships are provided by digital technologies. In the author’s own work on
Mixed Reality Architecture, spatial connections between local and remote spaces are generated
across a shared virtual 3D space by inhabitants on the fly [32, 33]. Automatically adapting such virtual
connections based on physiological response is a directly feasible possibility. Decision could be
based on attempting to detect how much company a person might desire at a particular point in time
or with whom they might want to be in contact with.
Laban Dance Centre
The final case considers the Laban contemporary dance centre in South East London by Herzog and
De Meuron [21]. It is one of the largest dance academies in the world and a leading dance training
centre. The building combines rehearsal studios, theatres for performances and office spaces to
provide for the academy. One of the key features of the building is the semi-translucent façade, which
is punctuated by windows where a view is required in the interior. The level of translucency provided
affords the building media façade like quality as the normal activities during rehearsals and
workshops become visible from the outside, whereas during the night the internal lighting is visible
from some distance across the local area. The internal activity of dance rehearsals and performance
are therefore carried to the outside of the building, augmenting the façade and embedding the internal
activity within the exterior appearance of the building.
International Adaptive Architecture Conference, Building Centre, London, March 2011: H. Schnädelbach 12
Figure 6 Laban Dance Centre (Image ©arcspace.com)
With the acquisition of data about the dancers, entirely new possibilities emerge outside of direct and
synchronous visual representation. Dancing will raise physiological responses such as the heart rate
and breathing rate of dancers, for example. It is also conceivable to acquire and record the movement
of a dancer by recording muscular responses or possibly through the use of other types of tracking
equipment. Such data could then be mapped to visual façade effects like lighting and this might be
live during rehearsals but it could also be replayed at times when there is no activity. In the particular
case of the Laban centre one might also think about playing with the translucency of the overall
facade in addition to the positions of the transparent façade elements, providing views in and out at
different locations depending on activity. Taking this a step further, the physical movement detected
during dance might be linked much more directly to the physical movement of the building or at least
parts of the building, creating a more fluid building form. Taken together the mappings proposed here
would allow the creation of a building the appearance of which emerges in a joint choreography
between live and recorded dance and live and recorded building responses.
Summary and outlook
This paper has concentrated on the role of personal data, more specifically physiological data, and
the digital technologies mapped to personal data in making buildings adaptive. Through the iterative
prototyping of a built structure and the exploration of the possible mappings between physiological
data to that structure in the context of related work, the concept has been demonstrated to be
feasible. The possibility of a biofeedback loop has emerged through this work and the effectiveness of
this is currently being explored by the author. Work with the ExoBuilding prototype has then also led
to the formulation of an initial model of physiological data in adaptive architecture. This model
illustrates the circular relationship between inhabitant, physiological data, actuation and architectural
effect, with biofeedback being one special type of feedback. The discussion of this broader model has
then allowed the consideration of a number of possible future case studies in this context.
Reyner Banham argued that architecture cannot be understood without understanding its service
technology and presented a history of the influence of such technologies on the design of the built
environment [1]. Technology in general as well as technology specifically designed for buildings has
become digital in the mean time. Beyond creating a well-tempered environment such technologies are
supporting the drive to make buildings adaptive to their environment and to their inhabitants, which in
turn can be motivated by a whole variety of different issues [30]. Architecture can also clearly not be
understood without its inhabitants and the nature of physiological data provides new challenges as
well as opportunities in this area.
One of the key challenges is that physiological data is inherently private and not generally available
for public consumption. Previous work in the public arena has already raised associated issues in this
International Adaptive Architecture Conference, Building Centre, London, March 2011: H. Schnädelbach 13
context for example around the possible ‘serendipitous’ diagnosis of medical problems and in
situations that are out of control of the observing party [34]. In addition, apart from people who have
an inherent interest in it (such as sports people or medically trained people), the nature of
physiological data and its associated patterns is unknown to most people. This even applies to their
own responses. For most people, they will only be in contact with their own externalised physiological
responses when they see their GP or in a medical emergency. Just as an example, previous projects
have shown that the variability of heart rate caused real concerns with participants, although it is
entirely natural, simply because they were not aware of the facts. Making such data available more
generally within the building raises additional issues, especially as this could and probably would
include some recording functionality, with all the related issues around data leaks to third parties. In
this context, it will be absolutely essential to consider entirely new strategies for giving inhabitants
control over their personal data, whether physiological or otherwise, in addition to providing full
transparency over how any such data might be used within the built environment.
Finally, it is proposed here that orienting the understanding of Adaptive Architecture in relation to its
inhabitants could draw on the established (in HCI) concept of socio-technical systems, as a useful
vehicle for further discussion. In their original form, socio-technical systems describe organisations in
terms of the technology they use as well as in terms of the social organisation that is employed [7],
[11]. Critically, changing one of the two core parts cannot be achieved without impacting on the other.
In the same sense, buildings must be understood as complex socio-technical system where they
operate or don’t operate, because of the quality of the relationship between technology (the building
fabric and technical systems) and social organisation (organisational decisions, preferences and
knowledge). In this context the need for a better understanding of that relationship arises and the
general availability of physiological data in the built environment provides a real opportunity here. Its
acquisition and recording makes the ‘direct’ study of inhabitants possible which could be integrated
into a broader methodology set for post-occupancy evaluations. Such a methodology would help to
build up a much richer picture of the inhabitation of Adaptive Architecture, whether it is driven by
physiological or by other data.
Acknowledgements
ExoBuilding was developed at the Mixed Reality Lab, The University of Nottingham with the generous
support from the Leverhulme Trust. The author also would like to acknowledge funding for this work
which was provided by the UK Engineering and Physical Sciences Research Council through the ‘The
Challenge of Widespread Ubiquitous Computing’ grant (EPSRC grant EP/F03038X/1). The software
infrastructure is built around the open-source Equator Component Toolkit, developed during the
EQUATOR IRC.
International Adaptive Architecture Conference, Building Centre, London, March 2011: H. Schnädelbach 14
References
1. Banham, R. The architecture of the well-tempered environment. University of Chicago Press, Chicago, 1969.
2. Beesley, P. and Khan, O. Responsive Architecture / Performing Instruments. Situated Technologies Pamphlets, 2 (4).
3. Berger, E. A Sophisticated Soirée Ars Electronica Festival, 2001, 352-353.
4. Bonnemaison, S. and Eisenbach, R. Installations by architects : experiments in building and design. Princeton
Architectural Press, New York, 2009.
5. Bullivant, L. (ed.), 4dspace: Interactive Architecture. Wiley-Academy, 2005.
6. Cacioppo, J.T., Tassinary, L.G. and Berntson, G.G. (eds.). Handbook of Psychophysiology. Cambridge University
Press, Cambridge, UK, 2007.
7. Coakes, E., Willis, D. and Lloyd-Jones, R. (eds.). The New Socio Tech: Graffiti on the Long Wall. Springer, Berlin,
Germany, 2000.
8. Cohn, J.F. Foundations of human computing: facial expression and emotion 8th international conference on
Multimodal interfaces, ACM Press, Banff, Alberta, Canada, 2006.
9. Egglestone, S.R. Equip Project Homepage. http://equip.sourceforge.net/, The Mixed Reality Laboratory, University of
Nottingham, accessed 13 06 2006
10. Fox, M.A. and Kemp, M. Interactive architecture. Princeton Architectural Press, New York, 2009.
11. Fox, W.M. Sociotechnical System Principles and Guidelines: Past and Present. Journal of Applied Behavioral
Science, 31 (1). 14.
12. Guerra, L.C., Todoroff, T. and Sebti, Y. Lungs The Breather. http://www.thebreather.org/4 February 2011
13. Harper, R. Inside the smart home. Springer, London ; New York, 2003.
14. Hillier, B. and Hanson, J. The social logic of space. Cambridge University Press, Cambridge, UK, 1984.
15. Jacobs, M. and Findley, J. Breathe.
http://www.fundacion.telefonica.com/at/vida/vida10/paginas/v4/ebreathe.html23/04 2010
16. Jencks, C. and Heathcote, E. The architecture of hope: Maggie's cancer caring centres. Frances Lincoln, London,
2010.
17. Khan, O. Open Columns | A CO2 actuated responsive architecture.
http://cva.ap.buffalo.edu/opencolumns/?page_id=31, University of Buffalo, accessed 10/02 2011
18. Khut, G. and Tonkin, J. Drawing Breath, 2005-2006. http://georgekhut.com/2005/09/drawing-breath/4 February 2011
19. Kronenburg, R. Flexible : architecture that responds to change. Laurence King, London, 2007.
20. Lawson, B. and Phiri, M. The Architectural Healthcare Environment and its Effects on Patient Health Outcomes,
2003.
21. Mack, G. Herzog and De Meuron 1997 - 2001: The Complete Works, Volume 4. Birkhauser Verlag AG, 2008.
22. Mind Media B.V. Mind Media B.V. Information about BioTrace+ Software.
http://www.mindmedia.nl/english/biotrace.php, Mind Media B.V., accessed 02/09 2010
23. Mind Media B.V. Mind Media B.V. Information about the Nexus-10. http://www.mindmedia.nl/english/nexus10.php,
Mind Media B.V., accessed 09 02 2010
24. Moodjam Research Group Moodjam Research Group. http://www.moodjam.org 10 08 2007
25. Phidgets INC. Phidgets INC. : Unique and Easy to Use USB Interfaces. www.phidets.com, Phidgets INC, accessed
12 06 2006
26. Picard, R.W. Affective computing. MIT Press, Cambridge, Mass., 1997.
27. Picard, R.W., Tan, T., Tao, J. and SpringerLink (Online service). Affective Computing and Intelligent Interaction First
International Conference, ACII 2005, Beijing, China, October 22-24, 2005, Proceedings, Springer-Verlag GmbH.,
Berlin Heidelberg, 2005.
28. Price, C. The Square Book. Wiley&Sons, Chichester, UK, 2003.
International Adaptive Architecture Conference, Building Centre, London, March 2011: H. Schnädelbach 15
29. Rogers Stirk Harbour + Partners. Millennium Experience, Rogers Stirk Harbour + Partners, London, 2007.
30. Schnädelbach, H. Adaptive Architecture - A Conceptual Framework MediaCity, Bauhaus-Universität, Weimar,
Germany, 2010.
31. Schnädelbach, H., Glover, K. and Irune, A. ExoBuilding - Breathing Life into Architecture NordiCHI, ACM Press,
Reykjavik, Iceland, 2010.
32. Schnädelbach, H., Penn, A. and Steadman, P., Mixed Reality Architecture: A Dynamic Architectural Topology. in
Space Syntax Symposium, (Istanbul, Turkey, 2007), Technical University Istanbul, 10601-10614.
33. Schnädelbach, H., Penn, A., Steadman, P., Benford, S., Koleva, B. and Rodden, T., Moving Office: Inhabiting a
Dynamic Building. in CSCW, (Banff, Canada, 2006), ACM Press, 313-322.
34. Schnädelbach, H., Rennick Egglestone, S., Reeves, S., Benford, S., Walker, B. and Wright, M., Performing Thrill:
Designing Telemetry Systems and Spectator Interfaces for Amusement Rides. in CHI, (Boston, USA, 2008), ACM
Press, 1167-1176.
35. Schwartz, M. L’Institut du Monde Arabe Architecture in the Light of Day Southfield, USA, 2010.
36. Sengers, P., Boehner, K., Mateas, M. and Gay, G. The Disenchantment of Affect. Personal and Ubiquitous
Computing, 12 (5). 347-358.
37. Steenson, M.W. Cedric Price's Generator. Crit - Journal of the American Institute of Architecture Students.
38. Thomsen, M.R. Robotic Membranes - Exploring a Textile Architecture of Behaviour. in Protoarchitecture - Analogue
and Digital Hybrids, John Wiley & Sons, London, 2008, 92-97.
39. Walker, B., Schnädelbach, H., Rennick Egglestone, S., Clark, A., Orbach, T., Wright, M., Ng, K.H., Rodden, T.,
Benford, S. and French, A., Augmenting Amusement Rides with Telemetry. in ACE, (Salzburg, Vienna, 2007), ACM
Press.
40. Weiser, M. The Computer for the Twenty-First Century. Scientific American, 265 (3). 94-104.
41. Wiesner, S. and Yaniv, Z., Monitoring Patient Respiration using a Single Optical Camera. in Engineering in Medicine
and Biology Society, 2007. EMBS 2007. 29th Annual International Conference of the IEEE, (2007), 2740-2743.
42. Wilhilde, E. The Millennium Dome: The Official Book of the Dome. HarperCollins Illustrated, London, UK, 1999.
