‘Cognitive facility management’: Definition, system architecture, and
example scenario
Jinying Xu1, Weisheng Lu1*, Fan Xue1, and Ke Chen2
* Corresponding author, email: wilsonlu@hku.hk
This is the peer-reviewed post-print version of the paper:
Xu, J., Lu, W., Xue, F., & Chen, K. (2019). ‘Cognitive facility management’:
Definition, system architecture, and example scenario. Automation in Construction,
107, in press. Doi: 10.1016/j.autcon.2019.102922
The final version of this paper is available at: https://doi.org/10.1016/j.autcon.2019.102922.
The use of this file must follow the Creative Commons Attribution Non-Commercial No
Derivatives License, as required by Elsevier’s policy.

Abstract:
Prevailing facility management (FM) discourses have recognized the importance of efficient
facilities to improve the quality of life and the productivity of business. Advanced technologies
have elevated facilities from ‘brick and mortar’ to ‘intelligent beings’. To date, facilities have
become more anthropomorphized, imbued with cognitive capability akin to humans, e.g., able
to perceive, learn, and act. However, the development of ‘cognitive FM’ remains in its infancy.
This paper attempts to put forward the concept of cognitive FM by providing a working
definition and articulating its key elements, including perception, learning, and action. An
eight-layer system architecture is proposed to facilitate the implementation of cognitive FM.
Following that, a demonstrative scenario called ‘Event Manager’ is utilized to showcase the
potential applications of such cognitive FM. The paper contributes to the body of knowledge
by advancing the stagnant FM discourses defined and subsequently confined by the
smart/intelligent building language three decades ago. It opens a new avenue for both
researchers and practitioners to better investigate and value FM as a cognitive system.
Keywords: Facility; Facility management; Cognitive system; Cognitive facility management;
Proactive intelligence
1. Introduction
A facility is an asset that is built, installed, or established to serve our social and economic
activities (Kaplan et al., 2004). Facilities include different types of physical infrastructures (e.g.,
bridges, roads, and stadiums) and buildings (e.g., hospitals, shopping malls, and housing). The
growing complexity of modern facilities increases the responsibilities and importance of
1 Department of Real Estate and Construction, Faculty of Architecture, The University of Hong Kong,
Pokfulam, Hong Kong
2 Department of Construction Management, Huazhong University of Science and Technology, China
1

management. The International Standard Organization (ISO) (2017) defines facility
management (FM) as an organizational function which integrates people, place and process
within the built environment for the purpose of improving quality of life and productivity of
business. Drion et al. (2012) asserts that FM is a management function that focuses on how to
develop, maintain and improve the physical assets needed to support and add value to the
business processes of organizations, as well as create and maintain a physical workplace that
provides optimal support to the people and work of the organizations. One of the legitimate
scopes of FM is to facilitate resources use, enhance organizational effectiveness, and improve
our living and working environment for social and economic activities (Rondeau et al., 2012).
Compared with other phases in a facility’s life cycle, FM can last for a relatively longer period,
roughly thirty to fifty years if not longer (Atkin and Brooks, 2015). FM must be well conducted
to realize a facility’s designated functions. The efficiency and effectiveness of FM significantly
affects operation cost (Teicholz, 2013), energy consumption (Bilgen, 2014), indoor comfort
and in grander scheme global climate (Nielsen et al., 2016). The importance of FM is
underpinned in Winston Churchill’s famous quote (Cited in O'Toole and Lawler, 2007), “We
shape our buildings, thereafter they shape us”.
Over the last three decades or so, smart facilities and their management captured the basis of
FM discourse, especially concerning the lasting and heated discussion around smart building
and intelligent building. Adding automation and intelligence features to tradition, standard,
physical and largely dumb ‘brick and mortar’ renders them ‘intelligent’, providing more
efficient, responsive, and environmentally friendly services to occupants. Advanced
technologies, software, and systems directed at FM stimulated a boom in ‘intelligent’ initiatives
and innovations. Examples of such systems include the widely adopted CMMS (computerized
maintenance management system), CAFMS (computer-aided facility management systems),
BAS (building automation system), CAS (communication automation system), SAS (security
automation system), FAS (firefighting alarm system), pervasively discussed BIM (building
information modelling) technology, and smart public transportations system.
Nevertheless, current FM practices, be they routine maintenance or ad hoc event management,
are largely backward and characterized by a low level of automation and intelligence (Pärn et
al., 2017; Wu et al., 2014). Only able to operate on preset conditions, they cannot adapt to
complex scenarios. They often perform passively and in an isolated manner (Lau et al., 2013).
Different FM systems (e.g., CAS, BAS, FAS, and SAS) do not share information with each
other. Decision-making involved in current FM relies heavily on manual input, setting and
operation, which can be fallible due to the limited computation capacity and bounded rationality
of human beings (Niu et al., 2016). Moreover, current FM generally pays scant attention to the
differentiated and changing requirements of occupants in a facility. Therefore, it is not
uncommon that a facility fails to provide satisfactory services to people, organizations, and
businesses it serves (Wang et al., 2018). These predicaments are resulted from the passiveness
2

of current FM systems. Various passive FM systems are pre-programmed, which cannot
respond to complicated, flexible, changing situations in real life. FM needs to be updated with
intelligence to a higher level akin to human beings’ cognitive capability (e.g., to perceive, to
learn, and to act). Although some recent research efforts have explored the application of
cognitive capability to FM, none have been able to provide a detailed, systematic account.
This study aims to take the very first step to define a new concept, cognitive FM, which
empowers cognitive capability into traditional facilities and their management. It tries to answer
the following research questions: (1) what is cognitive FM? (2) what are its key elements? (3)
what is its system architecture? and (4) how does it influence FM? The remainder of this paper
is organized as follows. Section 2 reports related work on concepts related to cognitive FM and
cognitive computing, and emerging computation technology. Section 3 discusses the definition,
key elements and enabling technologies of cognitive FM. The proposed system architecture of
cognitive FM is illustrated in Section 4. Section 5 proposes an application scenario of cognitive
FM. Discussion and conclusions are made in Sections 6 and 7, respectively.
2. Literature review
2.1 Concepts related to cognitive facility management
Smart facilities and their management form the crux of FM literature and practices, especially
as relates to contentious concepts like the smart home, ambient intelligence, smart building, and
intelligent building. A ‘smart home’ constitutes a residence able to anticipate and respond to
the needs of its occupants, enhancing their comfort, convenience, entertainment, and security
with the support of technology (e.g., Information Communication Technologies [ICTs] or
Internet of Things [IoT]) in the home environment and connects to the world beyond (Harper,
2006). A wealth of research and development has been dedicated to smart electronic devices
and home appliances for smart homes. However, smart homes, by and large, can only react to
given requirements and conditions. They cannot effectively perform functions outside their
programmed range.
‘Ambient intelligence’ characterizes an environment embedded with technologies but in natural
surroundings. Sensitive, adaptive, and responsive to the presence of people and objects, it
exploits smart non-explicit assistance to augment activities (Weber and Rabaey, 2005). Using
sensors or sensor networks, ambient intelligence perceives the state of a setting, situation and
the people within them. Ambient intelligence can greatly influence people’s lives through
applications such as health monitoring assistances, emergency services, and intelligent
workplaces. However, ambient intelligence systems have to interact sensibly with users in a
variety of sophisticated ways to learn their preferences, intentions, and needs, which brings
intervention, privacy and security challenges (Cook et al., 2009).

3

‘Smart building’ and ‘intelligent building’ are often used interchangeably. Smart buildings
integrate and account for intelligence, enterprise, control, and materials and construction as an
entire building system, with adaptability, for building progression (Buckman et al., 2014).
Whereas intelligent building refers more to a dynamic and responsive architecture that provides
every occupant with productive, cost-effective and environmentally approved conditions
through a continuous interaction among its basic elements, i.e., structures, systems, services
and management, as well as their interrelations (Ghaffarianhoseini, et al., 2016). These
definitions suggest that the two terms share similar vision. Their implementations are more on
the updating of hardware and the adoption of ICTs, and still largely rely on passive control by
human beings.
It is difficult to trace the origin of the phrase ‘cognitive building’. In 2016, IBM applied it in
relation to machine learning, which makes it more of an industry concept than a scholarly
defined one. Cognitive buildings exhibit an integrated approach to the deployment of
technologically advanced solutions. They turn data into actionable insights via cognitive
computing and comprehensive facilities management capability to better manage buildings
(Ploennigs et al., 2018). Though the concept has propagated widely, no theoretical concepts,
foundations, or framework have been proposed to formalize it. Nor does the ‘building’ adage
differentiate unique types of facilities from one another.
A brief review of the concepts related to cognitive FM helped develop the Venn diagram shown
in Figure 1 illustrating the relationship between these concepts. Smart home is a subset of smart
building and intelligent building, which has large overlap between each other (Ghayvat et al.,
2015). Cognitive building shares a similar scope with smart and intelligent building, but with
more consideration for people and more intelligence in the loop (Pasini et al., 2016). Ambient
intelligence, focusing on the ambient environment, somewhat overlaps with the four other
concepts, but does not solely operate at the home and building level (Hui et al., 2017). FM
covers smart home, smart building, intelligent building, cognitive building and a large part of
ambient intelligence (Fairchild, 2019).

4

Figure 1. The relationship between advanced technologies related to cognitive FM
All these initiatives are stimulated by and rooted in the advancement of IoT, ICT, and the
growing power of computing. Adopting advanced technologies at the home/building/facility
level of intelligence succinctly falls into three purpose categories:
(1) Using data sensing technologies such as sensors, sensor networks, laser scanning and
photogrammetry to capture the status data of buildings or facilities (Wong et al., 2018).
They imitate the five senses of human beings, i.e., touch, sight, hearing, taste and smell.
Typically non-destructive technologies, they collect environmental parameters like
temperature, humidity, luminance, and air quality (Wu et al., 2016), estimate occupancy
(Candanedo and Feldheim, 2016), detect events (Sixsmith and Johnson, 2004), and so on.
Laser scanning and photogrammetry are widely used to access a built environment’s
geometric information. Furthermore, the information they collect help fabricate as-built
BIM reconstructions for existing buildings and facilities (Chen et al., 2018).
(2) Using advanced ICTs such as WiFi, Bluetooth, Zigbee, and NFC (Near Field
Communication) to report the status of home appliances and control them remotely. Via
Bluetooth or WiFi, smart devices like mobile phones, watches, or computers can connect
with televisions, sound devices, air conditioners, lights, refrigerators, and washing machines
to check their condition (e.g., on, off, operating at certain mode) in real time and change
that condition (e.g., turn on/off, set to a new mode) with a tap or click (Robles and Kim,
2010). With these technologies, users can monitor the utilization status, energy
consumption, or environmental conditions of buildings or facilities nearly instantaneously.
Moreover, humans and the machines can interact with each other naturally through text,
voice, gesture, motion, even emotion (Jaimes and Sebe, 2007).
(3) A clear trend is to develop various smart applications and integrate them into various BIM
systems/platforms (Lu et al., 2018). The Architecture, Engineering, Construction and
5

Facilities Management (AEC/FM) domains rely heavily on BIM technology (Liao and Teo,
2019). It involves digitally constructing an accurate virtual model of a building via BIM
technology (Azhar, 2011). The model can help improve the performance and productivity
of design, construction, operation, and maintenance (Love et al., 2013). For new buildings
or facilities, their BIM models can be created via BIM software in the design stage. While
for existing buildings or facilities, their BIM models are generated from measurement data
(e.g., 2D images or 3D point clouds) captured by cameras and laser scanners, respectively
(Volk et al., 2014). BIM-enabled platforms are increasingly integrating smart applications
into their features.
Although the concepts of smart home, ambient intelligence, smart building, intelligent building,
and cognitive building share some similarities, they also have several obvious differences. First,
the scopes of FM and smart/intelligent home/building, cognitive building are different. FM
focuses on not only buildings but also other types of facilities such as highways, utility grids,
and municipal pipelines. Second, smart/intelligent home/building, cognitive building, and
ambient intelligence are more from the perspective and ownership of residents, while FM
consigns the decision-making and control with the facility manger on behalf of clients,
occupants and tenants. Third, smart/intelligent building/home, allows the automation of
building systems, ICTs, and security systems separately (Chen et al., 2014) through emerging
pervasive technologies, while FM emphasizes the integrated management of entire facilities.
Nevertheless, all these concepts place a solid cornerstone for cognitive FM for this paper to
explore.
2.2 Cognitive computing and related technologies
Cognitive science is the foundation of cognition-related studies including cognitive FM.
Encompassing psychology, epistemology, linguistics, computer science, economics, and the
social sciences more generally, cognitive science is the interdisciplinary study of brain
cognition and its information processing mechanisms (Simon, 1980). Human cognition has
three accepted main components including perception, action, and learning (Fuster, 2003;
Haykin, 2012). A cognitive system can reason, use represented knowledge, learn from its
experience, accumulate knowledge, explain itself, accept directions, be aware of its own
behavior and capability, as well as respond in a way that loosely mimics the process of cognition
in the human mind (Sheth, 2016). Cognitive systems have been tested in construction
engineering and management fields such as construction safety management (Saurin et al., 2008)
and intelligent business information management (Ogiela and Ogiela, 2014).
Merging human cognizance with computing machines breeds cognitive computing (CC), which
basically enables cognitive systems the fuzzy computing ability. An interdisciplinary product,
it combines cognitive science and computer science to mimic the human thinking process in a
computerized environment. Cognitive computing manages the low power, small volume, mind6

like function, and real-time performance of the human brain (Preissl et al., 2012). It is trained,
rather than programmed, to forecast future events with purpose and estimated strength
(Marchevsky et al., 2017). It has the capability to extract concepts, emotions, objects, keywords,
and relationships from massive unstructured data, and use artificial intelligence and enhanced
machine learning algorithms to process analytics, make predictions and hypothesize in a
scalable way (Dessì et al., 2019). With cognitive computing, a cognitive system can quickly
learn and improve as it discovers knowledge and acquires profundity in complex environments.
Cognitive computing is permeating and evolving into such scopes as machine learning (ML),
deep learning (DL), natural language processing (NLP), computer vision (CV), and artificial
intelligence (AI) (Capuano and Toti, 2019). Machine learning concerns technology that can
learn from examples and teach itself how to solve problems without being programmed
manually (Stanescu et al., 2018). Deep learning, the forefront of ML, constitutes simple but
non-linear modules that each transform the representation of raw inputs into a representation at
a higher, slightly more abstract level (LeCun et al., 2015). Natural language processing converts
human written or spoken language into a formal representation easy for computers to
manipulate (Hirschberg and Manning, 2015). Computer vision is the transformation of data
from a still or video camera into either a decision or a new representation (Bradski and Kaehler,
2008). AI is defined as the simulation of human intelligence onto a machine, so as to make the
machine proficient at identifying and applying the right piece of “Knowledge” at a given step
to solve a problem (Konar, 1999). The Venn diagram shown in Figure 2 displays their
relationships according to the ascertains: DL is a subnet of ML which is a subnet of AI (LeCun
et al., 2015); NLP and CV are also subnets of AI while using some ML and DL techniques
(Jordan & Mitchell, 2015; Young et al., 2018); AI falls into a significant part of CC (Chen et
al., 2016). Embodying the major natural intelligence behaviors of the brain such as thinking,
inference, learning, and perceptions, cognitive computing is an intelligent technology past the
imperative and automatic (Wang et al., 2010).

7

Figure 2. The relationship between advanced technologies related to cognitive computing
A critical review of various concepts discovered that current developments in smart FM have
stagnated. Passive intelligence, which can only deal with given problems based on pre-defined
rules, repeatable patterns, and pre-programmed algorithms, remains the norm. FM nowadays
has to face complex and dynamic situations affected by energy, information, the external and
internal environment, and occupants’ changing requirements. An elevation in FM is highly
advantageous and desired from its current state of passive intelligence to one of cognitive
capability. The advances in cognitive computing provide a strategic opportunity to pursue such
cognitive FM. However, little research makes up this emerging field, demanding a clear
definition of cognitive FM and a theory or theoretical framework behind the new concept.
Espoused by the theories of cognition and FM, this paper aims to fill the gaps by putting forward
the concept of cognitive FM, presenting a potential system architecture, exemplifying an
application scenario, and exploring its prospects and challenges.
3. Cognitive facility management
3.1 A working definition
Based on current FM predicaments, emerging technologies development, and the theories
borrowed from cognitive system, the idea of ‘cognitive FM’ is sprouted. It aims to advance the
stagnant FM status and add advanced active intelligence that mimics the mechanism of human
brain information processing to FM. Perception, action, and learning are the primary elements
in a cognitive system to allow for active intelligence, as illustrated in Figure 3. There are two
parts in the system architecture, the core system and the external environment that the system
was contextualized in. Sensors and actuators bridge the system and the environment. Sensors
take in perceptions of environment and communicate the signals to performance element which
is responsible for selecting actions based on perceptions and evaluation component. The
8

evaluation component takes performance standard as input and gives out feedbacks to the
learning element, which forms the most important distinction between a cognitive agent and
other general agents. It is responsible for making improvements using feedback from the
evaluation component on how the system is doing to solve application problems and determines
how the performance element should be modified to do better in the future (Russell and Norvig,
2016). To sum up, perception and action are the two primary components which bridges the
environment and the system itself. Learning element is the core component which connects
application, performance, and evaluation.

Figure 3. The architecture of a learning agent in cognitive system (Adapted from Russel and
Norvig, 2016)
Based on the theories of cognitive system and facility management, ‘cognitive FM’ is defined
as:
The active intelligent management of a facility, which have the ability of perception
through cognitive systems, learning in the manner of human cognition with the
power of cognitive computing, and adaptive, adaptive, and efficient action via
automated actuators, to improve the quality of people’s life and productivity of
core business.
There are many elements of cognitive FM, but perception, learning and action are argued as
the three key definitional characteristics of the new concept.

9

3.2 Key characteristics of cognitive FM
3.2.1 Perception
Perception is an active process involving the recognition and interpretation of stimuli which
register senses made to the environment (Rookes and Willson, 2005). It requires a form of
expectation, of knowing what is about to confront us and preparing for it (Ratey, 2001).
Perception provides a system with information about the world it inhabits by interpreting
sensors (Russell and Norvig, 2016). Perception can be empowered by Cognitive IoT, which
supports cooperation and interaction between IoT and human beings. Wired or wireless sensors,
cameras, Auto-IDs, and GPS, smart devices and their connected utilities are the ‘things’ that
sense the status and environments of facilities. Wireless sensor networks (WSNs) in buildings,
utilities, and transportation systems can enable real-time collection of sensory data (Huang and
Mao, 2017). Videos or images captured by cameras are extensively used to gather movement
and behavioral information for health, safety and security. They are also pervasive sources of
spatial and geometric data that can help reconstruct digital models of in-use facilities. Auto-IDs
such as 2D barcode and RFID make real-time localization, tracking and navigation available at
facility managers’ fingertips (Xue et al., 2018). Capable of collecting different types of data,
like the ambient environment and user behavior patterns, smart devices have proved vital
sources of information.
3.2.2 Learning
Learning is the accumulation of information in memories, the acquisition of access routes for
its retrieval, and the discovery of knowledge based on it (Russell and Norvig, 2016). In a
cognitive system, learning helps recognize objects, categorizations, relationships, events,
procedures for executing actions, new actions and action sequences to accomplish new tasks
(Franklin et al., 2014). It provides an enormous source for variation, adaptation, and evolution
in system action. For example, learning from historic data of indoor temperatures, air
conditioning controls, and weather helps the system realize occupant preferences and anticipate
the outside forces that disrupt equilibrium. Disruptions can be natural or human, for example
an individual changing a room’s temperature controls one time to suit a unique and short-term
situation. A passive system would stay in that new condition indefinitely. Learning is achieved
through four approaches, i.e., observation, experience, feedback, and reinforcement (e.g,
through watching other’s performance and imitating, concluding experiences and recognizing
patterns, reflecting feedbacks and discovering knowledge, and reinforcing by punishment or
rewards, respectively) (Illeris, 2004). In real cases, learning is a multimodal process which
integrates various approaches and mechanisms. Multimodal integration learning combines
multiple sources of sensory information such as vision, sound and movement to help a cognitive
system achieve enhanced perceptual clarity and reduced ambiguity regarding its environment
(Noda et al., 2014).

10

3.2.3 Action
Action in human cognition instantiates in the brain as motor planning and representation
(Immordino-Yang, 2008). Action in cognitive FM may include decision-making, sending alerts,
recommending action, analyzing or visualizing statistics, reconstructing models, service
provisioning, device actuating, and a combination thereof. Simply being in action can facilitate
perception of affordances, i.e., what the environment affords the individual, because actions
provide information about the dynamic capability of the facility (Franchak et al., 2010). In
cognitive FM, action executes in three ways, i.e., passively, semi-actively and actively (Casciati
et al., 2012). Passive action aims to assist people to make decisions with an optimum action
plan based on the real-time situation it perceives. A typical example would be issuing an alert
that demands a response from facility managers or workers. Semi-active action is an action
cooperated by humans and machined via human-machine interaction. Active action can actually
execute the optimum action plan via the facility itself without requiring humans to be in the
loop. For example, controlling the lights of a room based on occupancy, natural light condition
and the type of activity in the room is active action. Realization of the action, especially semiactive and active action, is dependent on cognitive computing and automatically controlled
through actuating devices. Automatically controlled actuating devices comprise motorized
built-in controllers and robotic actuators, and as this technology matures, they become more
affordable.
3.3 Interrelationships of the three key elements of cognitive FM
In cognitive FM, three key elements have direct and/or indirect influence on one another as
regards a constantly changing environment and requirements from users. Figure 4 shows the
details and relationships between these three key elements. Perception and action intricately
connect. Perception provides the requisite information for selecting and guiding actions
adaptively (Franchak et al., 2010). Action in turn has concurrent effects on perception (Hecht
et al., 2001). They form a perception-action cycle through connections between sensory and
motor structures (Fuster, 2003). The perception-action cycle governs the cognitive interactions
of a system, be they animate or artificial. This interactive cycle is the basic principle that
characterizes the dynamic adaptation of the organism to its environment (Fuster, 2003).

11

Figure 4. The interrelationships between the key elements of cognitive FM
What hides behind the perception-action cycle to make a system cognitive is learning. Learning
helps the system improve its performance on future actions after making perceptions about the
internal and external environment (Russell and Norvig, 2016), as well as adapts perception
modes to the found patterns of actions, reducing sensing, computing, and memory consumption
(Noda et al., 2014).
Learning from perceptions empowers reasoning and knowledge discovery about changes in the
internal and external environment, and in turn shapes the way a system perceives, forming a
perception-learning cycle (Fiser et al., 2010). This perception-learning cycle facilitates a
cognitive system’s predictive perception, i.e., ability to predict what an incoming signal might
mean, respond, or adjust the perception modes. Predictive perception is what differentiates a
cognitive system from a traditional artificial intelligent system, which perceives based on preset rules or reinforced learning (Otsu et al., 2017).
Learning from a collection of actions and perception-action cycles enables the system to predict
actions and select the most appropriate one for a new perception. Learning can also facilitate
action selection and execution of the system in order to adapt to uncertainties and make
decisions that maximize the benefits while minimizing the risks (Arashpour et al., 2017). The
action-learning cycle enables a cognitive system to alter its strategies and behaviors according
to an ever-changing environment and requirements (Lee et al., 2018). This cycle embodies the
proactive intelligence and automation of a cognitive system by learning from its own
12

experiences. Robotics can realize and perfect proactive intelligence and automation being able
to both read conditions, take actions and learn from experiences (Lee et al., 2018).
In FM, perception, learning and action form a stable triangle that helps a facility to accurately,
effectively, and proactively adapt to dynamic environments and various user needs. This
triangle empowers the proactive intelligence and automation of a facility to better serve users
in a manner more acceptable to human beings.
4. System architecture of cognitive FM
Figure 5 illustrates a system architecture for cognitive FM according to the definition of
cognitive FM and the interrelationships of its three key elements. Its eight-layer structure
consists of an (1) environment layer, (2) perception layer, (3) data layer, (4) communication
layer, (5) computation layer, (6) application layer, (7) action layer, and (8) evaluation layer.
Each layer is elaborated as follows:

Figure 5. The system architecture of cognitive FM

13

(1) The environment layer entails the internal and external environments of a facility. The
external environment, including natural and social components, forms the context which
the facility is positioned in and exposed to. The facility as a whole is the object of cognitive
FM, therefore its internal environment is the utilities and users within the facility. It is the
layer where all commercial, residential, and social activities take place, and also the layer
the system needs to perceive.
(2) The perception layer consists of various sensing devices that operate in a connected fashion
to capture the features of the environment layer. Perceptors (e.g., sensors, cameras, LiDAR,
Auto-IDs, and GPS) sense the environment by processing the incoming stimuli and feeding
observations to the upper layer (Jung et al., 2019). They can detect the status or
characteristics of different objects. The information collected by the perceptors is
interpreted by the cognitive computing behind.
(3) The data layer denotes various data types generated after perception, including
environmental indicators, such as temperature, humidity, luminosity, air pressure, or
particulate matter density. It may also comprise pictures, video or 3D point cloud data of a
facility or area, the identities of building components, furniture and users, location,
operation status, user behavior, and other characteristics.
(4) The communication layer helps upload the sensing data to the database, which acts as base
memory. Different types of communication protocols may be used. Local and wireless
communication protocols offer a lot of current applications as well as promise for future
growth (Tolman et al., 2009). Emerging personal networks (e.g., WPAN) as well as new
interoperable networks (e.g., WiMAX) will likely popularize as social communication
happens more and more virtually. BIM is increasingly becoming the hub of this
communication layer (Chen and Lu, 2019).
(5) The computation layer effectively analyzes the sensed data from the lower layers to generate
decision support information through cognitive computing. This essentially embodies
behaviors like reasoning, interpreting, inferencing, and learning in uncertain, fuzzy and
complex environments with the help of ML, CV, NLP, SA and NN discussed in Section 2.2.
(6) The application layer uses the information and knowledge abstracted from the lower layers
to enable multiple or even massive interactive agents (e.g., facilities or users) to reason,
plan, and select the most suitable action to support corresponding FM applications.
(7) The action layer is where actuators proceed to facilitate the applications, control the objects
and change the perceptron via human-machine interface, robotics, or actuators. The
actuators directly control the facilities in the object layer, issuing alerts or guidance to
facility users or managers. Robots with caring, operation or maintenance abilities can also
execute actions.
(8) The evaluation layer shares import interfaces with social networks, in which on-demand
service provision can be shared by facility managers and users to social networks. Novel
performance metrics are designed to evaluate the services provided. Moreover, the
14

evaluation results feedback to the application layer to reset the application variables and
objectives, or to the computation layer to adjust the computing algorithms.
The eight-layer system architecture of cognitive FM helps recognize and explain its key
elements, i.e., perception, learning, and action. Cognitive FM energizes multi-sourcing and
cognitive computing power to support various applications. Through cognitive FM, data
sensing informs users, facility managers/owners, and facility manufacturers/designers, enabling
more profound information analysis, decision-making, and application execution. Meanwhile,
many tedious choices and actions can been performed automatically. The system architecture
can be put into the complex real-world FM scenarios.
5. Example scenario of applying cognitive FM
This section presents an applied scenario of cognitive FM called “Event Manager”. Such “Event
Manager” is especially important for public and commercial buildings, where events bring
considerable people flows within a short time, leading to periodic but rapidly growing needs
from certain facilities (e.g., elevators). Efficient scheduling and utilization of these facilities
aligned with events is crucial to sustainability, productivity and safe facility operations.
Through cognitive FM, the “Event Manager” can perceive data (perception ability) including
facilities operations processes (e.g., elevator traveling to a certain floor), the number of people
and their status (e.g., in a hurry, anxious, or claustrophobic), even their intentions (e.g., turning
the thermostat down) with cognitive IoT and other technologies at the perception layer. Such
data and information will be automatically uploaded into the database via the Internet. Learning
from experience, best practices, and current information empowered by cognitive computing
helps attain an optimal schedule plan, which represents the learning ability of cognitive FM
system. The facilities at a specific place will automatically schedule and take actions to satisfy
the requirements of users at that time and location (active action). Moreover, the digital model
will display the real-time statistics on a dashboard or APP for users and facility managers to
better visualize the data, these applications can intelligently guide users to better plan and
conduct work, as well as for their evaluation.
To further illustrate how cognitive FM works, an “Event Manager” of a campus building with
three elevators is presented. In this scenario, the general system architecture in Figure 5 is
specified to the one as displayed in Figure 6 (a). The elevators and users belong to the
environment layer (Figure 6 (b)). A UWB (ultra-wideband) local positioning system serves as
the perception layer (Figure 6 (c)). The system tracks the real-time location of the users by
UWB with a precision to 20cm. By installing UWB anchors in the building, the locations of the
UWB tags can be tracked (Ruiz and Granja, 2017). With the Users’ IDs attached to the UWB
tags allocated to them when they register the event, the location of the users can be tracked.
Once a user walked into the area of the building, the UWB tracking system can detect his/her
location. The most important inputs at the data layer include the event schedule of the building
15

(e.g., at what time a significant number of people will attend or exit an event on a given floor),
users’ location, elevators’ features, and floor function. The four major types of entities in the
system database and their relationships are shown in Figure 6 (d). All the information items are
connected in the database for the perception of the event and further computing. The user
location will be communicated via WLAN while elevators’ status can be communicated to the
database through LAN. The “Event Manager” in the elevator system works following the
rationality shown in Figure 6 (e). Finally, with the data in the database and the memory and
fuzzy computing ability, the “Event Manager” can better serve the users with cognition by
considering the lead time of the elevators to travel from floors to floors, the distribution of users’
location and arrival time from short-term memory, and the user behavior pattern from longterm memory. The “Event Manager” will take actions by calling the elevators in advance for
the users and automatically taking the users to their destinations. It will optimize the
performance of the facility by considering three objectives, i.e., shortening the waiting time,
improving the operation efficiency of the elevators and building, and saving energy through
learning. A visualization interface will also show the location of the users and elevators as
illustrated at Figure 6 (f). The actuation layer is the controlling system of the elevators, see
Figure 6 (g). The system can generate daily self-report for evaluation; users may also provide
their experience comments as feedbacks for future improvement, as illustrated by Figure 6 (h).
This cognitive change, albeit minor, can provide privileged services for the users and upgrade
the performance of the facilities.

16

Figure 6 Elevator management system of the “Event Manager”
It should be pointed out that the “Event Manager” as illustrated above is just one of the very
primitive scenarios of cognitive FM. Future studies are encouraged to develop other scenarios.
The system architecture proposed in this study provides a general guideline by following which
other more sophisticated scenarios can be developed. New developments in different layers,
particularly perception, data, communication, computing, and actuating technologies will
provide unlimited opportunities to realize cognitive FM. Due efforts should be paid to validate
such system architecture. The key to this validation work is to develop a real-life cognitive FM
17

system and define their key performance indicators, including both tangible (e.g., energy saving,
time saving, or system reliability) and intangible ones (e.g., users satisfactory, ambient
environment, or productivity enhancement).
6. Discussion
This paper put forward a bold proposition – cognitive FM – in the inert body of FM literature.
It provided the working definition of cognitive FM, and articulated its three key definitional
characteristics, i.e., perception, learning, and action. A system architecture of eight layers
provides a cognitive FM framework for the structure and operation of human-like proactive
intelligence in FM. The highlighted interrelationships and the real-world case, i.e., the ‘Event
Manager’, were presented in an attempt to illustrate how cognitive FM could push boundary of
both FM theory and practice. Beyond the ‘Event Manager’ example portrayed, readers of this
paper should note the wealth of other compelling scenarios such as customized service
provision and predictive maintenance which fall under the realm of cognitive FM that can be
explored in the future.
The advancement of new technologies provides boundless challenges and opportunities to
realizing cognitive FM. The technologies on the perception layer, e.g., sensor networks, camera,
Auto-ID, smart devices, and cognitive IoT are developing in an unprecedented pace, which
allows quickly capturing numerous and various ‘big data’ (known for its 3V, namely, volume,
velocity, and variety) related to facilities and their occupiers and managers with an affordable
cost. The communication technologies such as WAN, LAN, and particularly recent fifth
generation (5G) cellular network, allow quick communication of the facility big data between
the data layer and the computing layer. Quick development of computing technologies such as
GPU and clouding computing enables the harness of the power of big data to realize perception
and learning capability, while the advancement of human-computer interface, robotics, and
other actuating technologies will allow the realization of autonomous actions of cognitive FM.
Not only the individual technologies developed (e.g. emerging pervasive technologies (Niu et
al., 2019)), but also the new system integration philosophies (e.g., edge computing (Shi et al.,
2016), smart construction objects (Niu et al., 2016)) make the bold proposition of cognitive FM
feasible than ever. Information and industrial manufacturing conglomerates like Schneider,
Siemens, and IBM continuously offer more and more cognitive building services (e.g., Watson
IoT for Buildings and IBM TRIRIGA), which integrate the physical and digital worlds.
This research deems FM remained dormant and without a literature base to elevate the
discipline to the science it should be, especially given the new technology and capacities
available to it. Challenges and opportunities also exist in the development of cognitive science,
which is the interdisciplinary study of cognition in humans, animals, and machines (Norman,
1980). Despite the great achievements in cognitive science, very limited actually we know
about our cognitive capability and process. Cognitive capability such as perception and learning
18

exclusively possessed either innate or learned by human beings are now facing the challenge to
be anthropomorphized to facilities to realize cognitive FM. The resurgence of Artificial
Intelligence (AI), and the technological advancement as exhibited above, all underpin the best
time to pursue cognitive FM. Research and academia should provide more solid science and
technological foundation for promoting cognitive FM as a discipline of science.
While this paper so far has mainly explored the benefits of cognitive FM, its potential
drawbacks should not be neglected. For example, the benefits (e.g., convenience) that cognitive
FM brings to occupants should never be achieved at the cost of their security and privacy. If we
take the pioneering modernist architect Le Corbusier’s view that ‘a house is a machine for living
in’ in 1923 (Kornberger and Clegg, 2004), now the machine/facility is of human’s cognitive
capability of perception, learning, and action, all the ethics issues related to human dignity,
transparency, accountability, or unintended consequences would surface from this new domain
of FM. Cognitive FM can join the ever-heated debate on AI and robotics ethics to gain a better
understanding of its domain specific ethics issues.
Ultimately, development of cognitive FM is determined by the changing dyadic relationship
between us and the facilities we built. Man has never stopped exploring such relationship. Le
Corbusier’s view in the 1920s represented the industrialization era where humans have started
to massively engage machine in production, life, and other activities. We built buildings as a
machine to serve us. The sociotechnical systems coined at Tavistock Institute (1966), referring
to the interaction between society's complex infrastructures and human behavior, provides a
perspective of the dyadic relationship between facilities and humans. The facilities, be they
machines or intelligent/smart buildings, are alien to us. They have never been treated as a part
of mankind. Humans never stop trying to make them as cognitive and intelligent as possible to
free us from routine work and regular decisions. We imbue facilities with cognitive capability.
We treat a way humans taking care of each other is the way humans need from facility services.
It reaches to the level that Churchill has seen it, i.e., facilities can shape us. The issue is whether
humans are ready for it, will allow it happen, or can stop it happen.
7. Conclusion
With all due respect, the researchers of this paper are unsatisfied with the stagnant FM
discourses that have been demarcated by smart/intelligent building for already three decades or
so. To this end, they put forward the concept of ‘cognitive FM’. Inspired by human cognition
systems, the definition of cognitive FM was given as actively intelligent management of
facilities within the built environment, where a facility perceives through a cognitive IoT, learns
in the manner of human cognition with the power of cognitive computing, and acts actively,
adaptively, and efficiently via automated actuators to improve the quality of life or productivity
of business. Unlike passive, pre-setting, and nonreciprocal intelligence of smart/intelligent
19

building, cognitive FM proposes an active, adaptive and naturally interactive way to give
intelligence to facilities through three key characteristics, i.e., perception, learning, and action.
An eight-layered system architecture comprising of environment, perception, data,
communication, computation, application, actuation and evaluation was developed for better
realization of cognitive FM. This architecture puts most of the advanced technologies of
information, communication, computation, and action together in an organic and dynamic
fashion to realize the cognitive capability of perception, learning, and action. To better illustrate
what cognitive FM is and how the architecture can guide its implementation, an illustrative
scenario called ‘Event Manager’ was elaborated. The case study showed that cognitive FM can
enable human-like cognition, enhancing the value of facilities, and ultimately, better serving
our health, economic activities, and social behaviour. Other than the ‘Event Manager’, there are
other compelling scenarios of cognitive FM awaiting researchers and practitioners to explore.
Cognitive FM tries to imbue existing facilities with human-like cognition in order to deliver
their designated functions (e.g., keep us warm/cold; save energy; make us comfortable; keeping
us secure; keeping us healthy; improving quality of life; and enhancing productivity of our
business). The facility can know what its users want and need, understand their needs and
preferences, learn and give feedbacks, catch the details the users might miss, distinguish
between different users, interact with and serve for them like an experienced but agile manager
with the ability of perception, learning, and action. The definition of cognitive FM, its system
architecture, and the pervasive supporting technologies as presented in this paper would open
up a new avenue to empower the traditional FM in the era of AI resurgence. Our understanding
towards facilities has evolved from the 1920s when Le Corbusier perceived them as machines
for living in, to intelligent/smart building 30 years ago, and to this cognitive system akin to
human now. Technological advancement provides unlimited possibilities to realizing cognitive
FM. But whether a cognitive facility can be a human-like one, or whether we human want it to
be indistinguishable from us, are the philosophical questions to be answered on this journey to
develop cognitive FM.
8. References
Azhar, S. Building information modeling (BIM): Trends, benefits, risks, and challenges for the
AEC industry. Leadership and Management in Engineering, 11(3) (2011)241-252,
10.1061/(ASCE)LM.1943-5630.0000127.
Bilgen, S. Structure and environmental impact of global energy consumption. Renewable and
Sustainable Energy Reviews, 38 (2014)890-902, 10.1016/j.rser.2014.07.004.
Bradski, G., & Kaehler, A. Learning OpenCV: Computer vision with the OpenCV library.
O'Reilly Media, Inc, 2008, ISBN 97806596616130.
Buckman, A. H., Mayfield, M., & Beck, BM. S. What is a smart building?. Smart and
Sustainable Built Environment, 3(2) (2014)92-109, 10.1108/SASBE-01-2014-0003

20

Candanedo, L. M., & Feldheim, V. Accurate occupancy detection of an office room from light,
temperature, humidity and CO2 measurements using statistical learning models. Energy
and Buildings, 112(2016) 28-39, 10.1016/j.enbuild.2015.11.071.
Capuano, N., & Toti, D. Experimentation of a smart learning system for law based on
knowledge discovery and cognitive computing. Computers in Human Behavior, 92
(2019)459-467, 10.1016/j.chb.2018.03.034.
Casciati, F., Rodellar, J., & Yildirim, U. Active and semi-active control of structures–theory
and applications: A review of recent advances. Journal of Intelligent Material Systems and
Structures, 23(11) (2012)1181-1195, 10.1177/1045389X12445029.
Chen, C. Y., Fu, J. H., Sung, T., Wang, P. F., Jou, E., & Feng, M. W. Complex event processing
for the internet of things and its applications. In 2014 IEEE International Conference
on Automation
Science
and
Engineering,
2014,
pp
1144-1149,
10.1109/CoASE.2014.6899470.
Chen, K., & Lu, W. Bridging BIM and building (BBB) for information management in
construction: The underlying mechanism and implementation. Engineering, Construction
and Architectural Management, (2019), 10.1108/ECAM-05-2018-0206.
Chen, K., Lu, W., Xue, F., Tang, P., & Li, L. H. Automatic building information model
reconstruction in high-density urban areas: Augmenting multi-source data with
architectural
knowledge. Automation
in
Construction, 93(2018)
22-34,
10.1016/j.autcon.2018.05.009.
Chen, Y., Argentinis, J. E., & Weber, G. IBM Watson: how cognitive computing can be applied
to big data challenges in life sciences research. Clinical therapeutics, 38(4) (2016) 688701, 10.1016/j.clinthera.2015.12.001.
Coccoli, M., Maresca, P., & Stanganelli, L. The role of big data and cognitive computing in the
learning process. Journal of Visual Languages & Computing, 38 (2016) 97-103,
10.1016/j.jvlc.2016.03.002.
Cook, D. J., Augusto, J. C., & Jakkula, V. R. Ambient intelligence: Technologies, applications,
and opportunities. Pervasive and Mobile Computing, 5(4) (2009) 277-298,
10.1016/j.pmcj.2009.04.001.
Dessì, D., Fenu, G., Marras, M., & Recupero, D. R. Bridging learning analytics and Cognitive
Computing for Big Data classification in micro-learning video collections. Computers in
Human Behavior, 92 (2019)468-477, 10.1016/j.chb.2018.03.004.
Ding, G., Wu, Q., Zhang, L., Lin, Y., Tsiftsis, T. A., & Yao, Y. D. An amateur drone
surveillance system based on the cognitive Internet of Things. IEEE Communications
Magazine, 56(1) (2018) 29-35, 10.1109/MCOM.2017.1700452.
Drion, B., Melissen, F., & Wood, R. Facilities management: lost, or
regained?. Facilities, 30(5/6) (2012) 254-261, 10.1108/02632771211208512.
Fairchild, A. Twenty-First-Century Smart Facilities Management: Ambient Networking.
In Guide to Ambient Intelligence in the IoT Environment: Principles, Technologies and
Applications. Springer, Cham, 2019, ISBN 9783030041724.
Fiser, J., Berkes, P., Orbán, G., & Lengyel, M. Statistically optimal perception and learning:
from behavior to neural representations. Trends in cognitive sciences, 14(3) (2010) 119130, 10.1016/j.tics.2010.01.003.
Franchak, J. M., van der Zalm, D. J., & Adolph, K. E. Learning by doing: Action performance
facilitates
affordance
perception. Vision
Research, 50(24)
(2010)2758-2765,
10.1016/j.visres.2010.09.019.
Franklin, S., Madl, T., D'Mello, S., & Snaider, J. LIDA: A systems-level architecture for
cognition, emotion, and learning. IEEE Transactions on Autonomous Mental
Development, 6(1) (2014) 19-41, 10.1109/TAMD.2013.2277589.
21

Fuster, J. M. Cortex and mind: Unifying cognition. Oxford university press, New York, 2003,
ISBN .
Ghaffarianhoseini, A., Berardi, U., AlWaer, H., Chang, S., Halawa, E., Ghaffarianhoseini, A.,
& Clements-Croome, D. What is an intelligent building? Analysis of recent interpretations
from an international perspective. Architectural Science Review, 59(5) (2016)338-357,
10.1080/00038628.2015.1079164.
Ghayvat, H., Mukhopadhyay, S., Gui, X., & Suryadevara, N. WSN-and IOT-based smart homes
and their extension to smart buildings. Sensors, 15(5) (2015) 10350-10379,
10.3390/s150510350.
Gibson, J. J. The ecological approach to visual perception: classic edition. Psychology Press,
New York, 2014, ISBN 9781315740218.Harper, R. (Ed.). Inside the smart home. Springer
Science & Business Media, London, 2006, ISBN 1852336889.
Haykin, S. Cognitive dynamic systems: perception-action cycle, radar and radio. Cambridge
University Press, Cambridge, 2012, ISBN 9780521114363.
Hecht, H., Vogt, S., & Prinz, W. Motor learning enhances perceptual judgment: A case for
action-perception
transfer. Psychological
Research, 65(1)
(2001)3-14,
10.1007/s004260000043.
Hirschberg, J., & Manning, C. D. Advances in natural language processing. Science, 349(6245)
(2015) 261-266, 10.1126/science.aaa8685.
Huang, Q., & Mao, C. Occupancy estimation in smart building using hybrid CO2/light wireless
sensor
network. Journal of
Applied
Sciences and Arts, 1(2)
(2017)5,
https://opensiuc.lib.siu.edu/jasa/vol1/iss2/5, Accessed date: 31 May 2019.
Hui, T. K., Sherratt, R. S., & Sánchez, D. D. Major requirements for building Smart Homes in
Smart Cities based on Internet of Things technologies. Future Generation Computer
Systems, 76 (2017) 358-369, 10.1016/j.future.2016.10.026.
Illeris, K. Transformative learning in the perspective of a comprehensive learning
theory. Journal
of
Transformative
Education, 2(2)
(2004)79-89,
10.1177/1541344603262315.
Immordino‐Yang, M. H. The smoke around mirror neurons: Goals as sociocultural and
emotional organizers of perception and action in learning. Mind, Brain, and Education, 2(2)
(2008) 67-73, 10.1111/j.1751-228X.2008.00034.x.
ISO (International Organization for Standardization). Facility management - Vocabulary (ISO
41011: 2017). https://bit.ly/2MEuxRY, Accessed date: 31 May, 2019.
Jaimes, A., & Sebe, N. Multimodal human–computer interaction: A survey. Computer Vision
and Image Understanding, 108(1-2) (2007) 116-134, 10.1016/j.cviu.2006.10.019.
Jordan, M. I., & Mitchell, T. M. Machine learning: Trends, perspectives, and
prospects. Science, 349(6245) (2015) 255-260, 10.1126/science.aaa8415.
Jung, Y., Oh, H. & Jeong, M. M. An approach to automated detection of structural failure using
chronological image analysis in temporary structures. International Journal of
Construction Management, 19 (2) (2019) 178-185, 10.1080/15623599.2017.1411457.
Kaplan, R. S., Kaplan, R. E., Norton, D. P., Norton, D. P., and Davenport, T. H. Strategy Maps:
Converting Intangible Assets into Tangible Outcomes. Harvard Business Press, Boston,
2004, ISBN 1591391342.
Konar, A. Artificial intelligence and soft computing: behavioral and cognitive modeling of the
human brain. CRC press, Boca Raton, 1999, ISBN: 9781315219738.
Kornberger, M., & Clegg, S. R. Bringing space back in: Organizing the generative
building. Organization Studies, 25(7) (2004) 1095-1114, 10.1177/0170840604046312.
Kotseruba, I., & Tsotsos, J. K. 40 years of cognitive architectures: core cognitive abilities and
practical applications. Artificial Intelligence Review, (2018)1-78, 10.1007/s10462-0189646-y.
22

Lau, D., Liu, J., Majumdar, S., Nandy, B., St-Hilaire, M., & Yang, C. S. A cloud-based
approach for smart facilities management. In Prognostics and Health Management (PHM),
2013 IEEE Conference, 2013, pp. 1-8, 10.1109/ICPHM.2013.6621459.
LeCun, Y., Bengio, Y., & Hinton, G. Deep learning. Nature, 521(7553) (2015) 436-444,
10.1038/nature14539.
Lee, B. J., Choi, J., Lee, C. Y., Park, K. W., Choi, S., Han, C., ... & Zhang, B. T. Perceptionaction-learning system for mobile social-service robots using deep learning. In ThirtySecond AAAI Conference on Artificial Intelligence, 2018, accessed at:
https://www.aaai.org/ocs/index.php/AAAI/AAAI18/paper/view/16337/16446.
Liao, L. H. & Teo, E. A. L. Managing critical drivers for building information modelling
implementation in the Singapore construction industry: an organizational change
perspective. International Journal of Construction Management, 19 (3) (2019) 240-256,
10.1080/15623599.2017.1423165
Love, P. E., Simpson, I., Hill, A., & Standing, C. From justification to evaluation: Building
information modeling for asset owners. Automation in Construction, 35 (2013) 208-216,
10.1016/j.autcon.2013.05.008.
Lu, Q., Chen, L., Lee, S., & Zhao, X. Activity theory-based analysis of BIM implementation in
building O&M and first response. Automation in Construction, 85 (2018) 317-332,
10.1016/j.autcon.2017.10.017.
Marchevsky, A. M., Walts, A. E., & Wick, M. R. Evidence-based pathology in its second
decade: toward probabilistic cognitive computing. Human pathology, 61 (2017 1-8,
10.1016/j.humpath.2016.09.002.
Mehrdad, A., Abbasi, B., Arashpour, M., Hosseini, M. R., and Yang, R. Integrated management
of on-site, coordination and off-site uncertainty: theorizing risk analysis within a hybrid
project setting. International Journal of Project Management, 35(4) (2017) 647-655,
10.1016/j.ijproman.2017.02.016.
Nielsen, S. B., Sarasoja, A. L., & Galamba, K. R. Sustainability in facilities management: an
overview of current research. Facilities, 34(9/10) (2016) 535-563, 10.1108/F-07-20140060.
Niu, Y., Anumba, C., & Lu, W. Taxonomy and Deployment Framework for Emerging
Pervasive Technologies in Construction Projects. Journal of Construction Engineering and
Management, 145(5) (2019) 04019028, 10.1061/(ASCE)CO.1943-7862.0001653.
Niu, Y., Lu, W., Chen, K., Huang, G. G., & Anumba, C. Smart construction objects. Journal of
Computing in Civil Engineering, 30(4) (2016) 04015070, 10.1061/(ASCE)CP.19435487.0000550.
Noda, K., Arie, H., Suga, Y., & Ogata, T. Multimodal integration learning of robot behavior
using deep neural networks. Robotics and Autonomous Systems, 62(6) (2014) 721-736,
10.1016/j.robot.2014.03.003.
Norman, D. A. Twelve issues for cognitive science. Cognitive science, 4(1) (1980) 1-32,
10.1207/s15516709cog0401_1.
Ogiela, L., & Ogiela, M. R. Cognitive systems for intelligent business information management
in cognitive economy. International Journal of Information Management, 34(6) (2014)
751-760, 10.1016/j.ijinfomgt.2014.08.001.
O'Toole, J., & Lawler, E. E. The new American workplace. Palgrave Macmillan, New York,
2007, ISBN 1403969590.
Otsu, K., Agha-Mohammadi, A. A., & Paton, M. Where to look? Predictive perception with
applications to planetary exploration. IEEE Robotics and Automation Letters, 3(2) (2017)
635-642, 10.1109/LRA.2017.2777526.
23

Pärn, E. A., Edwards, D. J., & Sing, M. C. P. The building information modelling trajectory in
facilities management: A review. Automation in Construction, 75 (2017) 45-55,
10.1016/j.autcon.2016.12.003.
Pasini, D., Ventura, S. M., Rinaldi, S., Bellagente, P., Flammini, A., & Ciribini, A. L. C.
Exploiting Internet of Things and building information modeling framework for
management of cognitive buildings. In 2016 IEEE International Smart Cities Conference
(ISC2), 2016, pp. 1-6, 10.1109/ISC2.2016.7580817.
Ploennigs, J., Ba, A., & Barry, M. Materializing the promises of cognitive iot: How cognitive
buildings are shaping the way. IEEE Internet of Things Journal, 5(4) (2018) 2367-2374,
10.1109/JIOT.2017.2755376.
Preissl, R., Wong, T. M., Datta, P., Flickner, M., Singh, R., Esser, S. K., Risk, W. P., Simon,
H. D., & Modha, D. S. Compass: A scalable simulator for an architecture for cognitive
computing. In Proceedings of the International Conference on High Performance
Computing, Networking, Storage and Analysis, 2012, no. 54, ISBN 978-1-4673-0804-5.
Pulvermüller, F., Moseley, R. L., Egorova, N., Shebani, Z., & Boulenger, V. Motor cognition–
motor
semantics:
action
perception
theory
of
cognition
and
communication. Neuropsychologia, 55
(2014)71-84,
10.1016/j.neuropsychologia.2013.12.002.
Ratey, J. J. A user's guide to the brain: Perception, attention, and the four theatres of the brain.
Vintage Books, New York, 2001, ISBN 0679453091.
Robles, R. J., & Kim, T. H. Applications, Systems and Methods in Smart Home Technology:
A Review. International Journal of Advanced Science and Technology, 15 (2010) 37-48, .
Rondeau, E. P., Brown, R. K., & Lapides, P. D. Facility management. John Wiley & Sons,
Hoboken, 2012, ISBN 0471210617.
Rookes, P., & Willson, J. Perception: Theory, Development and Organisation. Routledge,
London, 2005, ISBN 9780203977408.
Ruiz, A. R. J., & Granja, F. S. Comparing ubisense, bespoon, and decawave uwb location
systems: Indoor performance analysis. IEEE Transactions on instrumentation and
Measurement, 66(8) (2017) 2106-2117, 10.1109/TIM.2017.2681398.
Russell, S. J., & Norvig, P. Artificial intelligence: a modern approach. Pearson Education
Limited, Harlow, 2016, ISBN 1292153962.
Saurin, T. A., Formoso, C. T., & Cambraia, F. B. An analysis of construction safety best
practices from a cognitive systems engineering perspective. Safety science, 46(8) (2008)
1169-1183, 10.1016/j.ssci.2007.07.007.
Sheth, A. Internet of things to smart iot through semantic, cognitive, and perceptual computing.
IEEE Intelligent Systems, 31(2) (2016) 108-112, 10.1109/MIS.2016.34.
Shi, W., Cao, J., Zhang, Q., Li, Y., & Xu, L. Edge computing: Vision and challenges. IEEE
Internet of Things Journal, 3(5) (2016) 637-646, 10.1109/JIOT.2016.2579198.
Simon, H. A. Cognitive science: The newest science of the artificial. Cognitive science, 4(1)
(1980) 33-46, 10.1016/S0364-0213(81)80003-1.
Sixsmith, A., & Johnson, N. A smart sensor to detect the falls of the elderly. IEEE Pervasive
computing, (2) (2004) 42-47, 10.1109/MPRV.2004.1316817.
Somov, A., Dupont, C., & Giaffreda, R. Supporting smart-city mobility with cognitive Internet
of Things. In Future Network and Mobile Summit, 2013, pp. 1-10, ISBN 9781905824373.
Stanescu, A., Mata-Toledo, R. A., & Gupta, P. Machine learning. In AccessScience. McGrawHill Education, 2018. 10.1036/1097-8542.395250
Tavistock Institute of Human Relation. Interdependence and Uncertainty: A Study of the
Building Industry. Tavistock Publications, London, 1966, 0415264375.
Teicholz, P. (Ed.). BIM for facility managers. John Wiley & Sons, Hoboken, 2013, ISBN
9781118382813.
24

Tolman, A., Matinmikko, T., Möttönen, V., Tulla, K., & Vähä, P. The benefits and obstacles of
mobile technology in FM service procurement. Facilities, 27(11/12) (2009) 445-456,
10.1108/02632770910980736.
Volk, R., Stengel, J., & Schultmann, F. Building Information Modeling (BIM) for existing
buildings—Literature review and future needs. Automation in Construction, 38(2014) 109127, 10.1016/j.autcon.2013.10.023.
Wang, Y., Baciu, G., Yao, Y., Kinsner, W., Chan, K., Zhang, B., ... & Miao, D. Perspectives
on cognitive informatics and cognitive computing. International Journal of Cognitive
Informatics and Natural Intelligence, 4(1) (2010)1-29, 10.4018/jcini.2010010101.
Wang, Z., de Dear, R., Luo, M., Lin, B., He, Y., Ghahramani, A., & Zhu, Y. Individual
difference in thermal comfort: A literature review. Building and Environment, 1382018)
181-193, 10.1016/j.buildenv.2018.04.040.
Weber, W., & Rabaey, J. Ambient intelligence. Springer Science & Business Media, Heidelberg,
2005 , ISBN 3540238670.
Wong, J. K. W., Ge, J., & He, S. X. Digitisation in facilities management: A literature review
and future research directions. Automation in Construction, 92 (2018) 312-326,
10.1016/j.autcon.2018.04.006.
Wu, M., Tan, L., & Xiong, N. Data prediction, compression, and recovery in clustered wireless
sensor
networks
for
environmental
monitoring
applications. Information
Sciences, 329(2016) 800-818, 10.1016/j.ins.2015.10.004.
Wu, Q., Ding, G., Xu, Y., Feng, S., Du, Z., Wang, J., & Long, K. Cognitive internet of things:
a new paradigm beyond connection. IEEE Internet of Things Journal, 1(2) (2014) 129-143,
10.1109/JIOT.2014.2311513.
Xue, F., Chen, K., Lu, W., Niu, Y., & Huang, G. Q. Linking radio-frequency identification to
Building Information Modeling: Status quo, development trajectory and guidelines for
practitioners. Automation
in
Construction, 93(2018)
241-251,
10.1016/j.autcon.2018.05.023.
Young, T., Hazarika, D., Poria, S., & Cambria, E. Recent trends in deep learning based natural
language processing. IEEE Computational Intelligence Magazine 13(3) (2018) 55-75,
10.1109/MCI.2018.2840738.
Zhang, M., Zhao, H., Zheng, R., Wu, Q., & Wei, W. Cognitive internet of things: concepts and
application example. International Journal of Computer Science Issues, 9(6) (2012) 151158, Retrieved from https://search.proquest.com/docview/1315193015?accountid=14548.

25