INCLUSION OF TEMPORAL DATABASES WITH

INDUSTRY FOUNDATION CLASSES

A Basis for Adaptable Intelligent Buildings

Hubert Grzybek*, Stephen Gulliver and Zhuotao Huang

Informatics Research Centre, University of Reading, Ground floor, Building 42, Whiteknights, Reading,RG6 6WB, U.K.

Keywords: Intelligent buildings, Industry Foundation Classes, Temporal database, Dynamic Building Model.

Abstract: Current ‘intelligent’ buildings (IB) lack the crucial requirement of being adaptable. Without being dynamic

and adaptable, buildings can only be called ‘responsive’ or ‘automated’, and therefore continue to perform

sub-optimally due to their reliance on manual calibration and inability to automatically recognise patterns in

building usage. It is proposed that the development of an open, dynamic and temporal building model will

alleviate many of these problems. This will facilitate IB software to automatically and continuously re-

calibrate itself in order to achieve personalised occupant, environmental and energy consumption goals. The

aim of this paper is to examine the main standards and technologies that are currently used in the field of

intelligent buildings, and highlight deficiencies in their ability to support adaptable intelligence. In addition

we propose a novel solution and prototype, combining IFC and temporal database theory, in order to discuss

the implications and application of practical deployment.

1 INTRODUCTION

There is currently no universally accepted definition

for ‘intelligent buildings’ (IB). Various definitions

differ in focus and emphasis, however most IB

authors share the belief that an IB should use

modern technology to improve resource efficiency

and provide a comfortable environment for

occupants. The concept of ‘intelligent’ systems,

however, stipulates that such systems should be

capable of ‘learning’ and automatically calibrating

itself to react to changes in its environment, in order

to continuously ensure optimal running (Fritz,

2010). This notion contrasts, however, with current

IB software, which focuses entirely on building

automation and visualisation of building system

states by responding to a given set of conditions with

pre-programmed actions. Current IB software is

subsequently unable to dynamically respond to

changes in its environment and adapt its behaviour

to varying usage patterns over time. For the purpose

of this paper, adaptability and flexibility relate to a

building’s ability to automatically recognise and

respond to changes in its environment, including the

‘learning’ of recognised patterns in recorded data.

Within a dynamic environment, responsive systems

will therefore require continual reconfiguration if

they are not to become ineffectual or obsolete.

If dynamic adaptability is essential within the

'intelligent building' definition, we would have to

agree that the first IB is yet to be designed. Current

IB software is incapable of supporting such

intelligent buildings, due to their inability to create

and make use of historical data relating to the

changing state of the building, its lack of

information concerning objects with the space, and a

lack of a machine-readable semantic building model

to allow information to be placed in context of the

building space. Historical data is crucial to

supporting ‘learning’, as it supports data mining and

the recognition of patterns in use of resources and

space (Cheng, 2006). This lack of historical data and

its link with a semantic building model is currently a

huge obstacle preventing buildings from responding

dynamically to the changing patterns of building

usage and dynamic occupant behaviour.

Whilst current intelligent building technology is

capable of improving building performance in terms

of resource efficiency and the provision of a

comfortable environment for occupants, they do not

meet the requirements of adaptability and flexibility.

Changes in business process, organisational

Grzybek H., Gulliver S. and Huang Z.

INCLUSION OF TEMPORAL DATABASES WITH INDUSTRY FOUNDATION CLASSES - A Basis for Adaptable Intelligent Buildings.

DOI: 10.5220/0003259000240031

In Proceedings of the Twelfth International Conference on Informatics and Semiotics in Organisations (ICISO 2010), page

ISBN: 978-989-8425-26-3

structure and/or building use risks impacting

building performance, yet can only be optimised via

constant manual reconfiguration. To automatically

recognise and respond to change, buildings must be

capable of benchmarking critical performance

metrics, in relation to a set of building goals, and

have the ability to recognise when the building’s

performance is sub-optimal as a result of changes in

its environment. The authors believe that, to

understand and adapt to such changes, IB software

must maintain an accurate temporal model of the

building to provide a complete current and historical

representation of: the building’s structure; the

location and state of its contained spaces; occupants

or objects of interest; as well as the changing

relationships between these occupants and objects.

Such records support data mining and the

recognition of patterns in occupant behaviour and

space usage, which is critical to supporting building

adaptability and learning.

This paper examines current standards and

technologies in the field of IB and explains why the

current technology is insufficient. Moreover we

propose a novel solution, via the introduction of a

temporal building model prototype that we hope will

stimulate research ideas in this area. The rest of the

paper is divided into the following sections: in

section 2 we discuss current building standards and

technologies (including BMS, BIM and IFC) and

relate them to the requirements of adaptability. In

section 3 we give a summary of temporal database

theory, its applicability to IB, and a description of

the proposed building model standard - including

expected advantages and potential drawbacks.

Finally, in sections 4 and 5, we present our

prototype, which showcases the feasibility of the

proposed IB model. We conclude by summarising

the key points of the paper and expanding the

implications of this technology.

2 CURRENT BUILDING

STANDARDS AND

TECHNOLOGIES

2.1 Building Management Systems

(BMS)

BMS products support the integration of the

building’s current systems, i.e. HVAC, lighting

security, etc. and the automation of specific actions,

such as the locking all entrances at a specified time.

A BMS can be defined as a system for centralising

and optimising the monitoring, operating, and

managing of a building. Services may include

heating, cooling, ventilation, lighting, security, and

energy management.

The functional focus of BMS products is on

peripheral connectivity and actuator responses to

pre-defined sensor inputs or timers. BMS systems

support communication between the software system

and the building’s sensors and actuators through a

communication bus, and provide building managers

with the capability of setting automated behaviours

related to these peripherals, and viewing the state of

the building’s systems through a friendly user-

interface (Knibbe, 1996).

Whilst standards in communications protocols

exist, for example LonWorks and BACnet, BMS

vendors still focus on the development of rule-based

systems, and do not support self-adapting model

based systems. The installation of typical BMS

products involves manual calibration, typically

involving the setting of building-specific rules and

the maintenance of the various parts of the system.

Once running, the system is incapable of

benchmarking its own performance, recognising

patterns in the building’s usage and suggesting

improvements that will meet the occupant’s

objectives in terms of resource usage and occupant

comfort. As a result, current BMS products do not

fulfil the adaptable/flexible requirement of truly

intelligent buildings (Clements-Croome, 2004). It is

possible, however, that existing BMS system data

could be used by a temporal building model to track

building usage over time, and store this historical

data to support data mining and learning.

2.2 Building Information Modelling

(BIM)

BIM aims to shift the construction industry away

from 2D CAD (Computer Aided Design) building

blueprints to full 3D, object-based models backed by

semantic information. It also aims to provide a

paradigm shift in the way that building project

stakeholders interact, distribute responsibility, share

information and collaborate in order to reduce costs,

increase value for the customer and remove the

current legislative culture (Smith et al, 2009).

BIM tools and techniques aim to support

construction project stakeholders in collaboratively

developing a complete virtual building model, which

includes all details relating to project management,

e.g. the building’s structure, construction schedule,

tasks and deadlines, etc. The benefits include

increased time and resource efficiency, as all

INCLUSION OF TEMPORAL DATABASES WITH INDUSTRY FOUNDATION CLASSES - A Basis for Adaptable

Intelligent Buildings

stakeholders can create and share information that

would otherwise lead to repeated data collection.

BIM tools typically support the development of 3D

visualisations of buildings through the open and

object-based IFC format.

By moving away from 2D drawings, which

include basic CAD objects such as lines and arcs,

towards software objects, which represent physical

parts of a building, its contents, and how objects

relate to each other, the model is able to carry much

more semantic information for use by both people

and computers. It is this semantic information, not

the 3D visualisation, which we believe to be critical

to achieving truly adaptable systems. By allowing

computers to understand building models, BIM

software is capable of achieving advanced features,

such as automated rule-based building design and

standards checking. Parametric objects, for example,

can automatically re-build themselves according to

simple embedded rules, such as requiring a window

to be wholly within a wall. As the objects are

machine readable, spatial conflicts in a building

model can be checked automatically, resulting in

greatly reduced quantity of errors and change orders

(Eastman, 2009).

Object-based semantic models have great

potential in terms of adding dynamic properties to

BMS, and support the potential of truly intelligent

buildings, however current use of BIM is limited to

use by Facilities Management (FM) in tasks such as:

scheduling routine maintenance, finding objects

within a building, and checking maintenance-related

costs.

It is clear that an ‘as-built’ BIM model could

provide most of the data required for a dynamic

building model. However, to support the adaptable

requirement this data would first need to be

extracted into software that supports the

management of this information in a temporal

fashion, i.e. to allow for the information to change

whilst recording information relating to all past

states to support learning. The structural information

for such a model is defined in BIM IFC files, which

will be discussed in more depth in section 2.3.

2.3 Industry Foundation Classes (IFC)

To reduce data duplication and inconsistency, BIM

proponents use a single building model per project,

which is then distributed amongst multiple

specialised applications which are designed to

collaborate. BIM applications typically store this

building data using the open IFC object model. The

IFC model provides an object-oriented description

of the building and related services, enabling

interoperability between different vendors of

Architectural, Engineering and Construction /

Facilities Management software (Spearman, 2007).

The model supports both step formatted text and

XML, and therefore it can be used by software tools

across all platforms. The format is supported by

numerous CAD software vendors, including

Graphisoft , AutoDesk and Bentley Systems . These

CAD tools allow for the development of 3D building

models, which are semantically marked-up, i.e.

window objects are added to the model instead of

abstract block references made up of individual

“polylines”.

An excellent example of the potential of

semantic models within the construction industry is

described by Wu et al (2004) who developed a

system for generating space model graphs based on

data from IFC files. This was possible primarily

because the IFC format describes the relationships

between objects, e.g. doors and spaces – a function

that is not possible without semantic modelling. This

functionality allows for quicker and easier analysis

of building space, but the same technique can be

extended to support automated path creation,

facilitating user navigation support of complex

buildings. Whilst the design of the IFC format

clearly provides potential for sophisticated and

powerful functionality, it is currently severely

hampered by the support of a static building model,

i.e. it is not currently possible to include the

dimension of time, thus facilitating the storage and

usage of historical data. Spearpoint (2003) identified

that the IFC model only provides a static view of a

building and that this is insufficient for the purpose

of running building simulations.

We propose a solution that merges the

semantically rich, yet static, building information

provided within IFC files with the innately dynamic

capabilities of temporal databases and object-

oriented software, thus supporting the development

of truly intelligent buildings. Such models would

allow IB software to mine, learn and dynamically

respond to all sensed building state changes as well

as user interactions.

3 TEMPORAL DATABASES

Temporal databases differ from typical ‘snapshot’

databases by supporting the management and

querying of historical data. Snapshot databases store

only the current state of records i.e. each change in

an object’s state replacing the previous state through

SQL (Structured Query Language) ‘update’

statements. Likewise, SQL delete statements are

used in order to remove records that are no longer

ICISO 2010 - International Conference on Informatics and Semiotics in Organisations

current, e.g. a customer cancelling their subscription

to a particular service which effectively implies that

the customer never had a relationship with the

company. In contrast, historical databases do not

delete records (without administrator intervention)

and only update and insert new records in order to

maintain an object’s historical audit trail. This

supports data mining and the recognition of trends as

results of object states and aggregate statistics can be

compared at different time points.

Temporal databases manage time values in one

of two ways: i) valid time and ii) transaction time.

Valid time denotes the time period during which a

fact is true with respect to the real world.

Transaction time is the time period during which a

fact is stored in the database. Some databases also

support a bi-temporal mode by including both valid

time and transaction time support.

In terms of building information, a temporal

database could be used, for example, to record

temperature changes within defined spaces. By

assigning a ‘start’ and ‘end’ (valid time) timestamp

to each record, it is possible to develop a complete

history of temperature fluctuations in different

rooms, and these records could be searched in order

to find the temperature of a given room at any past

point in time, or used to find the average

temperature during a period of time. This is one,

fairly simplistic example, however, the potential to

creating temporal relationships gives rise to much

more advanced functionality, including the ability to

support mobile sensors. If sensors in a building were

able to move between spaces, their readings could be

linked to the correct space objects. In this instance it

would be possible for the sensor to create records for

each temperature reading, but for each reading to be

attributed to the correct room. As a result of the

potential to create temporary relationships between

objects, and track the full history of all object

interactions, IB software would be provided with a

wealth of knowledge that could be mined to

recognise patterns to support resource efficiency and

user well-being optimisation.

There are, however, several difficulties to be

overcome in the development of a temporal database

based IB model, including: the complexity of design,

implementation and mining of data, as well as the

potential performance problems due to the complex

requirements of managing records. Perhaps the

greatest obstacle to the adoption of temporal

databases, however, is the lack of support from

database vendors.

Whilst some vendors, particularly Oracle, have

recently made efforts to include temporal features

into their database products, no single product

supports all the required features which are required

for a true temporal database. As a result the

management and querying of temporal data is still a

cumbersome and complex process due to the lack of

supported temporal features in existing RDBMS

(Relational Database Management System)

products. This state of affairs is due, at least in part,

to the exclusion of temporal features from the SQL

ISO standard.

In addition to the technical difficulties, there are

also social obstacles to be overcome, especially

when attempting to use such a database for the

purpose of tracking and recording occupant

behaviour, as such a system threatens damaging the

trust relationship between staff and management.

Moran et al. (2010) investigated this problem and

provided suggestions as to how organisations should

deal with the issues including: how data is collected

and for what purpose, access rights to the data and

the length of time temporal data is held by the

organisation.

4 PROTOTYPE OF THE

PROPOSED SOLUTION

In response to the need for a dynamic building

model standard, we propose the use of an adaptive

building model based on IFC and temporal

databases. The IFC model is capable of providing a

complete, yet static, representation of a building’s

structure and content, whilst the temporal database

will handle the recording, management and retrieval

of all object state changes. Figure 1 shows a diagram

of the proposed solution including the relationship

and differences between the proposed building

model and the currently available BMS technology.

A prototype system was developed using object-

oriented technology in order to support a continuous

simulation of the building, where each item has a

corresponding software object. Any change to these

objects, as reported by building sensors, is

immediately updated in the run-time software

objects and the related database records. The current

version of the prototype focuses on the simulation of

sensor instances linked to spaces and supports the

recording of sensor readings in a temporal database

as well as the visualisation of these readings.

For the sake of simplicity, the prototype

currently supports a variable number of temperature

sensors that can be set to create a new reading every

n number of seconds for a specified time period. The

readings are recorded in a temporal database that can

be visualised with a simple web-based interface. The

simulation was developed using an open-source

platform including linux, Java and MySQL. The

visualisation was developed with Apache, PHP and

INCLUSION OF TEMPORAL DATABASES WITH INDUSTRY FOUNDATION CLASSES - A Basis for Adaptable

Intelligent Buildings

Figure 1: Abstract System Architecture.

the PHP GDGraph library, which are connected to

the MySQL database. By using the open-source

platform we have promoted the potential for

distributing the solution to multiple servers, each of

which would be responsible for managing a part of a

building, which would likely be prohibitively

expensive if commercial technologies were used.

The current version of the database schema is

presented in figure 2. The sensor table includes the

following fields: id (The unique id key for the

sensor), current location (the unique id of the

sensor’s current space), sensor type (e.g. fire

detector/ temperature sensor) and start and end

fields, which record the time of the sensor’s

installation and de-commissioning.

The ‘sensor_temp_hist’ table records the state

changes of temperature sensors. Every time a sensor

object reports a different temperature, the database is

updated by ending the old state, i.e. updating the

‘end time’ field and creating a new record which

includes the sensor ID, current temperature reading

and the new start time. The current state of a sensor

always has a ‘NULL’ value in the ‘end time’ field.

The ‘curstate_temp’ table contains one tuple per

sensor object and records that sensor’s current state.

This information is also contained by the

‘sensor_temp_hist’ but has been included in the

‘curstate_temp’ table in order to prevent length full

table scans when searching for current object states.

The same pattern has been used with the

‘sensor_state_hist’ and ‘curstate_state’ tables, which

record the state of the sensors e.g.

active/inactive/malfunctioning, as well as the

‘location’ and ‘loc_history’ tables, which store the

changes in the location of sensors.

The prototype’s functionality is provided by the

following classes:

1. DbManager: provides database Application

Programming Interface (API) functionality

2. EventReading: represents a sensor reading event

3. ReadingManager: An object which manages

received EventReading objects. In future, multiple

ReadingManager objects could be useful in

buffering or distributing the application among a

number of application and/or database servers

4. Sensor: Sensor objects represent real-world

sensors which may be of a number of types and

create EventReading objects which represent their

readings.

5. SensorManager: An object which manages a

group of sensors.

6. SensorManagerTask: An object which extends

‘TimerTask’ allowing the simulation to trigger a task

every n seconds

7. SensorSim: The main class which starts and

controls the simulation

The visualisation was developed using

GDGraph, an open source PHP graph library. The

script reads the temporal database records and draws

a temperature graph as a PNG image (see figure 3).

Figure 2: Abstract Database Schema.

ICISO 2010 - International Conference on Informatics and Semiotics in Organisations

5 DYNAMIC BUILDING MODEL:

IMPLICATIONS AND

APPLICATIONS

The following example demonstrates the

applicability of combining temporal databases and

the IFC file content to the dynamic building model;

assuming that the model were to be continuously

updated by accurate sensor data. For this example,

we will use an imaginary office building termed

Building A, which has a dynamic building model

based on a temporal database and semantic

information from the building’s ‘as-built’ IFC file.

Building A is occupied during most of every day

by Bob who is monitored as he enters, moves around

and leaves the building. Bob’s movement between

spaces (possibly rooms or desks in terms of an open-

plan area) as well as the spaces’ temperatures are

recorded in the temporal database and the

relationships between temporal entities, a uniquely

identifiable item object or object state, is maintained

e.g. during a specified period of time Bob is linked

to the record of the space which he occupied during

that period, and, temperature readings are linked to

each space.

Given sufficient time, the system will build up

enough records to support the mining of data in

order to recognise Bob’s typical pattern of

behaviour. For example, the time he normally enters

or leaves the building could be calculated by taking

the average entry and exit times. Likewise, Bob’s

main space (his desk or office) could be

automatically inferred based on the space in which

he spends the highest percentage of his time. The

typical temperature of his most-occupied space

could also be calculated, and assuming that there are

no recorded instances of Bob changing the

temperature settings, we can deduce that the

recorded temperatures are within the range at which

Bob feels comfortable. These simple measurements

can be semantically linked to automatically provide

the system with accurate information relating to

individual occupant behaviour and preferences. This

information could then be used to automatically

configure spaces to meet occupant comfort settings,

in this case: temperature. This principle could,

however, be applied to other factors such as: lighting

and HVAC. Such knowledge could then be used to

plan building ecological strategies, seating plans,

etc., to ensure efficient use of building space.

Other recognised patterns can be used to directly

affect the building’s response to the occupant in

order to provide personalised settings to ensure their

comfort or to ‘intelligently’ respond to the user’s

behaviour in order to save resources (both financial

and energy). For example, if the occupant is

recognised as frequently leaving and returning to his

main space, it would be inefficient for the system to

continuously switch his space’s lighting on and off.

Likewise if the person leaves the building around

their expected leaving time, the system can

automatically switch off the lighting and electronic

appliances in their space in order to reduce energy

wastage. Crucially, by continuously re-examining

these patterns in the context of recent data, the

system can recognise when patterns change and

respond accordingly. For example, if Bob moves to

a different office, even if uninformed the system will

recognise this based on his recent behaviour. Any

automated system actions relating to Bob, such as

the setting of his heating or lighting preferences, will

then take place in the new space.

The recording of users in space and the

recognition of patterns in this data could also be very

beneficial for facilities managers (FM). The role of

FM staff includes space analysis, space usage

optimisation, support for the cleaning and

maintenance of facilities, which is often reliant on

accurate information relating to the number of

occupants in the building, the changes in such trends

and the availability of this information in a timely

manner. Currently, many FM managers struggle to

keep up with these duties due to the varying number

of occupants as a result of the increasing adoption of

flexible working (Lindkvist, 2009). This has lead to

unreliable predictions of occupant numbers. The

system could support these requirements to some

degree, but may struggle with highly dynamic

environments as historical data may be insufficient

to predict occupant behaviour.

Table 1: Sample Data.

Table 1 displays a sample of test data that was

automatically generated by the prototype. The sensor

objects have been set to automatically generate a

random new temperature reading every 3 seconds.

The start and end time fields delimit the various

temperature states. This arrangement can be applied

to all other temporal relations including location,

state (e.g. active/inactive etc) as well as relationships

between objects. The visualisation of recorded data

(see figure 3) is important as it supports human

operators examining changes that take place in the

Sensor

Temp

(C)

Start Time End

Time

1 18 17:09:52 17:09:54

2 16 17:09:52 17:09:54

3 18 17:09:52 17:09:54

4 17 17:09:52 17:09:54

1 16 17:09:55 17:09:57

2 15 17:09:55 17:09:57

3 19 17:09:55 17:09:57

4 21 17:09:55 17:09:57

INCLUSION OF TEMPORAL DATABASES WITH INDUSTRY FOUNDATION CLASSES - A Basis for Adaptable

Intelligent Buildings

building and for identifying trends. Table 2: Sample

Data

The main purpose of the prototype was to

demonstrate the applicability of temporal databases

to the built environment and intelligent spaces for

the purpose of supporting IB adaptability. We also

wanted to explore the issues and problems related to

their use in building management and learning.

The temporal database was designed in a manner

that supports frequent access to both historical

records and the current states of objects, as based on

temporal database literature (Gregerson, 1999), with

an extra table holding only the current states to

prevent lengthy full-table-scan database queries

when selecting current states.

The database structure includes the concept of

‘spaces’ which can easily be populated from IFC

files. By establishing a temporal relationship

between spaces and their contained sensors, it is

possible for sensors to be moved between spaces and

for their subsequent readings to be linked with

spaces at the time the readings were taken. This

assumes of course that the location of the sensors is

updated accurately and in a timely fashion. This

prototype functionality, to our knowledge, is

currently unavailable within IB software.

The richness of the building model relies heavily

on accurate input from a wide range of sensors and

sensor types. This has several implications,

including the cost of sensors and related technology,

data security and occupant trust. We believe that the

potential benefits of increased energy efficiency and

provision of personalised space settings will over

time outweigh the disadvantages related to these

factors.

6 FUTURE WORK

Since the demonstrated prototype is still in its

infancy, we plan to extend it in several ways. We

need to extend the database schema to include more

temporal entities, develop test queries that are time-

bound e.g. “In which space was Bob at 11am

yesterday?” as well as develop IFC import

functionality. We are considering the development

of a more abstract API, since this would be useful

when linked to building simulation software (e.g.

fire or lighting). Such an API would support the use

of a dynamic building model for running building

simulations, which has been commented as a

deficiency when using IFC files for such simulations

(Spearpoint, 2007). In addition we also plan to link

this work with a multi-agent system, which will

support various intelligent agents for performing a

multitude of tasks, including data mining.

Other proposed areas of application may also

include development and testing of the system to

support a distributed data structure. This may

involve distributing the model among multiple

servers, for parts of a building, and required research

concerning communication relating to building state

changes. Each server might also be responsible for

mining it’s own set of recorded data for patterns.

The currently proposed prototype is currently

only partly dynamic, as it does not support changes

to the building’s physical structure. This structure is

fixed based on the content of the imported IFC file.

In the future, we plan to allow for the import of

updated IFC files which will allow the dynamic

building model to remain up-to-date with changes in

the building’s structure.

Figure 3: Temperature Graph (X: Time Y: Temperature).

7 CONCLUSIONS

This paper has reviewed a number of state of the art

technologies and tools in intelligent buildings in

relation to their current inability to support dynamic

and historical building models. Such technologies

provide a starting point, and we have shown can be

modified and extended to support the required

functionality. Finally we proposed a novel prototype

based on IFC and a temporal database, and

demonstrated the feasibility of this approach.

The prototype fulfils a number of roles in our

research. These roles include: the development of an

ideal temporal database structure for storing building

information, a test-bed for the creation, recording,

management and mining of temporal data as well as

the model’s software objects. The prototype will aid

us in discovering potential issues and problems

related to the management of temporal building data

and will support us in providing recommendations

for a dynamic building model standard.

We hope that our development of this model will

support future intelligent buildings by allowing them

ICISO 2010 - International Conference on Informatics and Semiotics in Organisations

to dynamically monitor and respond to

environmental changes and thus meet the

requirement of adaptability.

ACKNOWLEDGEMENTS

We wish to thank CDC (Central Data Control) and

the EPSRC (Engineering and Physical Sciences

Research Council) for supporting and funding our

work.

REFERENCES

Cheng T. And Wang J., “Applications of spatio-temporal

data mining and knowledge discover (STDMKD) for

forest fire prevention”, ISPRS Commission VII Mid-

term Symposium "Remote Sensing: From Pixels to

Processes", Enschede, the Netherlands, 8-11 May

2006, Reed Business Geo. 2006

Clements-Croome, D. (ed.) Intelligent Buildings: Design,

Management and Operation. Thomas Telford P. Heron

Quay, London, 2004. pp. 39

Eastman C., Aug. 2009, “Building Information Modeling,

What is BIM?” (Online) Available:

http://bim.arch.gatech.edu/content_view.asp?id=402

(Accessed: 16 May 2010).

Fritz W., “Intelligent Systems and their Societies:

Glossary”, “27 Jan 2010”, (Online) Available:

http://www.intelligent-

systems.com.ar/intsyst/glossary.htm (Accessed: 16

May 2010).

Gregerson H. And Jensen S.C. “Temporal Entity-

Relationship Models – A survey”, IEEE Transactions

on Knowledge and Data Engineering. Vol 11, No. 3,

May/June 1999, pp.484

Knibbe, E.J., “Building Management System” U.S. Patent

5,565,855, Oct. 15, 1996. (Online) Available:

http://www.google.co.uk/patents (Accessed: 16 May

2010).

Lindkvist C and Elmualim A, “Pervasive technologies for

workspace management”, Journal of Facilities

Management, Vol. 7, No.2, 2009, pp 98-110.

Moran S. and Nakata K., “Analysing the factors affecting

users in intelligent pervasive spaces”, Intelligent

buildings International 2 (2010), 57-71,

doi:10.3763/inbi.2009.0040, 2010 Earthscan ISSN:

1750-8975(print), 1756-6932 (Online) Available:

www.earthscan.co.uk/journals/inbi (Accessed: 16 May

2010).

Smith K. Dana and Tardif M., “Building Information

Modeling A Strategic Implementation Guide for

Architects, Engineers, Constructors, and Real Estate

Asset Managers” New Jersey: John Wiley & Sons,

2009 pp. 35.

Snodgrass R., “Temporal Databases Status and Research

Directions”, Sigmod Record, Vol 19, No. 4, 1990

(Online)

Available: http://citeseerx.ist.psu.edu/viewdoc/

(Accessed: 16 May 2010).

Spearpoint, M.J., “Integrating the IFC Building Product

Model with Zone Fire Simulation Software”, In: Proc.

International Conference on Building Fire Safety,

QUT, Gardens Point Campus, Brisbane, Australia,

20—21 November 2003, pp. 56—66 (Online)

Available: http://ir.canterbury.ac.nz/ (Accessed: 16

May 2010).

Spearpoint M J, Dimyadi J A W (2007) Sharing fire

engineering simulation data using the IFC building

information model. Christchurch, NZ: MODSIM07,

International Congress on Modelling and Simulation,

10-13 December 2007.

Wu S. Et al, “An IFC-based space analysis for building

accessibility layout for all users”, Construction

Innovation 2004; 4: 129–141.

INCLUSION OF TEMPORAL DATABASES WITH INDUSTRY FOUNDATION CLASSES - A Basis for Adaptable

Intelligent Buildings