Butler-ising HomeManager: A Pervasive Multi-Agent System for Home

Intelligence

Enrico Denti and Roberta Calegari

Dipartimento di Informatica - Scienza e Ingegneria (DISI), Alma Mater Studiorum—Universit

a di Bologna, Bologna, Italy

Keywords:

Home Management, Ambient Intelligence, Pervasive Computing, Energy Saving, Smart Homes, Smart

Living, Coordination Infrastructures, Multi-Agent Systems, Gamiﬁcation.

Abstract:

Home Manager is an agent-based application for the control of an intelligent home, where the house is seen

as an intelligent environment made of independent devices that participate to an agent society. The society

is governed by a coordination infrastructure aimed at satisfying the users goals and preferences (lighting,

temperature, etc.) while achieving the global house policies and objectives (e.g. energy saving) in a highly-

conﬁgurable way. In the existing prototype, designed mostly to prove the feasibility and effectiveness of the

above approach, the testbed house was kept intentionally simple, with a limited number of rooms, user types,

control devices and policies, and the infrastructure implementation lacked some features.

The recent, widespread adoption of smart mobile devices (smartphones, tablets) enabling mobile connectivity

has dramatically changed the reference scenario: users now expect at least to be able to monitor, and possibly

control, their home devices in mobility, and in fact all major vendors now offer some app for this purpose. Yet,

this is just the basic step: exploiting the situated connectivity enabled by GPS and the other geo-localisation

techniques embedded in today’s smartphones, novel pervasive scenarios can be devised that could not even

be imagined in the past years. This aspect is developed in the Butlers architecture, which provides a general

framework and reference model for intelligent home management where the smart home is managed by an

intelligent butler and interacts with its inhabitants taking into account their habits, behavior, location, prefer-

ences and any other sort of information to anticipate their needs and support their goals.

In this context, this paper presents the novel “Butler-ised” Home Manager, that evolves the previous system

in the Butlers perspective: the new prototype not only supports the remote control of the house appliances via

an Android app, but exploits the user position, tracked via geo-localisation, to anticipate the user’s needs in a

simple, yet signiﬁcant, scenario – namely, autonomously switching the house oven on when discovering that

the user has just bought a take-away pizza in his/her way back home.

1 INTRODUCTION

Technology evolution is making the appliances that

populate our homes smarter and connected: it is

common to ﬁnd TV sets, air conditioners, washing

machines, refrigerators, etc. networked and often

remote-controlled, typically via an app for Android or

iOS smartphones so that integration frameworks are

being proposed (Google, 2014; Apple, 2014).

These trends are enlarging the application per-

spectives of “traditional” home automation in the con-

sumer market, merging aspects from Ambient Intel-

ligence (AmI) (Ducatel et al., 2001; Ayala et al.,

2013), smart environments (Bartolini et al., 2012), en-

ergy monitoring (Innova, 2012; World, 2013) and do-

motics (Ch

e et al., 2010; Menon et al., 2013) to focus

onto people’s needs from different viewpoints (Chong

and Mastrogiovanni, 2011).

In Ambient Intelligence, the emphasis is on user-

friendliness, user-empowerment, and support for hu-

man interaction: the home environment is “smart” in

that it can handle some aspects (e.g. lighting, heating)

based on the user’s preferences, so as to improve the

quality of life. This is why it is often found in “as-

sisted living” applications, whose aim is to support

speciﬁc user categories (elderly, disabled, etc).

Recent works, like (Denti, 2014), emphasise that

there is an extra value in considering all these aspects

together: novel, intriguing scenarios are devised that

could not even be imagined in the past years. In such

new scenarios, the smart home can interact with its

inhabitants not only to monitor and remote-control

the home appliances, but also to take into account

the users’ habits, behavior, location, preferences and

any other sort of information to anticipate their needs

and overall support their goals. The basic idea is that

knowing the user’s habits and observing his/her daily

behavior via mobile devices – in particular his/her

Denti E. and Calegari R..

Butler-ising HomeManager - A Pervasive Multi-Agent System for Home Intelligence.

DOI: 10.5220/0005284002490256

In Proceedings of the International Conference on Agents and Artiﬁcial Intelligence (ICAART-2015), pages 249-256

ISBN: 978-989-758-073-4

 2015 SCITEPRESS (Science and Technology Publications, Lda.)

location, thanks to the situated connectivity enabled

by GPS and the other geo-localisation techniques –

can enable an intelligent system (the home butler)

to take autonomous decisions and possibly anticipate

the users’ needs, managing the home devices on the

user’s behalf. In (Denti, 2014), these aspects are

put in context with other research results from in-

telligent agents, multi-agent systems (mainly in the

area of power generation and consumption (Tolbert

et al., 2001), power restoration (Nagata et al., 2000),

load management in power grid systems (Zhang et al.,

2011)), and coordination technologies (Papadopou-

los and Arbab, 1998; Omicini and Papadopoulos,

2001; Busi et al., 2001), on the one side, and “enter-

tainment” aspects, in the gamiﬁcation (Gamiﬁcation

Community, 2013) perspective, on the other – the lat-

ter being more and more recognised as a key factor

between success and failure in technology acceptance

from the consumers’ viewpoint.

Home Manager (Molesini et al., 2009) is a pro-

totype agent-based application for the control of an

intelligent home, where the house is seen as an intel-

ligent environment made of independent devices that

participate to an agent society: the system aims to

manage the overall energy consumption while sup-

porting the user’s living and activities inside the

house. The agent society is governed by the TuCSoN

coordination infrastructure (Omicini and Zambonelli,

1999; Ricci et al., 2002; Omicini and Rimassa, 2004;

Mariani and Omicini, 2014; TuCSoN, 2008) and tries

to satisfy the user’s goals and preferences as concerns

the lighting and room temperature, while achieving

the global house policies and objectives. Since the

above prototype was designed mostly to prove the fea-

sibility and effectiveness of the approach, the testbed

house was kept intentionally simple, with a limited

number of rooms and user categories, and a small set

of home appliances and policies; moreover, the im-

plementation was rooted on TuCSoN 1.4, where some

relevant features were not fully available.

In the Butlers (Denti, 2014) perspective, Home

Manager can be seen as an early implementation of

a small subset of the full architecture – in particular,

with no support for mobility and pervasive aspects.

Accordingly, this paper presents the new “Butler-

ised” Home Manager—the evolution of the previous

system in the Butlers perspective: the new prototype

not only supports the remote control of the house ap-

pliances via an Android app, but exploits the user po-

sition, tracked via geo-localisation, to anticipate the

user’s needs in a simple, yet signiﬁcant, scenario –

namely, autonomously switching the house oven on

when discovering that the user has just bought a take-

away pizza in his/her way back home.

So, after shortly summarising Home Manager

(Section 2), the Butlers vision (Section 3), and the

TuCSoN infrastructure (Section 4), we present the

new Home Manager system (Section 5), from the

main requirements to the implementation, with the

related discussion (Section 6). Related work and con-

clusions are reported in Sections 7 and 8, respectively.

2 HOME MANAGER

Home Manager (Molesini et al., 2009) is a prototype

application for the control of an intelligent home, de-

signed as a multi-agent system via the SODA method-

ology (SODA, 2008) and implemented on top of the

TuCSoN coordination infrastructure (TuCSoN, 2008).

The system considers a house with independent

devices (air conditioners, lights, etc.), each equipped

with an agent to participate to the agent society. The

coordination infrastructure, programmable via tuple

centres, embeds the coordination laws required both

to mediate among the different user’s preferences and

to pursue the overall system goals – in this case, to

manage (limit) the overall energy consumption.

The system tries to satisfy the user preferences

in terms of room lighting and temperature, unless

higher-order energy constraint are violated; if multi-

ple users with different preferences are in the same

room, it also mediates among them by applying some

suitable global policy. It features a rather complete

role-based model, and a (pc-based) graphical user in-

terface to conﬁgure and use the system with no need

to operate directly on the underlying infrastructure.

Coherently with its proof-of-concept nature, the

testbed was kept intentionally simple: no real mobil-

ity aspects were included (users were only considered

inside the house), no remote control capabilities were

supported, and the user’s proﬁle was limited. Also,

there was no idea of anticipating any user need or de-

sire – because “anticipating” an action requires to be

informed of what is occurring, and where.

More recently, the Home Manager system has

been re-interpreted, given its goals and features, in the

Butlers perspective (brieﬂy summarised in the next

section), positioning the prototype in the Butlers con-

ceptual reference layers (reported, with the author’s

permission, in Figure 1 for the reader’s convenience).

3 THE BUTLERS VISION

The Butlers architecture (Denti, 2014) deﬁnes a

framework with seven conceptual layers, which re-

late the availability of physical devices and enabling

Figure 1: Butlers multi-layer reference architecture.

technologies with the set of features that a home man-

agement system can expectedly provide, on the one

hand, and with the corresponding value-added for

users, on the other (Figure 1). The bottom layers

concern mainly enabling technologies (mono or bi-

directional communication-enabled sensors, meters,

actuators, etc.), while the middle layers are mainly in-

frastructural / middleware layers (aimed at providing

coordination and geographical information services);

the top layers, instead, are not necessarily to be taken

in the sequence, for they focus on speciﬁc aspects like

intelligence, social aspects, and gamiﬁcation (the lat-

ter seen as a key success factor to promote technol-

ogy acceptance in the mass market). The resulting

conceptual map can be used both to locate a given

system based on its feature – for instance, most of to-

day’s remote-controllable appliances, accessible via

Android or iOS apps, are clearly located at level 2,

while the Home Manager prototype above can be eas-

ily located at level 3, with minor aspects from level 5

– and, conversely, to identify the unexplored market

niches – that is, possibly-interesting systems that are

not currently available, suggesting their development.

This approach suggests new, intriguing scenarios.

While a complete discussion is outside the scope of

this paper, what is relevant here is that a smart home

could interact with its inhabitants not only to monitor

(level 1) and remote-control (level 2) the home appli-

ances, but - provided that a suitable coordination in-

frastructure is available (level 3) - also to take into ac-

count the users’ habits, behavior, location, and prefer-

ences (level 4) to reason on the overall situation (level

5) so as to possibly anticipate the user’s needs. In

principle, the social networks (level 6) and the gam-

iﬁcation perspective (level 7) could also be exploited

as a further source of information; however, these as-

pects are also outside the scope of this paper and will

not be further considered in the following.

The resulting architecture, where the envisioned

system is represented as an intelligent “home direc-

tor” – the butler –, is shown in Figure 2(a): at this

stage, the architecture is still technology-neutral—no

speciﬁc coordination technology is selected, nor is the

architecture necessarily agent-based. However, if an

agent-based approach is adopted and the TuCSoN co-

ordination technology is selected, the architecture can

be reﬁned as in Figure 2(b), where the TuCSoN coor-

dination artifacts (tuple centres, TCs, and agent coor-

dination contexts, ACCs) are explicitly highlighted.

This process is described in full in (Denti, 2014):

here, we just provide the essential details in Section 5

after summarising the basics of the TuCSoN model

and infrastructure.

4 TuCSoN IN A NUTSHELL

TuCSoN (Tuple Centres Spread over the Network)

(TuCSoN, 2008) is a coordination model and in-

frastructure based on a programmable coordination

medium, the tuple centre, which is a tuple space en-

hanced with the notion of behaviour speciﬁcation.

Since the speciﬁcation language, ReSpecT (Casadei

and Omicini, 2009; Casadei and Omicini, 2010), is

Turing-equivalent and both time-situated and space-

situated, so as to perceive the environment as needed,

any coordination-related computation can potentially

be expressed, including those that need to perceive/act

on the surrounding environment. As a result, the co-

ordination tasks can be charged on top of the coordi-

nation media, where they conceptually belong, rather

than onto the agents’ shoulders. Agents coordinate

by accessing and consuming tuples – ordered sets of

data chunks – to/from the tuple centres, by means of

the three basic read, in and out primitives.

Figure 2: The Butlers architecture in general (a) and in the TuCSoN-based concretisation (b).

The Agent Coordination Context (ACC) is the ba-

sic abstraction for modeling the space of interaction

and agents communication, as well as the perception

of the environment in which it is located. Roughly

speaking, it can be seen as the conceptual boundary

between the agent and the infrastructure: its aim is

to bind and govern the interaction between the agent

and the infrastructure, based on a customisable set of

rules that deﬁne what the agent is allowed/denied to

do. Any agent entering a TuCSoN-governed system

must ﬁrst ask for (negotiate) a suitable ACC, which is

released with the rules granting the agent the appro-

priate access rights; in their turn, the ACC conﬁgura-

tion is controlled by higher-level policies, set by the

system administrator in the conﬁguration phase.

So, properly-conﬁgured ACCs and tuple centres

together can be exploited also to design and enforce

the security aspects required in a distributed system,

encapsulating authorisation and access control poli-

cies with the desired granularity.

5 BUTLER-ISING THE SYSTEM

Our goal is to extend Home Manager towards the

Butlers layer 4 and 5, exploiting the user’s loca-

tion – tracked in real time thanks to the GPS and

the other geo-localisation techniques embedded in

modern smartphones – to enable an intelligent rea-

soner agent to take some autonomous decisions (for

instance, adjusting the air conditioner temperature),

possibly even anticipating some user’s needs, man-

aging the related devices on the user’s behalf (for

instance, deducing the opportunity to switch on the

oven, or post-pone the washing machine, etc.).

More precisely, we mean i) to add the support for

mobility, enabling users to operate on the home ap-

pliances remotely via a suitable Android application;

ii) to geo-localise the user in real time, reifying this

information into the Home Manager system as a suit-

able tuple; iii) to exploit this information to detect rel-

evant user patterns – for instance, whether he/she is

buying a take-away pizza when coming home – and

consequently anticipate his/her needs proactively – in

this case, to switch on the oven at 150

◦

C so as to

warm the pizza (possibly rescheduling other energy-

consuming tasks if necessary).

Figure 3 highlights the architectural changes. On

the left side (a), the TuCSoN-based architecture of

the general Butlers system depicted in Figure 2 (b)

is downsized to the limited set of features supported

by the original Home Manager prototype (Section 2).

With respect to the full system, the selected sub-

system considers only lighting, air-conditioning and

heating, has no intelligent Butler recognisable as

Figure 3: Home Manager architecture: current (a) vs Butlers-oriented (b).

Figure 4: Home Manager Remote Control Screen.

such, user localisation is limited to the interior of

the house and is simulated via GUI, and no so-

cial/gamiﬁcation aspects are supported. Instead, the

new “butler-ised” prototype (b) aims to introduce

an Android app to achieve the goals (i) and (ii)

above, and a “prompter” (reasoner) agent, in charge

of providing suggestions for autonomous actions (is-

sue (iii)).

As for issue (i), the app allows the user, after au-

thentication, to remote-control the home appliances:

Figure 4 shows the case of the air conditioner.

As for issue (ii), the user position is reiﬁed as a

tuple of the form:

geo position(user, lat, lon)

Such tuples are stored in a speciﬁc tuple centre, which

is programmed to generate a time-stamped version of

the same tuple, like the following:

geo position(user, lat, lon, time)

used to record the user position history.

This information is then exploited, together with

any other relevant information available in the user

Figure 5: Notiﬁcation of the prompter agent activity.

preferences, proﬁle, etc., by the prompter agent to for-

mulate its suggestions (issue (iii)): the consequent ac-

tions are autonomously performed, notifying the user

via a text message on his/her phone (Figure 5). Of

course, the user can change/cancel the action if, for

any reason, he/she does not like it.

6 DISCUSSION

With respect to the original Home Manager described

in (Molesini et al., 2009), the “Butler-ised” version

overcomes the pc-based simulated control, in favor

of an actual mobile support: users can interact with

the system dynamically, in mobility, and are free to

change their mind and programs any time and any-

where. On the other hand, the smart home is still sim-

ulated, and no real device is actually connected.

The value-added of this prototype, therefore, lies

mainly in its step towards the Butlers vision: by

adding the ability to reason over the user position, it

shows – albeit in a very simpliﬁed, preliminary form

– the potential applications and the innovative perva-

sive scenarios envisioned in (Denti, 2014).

Clearly, the current implementation is just a proof

of concept: a lot of work remains to be done to ac-

tually support the Butlers upper layers, both on the

Home Manager system itself, and on the interface

with actual (non-simulated) appliances. For instance,

just to mention some aspects, the user authentica-

tion is currently un-ciphered, the reasoning of the

prompter agent is barely trivial, no user proﬁle infor-

mation is actually exploited to customise the sugges-

tions, etc.—not to mention the social and gamiﬁca-

tion aspects, which are totally missing at this stage.

At the same time, the choice of the TuCSoN infras-

tructure as the underlying mechanism conﬁrms to be

a winning point: its ability to bridge between the dif-

ferent agent perceptions, particularly in the handling

of the geo position tuples, turned out to be essential

to add the new features smoothly, and with virtually

no impact on the existing system.

Widening the view, the aspect of situatedness, in-

tended as the strict coupling with the environment,

is emphasised as more and more crucial in today’s

complex computational systems (Weyns et al., 2007;

Omicini and Mariani, 2013).

Accordingly, one further dimension, deﬁnitely

worth exploring, is the speciﬁc support introduced

in the latest TuCSoN edition (Mariani and Omicini,

2014), which explicitly considers environment probes

and transducers as native metaphors. Moving from

the basic consideration that the user’s daily activi-

ties are unpredictable, but mostly depend on the en-

vironment being in some proper “enabler state” (and,

conversely, impact on / are affected by the environ-

ment), the Home Manager/Butlers scenario could be

matched on the TuCSoN meta-model by looking at en-

vironment changes as generated by probes and medi-

ated by transducers, thus enabling a uniform represen-

tation of environmental properties despite the appli-

ances’ heterogeneity. Physically, agents would then

be deployed to personal devices (like smartphones or

desktop PCs, depending on their role and function),

as in our new prototype, and probes to home appli-

ances (as assumed by the Butlers’ architecture), while

ACCs, tuple centres and transducers could be put ei-

ther on the user smartphone or remotely (the desktop

PC), based on design considerations.

7 RELATED WORK

Given the broad scope of this paper, the relevant liter-

ature spreads onto several different research areas—

from smart homes, domotic and home automation to

energy monitoring and saving, ambient intelligence,

assisted living and healthcare, and of course multi-

agent systems, coordination models and infrastruc-

tures.

Due to space restrictions, the references that have

already be discussed above will not be reported again.

Domotic and home automation systems have al-

ways aimed at automating the user interactions with

home appliances, in several ways (Ch

e et al., 2010;

Menon et al., 2013). Assisted living applications also

aim to support the user, though from another per-

spective and context: in (Coronato and Pietro, 2010),

for instance, a pervasive application is presented that

aims to assist people everywhere and at any time.

In the smart home context, all the major players

are introducing smarter appliances, architecture and

applications. For instance Google in 2014 proposed

‘The Works With Nest’ approach, whose ecosys-

tem aims to integrate heterogeneous apps and ser-

vices from different vendors in an unique framework

(Google, 2014), while Apple deﬁned the Home Kit

software platform aiming to integrate third party com-

ponents and tasks with voice recognition capabilities

for their activation (Apple, 2014). More generally

speaking, appliance monitoring ranges from energy

meters such as (Innova, 2012; World, 2013) to high-

end air conditioners, fridges, etc. which today are

often provided with Android or iOS apps for remote

monitoring and control purposes.

In the multi-agent systems area, literature mainly

concerns on energy-management (mostly about en-

ergy balancing and control, from the viewpoint of

the energy provider than of the ﬁnal consumer): very

little exists on MAS applied to the management of

home-related aspects. In (Tolbert et al., 2001), a

scalable MAS controls the energy resources for re-

liability and efﬁciency purposes in power generation

and consumption, while in (Nagata et al., 2000) the

MAS handles the power restoration after faults. In

(Zhang et al., 2011) the MAS combines local intelli-

gence with coordination issues for load management

in power grid systems, integrating the advantages of

centralized and decentralized architectures. (Conte

and Scaradozzi, 2007) is perhaps the most interest-

ing example of a MAS used for home automation:

the aim is to schedule the operation of different appli-

ances according to the user deﬁned priorities, keeping

the global electric load below a given power thresh-

old, to prevent a shut-down.

8 CONCLUSIONS

In this paper we presented the new Home Manager

system, extended in the Butlers perspective.

Despite its many limitations and strong simpliﬁca-

tions, due to its proof-of-concept nature, the “Butler-

ised” prototype makes one ﬁrst, yet fundamental, step

towards the Butlers upper layers: by reifying the user

position via geo-localisation and adding the ability

to reason over such information, enabling the antic-

ipation of the user’s needs, it highlights the poten-

tial of the innovative pervasive scenarios envisioned

in (Denti, 2014), and shows their feasibility in princi-

ple.

For this reason, this result is more a starting point

than the end of the story: as discussed above, a lot

of work remains to be done, and many extensions, in

several different directions, are worth considering.

For instance, if the above-cited frameworks by the

major players become eventually popular, an integra-

tion with our approach could be studied, although

their goals refer mainly to the Butlers lower layers (2,

3 and possibly 4) while we mean to incrementally en-

able advanced features from the Butlers upper layers.

At the same time, serious difﬁculties have to be

expected when moving on, especially towards the up-

per Butlers layers: apart from technical difﬁculties

in enabling the interaction among so many heteroge-

neous entities, the effectiveness and the actual scala-

bility of the approach are still to be proved, and so are

the robustness and reliability of the TuCSoN infras-

tructure in such a challenging scenario. Proper knowl-

edge representation is also likely to be a critical issue

when the application scenario goes beyond the toy ex-

ample presented in this paper. Moreover, in the per-

spective, proper design metaphors and user metaphors

will be needed to deal with the complexity of layers 6

and 7, with their challenging goals of supporting ex-

perience sharing, advanced user involvement, etc.

So, our approach will be to proceed stepwise: the

ﬁrst planned step is to go beyond the single, pre-

deﬁned take-away pizzeria considered in this exam-

ple, adding the notion of user’s favorites and possibly

importing them from existing social networks (Face-

book, Google+, and others), thus adding some (min-

imal) layer 6 feature. The next step will be to inte-

grate the system with Google Maps (at least partially),

so as to grab the user’s check-in location (take-away

pizzeria, restaurant, chemist’s, etc) from that source,

removing the current requirement that it is known a-

priori. Apart from the better usability, this step would

strongly enhance the prompter agent role, while con-

stituting a major testbed both for the Butlers architec-

ture and the Home Manager system design.

ACKNOWLEDGEMENTS

Authors would like to thank Dipl. Eng. Ilaria Berto-

letti for her contribution to this project and her work

to the new prototype and the Android app, and Dr.

Eng. Leo Di Carlo for his work to the extension of

basic the prototype.

REFERENCES

Apple(2014). Home kit https://developer.apple.com/homekit/

Ayala, I., Amor, M., and Fuentes, L. (2013). Self-

conﬁguring agents for ambient assisted living ap-

plications. Personal and Ubiquitous Computing,

17(6):1159–1169.

Bartolini, S., Milosevic, B., D’Elia, A., Farella, E., Benini,

L., and Cinotti, T. (2012). Reconﬁgurable natural in-

teraction in smart environments: approach and proto-

type implementation. Personal and Ubiquitous Com-

puting, 16(7):943–956.

Busi, N., Ciancarini, P., Gorrieri, R., and Zavattaro, G.

(2001). Coordination models: a guided tour. Spinger.

Casadei, M. and Omicini, A. (2009). Situated tuple centres

in ReSpecT. In Shin, S. Y., Ossowski, S., Menezes,

R., and Viroli, M., editors, 24th Annual ACM Sympo-

sium on Applied Computing (SAC 2009), volume III,

pages 1361–1368, Honolulu, Hawai’i, USA. ACM.

Casadei, M. and Omicini, A. (2010). Programming agent-

environment interaction for mas situatedness in re-

spect. The Knowledge Engineering Review. Regular

Submission.

e, N., Pardons, N., Vanrompay, Y., Preuveneers, D., and

Berbers, Y. (2010). An intelligent domotics system to

automate user actions. In Ambient Intelligence and

Future Trends-International Symposium on Ambient

Intelligence (ISAmI 2010), volume 72 of Advances

in Intelligent and Soft Computing, pages 201–204.

Springer Berlin Heidelberg.

Chong, N.-Y. and Mastrogiovanni, F. (2011). Ambient In-

telligence and Smart Environments: Trends and Per-

spectives. IGI Global.

Conte, G. and Scaradozzi, D. (2007). Applying mas theory

to complex home automation systems. Modeling and

Control of Complex Systems.

Coronato, A. and Pietro, G. D. (2010). Pervasive and Smart

Technologies for Healthcare: Ubiquitous Methodolo-

gies and Tools. IGI Global.

Denti, E. (2014). Novel pervasive scenarios for home

management: the butlers architecture. SpringerPlus,

3(52):1–30.

Ducatel, K., Bogdanowicz, M., Scapolo, F. Leijten, J., and

Burgelman, J.-C. (2001). Scenarios for ambient intel-

ligence in 2010. ﬁnal, IPTS – European Commission’s

Joint Research Centre.

Gamiﬁcation Community (2013). The gamiﬁcation com-

munity portal. http://www.gamiﬁcation.org.

Google (2014). Works with nest.

http://techcrunch.com/2014 /06/23/google-makes-its-

nest-at-the-center-of-the-smart-home/.

Innova (2012). Regolazione automatica e telegestione, do-

motica. http://www.innova-srl.eu.

Mariani, S. and Omicini, A. (2014). TuCSoN coordina-

tion for MAS situatedness: Towards a methodology.

In Santoro, C. and Bergenti, F., editors, WOA 2014 –

XV Workshop Nazionale “Dagli Oggetti agli Agenti”,

volume 1260 of CEUR Workshop Proceedings, pages

62–71. Sun SITE Central Europe, RWTH Aachen

University.

Menon, V., Jayaraman, B., and Govindaraju, V. (2013).

Enhancing biometric recognition with spatio-temporal

reasoning in smart environments. Personal and Ubiq-

uitous Computing, 17(5):987–998.

Molesini, A., Denti, E., and Omicini, A. (2009). An agent-

based application for home intelligence. In Filipe, J.,

Fred, A., and Sharp, B., editors, International Con-

ference on Agents and Artiﬁcial Intelligence (ICAART

2009), pages 377–384, Porto, Portugal. INSTICC.

Nagata, T., Watanabe, H., Ohno, M., and Sasaki, H. (2000).

A multi-agent approach to power system restoration.

Power System Technology.

Omicini, A. and Mariani, S. (2013). Coordination for

situated MAS: Towards an event-driven architecture.

In Moldt, D. and R

olke, H., editors, International

Workshop on Petri Nets and Software Engineering

(PNSE’13), volume 989 of CEUR Workshop Proceed-

ings, pages 17–22. Sun SITE Central Europe, RWTH

Aachen University.

Omicini, A. and Papadopoulos, G. A. (2001). Editorial:

Why coordination models and languages in AI? Ap-

plied Artiﬁcial Intelligence: An International Journal,

15(1):1–10. Special Issue: Coordination Models and

Languages in AI.

Omicini, A. and Rimassa, G. (2004). Towards seamless

agent middleware. In IEEE 13th Inter. Workshops on

Enabling Technologies: Infrastructure for Collabo-

rative Enterprises (WET ICE 2004), pages 417–422.

IEEE CS.

Omicini, A. and Zambonelli, F. (1999). Coordination for

Internet application development. Autonomous Agents

and Multi-Agent Systems, 2(3):251–269.

Papadopoulos, G. A. and Arbab, F. (1998). Coordina-

tion models and languages. Advances in Computers,

46:330–401.

Ricci, A., Omicini, A., and Denti, E. (2002). Virtual en-

terprises and workﬂow management as agent coordi-

nation issues. International Journal of Cooperative

Information Systems, 11(3/4):355–379.

SODA (2008). Home page. http://soda.apice.unibo.it/.

Tolbert, L., Qi, H., and Peng, F. (2001). Scalable multi-

agent system for real-time electric power manage-

ment. IEEE Xplore Digital Library.

TuCSoN (2008). Home page. http://tucson.apice.unibo.it/.

Weyns, D., Omicini, A., and Odell, J. J. (2007). Environ-

ment as a ﬁrst-class abstraction in multi-agent sys-

tems. Autonomous Agents and Multi-Agent Systems,

14(1):5–30. Special Issue on Environments for Multi-

agent Systems.

World, A. (2013). Artie world. http://www.artie-world.com.

Zhang, M., Ye, D., Bai, Q., Sutanto, D., and Muttaqi, K.

(2011). A hybrid multi-agent framework for load

management in power grid systems. In Bai, Q. and

Fukuta, N., editors, Advances in Practical Multi-

Agent Systems, volume 325 of Studies in Computa-

tional Intelligence, pages 129–143. Springer Berlin

Heidelberg.