How ´ambient intelligence´ will enhance our experience of
everyday life
In World News 8.4, Stephano Marzano, managing director of Philips
Design, talked about the challenge of finding new opportunities that
will influence the shape our future.
In this article, World News talks to another of Philips' key
figures, Emile Aarts, head of New Media Systems & Applications at
Philips Research. Emile Aarts task is to turn visions of the future
into everyday reality. Ambient intelligence is fast becoming one of
these realities.
What is ambient intelligence?
Emile Aarts describes ambient intelligence as the integration of
technology into our environment, so that people can freely and
interactively utilize it. In concrete terms, ambient intelligence is
provided by a large number of small, intelligent devices, ´in-built´
into our surroundings. These devices have three important
characteristics: they can be personalized, they are adaptive and
they are anticipatory.
"A personalized device allows people to change its behaviour to suit
their own individual wishes, like changing the preferred TV channel
or adjusting the display colour," Aarts explains. "It requires a
conscious decision. Adaptivity means that activity is also tailored
to individual needs, but in this case behaviour is recorded and the
activity automatically adapts to your preferences. For example, your
TV would come to ´know´ your favourite programmes and automatically
record them for you." The advantages are apparent: less wasted time
for adjustment. "The anticipatory nature of these devices means that
actions can be anticipated from a profile of the user which is
created from observed behaviour," Aarts continues. "Since your
requirements can then be anticipated, the system will suggest
programmes you might want to see."
Separating function from the box
A TV set is just a ´black box´ with two signals in, video and power,
and two signals out, images and sound. But we can actually separate
the functionality from the box. Except for the screen - which could
be any flat surface - all processing can easily be done elsewhere.
It becomes a hidden ´virtual´ device that we must be able to
interact with.
How will this work? The important point about interaction with
distributed, hidden virtual devices is that it must happen in a
natural way. "The required interaction between people and machines
must become social, based on the same principle as between people.
We must make it possible to interact with systems naturally, using
gesture, smell, body language, touch and speech. It´s what people
intuitively want. Look at how they currently talk to their
computers!"
Another issue has been raised about today´s computer user interface,
which uses the mouse as the pointing device. Aarts says: "In an
ambient intelligence system, you won´t need a mouse. You will be the
pointing device of the future!"
The networked home
"The concept of electronic home networking is not in itself new; we
have been working towards it for some time. Philips Research has a
long history in the evolution of consumer electronics. Over the last
90 years, the company began with bulbs and tubes, and then developed
components, from which they moved into systems, and then into
functions. Now, we are moving into the new world of functions
separated from the boxes - as witnessed by our latest patents like
Incredible Sound and Natural Motion. The future scenario for ambient
intelligence incorporates computing functions into the walls of our
homes and the fabric of our lives.
"The networked home will consist of clusters of embedded devices
with a user interface that is an extension of our natural movements,
through touch panels, heat and weight sensors, and 'intelligent'
cameras with eyeball tracking, for example. Simply looking at
something and speaking a command will be enough to activate it. We
already have vision systems that can track eyeball movement. Simply
by looking at Chicago on a flat screen display of the world we could
find out local statistics, the weather or flight information -
whatever we wanted to know." ´Ubiquitous computing´ is a novel
architecture that supports this vision.
"These technologies are already being researched in Philips.
Networking their functions in the home will ultimately mean
interconnecting them so that the individual is freed from cumbersome
input and operation tasks, and is presented with information
associated with everyday tasks. We will be able to access the
information we need and want much more easily, naturally and quickly
- literally ´in the blink of an eye´!"
Three times more productivity - or three times more freedom?
What Aarts and the Philips Research computer science team are
proposing to achieve with ambient intelligence is different from
what most of the ICT (Information & Communication Technology) world
is trying to achieve! The usual aim is increased productivity. Three
times what we now achieve is generally believed to be possible. But
as Aarts sees it, that also means that we could do the same as we do
now in a third of the time - and since the trend in (western)
society is to trade-in working hours in favour of more free time for
experiences - this fits in with our vision of ambient intelligence.
The ´experience economy´ business model
The idea behind the ´experience economy´ is simple. Essentially, it
works like this. Trading started with ´commodities´ - say coffee
beans. Someone realized that coffee beans could be roasted and
packaged - that we could add value and create ´products´. Of course,
the sales price per unit then goes up. The next step was to provide
´services´: pour hot water on ground coffee and serve it - and again
prices per unit rise. The next step is to provide ´experiences´ -
serve the coffee in an exotic ambience at an even higher price.
According to Aarts, it is obviously true that experiences are
marketable: just look at how much people spend on holidays, for the
sake of experience. "With ambient intelligence, we can provide
experiences in the home. With windows as flat screens as well as
transparency, you can have a view of the beach, even from the
twentieth floor of a New York apartment block!"
Experiences are always real, even if generated by something virtual.
It's similar to how we use personal recollection. We take
photographs, for instance, to share with others later and recall the
events. But we still need to make relevant associations. We could
have large storage devices for all our electronic photos and a smart
system that is capable of displaying them, but because the material
is in the system, not physically available, there's still a need for
a trigger.
This is where ambient intelligence could provide an interactive
´experience´ that goes much further than visual stimulation. Aarts
envisages a souvenir with a tag in it, for example, that links to
images and information related to holiday experience. The souvenir
placed in front of the system immediately recalls all the related
pictures and information, perhaps local food recipes or current news
events. Images of people you knew would trigger suggestions from the
system of other events and pictures in which they appear. Equally,
homes equipped with ambient intelligence could automatically collect
images of visitors and associate them with previously recorded
visits. Your system would ´remind´ you of events when the same
person was present. Your associations are quickly and easily
retrieved.
A shift in TV
A current shift in broadcasting demonstrates how interested people
are in participative experience. An example is the Dutch reality TV
programme 'Big Brother'. There is no script, no story to be told.
People simply interact with each other - and millions watch, simply
for the experience. Many people are starting to make and broadcast
their own material on the Internet, so that they can see and be seen
in 'connected communities'. In TV, technology already enables us to
compose our own programmes from different video streams. We can
observe a game from different vantage points in a sports stadium, or
choose specific tennis players to follow, rather than watching games
selected for ´mainstream´ broadcasting. We can go back to re-view
something - effectively going back in time. Ambient intelligence
systems go even further. They 'know' the viewer's profile (their 'D'ego'
or digital ego) and make an automatic selection. The players, the
viewing angle, the replays you want are presented, without you
having to select them manually each time.
Interoperability and transparency
One of the most important research and development projects for
ambient intelligence is the interoperability of a communications
network. One example of interoperability is the IEEE 1394- based
HAVi (Home Audio Video interoperability) standard that has been
developed by a consortium of consumer electronics manufacturers,
including Philips. The standard ensures that a new TV set with this
standard, for example, would be added to your system and
´intelligently´ incorporated into the system, without the necessity
for additional programming or resetting. Devices on a HAVi network
can even share functionality.
Ambient intelligence is also characterized by its transparency.
Transparency means that the interconnected embedded systems are
invisible, because they are actually incorporated into objects, or
´painted´ onto them. Examples of use are T-shirts that respond to
body heat and give feedback as to heart- and pulse-rate for athletes
or keep-fit enthusiasts; or alarm activators embedded in clothing
that alert security systems in case of aggressive attack. The
technology underlying such inventions is ´silicon on anything´, in
which programmable silicon is applied directly to the existing
environment, rather than contained in a separate ´black box´.
Virtual devices, indeed.
We are the missing link!
Virtual devices use virtual software. So how does this affect our
audio and video ´collections´ such as cassettes or CDs? We already
have the memory technology for ´virtual´ rather than physical
collections, with hundreds of hours of visual and audible
information (programmes, software, or music) stored in the system.
However, this introduces interesting social/behavioural
considerations. We're accustomed to keeping collections on shelves.
Will we not want to actually have things ´in hand´? And will we be
able to trust ambient intelligence to store all the things we want?
Aarts believes we will overcome these natural inhibitions over time,
pointing to how we handle money as an example. 250 years ago, people
stored their wealth at home, in gold. Then came banks, where they
could store this visible, tangible wealth in a vault. This was
replaced with paper money, representations of an amount somewhere,
in a bank. Now we use digital transactions with plastic cards; we no
longer need to touch money. It´s a question of trust. When it comes
to accessing our collections, the simple solution could be a virtual
collection with visual representations of the music or films or
whatever.
Do we want to go this far?
Aarts points out that the technology is already available to turn
ambient intelligence into reality. But he also emphasizes that there
are ethical issues that need to be addressed. We should first think
whether we would like to have real ambient intelligence systems.
There are obvious matters like protection of privacy. But there are
also considerations like will the ambient system ´learn´ things
about you that may differ from how you envisage yourself? In a way,
it could be seen as a threat. If the electronics are invisible and
anticipate your behaviour, would they take over tasks you prefer to
do yourself, without first asking for it? Do we want that?
Clearly, ambient technology doesn´t only embody behavioural changes
but it is also brings ethical responsibilities. Just like all
technological breakthroughs, ambient intelligence must be used
within the boundaries of social expectations. How far we go must be
the decision of society as a whole. The responsibility of computer
scientists like Emile Aarts and his team is ensuring that consumers´
needs and wishes are turned into a reality we can all value and
enjoy.