Aware Home4, MavHome5, and Gator Tech
The levels of intelligence and proactivity a Smart
Home (and an AmI system in general) should exhibit is a
trade-off between provided services and user interaction.
Because the general behavior of a Smart Home is based on
rules that are encoded by human experts, the cognitive
capabilities offered by an AmI system largely depend on the
knowledge provided by these experts and on their ability to
encode it into reasoning systems.
Consider an automated system for turning on/off the
lights. Let’s assume that a room is provided with sensors
that are read to decide if there are people inside. If sensors
and the corresponding logic are based exclusively on movement detection, someone relaxing on a couch could easily
trick the system, which would immediately turn off the
lights for absence of any movement. From the sensory
perspective, the system would misinterpret absence of
movement as absence of people. The fault lies in the logic
provided by system designers, not the actual system behavior.
One solution to the automatic lighting system dilemma
is to use different and heterogeneous sensors to obtain a
clearer picture of the situation. For instance, by adding a
number of sensors able to detect door crossings, it would
be possible to infer the current amount of people inside the
room, and to turn on/off the lights independently from
actual human behavior. On the other hand, a system
capable of detecting if someone is sleeping could purposely
decide to turn the lights off in the environment.
From a functional perspective, a typical Smart Home is
very similar to what is depicted in Figure 1. From a human
A good example of something that may be in a Smart Home —
a memory browsing system that allows users to access and
display their photos anywhere in the house.
FIGURE 2. Basic information involved in Example 2.
FIGURE 3. Basic information involved in Example 3.
perspective — with the exception of non-obtrusive sensors —
the home is virtually identical to a standard home. Typical
sensors include: PIR (passive infra-red) sensors for
movement detection; smoke, gas, and light sensors;
temperature sensors; surveillance cameras; and pressure
sensors in chairs, couches, and beds. Additional devices
include automated doors and windows, and control systems
for temperature and light regulation.
But this is now. Soon, highly integrated
intelligent home appliances will be available that
are able to autonomously provide information
about their own use. In addition, there are
wearable sensors that can monitor biometric
data, such as human posture and blood pressure.
However, wearable sensors require active
participation from the user, which is against the
initial requirements of AmI.
What Have We Done
Up To Now?
Figure 1 shows a bedroom, a bathroom, and
a living room that are divided into two areas —
namely entrance and kitchen. A number of
PIRs are placed both in selected areas and in
doorways. Furthermore, we assume the availability
of bed, seat, and tap sensors, along with a series
of intelligent appliances (such as the oven, the
fridge, the shower, and the TV).
● EXAMPLE 1: The user is in bed while the
shower tap is active. This situation can be
46 SERVO 12.2008