Background - Smartphone-Based Human Activity Recognition - page 9

the use of ML techniques that take past experience to forecast the required actions
for similar situations in the future.

• They allow seamless human-machine interaction . They are provided with mech-

anisms to naturally communicate with its users (e.g. by interpreting body motion,
gestures, location and physiological signals). Ideally, without the user having to
interact through custom-made devices.

• They are adaptive : able to learn and respond to new or unusual situations. This

characteristic allows systems to become more robust and able to react against
unexpected conditions from the users and the environment.

2.2.2 AmI Applications

AmI has been already employed in a wide range of applications. Below, three sig-
nificant areas where it has been already applied are described:

• Smart Homes . Home automation is changing the way households and their res-

idents interact. Novel households are arranged with multimodal technologies
involving environmental sensors, user interfaces, computing devices and actua-
tors; which aim to guarantee supervision and fast response to people in order to
improve their QoL. Moreover, they are incorporating centralized controls for most
of their basic services such as heating, lighting, security, entertainment, and, in
general, all the available electronic appliances (García-Vázquez et al. 2010 ;José
et al. 2010 ; De Silva et al. 2012 ).

• Vital Signs Monitoring for Healthcare . Wearable body sensors (Yang and Yacoub

2006 ) offer an alternative for monitoring people's physiological and vital signs.
These include, for example, measurements of heart rate, blood pressure, oxygen
saturation, bodymotion, skin temperature and conductivity. AmI systems are capa-
ble of processing these data to supervise the health condition of a target group (e.g.
the elderly). Moreover, this information can be communicated and distributed in
real-time to interested parties such as family members and emergency services
when urgent assistance is required.

• Unifying heterogeneous operation interfaces . It is nowadays common to see envi-

ronments (e.g. living rooms, offices, conference rooms) composed of computing
resources and electronic equipment for all kinds of purposes. These include: TVs,
beamers, sound systems, lighting control systems, personal computers and portable
devices such as tablets and smartphones. Even though they share the same space,
their control mechanisms and interfaces are rather different and there is gener-
ally no unique approach to operate them. AmI contributes to unify the way these
devices interact with people by creating more natural and simpler approaches to
interface them. It also allows the automatic incorporation of new devices in a more
intuitive way.