Page 16 - Innovations in Intelligent Machines

Intelligent Machines: An Introduction 3
learning, researchers have proposed a multi-learning method that makes use
of more than one learning techniques [3]. Besides, different aspects of research
in robotics have been conducted, which include robot mobility and control [7],
robot perception [8], as well as the use of soft computing techniques for intelli-
gent robotic systems [9]. On the other hand, the divide and conquer principle
is applied to the learning tasks [10]. Each algorithm is given a specific task
to handle. The learning algorithms are chosen carefully after considering the
characteristics of the specific task. Another potential solution to learning is
intelligent agents. Agents collect data and learn about the surrounding envi-
ronment, and adapt to it [11]. The learning process in agents also requires
a self-organizing mechanism to control a society of autonomous agents [12].
It should be noted that the task of imparting learning into intelligent machines
is not an easy one; however the learning capability is what makes a machine
intelligent.

3 Application of Intelligent Machines

The applications of intelligent machines are widespread nowadays, extending,
for example, from Mars rover invented by NASA to intelligent vacuum cleaners
found in our homes. Some examples of intelligent machines are as follows.

3.1 Unmanned Aerial Vehicle (UAV)

There are some aerial missions and tasks that are not suitable for human pilots
either because it is too dangerous like military operations, or it takes a long
time in the air like mapping tasks. Yet, these tasks are important. UAVs have
been invented to carry out such mission-critical tasks [13]. Typically, an UAV
comprises onboard processing capabilities, vision, GPS (Global Positioning
System) navigation, and wireless communication. One of the main functions
of an UAV is to navigate in an uncontrolled environment, which also is often an
unknown environment, safely, and, at the same time, to perform its required
task [14]. What makes an UAV intelligent is the ability to fly to its target
under varying conditions. As it is not possible to predict all possible navigation
scenarios in one program, the UAV has to learn from its environment, and
adapt to the changes as they occur in order to reach the destination.
An UAV used to collect data in the atmosphere between satellite and the
ground base is created by National Oceanic and Atmospheric Administration
(NOAA), USA. The UAV is able to fill the gap where land-based and satellite-
based observations fall short, thus giving a view of the planet never seen before
[15]. Another UAV, a version of the military MQ9 Predator B, is used by the
Department of Homeland Security, USA to monitor remote and inaccessible
regions of the border. The UAV is equipped with special cameras and other
sensors, and is able to stay in the air for up to 30 hours [16].