An autonomous mobile robot closes the
loop between perception and action:it is capable of perceiving its environment
through its sensors, processing that information using its on-board computer,
and responding to it through movement.
This raises some interesting
questions, for example the question of how to achieve "intelligent” behavior.
What are the foundations of task-achieving behaviors; by what mechanism can
behaviors be achieved that appear "intelligent”to the human observer? Second, there
is a clear parallel between a robot’s interaction with the environment and that
of animals. Can we copy animal behavior to make robots more successful? Can we
throw light on the mechanisms governing animal behavior, using robots?
Such questions concerning behavior,
traditionally the domain of psychologists, ethnologists and biologists, we
refer to as "science”. They are not questions of hardware and software design, i.e. questions that
concern the robot itself,but questions that use the mobile as a tool to investigate
other questions. Suchuse of mobile robots is continuously increasing, and a
wide body of literature exists in this area, ranging from "abstract” discussions
of autonomous agents([Braitenberg, 1987, Steels, 1995, von Randow, 1997, Ritter
et al., 2000]) to the application of Artificial Intelligence and Cognitive
Science to robotics
([Kurz, 1994, Arkin, 1998, Murphy,
2000]
[Dudek and Jenkin, 2000]). Journals
such as Adaptive Behavior or IEEE Transactions
on Systems, Man, and Cybernetics also address
issues relevant to this
topic.
1.2.3 (Commercial) Applications
Mobile robots have fundamental
strengths, which make them an attractive option for many commercial
applications, including transportation, inspection, surveillance,health care
[Katevas, 2001], remote handling, and specialist applications like operation in
hazardous environments, entertainment robots ("artificial pets”)or even museum
tour guides [Burgard et al., 1998].
Like any robot, mobile or fixed,
mobile robots can operate under hostile conditions,continuously, without
fatigue. This allows operation under radiation, extreme temperatures, toxic
gases, extreme pressures or other hazards. Because of their capability to
operate without interruption, 24 h of every day of the week,even very high investments
can be recovered relatively quickly, and a robot’s ability to operate without
fatigue reduces the risk of errors.
In addition to these strengths, which
all robots share, mobile robots have the additional advantage of being able to
position themselves. They can therefore attain an optimal working location for
the task at hand, and change that position during operation if required (this is
relevant, for instance, for the assembly
Of large structures). Because they
can carry a payload, they are extremely flexible:
mobile robots, combined with an
on-board manipulator arm can carry a
range of tools and change them on
site, depending on job requirements. They
can carry measurement instruments and
apply them at specific locations as required
(for example measuring temperature,
pressure, humidity etc. at a precisely
defined location). This is exploited,
for instance, in space exploration
[Iagnemma and Dubowsky, 2004].
Furthermore, cooperative mobile robot
systems can achieve tasks that are
not attainable by one machine alone,
for example tasks that require holding an
item in place for welding, laying
cables or pipework, etc. Cooperative robotics is
therefore a thriving field of
research. [Beni and Wang, 1989, Ueyama et al., 1992]
[Kube and Zhang, 1992, Arkin and
Hobbs, 1992, Mataric, 1994] and
[Parker, 1994] are examples of
research in this area.
There are also some weaknesses unique
to mobile robots, which may affect
their use in industrial application.
First, a mobile robot’s distinct
advantage of being able to position itself introduces
the weakness of reduced precision.
Although both manipulators and
mobile robots are subject to sensor
and actuator noise, a mobile robot’s position
is not as precisely defined as it is
in a manipulator that is fixed to a permanent
location, due to the additional
imprecision introduced by the robot’s chassis
movement. Furthermore, any drive
system has a certain amount of play, which
affects the theoretical limits of
precision.
Second, there is an element of
unpredictability in mobile robots, particularly
if they are autonomous, by which is
meant the ability to operate without external
links (such as power or control).
With our current knowledge of the process of
robot-environment interaction it is
not possible to determine stability limits and
behavior under extreme conditions
analytically. One of the aims of this book
is to develop a theory of
robot-environment interaction, which would allow a
theoretical analysis of the robot’s
operation, for example regarding stability and
behavior under extreme conditions.
Third, the payload of any mobile
robot is limited, which has consequences for
on-board power supplies and operation
times. The highest energy density is currently
achieved with internal combustion
engines, which cannot be used in many
application scenarios, for example
indoors. The alternative, electric actuation,
is dependent on either external power
supplies, which counteract the inherent
advantages of mobility because they
restrict the robot’s range, or on-board batteries, which currently are very
heavy. As technology progresses, however, this
disadvantage will become less and
less pronounced.
1.3 The Emergence of Behaviors
Why is it that a mobile robot,
programmed in a certain way and placed in some environment to execute that
program, behaves in the way it does? Why does it follow exactly the trajectory
it is following, and not another?
The behavior of a mobile robot—what
is observed when the robot interacts with its environment — is not the result of
the robot’s programming alone, but results from the makeup of three fundamental
components:
1. The program running on the robot
(the "task”)
2. The physical makeup of the robot
(the way its sensors and motors work,battery charge, etc.)
3. The environment itself (how
visible objects are to the robot’s sensors, how
good the wheel grip is, etch)
The robot’s behavior emerges from the
interaction between these three fundamental