Page 124 - Designing Autonomous Mobile Robots : Inside the Mindo f an Intellegent Machine

7
CHAPTER

Basic Navigation Philosophies

Before we plunge into the subject of navigation, it is helpful to put the whole topic
into perspective. Thousands of papers have been written about robot navigation,
and many very compelling concepts have been developed. When one looks at the
actual robotic products that have been produced, however, most have used much
simpler approaches. Two schools of thought about robot navigation have emerged:
the academic school and the industrial school.

The academic school of thought
There is a recurring theme in many academic papers on robot navigation that goes
like this: “Before robots can be widely accepted, they must be capable of learning
their environments and programming themselves without assistance.” So consis-
tently is this line repeated, that one is prone to believe it must have been chiseled
on the back of the tablets that Moses brought down from the mountain. Whether it
will turn out to be a fact, however, is yet to be determined. There is evidence on
both sides of this argument.

In support of the academic argument, it is true that most potential customers (other
than the academic and hobby markets) do not want to program their robots. These
customers, quite rightly, do not want to be distracted from the primary mission focus
of their business or department. On the other hand, they do not care whether a
robot is programmed or not, as long as it is installed and does the job they expect.
The yardstick for such customers is one of return on investment (ROI). If program-
ming takes a high degree of expertise on site, then this will add to the cost of the
initial installation and certainly reduce the ROI. It is also important that the system
be readily changed when the environment changes. If changes add to operational
costs, then the ROI is reduced again.

107