Page 258 - Concise Encyclopedia of Robotics

Proprioceptor
starting paths: A, B, C, and D. Two of these, B and C, lead to the desired
result; the other two do not. But even if TPM starts out along B or C,
there are many possible dead ends.
In this example, there is a crossover between paths B and C. One of the
sidetracks from path B can lead to the desired result indirectly, by finishing
through path C. Also, a sidetrack from path C can take TPM to the proof
by moving over to path B. However, these crossings-over can also lead
TPM back toward the starting point, and possibly even to dead ends on
the way back there.
Dead ends
When TPM runs into a dead end, it can stop, turn around, and backtrack.
But how can TPM know that it has come to a dead end? It might keep
trying repeatedly to break through the barrier without success. As you
know from real-life experience, sometimes persistence can get you over a
difficult hurdle, and in other cases all your effort cannot break through the
barrier.After you try for a long time to get out of a dead end, you will give
up from exasperation and turn back. At what point should TPM give up?
The answer to this quandary lies in the ability of TPM to learn from
experience. This is one of the most advanced concepts in AI.
A true TPM that can always solve proofs of true propositions will never,
and in fact can never, be constructed. This is because there are statements
in any logical system that cannot be proven true or false in a finite number
of steps. This was proven by logician Kurt Gödel in 1930, and is called the
incompleteness theorem.
See also INCOMPLETENESS THEOREM.
PROPRIOCEPTOR
If you close your eyes and move your arms around, you can always tell
where your arms are. You know if your arms are raised or whether they
are hanging at your sides. You know how much your elbows are bent, the
way your wrists are turned, and whether your hands are open or closed.
You know which of your fingers are bent and which ones are straight.You
know these things because of the nerves in your arms, and the ability of
your brain to interpret the signals the nerves send out.
There are advantages in a robot having some of this same sense, so that
it can determine, and act according to, its positioning relative to itself. A
proprioceptor is a system of sensors that allows this.
See also the following definitions: COMPUTER MAP, DIRECTION FINDING, DIRECTION
RESOLUTION, DISPLACEMENT TRANSDUCER, DISTANCE MEASUREMENT, DISTANCE RESOLUTION, EDGE
DETECTION, EPIPOLAR NAVIGATION, EYE-IN-HAND SYSTEM, GUIDANCE SYSTEM, LANDMARK, LOCAL
FEATURE FOCUS, ODOMETRY, PARALLAX, PHOTOELECTRIC PROXIMITY SENSOR, PROXIMITY SENSING,
SONAR, and VISION SYSTEM.

