Page 122 - Robots Androids and Animatrons : 12 Incredible Projects You Can Build

Intelligence 6

INTELLIGENCE PACKAGED IN A ROBOT TAKES ONE OF TWO
forms: rule-based (expert) or neural. It’s possible for both forms of
intelligence to work in tandem. This synthesis of intelligence will be
commonly used in robotics to create a robust intelligence system.

Expert (rule-based) intelligence programs are familiar to most
people; these are programs written in high-level or low-level lan-
guages like C , BASIC, and assembly. Neural systems on the other 101
hand use electronic neurons and feedback to control (generate
behavior of) the robot. This neural behavior-based robotic archi-
tecture was pioneered in the late 1940s and early 1950s by William
Grey Walter. More recently, Rodney Brooks at the Massachusetts
Institute of Technology (MIT) has been developing behavior-based
robotic architecture under the name of “subsumption architec-
ture.” We will look at behavior-based robotics in Chap. 8.
In this chapter we will focus on rule-based systems and microcon-
trollers. Keep in mind that it is possible to mimic neural systems
using rule-based systems programming. It is also noteworthy to
know that almost all neural network software on the market today
runs on existing rule-based computers, using rule-based program-
ming that simulates neural networks.

Microchip’s PIC microcontroller

Adding intelligence in the form of a computer to a small robot or
robotic system has never been easier. There are numerous single-
chip computers (commonly know as microcontrollers) available
that can do the job.

Team LRN Intelligence

Copyright 2002, 1998 The McGraw-Hill Companies, Inc. Click Here for Terms of Use.