Computing Reviews
Today's Issue Hot Topics Search Browse Recommended My Account Log In
Review Help
Search
A general architecture for robotics systems: a perception-based approach to artificial life
Young R. Artificial Life23 (2):236-286,2017.Type:Article
Date Reviewed: Apr 10 2018

“Conventional views of intelligence and behavior are fundamentally flawed.” That is the inflection point that motivates this paper. Most robotic systems depend on a digitized model of their environment that they can use to reason about that environment. The model is never complete enough, and it’s always bounded by resource limits like storage or processor speed. There’s always more information that should be included. The author of this paper argues that that approach is unnecessary, that it’s not necessary to understand the universe in order to navigate through it.

Young embraces perceptual control theory (PCT) as the underlying behavioral principle of artificial life and perceptual control as the central process supporting all behavioral systems. Take the example of catching a ball. Traditional robotics attempts to model a ball moving through the air, and to understand the underlying gravity and physics in order to catch it. PCT is vastly simplified: perceive the position of the ball, and then move left or right, forward or backward, to adjust perception. At an abstract level, the robot changes its outputs (movement) to improve its input (perception of the ball) rather than attempting to generate specific outputs from environmental input. As the author states, this is the beginning of a whole new line of behavioral research. Young argues that layers of perception control circuits can explain all intelligent behavior.

The paper itself is longer than most at 51 pages. However, given that we’re being persuaded to believe a new theory of intelligence, the length is probably about right. The paper is divided into 11 sections of text, citation references, and an appendix that explains his experiment configuration using Lego NXT robotics components. The sections are:

1. Introduction

2. Perceptual Control Theory Overview

3. A Brief History of Conventional Approaches

4. Perception-Based Robotics

5. Key Attributes and Benefits

6. The System (explanation of Lego NXT configuration)

7. Behaviors and Perceptions

8. The Model

9. High-Level Perceptions

10. Results

11. Discussion and Conclusions

It’s always rewarding to challenge your understanding of the universe. Obviously this paper is meant for, and valuable to, roboticists, but I think any scientist, engineer, or even philosopher would enjoy thinking through this presentation of behavioral theory. It’s written in a very clear and understandable manner, and no particular background is needed to appreciate the major points.

Reviewer:  Bayard Kohlhepp Review #: CR145963 (1806-0339)
Bookmark and Share
  Featured Reviewer  
 
Robotics (I.2.9 )
 
Would you recommend this review?
yes
no
Other reviews under "Robotics": Date
Movement problems for 2-dimensional linkages
Hopcroft J. (ed), Joseph D., Whitesides S. SIAM Journal on Computing 13(3): 610-629, 1984. Type: Article
Feb 1 1985
Robot motion planning with uncertainty in control and sensing
Latombe J. (ed), Lazanas A., Shekhar S. Artificial Intelligence 52(1): 1-47, 1991. Type: Article
Oct 1 1992
Dictionary of robot technology in four languages: English, German, French, Russian
Bürger E., Korzak G., Elsevier North-Holland, Inc., New York, NY, 1986. Type: Book (9789780444995193)
Mar 1 1988
more...

E-Mail This Printer-Friendly
Send Your Comments
Contact Us
Reproduction in whole or in part without permission is prohibited.   Copyright 1999-2024 ThinkLoud®
Terms of Use
| Privacy Policy