Information Requirements for Self-Reproducing Systems in Lunar Robotic Colonies

Paper number

IAC-06-A5.P.04

Author

Mr. Amor Menezes, University of Michigan, United States

Coauthor

Dr. Pierre Kabamba, University of Michigan, United States

Year

2006

Abstract

This paper is devoted to the use of self-reproducing systems in the International Lunar Robotic Village, and examines one of the feasibility requirements for such systems.  Robots endowed with the capacity for self-reproduction will be able to use available resources on-site to enlarge their numbers if it is deemed necessary for a given task.  Such technology will not be dependent on either launch capabilities or the fiscal constraints surrounding the multiple launches of robots required for the colony.  While colonies of robots have been extensively studied in the past, they have not been examined as a population capable of expansion.  The CREATORS Project at the Aerospace Engineering Department of the University of Michigan has been investigating a Colony of Robotic Entities for Autonomous Technologies Of Robust Self-reproduction, and this paper will highlight some preliminary results.

The technical approach employed in this work is to utilize concepts from a novel theory called Generation Theory, and, with the aid of Claude Shannon’s Mathematical Theory of Communication (Information Theory), determine the information required to specify an entity that can self-reproduce.  In a series of lectures in the 1940s, John von Neumann first postulated the existence of a threshold of complexity below which any attempt at self-reproduction was doomed to degeneracy.  However, he did not compute its value, and no one published an evaluation of this threshold in the following 60 years.  One of the original contributions contained in this work is the explicit computation of the von Neumann threshold.  The accompanying conclusion that a machine must be specifically designed if self-reproduction is to occur is also derived.

In general, robots require mass, energy, and information or knowledge in order to perform their assigned tasks.  While a seed robot armed with some start-up resources is needed to initiate a colony, it is not feasible to deploy the robot with all the mass and energy necessary for self-reproduction, nor is it expected that the first seed robot arrives with all the knowledge it requires for survival, since it can learn from its actions within the environment.  Accordingly, there are bounds on the minimum requirements of information, mass, and energy for the seed of a self-reproducing system.  This work focuses solely on the information that is a fundamental requirement for the seed.  Indeed, this minimum information requirement is the aforementioned von Neumann threshold, which can now be determined.

Abstract document

IAC-06-A5.P.04.pdf

Manuscript document

IAC-06-A5.P.04.pdf (🔒 authorized access only).

To get the manuscript, please contact IAF Secretariat.