Kinematic Self-Replicating Machines
© 2004 Robert A. Freitas Jr. and Ralph C. Merkle. All Rights Reserved.
Robert A. Freitas Jr., Ralph C. Merkle, Kinematic Self-Replicating Machines, Landes Bioscience, Georgetown, TX, 2004.
3.25.2 Chirikjian Self-Replicating Lunar Factories (2003-2004)
As a continuation of the work described in Section 3.24, Gregory Chirikjian at Johns Hopkins University was awarded a Phase I NIAC contract to study “Architecture for Unmanned Self-Replicating Lunar Factories” during 1 October 2003 through 31 March 2004 .
According to the study abstract: “The goal of this proposal is to analyze the feasibility of a fully automated robotic factory system for the development of lunar resources, and the transportation of those resources to low-Earth orbit. The key issue that will determine the feasibility of this approach is whether or not an autonomous robotic factory can be devised such that it is small enough to be transported to the moon, yet complete in its ability to self-replicate with no other inputs than what is available on the lunar surface. Self-replication leads to exponential growth, and would allow as few as one initial factory to spawn lunar production of materials and energy on a massive scale. Such capacity would dramatically impact man’s ability to explore and colonize space, as well as to deliver hydrogen and oxygen to fuel the fledgling industries that will develop in low-Earth orbit over the next few decades.
“Our architecture for a self-replicating robotic factory system consisted of five subsystems: (1) multi-functional robots for digging and transportation of materials, and assembly of components during the replication process; (2) materials refining plant; (3) parts manufacturing facility; (4) solar energy conversion, storage and transmission; (5) electromagnetic guns for long-distance transportation (e.g., for sending materials to low-Earth orbit, or transporting replicated factories to distal points on the moon). We envision that a fully functional lunar factory site will occupy approximately one square kilometer. However, the precursor that is launched from the earth will be a minimalist system consisting only of two robots, a small furnace, molds, mirrors and solar panels and weighing between five and ten metric tons. The full self-replicating robotic factory will be constructed under remote control from the earth using the precursor system.”
After the Phase I study was completed in March 2004, the Final Report  concluded: “The key issue to determine the feasibility of this approach in the NIAC time frame was whether or not complementary technologies expected over the next 10 to 40 years would exist for an autonomous robotic factory to function. In particular, it was not clear a priori whether such a system could be devised such that it would be small enough to be transported to the moon, yet complete in its ability to self-replicate with little input other than what is available on the lunar surface. Minimalist systems which can be launched at low cost, harvest lunar resources, and bootstrap up to a substantial production capability are appealing. Self-replication leads to exponential growth, and would allow as few as one initial factory to spawn lunar production of materials and energy on a massive scale. Such capacity would dramatically impact man’s ability to explore and colonize space, as well as to deliver hydrogen and oxygen to fuel interplanetary spacecraft and the fledgling industries that will develop in space over the next few decades. This report has shown using a combination of prototype implementations, analysis, and literature survey that self-replicating lunar factories do in fact appear to be feasible. A road map for the novel recombination and integration of existing technologies and systems into an architecture that can be implemented in the next ten to forty years is provided.
“Many technological hurtles must be overcome before self-replicating robots can become a reality, and current knowledge from many diverse disciplines must be recombined in new ways. In this Phase I feasibility study we examined what lunar resources can be exploited, and investigated “toy” designs for robots with the ability to self-replicate. To this end, we examined how each subsystem of a robotic factory (motors, electronic components, structural elements, etc.) can be constructed from lunar materials, and demonstrated these ideas in hardware. We did this at several levels. For example, robots that assemble exact functional copies of themselves from pre-assembled subsystems were demonstrated. The feasibility of assembling an actuator from castable shapes of structural material and molten metal was demonstrated with proxy materials. The assembly of simple self-replicating computers made of individual logic gates was demonstrated. A strategy was developed for how the lunar regolith can be separated into ferrous, nonferrous conducting, and insulating materials in the absence of water was studied and partially demonstrated. In addition, the energy resources available at the lunar surface were evaluated, means for using this energy were developed, and the energetic requirements of various subsystems were computed. Our conclusion is that the proposed system architecture indeed appears to be feasible provided certain existing technologies can be integrated in new ways.”
The control systems for the mobile lunar robots are to be constructed using electromechanical relays and vacuum tubes, providing simple (though bulky) electronic logic circuits that are readily fabricated using lunar materials. As a preliminary demonstration, Chirikjian’s group has built several exemplar “self-replicating circuits” using transistors embedded in LEGO® blocks [2339-2341]. In one case , the physical implementation consisted of a robotic system and a conveyor belt system for replicating its code and circuit, plus the initial code and the circuit to be replicated. During replication, “a conveyor belt feeds the code one line at a time to the reader array. The array then sends the code signal to the control circuit. The signal is decoded as any of seven robot arm or position movements. The robot then carries out the movement command. The commands would tell it to go to one of three feeder positions, pick up the items at that position, move to the assembly line, and drop the items into place. The next set of codes is then fed to the readers, and the process would repeat.”
Last updated on 1 August 2005