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.
3.22 Brooks Living Machines Program (2001-present)
Rodney Brooks is both the Director of the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL) and the Chairman and CTO of iRobot, a 120-person robotics company. Well known for his pioneering work [1264-1269] in making small, simple workable robots using the subsumption architecture (e.g., Attila [1270, 1271] and Genghis [1271, 1272]) and later for his investigations of basic learning processes in humanoid robots (e.g., Cog [1268, 1273] and Kismet [1268, 1274]), in 2001 Brooks began an entirely new research program [1275] called “Living Machines” – moving away from building humanoid robots to considering instead the differences between living matter and non-living matter [1276]. “At one level we’re trying to build robots that have properties of living systems that robots haven’t had before,” he explained in a June 2002 interview [1277]. “We’re trying to build robots that can repair themselves, that can reproduce (although we’re a long way from self-reproduction), that have metabolism, and that have to go out and seek energy to maintain themselves.”
Brooks is very interested in replicating systems [1277]: “We’re also trying to build self-reproducing robots. We’ve been doing experiments with Fischer Technik and LEGO®. We’re trying to build a robot out of LEGO® which can put together a copy of itself with LEGO® pieces. Obviously you need motors and some little computational units, but the big question is to determine what the fixed points in mechanical space are to create objects that can manipulate components of themselves and construct themselves. There is a deep mathematical question to get at there, and for now we’re using these off-the-shelf technologies to explore that. Ultimately we expect we’re going to get to some other generalized set of components which have lots and lots of ways of cooperatively being put together, and hope that we can get them to be able to manipulate themselves. You can do this computationally in simulation very easily, but in the real world the mechanical properties matter. What is that self-reflective point of mechanical systems? Biomolecules as a system have gotten together and are able to do that.” The specific projects being studied in Brooks’ lab are in a constant state of flux, but the following are a few examples of some efforts that were being pursued in early 2003.
One of Brooks’ graduate students, Jessica Banks [1278], was directly investigating the biological mechanisms of reproduction from a kinematic cellular automaton (Section 3.8) approach, using LEGO® parts as the primitive building blocks. According to her original research plan: “The immediate goal is to build a robot that joins together the same pieces out of which it is built. To simplify this problem, we chose to construct the robot out of a limited set of LEGO® parts analogous to nature’s atomic building blocks: carbon, hydrogen, oxygen, nitrogen, sulfur, and phosphorous. The robot is designed to assemble these blocks by combining minimal sensing and actuation with the passive incorporation of the environment in which it is situated. As such, we are hoping to draw an analogy between the robotic system and that of molecule structures which organize due to the energy flow of and reactions with the liquid water medium in which they are suspended. The future goal is to try to answer questions about whether it is possible for machines to beget machines. We would like a robot to autonomously assemble copies of itself, either directly or through a sequence of intermediate robotic constructions. What does reproduction mean for a robot and what is required for this process? Can we extract a fixed point for robotic self-assembly that we can apply to other inorganic and organic systems?”
Another graduate student, Lijin Aryananda [1279], was pursuing a slightly different approach: “We seek to explore the ambitious question of how to construct self-replicating non-trivial robots. In nature, we observed that evolution first generated self-replicating single-celled organisms. Multicellular organisms didn’t appear in abundance until approximately 550 million years ago, during the Cambrian explosion. Based on this observation, we propose to divide our target problem into two parts: how to construct self-replicating simple robots and how these units may aggregate to form self-replicating non-trivial robots. In this project, we plan to carry out various computational experiments to study the following issues: how do unicellular aggregates form and under what conditions are they more beneficial? Why do individuals in aggregates surrender their ability to reproduce? How does a unicellular organism’s self-reproducing mechanism evolve to the self-reproducing mechanism in multicelled organisms? How does differentiation emerge in multicellular organisms? Our hope is that the answers to these questions can be ultimately applied in designing complex robots that self-replicate.”
A postdoc in Brooks’ lab, Martin C. Martin [1280], was pursuing an evolutionary approach to self-replicating machines more akin to the work of Lipson and Pollack (Section 3.20): “Evolved bodies will be constrained to use parts typical of machines, such as rigid cylinders, metallic plates and electric motors. Existing rigid body simulators are well suited to this task. The world will be much richer than existing work, containing areas of water, land and air, as well as varied terrain in each area. Later, other variations may be added, such as day/night cycles and tides. Previous work has largely focused on neural networks as the representation for brains, but an alternate representation could lead to behaviors of a much greater complexity. Reinforcement learning will be used to allow the creatures to adapt to their environment, with the details of the learning framework under genetic control. This will allow the complexity of the robot to mirror the complexity of the world, rather than forcing that complexity to be present in the genome. It will also allow the brain to better adapt to changes in the body due to mutation. This will allow more mutations to be explored, ideally allow evolution to be more efficient. A success in this work could provide a new way to design machines of a greater complexity than is possible at the moment. Successful creatures could be reverse engineered to determine how they work. This could lead to insights into the proper method of combining development and learning, possibly providing new paradigms for traditional hand-designed machines. The work could also point the way to giving important properties of living systems to machines, and shed light on the nature of those qualities.”
Last updated on 1 August 2005