chirikjian

Robotic Self-Replication in Structured and Adaptable Environments: 

Robotic Self-Replication in Structured and Adaptable Environments Greg Chirikjian, Kiju Lee, Matt Moses Dept. of Mechanical Engineering Johns Hopkins University

Contents: 

Contents Our First Self-Replicating Robots in Structured Environments Self-Replicating Robots in Semi-Structured Environments Self-Assembling Computers Self-Replicating Robots with Discrete Electronics Self-Replicating Robots that Structure Their Environment Future Directions Develop Principles of Robotic Entropy/Disorder Robotic Self-Replication by Mitosis

A Remote Controlled Design: 

A Remote Controlled Design

A Fully Autonomous SRR: 

A Fully Autonomous SRR

Robotic Self-Repair in a Semi-Structured Environment: 

Robotic Self-Repair in a Semi-Structured Environment

Robotic Self-Repair in a Semi-Structured Environment: 

Robotic Self-Repair in a Semi-Structured Environment

Computers that Copy Themselves: 

Computers that Copy Themselves To make computers from logic elements using a von Neumann Universal Constructor would require a very complicated code. An alternative is Laing’s paradigm of self-replication by self-inspection. We developed three prototypes to demonstrate this idea physically for the first time.

Self-Replication by Self-Inspection 1: 

Self-Replication by Self-Inspection 1

Self-Replication by Self-Inspection 2: 

Self-Replication by Self-Inspection 2

Self-Replication by Self-Inspection 3: 

Self-Replication by Self-Inspection 3

Combining Previous Concepts: : 

Combining Previous Concepts: Self-Replicating Robots with Distributed Computing Elements

Slide12: 

SRR with Distributed Circuits

Slide13: 

SRR with Distributed Circuits

SRR in Adaptable Environment: 

SRR in Adaptable Environment

Slide15: 

SRR in Adaptable Environment

Remainder of the Talk: 

Remainder of the Talk Principles of Robotic Self-Replication (RSR) Universality, High Relative Complexity, Robustness, and Efficiency Entropy/Information

Principles: 

Principles Universality What can it become? What can it produce? What functions can it perform? High Relative Complexity Complexity of the Robot vs. Complexity of the Individual Parts Robustness Error tolerance in sensing and manipulation Efficiency How efficiently can it reproduce?

Entropy: 

Entropy Entropy is: A potential useful tool to describe the sophistication of tasks in robotic self-replication; A measure useful when the environment can be defined by any element of an ensemble of different configurations; A property of the collection of all environmental objects over all of their possible arrangements.

Entropy: 

Entropy

Pose Entropy: 

Pose Entropy Assuming that each body can move independently: An estimated overlap of rigid bodies i and j : Then we have,

Future: A Self-Replicating Lunar Factory System: 

Future: A Self-Replicating Lunar Factory System

Selected References: 

Selected References Suthakorn, J., Andrew B. Cushing, and Chirikjian, G.S. ``An Autonomous Self-Replicating Robotic System,’’ Proceedings of the 2003 IEEE/ASME International Conference on Advanced Intelligent Mechatronics, 2003 Chirikjian, Zhou, Y. and G.S, Suthakorn, ``Self-Replicating Robots for Lunar Development,’’ ASME & IEEE Transactions on Mechatronics Vol. 7, Issue: 4, Dec 2002. For more: http://custer.me.jhu.edu/publication/self_replicating.html