Materials as robots- Toward functional mechanical computing
Abstract
There has been great interest in unconventional computing in recent years, and how physical systems can represent and process information. Numerous mechanical computing systems have been proposed that use analog or digital representations of information, as well as both static and dynamic mechanisms. However, most research related to mechanical computing has remained abstract and theoretical, focusing on how information is abstracted in a particular mechanism rather than how and when these systems would be beneficial to system function and performance. In contrast with traditional electronic computing systems, which store information as isolated bits (transistors), in mechanical computing, the information state is generally represented by deformation. As a result, the shape and mechanical properties of the mechanical system can change significantly as information is processed. This fact has been mostly neglected in past research, which has tended to treat mechanical computers as direct analogs to electronic computers. The objective of the proposed work is to create a formalism for mapping information processing in mechanical systems to changes in their shape, properties, and behavior, and to devise the design tools and procedures necessary to demonstrate the utility of this formalism in autonomous pneumatic robots. We will use experimental and numerical approaches to design mechanical computing systems for which different information states correspond to different properties and behaviors (e.g., a robot that jumps or grasps depending on the result of a mechanical computation) rather than mere abstractions. Finally, we will create inverse design tools to enable algorithmic design of metamaterials that perform specified robotic functions via sequential deformation (e.g., translation). This work will provide a fundamental framework for mechanical computing and its application in autonomous robotics, blurring the distinction between materials and robots.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Mar 06, 2024
- Source ID
- FA95502310416
Entities
People
- Jordan R. Raney
Organizations
- Air Force Office of Scientific Research
- United States Air Force
- University of Pennsylvania