To demonstrate robust construction and adaptive reconfiguration of physical systems from elementary components, under uncertainty and variability of material resources.
This project envisions a manufacturing process where the relationship between the target structure and the source materials is defined indirectly, and where the path between them is subject to fluctuations requiring strategic decisions. Consider the challenge of constructing a structure that needs to support a specified load, given a pile of components of a known statistical distribution. While the target structure is being constructed, the distribution of raw components may change, requiring a change in strategy. This may in turn involve the use of different components, the redesign of the target structure, or even the disassembly of existing yet less critical structures. The project addresses these challenges in the following aims:
- Aim 1: Develop a theoretical foundation for analyzing the governing dynamics of stochastic factories.
- Aim 2: Develop design tools, analytical approaches, and reconfiguration algorithms for stochastic environments.
- Aim 3: Implement and demonstrate the algorithms and theoretical principles on a physical testbed.
A stochastic factory example in which two robots, one wheeled and one climbing, are in the process of reconfiguring two structures in a stochastic environment.
As an example, the video below shows a stochastic reconfiguration task in which one type of tower is disassembled on one end of the factory floor and a different type of tower is assembled on the other. Each factory floor module (white square) checks local conditions on its environment and probabilistically moves raw materials, nodes (turquoise) and trusses (red), in response to these conditions. Yellow indicates that the lifting mechanism of a module is up (Courtesy: Klavins Research Group).