Self-Organized Construction by Minimal Surprise

TL;DR

This paper explores how robots driven by a minimal surprise principle can autonomously organize and construct environments by making their world more predictable, demonstrated through simulated block-pushing tasks.

Contribution

It introduces a minimal surprise-based approach for collective construction, enabling robots to self-organize and adapt behaviors for environment structuring.

Findings

Robots successfully organized blocks to create predictable environments.

Emergent behaviors were observed in the collective construction process.

Controllers evolved for collective construction can be applied in real robot systems.

Abstract

For the robots to achieve a desired behavior, we can program them directly, train them, or give them an innate driver that makes the robots themselves desire the targeted behavior. With the minimal surprise approach, we implant in our robots the desire to make their world predictable. Here, we apply minimal surprise to collective construction. Simulated robots push blocks in a 2D torus grid world. In two variants of our experiment we either allow for emergent behaviors or predefine the expected environment of the robots. In either way, we evolve robot behaviors that move blocks to structure their environment and make it more predictable. The resulting controllers can be applied in collective construction by robots.