Smart Cellular Bricks for Decentralized Shape Classification and Damage Recovery

A system of hundreds of physical 3D bricks capable of inferring global shape class and detecting structural damage through local interactions. The dataset, authored by Rodrigo Moreno and last updated in May 2026, leverages a Neural Cellular Automata algorithm for fully-distributed processing. It demonstrates strong robustness to shape variations and high tolerance to communication faults and failed modules.

Use Cases

• Training decentralized shape recognition models based on local sensing and communication.
• Benchmarking fault tolerance in modular robotic systems based on the described communication fault tolerance.
• Developing damage localization algorithms for physical structures based on the framework for detecting missing components.
• Studying bio-inspired self-recognition and morphological regeneration in engineered systems.

Strengths

• Demonstrates capabilities on collections of hundreds of physical 3D bricks.
• Shows strong robustness to out-of-distribution shape variations and high tolerance to communication faults.
• Extends the same decentralized framework to both shape inference and damage detection.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

figshare

Collection Method

Likely contains experimental results from a physical realization of a decentralized system.

Freshness

Last updated 2026-05-25 11:03:06; freshness should be verified.