Reproduce “Morphogenetic Assembly and Adaptive Control for Heterogeneous Modular Robots” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Chongxi Meng, Da Zhao, Yifei Zhao, Minghao Zeng, Yanmin Zhou, Zhipeng Wang, Bin He

This research introduces a closed-loop automation framework for heterogeneous modular robots, encompassing both their morphogenetic construction and adaptive control. It features a hierarchical planner for dynamic assembly of diverse module types and a GPU-accelerated, morphology-agnostic controller that enables real-time locomotion for newly formed configurations.

Key insights

  • The inclusion of a type-penalization term within the high-level planning heuristics is essential for ensuring robustness and scalability when planning reconfigurations with heterogeneous module types.
  • A Greedy heuristic, while not globally optimal in abstract assignment, practically outperforms the Hungarian heuristic in generating more cost-efficient physical execution plans for the assembler robot.
  • Combining multi-stage variance annealing with GPU acceleration in the MPPI controller enables real-time (50 Hz) and highly accurate locomotion control for novel robot morphologies without specific retraining.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.