Driving · 2026-10-19

CoEvolve: Self-Evolving Code Policies from Skills to Strategies for Humanoid Robots

Zicheng Liu, Haochen Wu, Jiawei He, Suyun Zhao, Huan Yang, Cheng Chi, Hao Tang

arXiv:2610.23288PDFCode

AI-assisted summary based on the paper. Consult the linked source for methods and limitations. Benchmark numbers appear after editor verification.

TL;DR

CoEvolve represents multi-robot team coordination as executable code over fixed humanoid motor skills, and adapts it during execution using Verifier Guided Coordination Evolution. A verifier accepts/rejects each transition; rejected transitions trigger typed, verifier-scoped strategy/coordination/parameter revision from an accepted checkpoint, and verified outcomes are stored for later reuse. In simulation with two Unitree G1 robots transporting a rigid payload through corridors, CoEvolve completes 8 tasks across 15 runs (base SR@80).

Why it matters

The authors address adapting executable inter-robot coordination when physical execution deviates under shared physical constraints. By combining executable code coordination, per-transition physical verification with checkpoint recovery, and verified experience reuse, CoEvolve targets long-horizon coordination repair beyond locally valid or verifier-accepted motions.

Method

  • Represent team coordination as executable code grounded on fixed motor skills: the orchestrator commits coordination primitives, semantic objectives, temporary roles, coordination mode, and bounded execution parameters to generate paired robot policies.
  • Verifier Guided Coordination Evolution: execute from the latest accepted checkpoint; the verifier evaluates measured task progress, payload support/stability, robot configuration, and contacts; rejected transitions trigger typed revision limited to Strategy/Coordination/Parameter.
  • Verified cooperative experience: maintain strategy-record memory and executable action exemplar memory; retrieve strategy by active phase and failure context, retrieve exemplars compatible with committed coordination; update only after physical execution outcomes; unexecuted repairs are not promoted.

Limitation

The authors report persistent limitations in motion precision and contact stability.

Abstract (from arXiv)

Recent advances in Vision-Language-Action (VLA) models and World Action Models (WAMs) have expanded the manipulation and whole body capabilities of individual robots. Yet a single robot remains limited by its embodiment, and many real world tasks require capabilities that can only emerge through cooperation. Existing multi robot frameworks largely emphasize task decomposition, allocation, and replanning, while providing limited support for adapting executable coordination when physical execution deviates from the intended behavior. We present CoEvolve, a framework that represents team coordination as executable code over fixed robot motor skills. CoEvolve introduces Verifier Guided Coordination Evolution, where a multimodal foundation model instantiates coordination strategies through motion primitives, semantic objectives, temporary robot roles, and bounded execution parameters. Physical verification evaluates each executed transition, rejected behaviors trigger structured coordination revision from an accepted checkpoint, and verified experience is retained to guide subsequent decisions. We evaluate CoEvolve with two simulated Unitree G1 humanoids transporting a freely supported rigid payload through constrained corridors. Across 15 base task runs with three multimodal backbone families, CoEvolve completes 8 tasks, with every backbone achieving end to end success. Additional evaluations yield verified placement on two additional corridor layouts and after both same-backbone and cross-backbone experience initialization, while exposing persistent limitations in motion precision and contact stability. These results support executable coordination as a practical interface for adapting long horizon multi robot collaboration under shared physical constraints.

Related papers

Publication record

Bibliographic evidence describes where and when a record appeared. It does not validate scientific claims.

First public release: 2026-10-16 · Indexed by Pub-AI: 2026-10-19 UTC

preprint

CoEvolve: Self-Evolving Code Policies from Skills to Strategies for Humanoid Robots

Zicheng Liu, Haochen Wu, Jiawei He, Suyun Zhao, Huan Yang, Cheng Chi, Hao Tang

Preprint posted: 2026-10-16 · Original record ↗

arxiv
2610.23288
Explicit source versions (1)
  • v2 · 2026-10-19
  1. Explicit source version detected · v2

    Source date: 2026-10-19 · First detected: 2026-10-25 UTC · Historical baseline

    Source evidence (1)
  2. Preprint available

    Source date: 2026-10-16 · First detected: 2026-10-25 UTC · Historical baseline

    Source evidence (1)
Abstract · arxiv

Recent advances in Vision-Language-Action (VLA) models and World Action Models (WAMs) have expanded the manipulation and whole body capabilities of individual robots. Yet a single robot remains limited by its embodiment, and many real world tasks require capabilities that can only emerge through cooperation. Existing multi robot frameworks largely emphasize task decomposition, allocation, and replanning, while providing limited support for adapting executable coordination when physical execution deviates from the intended behavior. We present CoEvolve, a framework that represents team coordination as executable code over fixed robot motor skills. CoEvolve introduces Verifier Guided Coordination Evolution, where a multimodal foundation model instantiates coordination strategies through motion primitives, semantic objectives, temporary robot roles, and bounded execution parameters. Physical verification evaluates each executed transition, rejected behaviors trigger structured coordination revision from an accepted checkpoint, and verified experience is retained to guide subsequent decisions. We evaluate CoEvolve with two simulated Unitree G1 humanoids transporting a freely supported rigid payload through constrained corridors. Across 15 base task runs with three multimodal backbone families, CoEvolve completes 8 tasks, with every backbone achieving end to end success. Additional evaluations yield verified placement on two additional corridor layouts and after both same-backbone and cross-backbone experience initialization, while exposing persistent limitations in motion precision and contact stability. These results support executable coordination as a practical interface for adapting long horizon multi robot collaboration under shared physical constraints.