CoRE: Learning Collaboration-Role Experts for Decentralized Collaborative Manipulation with One Policy

Key points

• Collaborative manipulation requires robots to perform complementary actions as interactions unfold.1
• We study single-policy decentralized collaboration: every robot runs the same policy from its visual observations and proprioception, without task prompts, identity labels, or inter-robot messages.1
• The challenge is to learn complementary team behaviors within shared parameters and select appropriate actions from each robot's local observations.1
• During training, an action-expert alignment loss supervises expert selection using relative forced-route prediction errors against demonstrations under fixed inputs, without role labels.1

Sources (1)

• [1]CoRE: Learning Collaboration-Role Experts for Decentralized Collaborative Manipulation with One Policy
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 6, 04:49 AM

  We introduce CoRE, which learns Collaboration-Role Experts from pooled multi-task, multi-robot demonstrations.

  Collaborative manipulation requires robots to perform complementary actions as interactions unfold.

Extractive summary: sentences quoted from the sources.