Event
InstAgent: A Recursive Self-Improving Operating System for Heterogeneous Robot Collaboration
1 reports · 1 sources
Haoyang Luo et al. propose InstAgent, a recursive self-improving (RSI) agent operating system for heterogeneous robot collaboration, which explicitly models capability allocation, inter-robot dependencies, and handoff checks using agent behavior trees (BT), and updates collaboration programs through execution feedback, retaining successful revisions for reuse in subsequent tasks. Experiments on 20 real multi-robot tasks and 3 high-level models show consistent improvements in task success rate, first-time completion rate, and execution efficiency.
Report timeline
- InstAgent: A Recursively Self-Improving Agent Operating System for Heterogeneous Robot Collaboration
InstAgent proposes a recursively self-improving (RSI) agent operating system that uses agentic behavior trees (BTs) to explicitly model capability assignment, inter-robot dependencies, and handoff checks for heterogeneous robots, and updates collaboration programs from execution feedback while preserving successful revisions for reuse in later tasks. Experiments on 20 real-world multi-robot tasks with 3 high-level models show consistent gains in task success rate, first-attempt completion rate, and execution efficiency.
arXiv cs.RO ·