On-Device Robotic Planning: Eliminating Inference Redundancy for Efficient Decision-Making
This work tackles the problem of high inference latency in reasoning-based robotic policies, which limits real-time deployment for practical robotic applications, offering an incremental improvement.
This paper addresses the high inference latency of reasoning-based robotic policies by identifying temporal redundancy in robotic reasoning workloads. They propose REIS, a framework that minimizes unnecessary reasoning while maintaining semantic adaptability, achieving significant suppression of reasoning overhead on ALFRED and real-world robotic tasks.
Reasoning-based robotic policies using large language and vision-language models achieve strong semantic planning capabilities but mostly suffer from a high inference latency that limits practical real-time deployment. In this work, we observe that robotic reasoning workloads contain substantial temporal redundancy, where consecutive observations frequently produce identical actions and subgoals. Based on this insight, we present REIS, a human cognition inspired robotic decision-making framework that minimizes unnecessary reasoning while preserving semantic adaptability. REIS combines lightweight scene gating, KV-steered affordance routing, and deliberative reasoning to accelerate robotic control under embodied constraints. Experiments on ALFRED, and real-world robotic tasks demonstrate that REIS significantly suppresses reasoning overhead while maintaining competitive task performance.