On-Device Robotic Planning: Eliminating Inference Redundancy for Efficient Decision-Making
Joonhee Lee, Hyunseung Shin, Hyunmi Kim, Pei Zhang, Jeonggil Ko
THE PROBLEM
This paper focuses on Core ConceptsTaskThe job the robot is supposed to complete, such as pick-and-place, navigation, or drawer opening. Control & PlanningPlanningFiguring out what the robot should do before or during movement.. REIS cuts Robot LearningInferenceUsing a trained model to make predictions or choose actions. Simulation & Sim-to-RealLatencyDelay between input, computation, and action. for LLM/VLM-based robotic policies by ~50% or more by skipping redundant reasoning steps when consecutive frames would produce the same Core ConceptsActionA command the robot sends to its motors, controller, or low-level system.. This makes real-time Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. Control & PlanningControlThe method used to make the robot move the way you want. feasible on edge devices without sacrificing semantic Control & PlanningPlanningFiguring out what the robot should do before or during movement. capability—you get fast, smart decision-making instead of choosing between speed and reasoning. Read the paper by tracking the Core ConceptsTaskThe job the robot is supposed to complete, such as pick-and-place, navigation, or drawer opening. definition, the Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. or data assumptions, and the evidence that supports the claimed improvement.
HOW IT WORKS
Task framing
Core method
Data and supervision
Evaluation evidence
FIGURES
KEY RESULTS
REIS cuts Robot LearningInferenceUsing a trained model to make predictions or choose actions. Simulation & Sim-to-RealLatencyDelay between input, computation, and action. for LLM/VLM-based robotic policies by ~50% or more by skipping redundant reasoning steps when consecutive frames would produce the same Core ConceptsActionA command the robot sends to its motors, controller, or low-level system.. This makes real-time Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. Control & PlanningControlThe method used to make the robot move the way you want. feasible on edge devices without sacrificing semantic Control & PlanningPlanningFiguring out what the robot should do before or during movement. capability—you get fast, smart decision-making instead of choosing between speed and reasoning.
WHY DEVELOPERS SHOULD CARE
REIS cuts Robot LearningInferenceUsing a trained model to make predictions or choose actions. Simulation & Sim-to-RealLatencyDelay between input, computation, and action. for LLM/VLM-based robotic policies by ~50% or more by skipping redundant reasoning steps when consecutive frames would produce the same Core ConceptsActionA command the robot sends to its motors, controller, or low-level system.. This makes real-time Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. Control & PlanningControlThe method used to make the robot move the way you want. feasible on edge devices without sacrificing semantic Control & PlanningPlanningFiguring out what the robot should do before or during movement. capability—you get fast, smart decision-making instead of choosing between speed and reasoning.
LIMITATIONS
The main limitation to check is whether the claimed behavior holds outside the paper's reported setup. That means testing across different Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. embodiments, scenes, objects, and data distributions.
WHAT COMES NEXT
The practical next step is independent reproduction with clear baselines, ablations, and stress tests. For a developer, the useful follow-up is to map the paper's Core ConceptsTaskThe job the robot is supposed to complete, such as pick-and-place, navigation, or drawer opening. Control & PlanningPlanningFiguring out what the robot should do before or during movement. assumptions onto a concrete Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. stack, then test the smallest version of the method that could run end to end.