New method speeds up robot action prediction while improving success
Accelerating VLA Inference with Perturbation Informed Execution
Robotics
Summary
Robots often plan several actions ahead but only carry out part of their plan before rethinking. The authors found that always choosing long or short plans can slow things down or reduce accuracy. They created a method called PICES that looks at small differences in possible actions to decide how many steps a robot should do before planning again. This approach makes robots work faster and more successfully in tests and real tasks without extra training.
What this means in practice
- •For robotics engineers: Improve robot motion planning efficiency by dynamically choosing action lengths during execution to reduce computation time and increase task success.
- •For embedded systems developers: Deploy faster action inference on edge devices like Jetson Orin by selecting varying action horizons based on perturbation differences to optimize runtime.
Authors
Jehun Lee, Hajun Lee, Jimin Ryu, Youhan Lee, Beomseok Kang, Jonghyun Choi, Jae-Joon Kim
Abstract
Action-chunking vision-language-action (VLA) models predict sequences of future robot actions. Under fixed-horizon execution, the robot executes a fixedlength prefix of each predicted chunk before invoking the policy again, leaving the remaining predictions unused. For a given number of executed actions, short execution horizons require more policy calls. Extending the execution horizon amortizes policy inference cost over more actions. However, uniformly applying long horizons across episodes can limit opportunities for precise corrections. To address this limitation, we propose PICES (Perturbation-Informed Chunk and Execution Horizon Selection), a training-free method that jointly selects an action chunk and the number of actions executed before the next policy call. Within a single policy call, the action expert generates a batch of action chunks from multiple initial noise samples and their locally perturbed counterparts, using the same conditioning features across all samples. PICES exploits differences between corresponding actions in each nominal chunk and its counterpart to determine candidate-specific execution horizons and select a nominal chunk for execution. Experiments on simulation benchmarks and real-world manipulation tasks demonstrate that PICES achieves speedup by up to 2.3x in wall-clock policy inference time on Jetson Orin and Thor relative to fixed horizon baselines while improving the success rate.