SEAM: Smooth Execution of Action-Chunked Motion for Vision-Language-Action Policies
2026-07-06 • Robotics
Robotics
AI summaryⓘ
The authors studied how robots following vision-language instructions can have sudden jerky movements between chunks of planned actions because each chunk is generated separately and might not line up well. They introduced SEAM, a new method that smooths these action transitions without extra training or slow computations by using what the robot already did as a reference. SEAM applies small adjustments during execution to avoid abrupt jumps, improving motion smoothness while keeping task success rates and computational cost about the same.
Vision-Language-Action (VLA)action chunkstrajectory discontinuityGaussian latentsdenoisingflow matchinginference-time methodEuler steprobot motion smoothing
Authors
Dijia Zhan, Xuemiao Xu, Jinyi Li, Jie Tang
Abstract
Vision-Language-Action (VLA) policies that execute fixed-length action chunks can exhibit multimodal bifurcation: a cross-chunk inconsistency in which adjacent chunks generated from independent Gaussian latents can converge to incompatible trajectory modes, producing abrupt discontinuities at chunk boundaries. Existing remedies either require backpropagation through the policy at each denoising step, rely on rejection sampling, or require retraining, each trading computational cost or task reliability for smoother transitions. We propose SEAM (Smooth Execution of Action-Chunked Motion), a training-free inference-time method for flow matching VLAs. SEAM exploits a simple synchronous-execution insight: after the robot consumes the executed prefix, the previous chunk's unexecuted tail is already available as an analytic consistency reference. Its core mechanism, Velocity-guided Loss Steering (VLS), derives a time-dependent target from this tail and applies a closed-form correction after each Euler step without backpropagating through the policy network. On LIBERO-10 with pi_0.5, SEAM reduces boundary jerk by 28%, reduces chunk transition discontinuity by 27%, preserves baseline-level task success, and keeps denoising-loop cost near the unguided baseline.