KineVLA: Towards Kinematics-Aware Vision-Language-Action Models with Bi-Level Action Decomposition

TL;DR

KineVLA introduces a framework for fine-grained, kinematics-aware vision-language-action tasks, enabling robots to interpret and execute complex, instruction-level kinematic commands with improved precision and adaptability.

Contribution

The paper proposes a novel bi-level action representation and reasoning approach for kinematics-rich VLA tasks, along with new datasets and benchmarks for evaluation.

Findings

KineVLA outperforms existing baselines on kinematics-sensitive benchmarks.

The framework achieves more precise and controllable robotic manipulation.

Experiments validate the effectiveness of bi-level reasoning tokens in aligning language and action.

Abstract

In this paper, we introduce a novel kinematics-rich vision-language-action (VLA) task, in which language commands densely encode diverse kinematic attributes (such as direction, trajectory, orientation, and relative displacement) from initiation through completion, at key moments, unlike existing action instructions that capture kinematics only coarsely or partially, thereby supporting fine-grained and personalized manipulation. In this setting, where task goals remain invariant while execution trajectories must adapt to instruction-level kinematic specifications. To address this challenge, we propose KineVLA, a vision-language-action framework that explicitly decouples goal-level invariance from kinematics-level variability through a bi-level action representation and bi-level reasoning tokens to serve as explicit, supervised intermediate variables that align language and action. To support this task, we construct the kinematics-aware VLA datasets spanning both simulation and real-world robotic platforms, featuring instruction-level kinematic variations and bi-level annotations. Extensive experiments on LIBERO and a Realman-75 robot demonstrate that KineVLA consistently outperforms strong VLA baselines on kinematics-sensitive benchmarks, achieving more precise, controllable, and generalizable manipulation behaviors.