PEAfowl: Perception-Enhanced Multi-View Vision-Language-Action for Bimanual Manipulation

TL;DR

PEAfowl introduces a perception-enhanced multi-view vision-language-action policy that improves bimanual manipulation by incorporating 3D spatial reasoning, iterative instruction grounding, and depth distillation, leading to significant success rate improvements.

Contribution

The paper presents a novel multi-view VLA policy with 3D spatial reasoning, iterative instruction grounding, and depth distillation, enhancing generalization and robustness in bimanual manipulation tasks.

Findings

23.0 percentage points success rate improvement over baseline

Effective sim-to-real transfer demonstrated on real robots

Depth distillation enhances perception accuracy without inference overhead

Abstract

Bimanual manipulation in cluttered scenes requires policies that remain stable under occlusions, viewpoint and scene variations. Existing vision-language-action models often fail to generalize because (i) multi-view features are fused via view-agnostic token concatenation, yielding weak 3D-consistent spatial understanding, and (ii) language is injected as global conditioning, resulting in coarse instruction grounding. In this paper, we introduce PEAfowl, a perception-enhanced multi-view VLA policy for bimanual manipulation. For spatial reasoning, PEAfowl predicts per-token depth distributions, performs differentiable 3D lifting, and aggregates local cross-view neighbors to form geometrically grounded, cross-view consistent representations. For instruction grounding, we propose to replace global conditioning with a Perceiver-style text-aware readout over frozen CLIP visual features, enabling iterative evidence accumulation. To overcome noisy and incomplete commodity depth without adding inference overhead, we apply training-only depth distillation from a pretrained depth teacher to supervise the depth-distribution head, providing perception front-end with geometry-aware priors. On RoboTwin 2.0 under domain-randomized setting, PEAfowl improves the strongest baseline by 23.0 pp in success rate, and real-robot experiments further demonstrate reliable sim-to-real transfer and consistent improvements from depth distillation. Project website: https://peafowlvla.github.io/.