aiqbench analysis
PEAfowl advances bimanual robotic manipulation by integrating multi-view depth distributions and iterative text-aware readouts to handle occlusions and viewpoint shifts more robustly than standard token concatenation methods. The approach leverages frozen CLIP features and a depth distillation technique to refine geometric priors without increasing inference latency, achieving a 23 percentage point improvement in success rates on the RoboTwin 2.0 benchmark. This work represents a significant step forward in building reliable Vision-Language-Action policies for complex physical tasks.
PEAfowl: Perception-Enhanced Multi-View Vision-Language-Action for Bimanual Manipulation
arXiv:2601.17885v2 Announce Type: replace-cross Abstract: Bimanual manipulation in cluttered scenes requires policies that remain stable under occlusions, viewpoint changes and scene variations. Existing vision-language-action models often lack such robustness because (i) multi-view features are fused via view-agnostic token concatenation, yielding limited cross-view spatial representations, 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 perception, PEAfowl predicts per-token depth distributions, performs differentiable 3D lifting, and aggregates local cross-view neighbors to form geometrically grounded, cross-view aligned representations. For language utilization, we propose to replace global conditioning with a Perceiver-style text-aware readout over frozen CLIP visual features, enabling iterative evidence accumulation. To better exploit commodity RGB-D sensing despite noisy and incomplete depth, PEAfowl's depth-distribution lifting naturally supports training-only depth distillation, where a pretrained depth teacher supervises the depth-distribution head to inject refined geometric priors without adding inference overhead. On RoboTwin 2.0 under domain-randomized setting, PEAfowl improves the strongest baseline by 23.0 pp in success rate, and physical experiments further demonstrate improved performance on the evaluated real-robot tasks. Project website: https://peafowlvla.github.io/.
Abstract:Bimanual manipulation in cluttered scenes requires policies that remain stable under occlusions, viewpoint changes and scene variations. Existing vision-language-action models often lack such robustness because (i) multi-view features are fused via view-agnostic token concatenation, yielding limited cross-view spatial representations, 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 perception, PEAfowl predicts per-token depth distributions, performs differentiable 3D lifting, and aggregates local cross-view neighbors to form geometrically grounded, cross-view aligned representations. For language utilization, we propose to replace global conditioning with a Perceiver-style text-aware readout over frozen CLIP visual features, enabling iterative evidence accumulation. To better exploit commodity RGB-D sensing despite noisy and incomplete depth, PEAfowl's depth-distribution lifting naturally supports training-only depth distillation, where a pretrained depth teacher supervises the depth-distribution head to inject refined geometric priors without adding inference overhead. On RoboTwin 2.0 under domain-randomized setting, PEAfowl improves the strongest baseline by 23.0 pp in success rate, and physical experiments further demonstrate improved performance on the evaluated real-robot tasks. Project website: this https URL.
Accepted by IEEE Robotics and Automation Letters (RA-L), 2026. This version includes an extended appendix with additional implementation details, deployment analysis, robustness evaluation, and real-world evaluation protocol. DOI: https://doi.org/10.1109/LRA.2026.3726379
Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Robotics (cs.RO)
arXiv:2601.17885 [cs.CV]
(or arXiv:2601.17885v2 [cs.CV] for this version)
https://doi.org/10.1109/LRA.2026.3726379 DOI(s) linking to related resources
Full text under the publisher's open licence · 238 words
Related coverage
NVIDIA GTC Taipei at COMPUTEX: Live Updates on What’s Next in AI
NVIDIA · May 21, 2026
Vertical Advantage: Transforming Industries with Lakebase and Agentic AI
Databricks · Aug 27, 2026
The Machines that Make the Machines
NVIDIA · Oct 7, 2026
Does Adversarial Training Improve Generalization in Multi-View VLAs? Revealing and Mitigating View Collapse
arXiv · Sep 29, 2026