FailBench is a new benchmark for robot failure detection, containing 2,197 manipulation attempts from 14 public sources, with 75% of failures occurring naturally. The study evaluates 13 vision‑language model (VLM) detectors, finding the best model achieves only 0.77 mean balanced accuracy, and that fine‑tuned failure detectors often underperform general‑purpose VLMs. Performance varies with visual evidence, excelling when object motion is observable but dropping to near chance on contact‑intensive assembly tasks, and input‑level cropping of outcome‑relevant regions improves the top detector by 2.4 percentage points.
arXiv:2512.01946v4 Announce Type: replace-cross Abstract: Robust robotic manipulation requires reliable failure detection and recovery. Although recent Vision-Language Models (VLMs) show promise in r...
arXiv:2606. 28385v1 Announce Type: cross Abstract: Recent advances in robot world models enable synthetic video generation for embodied prediction and planning.
The paper introduces FailureSpot, a label‑efficient method for detecting failures at the timestamp level in vision‑language‑action (VLA) policies. It first generates weak supervision from unlabeled VLA action chunks by identifying abnormal patterns, then employs active learning to annotate only the most uncertain trajectories. Experiments on multiple VLA policies demonstrate improved performance for both timestamp‑level and trajectory‑level failure detection.
VLA-Scope is a two‑stage framework designed to predict failures in vision‑language‑action models under distribution shifts. The first stage detects out‑of‑distribution inputs and classifies their shift categories using pooled image and language representations. For OOD inputs, the second stage updates failure risk during execution by combining shift category, action‑prefix features, and execution progress, achieving a ROC‑AUC of 0.8497 after 60 actions and outperforming baseline methods.
arXiv:2606. 08881v1 Announce Type: cross Abstract: Vision-Language-Action (VLA) models have demonstrated strong generalization in robotic manipulation, yet existing evaluations are primarily conducted in simulation or on expensive robotic platforms, leaving their robustness on affordable real-world robots largely unexplored.
arXiv:2603. 15600v2 Announce Type: replace-cross Abstract: Accurate process supervision remains a critical challenge for long-horizon robotic manipulation.
ProTracer is a training‑free framework that uses Vision‑Language Models (VLMs) together with proprioceptive signals to analyze robot manipulation failures. It performs binary failure detection, categorization, explanation generation, and introduces failure onset localization—identifying the earliest moment a robot deviates from a valid trajectory leading to failure. The method leverages proprioceptive dynamics to pinpoint informative action boundaries and converts robot‑state signals into natural‑language descriptions for joint multimodal reasoning, achieving strong performance on both conventional failure diagnosis and the new failure onset localization task.
The paper introduces Configured Failure Trapping, a new backdoor attack targeting Vision‑Language‑Action (VLA) models that activates through subtle textual triggers and forces the robot to fail in a specific, controlled manner. It presents a data engine for generating high‑quality target trajectories, an automated evaluation suite, and two benchmarks—Trap‑LIBERO and Trap‑RoboTwin—covering four failure modes. The authors propose TrapVLA, a method that learns trigger‑induced action residuals to steer policies toward the desired failure behavior, demonstrating effectiveness in both simulation and real‑world robotic experiments while maintaining performance on clean data.
arXiv:2608.29967v1 Announce Type: cross Abstract: Vision-Language-Action (VLA) models demonstrate strong semantic understanding yet exhibit systematic failures during deployment. The conditions under...
arXiv:2606. 08508v1 Announce Type: cross Abstract: Generative robot policies fail unpredictably at deployment: they hesitate at critical moments, drift off-task, or commit to unrecoverable actions.
LIBERO-VPro is a benchmark designed to assess the closed‑loop visual robustness of robotic foundation models by systematically perturbing visual inputs during task execution. It spans four dimensions—Visual Evidence Degradation, Camera Staleness, Visual Source Consistency, and Task‑Relevant Scene Variation—across 12 challenge categories, 96 settings, and 3,296 task‑condition cases. Evaluations on six models over 196,000 simulated episodes and 200 real‑world rollouts show that high nominal performance can hide significant weaknesses in visual grounding, adaptation, and sensitivity to stale or missing observations.
Vision-language-action models (VLAs) combine vision-language backbones with expressive generative action heads trained via flow matching on large-scale robotic datasets. Despite their strong empirical performance in robotic manipulation, VLAs lack mechanisms to quantify confidence in their predictions and to detect when their actions may be unreliable.