Search papers, labs, and topics across Lattice.
This paper introduces HUGIN, a training framework designed to enhance vision-language planning for autonomous logistics sorting systems (ALSS) by addressing the challenges of Joint Multi-Scene Understanding (JMSU). HUGIN employs Endogenous Data Augmentation and Global Context Ranking to improve the alignment of instruction representations with complete visual contexts, significantly boosting performance across five open vision-language models (VLMs). Notably, the Qwen3-VL-8B model's accuracy on the newly constructed SortingBench dataset increases from 63.6% to 78.8%, demonstrating the framework's effectiveness and practical applicability in real-world logistics scenarios.
HUGIN boosts vision-language model accuracy for logistics sorting by over 15 percentage points, showcasing a significant leap in performance through innovative training techniques.
Autonomous logistics sorting systems (ALSS) are an important industrial application of embodied AI, which requires joint planning over spatially disjoint camera views. We formulate this setting as Joint Multi-Scene Understanding (JMSU). With open-world visual understanding and task-planning capabilities, vision-language models (VLMs) are promising candidates for JMSU. However, directly applying existing VLMs to JMSU is non-trivial due to scarce cross-scene supervision and attention dispersion caused by long visual context in JMSU. To address these challenges, we propose HUGIN, a training framework with two complementary components. Endogenous Data Augmentation recombines verified atomic facts under operating constraints, while Global Context Ranking aligns the instruction representation more strongly with the complete visual context than with a partial visual context. To support ongoing research, we construct a high-quality industrial sorting dataset and benchmark named SortingBench from four layouts of autonomous logistics sorting systems. Across five open VLMs, HUGIN consistently outperforms matched baselines; for example, the accuracy on SortingBench of Qwen3-VL-8B increases from 63.6% to 78.8%. Additional experiments verify the effectiveness of each component and JMSU's spillover benefits in embodied tasks. Deployment tests involving more than 15,000 packages support the practical viability of VLM-based planning for autonomous logistics sorting.
