A Generative Self-Diagnosis Disengagement Reporting System for Autonomous Shuttles

The increasing presence of autonomous vehicles on public roads has highlighted the limitations of traditional incident reporting systems, which rely on human-generated tables and descriptions. To address this, we propose a generative reporting framework that integrates a large language model (LLM) with semantic scene generation models. This framework utilizes perception snapshots and self-diagnostic data to generate detailed incident reports, addressing environmental blind spots. Our 3D scene completion network, combining diffusion and state-space models, reconstructs blind zones undetected by exterior sensors, achieving IoU scores of 41.92 on SSCBench-KITTI360 and 44.13 on SemanticKITTI. Public road experiments validate the system's ability to improve incident report quality while maintaining performance.