血管内导航多任务世界模型控制的渐进经验融合
原标题：Progressive Experience Fusion for Multi-Task World Model Control in Endovascular Navigation
来源：arXiv cs.RO发布时间待核实
用渐进经验融合（PEF）训练多任务 TD-MPC2 控制器做血管内导航：30 例血管训练后在 10 例留出血管上成功率 90%，迁移到体外卒中患者血管模型时经微调将路径完成率从 63% 提升至 80%。
AI 中文摘要 · AI 生成
arXiv:2608.18647v1 Announce Type: new Abstract: Autonomous endovascular navigation could support the delivery of mechanical thrombectomy to underserved areas, but controllers must navigate long, multi-stage paths across varying vascular anatomies. This study investigates Progressive Experience Fusion (PEF) to train a multi-task TD-MPC2 controller.