Safe Task Planning with Long-Term Graph Memory for Embodied Agents
Safe task planning for embodied agents: keeps a long-term semantic graph of the scene so an LLM risk predictor can catch hazards outside the current view and trigger replanning. IS-Bench + real robot.具身智能体的安全任务规划：持续维护场景的长期语义图记忆，让 LLM 风险预测器发现视野外的隐患并触发重规划。在 IS-Bench 与真机上验证。
BibTeX
@misc{li2026safe,
title = {{Safe Task Planning with Long-Term Graph Memory for Embodied Agents}},
author = {Siyuan Li and Taiyan Lang and Aoqi Yan and Jia Yu and Feifan Liu and Yihan Du and Yu Zheng and Xun Wang and Peng Liu},
year = {2026},
eprint = {2609.08444},
archivePrefix = {arXiv},
primaryClass = {cs.RO},
url = {https://arxiv.org/abs/2609.08444},
}