Active Localization problem
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/ronbenc/Anytime-Incremental-rho-POMDP-Planning-in-Continuous-Spaces
下载链接
链接失效反馈官方服务：
资源简介：
该数据集旨在让智能体在一个充满障碍的连续二维环境中，通过信标观测来最小化对其位置的不确定性。奖励机制完全基于信息增益，以此推动不确定性的降低。该实验进行了1000次试验，任务目标是尽可能地减少智能体位置的不确定性。
This dataset is designed to enable AI Agents to minimize their positional uncertainty via beacon observations in a continuous 2D environment cluttered with obstacles. Its reward mechanism is exclusively based on information gain, which incentivizes the reduction of positional uncertainty. A total of 1000 trials are conducted for this experiment, with the task objective being to minimize the AI Agent's positional uncertainty as much as possible.
提供机构：
Authors of the paper
