Adaptive Information Belief Space Planning
Abstract
Reasoning about uncertainty is vital in many real-life autonomous systems. However, current state-of-the-art planning algorithms either cannot reason about uncertainty explicitly, or do so with high computational burden. Here, we focus on making informed decisions efficiently, using reward functions that explicitly deal with uncertainty. We formulate an approximation, namely an abstract observation model, that uses an aggregation scheme to alleviate computational costs. We derive bounds on the expected information-theoretic reward function and, as a consequence, on the value function. We then propose a method to refine aggregation to achieve identical action selection in a fraction of the computational time.
Cite
Text
Barenboim and Indelman. "Adaptive Information Belief Space Planning." International Joint Conference on Artificial Intelligence, 2022. doi:10.24963/IJCAI.2022/637Markdown
[Barenboim and Indelman. "Adaptive Information Belief Space Planning." International Joint Conference on Artificial Intelligence, 2022.](https://mlanthology.org/ijcai/2022/barenboim2022ijcai-adaptive/) doi:10.24963/IJCAI.2022/637BibTeX
@inproceedings{barenboim2022ijcai-adaptive,
title = {{Adaptive Information Belief Space Planning}},
author = {Barenboim, Moran and Indelman, Vadim},
booktitle = {International Joint Conference on Artificial Intelligence},
year = {2022},
pages = {4588-4596},
doi = {10.24963/IJCAI.2022/637},
url = {https://mlanthology.org/ijcai/2022/barenboim2022ijcai-adaptive/}
}