Mapless navigation via Hierarchical Reinforcement Learning with memory-decaying novelty
Lists
Only you see this
Abstract
Highlights•HRL-based mapless navigation faces challenges of local minimum problems•A novel reward combining extrinsic and intrinsic rewards is proposed•The intrinsic reward is formed by novelty theory, episode memory, and memory decaying•A network incorporates LSTM to augment the robot’s memory and reasoning capabilities•Experiments demonstrated remarkable results in success rate and local minimum issue
BibTeX
@inproceedings{gao2024mapless,
author = {Yan Gao and Feiqiang Lin and Boliang Cai and Jing Wu and Changyun Wei and Raphael Grech and Ze Ji},
title = {{Mapless navigation via Hierarchical Reinforcement Learning with memory-decaying novelty}},
booktitle = {Robotics Auton. Syst},
year = {2024},
url = {https://openreview.net/forum?id=8flhf9UG3g},
}Loading related content...