Nonholonomic Narrow Dead-End Escape with Deep Reinforcement Learning
Published in Proceedings of the 2025 9th International Conference on Computer Science and Artificial Intelligence, 2025
This work studies escape behavior for Ackermann-style nonholonomic robots in narrow dead-end environments. It connects my robotics and reinforcement-learning background with broader questions in spatial decision making, embodied intelligence, and policy learning under action constraints.
The official citation should use the ACM DOI record. The arXiv link is provided as an open version.
Recommended citation: Denghan Xiong, Yanzhe Zhao, Yutong Chen, and Zichun Wang. 2025. Nonholonomic Narrow Dead-End Escape with Deep Reinforcement Learning. In Proceedings of the 2025 9th International Conference on Computer Science and Artificial Intelligence. ACM, 126-131. https://doi.org/10.1145/3788149.3788204
Download Bibtex