Policy-Splitting Proximal Policy Optimization for Small Unmanned Surface Vehicle Navigation: A Simulation Study
Applied Sciences · 12.11.2026 · 16 (22) · 11218
DOI: 10.3390/app162211218 ↗Kurz gefasst
Fachartikel in der Zeitschrift Applied Sciences, online erschienen am 12.11.2026. Verfasst von Liang He und 3 weiteren Autor:innen. Die Arbeit nennt 30 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Jiangxi Polytechnic University, Jiangxi Provincial Department of Human Resources and Social Security.
Abstract (Originalsprache)
This study examines whether an explicit separation of navigation and avoidance yields useful control structure and performance gains for small unmanned surface vehicles in simulation. The policy-splitting proximal policy optimization actor (PSN-PPO) combines a navigation mean, a bounded avoidance residual, and a geometric risk gate with a learnable exponent, while retaining the PPO update and a separate critic. Five independent models per method are trained in each of five scenarios and evaluated on common nominal and perturbed cases. Crossing arrival is 100% for PSN-PPO versus 80% for PPO, but arrival without soft-domain intrusion is 25.6% versus 37.6%; static-obstacle arrival is 50.4% versus 80%. A 48-run equal-interaction study in static and crossing tasks finds no consistent advantage of full gating over always-on correction. Navigation-only control also achieves full crossing arrival, and the learned gate exponent remains close to its initial value. Simulated braking, turning, and paired threshold tests characterize the numerical maneuvering and risk scales. The contribution is an explicitly constrained actor decomposition and an empirical assessment of its limits; the results do not establish an additional performance benefit from the complete adaptive gate in these benchmark families.
Förderung
- Jiangxi Polytechnic University
- Jiangxi Provincial Department of Human Resources and Social Security
Zitieren
APA
He, L., Xiao, X., Li, M., & Wang, W. (2026). Policy-Splitting Proximal Policy Optimization for Small Unmanned Surface Vehicle Navigation: A Simulation Study. Applied Sciences, 16(22), 11218. https://doi.org/10.3390/app162211218
Vancouver
He L, Xiao X, Li M, Wang W. Policy-Splitting Proximal Policy Optimization for Small Unmanned Surface Vehicle Navigation: A Simulation Study. Applied Sciences. 2026;16(22):11218. doi:10.3390/app162211218
BibTeX
@article{He2026,
author = {He, Liang and Xiao, Xiong and Li, Mingjie and Wang, Weijun},
title = {Policy-Splitting Proximal Policy Optimization for Small Unmanned Surface Vehicle Navigation: A Simulation Study},
journal = {Applied Sciences},
year = {2026},
volume = {16},
number = {22},
pages = {11218},
doi = {10.3390/app162211218}
}
EUVIMED