Multi-UAV encirclement training optimization method based on implicit curriculum experience partitioning
This paper proposes an implicit curriculum learning framework for multi-drone coordination, employing autonomous phased training via dynamic experience replay. By integrating stage-critical sample prioritization and role-adaptive exploration policies, the method resolves local-global strategy confli...
| Published in: | Инноватика-2025 : сборник материалов XXI Международной школы-конференции студентов, аспирантов и молодых ученых, 28-30 апреля 2025 г., г. Томск, Россия С. 551-558 |
|---|---|
| Main Author: | |
| Format: | Book Chapter |
| Language: | English |
| Subjects: | |
| Online Access: | https://vital.lib.tsu.ru/vital/access/manager/Repository/koha:001272935 Перейти в каталог НБ ТГУ |
| Summary: | This paper proposes an implicit curriculum learning framework for multi-drone coordination, employing autonomous phased training via dynamic experience replay. By integrating stage-critical sample prioritization and role-adaptive exploration policies, the method resolves local-global strategy conflicts without manual curriculum design. Validated in dynamic environments, the framework demonstrates enhanced collaborative efficiency and hierarchical strategy learning capabilities, providing systematic solutions for phased multi-agent tasks. |
|---|---|
| Bibliography: | Библиогр.: 12 назв. |
| ISBN: | 9785936297311 |