星团(航天器)
运动规划
路径(计算)
计算机科学
人工智能
计算机网络
机器人
作者
Bowen Xing,Hanzheng Chen,Zhenchong Liu,Xiao Wang
标识
DOI:10.1002/adts.202401001
摘要
Abstract The path planning of deep‐sea mining vehicle clusters and the spatial layout of pipeline systems are critical for mining efficiency and safety. Many existing path planning strategies overlook hose entanglement issues, limiting their applicability in complex environments. This paper presents a novel full‐coverage path‐planning method based on an improved Deep Q‐Network (DQN) algorithm. The proposed algorithm optimizes sample selection and incorporates a backtracking mechanism to improve learning efficiency and correct erroneous actions. Moreover, an innovative spatial constraint mechanism is designed to transform the hose entanglement problem into a path optimization problem, thereby proactively avoiding entanglement risks during planning. The experiments show that the algorithm proposed in this paper can achieve a coverage rate of 100% of the target area within 200 steps, with no instances of hose entanglement. Furthermore, the algorithm handles dynamic obstacles and flexibly adjusts vehicle numbers, proving its adaptability and robustness in changing environments. Overall, the paper provides a highly practical and innovative solution for intelligent path planning in deep‐sea mining operations.
科研通智能强力驱动
Strongly Powered by AbleSci AI