蚁群优化算法
运动规划
计算机科学
局部最优
启发式
路径(计算)
算法
数学优化
避障
网格
人工智能
移动机器人
数学
机器人
几何学
程序设计语言
作者
Yanli Chen,Guiqiang Bai,Yin Zhan,Xinyu Hu,Jun Liu
出处
期刊:IEEE Access
[Institute of Electrical and Electronics Engineers]
日期:2021-01-01
卷期号:9: 40728-40742
被引量:75
标识
DOI:10.1109/access.2021.3062375
摘要
Path planning is important to the efficiency and navigation safety of USV autonomous operation offshore.To improve path planning, this study proposes the improved ant colony optimizationartificial potential field (ACO-APF) algorithm, which is based on a grid map for both local and global path planning of USVs in dynamic environments.The improved ant colony optimization (ACO) mechanism is utilized to search for a globally optimal path from the starting point to the endpoint for a USV in a grid environment, and the improved artificial potential field (APF) algorithm is subsequently employed to avoid unknown obstacles during USV navigation.The primary contributions of this article are as follows: (1) this article proposes a new heuristic function, pheromone update rule, and dynamic pheromone volatilization factor to improve convergence and mitigate finding local optima with the traditional ant colony algorithm;(2) we propose an equipotential line outer tangent circle and redefine potential functions to eliminate goals unreachable by nearby obstacles (GNRONs) and local minimum problems, respectively; (3) to adapt the USV to a complex environment, this article proposes a dynamic early-warning step-size adjustment strategy in which the moving distance and safe obstacle avoidance range in each step are adjusted based on the complexity of the surrounding environment; (4) the improved ant colony optimization algorithm and artificial potential field algorithm are effectively combined to form the algorithm proposed in this article, which is verified as an effective solution for USV local and global path planning using a series of simulations.Finally, in contrast to most papers, we successfully perform field experiments to verify the feasibility and effectiveness of the proposed algorithm.
科研通智能强力驱动
Strongly Powered by AbleSci AI