摘要
新型海洋机器人能够有效捕获能量,减少能源消耗,提高续航能力。论文基于A3C算法,提出一种考虑能量捕获的海洋机器人节能局部路径规划算法。根据“驭浪者”号的能源系统和运动特性,考虑海流和海风对海洋机器人的影响,建立能耗模型,结合A3C算法设计趋近目标奖励函数和避障奖励函数,得到距离最优的局部路径规划(DOA3C)算法;考虑新型海洋机器人对风能和太阳能的捕获,建立能量捕获模型。依据能耗模型和能量捕获模型,设计节能奖励函数,采用A3C算法得到能源最优的局部路径规划(EOA3C)算法。通过仿真试验,对比DOA3C和EOA3C算法在相同海洋环境下的能耗情况,验证算法的可行性和有效性。
Ocean-powered robots can effectively capture ocean energy,reduce energy consumption,and improve endurance.In this paper,an energy-saving A3C local path planning algorithm for ocean energy-driven robots considering energy capture is proposed.According to the energy system and motion characteristics of the"Wave Rider",considering the influence of ocean currents and sea breeze on the marine energy driving robot,an energy consumption model,combined with the Asynchronous Advantage Actor-Critic(A3C)algorithm is established to design the approach target reward function and obstacle avoidance reward function and obtain the Distance Optimal Asynchronous Advantage Actor-Critic(DOA3C)algorithm.Considering the capture of wind energy and solar energy,a model of ocean energy capture is established.According to the energy consumption model and the ocean energy capture model,an energy-saving reward function was designed,and Energy Optimal Asynchronous Advantage Actor-Critic(EOA3C)was obtained by using the A3C algorithm.Through the simulation test,the energy consumption calculated with the DOA3C algorithm and the EOA3C algorithm in the same marine environment is compared to verify the feasibility and effectiveness of the algorithm.
作者
廖煜雷
李可
赵永波
刘骁锋
张家华
LIAO Yulei;LI Ke;ZHAO Yongbo;LIU Xiaofeng;ZHANG Jiahua(Science and Technology on Underwater Vehicle Laboratory,Harbin Engineering University,Harbin 150001,China;Nanhai Institute of Harbin Engineering University,Sanya 572024,China)
出处
《中国造船》
EI
CSCD
北大核心
2023年第6期225-239,共15页
Shipbuilding of China
基金
国家自然科学基金(52071097)
水下机器人技术重点实验室研究基金(2021JCJQ-SYSJJ-LB06910)
海南省自然科学基金(522MS162)。
关键词
海洋机器人
局部路径规划
A3C算法
能源最优
ocean energy driven robot
local path planning
Asynchronous Advantage Actor-Critic algorithm
energy optimization