摘要
水下动态对接可帮助自主水下机器人(AUV)实现与母艇之间的水下能源补充及数据交换,提高AUV作业能力,捕获式动态对接要求AUV具有准确估计与母艇间相对位姿的能力。针对导引灯中心提取误差导致AUV相对母艇位姿估计不准的问题,提出了一种基于随机惯性权重粒子群(SIWPSO)优化算法的位姿估计方法,首先建立导引灯成像误差函数,对其优化求解取最小值得到AUV相对母艇位姿估计模型,以抑制像素中心提取误差的干扰,然后利用随机惯性权重粒子群算法解算位姿估计模型来得到最优相对位姿估计值。通过实验验证了所提出位姿估计方法的有效性,相对姿态误差在0.1°以内,相对位置误差在0.50m以内;其中各姿态角估计误差均值在2°以内,各位置估计误差均值在0.20m以内,估计精度较高,稳定性较好。
Underwater dynamic docking can help the autonomous underwater vehicle(AUV)to realize underwater energy supplement and data exchange with the mother boat,and improve the AUV’s operation ability.The capture dynamic docking requires that the AUV has the ability to accurately estimate the relative pose between the mother boat and the AUV.A pose estimation method based on stochastic inertial weighted particle swarm optimization(SIWPSO)was proposed to solve the problem of incorrect estimation of AUV relative to the mother boat caused by the center extraction error of homing light.Firstly,the pose estimation model of AUV relative to the mother boat was established by optimizing the imaging error function of the guidance lamp to take the minimum value,so as to suppress the interference of pixel center extraction error.Then,the pose estimation model was solved by using the stochastic inertial weighted particle swarm optimization algorithm to obtain the optimal relative pose estimation value.Experimental results show that the proposed pose estimation method is effective,and the relative pose error is within 0.1°and the relative position error is within 0.50m.The average estimation error of each attitude Angle is less than 2°,and the average estimation error of each position is less than 0.20m,indicating high estimation accuracy and good stability.
作者
杨本超
徐会希
吕凤天
石凯
YANG Benchao;XU Huixi;LYU Fengtian;SHI Kai(State Key Laboratory of Robotics,Shenyang Institute of Automation,Chinese Academy of Sciences,Liaoning Shenyang 110016,China;Institutes for Robotics and Intelligent Manufacturing,Chinese Academy of Sciences,Liaoning Shenyang 110169,China;Key Laboratory of Marine Robotics,Liaoning Shenyang 110169,China;University of Chinese Academy of Sciences,Beijing 100049,China)
出处
《机械设计与制造》
北大核心
2024年第12期318-322,327,共6页
Machinery Design & Manufacture
基金
中国科学院A类战略性先导科技专项(XDA22040103)
广东省重点领域研发计划(2020B1111010004)。
关键词
自主水下机器人
动态对接
位姿估计
粒子群优化
AUV(Autonomous Underwater Vehicle)
Dynamic Docking
Pose Estimation
Particle Swarm Optimization
