摘要
针对机器人视觉伺服中的位姿估计问题,提出了一种基于滚动时域优化的位姿估计方法.根据基于位置的视觉伺服结构以及相机投影模型给出了特征点的坐标变换,建立具有特征点丢失下的运动学离散时间模型.采用滚动优化策略,通过最小化位姿误差代价函数设计了最优位姿估计器,并给出了次优滚动时域位姿估计器设计方法以提高计算效率.进一步,证明了最优估计器和次优估计器的估计误差上界收敛.最后,通过仿真对比验证了文章所提算法的有效性和优越性.
In this paper,a pose estimation method based on moving horizon optimization is proposed to deal with robotic visual servoing problem.Firstly,according to the position-based visual servoing structure and camera projection model,the coordinates of feature points are transferred and the kinematics discrete time model with the loss of feature points is established.The optimal moving horizon estimator is designed by minimizing the pose error cost function with moving horizon policy,and the suboptimal moving horizon pose estimator is given to improve the calculation efficiency.Furthermore,the upper bounds of the estimation error of the optimal estimator and the suboptimal estimator converge are proved.Finally,the effectiveness and superiority of the proposed algorithm is verified by comparative simulation.
作者
卢威威
刘安东
仇翔
俞立
LU Weiwei;LIU Andong;QIU Xiang;YU Li(College of Information Engineering,Zhejiang University of Technology,Hangzhou 310023)
出处
《系统科学与数学》
CSCD
北大核心
2021年第7期1772-1787,共16页
Journal of Systems Science and Mathematical Sciences
基金
NSFC-浙江两化融合联合基金(U1709213)
国家自然科学基金(61973275)资助课题
关键词
机器人视觉伺服
位姿估计
滚动时域估计
代价函数
Robotic visual servoing
pose estimation
moving horizon estimation
cost function
