基于鲁棒信息滤波器的图像雅可比矩阵在线估计

摘要

This paper reviews present mainstream uncalibrated estimations of image Jacobian matrix (UM) , and analyzes methods based on Kalman Filter, Fuzzy Adaptive Kalman Filter and Particle Filter, as well as their advantages and disadvantages in detail. In order to improve system estimation precision in unknown environment, this paper selects the estimation framework based on filtering theory, adjusts the system model, utilizes Robust Information Filter (RIF) to estimate IJM, and realizes moving-object-tracking accurately in simulation and industrial robot system respectively. RIF is robust to bounded noise with any distribution, both simulation and experimental results verify the effectiveness of the introduced method.%针对现有的图像雅可比矩阵无标定求解方法,分析了基于Kalman滤波、模糊自适应Kalman滤波和粒子滤波的图像雅可比矩阵在线估计的优缺点.为了进一步提高未知环境下的系统估计精度,选择基于滤波理论的估计框架,对系统模型进行调整,用鲁棒信息滤波器在线估计图像雅可比矩阵,该滤波算法对任意分布的有界噪声都具有较强的鲁棒性.仿真和实验结果表明,在未知系统噪声的情况下,该算法仍可以实现图像雅可比矩阵的精确估计.