水轮机叶片坑内修复机器人姿态估计改进算法  

Improved pose estimation algorithm for an on-site turbine blade repair robot

作　　者：马献德[1] 陈强[1] 孙振国[1] 张文增[1] 

机构地区：[1]清华大学机械工程系先进成形制造教育部重点实验室,北京100084

出　　处：《清华大学学报（自然科学版）》2011年第5期697-701,共5页Journal of Tsinghua University(Science and Technology)

基　　金：国家"八六三"高技术项目(2007AA04Z258);国家自然科学基金资助项目(50875147);2009年中国焊接学会预研奖学金A类创新课题资助项目(09-09-002)

摘　　要：水轮机叶片修复机器人的姿态估计,可能出现姿态矩阵退化和Euler角奇异的问题。该文采用四元数建立了"一次转动姿态误差"(one-rotation pose error,ORPE),以更好地描述机器人姿态估计误差。基于ORPE研究了低精度惯性器件长期运行的积累误差,并采用倾角传感器校正姿态估计。理论上,仅有倾角传感器本身不能校正姿态矩阵,该文以ORPE和Frobenius范数为准则,给出机器人姿态的有效校正值。仿真结果验证了所提算法的有效性。基于低精度惯性器件与低精度倾角传感器的组合,该算法可以给出相对较高的姿态估计精度。Pose estimates of mobile robots for repairing hydraulic turbine blades have problems with degenerated pose matrices and singular solutions of the Euler angles.A one-rotation pose error（ORPE） based on the quaternion was defined to better describe pose errors of mobile robots.The ORPE was used to evaluate the accumulated errors resulting from long-term use of a low-precision inertial guidance system,with inclinometers then used to adjust the pose.Theoretically,inclinometers alone can not be used to adjust the pose matrix.However,this paper shows that ORPE or the Frobenius norm of the matrix can be used to give effective adjustments.Simulations show the validity of these approaches and that the combination of a low-precision inertial guidance system and low-precision inclinometers can give rise to relatively high precision pose estimates with these algorithms.

关 键 词：移动机器人 姿态估计 校正 水轮机叶片 倾角传感器 四元数 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]