管道地理坐标内检测的里程校正算法  

Odometer correction algorithm for geographic coordinate detection in pipeline

作　　者：杨理践[1] 杨洋[1,2] 高松巍[1] 李瑞强[1] 

机构地区：[1]沈阳工业大学信息科学与工程学院,沈阳110870 [2]辽宁大学物理学院,沈阳110036

出　　处：《仪器仪表学报》2013年第1期26-31,共6页Chinese Journal of Scientific Instrument

基　　金：国家自然科学基金(60927004;61141004);十二五国家科技部支撑计划(2011BAK06B01-03)资助项目

摘　　要：利用管道清管器(pipeline cleaner,PIG)搭载捷联惯导(strap down intertial navigation system,SINS)器件实现管道地理坐标的定期测量。SINS计算的结果发散是测量系统实用化的瓶颈,为此将PIG里程校正的卡尔曼(Kalman)滤波方法应用于管道地理坐标测量中。根据管道内检测环境的需要选择霍尔式里程轮和采用微机电系统(micro-electromechanical systems,MEMS)的惯性测量单元(inertial measurement unit,IMU);针对MEMS器件精度较低的问题,利用分离滤波方法减小信号的零偏噪声,使惯性信号满足Kalman滤波的需要;建立了9维状态误差的数学模型,将里程轮的速度与SINS计算的速度之差作为观测量,通过Kalman滤波对管道定位的状态误差进行估计和补偿;搭建实验装置进行实验验证。结果表明,校正算法解决了计算结果发散的问题,检测160 m长度的管道精度达到3.8%,具有一定的实用性。The pipeline cleaner （PIG） equipped with strap down inertial navigation system （SINS） is utilized to realize the periodic measurement of pipeline geographic coordinates. The divergence of SINS calculation result is the bottleneck of the measurement system application. Hence, the Kalman filtering algorithm with odometer correction is applied to the pipeline geographical coordinate measurement. The Hall type odometer and micro-electromechanical system （MEMS） type inertial measurement unit （IMU） are selected according to the requirement of in pipeline detection environment; aiming at the low accuracy of the MEMS, the separated filtering method is utilized to diminish the zero random noise, and make inertial signal meet the requirement of the Kalman filtering; and a nine-dimensional state error mathematic model is established. Taking the difference between the odometer velocity and SINS calculated velocity as the observed quantity, Kalman filtering is used to estimate and compensate the state error of pipeline positioning. An experimental apparatus was buiXt to carry out experimental verification. Experiment results indicate that the correction algorithm resolves the divergence problem of the calculation results, the detection precision achieves 3.8% for a 160 m pipeline,and the algorithm has practicability.

关 键 词：管道地理坐标 管道内检测 捷联惯导 卡尔曼滤波 姿态角 误差模型 里程轮 

分 类 号：TP274.5[自动化与计算机技术—检测技术与自动化装置] TH82[自动化与计算机技术—控制科学与工程]