基于N次K-NN分类算法的管道机器人定位技术研究

Pipeline Robot Localization Technique Based on CSP and N Order K-NN Algorithm

  • 摘要: 分析了低频电磁波在均匀介质中的磁场分布,其分布与介质的介电常数、磁导率密切相关.根据管道机器人定位的实际工程需要,给出了三传感器低频电磁波定位模型.为了减小传输介质介电常数、磁导率参数对管道机器人定位的影响,提出了N次K-NN分类算法.实验结果表明,该算法分类的正确率可达97.5%,定位精度可达±10 cm,在传输介质介电常数、磁导率等参数不确定条件下,可有效地求解低频电磁波发射源的位置参数.

     

    Abstract: Magnetic field distribution of low frequency electromagnetic wave in even medium is analyzed,which is related to the medium's dielectric coefficient and magnetic inductive capacity.According to the localization requirement of the pipeline robot for a pipeline inspection project,a localization model based on tri-sensors is established.In order to reduce the effect of dielectric coefficient and magnetic inductive capacity on pipeline robot's localization,an N order K nearest neighbor(K-NN) electromagnetic wave localization algorithm is presented.Experimental results show that the correct sort percentage of the algorithm can reach 97.5% and the location precision can reach ±10 cm.The low frequency electromagnetic wave transmit antenna's location can be calculated effectively by the algorithm when the medium's dielectric coefficient and magnetic inductive capacity is uncertain.

     

/

返回文章
返回