摘要
为了实现电磁跟踪系统无线精准定位,设计一种新型五自由度无线电磁定位系统实现方法.首先,通过磁偶极子模型分析9个阵列线圈的磁场分布规律,建立感应线圈的定位模型.其次,由STM32单片机产生的正弦信号通入发射线圈产生电磁场,接收传感器线圈耦合出感应电动势,经过放大滤波电路传入PC机进行幅值提取,导入定位方程组.最后,使用先验知识的反向传播神经网络(PKBPNN)标定初值,结合LM优化算法得出感应线圈的3维位置信息和2维方向信息,在软件定位界面显示空间定位信息.在实际实验中该系统的平均位置误差在±2 mm内,平均方向误差±0.5°内.该方法具有定位精度高、工作稳定、不易受外界干扰等优势,可应用于计算机辅助微创手术.
In order to realize the wireless precise positioning of the electromagnetic tracking system,a new five-degree-offreedom wireless electromagnetic positioning system was proposed.First,the magnetic dipole model was used to analyze the distri⁃bution rule of the magnetic field,and the positioning model of the induction coil was established.Secondly,the sinusoidal signal generated by the STM32 single-chip was passed into the transmitting coil to generate an electromagnetic field,and the receiving sensor coil was coupled to the induced electromotive force,and then passed into the PC through the amplification filter circuit to extract the amplitude and import the positioning equations.Finally,using the prior knowledge back propagation neural network(PKBPNN)to calibrate the initial value,combined with the LM optimization algorithm to obtain the 3D position information and 2D direction information of the induction coil,the spatial positioning information was displayed on the software positioning interface.In actual experiments,the average position error of the system is within±2mm,and the average direction error is within±0.5°.The system has the advantages of high positioning accuracy,stable operation,and low vulnerability to external in⁃terference.It can be applied to computer-assisted minimally invasive surgery.
作者
张超洋
阳万安
周桂宇
ZHANG Chaoyang;YANG Wanan;ZHOU Guiyu(Faculty of Intelligence Manufacturing,Yibin University,Yibin,Sichuan 644000,China;Faculty of Artificial Intelligence and Big Data,Yibin Uni-versity,Yibin,Sichuan 644000,China)
出处
《宜宾学院学报》
2023年第6期39-46,共8页
Journal of Yibin University
基金
宜宾学院培育项目(2020PY08)
四川省高等教育人才培养质量和教学改革项目(JG2021-1301)
四川省科技创新苗子工程培育项目(2021009)。
关键词
无线电磁定位
磁偶极子
LM算法
反向传播神经网络
wireless electromagnetic positioning
magnetic dipole
LM algorithm
back propagation neural network
