本論文提出一個基於類神經網路與互補式開口共振腔(complementary split-ring planar resonator, CSRR)來量測三層待測物(Material under the test, MUT)之方法，以往使用之微波平面式共振腔量測介電值方法多藉由微擾理論，CSRR已發展出可準確量測雙層材料之方法，結合等效電路模型建立可準確分析雙層材料之介電常數及厚度之公式，根據其公式使用多個頻率同時量測可分析同等數量之目標參數，藉由多頻率之間求解聯立方程將其包含的參數解出。本論文基於其方法改為藉由類神經網路分析量測的資訊，可使用更多頻率同時量測多個參數且無需重新建立公式。而CSRR量測多層材料因其穿透性質，使得不同層待測物之間所造成頻率偏移有極大差距，神經網路可針對單一目標進行分析，將不同層之介電值分別計算，達到將問題簡化的效果。實際量測具0.1mm空氣間隙之單層待測物，平均介電常數與厚度誤差分別為11.83%與6.59%。量測具0.2mm空氣間隙之單層待測物時，平均介電常數與厚度誤差分別為12.30%與6.51%。量測各層厚度固定之三層待測物，平均介電常數誤差ε_1為2.74%、ε_2為6.15%、ε_3為35.62%。

This paper proposes a method based on neural network and complementary split-ring planar resonator (CSRR) to measure the three-layer material under the test (MUT). In the past, the microwave planar resonator measurement method used to measure the dielectric value mostly based on the perturbation theory. CSRR method that can accurately measure the double-layer material has been developed, combined with the equivalent circuit model to establish a formula that can accurately analyze the permittivity and thickness of the double-layer material. According to its formula, using multiple frequencies to measure at the same time can analyze the same number of target parameters, and solve the included parameters by solving simultaneous equations between multiple frequencies. This paper is based on its method to analyze the measured information through a neural network, which can use more frequencies to measure multiple parameters at the same time without the need to re-create the formula. When CSRR is used to measure multi-layer materials, due to its penetrating properties, there is a huge difference in frequency shifting between different layers of MUTs. The neural network can analyze for a single target and calculate the permittivity of different layers separately to achieve the effect of simplifying the problem. In the measurement of a single-layer MUT with a 0.1mm air gap, the average permittivity and thickness error are 11.83% and 6.59%, respectively. When measuring a single-layer MUT with a 0.2mm air gap, the average permittivity and thickness error are 12.30% and 6.51%, respectively. Measure the three-layer MUT with a fixed thickness of each layer, and the average permittivity error ε_1 is 2.74%, ε_2 is 6.15%, and ε_3 is 35.62%.

[1] S. Bakhtiari, N. Qaddoumi, S. I. Ganchev, and R. Zoughi, "Microwave noncontact examination of disbond and thickness variation in stratified composite media," IEEE Transactions on Microwave Theory and Techniques, vol. 42, no. 3, pp. 389-395, 1994.
[2] D. Misra, M. Chabbra, B. R. Epstein, M. Microtznik, and K. R. Foster, "Noninvasive electrical characterization of materials at microwave frequencies using an open-ended coaxial line: Test of an improved calibration technique," IEEE transactions on microwave theory and techniques, vol. 38, no. 1, pp. 8-14, 1990.
[3] Y. Li and N. Bowler, "Resonant frequency of a rectangular patch sensor covered with multilayered dielectric structures," IEEE Transactions on Antennas and Propagation, vol. 58, no. 6, pp. 1883-1889, 2010.
[4] A. V. Mamishev, S. R. Cantrell, Y. Du, B. C. Lesieutre, and M. Zahn, "Uncertainty in multiple penetration depth fringing electric field sensor measurements," IEEE Transactions on instrumentation and measurement, vol. 51, no. 6, pp. 1192-1199, 2002.
[5] H.-C. Wang, A. Zyuzin, and A. V. Mamishev, "Measurement of coating thickness and loading using concentric fringing electric field sensors," IEEE Sensors Journal, vol. 14, no. 1, pp. 68-78, 2013.
[6] C.-S. Lee and C.-L. Yang, "Single-compound complementary split-ring resonator for simultaneously measuring the permittivity and thickness of dual-layer dielectric materials," IEEE Transactions on Microwave Theory and Techniques, vol. 63, no. 6, pp. 2010-2023, 2015.
[7] C.-L. Yang, C.-S. Lee, K.-W. Chen, and K.-Z. Chen, "Noncontact measurement of complex permittivity and thickness by using planar resonators," IEEE Transactions on Microwave Theory and Techniques, vol. 64, no. 1, pp. 247-257, 2015.
[8] E. C. Green, B. R. Jean, and R. J. Marks, "Artificial neural network analysis of microwave spectrometry on pulp stock: Determination of consistency and conductivity," IEEE transactions on instrumentation and measurement, vol. 55, no. 6, pp. 2132-2135, 2006.
[9] Y. Yan, L. Xu, and P. Lee, "Mass flow measurement of fine particles in a pneumatic suspension using electrostatic sensing and neural network techniques," IEEE Transactions on instrumentation and measurement, vol. 55, no. 6, pp. 2330-2334, 2006.
[10] K. Hornik, "Approximation capabilities of multilayer feedforward networks," Neural networks, vol. 4, no. 2, pp. 251-257, 1991.
[11] M. Leshno, V. Y. Lin, A. Pinkus, and S. Schocken, "Multilayer feedforward networks with a nonpolynomial activation function can approximate any function," Neural networks, vol. 6, no. 6, pp. 861-867, 1993.
[12] D. Ghodgaonkar, V. Varadan, and V. Varadan, "Free-space measurement of complex permittivity and complex permeability of magnetic materials at microwave frequencies," IEEE Transactions on instrumentation and measurement, vol. 39, no. 2, pp. 387-394, 1990.
[13] D. M. Hagl, D. Popovic, S. C. Hagness, J. H. Booske, and M. Okoniewski, "Sensing volume of open-ended coaxial probes for dielectric characterization of breast tissue at microwave frequencies," IEEE Transactions on Microwave Theory and Techniques, vol. 51, no. 4, pp. 1194-1206, 2003.
[14] Y. Li, N. Bowler, and D. B. Johnson, "A resonant microwave patch sensor for detection of layer thickness or permittivity variations in multilayered dielectric structures," IEEE Sensors Journal, vol. 11, no. 1, pp. 5-15, 2010.
[15] C.-S. Lee and C.-L. Yang, "Thickness and permittivity measurement in multi-layered dielectric structures using complementary split-ring resonators," IEEE Sensors Journal, vol. 14, no. 3, pp. 695-700, 2013.
[16] A. M. Albishi and O. M. Ramahi, "Microwaves-based high sensitivity sensors for crack detection in metallic materials," IEEE Transactions on Microwave Theory and Techniques, vol. 65, no. 5, pp. 1864-1872, 2017.
[17] J. Schmidhuber, "Deep learning in neural networks: An overview," Neural networks, vol. 61, pp. 85-117, 2015.
[18] H. Lee, P. Pham, Y. Largman, and A. Y. Ng, "Unsupervised feature learning for audio classification using convolutional deep belief networks," in Advances in neural information processing systems, 2009, pp. 1096-1104.
[19] E. K. Murphy and V. V. Yakovlev, "RBF network optimization of complex microwave systems represented by small FDTD modeling data sets," IEEE transactions on microwave theory and techniques, vol. 54, no. 7, pp. 3069-3083, 2006.
[20] H. Kabir, Y. Wang, M. Yu, and Q.-J. Zhang, "High-dimensional neural-network technique and applications to microwave filter modeling," IEEE Transactions on Microwave Theory and Techniques, vol. 58, no. 1, pp. 145-156, 2009.
[21] L. Xu, W. Zhou, X. Li, and S. Tang, "Wet gas metering using a revised Venturi meter and soft-computing approximation techniques," IEEE transactions on Instrumentation and Measurement, vol. 60, no. 3, pp. 947-956, 2010.
[22] D. M. Pozar, Microwave engineering. John Wiley & Sons, 2009.
[23] M. S. Boybay and O. M. Ramahi, "Material characterization using complementary split-ring resonators," IEEE Transactions on instrumentation and Measurement, vol. 61, no. 11, pp. 3039-3046, 2012.
[24] A. Ebrahimi, W. Withayachumnankul, S. Al-Sarawi, and D. Abbott, "High-sensitivity metamaterial-inspired sensor for microfluidic dielectric characterization," IEEE Sensors Journal, vol. 14, no. 5, pp. 1345-1351, 2013.
[25] J. D. Baena et al., "Equivalent-circuit models for split-ring resonators and complementary split-ring resonators coupled to planar transmission lines," IEEE Transactions on Microwave Theory and Techniques, vol. 53, no. 4, pp. 1451-1461, 2005.
[26] N. Qian, "On the momentum term in gradient descent learning algorithms," Neural networks, vol. 12, no. 1, pp. 145-151, 1999.
[27] 葉怡成, "類神經網路模式應用與實做," ed: 臺北市: 儒林出版社, 2004.
[28] 葉怡成, 應用類神經網路. 儒林, 2004.
[29] N.-Y. Liang, G.-B. Huang, P. Saratchandran, and N. Sundararajan, "A fast and accurate online sequential learning algorithm for feedforward networks," IEEE Transactions on neural networks, vol. 17, no. 6, pp. 1411-1423, 2006.
[30] N. Baba, Y. Mogami, M. Kohzaki, Y. Shiraishi, and Y. Yoshida, "A hybrid algorithm for finding the global minimum of error function of neural networks and its applications," Neural Networks, vol. 7, no. 8, pp. 1253-1265, 1994.
[31] S. Razavi and B. A. Tolson, "A new formulation for feedforward neural networks," IEEE Transactions on neural networks, vol. 22, no. 10, pp. 1588-1598, 2011.
[32] H. Zhang, W. Wu, F. Liu, and M. Yao, "Boundedness and convergence of online gradient method with penalty for feedforward neural networks," IEEE transactions on neural networks, vol. 20, no. 6, pp. 1050-1054, 2009.
[33] J.-T. Tsai, J.-H. Chou, and T.-K. Liu, "Tuning the structure and parameters of a neural network by using hybrid Taguchi-genetic algorithm," IEEE Transactions on Neural Networks, vol. 17, no. 1, pp. 69-80, 2006.
[34] Y. Liu, J. A. Starzyk, and Z. Zhu, "Optimized approximation algorithm in neural networks without overfitting," IEEE transactions on neural networks, vol. 19, no. 6, pp. 983-995, 2008.