[1] Electromagnetic-Thermal Analysis With FDTD and Physics-Informed Neural Networks

We present the coupling of the finite-difference time-domain (FDTD) method for electromagnetic field simulation, with a physics-informed neural network based solver for the heat equation. To this end, we employ a physics-informed U-Net instead of a numerical method to solve the heat equation. This approach enables the solution of general multiphysics problems with a single-physics numerical solver coupled with a neural network, overcoming the questions of accuracy and efficiency that are associated with interfacing multiphysics equations. By embedding the heat equation and its boundary conditions in the U-Net, we implement an unsupervised training methodology, which does not require the generation of ground-truth data. We test the proposed method with general 2-D coupled electromagnetic-thermal problems, demonstrating its accuracy and efficiency compared to standard finite-difference based alternatives.

[2] Deep Neural Networks for Rapid Simulation of Planar Microwave Circuits Based on their Layouts

We propose a deep learning based methodology for the rapid simulation of planar microwave circuits based on their layouts. We train convolutional neural networks to compute the scattering parameters of general, two-port circuits consisting of a metallization layer printed on a grounded dielectric substrate, by processing the metallization pattern along with the thickness and dielectric permittivity of the substrate. This approach harnesses the efficiency of convolutional neural networks with pattern recognition tasks and extends previous efforts to employ neural networks for the simulation of parameterized circuit geometries. Training is based on full-wave simulation data generated via the Finite-Difference Time-Domain (FDTD) method over a target frequency range. To accelerate the generation of such data, we build a hybrid neural network including recurrent neural network modules, to compensate numerical dispersion errors in coarse-grid FDTD.
This novel dispersion compensation scheme allows us to generate accurate FDTD training data from fast, coarse-grid simulations.

[3] 2D Electromagnetic Solver Based on Deep Learning Technique

Although the deep learning technique has been introduced into computational physics in recent years, the feasibility of applying it to solve electromagnetic (EM) scattering field from arbitrary scatters remains open. In this article, the convolutional neural network (CNN) has been employed to predict the EM field scattered by complex geometries under plane-wave illumination. The 2-D finite-difference frequency-domain (FDFD) algorithm, wrapped by a module to randomly generate complex scatters from basic geometries, is employed to produce training data for the network. The multichannel end-to-end CNN is modified and combined with residual architecture and skip connection, which can speed up convergence and optimize network performance, to form the EM-net. The well-trained EM-net has good performance in this problem since it is compatible with different shapes, multiple kinds of materials, and different propagation directions of the incident waves. The effectiveness of the proposed EM-net has been validated by numerical experiments, and the average numerical error can be as small as 1.23%. Meanwhile, its speedup ratio over the FDFD method is as large as 2000.

[4] Fabriccio: Touchless Gestural Input on Interactive Fabrics

We present Fabriccio, a touchless gesture sensing technique developed for interactive fabrics using Doppler motion sensing. Our prototype was developed using a pair of loop antennas (one for transmitting and the other for receiving), made of conductive thread that was sewn onto a fabric substrate. The antenna type, configuration, transmission lines, and operating frequency were carefully chosen to balance the complexity of the fabrication process and the sensitivity of our system for touchless hand gestures, performed at a 10 cm distance. Through a ten-participant study, we evaluated the performance of our proposed sensing technique across 11 touchless gestures as well as 1 touch gesture. The study result yielded a 92.8% cross-validation accuracy and 85.2% leave-one-session-out accuracy. We conclude by presenting several applications to demonstrate the unique interactions enabled by our technique on soft objects.

[5] ADS-B Message Authentication Using Features of Signal in Transition Regions

Automatic Dependent Surveillance - Broadcast (ADS-B) is an essential communication protocol used in modern air traffic control. However, lacking security measures such as authentication and encryption, ADS-B messages can be forged and modified easily by malicious attackers. This paper addresses the problem of authenticating an ADS-B message by means of specific emitter identification (SEI), employing the unintentional modulation on pulse (UMOP). A complete identification system is presented, including data acquisition, feature extraction and classification. In order to create the feature vector for classification, the transition region in a pulse is first delimited and then used for extraction of UMOP features through Hilbert-Huang Transform. The performance of the method is tested by real signal from 30 ADS-B transmitters. Our proposed method is shown to achieve an recognition ratio of over 94%, and performs better under low SNR compared with two previous techniques. It is demonstrated that the SEI technique proposed can be used as an additional tool to enhance the security of ADS-B protocol.