Abstract: Accurate assessment of human exposure to millimeter-wave (mmWave) electric fields (E-fields) has recently become critical for public health and safety. High-spatial-resolution E-field distribution is required for assessment of mmWave electromagnetic exposure according to the International Electrotechnical Commission (IEC) and the Institute of Electrical and Electronics Engineers (IEEE) (IEC/IEEE 63195-2 standard). This study proposes a generative adversarial network (GAN) integrated with field gradient loss, termed EFGraGAN, for superresolution reconstruction of mmWave E-fields. The incorporation of E-field gradient loss enables the network to learn both local field magnitudes and spatial structures, thereby enhancing the accuracy and fine structural details of reconstructed E-field maps. To improve generalization across antenna types, the training dataset is generated using plane-wave integral representation (PWIR) and randomized parametric incidence, simulating diverse field distributions. Combined with bilinear interpolation, the method achieves high-resolution reconstruction at 30 GHz and 60 GHz, meeting the requirements of the IEC/IEEE 63195-2 standard for exposure assessment. Numerical simulations show that EFGraGAN reconstructs E-field distributions in a skin phantom with a maximum mean relative error (MRE) of <9% up to 60 GHz in a 4×4 dipole array scenario, outperforming conventional interpolation and traditional GAN methods. The approach also demonstrates strong robustness to noise, enabling current measurement systems to achieve accurate and efficient evaluation of mmWave exposure.

基于梯度引导生成对抗网络与平面波积分表示的毫米波暴露评估电场超分辨率重建

[1]Chen Y, Lü X, Jia XJ, et al., 2025a. Barrier option pricing and volatility surface predicting with an extended physics-informed neural network. Expert Syst Appl, 291:128279.

[2]Chen Y, Lü X, Tian H, et al., 2025b. Physics-informed neural network for barrier option pricing in coupled financial quantitative system with varying interest rate and volatility. Eng Anal Bound Elem, 180:106457.

[3]Guo LT, Zhang Y, Li Y, 2021. An intelligent electromagnetic environment reconstruction method based on super-resolution generative adversarial network. Phys Commun, 44:101253.

[4]Guo MH, Lü X, Jin YX, et al., 2025. Extraction and reconstruction of variable-coefficient governing equations using Res-KAN integrating sparse regression. Phys D Nonl Phenom, 481:134689.

[5]IEC/IEEE, 2020. Measurement Procedure for the Assessment of Specific Absorption Rate of Human Exposure to Radio Frequency Fields from Hand-Held and Body-Mounted Wireless Communication Devices—Part 1528: Human Models, Instrumentation, and Procedures (Frequency Range of 4 MHz to 10 GHz). IEC/IEEE 62209-1528-2020.

[6]IEC/IEEE, 2022a. Assessment of Power Density of Human Exposure to Radio Frequency Fields from Wireless Devices in Close Proximity to the Head and Body (Frequency Range of 6 GHz to 300 GHz)—Part 1: Measurement Procedure. IEC/IEEE 63195-1-2022.

[7]IEC/IEEE, 2022b. Assessment of Power Density of Human Exposure to Radio Frequency Fields from Wireless Devices in Close Proximity to the Head and Body (Frequency Range of 6 GHz to 300 GHz)—Part 2: Computational Procedure. IEC/IEEE 63195-2-2022.

[8]Ledig C, Theis L, Huszár F, et al., 2017. Photo-realistic single image super-resolution using a generative adversarial network. Proc IEEE Conf on Computer Vision and Pattern Recognition, p.105-114.

[9]Lepcha DC, Goyal B, Dogra A, et al., 2023. Image super-resolution: a comprehensive review, recent trends, challenges and applications. Inform Fus, 91:230-260.

[10]Li CS, Yang L, Lu BS, et al., 2016. A reverberation chamber for rodents’ exposure to wideband radiofrequency electromagnetic fields with different small-scale fading distributions. Electromagn Biol Med, 35(1):30-39.

[11]Li K, Kodera S, Poljak D, et al., 2023. Calculated epithelial/absorbed power density for exposure from antennas at 10–90 GHz: intercomparison study using a planar skin model. IEEE Access, 11:7420-7435.

[12]Li K, Kodera S, Poljak D, et al., 2024. Spatially averaged epithelial/absorbed power density for nonplanar skin models exposed to antenna at 10–90 GHz. IEEE Access, 12:15379-15389.

[13]Lin JC, 2016. Human exposure to RF, microwave, and millimeter-wave electromagnetic radiation [health effects]. IEEE Microw Mag, 17(6):32-36.

[14]Liu ZC, Allal D, Cox M, et al., 2020. Discrepancies of measured SAR between traditional and fast measuring systems. Int J Environ Res Publ Health, 17(6):2111.

[15]Ma C, Rao YM, Cheng YA, et al., 2020. Structure-preserving super resolution with gradient guidance. Proc IEEE Conf on Computer Vision and Pattern Recognition, p.7766-7775.

[16]Mirza M, Osindero S, 2014. Conditional generative adversarial nets. https://arxiv.org/abs/1411.1784.

[17]Qamar Z, Naeem U, Khan SA, et al., 2016. Mutual coupling reduction for high-performance densely packed patch antenna arrays on finite substrate. IEEE Trans Antenn Propag, 64(5):1653-1660.

[18]Sharma P, Kumar M, Sharma HK, et al., 2024. Generative adversarial networks (GANs): introduction, taxonomy, variants, limitations, and applications. Multim Tools Appl, 83(41):88811-88858.

[19]Su YB, Lue X, Li SK, et al., 2024. Self-adaptive equation embedded neural networks for traffic flow state estimation with sparse data. Phys Fluids, 36(10):104127.

[20]Sun J, Sun J, Xu ZB, et al., 2011. Gradient profile prior and its applications in image super-resolution and enhancement. IEEE Trans Image Process, 20(6):1529-1542.

[21]Wang X, Sun LJ, Chehri A, et al., 2023. A review of GAN-based super-resolution reconstruction for optical remote sensing images. Remote Sens, 15(20):5062.

[22]Wang Z, Bovik AC, Sheikh HR, et al., 2004. Image quality assessment: from error visibility to structural similarity. IEEE Trans Image Process, 13(4):600-612.

[23]Wu TN, Hadjem A, Wong MF, et al., 2010. Whole-body new-born and young rats’ exposure assessment in a reverberating chamber operating at 2.4 GHz. Phys Med Biol, 55(6):1619.

[24]Wu TN, Shao Q, Yang L, 2013. Simplified segmented human models for whole body and localised SAR evaluation of 20 MHz to 6 GHz electromagnetic field exposures. Radiat Prot Dosim, 153(3):266-272.

[25]Wu TN, Chen YM, Li CS, 2024. Efficient evaluation of epithelial/absorbed power density by multiantenna user equipment with SAM head model. IEEE Antenn Wirel Propag Lett, 23(12):4059-4063.

[26]Xue WF, Zhang L, Mou XQ, et al., 2014. Gradient magnitude similarity deviation: a highly efficient perceptual image quality index. IEEE Trans Image Process, 23(2):684-695.

[27]Yan H, Wang ZX, Xu ZJ, et al., 2024. Research on image super-resolution reconstruction mechanism based on convolutional neural network. Proc 4^th Int Conf on Artificial Intelligence, Automation and High Performance Computing, p.142-146.

[28]Yang L, Zhang C, Chen ZY, et al., 2021. Functional and network analyses of human exposure to long-term evolution signal. Environ Sci Pollut Res, 28(5):5755-5773.

[29]Yin YH, Lü X, Li SK, et al., 2025. Graph representation learning in the ITS: car-following informed spatiotemporal network for vehicle trajectory predictions. IEEE Trans Intell Veh, 10(4):2642-2652.

Similar articles