In this letter, a dual-port antenna for multiband operation is reported. The prototype of the proposed design is fabricated using graphene-based conductive ink printed textile showing the surface resistance of 2.7 Ω/LI, and conductivity of 0.37 × 10 ⁵ S/m. The microstructural analysis of the conductive ink printed cotton fabric is done by field emission scanning electron microscopy (FESEM) and electric characteristic is extracted using a fourprobe measurement. The proposed antenna resonates at a triple-band and dual-band mode through two different feeding locations. Through Port-1, the antenna resonates only at higher order TM ₀₂ and TM ₂₀ modes, whereas through Port-2 three TM ₁₀ , TM ₀₂ , and TM ₂₀ modes are excited. Thereafter, the Coand Cross-Polarization (CP-XP) isolation of each band is investigated identifying only two useful frequency bands TM10, 2.9 GHz, and TM ₂₀ , 6 GHz. A defected ground structure (DGS) technique is deployed to improve the CP-XP isolation of TMm ₀ modes achieving CP-XP isolation better than 30 and 15 dB in both useful bands. The measured results have a fairly good agreement with simulated results.

The portable executable header (PEH) information is commonly used as a feature for malware detection systems to train and validate machine learning (ML) or deep learning (DL) classifiers. We propose to extract the deep features from the PEH information through hidden layers of a feed‐forward deep neural network (FFDNN). The extraction of deep features of hidden layers represents the dataset with a better generalization for malware detection. While feeding the deep feature of one hidden layer to the succeeding layer, the Gaussian error linear unit (GeLU) activation function is applied. The FFDNN is trained with the GeLU activation function using the deep features of individual layers as well as concatenated deep features of all hidden layers. Similarly, the ML classifiers are also trained and validated in with individual layer deep features and concatenated features. Three highly effective ML classifiers, random forest (RF), support vector machine (SVM), and k‐nearest neighbour (k‐NN) have been investigated. The performance of the proposed model is demonstrated using a statically significant large dataset. The obtained results are interesting and encouraging in terms of classification accuracy. The classification accuracy reaches 99.15% with the internal discriminative deep feature for the proposed FFDNN‐ML classifier with the GeLU activation function.

In this paper, a multiple-input–multiple-output (MIMO) antenna is reported for 5G frequency range-2 (FR-2), 28 GHz bands. The MIMO antenna is developed in multiple iterations, including single-element design, cross-polarization reduction, and mutual coupling reduction. Initially, a single-element coplanar edge feed rectangular patch antenna is designed and the E-plane cross-polarization is reduced by −13 dB by trimming the forward corners of the patch. The ground plane is truncated to improve the −3 dB half-power-beamwidth (HPBW). A multi-wavelength spiral inspired parasitic surrounding the single element antenna is loaded, and performance analysis is performed. This parasitic element is used for self-field cancelation for the MIMO configuration. Two MIMO configurations, one with linear and the second with inverted elements, are developed and investigated. The first configuration is found to have better isolation of less than −25 dB compared to the −20 dB of the second configuration. Similarly, the gain of 4.8 dBi, the bandwidth of 3 GHz, envelope correlation coefficient (ECC) of 0.01, and diversity gain (DG) of 9.99 dB are superior to the second configuration. To validate the work, one of two MIMO configurations is fabricated and good agreement is found between simulation and measurement results.