In this contribution, we analyze machine learning-assisted solutions to tackle real-time fault detection in large-scale active phased array antennas. The challenge of integrating the circuit component nonlinearities and mutual coupling effects in fault-finding methodologies is addressed. A novel machine learning (ML) solution based on an array theory-enhanced neural network (NN) is proposed. To address the practical constraints of large array measurements, sparse far-field (FF) measurements are considered. The method is experimentally verified by applying it to fault detection in a 64-element planar phased array prototype operating at 26 GHz with far-field measurements collected by a fixed high-speed multi-probe pattern acquisition setup, the Antenna Dome. Significant improvement in fault prediction performance over a conventional Genetic Algorithm (GA) based heuristic approach with improvements of up to 40%, 25%, and 20% in predicting 8-, 4-, and 2-element faults, respectively, is demonstrated. The robustness of the proposed performance with respect to the number of 3D spatial field sampling points is shown, offering efficient diagnostics with in-field pattern sampling compatibility and low hardware complexity.@en

The sunflower array topology concept is introduced, for the first time, to the constrained infinitesimal dipole modeling (IDM) technique to increase the computational efficiency and reduce the modeling errors. The concept is applied to embedded element pattern predictions via matrix inversion. A novel study on the impact of the type and orientation of the dipoles on the IDM performance in pattern mean square error (MSE) and stability against noise (linked to the matrix condition number) is conducted. A 5 by 5 patch antenna array modeled with 81 dipoles is used for demonstration. It is shown that using magnetic dipoles (oriented in the direction of a radiating edge of the patch) in IDM yields the optimal performance. Besides, the sunflower topology significantly lowers the MSE (by 5 dB, on average), while reducing the condition number by a factor of 10.@en

A novel dual-linear polarized active aperiodic millimeter-wave (26 GHz band) phased array is prototyped. Four-channel integrated circuits are used for beamforming. An asymmetrical signal distribution and antenna feeding network is designed. A heuristic post-calibration method is applied in beam steering. The measurements show that, as compared to the benchmark array with a conventional square-grid layout, the proposed aperiodic topology can significantly decrease the peak sidelobe level, the extent of which depends on the polarization and scan angle. This comes at the expense of larger array size, higher feed network losses and increased cross-polarization levels.

@en

A novel ensemble prediction technique is introduced to enhance the accuracy of far-field embedded element pattern (EEP) prediction under mutual coupling (MC) effects, while relaxing the training data size challenge in neural network (NN)-based algorithms. The proposed method integrates a two-stage NN for direct EEP prediction from full-wave simulated pattern data in spherical coordinates with a fully connected NN for the prediction of excitation coefficients of an array of infinitesimal dipoles, approximating the full-wave simulated EEPs via constrained infinitesimal dipole modeling (IDM). Quasi-randomly distributed five-element pin-fed S-band patch antenna arrays are used for demonstration purpose. It is shown that, for a large-sized (3500 topologies) and relatively small-sized (1500 topologies) dataset, incorporating IDM-NN with the benchmarked direct EEP-NN in an ensemble technique increases the pattern prediction accuracy by 11% and 60% on average, respectively.

@en

This paper examines how training data affects machine learning-assisted antenna pattern prediction under mutual coupling. For demonstration, a neural network-based approach is used to predict the embedded pattern of a central patch antenna element near randomly distributed patches. It is shown that when the full-wave simulated dataset size is excessively reduced, the high prediction error in the validation set may become a critical issue. Maintaining sufficient accuracy in pattern prediction with a relatively small dataset remains an open challenge.@en

Efficient prediction of embedded element patterns (EEPs) is including the mutual coupling (MC) effects in the optimization of irregular planar arrays is studied for the first time in the literature. An ANN-based methodology is used to predict the pattern of each element in the whole visible space for a flexible planar array topology in milliseconds. The technique is proposed is validated on a 4-element planar non-uniform sub-array structure. Excellent accuracy on the EEP prediction while providing great efficiency in computational time and load in comparison to the full-wave simulations is demonstrated.@en

The effects of multipath on the statistical cell-edge user service quality is for the first time investigated for mm-wave multi-user communication systems. The focus is given on setting the user spacing constraints and the transmit array topology via thinning, which can be used to enhance wireless security or decrease analog/digital complexity. A hybrid line-of-sight/non-line-of-sight channel is created by using a statistical model following the communication standards. The multipath signal components are included in the model by using non-coherent or coherent modes of operation. It is shown in simulation that selection, by the medium access control layer, of large angular spacings between the simultaneously served users and application of antenna array thinning at the array edges improves the system performance.@en

A heuristic calibration-before-operation method based on near-field-measurements of integrated circuit (IC) responses is proposed and tested for an active 64-element 5G millimeter-wave (24-28 GHz) phased array. The IC phase shifter and variable gain settings are digitally tuned to minimize the relative differences in complex transmission coefficients on broadside at each element. It is shown that, by means of calibration, it is possible to increase the array gain, decrease the peak side lobe level and remove the beam pointing errors.

@en

The optimization of mode excitation coefficients in linear periodic arrays of multi-mode antenna elements is studied for grating lobe reduction. A novel beamforming architecture is proposed with a new optimization problem based on equi-amplitude element excitations for optimal power efficiency. The capabilities of the proposed synthesis approach on suppressing the grating lobe for wide scan angles, and on maintaining the peak gain at the steering angle are analyzed. A 16-element 0.7-wavelength spaced array of dual-mode circular patch antenna elements is used for demonstration purposes. It is shown that a good performance trade-off is achieved when the excitation amplitude of the high order mode is restricted to a sufficiently large value. The ratio of the peak gain outside the main lobe to the gain at the angle of steering is reduced up to about −15 dB and −14 dB for scanning towards 30 and 45 degrees off-broadside, respectively.@en