No abstract

No abstract

This paper presents a wide-angle scanning phased array antenna using high gain pattern reconfigurable antenna (PRA) elements. Using PRA elements is an attractive solution for wide-angle scanning phased array antennas because the scanning range can be divided into several subspaces. To achieve the desired scanning performance, some characteristics of the PRA element such as the number of switching modes, tilt angle, and maximum half-power beamwidth (HPBW) are required. We analyzed the required characteristics of the PRA element according to the target scanning range and element spacing, and presented a PRA element design guideline for phased array antennas. In accordance with the guideline, the scanning range was set as ±70° and a high gain PRA element with three reconfigurable patterns was used to compose an 8x1 array antenna with 0.9 λ0 spacing. After analyzing whether the active element patterns meet the guideline, the array antenna was fabricated and measured to demonstrate the scanning performance. The fabricated array can scan its beam from -70° to 70° by dividing the scanning range into three subspaces. It shows that even if the array antenna has large element spacing, the desired scanning performance can be obtained using the elements designed under the guideline.

Semi-supervised learning and continuous learning are fundamental paradigms for human-level intelligence. To deal with real-world problems where labels are rarely given and the opportunity to access the same data is limited, it is necessary to apply these two paradigms in a joined fashion. In this paper, we propose Label Propagation Adaptive Resonance Theory (LPART) for semi-supervised continuous learning. LPART uses an online label propagation mechanism to perform classification and gradually improves its accuracy as the observed data accumulates. We evaluated the proposed model on visual (MNIST, SVHN, CIFAR-10) and audio (NSynth) datasets by adjusting the ratio of the labeled and unlabeled data. The accuracies are much higher when both labeled and unlabeled data are used, demonstrating the significant advantage of LPART in environments where the data labels are scarce.

No abstract

In this paper, we first demonstrate a metasurface that offers spectral filter performances not yet reported for a single-layer metasurface because of the limited number of geometric parameters. The configuration of the metasurface is determined by a conformational space annealing algorithm, which is a key toward being able to design for a desired frequency, bandwidth, and response type for various wireless communication devices, including the fifth-generation (5G) millimeter-wave communication devices. As a proof-of-concept, we fabricate a very thin (50 µm) single-layer metasurface combined with a 5G antenna array to experimentally demonstrate high frequency-selectivity with stable frequency responses with oblique incidences and polarizations. The gain of the 4 × 4 patch antenna array with the metasurface bandpass filter was reduced by only 0.48 dB compared with that of the 4 × 4 patch antenna array itself at 27.5 GHz, whereas the gain of the 4 × 4 patch antenna array with the metasurface bandstop was reduced by 20 dB compared with that of the 4 × 4 patch antenna array itself at 28.5 GHz.

This paper proposed a high gain spherical dielectric resonator antenna (DRA) operating on higher-order mode excited by microstrip patch. A microstrip patch was used to excite TE n01 mode on a dielectric sphere. The excited dielectric sphere operates as higher-order mode spherical DRA with high gain. Impedance matching method for conventional microstrip patch could be applied to the proposed antenna. Size minimization method for the excitation microstrip patch was also described on this paper. A prototype antenna operating on TE 301 mode at 5.8 GHz was fabricated by a ceramic dielectric material, which dielectric constant is 13, and showed the peak gain of 9.03 dBi.

No abstract