11 December 2017 in IEEE Antennas and Wireless Propagation Letters
This letter presents a 28 GHz switched five-beam antenna system based on a rectangular waveguide and a reconfigurable semiconductor circuit (RSC) with slots. Surface p-i-n diodes are used to close or open slots in order to achieve required configuration of an aperture. The RSC with a modified WR28 waveguide composes a reconfigurable radiating structure (RRS), which provides three beams directed toward 0°, 30°, and 45°. Beam switching is achieved by reconfigurable arrangement of a set of slots. Complementary RRSs connected to the electromechanical RF switch is used to generate remaining -30° and -45° beams. The gain of each generated beam is similar because different numbers of slots are opened, i.e., three, five, and nine slots for beams directed toward 0°, 30°, and 45°, respectively. Aperture of the antenna is controlled by a unit composed of high-current operational amplifiers. Measured results validate the performance of designed millimeter-wave reconfigurable antenna.
4 March 2019 in IEEE Transactions on Antennas and Propagation
This paper presents a design of a time modulated antenna array (TMAA) with substantially wider bandwidth and higher efficiency than have been already presented in the literature. The TMAA is designed for 5.4–5.8 GHz frequency band. More than 50 MHz of channel bandwidth is achieved by utilization of ultra fast microwave switches, digital delay lines with 60 ps step, and fast control circuitry. The efficiency is improved by alternate switching between array elements instead of switching between the array element and a matched load. Moreover, a mathematical derivation of optimal switching sequence is given. The beam-steering at the first sideband frequency was obtained with a 1 degree step in a range from –50 to 50 degrees. The TMAA has been evaluated by means of the radiation pattern measurements at different sideband frequencies.
2 May 2024 in IEEE Access
Real-world multiple or streaming tabular datasets, such as electronic health records from various sources and internet-of-things data generated from different devices, typically exhibit varied feature spaces depending on the datasets. Batch-mode learning with these types of datasets is often inefficient or impractical due to time constraints or privacy regulations. Therefore, an incremental-learning model capable of handling dynamically varying feature spaces, without relying on previous data, is required. To address this need, we propose a new incremental-learning method called Robust and Adaptive Incremental Learning (RAIL). RAIL comprises two core components: an incremental classifier based on naïve Bayes, and a novel adaptive feature-weighting component that utilizes feature-to-feature and feature-to-class relations. RAIL robustly handles missing and new features and adaptively assigns feature weights to improve representation capability while maintaining robustness. Based on public tabular datasets from diverse categories, we demonstrate that RAIL exhibits effective incremental-learning performance for various scenarios where the feature space regularly or arbitrarily varies. Furthermore, we validate that the proposed adaptive feature-weighting method significantly improves prediction accuracy. Additionally, we show that RAIL is more robust in preserving acquired knowledge than the existing state-of-the-art methods. Thus, our approach provides a viable incremental-learning solution for dynamic environments involving varying features.
8 November 2023 in IEEE Access
We present a theoretical and experimental analysis of the use of a reversed uneven power splitting (RUPS) technique for asymmetric Doherty power amplifiers (PAs). The RUPS technique, which has been employed to enhance the performance of GaN monolithic microwave integrated circuit (MMIC) PAs, utilizes an uneven power splitter that drives more input power into the carrier amplifier, enabling shallow class-C operation of the peaking amplifier. Nevertheless, there has been a lack of comprehensive research examining the fundamental factors that contribute to the effectiveness of the RUPS technique. We conducted numerical and experimental investigations to demonstrate that the RUPS Doherty PA exhibits significant improvements in efficiency, gain, and linearity compared to conventional Doherty PAs with even power splitting (EPS). For the experiments, the EPS and RUPS networks were developed using lumped-element directional couplers. The fabricated RUPS Doherty PA, based on a 0.25-μm GaN HEMT MMIC process, achieves superior overall performance at 2.14 GHz compared to the conventional EPS Doherty PA, without requiring any additional circuitry. The results verify that the RUPS technique can enhance the performance of asymmetric Doherty PAs.