30 GHz band double voltage rectifier MMIC with the 0.1 μm E-pHEMT gated anode diode

IEICE Technical Report; IEICE Tech. Rep.

Published On 2021/12/9

(in English) In this paper, the 30 GHz band double voltage rectifier MMIC with the 0.1 μm E-pHEMT is demonstrated for the millimeter wave wireless power transfer system. The gated anode diode (GAD) that is the 0.1 μm E-pHEMT with the connected gate-drain electrodes as the anode terminate is employed as a rectifier diode. To improve the rectification efficiency, the finger width of the E-pHEMT is optimized for the higher cut-off frequency. As the result, the GAD has the cut-off frequency of 810 GHz. The developed rectifier MMIC achieves rectification efficiency of 59.6% at an input power of 18.5 dBm (0.07 W) that is the top performance in 30 GHz band.

Journal

IEICE Technical Report; IEICE Tech. Rep.

Published On

2021/12/9

Volume

121

Issue

303

Page

31-36

Authors

Naoki Sakai

Naoki Sakai

Kanazawa Institute of Technology

Position

H-Index(all)

8

H-Index(since 2020)

6

I-10 Index(all)

0

I-10 Index(since 2020)

0

Citation(all)

0

Citation(since 2020)

0

Cited By

0

Research Interests

Microwave circuit

Wireless power transfer

Rectifier

University Profile Page

Other Articles from authors

Naoki Sakai

Naoki Sakai

Kanazawa Institute of Technology

An Integrated Planar Guanella Balun with the Quasi-Coaxial Structure on GaAs Substrate

This paper proposes a planar Guanella balun with the quasi-coaxial structure. This is the first report of the Guanella balun fabricated on a GaAs substrate. The quasi-coaxial structure is configured with metal layers on three thin dielectric layers of a few micrometers thickness. These layers are laminated on the GaAs substrate. The metal on the second layer (the second metal) is an inner conductor, which is surrounded by the first and third metals correspond to an outer conductor. The first and third metals are split lines to widen the inner conductor. The fabricated Guanella balun has a spiral shape, and its size is by . The fabricated Guanella balun shows less than of amplitude imbalance and 7.3 degrees of phase imbalance from 5 GHz to 15GHz. The fabricated Guanella balun is realized with half elements of a Marchand balun and achieves good performance with small size.

Naoki Sakai

Naoki Sakai

Kanazawa Institute of Technology

IEEE Antennas and Wireless Propagation Letters

Graded Index Fiber-Inspired 3-D Printed Surface Focusing Porous Dielectric Structure With GaAs MMIC Rectenna Towards Millimeter Wave Wireless Power Transfer Application

Borrowing the idea of the graded-index (GRIN) fiber, we propose an all-dielectric porous structure with a unique surface focusing feature to improve the performance of the rectenna for millimeter-wave (mm-Wave) wireless power transfer (WPT) applications in this letter. The on- and off-axis focusing properties of the proposed structure are analyzed by the full-wave electromagnetic simulation. Moreover, benefitting from the low-cost 3-D printing technique, the designed GRIN structure prototype with different pore densities of the polylactic acid material is fabricated. The near-field phase transform experiment at 28 GHz was carried out to verify the excellent focusing performance of the fabricated structure prototype. Besides, the characteristics of multifocal spots on the surface of the proposed structure are further confirmed by the mm-Wave WPT experiments with a GaAs monolithic microwave integrated circuit rectenna …

Naoki Sakai

Naoki Sakai

Kanazawa Institute of Technology

IEICE Conferences Archives

高効率・大電力レクテナ

本報告では,金沢工業大学における高効率・大電力レクテナの研究状況について述べる.ドローンなどのUAVへの無線給電を行う場合,限られた受電面積から所望の直流電力を取り出すために,レクテナの大電力化,高効率化が必要となる.本稿ではその手法を述べる.

Naoki Sakai

Naoki Sakai

Kanazawa Institute of Technology

IEICE Conferences Archives

GaAs E-pHEMT gated anode diode を用いた 5.75 GHz 帯 1W レクテナ

空間伝送型ワイヤレス電力伝送の法制度化が進み,実用化に向けた開発が行われている.その中で,金沢工大ではダイポールアンテナとE-pHEMT gate anode diode(GAD)による整流素子を直接整合し,整合回路の損失を抑制した高効率レクテナを報告している[1].本報告は,レクテナを低姿勢化させる上で有利なパッチアンテナ用い,[1]と同様の直接整合構成の1Wレクテナを開発したので,その結果を報告する.[1] N. Sakai, K. Itoh et.al, "5.8GHz Band 10W Rectenna with GaAs E-pHEMT Gated Anode Diode on the Aluminum Nitride Antenna for Thermal Dispersion," IEEE MTT-S IMS2023, pp.1003 – 1005, 2023-8.

Naoki Sakai

Naoki Sakai

Kanazawa Institute of Technology

IEICE Conferences Archives

塗布型カーボンナノチューブ薄膜トランジスタによる Gated anode diode を用いる 920MHz 帯レクテナ

本報告では薄形で安価な塗布プロセスが適用可能なカーボンナノチューブ薄膜トランジスタ (carbon nanotube thin –film transistor: CNT-TFT) による920MHz帯レクテナについて述べる.ここでは高い移動度が得られるCNT-TFTによりgated anode diode (GAD)を構成し,整流素子としている.このCNT-TFT GADは強い周波数分散特性を有するため,直流特性と高周波特性を用いモデル化している.このCNT-TFT GADにより倍電圧整流器を構成し,さらに高利得特性を有するデュアルループアンテナと接続し,レクテナを構成している.試作した920MHz帯レクテナの大きさは46mm x 92mmであり,基準アンテナへの受電電力0dBmの条件で0.41Vの出力電圧を確認した.これはEIRP 2Wの送信機から1.48m離れた場所での受電に相当する測定条件である.

Naoki Sakai

Naoki Sakai

Kanazawa Institute of Technology

IEICE Conferences Archives

倍電圧整流器 IC の高調波輻射量の評価結果

本報告では,倍電圧整流器ICの高調波輻射量の評価結果を報告する.一般的な倍電圧整流器の等価回路をもとにしたレイアウト(Layout A)と平滑用MIMキャパシタの配置を変え,対称構造にしたレイアウト(Layout B)を比較する.その結果として,Layout Aは入力電力25.6dBmのとき,整流効率67.2%,出力電圧8.6Vを得ている.Layout Bも同等である.また,2倍波電力への変換損の測定値は,Layout Aでは-38.5dBc@Pin = 25.6dBm,Layout Bでは-41.3dBc@Pin = 25.6dBmである.レイアウトを対称とすることにより2倍波の高調波輻射レベルが2.8dB抑制することができる.

Naoki Sakai

Naoki Sakai

Kanazawa Institute of Technology

IEICE Conferences Archives

量子ビット駆動用 X 帯偶高調波リングミクサ MMIC の特性

量子ビット駆動用マイクロ波信号源では,演算の判定誤りを抑制するため低スプリアス特性が要求される.これに適用するため,アンチパラレルダイオードペア(APDP)を用いる偶高調波ミクサを検討している.ここでは X帯偶高調波リングミクサMMICの評価結果を示す.

Naoki Sakai

Naoki Sakai

Kanazawa Institute of Technology

IEICE Conferences Archives

インピーダンス変成微小ループアンテナを用いる 28GHz 帯 GaAs 高感度レクテナ MMIC

Beyond 5Gに向け,ミリ波における通信と無線電力伝送(WPT)の融合システムを検討している. WPTのサービスエリア拡大のためには高感度レクテナが必要となる.本報告では,小形に高インピーダンス特性を実現できる微小ループアンテナ(SLA)を適用した28GHz帯GaAs高感度レクテナMMICの試作結果を示す.

Naoki Sakai

Naoki Sakai

Kanazawa Institute of Technology

IEICE Conferences Archives

整合回路付き高周波倍電圧整流回路の整流効率の理論式

本報告は整合回路を含む高効率倍電圧整流回路の設計に向けて,整合回路を装荷した倍電流整流回路の整流効率の理論式を導出する.

Naoki Sakai

Naoki Sakai

Kanazawa Institute of Technology

IEICE Conferences Archives

準同軸構造グァネラバラン GaAs IC の試作結果

平衡-不平衡変換を行う高周波回路としてマーチャントバランが良く用いられている.ただし,2つのλ/4結合線路で構成されるため小型化の課題がある.ここではバランの小型化に向けて,1つのλ/4同軸伝送線路で構成可能なグァネラバランを準同軸構造によりGaAs IC上に形成した.試作の結果,4~15GHzで,3dBの分配損とI/O パッドの損失を含む挿入損失は6.3±1.4 dB,振幅インバランスは0.7dB以下,位相インバランスは10.9度以下が得られた.I/O パッドを除くグァネラバランのコアサイズは425μm×345μmである.併せて作製したマーチャントバランと比較して同等の性能で22.2%の小型化が図れた.

Naoki Sakai

Naoki Sakai

Kanazawa Institute of Technology

Novel Energy Harvesting and SWIPT System at 28 GHz with a Simple Phased Array

The realization of future radiative wireless power transfer products necessitates a simple and low-cost phased array with higher frequency to increase the beam efficiency. Further, simultaneous wireless power and information transfer (SWIPT) are also expected. This study developed a simple phased array at 28 GHz while considering energy harvesting from 5G radio waves. The number of antenna elements in the phased array with parasitic elements were reduced and novel, simple, and low-cost phase shifters were employed. Consequently, beam steering and SWIPT experiments were conducted using the developed phased array. This study expanded upon the developed phased array and the experimental results in Japan.

Naoki Sakai

Naoki Sakai

Kanazawa Institute of Technology

920 MHz band rectenna with the CNT-TFT GAD

This paper presents the first demonstration of rectification with the carbon nanotube thin film transistor (CNT-TFT) for wireless power transfer (WPT) in the 920 MHz band. The CNT-TFT that is a coated TFT using semiconducting single-walled carbon nanotubes has advantage of low-cost fabrication and planner integration with an antenna on a thin film. On the other, reported frequency ranges were within HF bands due to their wider channel size. For applying the CNT-TFT to GHz-order rectification, the gated anode diode (GAD) is employed and is characterized for its modeling. As the results, practical operations of 920 MHz band rectenna with the CNT-TFT GAD and the dual loop antenna can be confirmed in experimental investigations.

Naoki Sakai

Naoki Sakai

Kanazawa Institute of Technology

電子情報通信学会論文誌 C

AMC 動作基板上の誘導性高インピーダンスモノポールアンテナを用いる 5.8 GHz 帯大電力レクテナ

本論文では薄形レクテナを地導体上に配置するために,AMC(Artificial Magnetic Conductor)動作基板上に配置した誘導性高インピーダンスモノポールアンテナを用いる5.8 GHz帯大電力レクテナの設計,試作結果を示す.ここではAMCから想起され,同様の動作をするグリッド導体を有する薄形基板であるAMC動作基板を用いることにより,地導体から2 mmの高さでモノポールアンテナを実現する.これを誘導性かつ高インピーダンス動作するよう設計し,整流用ダイオードとして用いる0.5 μm GaAs E-pHEMTによるGated Anode Diode(GAD)と共役整合する.整合回路を不要とすることで,高効率化を図っている.試作したレクテナの大きさは15 mm × 17.2 mm × 2.1 mmであり,地導体・金属上に実装可能である.測定の結果,整流器への入力電力31.8 dBm(1.51 W)に対し整流効率84.4%,出力電圧37.7 Vの良好な結果が得られている.

Naoki Sakai

Naoki Sakai

Kanazawa Institute of Technology

5.8 GHz band 10 W rectenna with GaAs E-pHEMT gated anode diode on the aluminum nitride antenna for thermal dispersion

In this paper, the 5.8 GHz band 10 W rectenna is demonstrated. The GaAs bridge rectifier IC with GaAs E-pHEMT gated anode diodes (GADs) is employed to obtain input power of 10 W. The IC is directly connected with the inductive high-impedance antenna to reduce circuits’ loss. Furthermore, the antenna is implemented on the aluminum nitride (AlN) substrate for thermal dispersion. With migrated circuit functionalities, simulated radiation efficiency of the antenna is 99.2 %. Measured rectification efficiency of the rectifier is 83.7 % at input power of 10 W. This is the top performance among 10 W class rectifies

Naoki Sakai

Naoki Sakai

Kanazawa Institute of Technology

IEICE Technical Report; IEICE Tech. Rep.

5.8 GHz band 10 W rectenna on the aluminum nitride antenna for thermal dispersion

(in English) This paper presents the 5.8 GHz band 10W rectenna with GaAs E-pHEMT gated anode diode (GAD) on the aluminum nitride (AlN) antenna. The rectifier IC on the prototyped AlN antenna achieved a rectification efficiency of 83.7% at an input power of 10.5 W while maintaining a GAD junction temperature of 48 degrees. These results demonstrate that the rectenna can continuously receive RF power of 10 W with high rectification efficiency. Furthermore, we demonstrate a 3x3 rectenna array with the prototyped rectenna in a microwave power transfer (MPT) system. The MPT system with the rectenna array could transmit a DC power of 43.4 W, allowing a drone to fly continuously for ten minutes.

Naoki Sakai

Naoki Sakai

Kanazawa Institute of Technology

IEICE Conferences Archives

5.8 GHz 帯 10W 級倍電圧整流器 MMIC

マイクロ波での無線電力伝送(MPT)が法制化され,その実用化が進められている.筆者らは,MPT用レクテナの大電力化のため,GaAs E-pHEMTによるGated anode diode (GAD)を用いる大電力レクテナの検討を行っている.本報告では5.8GHz帯10W級倍電圧整流器MMICの評価結果を示す.

Naoki Sakai

Naoki Sakai

Kanazawa Institute of Technology

IEICE Conferences Archives

ダイオードの閾値と降伏電圧を考慮した倍電流整流回路の解析

無線電力伝送システムにおいて,整流回路は受電した高周波電力を直流に変換し負荷へ給電するために欠かせない回路である.倍電流整流回路はモーターや電池などの低電圧・大電流で動作する負荷へ直流電力を供給するのに向いている.過去の発表ではダイオードを理想的なスイッチと仮定し,倍電流整流回路の動作および理論効率をまとめている.本稿では閾値電圧ならびにブレイクダウン(降伏)電圧を含んだダイオードモデルを用い,倍電流整流回路の動作について報告する.

Naoki Sakai

Naoki Sakai

Kanazawa Institute of Technology

IEICE Conferences Archives

GaAs E-pHEMT gated anode diode を用いる 5.8 GHz 帯 5W レクテナ

大電力化に適するGaAs E-pHEMTによるgated anode diode(GAD)を用いる5.8GHz帯5Wレクテナの試作結果を示す.入力電力Pin=37dBm(5W)のときに整流効率=85.2%,出力電圧Vout=35.8Vである.

Naoki Sakai

Naoki Sakai

Kanazawa Institute of Technology

IEICE Technical Report; IEICE Tech. Rep.

Quasi millimeter wave high-power rectifier with GaAs E-pHEMT GADs

(in English) In this report, quasi-millimeter wave 1 W-class double-voltage rectifier MMIC with 0.18 μm GaAs E-PHEMT gated anode diodes (GADs) is presented. At first, high current characteristic of the 0.18 μm GaAs E-PHEMT GAD is indicated in its characterization, compared with the GaAs SBD. As the result, terminal capacitance of the GAD is made lower than that of the GaAs SBD, matching loss can be reduced especially in high-power rectification. In the experimental investigation of the developed double voltage rectifier MMIC, rectification efficiency of 62% and output voltage of 32.9 V are obtained at an input power of 30 dBm. This is the highest input power in reported rectifiers from 20 to 40 GHz band. And measured rectification efficiency is extreme higher than the reported efficiency trend versus input power. The reported circuit technique with the GaAs E-PHEMT GAD indicates good possibility for future …

Other articles from IEICE Technical Report; IEICE Tech. Rep. journal

Jin Nakazato

Jin Nakazato

Tokyo Institute of Technology

IEICE Technical Report; IEICE Tech. Rep.

User Scheduling based on Expanded Null-Space for Massive MIMO

(in English) This paper proposes a user scheduling scheme for multi-user massive MIMO (MU-mMIMO), particularly effective in time-varying channel environments. MU-MIMO enables spectral efficiency enhancement by multiplexing a number of user terminals (UTs) in the spatial domain.Due to the finite number of base station (BS) antennas, the simultaneous transmission capacity is constrained in MU-MIMO. Therefore, BS must divide UTs into distinct subsets and transmit signals for each subset. When UTs move at high speed, it causes inter-user interference (IUI) and subsequently reduces the spatial multiplexing capability.

Naoya Isoyama

Naoya Isoyama

Nara Institute of Science and Technology

IEICE Technical Report; IEICE Tech. Rep.

Influence of Temperature Stimulus to the Sitting Person's Seat Surface on the Center of Gravity

(in English) In this study, we focus on the deterioration of sitting posture due to VDT work that is spreading in modern society and the associated health problems, and investigate the possibility of improving sitting posture by means of thermal stimulation. Studies have been conducted on physical devices and information presentation methods to improve sitting posture. Physical devices promote effective posture improvement, but the physical deformation limits the user's degree of freedom and may cause stress. On the other hand, information presentation methods have the advantage of easily presenting information, but require the user to pay attention to the information presented, and spontaneous behavior change based on the presented information is not guaranteed to be effective depending on the will of the individual and the situation. In this study, we examine the effects of thermal stimuli on the actual and …

Yuki Matsuda (松田 裕貴)

Yuki Matsuda (松田 裕貴)

Nara Institute of Science and Technology

IEICE Technical Report; IEICE Tech. Rep.

Satisfaction Evaluation of Recommended Intake Calculation Method for Postprandial Hyperglycemia Prevention

(in English) Continuous self-management of environmental factors, including dietary habits to prevent postprandial hyperglycemia is essential. Previously, we proposed a framework for dietary behavior change that involves real-time monitoring of blood glucose levels, predicting postprandial hyperglycemia from planned meals, calculating foods that should be reduced in intake, and presenting the results to users. However, the optimal calculation method for selecting foods and quantities to induce behavioral change while maintaining meal satisfaction for preventing postprandial hyperglycemia remains unclear. To realize the proposed framework, it is necessary to establish calculation methods for recommended intake of each food and evaluate the satisfaction of meals when applying these recommended intake amounts.In this paper, we set multiple recommended intake calculation methods and conducted …

Koichi Kobayashi

Koichi Kobayashi

Hokkaido University

IEICE Technical Report; IEICE Tech. Rep.

Design of Block-Sparse Controllers for Cyber-Physical Systems

(in English) In sensor and actuator networks, it is important to consider not only the choice of sensors/actuators but also that of communication channels. In this paper, we propose an LMI-based design method for structured block-sparse output feedback controllers. First, the sparse reconstruction problems for vectors and matrices are summarized. Next, the design problem is formulated, and an LMI-based solution method is proposed. Finally, a simple example is presented to show the effectiveness of the proposed method.

Mohd Adib Sarijari

Mohd Adib Sarijari

Universiti Teknologi Malaysia

IEICE Technical Report; IEICE Tech. Rep.

An Evaluation of CNN Using Deep Residual Learning for Modulation, 5G, LTE, and WLAN System Classification

(in English) In this study, we investigate and present a deep residual learning for modulation classification. The simulation results show the degradation problem that was exposed due to an increase in network depth and the saturation of accuracy in the modified conventional CNN; however, the proposed CNN has no such degradation. Therefore, the processing burden of the conventional CNN is much larger than the proposed CNN. In the simulation results, the proposed CNN framework achieves almost the same modulation classification accuracy as the normal CNN framework when reducing the processing burden in the proposed one. The better simulation results are shown by adjustment of the parameters using the proposed method in the case of OFDM and single carrier modulation types.

Ching-Ming Lai, IET Fellow/AIE Fellow/IEEE Senior Member

Ching-Ming Lai, IET Fellow/AIE Fellow/IEEE Senior Member

National Chung Hsing University

IEICE Technical Report; IEICE Tech. Rep.

Over 98% Efficiency Floating Four-Phase Interleaved Charge-Pump Bidirectional DC-DC Converter with Wide Buck/Boost Voltage Ratio

(in English) A novel floating four-phase interleaved charge-pump bidirectional dc-dc converter (4F-ICPBDC) with wide Buck/Boost voltage ratio is proposed in this article. The interleaved structure is adopted to reduce the ripple of the current through the low-voltage side, and the floating structure is adopted to achieve the high Buck/Boost voltage conversion ratio. In addition, to achieve balance of the average inductor current within the full duty cycle range, an asymmetric duty limit control (DLC) strategy has been implemented. Furthermore, the operating principle of the converter is analyzed. Finally, a 1 kW prototype is developed with SiC-MOSFETs, which verifies the wide Buck/Boost voltage ratio of this converter between the constant low-voltage side (72 V) and the variable high-voltage side (400-800 V). The maximum efficiencies of the converter are recorded as 98.3% in the Buck mode and 97.6% in the Boost …

Kenta Ninomiya

Kenta Ninomiya

Kyushu University

IEICE Technical Report; IEICE Tech. Rep.

Prediction of Cancer Relapse for Non-Small Cell Lung Cancer Patients Using Clinical and Imaging Features Before Stereotactic Ablative Radiotherapy

(in English) The aim of this study is to predict locoregional relapse and distant metastasis in stage I non-small cell lung cancer patients using clinical and image features before radiotherapy. Image features were obtained from Betti number maps and wavelet decomposition images calculated from CT images. The predictive performances were evaluated by significant differences of the Kaplan-Meier curves for high-risk and low-risk patient groups predicted from the clinical and image features. The clinical and image combined features could have a potential to improve the relapse risk classification.

Keiichi Yasumoto

Keiichi Yasumoto

Nara Institute of Science and Technology

IEICE Technical Report; IEICE Tech. Rep.

A PoI Localization Method using Time-series SNS Posts with Photos

(in English) In recent years, with the easing of COVID-19 movement restrictions, there has been a trend towards the recovery of tourism demand. In tourist destinations, tourists increasingly use social media not only for disseminating information but also for gathering information. It is common for them to decide their destinations based on information posted by others. However, posts with location information typically involve tagging posts with locations registered in existing databases, which means unregistered locations cannot be tagged. This poses a challenge as there is also a trend in social media towards hiding location information for privacy protection. To apply these posts to tourist information services, it is necessary to estimate the Point of Interest (PoI) locations. This study proposes a method for estimating PoI locations based on the temporal sequence of photo and text posts on social media. Experiments …

Fumio Teraoka

Fumio Teraoka

Keio University

IEICE Technical Report; IEICE Tech. Rep.

Basic Evaluation of MECM2M, A M2M System Infrastructure Considering Mobile Nodes as Computing Servers

(in English) We are developing MECM2M, an M2M (Machine-to-Machine) system infrastructure suitable for MEC (Multi-access Edge Computing) environments, considering mobile nodes with both sensor/actuator and compute server functions, such as automobiles and smart phones.Based on the design concept of the Internet, MECM2M provides M2M applications with only the basic functions of acquiring sensor values and instructing actuators to move by specifying areas or nodes.

Shinpei Ogata

Shinpei Ogata

Shinshu University

IEICE Technical Report; IEICE Tech. Rep.

A Method to Create Scenario by Large Language Model for Input Amount Simulation

(in English) One of the means to reasonably reduce the burden of user input to the UI (User Interface) is to introduce usability-enhancing functions (usability functions), but it is difficult to find an appropriate configuration method for them. In order to improve this problem, we have proposed a simulation method to estimate the user input amount from screen transition models and scenarios for a usability function called an input storage function. However, there was a problem that the burden of manual scenario creation by the method user was large. To solve this problem, we propose a method to support scenario creation using a large language model. To evaluate the effectiveness of the proposed method, we attempted to create scenarios using the proposed method for three types of scenarios created in a previous experiment. The results showed that the proposed method was more effective than the methods used in …

Keiichi Yasumoto

Keiichi Yasumoto

Nara Institute of Science and Technology

IEICE Technical Report; IEICE Tech. Rep.

Satisfaction Evaluation of Recommended Intake Calculation Method for Postprandial Hyperglycemia Prevention

(in English) Continuous self-management of environmental factors, including dietary habits to prevent postprandial hyperglycemia is essential. Previously, we proposed a framework for dietary behavior change that involves real-time monitoring of blood glucose levels, predicting postprandial hyperglycemia from planned meals, calculating foods that should be reduced in intake, and presenting the results to users. However, the optimal calculation method for selecting foods and quantities to induce behavioral change while maintaining meal satisfaction for preventing postprandial hyperglycemia remains unclear. To realize the proposed framework, it is necessary to establish calculation methods for recommended intake of each food and evaluate the satisfaction of meals when applying these recommended intake amounts.In this paper, we set multiple recommended intake calculation methods and conducted …

ROZEHA BT. A. RASHID

ROZEHA BT. A. RASHID

Universiti Teknologi Malaysia

IEICE Technical Report; IEICE Tech. Rep.

An Evaluation of CNN Using Deep Residual Learning for Modulation, 5G, LTE, and WLAN System Classification

(in English) In this study, we investigate and present a deep residual learning for modulation classification. The simulation results show the degradation problem that was exposed due to an increase in network depth and the saturation of accuracy in the modified conventional CNN; however, the proposed CNN has no such degradation. Therefore, the processing burden of the conventional CNN is much larger than the proposed CNN. In the simulation results, the proposed CNN framework achieves almost the same modulation classification accuracy as the normal CNN framework when reducing the processing burden in the proposed one. The better simulation results are shown by adjustment of the parameters using the proposed method in the case of OFDM and single carrier modulation types.

Yuki Matsuda (松田 裕貴)

Yuki Matsuda (松田 裕貴)

Nara Institute of Science and Technology

IEICE Technical Report; IEICE Tech. Rep.

A PoI Localization Method using Time-series SNS Posts with Photos

(in English) In recent years, with the easing of COVID-19 movement restrictions, there has been a trend towards the recovery of tourism demand. In tourist destinations, tourists increasingly use social media not only for disseminating information but also for gathering information. It is common for them to decide their destinations based on information posted by others. However, posts with location information typically involve tagging posts with locations registered in existing databases, which means unregistered locations cannot be tagged. This poses a challenge as there is also a trend in social media towards hiding location information for privacy protection. To apply these posts to tourist information services, it is necessary to estimate the Point of Interest (PoI) locations. This study proposes a method for estimating PoI locations based on the temporal sequence of photo and text posts on social media. Experiments …

Akira Uchiyama

Akira Uchiyama

Osaka University

IEICE Technical Report; IEICE Tech. Rep.

A Preliminary Study on Parameter Optimization Using a Backpropagation Algorithm for a Neonatal Thermal Model

(in English) Neonates need temperature management in incubators due to their underdeveloped thermoregulatory functions. Traditional methods using skin-attached probes are challenging due to the neonates' delicate skin. Therefore, efforts are being made to realize non-contact temperature measurement using thermography. However, thermography can be hindered during medical procedures by obstructions like medical staff's hands. To address this, we utilize a human thermal model to simulate body temperature changes, integrating real-time sensor data, including thermography and incubator conditions, to estimate surface and core body temperatures. Our method employs backpropagation for rapid parameter optimization in the human thermal model. The effectiveness was confirmed using actual neonatal data, achieving core body temperature estimation with an average absolute error of 0.032$ degree …

Sinan Chen

Sinan Chen

Kobe University

IEICE Technical Report; IEICE Tech. Rep.

A Study of Promotion Method for Energy-Saving Behavior in Homes with Personalized Adaptive Interaction

(in English) In recent years, climate change, including global warming, has become a serious issue, and Japan is aiming to realize a zero carbon society. Zero carbon means reducing the net emissions of greenhouse gases to zero by balancing the amount emitted and absorbed, which is crucial for reducing CO2 emissions. This study focuses on energy reduction through in-home energy-saving actions, addressing issues such as the lack of acquisition of appliance-specific electricity consumption data, lack of personal adaptation, and oversight of notifications. To address these, a method that integrates a power consumption management service using IoT devices and baselines with a virtual agent (VA) was proposed and implemented. The evaluation experiment confirmed that the introduction of VA led to a reduction in power consumption, and an improvement in energy-saving consciousness and behavioral …

RUBITA SUDIRMAN

RUBITA SUDIRMAN

Universiti Teknologi Malaysia

IEICE Technical Report; IEICE Tech. Rep.

An Evaluation of CNN Using Deep Residual Learning for Modulation, 5G, LTE, and WLAN System Classification

(in English) In this study, we investigate and present a deep residual learning for modulation classification. The simulation results show the degradation problem that was exposed due to an increase in network depth and the saturation of accuracy in the modified conventional CNN; however, the proposed CNN has no such degradation. Therefore, the processing burden of the conventional CNN is much larger than the proposed CNN. In the simulation results, the proposed CNN framework achieves almost the same modulation classification accuracy as the normal CNN framework when reducing the processing burden in the proposed one. The better simulation results are shown by adjustment of the parameters using the proposed method in the case of OFDM and single carrier modulation types.

Shinji Watanabe

Shinji Watanabe

Carnegie Mellon University

IEICE Technical Report; IEICE Tech. Rep.

Evaluating speech generation based on objective measures for text generation

(in English) In the evaluation of speech generation, while subjective judgments have long been the gold standard, objective metrics such as Mel Cepstral Distortion (MCD) and Mean Opinion Score (MOS) prediction models have also been used. These objective metrics are valued for their lower time and financial costs and the ability to compare different results, driving the demand for metrics that correlate well with human subjective judgments. This paper proposes an automatic evaluation method for speech generation based on text generation metrics. Our proposed SpeechBERTScore computes BERTScores on self-supervised speech feature sequences derived from both generated and reference speech. In addition, SpeechBLEU and SpeechTokenDistance define metrics using self-supervised discrete speech tokens. Experimental evaluations of synthesized speech show that our SpeechBERTScore correlates …

Yohei Murakami

Yohei Murakami

Ritsumeikan University

IEICE Technical Report; IEICE Tech. Rep.

Cooperative Agents for Federated Learning of Neural Machine Translation

(in English) Accurate neural machine translation requires large amounts of high-quality bilingual data, however due to the copyright and confidentiality issues, it is difficult to share the bilingual data between different organizations. With federal learning, multiple clients can collaboratively build a neural machine translation model and share only the translation model while keeping their own data confidential. However, if the domain of bilingual data differs among clients, the translation accuracy is not necessarily improved by integrating all translation models. Therefore, we propose a cooperative agent that dynamically selects a cooperative partner and integrates translation models in each aggregation process in cooperative learning. Compared to conventional cooperative learning methods, our method can improve the translation accuracy of each client by 22.7% on average.

TUTOMU MURASE

TUTOMU MURASE

Nagoya University

IEICE Technical Report; IEICE Tech. Rep.

Segmented file delivery method by mobile APs using a combination of overhearing and transmission rate control

(in English) In the event of a large-scale disaster, there is a possibility that the communication infrastructure may not work properly. Therefore, a method of distributing evacuation information that does not depend on the communication infrastructure is required. Ad hoc networks and DTNs are available as communication methods that do not depend on the infrastructure in the event of a disaster. In addition, other research has proposed a method that enable information distribution to terminals over a wide area by using mobile vehicles equipped with wireless LAN access points. However, conventional researches consider information which size is a few Byte. Therefore, they did not consider a large amount of information which size is about 10Mbyte such as the evacuation map to deliver for multiple terminals. This paper proposes a method which combines an overhearing and a segmented file delivery method based …

Yuki Koizumi

Yuki Koizumi

Osaka University

IEICE Technical Report; IEICE Tech. Rep.

Evaluation of Large-capacity Content Transmission Equipment to Bundle12 and 21GHz-band Satellite Transmission Channels and IP Network--Laboratory Experiment for Flexible …

(in English) We have developed large-capacity content transmission equipment that bundles 21-and 12-GHz-band satellite channels and an IP network for providing immersive content such as volumetric content that is large capacity due to expanding spatial expressions. This equipment achieves a maximum transmission capacity of about 1 Gbps by transmitting a content to multiple transmission channels and network where the capacity and delay are different. For immersive media, we are developing video and audio formats and scene descriptions to facilitate AR/VR schemes and a variety of other content representations appropriate for any given viewer environment. Immersive content enables users to experience various viewer environments by producing content for each object. Therefore, we developed a function that specifies the transmission channels or network in accordance with the objects of the content …