関屋 大雄

This paper proposes a load-independent class-E-1 inverter with shunt capacitance. This inverter achieves the ZVS and the constant output current at any load resistances without any control, which is called load-independent mode. This paper provides waveform and design equations of the proposed inverter. We confirmed that the laboratory prototype worked with the load-independent characteristics, which denoted validities of the analytical equations.

This paper proposes the load-independent inverse class-E oscillator with Armstrong-oscillator based topology. The proposed oscillator achieves constant output voltage and soft switching against load variations due to the load-independent design. Moreover, the proposed circuit realizes the self-Tuning function using the Armstrong-oscillator topology. As a result, the oscillator can also have robustness against the variations in the resonant frequency. The experimental results showed the effectiveness of the proposed oscillator.

In this research, a novel numerical design strategy is presented for the class-DE inverter design, which is based on particle swarm optimization (PSO) algorithm. Compared with previous numerical design methods based on Newton's method, the proposed method shows convenience without restricting the number of design parameters. Additionally, no initial estimation is needed for ensuring convergence. Therefore, system-level optimization, which considers optimized inductor designs, can be realized. A design example of the class-DE inverter is shown in this paper. The numerical calculation proved the validity of the proposed design method.

This paper tackles the problem of using and paying a network service with guaranteed quality of service (QoS) at a fine granularity in the future wireless network. We propose the IOTA-based QoS guaranteed Flow system (IQF), where a user and service provider can exchange the IOTA cryptocurrency for the guaranteed network resources, such as latency and bandwidth. First, IQF uses the IOTA Tangle, a lightweight, efficient distributed ledger technology, for its payment and transaction record. Second, IQF adopts Software Defined Networking to realize the guaranteed delay and bandwidth provision. We have implemented the IQF systems using the mininet-wifi emulator, an SDN controller (i.e., POX), and IOTA clients. Moreover, we have evaluated the service provisioning algorithms with a micropayment of IQF with the test Tangle (i.e., comnet). The evaluation results show that IQF has successfully achieved the delay and bandwidth provision with all the transactions stored in comnet.

This paper presents a 2-hop wireless power transfer (WPT) system for the robot-arm application. The proposed WPT system consists of the class-E inverter on the transmitter side and a class-D rectifier on the receiver side. We obtain the analytical model of the WPT system for obtaining the class-E ZVS/ZDS conditions. The model includes the detailed physical information of the coil, namely coil diameter, height, wire thickness, winding turn number, and so on. Therefore, it is possible to optimize the WPT system for the theoretical highest power-delivery efficiency under the robot-arm coil-space restriction. The circuit experiment showed quantitative agreements with the theoretical predictions, which proved the validity of the proposed design method. The experimental prototype achieved 78.0 % power-delivery efficiency at 6.78 MHz operating frequency and 39.3 W output power.

Network softwarization, such as Software-Defined Networking (SDN), is a potential technology to manage networking resources and routing paths dynamically in 5G and beyond. However, SDN, which provides the control plane programmability, requires a de-factor protocol (i.e., OpenFlow) on network devices. In an SDN network, the load of network control traffic is likely concentrated on a network controller. Moreover, most of SDN advantages are not easily extensible in wireless networks. Programming Protocol-independent Packet Processors (P4), which enables the data plane programmability, potentially solves SDN issues. P4 allows programming new packet forwarding methods, which don't necessarily rely on the predefined protocols. In this paper, we show the feasibility of P4 in realizing a multipath communication mechanism in wireless networks. We first introduce and implement a P4-based multipath transmission mechanism of UDP traffic. We then evaluate the mechanism on an emulated wireless network (i.e., using mininetwifi) in comparison to the single path. The results show that the P4-based mechanism can effectively achieve the throughput aggregation with a slightly bigger latency variation.

Guaranteeing delay performance is an essential issue in a Wi-Fi network since it has increasingly conveyed a huge amount of Internet application traffic. This paper introduces our continuous effort in delay-guaranteed provision for Wi-Fi networks. More specifically, we have introduced the previously proposed delay-guaranteed mechanism that uses the advantage of SDN technology. In this work, we evaluate the mechanism in a new environment considering several TCP flows. The evaluation results show that the SDN-based algorithm effectively controls the delay values of multiple flows. Moreover, the sharing bandwidth between the flows is fair with negligible overhead.

Wi-Fi technology has become popular in our lives with various Wi-Fi capable products such as laptops, mobile phones, etc. Moreover, in Wi-Fi networks, a device communicates typically with a server using Transmission Control Protocol (TCP) for most applications. In such a context, the behavior of the TCP congestion control primarily affects network performance. Hence, it is essential to determine and select an appropriate TCP congestion control that fits different network conditions. In this paper, we present a performance evaluation of TCP congestion control algorithms in an emulated Wi-Fi network. More specifically, we investigate 14 TCP congestion control algorithms in the Wi-Fi network, considering the variation of queue size and packet loss. The evaluation results of Round Trip Time (RTT) and achieved throughput allows us to compare the TCP congestion control algorithms. In our investigation, the Bottleneck Bandwidth Roundtrip-Time algorithm (BBR) has the best performance.

This paper presents a design example of the class-Φ3 inverter. The design parameters can be theoretically obtained by applying the multi-frequency analysis. The class-Phi3 inverter is a special version of the class-E/F3 inverters, whose output reactance is zero. It is possible to obtain the circuit component values, which give zero output reactance of the output filter with consideration of low input inductance. The experimental results and PSpice-simulation results agreed with the analytical predictions quantitatively, which showed the validity of analytical expressions in this paper.

This paper proposes a parallel resonant power oscillator that has the load-independent characteristic. The proposed oscillator is the first parallel-resonant power oscillator for high frequency operations. In addition, it achieves the constant output voltage and the soft switching against the load variations without any control. The feedback network for generating a driving signal does not have a resonant structure and its function is not affected by the self-oscillation frequency. Therefore, the proposed oscillator also has a self-tuning function for the resonant-filter component tolerances. Because of the above characteristics, the proposed oscillator is expected to be used in many applications such as a transmitter of the wireless power transfer systems.

In the current and next generation of mobile networks, a mobile device intuitively has multiple wireless radios such as Wi-Fi, LTE, 5G New Radio, etc. Moreover, most likely, the mobile device continues to use the TCP/IP stack, which limits the device's network selection. The application traffic may stick to a default network, which may not have the best performance among the associated ones. That harms the performance of TCP short flows, such as web access. This paper proposes using TCP SYN duplication to relax the constraint, aiming to select the best network at the present moment. The device tries to initialize a TCP connection by duplicating and sending SYN packets via all the available networks. The network, which conveys the earliest SYN/ACK response, will be selected for the data transfer. We implement and evaluate the proposal in a testbed with real Internet connections to show its effectiveness.

A novel design of single-stage single-switch PFC converter topology is presented in this paper. The converter bases on class-E2 topology and operates at 1 MHz frequency. With only one wide-bandgap GaN FET without extra stages or cells, it is possible to reduce the size of the circuit. Frequency modulation control method is applied to the system to decrease the harmonic distortions of the input line current. In order to keep the ZVS characteristic achieved during the entire period, a numerical design procedure of the converter is also considered, which can reduce the switching loss of the switch efficiently. Moreover, experiment prototype is implemented to validate the high performance of the proposed system. The experimental result shows good agreement with the simulation, achieving high power conversion efficiency and low total harmonic distortions (THD), which proves the effectivity of the proposed design procedure.

The IEEE 802.11ad technology, which is a member of the IEEE 802.11 family, promisingly extends the operation of Wi-Fi networks to the unlicensed 60 GHz band. In such a case, the Wi-Fi network can support the throughput of multi-gigabit per second, potentially support various emerging applications. Since the Transport Control Protocol (TCP) plays a vital role in Wi-Fi networks, it is necessary to understand the TCP performance in that multi-Gbps condition. This paper investigates the performance of popular TCP variants on the IEEE 802.11ad wireless link. We conduct the experiments with TCP CUBIC, CDG, BBR, which are typically loss-based, delay-based, and model-based TCP congestion control algorithms, in different loss conditions. The results show that in a no-loss scenario, CUBIC and CDG keep good throughput at the cost of high TCP latency. Meanwhile, BBR has excellent performance even under the high loss scenario.

This paper presents an analytic expression of the class-E rectifier with low output-filter inductance. Compared with the traditional rectifiers, the larger power output capability can be obtained by varying the output-filter inductance. Also, the validity of the analytical expression and design strategy are presented by implementation of the class-E resonant DC-DC converter. The converter with low output-filter inductance achieved 11 % power output capability than the traditional class-E rectifier. 2

This paper presents an easy-to-use analytical design approach of the load-independent class-E inverter at any duty ratio. By using the analytical approach, it is possible to maintain the amplitude of the output voltage and the operation of the zero-voltage switching. Additionally, a design example is given along with its LTspice simulation and laboratory experiment. The LTspice simulation and laboratory experiment waveforms are in good agreement with the analytical ones even load variations. Moreover, all the switch-voltage waveforms satisfy the ZVS condition. These results showed that the proposed approach and analytical expressions are available and effective.

© 2019 IEEE. This paper presents a novel class-E2 dc-dc converter and its design procedure. The proposed converter can maintain high-frequency and high-efficiency operations while realizing low switch-voltage stresses. Additionally, by using the presented design procedure, the proposed converter satisfying the zero-voltage-switching and zero-derivative-switching (ZVS/ZDS) conditions can be implemented without applying any tuning processes. A design example of the proposed class-E2 converter is shown along with its LTspice simulation and laboratory experiment. The results from the LTspice simulator and laboratory experiment show quantitative agreement with the analytical predictions, which indicates the validity of the proposed topology and design procedure.