グエン キエン

This paper proposes a load-independent high-frequency wireless power transfer (WPT) system. The proposed system achieves mode switching between constant current (CC) and constant voltage (CV) output modes, which can be applied to rechargeable batteries. In the proposed system, a single switch component is added on the receiver side to change the topology of the compensation circuit. As a result, the switching between CC and CV modes is accomplished on the receiver side, which means wireless communication about the load-information feedback from the receiver to the transmitter becomes unnecessary. By achieving zero-voltage switching (ZVS) in both CC and CV modes through the load-independent operation, the proposed system achieves high power-delivery efficiency even at high frequencies. Therefore, the proposed WPT system contributes to cost reduction, system simplification, and circuit downsizing. The experimental results agreed with the analytical predictions quantitatively, which verified the effectiveness of the proposed circuits.

This paper presents the design of the class-Φ3 power oscillator. The topology of the proposed power oscillator features the feedback network from the resonant capacitance, which allows for the application of wireless power transfer (WPT) systems. Furthermore, the required phase shift necessary for applying the proposed feedback network is achieved by employing the class-Φ3 inverter, which has a finite input inductance. Closed-form design equations for the proposed power oscillator are provided, which allows a quick and intuitive design. An experimental verification is conducted with the prototype power oscillator. The measured power-conversion efficiency was 92.2 % with 70 W output power at 1 MHz operation. The experimental results agreed quantitatively with the analysis, which shows the validity of the design.

This study explores the intersection of neuroscience and computer science, focusing on the use of spiking neural networks (SNNs) to simulate the behavior of biological neurons. A neural network model based on the Izhikevich neuron model is proposed to simulate the local auditory network of crickets. The parameters of the neuron model are optimized based on evaluation functions and identified by Particle Swarm Optimization (PSO), aligning its input-output relationships with the observed cricket neuron responses. The results showed that the network successfully simulated the behavior of individual neurons, promising applications in fields like neural prosthetics.

This paper proposes a high-frequency multiple-receiver wireless power transfer (WPT) system with a load-independent class-E/F inverter. Each receiver has a post-regulator, which changes the equivalent resistance seen from the inverter to obtain the necessary power for output voltage regulation. Because the load-independent class-E/F inverter generates constant AC current ((Formula presented.)), the transmitter supplies the minimum required power to the receivers by the change of the equivalent resistances. Besides, the load-independent inverter consistently achieves zero-voltage switching (ZVS) without any control. As a result, no information feedback by wireless communication is necessary for output regulation and ZVS achievement, simplifying the system configuration and improving the transient response of the control. This paper presents analytical expressions of the proposed system. Besides, the experiment was carried out with a two-receiver WPT system. The implemented system worked well by individual and independent output regulation of the post regulator at each receiver. The implemented WPT system achieved the power-delivery efficiency of 83.4% at the 6.78 MHz transmission frequency and the total output power of 40 W.