• Journal of Power Electronics
• /
• 제16권3호
• /
• pp.1226-1235
• /
• 2016

For wearable health monitoring systems, a fundamental problem is the limited space for storing energy, which can be translated into a short operational life. In this paper, a highly efficient active voltage doubling rectifier with a wide input range for micro-piezoelectric energy harvesting systems is proposed. To obtain a higher output voltage, the Dickson charge pump topology is chosen in this design. By replacing the passive diodes with unbalanced-biased comparator-controlled active counterparts, the proposed rectifier minimizes the voltage losses along the conduction path and solves the reverse leakage problem caused by conventional comparator-controlled active diodes. To improve the rectifier input voltage sensitivity and decrease the minimum operational input voltage, two low power common-gate comparators are introduced in the proposed design. To keep the comparator from oscillating, a positive feedback loop formed by the capacitor C is added to it. Based on the SMIC 0.18-μm standard CMOS process, the proposed rectifier is simulated and implemented. The area of the whole chip is 0.91×0.97 mm², while the rectifier core occupies only 13% of this area. The measured results show that the proposed rectifier can operate properly with input amplitudes ranging from 0.2 to 1.0V and with frequencies ranging from 20 to 3000 Hz. The proposed rectifier can achieve a 92.5% power conversion efficiency (PCE) with input amplitudes equal to 0.6 V at 200 Hz. The voltage conversion efficiency (VCE) is around 93% for input amplitudes greater than 0.3 V and load resistances larger than 20kΩ.

• Journal of Multimedia Information System
• /
• 제8권4호
• /
• pp.203-210
• /
• 2021

Deep leaning convolutional neural networks (CNN) have successfully been applied to image super-resolution (SR). Despite their great performances, SR techniques tend to focus on a certain upscale factor when training a particular model. Algorithms for single model multi-scale networks can easily be constructed if images are upscaled prior to input, but sub-pixel convolution upsampling works differently for each scale factor. Recent SR methods employ multi-scale and multi-path learning as a solution. However, this causes unshared parameters and unbalanced parameter distribution across various scale factors. We present a multi-scale single-path upsample module as a solution by exploiting the advantages of sub-pixel convolution and interpolation algorithms. The proposed model employs sub-pixel convolution for the highest scale factor among the learning upscale factors, and then utilize 1-dimension interpolation, compressing the learned features on the channel axis to match the desired output image size. Experiments are performed for the single-path upsample module, and compared to the multi-path upsample module. Based on the experimental results, the proposed algorithm reduces the upsample module's parameters by 24% and presents slightly to better performance compared to the previous algorithm.

• 대한전자공학회논문지SD
• /
• 제43권4호
• /
• pp.56-65
• /
• 2006

Quantum-dot Cellular Automata (QCA)는 분자 혹은 원자 수준의 작은 크기의 소자이며 극도로 낮은 소모 전력 특성을 가지기 때문에 디지털 논리 구현에 있어 차세대 기술로 많은 주목을 받고 있다. 현재까지 다양한 QCA 설계가 발표되었지만 대부분 시뮬레이션에 의해 동작하지 않음이 확인되었으며 설계를 위한 일반적인 규범이나 지침도 제시되지 알았다. 동작하는 기본적인 구조를 간단히 확장하는 경우에도 시뮬레이션이 실패하였으며 대규모 회로 설계에는 엄청난 시간 소요가 예상되었다. 본 논문에서는 게이트 입력 경로의 불균형 및 배선구조의 숨은 잡음 경로등 기본적인 QCA 구조에서 나타난 치명적인 취약성에 대해 설명하고 이를 해결하기 위한 강건한 QCA 설계를 위해 규범 및 지침을 제시한다. 또한, 이 강건 설계 기법에 따라 설계되고 시뮬레이션에 의해 그 동작이 검증된 고속 가산기를 제시한다.