We propose a subthreshold Static Random Access Memory (SRAM) circuit architecture with improved write ability. Even though the circuits can achieve ultra-low power dissipation in subthreshold digital circuits, the performance is significantly degraded with threshold voltage variations due to the fabrication process and temperature. Because the write operation of SRAM is prone to failure due to the unbalance of threshold voltages between the nMOSFET and pMOSFET, stable operation cannot be ensured. To achieve robust write operation of SRAM, we developed a compensation technique by using an adaptive voltage scaling technique that uses an on-chip threshold voltage monitoring circuit. The monitoring circuit detects the threshold voltage of a MOSFET with the on-chip circuit configuration. By using the monitoring voltage as a supply voltage for SRAM cells, write operation can be compensated without degrading cell stability. Monte Carlo simulations demonstrated that the proposed SRAM architecture exhibits a smaller write operation failure rate and write time variation than a conventional 6T SRAM.

In this paper, we propose an error diagnosis technique based on clustering LUT elements to shorten the processing time. By grouping some elements as a cluster, our technique reduces the number of elements to be considered, which is effective to shorten the processing time for screening error location sets. First, the proposed technique partitions the circuit into FFR (fanout-free region) called cluster, which is a subcircuit composed of LUT elements without fanout. After screening the set of clusters including error locations, this technique screens error location sets composed of elements in the remaining set of clusters, where corrections should be made. Experimental results with benchmark circuits have shown that our technique shortens the processing time to 1/170 in the best case, and rectifies circuits including 6 errors which cannot be rectified by the conventional technique.

This paper proposes image super-resolution techniques with multi-channel convolutional neural networks. In the proposed method, output pixels are classified into K×K groups depending on their coordinates. Those groups are generated from separate channels of a convolutional neural network (CNN). Finally, they are synthesized into a K×K magnified image. This architecture can enlarge images directly without bicubic interpolation. Experimental results of 2×2, 3×3, and 4×4 magnifications have shown that the average PSNR for the proposed method is about 0.2dB higher than that for the conventional SRCNN.

An intermediate frequency (IF) variable gain amplifier (VGA) with exponential gain control for a radio receiver is fabricated in 0.25-µm CMOS technology. The techniques to improve the bandwidth and to reduce temperature dependence of gain are described. The complete VGA is composed of two stages of linearized transconductance VGA and three stages of fixed gain amplifier (FGA). The complete VGA provides a continuous 10 dB to 76.5 dB gain control range, an IIP3 of -11.5 dBm and an NF of 15 dB at 40 MHz.

This paper describes a programmable differential bandgap reference (PD-BGR) for ultra-low-power IoT (Internet-of-Things) edge node devices. The PD-BGR consists of a current generator (CG) and differential voltage generator (DVG). The CG is based on a bandgap reference (BGR) and generates an operating current and a voltage, while the DVG generates another voltage from the current. A differential voltage reference can be obtained by taking the voltage difference from the voltages. The PD-BGR can produce a programmable differential output voltage by changing the multipliers of MOSFETs in a differential pair and resistance with digital codes. Simulation results showed that the proposed PD-BGR can generate 25- to 200-mV reference voltages with a 25-mV step within a ±0.7% temperature inaccuracy in a temperature range from -20 to 100°C. A Monte Carlo simulation showed that the coefficient of the variation in the reference was within 1.1%. Measurement results demonstrated that our prototype chips can generate stable programmable differential output voltages, almost the same results as those of the simulation. The average power consumption was only 88.4 nW, with a voltage error of -4/+3 mV with 5 samples.

Extremely low-voltage charge pump (ELV-CP) and its dedicated multi-stage driver (MS-DRV) for sub-60-mV thermoelectric energy harvesting are proposed. The proposed MS-DRV utilizes the output voltages of each ELV-CP to efficiently boost the control clock signals. The boosted clock signals are used as switching signals for each ELV-CP and MS-DRV to turn switch transistors on and off. Moreover, reset transistors are added to the MS-DRV to ensure an adequate non-overlapping period between switching signals. Measurement results demonstrated that the proposed MS-DRV can generate boosted clock signals of 350 mV from input voltage of 60 mV. The ELV-CP can boost the input voltage of 100 mV with 10.7% peak efficiency. The proposed ELV-CP and MS-DRV can boost the low input voltage of 56 mV.

This paper presents a fully integrated voltage boost converter consisting of a charge pump (CP) and maximum power point tracking (MPPT) controller for ultra-low power energy harvesting. The converter is based on a conventional CP circuit and can deliver a wide range of load current by using nMOS and pMOS driver circuits for highly efficient charge transfer operation. The MPPT controller we propose dissipates nano-watt power to extract maximum power regardless of the harvester's power generation conditions and load current. The measurement results demonstrated that the circuit converted a 0.49-V input to a 1.46-V output with 73% power conversion efficiency when the output power was 348µW. The circuit can operate at an extremely low input voltage of 0.21V.

Reducing the power consumption of capsule endoscopy is essential for its further development. We introduce K-SVD dictionary learning to design a dictionary for sparse coding, and improve reconstruction accuracy of capsule endoscopic images captured using compressed sensing. At a compression ratio of 20%, the proposed method improves image quality by approximately 4.4 dB for the peak signal-to-noise ratio.

A fully on-chip CMOS relaxation oscillator (ROSC) with a PVT variation compensation circuit is proposed in this paper. The circuit is based on a conventional ROSC and has a distinctive feature in the compensation circuit that compensates for comparator's non-idealities caused by not only offset voltage, but also delay time. Measurement results demonstrated that the circuit can generate a stable clock frequency of 6.66kHz. The current dissipation was 320nA at 1.0-V power supply. The measured line regulation and temperature coefficient were 0.98%/V and 56ppm/°C, respectively.

We propose an ultra low power watch-dog circuit with the use of MOSFETs operation under subthreshold characteristics. The circuit monitors the amount of the product degradation because the subthreshold current of MOSFET emulates the rate of the general chemical reaction. Its operation was verified with both SPICE simulation and the measurement of the prototype chip. The new circuit embedded in a tag attached to any product could dynamically monitor the degradation regardless of storage conditions.

This paper presents an architecture of signal-dependent analog-to-digital converter (ADC) based on MINIMAX sampling scheme that allows achieving high data compression rate and power reduction. The proposed architecture consists of a conventional synchronous ADC, a timer and a peak detector. AD conversion is carried out only when input signal peaks are detected. To improve the accuracy of signal reconstruction, MINIMAX sampling is improved so that multiple points are captured for each peak, and its effectiveness is experimentally confirmed. In addition, power reduction, which is the primary advantage of the proposed signal-dependent ADC, is analytically discussed and then validated with circuit simulations.

A threshold-logic gate device consisting of subthreshold MOSFET circuits is proposed. The gate device performs threshold-logic operation, using the technique of current-mode addition and subtraction. Sample digital subsystems, i.e., adders and morphological operation cells based on threshold logic, are designed using the gate devices, and their operations are confirmed by computer simulation. The device has a simple structure and operates at low power dissipation, so it is suitable for constructing cell-based, parallel processing LSIs such as cellular-automaton and neural-network LSIs.

We designed subthreshold analog MOS circuits implementing an inhibitory network model that performs noise-shaping pulse-density modulation (PDM) with noisy neural elements, with the aim of developing a possible ultralow-power one-bit analog-to-digital converter. The static and dynamic noises given to the proposed circuits were obtained from device mismatches of current sources (transistors) and externally applied random spike currents, respectively. Through circuit simulations we confirmed that the circuit exhibited noise-shaping properties, and signal-to-noise ratio (SNR) of the network was improved by 7.9 dB compared with that of the uncoupled network as a result of noise shaping.

This paper proposes an ultra-low power fully on-chip CMOS relaxation oscillator (ROSC) for a real-time clock application. The proposed ROSC employs a compensation circuit of a comparator's non-idealities caused by offset voltage and delay time. The ROSC can generate a stable, and 32-kHz oscillation clock frequency without increasing power dissipation by using a low reference voltage and employing a novel compensation architecture for comparators. Measurement results in a 0.18-$mu$m CMOS process demonstrated that the circuit can generate a stable clock frequency of 32.55,kHz with low power dissipation of 472,nW at 1.8-V power supply. Measured line regulation and temperature coefficient were 1.1%/V and 120,ppm/$^{circ}$C, respectively.

In this paper, we propose a low-power circuit-shared static flip-flop (CS2FF) for extremely low power digital VLSIs. The CS2FF consists of five static NORs and two inverters (INVs). The CS2FF utilizes a positive edge of a buffered clock signal, which is generated from a root clock, to take data into a master latch and a negative edge of the root clock to hold the data in a slave latch. The total number of transistors is only 24, which is the same as the conventional transmission-gate flip flop (TGFF) used in the most standard cell libraries. SPICE simulations in 0.18-µm standard CMOS process demonstrated that our proposed CS2FF achieved clock-to-Q delay of 18.3ns, setup time of 10.0ns, hold time of 5.5ns, and power dissipation of 9.7nW at 1-MHz clock frequency and 0.5-V power supply. The physical design area increased by 16% and power dissipation was reduced by 13% compared with those of conventional TGFF. Measurement results demonstrated that our proposed CS2FF can operate at 0.352V with extremely low energy of 5.93fJ.

We present on-chip oscillator arrays synchronized by random noises, aiming at skew-free clock distribution on synchronous digital systems. Nakao et al. recently reported that independent neural oscillators can be synchronized by applying temporal random impulses to the oscillators [1],[2]. We regard neural oscillators as independent clock sources on LSIs; i.e., clock sources are distributed on LSIs, and they are forced to synchronize through the use of random noises. We designed neuron-based clock generators operating at sub-RF region (< 1 GHz) by modifying the original neuron model to a new model that is suitable for CMOS implementation with 0.25-µm CMOS parameters. Through circuit simulations, we demonstrate that i) the clock generators are certainly synchronized by pseudo-random noises and ii) clock generators exhibited phase-locked oscillations even if they had small device mismatches.

We propose a sampling method that incorporates a normally distributed sampling series for EEG measurements using compressed sensing. We confirmed that the ADC sampling count and amount of wirelessly transmitted data can be reduced by 11% while maintaining a reconstruction accuracy similar to that of the conventional method.

Convolutional neural network (CNN)-based image super-resolutions are widely used as a high-quality image-enhancement technique. However, in general, they show little to no luminance isotropy. Thus, we propose two methods, “Luminance Inversion Training (LIT)” and “Luminance Inversion Averaging (LIA),” to improve the luminance isotropy of CNN-based image super-resolutions. Experimental results of 2× image magnification show that the average peak signal-to-noise ratio (PSNR) using Luminance Inversion Averaging is about 0.15-0.20dB higher than that for the conventional super-resolution.

We developed a CMOS watchdog sensor that simulates the changes in quality of perishables such as farm and marine products. The sensor can imitate a chemical reaction that causes the changes in the quality of perishables, with a wide range of activation energy from 0.1 eV to 0.7 eV. Attached to perishable goods, the sensor simulates the deterioration of the goods caused by surrounding temperatures. By reading the output of the sensor, consumers can determine whether the goods are fresh or not. This sensor consists of subthreshold CMOS circuits with a low-power consumption of 5 µW or less.

This paper presents an image super-resolution technique using a convolutional neural network (CNN) and multi-task learning for multiple image categories. The image categories include natural, manga, and text images. Their features differ from each other. However, several CNNs for super-resolution are trained with a single category. If the input image category is different from that of the training images, the performance of super-resolution is degraded. There are two possible solutions to manage multi-categories with conventional CNNs. The first involves the preparation of the CNNs for every category. This solution, however, requires a category classifier to select an appropriate CNN. The second is to learn all categories with a single CNN. In this solution, the CNN cannot optimize its internal behavior for each category. Therefore, this paper presents a super-resolution CNN architecture for multiple image categories. The proposed CNN has two parallel outputs for a high-resolution image and a category label. The main CNN for the high-resolution image is a normal three convolutional layer-architecture, and the sub neural network for the category label is branched out from its middle layer and consists of two fully-connected layers. This architecture can simultaneously learn the high-resolution image and its category using multi-task learning. The category information is used for optimizing the super-resolution. In an applied setting, the proposed CNN can automatically estimate the input image category and change the internal behavior. Experimental results of 2× image magnification have shown that the average peak signal-to-noise ratio for the proposed method is approximately 0.22 dB higher than that for the conventional super-resolution with no difference in processing time and parameters. We have ensured that the proposed method is useful when the input image category is varying.