The Infona portal uses cookies, i.e. strings of text saved by a browser on the user's device. The portal can access those files and use them to remember the user's data, such as their chosen settings (screen view, interface language, etc.), or their login data. By using the Infona portal the user accepts automatic saving and using this information for portal operation purposes. More information on the subject can be found in the Privacy Policy and Terms of Service. By closing this window the user confirms that they have read the information on cookie usage, and they accept the privacy policy and the way cookies are used by the portal. You can change the cookie settings in your browser.
Neural Networks (NNs) have recently gained popularity in a wide range of modern application domains due to its superior inference accuracy. With growing problem size and complexity, modern NNs, e.g., CNNs (Convolutional NNs) and DNNs (Deep NNs), contain a large number of weights, which require tremendous efforts not only to prepare representative training datasets but also to train the network. There...
Neural Networks (NNs) have recently gained popularity in a wide range of modern application domains due to its superior inference accuracy. With growing problem size and complexity, modern NNs, e.g., CNNs (Convolutional NNs) and DNNs (Deep NNs), contain a large number of weights, which require tremendous efforts not only to prepare representative training datasets but also to train the network. There...
Resistive Memory (ReRAM) has emerged as a promising non-volatile memory technology that may replace a significant portion of DRAM in future computer systems. ReRAM has many advantages such as high density, low standby power and good scalability. ReRAM, when adopting crossbar architecture, has the smallest 4F2 planar cell size, which is ideal for constructing dense memory with large capacity. However,...
Modern Graphics Processing Units (GPUs) widely adopt large SRAM based register file (RF) to enable fast context-switch. A large SRAM RF may consume 20% to 40% GPU power, which has become one of the major design challenges for GPUs. Recent studies mitigate the issue through hybrid RF designs that architect a large STT-RAM (Spin Transfer Torque Magnetic memory) RF and a small SRAM buffer. However, the...
To eliminate the worst-case timing margins, a 13-transistor holosymmetrical transition detector (HTD) is proposed for use in timing variation resilient systems. The HTD achieves low overhead and wide-voltage-range operation via monitoring the discharge at the floating node of two-stage CMOS inverters. Using local detection and global clock stalling, the system is stalled immediately for one cycle...
This paper investigates the semi-global practical stabilization for second-order nonlinear systems subject to unknown control coefficients and completely unknown nonlinear function. In order to achieve the control objective, a state feedback controller is firstly constructed to ensure that systems under consideration satisfy uniform Lyapunov property. Then, based on output feedback domination method,...
STT-MRAM is a promising non-volatile memory technology for building large LLCs (Last Level Caches). Multi-level cell (MLC) STT-MRAM can further enlarge cache capacity with reduced per bit cost. However, due to fast technology scaling, STT-MRAM, in particular, MLC STT-MRAM, suffers from significant read errors, including read disturbance errors and sensing errors, which lead to unreliable accesses...
To reduce conservative timing margin, many timing-error detection techniques by monitoring selected critical paths had been proposed. However, traditional adaptive methods incur significant area overheads and cannot prevent the error that is forming in the current clock cycle. In this paper, a low-overhead Transition-Detector (TD) with a 9-transistor current sensing circuit is proposed. TDs are inserted...
The purpose of this study is to determine the sleep status of college students. 12 volunteers(mean ± SE: 22 ± 2 yr, 170 ± 10cm, 61 ± 16 kg) take part in the research. The sleep situation are got by the waistband sleep monitor system. The average heart rate of the whole night, the average breath rate of the whole night, bed time, get up time, onset latency, total sleep time are monitored. These important...
With the rapidly increasing applications of deep learning, LSTM-RNNs are widely used. Meanwhile, the complex data dependence and intensive computation limit the performance of the accelerators. In this paper, we first proposed a hybrid network expansion model to exploit the finegrained data parallelism. Based on the model, we implemented a Reconfigurable Processing Unit(RPU) using Processing In Memory(PIM)...
This paper proposes a new continuous terminal sliding mode guidance law for three-dimensional missile guidance system under maneuvering target with consideration of second-order autopilot dynamics. The proposed guidance law not only guarantees the azimuth rate and elevation rate of guidance system converge to zero in finite time but also ensures the continuity of control action. Simulations of a practical...
The estimation of state-of-charge (SOC) is the fundamental technology to improve battery working status and life. But severe external environment, dynamic nonlinearity of battery model, and measurement errors make SOC estimation challengable. In this paper, a multi-strategy probabilities based fusion method is proposed. It can combine the dynamic tracking capability of Proportional-Integral Observer...
Sensor networks have been widely used in environment monitoring. Although there are some related works about water pollution monitoring and water pollution source localization using sensor networks, there are many problems which have not been solved so far. In this paper, the current research status on water pollution monitoring and water pollution source localization in sensor networks is illustrated...
Antimonene is a new 2D material and its applications in gas sensors are still deficient. We took a first principles investigation of three small organic molecules (formaldehyde, acetone, and methane) adsorption on the antimonene. We found that only the adsorption of the acetone molecule could induce a moderate adsorption energy, obvious charge transfer, and large work function change, which are suitable...
Recent studies showed that DRAM restore time degrades as technology scales, which imposes large performance and energy overheads. This problem, prolonged restore time (PRT), has been identified by the DRAM industry as one of three major scaling challenges.This paper proposes DrMP, a novel fine-grained precision-aware DRAM restore scheduling approach, to mitigate PRT. The approach exploits process...
Sub-threshold circuit is a promising circuit design style for IoT application, but the timing closure, especially hold timing fixing is a big challenge for designers. This paper proposes a mathematical method to estimate the number of insertion inverters/buffers for hold timing fixing in each short path. Firstly, the distribution of path delay is rigorously proved to be lognormal distribution in the...
In this paper, a hybrid space time adaptive processing (STAP) algorithm of direct data domain (DDD) approach and cost function reconstruction is presented to provide a solution to sample support problem at a low cost of space-time aperture loss. The correlation matrix estimated in DDD approach is partitioned into sub-matrices and two equivalent cost functions are reconstructed. By iteratively solving...
Racetrack memory (RM, also known as DWM (domain wall memory)) is an emerging memory technology that has many advantages such as low power, high density, and low access latency. Recent studies have shown that it is promising to architect RM as last level cache (LLC). Given that a RM track consists of m domains (for storing m bits data) and n access heads (1≤n<m), one RM access often requires multiple...
In this paper, a two stage STAP algorithm in non-homogeneous environment is proposed to suppress discrete interference. The training data is pre-processed to eliminate the target component in the first stage. And then rank-reduced STAP algorithm based on sub-matrices is utilized to suppress the discrete interference at a low computational cost as the second stage. The effectiveness of the proposed...
A novel distributed approach based on a factor graph framework, using inter-satellite ranging measurements to achieve simultaneous autonomous orbit determination and time synchronization of networked satellites, is proposed. In order to reduce computation complexity as well as communication overhead considering limited resource constraints onboard, we present a parametric message passing scheme with...
Set the date range to filter the displayed results. You can set a starting date, ending date or both. You can enter the dates manually or choose them from the calendar.