Future processor chips will not be limited by the transistor resources, but mainly constrained energy efficiency. Reconfigurable fabrics bring higher efficiency than CPUs via customized hardware that adapts to user applications. Among different reconfigurable fabrics, coarse-grained arrays (CGRAs) can even more efficient fine-grained FPGAs when bit-level customization is necessary in target CGRAs were originally developed era resources critical Previous work shares among operations modulo...

Human physical activity recognition is widely used in medical diagnosis, well-being management, and rehabilitation treatment. In spite of various Internet Things (IoT) designs available the literature, power resources often limit lifetime IoT. Regarding this weakness, paper develops a new motion sensor system wearable IoT (WIoT) for human without any signal conditioning circuits. The triboelectricity-based model explored designing sensor. It enables to collect signals caused by activities...

The recent development of non-volatile memory (NVM), such as spin-torque transfer magnetoresistive RAM (STT-RAM) and phase-change (PRAM), with the advantage low leakage high density, provides an energy-efficient alternative to traditional SRAM in cache systems. We propose a novel reconfigurable hybrid architecture (RHC), which NVM is incorporated last-level together SRAM. RHC can be reconfigured by powering on/off SRAM/NVM arrays way-based manner. In this work, we discuss both circuit design...

Many researchers have argued that data mining can improve the performance of intrusion detection system. So as one important techniques mining, clustering is an means for detection. Due to disadvantages traditional methods detection, this paper presents a graph-based algorithm by using outlier method based on local deviation coefficient (LDCGB). Compared other clustering, unnecessary initial cluster number. Meanwhile, it robust in outlier's affection and able detect any shape rather circle...

The recent development of non-volatile memory (NVM), such as spin-torque transfer magnetoresistive RAM (STT-RAM) and phase-change (PRAM), with the advantage low leakage high density, provides an energy-efficient alternative to traditional SRAM in cache systems. We propose a novel reconfigurable hybrid architecture (RHC), which NVM is incorporated last-level together SRAM. RHC can be reconfigured by powering on/off SRAM/NVM arrays way-based manner. In this work, we discuss both circuit design...

In this paper, a combined static and dynamic scheme is proposed to optimize the block placement for endurance energy-efficiency in hybrid SRAM STT-RAM cache. With scheme, maximized while performance maintained. We use compiler provide hints guide initial data placement, hardware correct based on run-time cache behavior. Experimental results show that improves by 23.9x 5.9x compared pure optimizations respectively. Furthermore, system energy can be reduced 17% optimization through minimizing writes.

Scratchpad memory (SPM) has been utilized as prefetch buffer in embedded systems and parallel architectures to hide access latency. However, the impact of reuse pattern on SPM prefetching not fully investigated. In this paper we quantify efficiency propose a reuse-aware (RASP) scheme. The average performance energy improvements are 15.9% 22.0% over cache prefetching, 12.9% 31.2% prefetch-only management, 18.5% 10% DRDU [1] with support.

Electrocardiogram (ECG) sensing is an important application for the diagnosis of cardiovascular diseases. Recently, driven by emerging technology wearable electronics, massive ECG sensors are developed, which however brings additional sources noise contamination on signals from these sensors. In this paper, we propose a new low-distortion adaptive Savitzky-Golay (LDASG) filtering method denoising based discrete curvature estimation, demonstrates better performance than state art denoising....

<title>Abstract</title> <bold>Bcakground:</bold> Backcourt forehand clear stroke is a fundamental technique in badminton, yet the biomechanical differences between novice and experienced players remain in-completely understood. This study aimed to compare kinematic characteristics of body segment acceleration during backcourt these two groups. <bold>Methods: </bold>Forty-five participants (22 with less than 2 years experi-ence 23 over 5 competitive experience) were equipped wireless...

Backcourt forehand clear stroke is a fundamental technique in badminton, but the biomechanical differences between novice and experienced players are not fully understood. This study aimed to compare kinematic characteristics of body segment acceleration during backcourt these two groups. Forty-five participants (22 with less than 2 years experience 23 over 5 competitive experience) were placed wireless accelerometers on key segments (foot, calf, thigh, hip, shoulder, upper arm, forearm)....

By reconfiguring part of the cache as software-managed scratchpad memory (SPM), hybrid caches manage to handle both unknown and predictable access patterns. However, existing provide a flexible partitioning SPM without considering adaptation run-time behavior. Previous set balancing techniques are either energy-inefficient or require serial tag data array access. In this paper an adaptive is proposed dynamically remap blocks from high-demand sets low-demand sets. This achieves 19%, 25%, 18%...

By reconfiguring part of the cache as software-managed scratchpad memory (SPM), hybrid caches manage to handle both unknown and predictable access patterns. However, existing provide a flexible partitioning SPM without considering adaptation run-time behavior. Previous set balancing techniques are either energy-inefficient or require serial tag data array access. In this paper an adaptive is proposed dynamically remap blocks from high-demand sets low-demand sets. This achieves 19%, 25%, 18%...

Accelerator-rich platforms demonstrate orders of magnitude improvement in performance and energy efficiency over software, yet they lack adaptivity to new algorithms can see low accelerator utilization. To address these issues we propose CAMEL: Composable Microprocessor Enhanced for Longevity. CAMEL features programmable fabric (PF) extend the use ASIC composable accelerators supporting that are beyond scope baseline platform. Using a combination hardware extensions compiler support, on...

In cognitive networks, since the secondary users are not licensed owners of spectrum, they can only access spectrum opportunistically and cooperatively to exploit best effort throughput for themselves while guaranteeing that their transmissions would break QoS performance primary communications. this paper, we consider network in which have delay sensitive services, e.g., voice or video take average constraint as users' requirement. Three cooperative schemes under such proposed, based on...

A fundamental challenge in realizing body-worn sensors is providing an effective and long-lasting power supply. Issues regarding batteries have prompted researchers to investigate powering devices by extracting energy from the inertial movement of human body. While previous studies investigated effect generator location wearer activity on harvestable power, they not considered orientation generator; this focus paper. Acceleration data collected across a sample population (10 participants)...

The goal of this study is to design a novel triboelectric motion sensor in wearable body network for human activity recognition. Physical recognition widely used well-being management, medical diagnosis and rehabilitation. Other than traditional accelerometers, we system based on triboelectrification. can be easily attached collect signals caused by physical activities. experiments are conducted five common data: sitting standing, walking, climbing upstairs, downstairs, running. k-Nearest...

The insufficient contact between triboelectric surfaces has been a key problem faced by wind driven nanogenerator (TENG) devices. We have designed and fabricated an arc‐shaped‐TENG (AS‐TENG) based on the single‐side fixed working mode. AS‐TENG includes two arc structures with fluorinated ethylene propylene (FEP) film stainless steel layer. Due to inherent stiffness of layer structure design, bent can fully FEP films attached structure, enhancing effects triboelectrification electrostatic...

Low probability of intercept (LPI) radar waveform recognition is one the crucial functions in electronic intelligence systems. Advances artificial promote performance LPI with various signal features defined analytical expressions. However, noisy still a challenge for traditional approaches even noise elimination techniques, especially heavily contaminated signals. Recently, adaptive analysis such as variational mode decomposition (VMD) and empirical (EMD) provide potential methods that...

Accelerator-rich platforms demonstrate orders of magnitude improvement in performance and energy efficiency over software, yet they lack adaptivity to new algorithms can see low accelerator utilization. To address these issues we propose CAMEL: Composable Microprocessor Enhanced for Longevity. CAMEL features programmable fabric (PF) extend the use ASIC composable accelerators supporting that are beyond scope baseline platform. Using a combination hardware extensions compiler support, on...

Inertial energy harvesting is an emerging technology that can power electronic devices using scavenged from the motion of human body. Owing to relatively low frequencies associated with body (<3 Hz), generated electrical typically in range a few μW; hence transduction must be optimized. Previous studies have investigated effect activity and harvester location on obtained power; this work evaluates how also affected by harvester's orientation. Ten participants performed walking running...

Programming difficulty is a key challenge to the adoption of FPGAs as general high-performance computing platform. In this paper we present CMOST, an open-source automated compilation flow that maps C-code for acceleration. CMOST establishes unified framework integration various system-level optimizations and different hardware platforms. We also several novel techniques on integrating in including task-level dependence analysis, block-based data streaming, SDF generation. Experimental...