Low-cost Accelerator Research Articles

Hardware Accelerator Design of DCT Algorithm With Unique-Group Cosine Coefficients for Mel-Scale Frequency Cepstral Coefficients

This study presents a compact <i>L</i>-points discrete cosine transform (DCT) hardware accelerator for <i>M</i>-points Mel-scale Frequency Cepstral Coefficients (MFCC). The main contributions of this work can be summarized as 1) proposing an algorithm with lower complexity; 2) achieving higher accuracy performance; 3) implementing a low-cost accelerator with a unique group of cosine coefficients. For algorithm derivation, the proposed method converts the original formula into the type IV of discrete cosine transform (DCT-IV) with a preprocessing procedure. The kernel computation of DCT-IV can be further derived into the same cosine multiplication with the proposed preprocessing. Therefore, a total of (<i>M</i>-1) (<i>L</i>-1) additions, (<i>M</i>-1) <i>L</i> multiplications, and <i>L</i> coefficients are required for the computation. Compared with Jo <i>et al</i>.&#x2019;s algorithm, the proposed method respectively reduces the number of additions and multiplications by 42.32 % and 41.67 %. Instead, the number of coefficients is increased by 33.33 %. Moreover, the proposed algorithm exhibits a higher peak signal-to-noise ratio (PSNR) value which is achieved at 90.1dB with a 16-bit coefficient word length. For hardware realization, the FPGA implementation results show that it can operate at a clock rate of 135.85 MHz and requires only 113 combinational elements, 87 registers, 3 DSP multipliers, 64&#x00D7;16 bits RAM and 32&#x00D7;16 bits ROM. Overall, it would be a good choice for integrating MFCC applications in the future.

IoT device-based data acquisition system with on-board computation of variables for cow behaviour recognition

Improved raw data handling is one of the requirements that research should fulfil in the design of IoT device-based data acquisition systems for enhancing the overall system performance. In this study, the system was composed of low-cost acceleration sensors broadcasting to a Raspberry PI.The main aim of this study was to develop firmware for the acceleration sensors with the purpose of maximising the battery life-time and minimising the information loss during data transfer while allowing high accuracy in the discrimination of the cow lying-standing behaviour. The attainment of these goals was achieved by using an aggregated accelerometer variable, computed on-board of the IoT device, together with the idea of saving in the payload the most recent variables.The comparisons conducted between the base firmware installed on the device and the new firmware developed showed the outstanding performance of the latter in terms of Raspberry PI CPU usage, storage memory occupation, and packet loss. The algorithms of lying-standing behaviour discrimination were implemented and assessed by using the new firmware, producing excellent values of the most used accuracy measures.

Scaling laws for laser-driven ion acceleration from nanometer-scale ultrathin foils.

Laser-driven ion acceleration has attracted global interest for its potential towards the development of a new generation of compact, low-cost accelerators. Remarkable advances have been seen in recent years with a substantial proton energy increase in experiments, when nanometer-scale ultrathin foil targets and high-contrast intense lasers are applied. However, the exact acceleration dynamics and particularly the ion energy scaling laws in this novel regime are complex and still unclear. Here, we derive a scaling law for the attainable maximum ion energy from such laser-irradiated nanometer-scale foils based on analytical theory and multidimensional particle-in-cell simulations, and further show that this scaling law can be used to accurately describe experimental data over a large range of laser and target parameters on different facilities. This provides crucial references for parameter design and experimentation of the future laser devices towards various potential applications.

Dynamic Monitoring of an Existing Reinforced Concrete Building in Naples Port Area

Introduction: The conservation and preservation of existing buildings, in particular public infrastructures, is currently a significant issue in Italy and Europe, considering their strategic role and the risk represented by human losses, management issues and also economic disruption in case of collapse. In this context, the interest in conservation is not restricted to monumental or artistical buildings but also includes the several buildings composing the Italian infrastructural heritage, which in many cases are made of reinforced concrete and show signs of ageing after half a century and more from their construction. Methods: On the basis of these premises, in the present paper, a preliminary investigation on an infrastructural case study located in Naples port is presented. Such a study is part of a research activity aimed at defining critical structural issues of the central administrative building of the Central Tyrrhenian Sea Port System Authority. Results: A system of high-sensitivity and low-cost MEMS acceleration sensors were installed on the structure, with the aim of investigating its dynamic behaviour. A Finite Element model of the building was created, including information about material properties and cross-section details from prior experimental activities. A model updating procedure was carried out, based on the dynamic data collected by the monitoring system and post-processed to estimate the fundamental frequencies. Conclusion: This has allowed highlighting the main features of the dynamic response of the building, and the critical role played by deformability of infill panels and floors on the modal properties of the structure.

아두이노 기반 MEMS 센서의 동적 계측 성능 연구

Prior to applying the Arduino-based acceleration sensor for structural health monitoring of low- and mid-floor structures, we would like to confirm the vibration measurement performance of the Arduino-based low-cost acceleration sensor and clarify its limitations. A commercial acceleration sensor and a low-cost acceleration sensor were installed in the excitation device, and the response was measured by excitation of sin waves of various cases. The response results were compared with the measurement performance of commercial acceleration sensors and low-cost sensors. In the estimation of the excitation frequency, similar results were shown in most cases, but in the case of the magnitude estimation result, when the excitation amplitude is less than 1Hz/100mm, the estimated Magnitude value was inaccurate. In all subsequent cases, the results of the estimation of the magnitude of the low-cost acceleration sensor and the commercial acceleration sensor were similar. Based on these experimental results, it is judged that the Arduino-based low-cost acceleration sensor is generally effective in dynamic measurement of buildings.

Preliminary comparison of vibration measurement accuracy between a low cost, portable acceleration measurement unit and a gold-standard accelerometer system

Characterization of Whole-Body Vibration (WBV) exposure is important for the development and evaluation of mitigation strategies for occupational WBV. However, barriers associated with cost and complexity limit widespread use of current gold-standard accelerometer (GSA) measurement systems. Small, simple, low cost Acceleration Measurement Unit (AMU) devices with built-in batteries and memory potentially allow for more efficient collection of WBV data, but the measurement accuracy of these devices need to be evaluated. Using known acceleration inputs generated by an accelerometer calibrator and field collected vibration profiles simulating real-world vibration exposures, the measurements of an AMU device and GSA system were compared. Analysis of accelerometer calibrator data showed no significant difference in weighted acceleration (Aw) measurements between the systems (mean difference −0.001 m/s2, p = 0.95). In field collected vibration profile testing, differences in Aw measurements were small (0.06 m/s2, 4.6%). These results suggest the AMU evaluated in this study may be acceptable for measuring occupational WBV.

Design and Simulation of MEMS Accelerometer for the Application of Seismic vibrations

Aim: Using accelerometer to detect the vibrations produced during earthquake. Backgorund: The big challenge facing the Earthquake Early Warning Systems (EEWS) is to accurately detect seismic tremors at the edge of the early warning system or outside of the earthquake seismic system.[1]. During the last years in development, early warning earthquake (EEW) has proved to be one of the potential means of disaster reduction. Objectives: During the last years in development, early warning earthquake (EEW) has proved to be one of the potential means of disaster reduction. The EEW network intensity usually determines the efficiency of both the EEW system An upgraded low-cost Micro Electro Mechanical System MEMS based device called GL-P2B to reduce sensor costs and establish a dense EEW In the this study, a dense EEW network was also developed to maximize the density of the seismic network for EEW applications. A tri-axial sensor with high-dynamic MEMS called GL-P2B was developed. This sensor was upgraded from the previous version by enhancing the CPU's processing capability and correcting certain errors found during the initial test cycle. Methodology: An autonomous sensor with an acceleration sensor is launched in this article First; we systematically evaluated a series of acceleration sensors to select the most suitable acceleration sensor using mems by analyzing their quality and accuracy, and then created a dedicated tool that can monitor earthquakes. Our result shows that an earthquake can be detected with a low-cost acceleration detector, thereby enhancing the safety of vulnerable groups against earthquakes. Results: There have been two earthquakes recently: Gyeongju Earthquake in magnitude 5.6 and Pohang South Korea Earthquake in magnitude 5.4 respectively in 206 and 2017. As a result, earthquake detection and response in a relatively short period of time was highly demanded. The use of seismic smart phone systems is one solution, but it is expensive to use a smartphone to track seismic events and to allow participants to use their smartphones as just another seismic detector. However, because of the existence of smartphones which are used extensively in our everyday lives, a large percentage of smartphones are useless for earthquakes to be detected. We also built a clever tool in this paper that can be mounted to a wall or to a ceiling. The system is only fitted with sensors that include an accelerator and the price in comparison to intelligent telephones is very small. Conclusion: In this paper, we have used an appropriate capacitive sensing technology to create the accelerometer. We have designed and simulated the accelerometer to measure the seismic vibrations by FEM Tool. The simulated results show that the unit is modelled on a 3Hz resonant frequency and therefore it senses the acceleration between 2 and 8 Hz.

A Smart IoT Device for Detecting and Responding to Earthquakes

The advancement of hardware and software technologies makes it possible to use smartphones or Internet of things for monitoring environments in realtime. In recent years, much effort has been made to develop a smartphone based earthquake early warning system, where low-cost acceleration sensors inside a smartphones are used for capturing earthquake signals. However, because a smartphone comes with a powerful CPU, spacious memory, and several sensors, it is waste of such resources to use it only for detecting earthquakes. Furthermore, because a smartphone is mostly in use during the daytime, the acquired data cannot be used for detecting earthquakes due to human activities. Therefore, in this article, we introduce a stand-alone device equipped with a low-cost acceleration sensor and least computing resources to detect earthquakes. To that end, we first select an appropriate acceleration sensor by assessing the performance and accuracy of four different sensors. Then, we design and develop an earthquake alert device. To detect earthquakes, we employ a simple machine learning technique which trains an earthquake detection model with daily motions, noise data recorded in buildings, and earthquakes recorded in the past. Furthermore, we evaluate the four acceleration sensors by recording two realistic earthquakes on a shake-table. In the experiments, the results show that the developed earthquake alert device can successfully detect earthquakes and send a warning message to nearby devices, thereby enabling proactive responses to earthquakes.

Electromagnetic wave propagation in a fast pulse line ion accelerator.

The pulse line ion accelerator (PLIA) is a low-cost accelerator concept originally designed to accelerate heavy ions. Our group has been investigating the use of PLIA to accelerate light ions and believe a multi-stage PLIA could be useful for short half-life PET isotope production. The goal of this work was to develop a single prototype fast PLIA structure and demonstrate electromagnetic wave propagation using a high-voltage pulser. A 1.6m fast PLIA structure (wave speed>107 m/s) was constructed along with a high-voltage, sinusoidal pulse generator. The latter uses capacitive voltage doubling and spark gap switching. A step-up transformer couples voltage from the pulser to the PLIA coil. Voltage measurements on the coil were made in air using a high-voltage resistive probe, while capacitive probes placed along the length of the PLIA were used to measure wave propagation with the PLIA structure filled with transformer oil. Voltage measurements acquired on the primary and secondary coils of the transformer coupler in air demonstrated a peak-to-peak voltage step-up of 4.2 relative to the pulser DC charging voltage. The maximum voltage time-rate-of-change on the PLIA coil was 0.76×1013 V/s. Capacitive probe measurements indicated voltage oscillations on the PLIA coil with half-period equal to 43±0.9ns and wave speed (with oil) of 1.2×107 m/s. Average and peak accelerating gradients were conservatively estimated to be 0.44 and 0.60MV/m, respectively, with a charging voltage of 55kV. Wave propagation was demonstrated at these gradients without flashover at a vacuum pressure of 9×10-6 Torr. Submerging the pulser in oil would allow for charging voltages up to 150kV and produce accelerating gradients >1.2MV/m. Use of a multi-stage, fast PLIA for light ion acceleration could provide a low-cost complement to cyclotrons for the production of short half-life isotopes used for PET imaging, including carbon-11, nitrogen-13, oxygen-15, and fluorine-18.

IN-PROCESS CONTROL WITH A SENSORY TOOL HOLDER TO AVOID CHATTER

Hydro-, steam- and gas- turbines, aircraft components or moulds are milled parts with complex geometries and high requirements for surface quality. The production of such industry components often necessitates the use of long and slender tools. However, instable machining situations together with work pieces with thin wall thickness can lead to dynamic instabilities in the milling processes. Resulting chatter vibrations cause chatter marks on the work piece surface and have influence on the tool lifetime. In order to detect and avoid the occurrence of process instabilities or process failures in an early stage, the Institute for Production Engineering and Laser Technology (IFT) developed an active control system to allow an in-process adaption of machining parameters. This system consists of a sensory tool holder with an integrated low cost acceleration sensor and wireless data transmission under real time conditions. A condition monitoring system using a signal-processing algorithm, which analyses the received acceleration values, is coupled to the NC- control system of the machine tool to apply new set points for feed rate and rotational speed depending on defined optimisation strategies. By the implementation of this system process instabilities can be avoided.

Design considerations for a pulse line ion accelerator (PLIA)-based PET isotope generator.

Positron emission tomography (PET) imaging remains limited due to the cost associated with on-site production of short half-life, positron-emitting isotopes. In this work, we examine the use of a pulse line ion accelerator (PLIA) to accelerate protons for single-dose PET isotope production. Time-domain electromagnetic field and particle-in-cell (PIC) simulations were performed for a 1.5-m PLIA structure modeled in CST Microwave Studio and Particle Studio software. Scaled measurements from a kV ramp-pulse generator were incorporated into the simulations to accelerate a 1 A, 50 ns proton beam injected with initial kinetic energy of 100 keV. A uniform, 3 T, solenoidal magnetic field was used to provide external beam focusing. Electromagnetic fields and particle phase space were recorded with ns resolution for subsequent analysis. Applying a scaled 100 kV, 20 ns ramped voltage pulse to the PLIA input resulted in a travelling electric field wave inside the structure with accelerating gradient of 2.4 MV/m. The observed wave speed was 1.2 × 107 m/s and is in good agreement with theoretical predictions. Phase space monitors showed both acceleration and bunching of the proton beam, with a maximum kinetic energy of 2.5 MeV, observed at the exit of the single PLIA stage. Evaluation of beam position monitors at different locations in the accelerator showed bunch compression and minimal beam divergence, illustrating that the 3 T field is adequate to contain the beam over the length of the PLIA structure. Simulations performed in this work demonstrate the feasibility of using a PLIA structure to accelerate protons with MV/m level gradients. Combining several PLIA stages in series could allow for a low-cost accelerator suitable for dose-on-demand PET isotope production.

Designing Low-Cost Hardware Accelerators for CE Devices [Hardware Matters

Hardware accelerators for signalprocessing applications, such as finite impulse response (FIR) filters, MPEG motion vectors, and autoregressive filters, are frequently used in consumer electronics (CE) devices. Hardware accelerators are typically used to improve performance, but there are tradeoffs in their use, such as silicon area and energy. High-level synthesis (HLS) of hardware accelerator design involves critical decision-making steps, such as the design space exploration (DSE) of resources, scheduling, allocation, and binding. Exploring an optimal hardware design solution comprising an array of resource types is a complex problem and requires intelligent decision making in an automated manner. The exploration process of an optimal design solution is considered a difficult problem, as it involves tradeoffs between conflicting parameters of hardware area and execution time among innumerable candidate design variants.

CU Partition Mode Decision for HEVC Hardwired Intra Encoder Using Convolution Neural Network.

The intensive computation of High Efficiency Video Coding (HEVC) engenders challenges for the hardwired encoder in terms of the hardware overhead and the power dissipation. On the other hand, the constrains in hardwired encoder design seriously degrade the efficiency of software oriented fast coding unit (CU) partition mode decision algorithms. A fast algorithm is attributed as VLSI friendly, when it possesses the following properties. First, the maximum complexity of encoding a coding tree unit (CTU) could be reduced. Second, the parallelism of the hardwired encoder should not be deteriorated. Third, the process engine of the fast algorithm must be of low hardware- and power-overhead. In this paper, we devise the convolution neural network based fast algorithm to decrease no less than two CU partition modes in each CTU for full rate-distortion optimization (RDO) processing, thereby reducing the encoder's hardware complexity. As our algorithm does not depend on the correlations among CU depths or spatially nearby CUs, it is friendly to the parallel processing and does not deteriorate the rhythm of RDO pipelining. Experiments illustrated that, an averaged 61.1% intraencoding time was saved, whereas the Bjøntegaard-Delta bit-rate augment is 2.67%. Capitalizing on the optimal arithmetic representation, we developed the high-speed [714 MHz in the worst conditions (125 °C, 0.9 V)] and low-cost (42.5k gate) accelerator for our fast algorithm by using TSMC 65-nm CMOS technology. One accelerator could support HD1080p at 55 frames/s real-time encoding. The corresponding power dissipation was 16.2 mW at 714 MHz. Finally, our accelerator is provided with good scalability. Four accelerators fulfill the throughput requirements of UltraHD-4K at 55 frames/s.

Streaming Elements for FPGA Signal and Image Processing Accelerators

Field-programmable gate array (FPGA) devices boast abundant resources with which custom accelerator components for signal, image, and data processing may be realized; however, realizing high-performance, low-cost accelerators currently demands manual register transfer level design. Software-programmable soft processors have been proposed as a way to reduce this design burden, but they are unable to support performance and cost comparable to custom circuits. This paper proposes a new soft processing approach for FPGA that promises to overcome this barrier. A high-performance, fine-grained streaming processor, known as a streaming accelerator element, is proposed, which realizes accelerators as large-scale custom multicore networks. By adopting a streaming execution approach with advanced program control and memory addressing capabilities, typical program inefficiencies can be almost completely eliminated to enable performance and cost, which are unprecedented among software-programmable solutions. When used to realize accelerators for fast Fourier transform, motion estimation, matrix multiplication, and sobel edge detection, it is shown how the proposed architecture enables real-time performance and with performance and cost comparable with hand-crafted custom circuit accelerators and up to two orders of magnitude beyond existing soft processors.

A low-cost acceleration monitoring system based on dual fiber Bragg gratings

In this paper, we constructed a low-cost acceleration monitoring system based on a rectangle cantilever beam (RCB), two fiber Bragg gratings and the light intensity measurement method. To do this, we investigated the minimum natural frequency of RCB and the dependence relationship of the strain of the measurement point and its change rate with respect to the acceleration amplitude upon the inclination angle of the sensor, and given the quantitative relationship between the amplitude of AC voltage of monitoring system versus the acceleration amplitude and the inclination angle of the sensor. Finally, we have finished a series of vibration monitoring experiments, and the experiment data are consistent with the theoretical equations.

Seismic Data Collection with Shakebox and Analysis Using MapReduce

In this paper we study a seismic sensing platform using Shakebox, a low-noise and low-power 24- bit wireless accelerometer sensor. The advances of wireless sensor offer the potential to monitor earthquake in California at unprecedented spatial and temporal scales. We are exploring the possibility of incorporating Shakebox into California Seismic Network (CSN), a new earthquake monitoring system based on a dense array of low-cost acceleration seismic sensors. Compared to the Phidget/Sheevaplug sensors currently used in CSN, the Shakebox sensors have several advantages. However, Shakebox sensor collects 4K Bytes of seismic data per second, giving around 0.4G Bytes of data in a single day. Therefore how to process such large amount of seismic data becomes a new challenge. We adopt Hadoop/MapReduce, a popular software framework for processing vast amounts of data in-parallel on large clusters of commodity hardware. In this research, the test bed-generated seismic data generation will be reported, the map and reduce function design will be presented, the application of MapReduce on the testbed-generated data will be illustrated, and the result will be analyzed.

Design of Semi-Active Roller Guides for High Speed Elevators

Comfort is an important topic in the elevator industry, and among the different factors which affect it, vibration in the car is one of the most important. Passengers not only perceive this phenomenon as an unpleasant effect, but also as a signal of the elevator safety level, and therefore it is of great importance to mitigate it. Although mature technologies are available for reducing vibration in low speed elevators, the increased number of high buildings and sky-scrappers, forces the development of new technologies for medium and high speed elevators. Active vibration control techniques are widespread used for damping vibration in different industrial environments, but their application to the elevator industry normally imply expensive solutions, with bulky and heavy actuators which highly burden its application. As alternative, the cheaper and lighter semi-active vibration systems based on passive elements capable of modifying their properties in response to external commands represent an appealing option. In this paper, we describe a design method for a semi-active roller guide based on a commercial magneto-rheological damper. Different control strategies based on low cost acceleration sensors are also analyzed.

Dynamic structural health monitoring of a civil engineering structure with a POF accelerometer

Purpose – In this work, the paper aims to demonstrate the feasibility of plastic optical fiber (POF) based accelerometers for the structural health monitoring (SHM) of civil engineering structures based on measurements of their dynamic response, namely to estimate natural frequencies. These sensors use POFs, combining the advantages of the optical technology with the robustness of this particular kind of fiber. The POF sensor output is directly compared with the signal from an electrical sensor, demonstrating the potential use of such sensors in structural monitoring applications. Design/methodology/approach – Within this work, the paper demonstrates the feasibility of using a low-cost acceleration system based on a POF accelerometer on the dynamic monitoring of a civil engineering structure, aiming its natural frequency evaluation, which is a primary parameter to be used in SHM methods and numerical models calibration. Findings – A low-cost POF-based accelerometer was used in the characterization of a civil engineering structural component, located in a building at the University of Aveiro Campus, being used to estimate its natural frequency with a relative error of 0.36 percent, comparatively to the value estimated recurring to a calibrated electronic sensor. Originality/value – Optical fiber sensors take advantage of the fibers properties, such as immunity to electromagnetic interference and electrical isolation. They are very attractive for use in hostile environments, like submerse environments or flammable atmospheres where electrical currents might pose a hazard. The advantages of POF itself should also be considered, like resistance to hash environments, robustness, flexibility, low-cost interrogation units and high numeric aperture (lower cost components). The paper demonstrates the feasibility of using a low-cost acceleration system based on a POF accelerometer on the dynamic monitoring of a civil engineering structure, aiming its natural frequency evaluation.

MrBayes tgMC3: A Tight GPU Implementation of MrBayes

MrBayes is model-based phylogenetic inference tool using Bayesian statistics. However, model-based assessment of phylogenetic trees adds to the computational burden of tree-searching, and so poses significant computational challenges. Graphics Processing Units (GPUs) have been proposed as high performance, low cost acceleration platforms and several parallelized versions of the Metropolis Coupled Markov Chain Mote Carlo (MC3) algorithm in MrBayes have been presented that can run on GPUs. However, some bottlenecks decrease the efficiency of these implementations. To address these bottlenecks, we propose a tight GPU MC3 (tgMC3) algorithm. tgMC3 implements a different architecture from the one-to-one acceleration architecture employed in previously proposed methods. It merges multiply discrete GPU kernels according to the data dependency and hence decreases the number of kernels launched and the complexity of data transfer. We implemented tgMC3 and made performance comparisons with an earlier proposed algorithm, nMC3, and also with MrBayes MC3 under serial and multiply concurrent CPU processes. All of the methods were benchmarked on the same computing node from DEGIMA. Experiments indicate that the tgMC3 method outstrips nMC3 (v1.0) with speedup factors from 2.1 to 2.7×. In addition, tgMC3 outperforms the serial MrBayes MC3 by a factor of 6 to 30× when using a single GTX480 card, whereas a speedup factor of around 51× can be achieved by using two GTX 480 cards on relatively long sequences. Moreover, tgMC3 was compared with MrBayes accelerated by BEAGLE, and achieved speedup factors from 3.7 to 5.7×. The reported performance improvement of tgMC3 is significant and appears to scale well with increasing dataset sizes. In addition, the strategy proposed in tgMC3 could benefit the acceleration of other Bayesian-based phylogenetic analysis methods using GPUs.

Hybrid Positioning through Extended Kalman Filter withInertial Data Fusion

In wireless sensor networks (WSNs), hybrid algorithms are widely used in order to improve the final positioning accuracy. This paper presents a hybrid positioning algorithm which combines time of arrival (TOA) and received signal strength (RSS) measurements using two different radio technologies, ultra wide band (UWB) and ZigBee, respectively. The TOA measurements are used to estimate the distances between a mobile node and a set of anchor nodes. Both UWB- based distance estimates and RSS measurements based on ZigBee are simultaneously processed by an Extended Kalman Filter (EKF). Moreover, a low cost inertial device is also used to acquire acceleration measurements which proved to be useful in order to detect the motion of the mobile node. This information has also been integrated in the EKF algorithm accordingly. The performance of the final hybrid positioning algorithm is compared with the conventional EKF which uses a single type of range measurements, TOA or RSS. Simulation results based on a real measurements campaign, show that the hybrid algorithm significantly improves positioning accuracy. In addition, a further improvement has been achieved by applying the motion detection approach based on inertial measurements performed by the low cost acceleration sensor.