scholarly journals Mapping Neural Networks to FPGA-Based IoT Devices for Ultra-Low Latency Processing

Sensors ◽  
2019 ◽  
Vol 19 (13) ◽  
pp. 2981 ◽  
Author(s):  
Maciej Wielgosz ◽  
Michał Karwatowski
Keyword(s):  
Short Term Memory ◽  
Autonomous Driving ◽  
Dense Layer ◽  
Latency Reduction ◽  
Covariance Matrix Adaptation ◽  
Field Programmable ◽  
Programmable Gate Arrays ◽  
Iot Devices ◽  
Access To Knowledge

Internet of things (IoT) infrastructure, fast access to knowledge becomes critical. In some application domains, such as robotics, autonomous driving, predictive maintenance, and anomaly detection, the response time of the system is more critical to ensure Quality of Service than the quality of the answer. In this paper, we propose a methodology, a set of predefined steps to be taken in order to map the models to hardware, especially field programmable gate arrays (FPGAs), with the main focus on latency reduction. Multi-objective covariance matrix adaptation evolution strategy (MO-CMA-ES) was employed along with custom scores for sparsity, bit-width of the representation and quality of the model. Furthermore, we created a framework which enables mapping of neural models to FPGAs. The proposed solution is validated using three case studies and Xilinx Zynq UltraScale+ MPSoC 285 XCZU15EG as a platform. The results show a compression ratio for quantization and pruning in different scenarios with and without retraining procedures. Using our publicly available framework, we achieved 210 ns of latency for a single processing step for a model composed of two long short-term memory (LSTM) and a single dense layer.

scholarly journals FPGA-Aware Scheduling Strategies at Hypervisor Level in Cloud Environments

2016 ◽  
Vol 2016 ◽  
pp. 1-12
Author(s):  
Julio Proaño Orellana ◽  
Blanca Caminero ◽  
Carmen Carrión ◽  
Luis Tomas ◽  
Selome Kostentinos Tesfatsion ◽  
...  
Keyword(s):  
Quality Of Service ◽  
Energy Consumption ◽  
Service Level Agreement ◽  
Service Level ◽  
Consumption Ratio ◽  
Field Programmable ◽  
Related Service ◽  
Programmable Gate Arrays ◽  
Scheduling Strategies

Current open issues regarding cloud computing include the support for nontrivial Quality of Service-related Service Level Objectives (SLOs) and reducing the energy footprint of data centers. One strategy that can contribute to both is the integration of accelerators as specialized resources within the cloud system. In particular, Field Programmable Gate Arrays (FPGAs) exhibit an excellent performance/energy consumption ratio that can be harnessed to achieve these goals. In this paper, a multilevel cloud scheduling framework is described, and several FPGA-aware node level scheduling strategies (applied at the hypervisor level) are explored and analyzed. These strategies are based on the use of a multiobjective metric aimed at providing Quality of Service (QoS) support. Results show how the proposed FPGA-aware scheduling policies increment the number of users requests serviced with their SLOs fulfilled while energy consumption is minimized. In particular, evaluation results of a use case based on a multimedia application show that the proposal can save more than 20% of the total energy compared with other baseline algorithms while a higher percentage of Service Level Agreement (SLA) is fulfilled.

scholarly journals Identify a Spoofing Attack on an In-Vehicle CAN Bus Based on the Deep Features of an ECU Fingerprint Signal

Smart Cities ◽  
2020 ◽  
Vol 3 (1) ◽  
pp. 17-30
Author(s):  
Yun Yang ◽  
Zongtao Duan ◽  
Mark Tehranipoor
Keyword(s):  
Autonomous Vehicles ◽  
Short Term Memory ◽  
Can Bus ◽  
Area Network ◽  
Gate Arrays ◽  
Control Units ◽  
Field Programmable ◽  
Programmable Gate Arrays ◽  
Long Short Term Memory ◽  
High Computational Complexity

An in-vehicle controller area network (CAN) bus is vulnerable because of increased sharing among modern autonomous vehicles and the weak protocol design principle. Spoofing attacks on a CAN bus can be difficult to detect and have the potential to enable devastating attacks. To effectively identify spoofing attacks, we propose the authentication of sender identities using a recurrent neural network with long short-term memory units (RNN-LSTM) based on the features of a fingerprint signal. We also present a way to generate the analog fingerprint signals of electronic control units (ECUs) to train the proposed RNN-LSTM classifier. The proposed RNN-LSTM model is accelerated on embedded Field-Programmable Gate Arrays (FPGA) to allow for real-time detection despite high computational complexity. A comparison of experimental results with the latest studies demonstrates the capability of the proposed RNN-LSTM model and its potential as a solution to in-vehicle CAN bus security.

RESEARCH OF DIGITAL QUBITES FOR HETEROGENEOUS DIGITAL QUANTUM COPROCESSORS

2021 ◽  
Vol 4 (1) ◽  
pp. 91-99
Author(s):  
Valeriy S. Hlukhov
Keyword(s):  
Quantum Algorithms ◽  
Qubit State ◽  
Quantum Computers ◽  
Digital Version ◽  
Gate Arrays ◽  
Field Programmable ◽  
Programmable Gate Arrays ◽  
Random Codes ◽  
Functional Converter

Recently, interest is growing towards real quantum computers, which are analog and probabilistic devices by nature. The interest is also growing to their digital version, both software and hardware. One approach to the construction of real quantum computers is to use quantum chips. The hardware implementation of digital quantum computers involves the use of field programmable gate arrays. A digital quantum coprocessor has already been created which has over a thousand digital qubits and can perform such complex algorithms as a quantum Fourier transformation. The created and working digital quantum coprocessor can already be used to work out various quantum algorithms, algorithms for the interaction of a classic computer and its quantum coprocessor, as well as for research various options for building digital qubits. The purpose of this work is to study the effect of the accuracy of the presentation of the state of digital qubit on the probability of obtaining the correct results of the digital quantum coprocessor. For the study, a heterogeneous digital quantum coprocessor with thirty two digital qubits is selected, which will perform the Fourier quantum transformation. The article describes the basics of building digital quantum coprocessors. Schemes that illustrate the interaction of a classic computer and a quantum coprocessor, the architecture of the coprocessor and the possible structures of its digital qubits are given. Two variants of the coprocessor, homogeneous one with one pseudo-random codes generator and one comparator, and heterogeneous one, with a generator and a comparator in each digital quantum cell, from which digital qubits consist, are shown. Two options for comparators are also shown - with a direct functional converter and with reverse one. In this work, the influence of the length of the qubit state codes of heterogeneous digital quantum coprocessors on the probability of the correct results formation is investigated. It was shown that the probability of obtaining the correct results at the output of the digital heterogeneous coprocessor is sharply (up to fifty percent) improved with a decrease of the qubit state code length, that is, with a decrease in the coprocessor hardware cost. With a length of a code equal to two bits, the quality of the operation of the heterogeneous coprocessor becomes commensurate with the quality of the homogeneous one. The need for additional research in this direction, including with homogeneous coprocessors, is shown.

scholarly journals Resilient Multiuser Session Control in Softwarized Fog-Supported Internet of Moving Thing Systems

Sensors ◽  
2019 ◽  
Vol 19 (12) ◽  
pp. 2766 ◽  
Author(s):  
Helber Wagner da Silva ◽  
Augusto José Venâncio Neto
Keyword(s):  
Impact Analysis ◽  
Autonomous Driving ◽  
Smart Environments ◽  
Control Plane ◽  
Related Approach ◽  
Iot Devices ◽  
And Performance ◽  
Set Up ◽  
Qos Performance

The combination of IoT and mobility promises to open a new frontier of innovations in smart environments, through the advent of the Internet of Moving Things (IoMT) paradigm. In IoMT, an array of IoT devices leverage IP-based mobile connectivity to provide a vast range of data ubiquitously. The IoMT realization will foster smart environments at unprecedented levels, by efficiently affording services and applications whereby today’s technologies make their efficiency unfeasible, such as autonomous driving and in-ambulance remotely-assisted patient. IoMT-supported mission-critical applications push computing and networking requirements to totally new levels that must be met, raising the need for refined approaches that advance beyond existing technologies. In light of this, this paper proposes the Resilient MultiUser Session Control (ReMUSiC) framework, which deploys emerging softwarization and cloudification technologies to afford flexible, optimized and self-organized control plane perspectives. ReMUSiC extends our previous work through the following innovations. A quality-oriented resilience mechanism is capable of responding to network dynamics events (failure and mobility) by readapting IoMT multiuser mobile sessions. A softwarized networking control plane that allows to, at runtime, both fetch current network state and set up resources in the attempt to always keep affected IoMT multiuser mobile sessions best-connected and best-served. A cloudification approach allows a robust environment, through which cloud- and fog-systems interwork to cater to performance-enhanced capabilities. The IoMT’s suitability and performance impacts by ReMUSiC framework use are assessed through real testbed prototyping. Impact analysis in Quality of Service (QoS) performance and perceived Quality of Experience (QoE), demonstrate the remarkable abilities of the ReMUSiC framework, over a related approach, in keeping IoMT multiuser mobile sessions always best-connected and best-served.

Imaging through Scattering Media with a Learning Based Prior

2020 ◽  
Vol 2020 (14) ◽  
pp. 306-1-306-6
Author(s):  
Florian Schiffers ◽  
Lionel Fiske ◽  
Pablo Ruiz ◽  
Aggelos K. Katsaggelos ◽  
Oliver Cossairt
Keyword(s):  
Optimization Problem ◽  
Autonomous Driving ◽  
Scattering Media ◽  
Photon Scattering ◽  
Point Spread ◽  
Spread Function ◽  
Radiative Transport Equation ◽  
Spatio Temporal ◽  
Transient Imaging

Imaging through scattering media finds applications in diverse fields from biomedicine to autonomous driving. However, interpreting the resulting images is difficult due to blur caused by the scattering of photons within the medium. Transient information, captured with fast temporal sensors, can be used to significantly improve the quality of images acquired in scattering conditions. Photon scattering, within a highly scattering media, is well modeled by the diffusion approximation of the Radiative Transport Equation (RTE). Its solution is easily derived which can be interpreted as a Spatio-Temporal Point Spread Function (STPSF). In this paper, we first discuss the properties of the ST-PSF and subsequently use this knowledge to simulate transient imaging through highly scattering media. We then propose a framework to invert the forward model, which assumes Poisson noise, to recover a noise-free, unblurred image by solving an optimization problem.

scholarly journals FPGA-Based Controller for a Hybrid Grid-Connected PV/Wind/Battery Power System with AC Load

Energies ◽  
2021 ◽  
Vol 14 (8) ◽  
pp. 2108
Author(s):  
Mohamed Yassine Allani ◽  
Jamel Riahi ◽  
Silvano Vergura ◽  
Abdelkader Mami
Keyword(s):  
Fuzzy Logic ◽  
Hybrid System ◽  
High Voltage ◽  
Field Programmable Gate Arrays ◽  
Energy System ◽  
Maximum Power ◽  
Hybrid Energy ◽  
Gate Arrays ◽  
Field Programmable ◽  
Programmable Gate Arrays

The development and optimization of a hybrid system composed of photovoltaic panels, wind turbines, converters, and batteries connected to the grid, is first presented. To generate the maximum power, two maximum power point tracker controllers based on fuzzy logic are required and a battery controller is used for the regulation of the DC voltage. When the power source varies, a high-voltage supply is incorporated (high gain DC-DC converter controlled by fuzzy logic) to boost the 24 V provided by the DC bus to the inverter voltage of about 400 V and to reduce energy losses to maximize the system performance. The inverter and the LCL filter allow for the integration of this hybrid system with AC loads and the grid. Moreover, a hardware solution for the field programmable gate arrays-based implementation of the controllers is proposed. The combination of these controllers was synthesized using the Integrated Synthesis Environment Design Suite software (Version: 14.7, City: Tunis, Country: Tunisia) and was successfully implemented on Field Programmable Gate Arrays Spartan 3E. The innovative design provides a suitable architecture based on power converters and control strategies that are dedicated to the proposed hybrid system to ensure system reliability. This implementation can provide a high level of flexibility that can facilitate the upgrade of a control system by simply updating or modifying the proposed algorithm running on the field programmable gate arrays board. The simulation results, using Matlab/Simulink (Version: 2016b, City: Tunis, Country: Tunisia, verify the efficiency of the proposed solution when the environmental conditions change. This study focused on the development and optimization of an electrical system control strategy to manage the produced energy and to coordinate the performance of the hybrid energy system. The paper proposes a combined photovoltaic and wind energy system, supported by a battery acting as an energy storage system. In addition, a bi-directional converter charges/discharges the battery, while a high-voltage gain converter connects them to the DC bus. The use of a battery is useful to compensate for the mismatch between the power demanded by the load and the power generated by the hybrid energy systems. The proposed field programmable gate arrays (FPGA)-based controllers ensure a fast time response by making control executable in real time.

Sign in / Sign up

Export Citation Format

Share Document