Deep neural networks (DNNs) have played an increasingly important role in various areas such as computer vision and voice recognition. While training and validation become gradually feasible with high-end general-purpose processors such as graphical processor units (GPU), high throughput inferences in embedded hardware platforms with low hardware resources and power consumption efficiency are still challenging. Binarized neural networks (BNNs) are emerging as a promising method to overcome these challenges by reducing bit widths of DNN data representations with many optimal prior solutions. However, accuracy degradation is a considerable problem of the BNN, compared to the same architecture with full precision, while the binary neural networks still contain significant redundancy for optimization. In this paper, to address the limitations, we implement a streaming accelerator architecture with three optimization techniques: pipelining-unrolling for streaming each layer, weight reuse for parallel computation, and MAC (multiplication-accumulation) compression. Our method first constructs streaming architecture by pipelining-unrolling method to maximize throughput. Next, the weight reuse method with the K-mean cluster is applied to reduce the complexity of the popcount operation. Finally, MAC compression reduces hardware resources used for remaining computation on MAC operations. The implemented hardware accelerator integrated into a state-of-the-art field programable gate array (FPGA) provides the maximum performance of the classification at 1531k frames per second with 98.4% accuracy for the MNIST dataset and 205K frame per second with 80.2% accuracy for the Cifar-10 dataset. Besides, the proposed design’s ratio FPS/LUTs is approximately 57 (MNIST) and 0.707 (Cifar-10), which is much lower than the state-of-the-art design with a comparable throughput and inference accuracy.
16 December 2024 in IEEE Transactions on Network and Service Management
In this paper, cyber-attack prevention for the prosumer-based electric vehicle (EV) charging stations (EVCSs) is investigated, which covers two aspects: 1) cyber-attack detection on prosumers’ network traffic (NT) data, and 2) cyber-attack intervention. To establish an effective prevention mechanism, several challenges need to be tackled, for instance, the NT data per prosumer may be non-independent and identically distributed (non-IID), and the boundary between benign and malicious traffic becomes blurred. To this end, we propose an edge-assisted federated prototype knowledge distillation (E-FPKD) approach, where each client is deployed on a dedicated local edge server (DLES) and can report its availability for joining the federated learning (FL) process. Prior to the E-FPKD approach, to enhance accuracy, the Pearson Correlation Coefficient is adopted for feature selection. Regarding the proposed E-FPKD approach, we integrate the knowledge distillation and prototype aggregation technique into FL to deal with the non-IID challenge. To address the boundary issue, instead of directly calculating the distance between benign and malicious traffic, we consider maximizing the overall detection correctness of all prosumers (ODC), which can mitigate the computational cost compared with the former way. After detection, a rule-based method will be triggered at each DLES for cyber-attack intervention. Experimental analysis demonstrates that the proposed E-FPKD can achieve the largest ODC on NSL-KDD, UNSW-NB15, and IoTID20 datasets in both binary and multi-class classification, compared with baselines. For instance, the ODC for IoTID20 obtained via the proposed method is separately 0.3782% and 4.4471% greater than FedProto and FedAU in multi-class classification.
Partial discharge (PD) is a major indicator of various failures in power grid systems. PD exhibits a physical occurrence where a localized electrical discharge happens in insulation materials. This phenomenon causes damage to the insulating parts and, in various circumstances, leads to complete insulation breakdown. As a consequence, it can produce more costly outcomes such as abrupt outages or lost production. Therefore, PD detection plays a vital role in preventing insulation failure. In this work, we propose a system that utilizes heterogeneous sensors for the PD detection along with multi-sensor interface, real-time advanced denoise processing, flexible system operation, and Bluetooth-low-energy (BLE)-based ad hoc communication. Among the variety of heterogeneous sensors, several are developed by the application of nanomaterials and nanotechnology, thus outperforming the regular types. The proposed system successfully identifies the presence of PD from sensor signals using a microprocessor-based processing system and effectively performs an advanced denoising technique based on the wavelet transform through field-programmable-gate-array (FPGA)-based programmable logics. With the development of the system, we aim to achieve a solution with low cost, high flexibility and efficiency, and ease of deployment for the monitoring of power grid systems.
Deep Neural Networks (DNNs) have shown superior accuracy at the expense of high memory and computation requirements. Optimizing DNN models regarding energy and hardware resource requirements is extremely important for applications with resource-constrained embedded environments. Although using binary neural networks (BNNs), one of the recent promising approaches, significantly reduces the design’s complexity, accuracy degradation is inevitable when reducing the precision of parameters and output activations. To balance between implementation cost and accuracy, in addition to proposing specialized hardware accelerators for corresponding specific network models, most recent software binary neural networks have been optimized based on generalized metrics, such as FLOPs or MAC operation requirements. However, with the wide range of hardware available today, independently evaluating software network structures is not good enough to determine the final network model for typical devices. In this paper, an architecture search algorithm based on estimating the hardware performance at the design time is proposed to achieve the best binary neural network models for hardware implementation on target platforms. With the XNOR-net used as a base architecture and target platforms, including Field Programmable Gate Array (FPGA), Graphic Processing Unit (GPU), and Resistive Random Access Memory (RRAM), the proposed algorithm shows its efficiency by giving more accurate estimation for the hardware performance at the design time than FLOPs or MAC operations.
More than ever, autonomous systems that are critical to everyday life, social infrastructure, and social security are increasingly dependent on sophisticated hardware computer systems. As the computer hardware systems become increasingly complex and semiconductor manufacturing technology becomes continuously miniaturized (especially in several nanometer technologies), there is a corresponding increased vulnerability to system malfunction caused by runtime hardware defects. Detecting defects that will be generated in postdeployment time is impossible in the predeployment verification. This is the reason why some lifetime testabilities are necessary, especially for autonomous systems requiring extremely high-reliable functionalities. This article proposes a hardware architecture that extends the pre-deployment mass production verification (such as the scan-test) to the lifetime of the hardware systems, to guarantee the mass production verification level integrity for the hardware systems in their lifetime. In addition, this article explains how to recover hardware functionalities compromised by runtime defects.