Hardware Limitations Research Articles

A Fast and Accurate Lane Detection Method Based on Row Anchor and Transformer Structure.

Lane detection plays a pivotal role in the successful implementation of Advanced Driver Assistance Systems (ADASs), which are essential for detecting the road's lane markings and determining the vehicle's position, thereby influencing subsequent decision making. However, current deep learning-based lane detection methods encounter challenges. Firstly, the on-board hardware limitations necessitate an exceptionally fast prediction speed for the lane detection method. Secondly, improvements are required for effective lane detection in complex scenarios. This paper addresses these issues by enhancing the row-anchor-based lane detection method. The Transformer encoder-decoder structure is leveraged as the row classification enhances the model's capability to extract global features and detect lane lines in intricate environments. The Feature-aligned Pyramid Network (FaPN) structure serves as an auxiliary branch, complemented by a novel structural loss with expectation loss, further refining the method's accuracy. The experimental results demonstrate our method's commendable accuracy and real-time performance, achieving a rapid prediction speed of 129 FPS (the single prediction time of the model on RTX3080 is 15.72 ms) and a 96.16% accuracy on the Tusimple dataset-a 3.32% improvement compared to the baseline method.

Firmware Updates over the Air via LoRa: Unicast and Broadcast Combination for Boosting Update Speed.

The capacity to update firmware is a vital component in the lifecycle of Internet of Things (IoT) devices, even those with restricted hardware resources. This paper explores the best way to wirelessly (Over The Air, OTA) update low-end IoT nodes with difficult access, combining the use of unicast and broadcast communications. The devices under consideration correspond to a recent industrial IoT project that focuses on the installation of intelligent lighting systems within ATEX (potentially explosive atmospheres) zones, connected via LoRa to a gateway. As energy consumption is not limited in this use case, the main figure of merit is the total time required for updating a project. Therefore, the objective is to deliver all the fragments of the firmware to each and all the nodes in a safe way, in the least amount of time. Three different methods, combining unicast and broadcast transmissions in different ways, are explored analytically, with the aim of obtaining the expected update time. The methods are also tested via extensive simulations, modifying different parameters such as the size of the scenario, the number of bytes of each firmware chunk, the number of nodes, and the number of initial broadcast rounds. The simulations show that the update time of a project can be significant, considering the limitations posed by regulations, in terms of the percentage of airtime consumption. However, significant time reductions can be achieved by using the proper method: in some cases, when the number of nodes is high, the update time can be reduced by two orders of magnitude if the correct method is chosen. Moreover, one of the proposed methods is implemented using actual hardware. This real implementation is used to perform firmware update experiments in a lab environment. Overall, the article illustrates the advantage of broadcast approaches in this kind of technology, in which the transmission rate is constant despite the distance between the gateway and the node. However, the advantage of these broadcast methods with respect to the unicast one could be mitigated if the nodes do not run exactly the same firmware version, since the control of the broadcast update would be more difficult and the total update time would increase.

Resource Efficient Deep Learning Hardware Watermarks with Signature Alignment

Deep learning intellectual properties (IPs) are high-value assets that are frequently susceptible to theft. This vulnerability has led to significant interest in defending the field's intellectual properties from theft. Recently, watermarking techniques have been extended to protect deep learning hardware from privacy. These technique embed modifications that change the hardware's behavior when activated. In this work, we propose the first method for embedding watermarks in deep learning hardware that incorporates the owner's key samples into the embedding methodology. This improves our watermarks' reliability and efficiency in identifying the hardware over those generated using randomly selected key samples. Our experimental results demonstrate that by considering the target key samples when generating the hardware modifications, we can significantly increase the embedding success rate while targeting fewer functional blocks, decreasing the required hardware overhead needed to defend it.

Empirical validation of federated learning with YOLO v7-tiny for road sign detection: A simulation-based comparative study

Detecting road signs is a critical component in the development of intelligent driving systems. While centralized machine learning approaches have demonstrated potential in this field, the untapped potential of Federated Learning warrants exploration. This research aims to bridge this gap by examining the feasibility of applying Federated Learning within edge Artificial Intelligence (AI) computing environments for the purpose of road sign detection. Utilizing the You Only Look Once (YOLO) v7-tiny model and a range of experimental parameters demonstrates that Federated Learning is viable and outperforms centralized approaches under specific conditions. The study's empirical analysis highlights the sensitivity of detection accuracy to varying experimental parameters. The study contributes to the existing literature by establishing the efficacy of Federated Learning in road sign detection, particularly in edge AI settings constrained by hardware limitations and privacy concerns. However, the study acknowledges limitations, including the lack of deployment on actual edge AI devices and a restricted range of experimental parameters. Future research should aim for more exhaustive experiments with broader datasets, diverse parameters, and real-world edge AI environments. These findings offer valuable insights for future implementations in intelligent automotive systems.

An event-oriented diffusion-refinement method for sparse events completion

Event cameras or dynamic vision sensors (DVS) record asynchronous response to brightness changes instead of conventional intensity frames, and feature ultra-high sensitivity at low bandwidth. The new mechanism demonstrates great advantages in challenging scenarios with fast motion and large dynamic range. However, the recorded events might be highly sparse due to either limited hardware bandwidth or extreme photon starvation in harsh environments. To unlock the full potential of event cameras, we propose an inventive event sequence completion approach conforming to the unique characteristics of event data in both the processing stage and the output form. Specifically, we treat event streams as 3D event clouds in the spatiotemporal domain, develop a diffusion-based generative model to generate dense clouds in a coarse-to-fine manner, and recover exact timestamps to maintain the temporal resolution of raw data successfully. To validate the effectiveness of our method comprehensively, we perform extensive experiments on three widely used public datasets with different spatial resolutions, and additionally collect a novel event dataset covering diverse scenarios with highly dynamic motions and under harsh illumination. Besides generating high-quality dense events, our method can benefit downstream applications such as object classification and intensity frame reconstruction.

Sequentially Triggering “Time-Bomb” Trojan into Hardware Wired Microsequencer

This paper presents an analysis of a sequential Triggered “Time Bomb” hardware Trojan (HT). Major security concerns have been rising up since malicious modification of hardware or fabrication IC can lead to an altered functional behavior, potentially with disastrous consequences in safety-critical applications. Due to the stealthy nature of the hardware Trojans the conventional design and time verification and post manufacturing testing cannot be readily extended. There are a large number of possible instances and operating modes for the hardware Trojans in a digital system. Since the hardware Trojan insertion can modify the functionality of the digital integrated circuit (IC), alter its behavior, generate denial of service (DoS), the HT threats should be analyzed with maximum importance through the entire lifecycle of the IC.

Experimental and Computational Investigation of Flow Boiling in a 52 μm Hydraulic Diameter Microchannel Evaporator With Inlet Restrictions and Heat Spreading

Abstract Microchannel flow boiling presents an effective thermal management strategy for high heat flux (>1 kW/cm2) devices. Fundamental mechanisms of microchannel flow boiling behaviors are difficult to determine due to macroscopic limitations of experimental hardware. In addition, flow stabilizing features of microchannel evaporators such as inlet restrictions and heat spreading further complicate fluid flow and heat transfer dynamics. Computational models, when utilized with experiments, can provide a more detailed understanding of behaviors which cannot be determined experimentally. The present study developed a computational model for flow boiling heat transfer in a 52 μm silicon microchannel evaporator designed to cool a laser diode bar, with inlet restrictions and a nonuniform heating profile at the channel level. A conjugate heat transfer model along with a coupled level set and volume of fluid (CLSVOF) model was created in ansysfluent and compared with experimental flow boiling data to gain further insights into the performance of a realistic microdevice. Heat spreading in the channel outside of the heater footprint was observed due to the high thermal conductivity of the silicon substrate. The inlet orifices impacted local flow patterns by creating a large pressure drop and forming a recirculation zone immediately downstream. This behavior resulted in pressure recovery zones and regions of separated flow boiling behavior. Bubbly, slug, and churn flows were seen to be dominant flow regimes. The heat transfer coefficient was found to be dependent on heat flux and flow regime, and more weakly on mass flux and outlet vapor quality.

Advancements and Applications of Generative Artificial Intelligence and Large Language Models on Business Management: A Comprehensive Review

This comprehensive review delves into the landscape and recent advancements of Generative Artificial Intelligence (AI) and Large Language Models (LLMs), shedding light on their transformative potential and applications across various sectors. Generative AI, exemplified by models like ChatGPT, DALL-E, and Midjourney, has rapidly evolved and is driven by breakthroughs in deep learning architectures and the availability of vast datasets. Concurrently, LLMs have revolutionized natural language processing tasks, utilizing vast text corpora to generate human-like text. The study explores recent developments, including the introduction of advanced models like GPT-4 and PaLM2 and the emergence of specialized LLMs like small LLMs (sLLMs), aimed at overcoming hardware limitations and cost constraints. Additionally, the expanding applications of generative AI, from healthcare to finance, underscore its transformative potential in addressing real-world challenges. Through a comprehensive analysis, this research contributes to the ongoing discourse on AI ethics, governance, and regulation, emphasizing the importance of responsible innovation for the benefit of humanity.

Performance and first measurements of the MAGIC stellar intensity interferometer

ABSTRACT In recent years, a new generation of optical intensity interferometers has emerged, leveraging the existing infrastructure of Imaging Atmospheric Cherenkov Telescopes (IACTs). The MAGIC telescopes host the MAGIC-SII system (Stellar Intensity Interferometer), implemented to investigate the feasibility and potential of this technique on IACTs. After the first successful measurements in 2019, the system was upgraded and now features a real-time, dead-time-free, 4-channel, GPU-based correlator. These hardware modifications allow seamless transitions between MAGIC’s standard very-high-energy gamma-ray observations and optical interferometry measurements within seconds. We establish the feasibility and potential of employing IACTs as competitive optical Intensity Interferometers with minimal hardware adjustments. The measurement of a total of 22 stellar diameters are reported, 9 corresponding to reference stars with previous comparable measurements, and 13 with no prior measurements. A prospective implementation involving telescopes from the forthcoming Cherenkov Telescope Array Observatory’s Northern hemisphere array, such as the first prototype of its Large-Sized Telescopes, LST-1, is technically viable. This integration would significantly enhance the sensitivity of the current system and broaden the UV-plane coverage. This advancement would enable the system to achieve competitive sensitivity with the current generation of long-baseline optical interferometers over blue wavelengths.

Projection mapping technologies: A review of current trends and future directions.

This study summarizes current trends and future directions in projection mapping technologies. Projection mapping seamlessly merges the virtual and real worlds through projected imagery onto physical surfaces, creating an augmented reality environment. Beyond traditional applications in advertising, art, and entertainment, various fields, including medical surgery, product design, and telecommunications, have embraced projection mapping. This study categorizes recent techniques that address technical challenges in accurately replicating desired appearances on physical surfaces through projected imagery into four groups: geometric registration, radiometric compensation, defocus compensation, and shadow removal. It subsequently introduces unconventional projectors developed to resolve specific technical issues and discusses two approaches for overcoming the inherent limitations of projector hardware, such as the inability to display images floating above physical surfaces. Finally, this study concludes the discussion with possible future directions for projection mapping technologies.

Advancing Synergy of Computing and Artificial Intelligence with Innovations Challenges and Future Prospects

As the systematic revolution converges computing technologies with Artificial Intelligence (AI) applications, various industries have been transformed, ranging from healthcare to finance. This paper presents an extensive review of the progress achieved through this synergy, alongside the challenges and possible future perspectives. The methodology used in this study is a literature review supported by a theoretical model illustrating the integration of AI with computing. The findings reveal that computational capabilities, data processing speeds, and AI model efficiencies have drastically improved. However, significant roadblocks persist, particularly in data privacy, ethical considerations, and hardware limitations. These challenges underscore the need for addressing the field’s boundless potential for breakthroughs. Existing hurdles, such as the lack of robust regulations, call for timely intervention. As we move forward, this paper emphasizes the necessity for ongoing innovation and the development of comprehensive policies to navigate the evolving landscape. Ultimately, our conclusion highlights the critical importance of preparing for these challenges to fully realize the transformative potential of the AI and computing synergy.

SCFI: Efficient forward fine-grained control flow integrity based on coarse-grained ISA extensions

Code reuse attack (CRA) is a severe threat to computer systems. To mitigate this threat, control-flow integrity (CFI) was proposed to restrict control-flow transfers. Processor vendors developed Instruction Set Architecture (ISA) extensions to enforce forward CFI on processors, such as Indirect Branch Tracking (IBT) developed by Intel and Branch Target Identification (BTI) developed by ARM. However, these extensions are coarse-grained and unable to provide enough security protection.We combine existing coarse-grained ISA extension with novel software modifications and propose a new fine-grained forward CFI implementation called Slot-based CFI (SCFI). SCFI consists of two software modifications: Slot Policy and ID Policy. Slot policy utilises lower bits of code address as labels to implement fine-grained CFI. ID Policy provides multi-label support to solve the destination equivalence problem. SCFI strengthens security by improving coarse-grained ISA extension to fine-grained CFI (improving Average Indirect Targets Allowed (AIA) from 752 to 6.24). SCFI costs negligible performance overhead (on average 0.78% in SPEC2006). It also has good performance in complex programs (1.12% in 471.omnetpp and 1.79% in 483.xalancbmk). In addition, SCFI only needs lightweight software instrument without hardware modification and supports complex control-flow graphs.

Inclined Obstacle Recognition and Ranging Method in Farmland Based on Improved YOLOv8

Unmanned tractors under ploughing conditions suffer from body tilting, violent shaking and limited hardware resources, which can reduce the detection accuracy of unmanned tractors for field obstacles. We optimize the YOLOv8 model in three aspects: improving the accuracy of detecting tilted obstacles, computational reduction, and adding a visual ranging mechanism. By introducing Funnel ReLU, a self-constructed inclined obstacle dataset, and embedding an SE attention mechanism, these three methods improve detection accuracy. By using MobileNetv2 and Bi FPN, computational reduction, and adding camera ranging instead of LIDAR ranging, the hardware cost is reduced. After completing the model improvement, comparative tests and real-vehicle validation are carried out, and the validation results show that the average detection accuracy of the improved model reaches 98.84% of the mAP value, which is 2.34% higher than that of the original model. The computation amount of the same image is reduced from 2.35 billion floating-point computations to 1.28 billion, which is 45.53% less than the model computation amount. The monitoring frame rate during the movement of the test vehicle reaches 67 FPS, and the model meets the performance requirements of unmanned tractors under normal operating conditions.

One-vs-One, One-vs-Rest, and a novel Outcome-Driven One-vs-One binary classifiers enabled by optoelectronic memristors towards overcoming hardware limitations in multiclass classification

Deep neural networks have achieved considerable success over the past ten years in a variety of fields. However, current state-of-the-art artificial intelligence (AI) systems require large computing hardware infrastructure and high power consumption. To overcome these hurdles, it is required to adopt new strategies such as designing novel computation architectures and developing building blocks that can mimic the low energy consumption of biological systems. On the architecture level, implementing classification tasks by splitting the problem into simpler subtasks is a way to relax hardware constraints despite the less accuracy of the approach. On the computation unit level, memristive devices are a promising technology for low power neuromorphic computation. Hereby, we combine both these two approaches and present a novel algorithmic approach for multiclass classification tasks through splitting the problem into binary subtasks while using optoelectronics memristors as synapses. Our approach leverages the core fundamentals from the One-vs-One (OvO) and the One-vs-Rest (OvR) classification strategies towards a novel Outcome-Driven One-vs-One (ODOvO) approach. The light modulation of synaptic weights, fed in our algorithm from experimental data, is a key enabling parameter that permits classification without modifying further applied electrical biases. Our approach requires at least a 10X less synapses (only 196 synapses are required) while reduces the classification time by up to N2\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\frac{N}{2}$$\\end{document} compared to conventional memristors. We show that the novel ODOvO algorithm has similar accuracies to OvO (reaching over 60% on the MNIST dataset) while requiring even fewer iterations compared to the OvR. Consequently, our approach constitutes a feasible solution for neural networks where key priorities are the minimum energy consumption i.e., small iterations number, fast execution, and the low hardware requirements allowing experimental verification.

Autonomous navigation system for hexa-legged search and rescue robot using LiDAR

This study proposes an autonomous navigation system for hexapod robots, promising in complex rescue scenarios. The system is tested in simulations under three environments: rocky, cracked flooring, and inclined surfaces. utilizing light detection and ranging (LiDAR) and simultaneous localization and mapping (SLAM), the robot recognizes positions and constructs environmental maps. Implemented via robot operating system, the research successfully applies navigation and mapping using hector_slam. LiDAR mapping yields satisfactory accuracy, with average errors of 0.21% for general mapping and 5.34% for circular paths. Within a 2-meter range, navigation achieves good accuracy, averaging 1.2% error on the x-axis and 0.011% on the y-axis during linear motion. Navigational repeatability improves, with reliable results showing an average error of 4.33 cm on the x-axis and 0.5 cm on the y-axis when returning to starting points. Arena testing with varied obstacles demonstrates successful obstacle traversal. However, in the second test, limitations in hardware, notably the Raspberry Pi 4 CPU usage reaching 97% during navigation, hindered reaching the third target.

An end-cloud collaboration for state-of-charge estimation of lithium-ion batteries based on extended Kalman filter and convolutional neural network (CNN)—long short-term memory (LSTM)—attention mechanism (AM)

This paper introduces an innovative online state of charge (SOC) estimation method for lithium-ion batteries, designed to address the challenges of accurate and timely SOC estimation in electric vehicles under complex working conditions and computational limitations of on-board hardware. Central to this method is the concept of end-cloud collaboration, which harmonizes accuracy with real-time performance. The framework involves deploying a data-driven model on the cloud side for high-accuracy estimation, complemented by a fast model on the end side for real-time estimation. A crucial component of this system is the implementation of the extended Kalman filter on the end side, which fuses results from both ends to achieve high-accuracy and real-time online estimation. This method has been rigorously evaluated under various dynamic driving conditions and temperatures, demonstrating high accuracy, real-time performance, and robustness. The estimation results yield a root mean square error and mean absolute error of approximately 1.5% and 1%, respectively. Significantly, under the Cyber Hierarchy and Interactional Network framework, this method shows promising potential for extension to multi-state online cooperative estimation, opening avenues for advanced battery system management.

Microbial Community Profiling Protocol with Full-length 16S rRNA Sequences and Emu.

16S rRNA targeted amplicon sequencing is an established standard for elucidating microbial community composition. While high-throughput short-read sequencing can elicit only a portion of the 16S rRNA gene due to their limited read length, third generation sequencing can read the 16S rRNA gene in its entirety and thus provide more precise taxonomic classification. Here, we present a protocol for generating full-length 16S rRNA sequences with Oxford Nanopore Technologies (ONT) and a microbial community profile with Emu. We select Emu for analyzing ONT sequences as it leverages information from the entire community to overcome errors due to incomplete reference databases and hardware limitations to ultimately obtain species-level resolution. This pipeline provides a low-cost solution for characterizing microbiome composition by exploiting real-time, long-read ONT sequencing and tailored software for accurate characterization of microbial communities. © 2024 Wiley Periodicals LLC. Basic Protocol: Microbial community profiling with Emu Support Protocol 1: Full-length 16S rRNA microbial sequences with Oxford Nanopore Technologies sequencing platform Support Protocol 2: Building a custom reference database for Emu.

Quantum-based privacy-preserving techniques for secure and trustworthy internet of medical things an extensive analysis

The advent of Quantum Key Distribution (QKD) and quantum-based privacy-preserving techniques has ushered in a new era for securing communication channels within the Internet of Medical Things (IoMT) systems. We examine the fundamentals, uses, and ramifications of quantum cryptography in relation to healthcare data security in this thorough investigation. The journey commences with an in-depth overview of quantum cryptography, unraveling the concepts of superposition, entanglement, and quantum gates that form the bedrock of quantum computing. As we proceed, we investigate the uses of quantum cryptography, highlighting its contribution to resolving the particular issues brought about by the data-intensive and networked character of IoMT systems. The relevance of QKD in ensuring secure communication within IoMT is meticulously dissected, with case studies and experiments demonstrating the practicality and effectiveness of quantum-based privacy-preserving techniques. From telemedicine networks to wearable health devices, each case study offers valuable insights into the application of quantum-safe cryptography, showcasing its ability to fortify data integrity and confidentiality. A meticulous survey of existing research, coupled with an analysis of advancements in quantum cryptography, provides a panoramic view of the current landscape. From hardware limitations to distance constraints, the study navigates through challenges and breakthroughs, offering a roadmap for the future integration of quantum technologies into IoMT systems.Comparing quantum-based approaches with classical cryptography unveils a nuanced landscape of computational complexity, key distribution methodologies, and real-time encryption considerations. This comparative analysis serves as a guide for healthcare practitioners and technologists in making informed decisions regarding the adoption of quantum-based privacy-preserving techniques in IoMT environments. The case studies and experiments collectively paint a picture of the practicality and promise of quantum-based privacy-preserving techniques in IoMT scenarios. Anticipating future advancements, the exploration extends into quantum hardware improvements, standardization efforts, and the integration of quantum technologies with emerging trends like edge computing and blockchain. As the healthcare industry stands on the cusp of a quantum revolution, the comprehensive insights provided herein offer a foundation for understanding, implementing, and shaping the future of secure healthcare communication through the lens of quantum cryptography.

Analyzing computer vision models for detecting customers: a practical experience in a mexican retail

Computer vision has become an important technology for obtaining meaningful data from visual content and providing valuable information for enhancing security controls, marketing, and logistic strategies in diverse industrial and business sectors. The retail sector constitutes an important part of the worldwide economy. Analyzing customer data and shopping behaviors has become essential to deliver the right products to customers, maximize profits, and increase competitiveness. In-person shopping is still a predominant form of retail despite the appearance of online retail outlets. As such, in-person retail is adopting computer vision models to monitor store products and customers. This research paper presents the development of a computer vision solution by Lytica Company to detect customers in Steren’s physical retail stores in Mexico. Current computer vision models such as SSD Mobilenet V2, YOLO-FastestV2, YOLOv5, and YOLOXn were analyzed to find the most accurate system according to the conditions and characteristics of the available devices. Some of the challenges addressed during the analysis of videos were obstruction and proximity of the customers, lighting conditions, position and distance of the camera concerning the customer when entering the store, image quality, and scalability of the process. Models were evaluated with the F1-score metric: 0.64 with YOLO FastestV2, 0.74 with SSD Mobilenetv2, 0.86 with YOLOv5n, 0.86 with YOLOv5xs, and 0.74 with YOLOXn. Although YOLOv5 achieved the best performance, YOLOXn presented the best balance between performance and FPS (frames per second) rate, considering the limited hardware and computing power conditions.

Comparison of Bagging and Sparcity Methods for Connectivity Reduction in Spiking Neural Networks with Memristive Plasticity

Developing a spiking neural network architecture that could prospectively be trained on energy-efficient neuromorphic hardware to solve various data analysis tasks requires satisfying the limitations of prospective analog or digital hardware, i.e., local learning and limited numbers of connections, respectively. In this work, we compare two methods of connectivity reduction that are applicable to spiking networks with local plasticity; instead of a large fully-connected network (which is used as the baseline for comparison), we employ either an ensemble of independent small networks or a network with probabilistic sparse connectivity. We evaluate both of these methods with a three-layer spiking neural network, which are applied to handwritten and spoken digit classification tasks using two memristive plasticity models and the classical spike time-dependent plasticity (STDP) rule. Both methods achieve an F1-score of 0.93–0.95 on the handwritten digits recognition task and 0.85–0.93 on the spoken digits recognition task. Applying a combination of both methods made it possible to obtain highly accurate models while reducing the number of connections by more than three times compared to the basic model.