Hardware Approach Research Articles

Non-invasive transcranial volumetric ultrasound localization microscopy of the rat brain with continuous, high volume-rate acquisition

Rationale: Structure and function of the microvasculature provides critical information about disease state, can be used to identify local regions of pathology, and has been shown to be an indicator of response to therapy. Improved methods of assessing the microvasculature with non-invasive imaging modalities such as ultrasound will have an impact in biomedical theranostics. Ultrasound localization microscopy (ULM) is a new technology which allows processing of ultrasound data for visualization of microvasculature at a resolution better than allowed by acoustic diffraction with traditional ultrasound systems. Previous application of this modality in brain imaging has required the use of invasive procedures, such as a craniotomy, skull-thinning, or scalp removal, all of which are not feasible for the purpose of longitudinal studies. Methods: The impact of ultrasound localization microscopy is expanded using a 1024 channel matrix array ultrasonic transducer, four synchronized programmable ultrasound systems with customized high-performance hardware and software, and high-performance GPUs for processing. The potential of the imaging hardware and processing approaches are demonstrated in-vivo. Results: Our unique implementation allows asynchronous acquisition and data transfer for uninterrupted data collection at an ultra-high fixed frame rate. Using these methods, the vasculature was imaged using 100,000 volumes continuously at a volume acquisition rate of 500 volumes per second. With ULM, we achieved a resolution of 31 µm, which is a resolution improvement on conventional ultrasound imaging by nearly a factor of ten, in 3-D. This was accomplished while imaging through the intact skull with no scalp removal, which demonstrates the utility of this method for longitudinal studies. Conclusions: The results demonstrate new capabilities to rapidly image and analyze complex vascular networks in 3-D volume space for structural and functional imaging in disease assessment, targeted therapeutic delivery, monitoring response to therapy, and other theranostic applications.

A computing hardware approach aspires to emulate the brain

Neuromorphic computing promises energy savings, a deeper understanding of the human brain, and smarter sensors.

Hardware approach to brain tumour detection using AI concepts using real time embedded systems with Raspberry pi

Tumors have become the second leading cause of cancer nowadays, posing significant risks to numerous patients. The medical field is in urgent need of rapid, automated, efficient, and reliable techniques to detect tumors, especially brain tumors. Early and accurate detection plays a crucial role in successful treatment and keeping patients safe. To address this challenge, various image processing techniques are employed in the medical domain. These advancements have allowed doctors to administer appropriate treatments, leading to the successful recovery of many tumor patients. Tumors are characterized by the abnormal growth of cells, which proliferate uncontrollably. Brain tumors, in particular, can be devastating as they compete with healthy cells and tissues for essential nutrients, ultimately resulting in brain dysfunction. Traditionally, doctors have relied on manual examination of MR images to identify the location and extent of brain tumors. Unfortunately, this approach is prone to errors and can be extremely time-consuming. To overcome these limitations, we have implemented a cutting-edge deep learning architecture known as Convolution Neural Network (CNN), a type of Neural Network (NN) that utilizes Transfer Learning. This CNN-based model enables us to automatically detect the presence of brain tumors in medical images with high accuracy. If a tumor is detected, the model outputs a positive result; otherwise, it indicates the absence of a tumor. In our approach, we use K-means clustering in conjunction with Raspberry Pi to precisely pinpoint the location of the brain tumor. This ensures targeted and efficient treatment planning. Additionally, we have integrated an Arduino controller to facilitate the movement of robotic wheels, allowing for precise navigation to the exact location of the tumor during medical interventions. Overall, our system represents a significant advancement in tumor detection and localization, offering faster and more reliable results compared to manual methods. By leveraging state-of-the-art technology and innovative techniques, we strive to enhance patient outcomes, ultimately saving more lives in the fight against brain tumors.

Energy Consumption Monitoring with Evaluation of Hidden Energy Losses

This article presents a computational method for monitoring the energy consumption of technological systems with the assessment of their hidden energy losses caused by erroneous actions of personnel or equipment failures. Herewith, energy losses are calculated as the difference between the actual energy consumed and the minimum energy required to conduct the process in all operating modes. The minimum required energy is determined by the machine learning method based on stationary consumption precedents. Two approaches to the implementation of energy consumption monitoring with the assessment of hidden energy losses are considered – hardware and software. The hardware approach is based on the preliminary definition of normative, or minimum specific energy consumption in each technological mode. The software approach is based on the modeling of stationary areas of energy consumption in the form of precedents and their further analysis in the space of influential technological parameters. The paper notes the advantages and disadvantages of the proposed monitoring method, it is emphasized that the method is able to work with both linear and non-linear functions of energy dependence on the parameters of the technological process. It is noted in the paper that the advantage of the proposed method is the automated construction of the minimum energy function.

Physical deep learning with biologically inspired training method: gradient-free approach for physical hardware

Ever-growing demand for artificial intelligence has motivated research on unconventional computation based on physical devices. While such computation devices mimic brain-inspired analog information processing, the learning procedures still rely on methods optimized for digital processing such as backpropagation, which is not suitable for physical implementation. Here, we present physical deep learning by extending a biologically inspired training algorithm called direct feedback alignment. Unlike the original algorithm, the proposed method is based on random projection with alternative nonlinear activation. Thus, we can train a physical neural network without knowledge about the physical system and its gradient. In addition, we can emulate the computation for this training on scalable physical hardware. We demonstrate the proof-of-concept using an optoelectronic recurrent neural network called deep reservoir computer. We confirmed the potential for accelerated computation with competitive performance on benchmarks. Our results provide practical solutions for the training and acceleration of neuromorphic computation.

Near Infrared Illumination Optimization for Vein Detection: Hardware and Software Approaches

Venepuncture is one of the most crucial processes in many medical procedures. However, finding a real-time and vibrant visualization of the vein structures faces many difficulties. Several devices were introduced to solve this problem, yet, these devices shared common drawbacks, primarily when visualizing deep veins or veins in a thicker tissue of the human body. This study proposes a novel method for visualizing vein structures using a near-infrared (NIR) imaging technique enhanced with Hessian ridge detection. Several factors, including the wavelength of NIR light, square LED and ring LED arrangement and the effect of the diffuser and number of LEDs, were evaluated in the study. This study improves the overall quality of the acquired vein images and highlights the vein-morphological structure through image processing techniques. The study’s main aim is to achieve the highest number of visible veins. Based on the optical window, the maximum absorption range in the NIR spectrum was found from 700 to 950 nm. The NIR light absorption of human deoxygenated blood in the vein was highest at 850 nm peak of wavelength. The image processing further enhances the vein image by highlighting the extracted vein. The study also suggests that the square LED arrangements of NIR illumination are much more robust than the ring LED arrangement in ensuring excellent light penetration. The light diffuser further adds promising effects to the NIR illumination process. In terms of the square LED arrangement, increasing the square LED for enlarging the illumination area did not show any degradation effects in the visualization process. Overall, this paper presents an integrated hardware and software solution for the NIR image acquisition of a vein visualization system to cope with the image visualization of the vein for a thicker part of the human tissue, particularly on the arm and palm area.

Mitigating Power and Memory Constraints on a Venusian Seismometer

Abstract The nearest term pathway to the deployment of a seismometer on Venus is an instrument that can operate under ambient surface conditions on battery power. We conduct a series of studies on combined hardware and software approaches to maximize the quality of data returned under the likely restrictions of minimal on-board data storage and only being able to transmit in real time during a small fraction of a multimonth deployment. We assess likely Venus seismicity by examining different terrestrial analog settings; we find that likely Venus analog settings all fall within about an order of magnitude of mean Earth in terms of seismicity level. We use the seismic record from a station in central Alaska as a Venus surrogate for algorithm development. We tested various transmission triggers and developed a simple low-memory algorithm that mimics the common terrestrial long-term average/short-term average trigger. If the seismometer can operate in coordination with an orbiter that can remotely turn off data transmission, then the frequency content of a few seconds of data can be used to distinguish small, nearby earthquakes from large, distal ones, and total data transmission can be tuned to favor the latter. If an orbiter can also turn on transmission for other nearby seismometers, it would further enhance the ability to distinguish small- and large-magnitude earthquakes autonomously and increase the chances of capturing the initial onset of significant events.

Open-Source Hardware May Address the Shortage in Medical Devices for Patients with Low-Income and Chronic Respiratory Diseases in Low-Resource Countries.

Respiratory diseases pose an increasing socio-economic burden worldwide given their high prevalence and their elevated morbidity and mortality. Medical devices play an important role in managing acute and chronic respiratory failure, including diagnosis, monitoring, and providing artificial ventilation. Current commercially available respiratory devices are very effective but, given their cost, are unaffordable for most patients in low- and middle-income countries (LMICs). Herein, we focus on a relatively new design option—the open-source hardware approach—that, if implemented, will contribute to providing low-cost respiratory medical devices for many patients in LMICs, particularly those without full medical insurance coverage. Open source reflects a set of approaches to conceive and distribute the comprehensive technical information required for building devices. The open-source approach enables free and unrestricted use of the know-how to replicate and manufacture the device or modify its design for improvements or adaptation to different clinical settings or personalized treatments. We describe recent examples of open-source devices for diagnosis/monitoring (measuring inspiratory/expiratory pressures or flow and volume in mechanical ventilators) and for therapy (non-invasive ventilators for adults and continuous positive airway pressure support for infants) that enable building simple, low-cost (hence, affordable), and high-performance solutions for patients in LMICs. Finally, we argue that the common practice of approving clinical trials by the local hospital ethics board can be expanded to ensure patient safety by reviewing, inspecting, and approving open hardware for medical application to maximize the innovation and deployment rate of medical technologies.

An FPGA Based Energy-Efficient Read Mapper With Parallel Filtering and In-Situ Verification.

In the assembly pipeline of Whole Genome Sequencing (WGS), read mapping is a widely used method to re-assemble the genome. It employs approximate string matching and dynamic programming-based algorithms on a large volume of data and associated structures, making it a computationally intensive process. Currently, the state-of-the-art data centers for genome sequencing incur substantial setup and energy costs for maintaining hardware, data storage and cooling systems. To enable low-cost genomics, we propose an energy-efficient architectural methodology for read mapping using a single system-on-chip (SoC) platform. The proposed methodology is based on the q-gram lemma and designed using a novel architecture for filtering and verification. The filtering algorithm is designed using a parallel sorted q-gram lemma based method for the first time, and it is complemented by an in-situ verification routine using parallel Myers bit-vector algorithm. We have implemented our design on the Zynq Ultrascale+ XCZU9EG MPSoC platform. It is then extensively validated using real genomic data to demonstrate up to 7.8× energy reduction and up to 13.3× less resource utilization when compared with the state-of-the-art software and hardware approaches.

Computer-assisted fetal laser surgery in the treatment of twin-to-twin transfusion syndrome: Recent trends and prospects.

Fetal laser surgery has emerged as the preferred treatment of twin‐to‐twin transfusion syndrome (TTTS). However, the limited field of view of the fetoscope and the complexity of the procedure make the treatment challenging. Therefore, preoperative planning and intraoperative guidance solutions have been proposed to cope with these challenges. This review uncovers the literature on computer‐assisted software solutions focused on TTTS. These solutions are classified by the pre‐ or intraoperative phase of the procedure and further categorized by discussed hardware and software approaches. In addition, it evaluates the current maturity of technologies by the technology readiness level and enumerates the necessary aspects to bring these new technologies to clinical practice.

Water Warriors Living Lab: Towards an integrated “Heartware - Hardware – Software” Approach to Water Management

This paper presents the experience of the Water Warriors Living Lab, a bottom-up sustainable campus action research initiative in applying the integrated ‘Heartware - Hardware – Software’ approach for water management at the University of Malaya since the year 2013. It advances the proposition that a heartware approach is needed to enhance campus sustainability efforts, in addition to the more strategic hardware and software approaches. The paper describes how the Water Warrior’s experience in applying the heartware approach has been a strengthening factor in resolving water conservation issues at the university (specifically in lake management), and discusses important lessons learned to date. Our theoretical and empirical insights can hopefully provide renewed appreciation on how local, organic and humanistic factors can be essential ingredients for the long term ‘sustainability’ of campus sustainability efforts.Keyword: Campus sustainability, living lab, water management, heartware, bottom-up

Improving the Noise Floor and Speed of Your Detector: A Modular Hardware Approach for Under $1000

Improving the Noise Floor and Speed of Your Detector: A Modular Hardware Approach for Under $1000

Finite Element Modeling of Internally Actuated Triangular Lattice and Its Variants for Modular Active Cell Robots (MACROs)

Modular Active Cell Robots (MACROs) is an approach for modular robot hardware that leverages simple components such as actuators and compliant joints to make large deformable robotic structures. In this letter, we consider the planar triangular lattice and its sparser variants as the candidate mesh topologies for MACROs. Through Finite Element (FE) simulations, we quantify the deformations obtained for different choices of edges being actuated in the MACRO mesh. Further, we use this data to identify optimal actuation strategies for a target deformation. We demonstrate a trade-off between passive stiffness and actuation energy and thereafter, we refine the triangular lattice by removing edges to reduce the actuation energy without a significant reduction in stiffness. We also identify the refined meshes optimal for desired specifications.

In-Memory Principal Component Analysis by Crosspoint Array of Resistive Switching Memory: A new hardware approach for energy-efficient data analysis in edge computing

In-Memory Computing (IMC) is one of the most promising candidates for data-intensive computing accelerators of machine learning (ML). A key ML algorithm for dimensionality reduction and classification is principal component analysis (PCA), which heavily relies on matrix-vector multiplications (MVM) for which classic von Neumann architectures are not optimized. Here, we provide the experimental demonstration of a new IMC-based PCA algorithm based on power iteration and deflation executed in a 4-kbit array of resistive switching random-access memory (RRAM). The classification accuracy of the Wisconsin Breast Cancer data set reaches 95.43%, close to floating-point implementation. Our simulations indicate a 250× improvement in energy efficiency compared to commercial GPUs, thus supporting IMC for energy-efficient ML in modern data-intensive computing.

Past, Present, and Future of Deep Brain Stimulation: Hardware, Software, Imaging, Physiology and Novel Approaches.

Deep brain stimulation (DBS) has advanced treatment options for a variety of neurologic and neuropsychiatric conditions. As the technology for DBS continues to progress, treatment efficacy will continue to improve and disease indications will expand. Hardware advances such as longer-lasting batteries will reduce the frequency of battery replacement and segmented leads will facilitate improvements in the effectiveness of stimulation and have the potential to minimize stimulation side effects. Targeting advances such as specialized imaging sequences and “connectomics” will facilitate improved accuracy for lead positioning and trajectory planning. Software advances such as closed-loop stimulation and remote programming will enable DBS to be a more personalized and accessible technology. The future of DBS continues to be promising and holds the potential to further improve quality of life. In this review we will address the past, present and future of DBS.

A Unified Software/Hardware Scalable Architecture for Brain-Inspired Computing Based on Self-Organizing Neural Models.

The field of artificial intelligence has significantly advanced over the past decades, inspired by discoveries from the fields of biology and neuroscience. The idea of this work is inspired by the process of self-organization of cortical areas in the human brain from both afferent and lateral/internal connections. In this work, we develop a brain-inspired neural model associating Self-Organizing Maps (SOM) and Hebbian learning in the Reentrant SOM (ReSOM) model. The framework is applied to multimodal classification problems. Compared to existing methods based on unsupervised learning with post-labeling, the model enhances the state-of-the-art results. This work also demonstrates the distributed and scalable nature of the model through both simulation results and hardware execution on a dedicated FPGA-based platform named SCALP (Self-configurable 3D Cellular Adaptive Platform). SCALP boards can be interconnected in a modular way to support the structure of the neural model. Such a unified software and hardware approach enables the processing to be scaled and allows information from several modalities to be merged dynamically. The deployment on hardware boards provides performance results of parallel execution on several devices, with the communication between each board through dedicated serial links. The proposed unified architecture, composed of the ReSOM model and the SCALP hardware platform, demonstrates a significant increase in accuracy thanks to multimodal association, and a good trade-off between latency and power consumption compared to a centralized GPU implementation.

An Investigation on Spiking Neural Networks Based on the Izhikevich Neuronal Model: Spiking Processing and Hardware Approach

The main required organ of the biological system is the Central Nervous System (CNS), which can influence the other basic organs in the human body. The basic elements of this important organ are neurons, synapses, and glias (such as astrocytes, which are the highest percentage of glias in the human brain). Investigating, modeling, simulation, and hardware implementation (realization) of different parts of the CNS are important in case of achieving a comprehensive neuronal system that is capable of emulating all aspects of the real nervous system. This paper uses a basic neuron model called the Izhikevich neuronal model to achieve a high copy of the primary nervous block, which is capable of regenerating the behaviors of the human brain. The proposed approach can regenerate all aspects of the Izhikevich neuron in high similarity degree and performances. The new model is based on Look-Up Table (LUT) modeling of the mathematical neuromorphic systems, which can be realized in a high degree of correlation with the original model. The proposed procedure is considered in three cases: 100 points LUT modeling, 1000 points LUT modeling, and 10,000 points LUT modeling. Indeed, by removing the high-cost functions in the original model, the presented model can be implemented in a low-error, high-speed, and low-area resources state in comparison with the original system. To test and validate the proposed final hardware, a digital FPGA board (Xilinx Virtex-II FPGA board) is used. Digital hardware synthesis illustrates that our presented approach can follow the Izhikevich neuron in a high-speed state (more than the original model), increase efficiency, and also reduce overhead costs. Implementation results show the overall saving of 84.30% in FPGA and also the higher frequency of the proposed model of about 264 MHz, which is significantly higher than the original model, 28 MHz.

Software Hint-Driven Data Management for Hybrid Memory in Mobile Systems

Hybrid memory systems, comprised of emerging non-volatile memory (NVM) and DRAM, have been proposed to address the growing memory demand of current mobile applications. Recently emerging NVM technologies, such as phase-change memories (PCM), memristor, and 3D XPoint, have higher capacity density, minimal static power consumption and lower cost per GB. However, NVM has longer access latency and limited write endurance as opposed to DRAM. The different characteristics of distinct memory classes render a new challenge for memory system design. Ideally, pages should be placed or migrated between the two types of memories according to the data objects’ access properties. Prior system software approaches exploit the program information from OS but at the cost of high software latency incurred by related kernel processes. Hardware approaches can avoid these latencies, however, hardware’s vision is constrained to a short time window of recent memory requests, due to the limited on-chip resources. In this work, we propose OpenMem: a hardware-software cooperative approach that combines the execution time advantages of pure hardware approaches with the data object properties in a global scope. First, we built a hardware-based memory manager unit (HMMU) that can learn the short-term access patterns by online profiling, and execute data migration efficiently. Then, we built a heap memory manager for the heterogeneous memory systems that allows the programmer to directly customize each data object’s allocation to a favorable memory device within the presumed object life cycle. With the programmer’s hints guiding the data placement at allocation time, data objects with similar properties will be congregated to reduce unnecessary page migrations. We implemented the whole system on the FPGA board with embedded ARM processors. In testing under a set of benchmark applications from SPEC 2017 and PARSEC, experimental results show that OpenMem reduces 44.6% energy consumption with only a 16% performance degradation compared to the all-DRAM memory system. The amount of writes to the NVM is reduced by 14% versus the HMMU-only, extending the NVM device lifetime.

A Probabilistic Approach to Estimating Allowed SNR Values for Automotive LiDARs in "Smart Cities" under Various External Influences.

The paper proposes an approach to assessing the allowed signal-to-noise ratio (SNR) for light detection and ranging (LiDAR) of unmanned autonomous vehicles based on the predetermined probability of false alarms under various intentional and unintentional influencing factors. The focus of this study is on the relevant issue of the safe use of LiDAR data and measurement systems within the “smart city” infrastructure. The research team analyzed and systematized various external impacts on the LiDAR systems, as well as the state-of-the-art approaches to improving their security and resilience. It has been established that the current works on the analysis of external influences on the LiDARs and methods for their mitigation focus mainly on physical (hardware) approaches (proposing most often other types of modulation and optical signal frequencies), and less often software approaches, through the use of additional anomaly detection techniques and data integrity verification systems, as well as improving the efficiency of data filtering in the cloud point. In addition, the sources analyzed in this paper do not offer methodological support for the design of the LiDAR in the very early stages of their creation, taking into account a priori assessment of the allowed SNR threshold and probability of detecting a reflected pulse and the requirements to minimize the probability of “missing” an object when scanning with no a priori assessments of the detection probability characteristics of the LiDAR. The authors propose a synthetic approach as a mathematical tool for designing a resilient LiDAR system. The approach is based on the physics of infrared radiation, the Bayesian theory, and the Neyman–Pearson criterion. It features the use of a predetermined threshold for false alarms, the probability of interference in the analytics, and the characteristics of the LiDAR’s receivers. The result is the analytical solution to the problem of calculating the allowed SNR while stabilizing the level of “false alarms” in terms of background noise caused by a given type of interference. The work presents modelling results for the “false alarm” probability values depending on the selected optimality criterion. The efficiency of the proposed approach has been proven by the simulation results of the received optical power of the LiDAR’s signal based on the calculated SNR threshold and noise values.

ONCLE (One Clock Ensemble) for Galileo’s Next-Generation Robust Timing System

<h3>Abstract</h3> This paper presents the ONCLE (One Clock Ensemble) solution for the Galileo time and frequency reference system with advanced features in terms of robustness, performance, continuity, and simplicity. Each component clock is frequency-steered to the ensemble time, which itself creates an average of those steered clock outputs, while clock faults are detected and corrected in real time within an integrated system. The feasibility of algorithm and hardware approaches has been demonstrated on an elegant breadboard and verified by an extended test and validation campaign at the Engineering Model (EM) level, developed for the Galileo next-generation onboard timing system under European GNSS Evolutions Program. Based on the progress for space application and the heritage on the Galileo ground precise timing facility (PTF), we propose a robust solution for the upgrade of PTF for ground application aiming to provide a fully continuous and performance-improved timescale under automated operation. The capability is demonstrated by preliminary simulation results.