Low-cost Hardware Research Articles

Klessydra-T: Designing Vector Coprocessors for Multithreaded Edge-Computing Cores

Computation intensive kernels, such as convolutions, matrix multiplication and Fourier transform, are fundamental to edge-computing AI, signal processing and cryptographic applications. Interleaved-Multi-Threading (IMT) processor cores are interesting to pursue energy efficiency and low hardware cost for edge-computing, yet they need hardware acceleration schemes to run heavy computational workloads. Following a vector approach to accelerate computations, this study explores possible alternatives to implement vector coprocessing units in RISC-V cores, showing the synergy between IMT and data-level parallelism in the target workloads.

Software Phase Correction Technique for Passive Radar

A traditional radar actively transmits pulses and receive the corresponding echoes. By computing the time taken to receive the echoes the system is able to detect targets and estimate their ranges. However, mainly for military applications, an active radar has a major drawback: it can be detected by the enemy since it radiates electromagnetic pulses. As such, there is currently high research and development activity in the field of passive radar systems. In this alternative scenario, the radar does transmit any signal. Instead, it uses signals already present in the environment, such as TV, radio broadcasts and satellite digital video broadcast as transmitters of opportunity. By measuring the time difference between the signal received by the transmitter of opportunity and the signal echoed by the targets, the radar can detect targets and estimate their ranges. In scenarios which use very low power illuminators of opportunity, such as satellite DVB-S or GPS, relatively long integration times are required in order to obtain reasonable values of signal to noise ratio. This implies that, besides frequency correction, the coherency of local oscillators in the receivers need to be maintained during long time intervals, which can be very difficult. The paper presents a novel phase correction technique for passive radar which uses targets of opportunity, already present in the target area, as references to maintain the coherency of the oscillators for all the integration interval. The proposed methodology is quite simple and enables the use of low-cost hardware with independent oscillators for the reference and surveillance channels which can be geographically distributed. The obtained results illustrate the effectiveness and applicability of the method.

PEPSI++: Fast and Lightweight Network for Image Inpainting.

Among the various generative adversarial network (GAN)-based image inpainting methods, a coarse-to-fine network with a contextual attention module (CAM) has shown remarkable performance. However, due to two stacked generative networks, the coarse-to-fine network needs numerous computational resources, such as convolution operations and network parameters, which result in low speed. To address this problem, we propose a novel network architecture called parallel extended-decoder path for semantic inpainting (PEPSI) network, which aims at reducing the hardware costs and improving the inpainting performance. PEPSI consists of a single shared encoding network and parallel decoding networks called coarse and inpainting paths. The coarse path produces a preliminary inpainting result to train the encoding network for the prediction of features for the CAM. Simultaneously, the inpainting path generates higher inpainting quality using the refined features reconstructed via the CAM. In addition, we propose Diet-PEPSI that significantly reduces the network parameters while maintaining the performance. In Diet-PEPSI, to capture the global contextual information with low hardware costs, we propose novel rate-adaptive dilated convolutional layers that employ the common weights but produce dynamic features depending on the given dilation rates. Extensive experiments comparing the performance with state-of-the-art image inpainting methods demonstrate that both PEPSI and Diet-PEPSI improve the qualitative scores, i.e., the peak signal-to-noise ratio (PSNR) and structural similarity (SSIM), as well as significantly reduce hardware costs, such as computational time and the number of network parameters.

A Finite Precision Implementation of an Image Encryption Scheme Based on DNA Encoding and Binarized Chaotic Cores

Under finite precision implementation, one dimensional (1D) chaotic maps suffer from limited number of control parameters and converged periodicity, making them unsuitable for hardware based ciphering systems despite their simple implementation and low hardware cost. This paper first discusses the limited periodicity of 1D maps under fixed point precision representation, then, presents an image encryption algorithm based on DNA encoding and two specially configured binarized chaotic cores. The function of both cores is to perform the confusion and diffusion stages of the image by generating pseudo random numbers with excellent cryptographic properties. DNA encoding adds an extra layer of security to the algorithm by converting both the image and the chaotic stream to DNA sequences using specific DNA encoding rule. Initial values of both chaotic cores are image dependent based on a calculated hamming distance. These initial condition and the utilized DNA rules composes the overall secret key of the system with a total length of 336 bits. On the condition that all calculations involved in the scheme are based on binary integer arithmetic, all performed security analysis subjected to the scheme proved that the system could withstand known attacks with excellent encryption properties.

Potentials of Bluetooth Low Energy Beacons for order tracing in single and small batch production

The quality of production planning and control in production companies largely depends on the accuracy of the database regarding the status quo of the production process, as well as the time required for each production step. Especially in single and small batch production, it is difficult to create a reliable database due to the large product spectrum and the variation in the production process. For this reason, this paper examines the potential of Bluetooth Low Energy (BLE) Beacons to create a low-cost real-time locating system (RTLS). In comparison to other existing RTLS solutions like RFID, BLE stands out through very low hardware costs and a high compatibility to almost every mobile device. Through a signal strength based approach in combination with probabilistic neural networks, the RTLS has been able to achieve a positioning accuracy with a mean absolute error of up to 0,74 m in an industry-like environment.

Construction of Golay Complementary Matrices and Its Applications to MIMO Omnidirectional Transmission

Golay complementary sequences (GCS) have found wide applications in radar and communications thanks to their prominent property of complementary autocorrelation. This paper studies the extension of Golay complementary sequences to Golay complementary matrices (GCM), i.e, a pair of matrices whose two-dimensional autocorrelations add to be a two-dimensional <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\delta$</tex-math></inline-formula> -function. While the construction of (real) binary GCM are available in the literature, how to construct the quaternary GCM (and the more general unimodular GCM) remains unknown. We put together this missing piece, and show the feasible size that allows for binary and quaternary GCM. The feasible sizes can be further expanded if the phases of the matrices’ entries can be arbitrary. In addition, we consider the extension of the GCM to the set of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$N$</tex-math></inline-formula> ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$N\geq 2$</tex-math></inline-formula> ) matrices, whose two-dimensional autocorrelations add to be a two-dimensional <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\delta$</tex-math></inline-formula> -function, which we refer to as autocorrelation complementary matrices (ACM). For <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$N=4$</tex-math></inline-formula> , the ACM allows for matrices of arbitrary size. The GCM/ACM can be applied to omnidirectional transmission for a multi-input multi-output (MIMO) communication system with a uniform rectangular array (URA), which is useful for broadcasting cell-wide common messages. We propose two schemes for omnidirectional transmission, one is based on space-time block code (STBC) and the other is based on the dual cross-polarization of the antenna array. Both schemes can be implemented in the radio frequency (RF) analog domain, generating (theoretically) perfect omnidirectional beampattern with low hardware cost.

A Deep Learning-Based Low Overhead Beam Selection in mmWave Communications

Due to large amounts of available spectrum at high frequencies, millimeter-wave (mmWave) technology has gained extensive research attention in 5G communications, whereas mmWave links suffer from severe free space attenuation. Codebook-based beamforming techniques with multiple antennas can effectively alleviate this challenge with low computational complexity and low hardware cost. However, small delay and high-speed communications with beamforming techniques require beam alignment with small overhead so as to establish the wireless link quickly. In this context, this paper proposes a deep learning-based low overhead analog beam selection scheme by virtue of the super-resolution technology. To be concrete, deep neural networks are employed to conduct beam quality estimation based on partial beam measurements. Our proposed scheme can cover all the directions of arriving signals with low overhead by utilizing codebooks with different beam widths. Furthermore, for the purpose of further reducing the overhead, we formulate the beam quality prediction model based on the past beam sweepings. With these beam quality estimation and prediction model, the beam that achieves large signal-to-noise-power-ratio (SNR) can be selected based on partial beam measurements. Simulation results show that the proposed scheme can accurately estimate beam qualities and give high probability of optimal beam selections with low overhead.

An Experimental Study on the Use of LoRa Technology in Vehicle Communication

Among the profound changes that will come with the Internet of Things is vehicular communication. Shortly the vehicles will be connected between themselves (V2V) and their infrastructure. For the full achievement of these objectives, the challenges of new technologies are enormous due to the requirement of high reliability, high speed, and low latency. None of the technologies under development for this application has reached a satisfactory stage be assumed to as definitive. The LoRa technology, operating at frequencies below 1 GHz, presents a good signal spread and penetration in obstacles. It has a considerable range, open-source, simple, robust, and low-cost hardware, vast configuration possibilities, and applications, ranging from medicine to agriculture, do not use licensed bands. The tests proved to have good reach even in a dense urban environment. They can become a viable alternative in applications that require short message transmissions with few characters that do not require the constant sending of information packages. The purpose of this work is to evaluate the communication between V2I, V2V, and stationary vehicles using LoRa technology in field tests with measurements of signal strength, reception ratio, and signal-to-noise ratio. It will be using different SF (scattering factors) inherent to LoRa (SF7 and SF12) and evaluate the influence of the Doppler effect on communication.

Corrections to “Vision-Based Moving UAV Tracking by Another UAV on Low-Cost Hardware and a New Ground Control Station”

In the above article [1] , our aim was to automatically detect and track target moving Unmanned Aerial Vehicle (UAV) [4] by another UAV in real-time applications. For that purpose, we developed a vision-based low-cost hardware system and analyzed the tracking algorithm performances. In addition, we developed a ground control station that is integrated with the low-cost hardware system. We initially created a new public data set including different types of UAV images mostly recorded from experiments and search on Google images. The data set is available on the website linked as http://cogvi.atauni.edu.tr/ResearchLab/PageDetail/Our-ATAUAVs-Dataset-86 that is accessible to everyone.

Memory Utilization and Machine Learning Techniques for Compiler Optimization

Compiler optimization techniques allow developers to achieve peak performance with low-cost hardware and are of prime importance in the field of efficient computing strategies. The realm of compiler suites that possess and apply efficient optimization methods provide a wide array of beneficial attributes that help programs execute efficiently with low execution time and minimal memory utilization. Different compilers provide a certain degree of optimization possibilities and applying the appropriate optimization strategies to complex programs can have a significant impact on the overall performance of the system. This paper discusses methods of compiler optimization and covers significant advances in compiler optimization techniques that have been established over the years. This article aims to provide an overall survey of the cache optimization methods, multi memory allocation features and explore the scope of machine learning in compiler optimization to attain a sustainable computing experience for the developer and user.

Light-Weight CNN Enabled Edge-Based Framework for Machine Health Diagnosis

Condition Based Monitoring (CBM) leverages sensor measurements for measuring state of health of an asset and autonomous diagnosis of faults to trigger remedial actions. Countless deep learning architectures are available for cloud-based feature engineering, feature extraction and classification of data for CBM models. However, complex models pose memory and processing speed constraints for edge implementation, while cloud-based computing poses high latency, high cost of data transmission and storage and privacy threats. This calls for a data driven machine health diagnosis system that is effective yet edge-compatible and secure. In this work, we present a model that is light-weight and non-redundant by optimizing the model size and complexity for edge implementation onto resource constrained, low-cost hardware. The model is based on a light-weight, edge implementable Convolutional Neural Network (CNN) algorithm that utilizes vibration sensor measurements for fault event estimation of machines. The model was trained and tested on two publicly available and widely studied vibration datasets for rolling element bearing faults. This CNN based classification system using spectrograms is able to achieve near perfect accuracy for both binary as well as multi-class fault classification. Lastly, the trained model was implemented and tested on a Raspberry Pi single-board computer, thus realizing an edge-compatible implementation which is cost-saving, secure and reliable.

Low-Cost Precision Monitoring System of Machine Tools for SMEs

Digital manufacturing technology promotes the application of digital information to manufacturing processes and machine tools. However, the high entry cost and complexity of existing offerings limit the wide application of digital manufacturing in small and medium-sized enterprises (SMEs). For CNC machine tools, precision parameters can be acquired either by a visual inspection by the operator or by third-party connection software or hardware. To decrease human error, cost and time in data recording, this research proposes a machine tool data acquisition process for monitoring purposes using a novel monitoring system based on low-cost hardware (Arduino and low-cost camera) and open-source computation platforms (Node-RED). Remote data processing in a tablet is rendered possible by wireless data communication. Prototypes were tested with a HU40-T Five-axis machine tool. Tests show that the proposed system is capable of acquiring precision parameters such as the pitch error and backlash error of machine tools at a system cost of less than $200 CAD.

Energy Efficiency of Slotted LoRaWAN Communication With Out-of-Band Synchronization

Although the idea of using wireless links for covering large areas is not new, the advent of LPWANs has recently started changing the game. Simple, robust, narrowband modulation schemes permit the implementation of low-cost radio devices offering high receiver sensitivity, thus improving the overall link budget. The several technologies belonging to the LPWAN family, including the well-known LoRaWAN solution, provide a cost-effective answer to many Internet-of-things (IoT) applications, requiring wireless communication capable of supporting large networks of many devices (e.g., smart metering). Generally, the adopted MAC strategy is based on pure ALOHA, which, among other things, allows to minimize the traffic overhead under constrained duty cycle limitations of the unlicensed bands. Unfortunately, ALOHA suffers from poor scalability, rapidly collapsing in dense networks. This work investigates the design of an improved LoRaWAN MAC scheme based on slotted ALOHA. In particular, the required time dissemination is provided by out-of-band communications leveraging on FM-RDS broadcasting. An experimental setup based on low-cost hardware is used to characterize the obtainable synchronization performance and derive a timing error model. Consequently, improvements in success probability and energy efficiency have been validated by means of simulations in very large networks with up to 10000 nodes. It is shown that the advantage of the proposed scheme over conventional LoRaWAN communication is up to 100% when short update time and large payload are required. Similar results are obtained regarding the energy efficiency improvement, which is close to 100% for relatively short transmission intervals and long message duration; however, due to the additional overhead for listening to the time dissemination messages, efficiency gain can be negative for short-duration messages fastly repeating.

A Physical Unclonable Function Using a Configurable Tristate Hybrid Scheme With Non-Volatile Memory

The physical unclonable function (PUF) is a promising low-cost hardware security primitive. Recent advances in nanotechnology have provided new opportunities for nanoscale PUF circuits. The resistive random access memory (RRAM) is extensively used in nanoscale circuits due to its low cost, non-volatility and easy integration with CMOS. This paper proposes a novel tristate hybrid PUF (TH-PUF) design based on a one-transistor-one-RRAM (1T1R) cell; this cell can be configured into two weak PUFs and a strong PUF using few control signals. To assess the proposed PUF design, a compact RRAM model at UMC 65 nm technology is employed. Simulation results show that the proposed TH-PUF achieves good uniqueness, reliability as well as a higher gate usability compared with an entire CMOS PUFs. The number of challenge response pairs (CRPs) of the proposed TH-PUF is larger than other RRAM-based PUFs. Moreover, the TH-PUF is more resistant to a modeling machine learning attack than traditional PUF designs.

Development and Application of Cell-Phone-Based Internet of Things (IoT) Systems for Soil Moisture Monitoring

Active soil moisture monitoring is an important consideration in irrigation water management. A permanent and readily accessible record of changes in soil moisture can be used to improve future water management decision-making. Similarly, accessing stored soil moisture data in near-real-time is also essential for making timely farming and management decisions, such as where, when, and how much irrigation to apply. Access to reliable communication systems and delivery of real-time data can be affected by its availability near production fields. Therefore, soil moisture monitoring systems with real-time data functionality that can meet the needs of farmers at an affordable cost are currently needed. The objective of the study was to develop and fieldtest affordable cell-phone-based Internet of things (IoT) systems for soil moisture monitoring. These IoT systems were designed using low-cost hardware components and open-source software to transmit soil moisture data from the Watermark 200SS or ECH2O EC-5 sensors. These monitoring systems utilized either Particle Electron or Particle Proton Arduino-compatible devices for data communication. The IoT soil moisture monitoring systems have been deployed and operated successfully over the last three years in South Carolina.

High-Accuracy Multiply-Accumulate (MAC) Technique for Unary Stochastic Computing

Multiply-accumulate (MAC) operations are common in data processing and machine learning but costly in terms of hardware usage. Stochastic Computing (SC) is a promising approach for low-cost hardware design of complex arithmetic operations such as multiplication. Computing with deterministic unary bit-streams (defined as bit-streams with all 1s grouped together at the beginning or end of a bit-stream) has been recently suggested to improve the accuracy of SC. Conventionally, SC designs use multiplexer (MUX) units or OR gates to accumulate data in the stochastic domain. MUX-based addition suffers from scaling of data and OR-based addition from inaccuracy. This work proposes a novel technique for MAC operation on unary bit-streams that allows exact, non-scaled addition of multiplication results. By introducing a relative delay between the products, we control correlation between bit-streams and eliminate OR-based addition error. We evaluate the accuracy of the proposed technique compared to the state-of-the-art MAC designs. After quantization, the proposed technique demonstrates at least 37% and up to 100% decrease of the mean absolute error for uniformly distributed random input values, compared to traditional OR-based MAC designs. Further, we demonstrate that the proposed technique is practical and evaluate area, power and energy of three possible implementations.

Supporting the Digital Transformation: A Low-Threshold Approach for Manufacturing Related Higher Education and Employee Training

The technologies of digital transformation, such as the Internet-of-Things (IoT), artificial intelligence or predictive maintenance enable significant efficiency gains in industry and are becoming increasingly important as a competitive factor. However, their successful implementation and creative, future application requires the broad acceptance and knowledge of non-IT-related groups, such as production management students, engineers or skilled workers, which is still lacking today. This paper presents a low-threshold training concept bringing IoT-technologies and applications into manufacturing related higher education and employee training. The concept addresses the relevant topics starting from IoT-basics to predictive maintenance using mobile low-cost hardware and infrastructure.

Digital Vein Mapping Using Augmented Reality

Vein detection is an important issue for the medical field. There are some commercial devices for detecting veins using infrared radiation. However, most of these commercial solutions are cost-prohibitive. Recently, veins detection has attracted much attention from research teams. The main focus is on developing real-time systems with low-cost hardware. Systems developed to reduce costs suffer from low frame rates. This, in turn, makes these systems not suitable for real-world applications. On the other hand, systems that use powerful processors to produce high frame rates suffer from high costs and a lack of mobility. In this paper, a real-time vein mapping prototype using augmented reality is proposed. The proposed prototype provides a compromised solution to produce high frame rates with a low-cost system. It consists of a USB camera attached to an Android smartphone used for real-time detection. Infrared radiation is employed to differentiate the veins using 20 Infrared Light Emitting Diodes (LEDs). The captured frames are processed to enhance vein detection using light computational algorithms to improve real-time processing and increase frame rate. Finally, the enhanced view of veins appears on the smartphone screen. Portability and economic cost are taken into consideration while developing the proposed prototype. The proposed prototype is tested with people of different ages and gender, as well as using mobile devices of different specifications. The results show a high vein detection rate and a high frame rate compared to other existing systems.

Low-Cost and High-Throughput Hardware Design for the HEVC 16x16 2-D DCT Transform

This article presents the hardware design of the 16x16 2-D DCT used in the new video coding standard, the HEVC – High Efficiency Video Coding. The transforms stage is one of the innovations proposed by HEVC, since a variable size transforms stage is available (from 4x4 to 32x32), allowing the use of transforms with larger dimensions than used in previous standards. The presented design explores the 2-D DCT separability property, using two instances of the one-dimension DCT. The architecture focuses on low hardware cost and high throughput, thus the HEVC 16-points DCT algorithm was simplified targeting a more efficient hardware implementation. Operations and hardware minimization strategies were used in order to achieve such simplifications: operation reordering, factoring, multiplications to shift-adds conversion, and sharing of common sub-expressions. The 1-D DCT architectures were designed in a fully combinational way in order to reduce control overhead. A transposition buffer is used to connect the two 1-D DCT architectures. The synthesis was directed to Stratix III FPGA and TSMC 65nm standard cells technologies. The complete 2-D DCT architecture is able to achieve real-time processing for high and ultra-high definition videos, such as Full HD, QFHD and UHD 8K. When compared with related works, the architectures designed in this work reached the highest throughput and the lowest hardware resources consumption.

Classification of forearm EMG signals for 10 motions using optimum feature-channel combinations

Electromyography (EMG) is the study of electrical activity in the muscles. We classify EMG signals from surface electrodes (channels) using Artificial Neural Network (ANN). We evaluate classification performance of 10 different hand motions using several feature-channel combinations with wrapper method. Highest classification accuracy of 98.7% is achieved with each feature-channel combination. Compared to previous studies, we achieve the highest accuracy for 10 classes with lower number of feature-channel combination. We reduce ANN complexity without compromising the classification accuracy for deployment in low-end hardware with limited computational power along with improving the design of a low-cost hardware for EMG signal acquisition.