Search Results (183)

Mango, as an important economic crop in tropical and subtropical regions, suffers from chilling injuries caused by postharvest low-temperature storage, which seriously affect its quality and economic benefits. Traditional detection methods have limitations such as low efficiency and strong destructiveness. This study designs and implements a flexible visible light spectral sensing system based on visible light spectral sensing technology and low-cost environmentally friendly flexible circuit technology. The system is structured based on a perception-analysis-warning-processing framework, utilizing laser-induced graphene electroplated copper integrated with laser etching technology for hardware fabrication, and developing corresponding data acquisition and processing functionalities. Taking Yunnan Yumang as the research object, a three-level chilling injury label dataset was established. After Z-Score standardization processing, the prediction accuracy of the SVM (Support Vector Machine) model reached 95.5%. The system has a power consumption of 230 mW at 4.5 V power supply, a battery life of more than 130 days, stable signal transmission, and a monitoring interface integrating multiple functions, which can provide real-time warning and intervention, thus offering an efficient and intelligent solution for chilling injury monitoring in mango cold chain storage. Full article

Continuous-time Recurrent Neural Networks (CTRNNs) are well-suited for modeling temporal dynamics in low-power neuromorphic and analog computing systems, making them promising candidates for edge-based human activity recognition (HAR) in healthcare. However, training CTRNNs remains challenging due to their continuous-time nature and the need to respect physical hardware constraints. In this work, we propose G-CTRNN, a novel gradient-based training framework for analog-friendly CTRNNs designed for embedded healthcare applications. Our method extends Backpropagation Through Time (BPTT) to continuous domains using TensorFlow’s automatic differentiation, while enforcing constraints on time constants and synaptic weights to ensure hardware compatibility. We validate G-CTRNN on the WISDM human activity dataset, which simulates realistic wearable sensor data for healthcare monitoring. Compared to conventional RNNs, G-CTRNN achieves superior classification accuracy with fewer parameters and greater stability—enabling continuous, real-time HAR on low-power platforms such as MEMS computing networks. The proposed framework provides a pathway toward on-device AI for remote patient monitoring, elderly care, and personalized healthcare in resource-constrained environments. Full article

This paper presents a real-time remote-control platform for small wind turbines (SWTs) equipped with a permanent magnet synchronous generator (PMSG). The proposed system integrates a DC–DC boost converter controlled by an Arduino^® microcontroller, a Raspberry Pi^® hosting a WebSocket server, and a desktop application developed using MATLAB^® App Designer (version R2024b). The platform enables seamless remote monitoring and control by allowing upper layers to select the turbine’s operating mode—either Maximum Power Point Tracking (MPPT) or Power Curtailment—based on real-time wind speed data transmitted via the WebSocket protocol. The communication architecture follows the IEC 61400-25 standard for wind power system communication, ensuring reliable and standardized data exchange. Experimental results demonstrate high accuracy in controlling the turbine’s operating points. The platform offers a user-friendly interface for real-time decision-making while ensuring robust and efficient system performance. This study highlights the potential of combining open-source hardware and software technologies to optimize SWT operations and improve their integration into distributed renewable energy systems. The proposed solution addresses the growing demand for cost-effective, flexible, and remote-control technologies in small-scale renewable energy applications. Full article

Large language models (LLMs) have revolutionized natural language processing (NLP), yet their potential in Internet of Things (IoT) and embedded systems (ESys) applications remains largely unexplored. Traditional IoT interfaces often require specialized knowledge, creating barriers for non-technical users. We present Vega, a modular system that leverages LLMs to enable intuitive, natural language control and interrogation of IoT devices, specifically, a Raspberry Pi (RPi) connected to various sensors, actuators, and devices. Our solution comprises three key components: a physical circuit with input and output devices used to showcase the LLM’s ability to interact with hardware, an RPi integrating a control server, and a web application integrating LLM logic. Users interact with the system through natural language, which the LLM interprets to remotely call appropriate commands for the RPi. The RPi executes these instructions on the physically connected circuit, with outcomes communicated back to the user via LLM-generated responses. The system’s performance is empirically evaluated using a range of task complexities and user scenarios, demonstrating its ability to handle complex and conditional logic without additional coding on the RPi, reducing the need for extensive programming on IoT devices. We showcase the system’s real-world applicability through physical circuit implementation while providing insights into its limitations and potential scalability. Our findings reveal that LLM-driven IoT control can effectively bridge the gap between complex device functionality and user-friendly interaction, and also opens new avenues for creative and intelligent IoT applications. This research offers insights into the design and implementation of LLM-integrated IoT interfaces. Full article

Artificial intelligence of things (AIoT) has become a pivotal enabler of precision agriculture by supporting real-time, data-driven decision-making at the edge. Deep learning (DL) models are central to this paradigm, offering powerful capabilities for analyzing environmental and climatic data in a range of agricultural applications. However, deploying these models on edge devices remains challenging due to constraints in memory, computation, and energy. Existing model compression techniques predominantly target large-scale 2D architectures, with limited attention to one-dimensional (1D) models such as gated recurrent units (GRUs), which are commonly employed for processing sequential sensor data. To address this gap, we propose a novel three-stage coarse-to-fine compression framework, termed SUQ-3 (Structured, Unstructured Pruning, and Quantization), designed to optimize 1D DL models for efficient edge deployment in AIoT applications. The SUQ-3 framework sequentially integrates (1) structured pruning with an $M\times N$ sparsity pattern to induce hardware-friendly, coarse-grained sparsity; (2) unstructured pruning to eliminate low-magnitude weights for fine-grained compression; and (3) quantization, applied post quantization-aware training (QAT), to support low-precision inference with minimal accuracy loss. We validate the proposed SUQ-3 by compressing a GRU-based crop recommendation model trained on environmental and climatic data from an agricultural dataset. Experimental results show a model size reduction of approximately 85% and an 80% improvement in inference latency while preserving high predictive accuracy (F1 score: 0.97 vs. baseline: 0.9837). Notably, when deployed on a mobile edge device using TensorFlow Lite, the SUQ-3 model achieved an estimated energy consumption of 1.18 μJ per inference, representing a 74.4% reduction compared with the baseline and demonstrating its potential for sustainable low-power AI deployment in agricultural environments. Although demonstrated in an agricultural AIoT use case, the generality and modularity of SUQ-3 make it applicable to a broad range of DL models across domains requiring efficient edge intelligence. Full article

Microfluidic devices are tiny tools used to manipulate small volumes of liquids in various fields. However, these devices frequently require additional equipment to control fluid flow, increasing the cost and complexity of the systems and limiting their potential for widespread use in low-resource biomedical applications. Here, we present a cost-effective and simple fabrication method for PMMA microfluidic chips using laser cutting technology, along with a low-cost and open-source peristaltic pump constructed with common hardware. The pump, programmed with an Arduino microcontroller, offers precise flow control in microfluidic devices for small volume applications. The developed application for controlling the peristaltic pump is user-friendly and open source. The microfluidic chip and pump system was tested using Jurkat cells. The cells were cultured for 24 h in conventional cell culture and a microfluidic chip. The LDH assay indicated higher cell viability in the microfluidic chip (111.99 ± 7.79%) compared to conventional culture (100 ± 15.80%). Apoptosis assay indicated 76.1% live cells, 18.7% early apoptosis in microfluidic culture and 99.2% live cells, with 0.5% early apoptosis in conventional culture. The findings from the LDH and apoptosis analyses demonstrated an increase in both cell proliferation and cellular stress in the microfluidic system. Despite the increased stress, the majority of cells maintained membrane integrity and continued to proliferate. In conclusion, the chip fabrication method and the pump offer advantages, including design flexibility and precise flow rate control. This study promises solutions that can be tailored to specific needs for biomedical applications. Full article

To meet the demands of real-time and high-definition (HD) image processing applications, denoising methods must be both computationally efficient and hardware friendly. Traditional image denoising techniques are typically simple, fast, and resource-efficient but often fall short in terms of denoising performance and adaptability. This paper proposes an adjustable-threshold denoising method along with a corresponding hardware implementation designed to support the real-time processing of large-array images commonly used in image signal processors (ISPs). The proposed technique employs a LeGall 5/3 wavelet with a row-transform structure and multilevel decomposition. A 2D Pyramid VisuShrink thresholding algorithm is introduced, where the threshold is derived from the median value of the HH sub-band using a multi-stage segmentation approach. To further optimize performance, a quantization strategy with fixed-point parameter design is applied to minimize storage requirements and computational errors. A specialized hardware architecture is developed to enable the real-time denoising of 4K images while adhering to constraints on speed and resource utilization. The architecture incorporates a finite state machine (FSM) and a reusable median calculation unit to efficiently share threshold-related storage and computational resources. The system is implemented and verified on an FPGA, achieving real-time performance at a maximum frequency of 230 MHz. It supports flexible input data formats with resolutions up to 4096×4096 pixels and 16-bit depth. Comprehensive comparisons with other real-time denoising methods demonstrate that the proposed approach consistently achieves better PSNR and SSIM across various noise levels and image sizes. In addition to delivering improved denoising accuracy, the hardware implementation offers advantages in processing speed and resource efficiency while supporting a wide range of large-array images. Full article

(1) Background: Brain–computer interfaces (BCIs) enable direct communication between the brain and external devices using electroencephalography (EEG) signals, offering potential applications in assistive technology and neurorehabilitation. Code-modulated visual evoked potential (cVEP)-based BCIs employ code-pattern-based stimulation to evoke neural responses, which can then be classified to infer user intent. While increasing the number of EEG electrodes across the visual cortex enhances classification accuracy, it simultaneously reduces user comfort and increases setup complexity, duration, and hardware costs. (2) Methods: This online BCI study, involving thirty-eight able-bodied participants, investigated how reducing the electrode count from 16 to 6 affected performance. Three experimental conditions were tested: a baseline 16-electrode configuration, a reduced 6-electrode setup without retraining, and a reduced 6-electrode setup with retraining. (3) Results: Our results indicate that, on average, performance declines with fewer electrodes; nonetheless, retraining restored near-baseline mean Information Transfer Rate (ITR) and accuracy for those participants for whom the system remained functional. The results reveal that for a substantial number of participants, the classification pipeline fails after electrode removal, highlighting individual differences in the cVEP response characteristics or inherent limitations of the classification approach. (4) Conclusions: Ultimately, this suggests that minimal cVEP-BCI electrode setups capable of reliably functioning across all users might only be feasible through other, more flexible classification methods that can account for individual differences. These findings aim to serve as a guideline for what is currently achievable with this common cVEP paradigm and to highlight where future research should focus in order to move closer to a practical and user-friendly system. Full article

Significant vulnerabilities in traditional authentication systems have been demonstrated due to the high dependence on smartphone hardware devices to execute many different and complicated tasks. PINs, passwords, and static biometric techniques have been shown to be subjected to various serious attacks, such as environmental limitations, spoofing, and brute force attacks, and this in turn mitigates the security level of the entire system. In this study, a robust framework for smartphone authentication is presented. Touch dynamic pattern recognitions, including trajectory curvature, touch pressure, acceleration, two-dimensional spatial coordinates, and velocity, have been extracted and assessed as behavioral biometric features. The TOPSIS (Technique for Order of Preference by Similarity to Ideal Solution) methodology has also been incorporated to obtain the most affected and valuable features, which are then fed as input to three different Machine Learning (ML) algorithms: Random Forest (RF), Gradient Boosting Machines (GBM), and K-Nearest Neighbors (KNN). Our analysis, supported by experimental results, ensure that the RF model outperforms the two other ML algorithms by getting F1-Score, accuracy, recall, and precision of 95.1%, 95.2%, 95.5%, and 94.8%, respectively. In order to further increase the resiliency of the proposed technique, the data perturbation approach, including temporal scaling and noise insertion, has been augmented. Also, the proposal has been shown to be resilient against both environmental variation-based attacks by achieving accuracy above 93% and spoofing attacks by obtaining a detection rate of 96%. This emphasizes that the proposed technique provides a promising solution to many authentication issues and offers a user-friendly and scalable method to improve the security of the smartphone against cybersecurity attacks. Full article

A STEM education provides students with a friendly and efficient environment for learning science, technology, engineering, and math. According to the needs of STEM programs and activities, humanoid, biped, and quadruped robots have been developed. Those robots are used as a learning tool supporting students in exploring the principles and theory of robotics and their related applications. In addition, those robots adapt open sources to provide free instructions for the students to build their own low-cost robots. To enhance the effects, a low-cost, two-wheel robot was created in this study. Unlike other robots, two-wheel robots usually require a gyroscope sensor and a motion controller to keep them balanced. The developed robot is an integrated system including hardware and software. Its hardware consists of an ESP32 microcontroller, a pair of DC motors, a gyroscope sensor MPU6050, and a driver for DC motors. The robot receives signals “angle” from the gyroscope, and then depends on the PID approach to drive the DC motors precisely in order to achieve balanced and smooth motions. The results of this study present the design of the robot, sensor calibration methods, and proportional-integral-derivative tuning. Full article

Action representations are essential for developing mutual cognition toward efficient human–AI collaboration, particularly in human–robot collaborative (HRC) workspaces. As such, it has become an emerging research direction for robots to understand human intentions with video Transformers. Despite their remarkable success in capturing long-range dependencies, local redundancy in video frames can add up to the inference latency of Transformers due to overparameterization. Recently, token pruning has become a computationally efficient solution that selectively removes input tokens with minimal impact on task performance. However, existing sparse coding methods often have an exhaustive threshold searching process, leading to intensive hyperparameter search. In this paper, Bayesian Prototypical Pruning (ProtoPrune), a novel end-to-end Bayesian framework, is proposed for token pruning in video understanding. To improve robustness, ProtoPrune leverages prototypical contrastive learning for fine-grained action representations, bringing sub-action level supervision to the video token pruning task. With variational dropout, our method bypasses the exhaustive threshold searching process. Experiments show that the proposed method can achieve a pruning rate of $37.2\%$ while retaining $92.9\%$ of task performance using Uniformer and ActionCLIP, which significantly improves computational efficiency. Convergence analysis ensures the stability of our method. The proposed efficient video understanding method offers a theoretically grounded and hardware-friendly solution for deploying video Transformers in real-world HRC environments. Full article

The Transformer-based target detection model, DETR, has powerful feature extraction and recognition capabilities, but its high computational and storage requirements limit its deployment on resource-constrained devices. To solve this problem, we first replace the ResNet-50 backbone network in DETR with Swin-T, which realizes the unification of the backbone network with the Transformer encoder and decoder under the same Transformer processing paradigm. On this basis, we propose a quantized inference scheme based entirely on integers, which effectively serves as a data compression method for reducing memory occupation and computational complexity. Unlike previous approaches that only quantize the linear layer of DETR, we further apply integer approximation to all non-linear operational layers (e.g., Sigmoid, Softmax, LayerNorm, GELU), thus realizing the execution of the entire inference process in the integer domain. Experimental results show that our method reduces the computation and storage to 6.3% and 25% of the original model, respectively, while the average accuracy decreases by only 1.1%, which validates the effectiveness of the method as an efficient and hardware-friendly solution for target detection. Full article

In response to the issue that post-training quantization leads to performance degradation in mobile deployment, as well as the problem that the balanced consideration of quantization deviation by Clipping optimization techniques limits the improvement of quantization accuracy, this article proposes a novel clipping optimization method named ClipQ, which pays different attention to the parameters, aiming to preferentially reduce the quantization deviation of important parameters. The attention of the weight is positively related to its absolute value. Channel information entropy and principal component analysis are used to characterize the channel attention and spatial attention of activations, respectively. In addition, the particle swarm algorithm is applied in weight clipping to adjust the search step size and direction adaptively. ClipQ achieves high-precision quantization with very few calibration samples (<=50) and low time cost. Meanwhile, it does not bring extra computation, which is friendly to hardware. The experimental evaluation on image classification, semantic segmentation, and object detection shows that ClipQ outperforms other state-of-the-art clipping techniques, such as KL, ACIQ, and MSE. In 8-bit quantization, the average precision loss is 0.31% for image classification and 0.22% for object detection. More notably, it achieves almost lossless accuracy in semantic segmentation tasks. Full article

A simple algorithm for respiratory activity detection in data produced by Frequency-Modulated Continuous-Wave (FMCW) radars is presented in this paper. The proposed computational architecture can be directly mapped onto custom digital–analog VLSI hardware, which is a unique approach in research on intelligent FMCW sensor development, offering a potential energy-efficient data analysis solution for target applications, such as preventing human trafficking or providing life-sign detection under limited visibility. The algorithm comprises two main modules. The first one summarizes radar-produced data into a descriptor reflecting the amount of motion that occurs within appropriately determined time intervals. The second one classifies a sequence of the produced descriptors using a recurrent neural network composed of gated recurrent units. To ensure the algorithm’s implementation feasibility, an analog VLSI circuit comprising its main functional blocks has been designed, manufactured, and tested, providing constraints for neural model derivation. The adverse effects of the primary constraint, the severe restriction on admissible weight resolution, have been handled by introducing a novel training loss component and a simple mechanism for diversifying the effective weight sets of different network neurons. Experimental evaluation of the presented method, performed using the dataset of indoor recordings, indicates that the proposed simple, hardware implementation-friendly algorithm provides over 94% human detection accuracy and similar F1 scores. Full article

Background/Objectives: Multi-dimensional liquid chromatography coupled with mass spectrometry (mD-LC-MS) has emerged as a powerful technique for the in-depth characterization of biopharmaceuticals by assessing chromatographically resolved product variants in a streamlined and semi-automated manner. The study aims to demystify and enhance the accessibility to this powerful but inherently complex technique by detailing a robust and user-friendly instrument platform, allowing analysts to switch seamlessly between intact, subunit, and peptide mapping workflows. Methods: Starting from a commercially available Two-Dimensional Liquid Chromatography (2D-LC) system, we introduce specific hardware and software extensions leading to two versatile mD-LC-MS setups, in slightly different configurations. The technique’s efficacy is demonstrated through a case study on a cation exchange chromatography method assessing the charge variants of a bispecific antibody, isolating peak(s) of interest, followed by online sample processing, including reduction and enzymatic digestion, and subsequently mass spectrometry analysis. Results: The accuracy and reproducibility of both mD-LC-MS setups proposed in this study were successfully tested. Despite the complex peak patterns in the first dimension, the systems were equally effective in identifying and quantifying the underlying product species. This case study highlights the routine usability of mD-LC-MS technology for the characterization of (ultra) high-performance liquid chromatography (UHPLC) of therapeutic biomolecule. Conclusions: The demonstrated reliability and accuracy underscore the practicality of mD-LC-MS for routine use in biopharmaceutical analysis. Our detailed description of the mD-LC-MS systems and insights simplify access to this advanced technology for a broader scientific community, regardless of expertise level, and lower the entry barrier for its use in various research and industrial settings. Full article