Search Results (1,026)

Integrating road vehicles into broader Internet of Things (IoT) ecosystems is an important step in the development of fully connected and smart transportation systems. This research explores the potential of using communication technologies that achieve a balance between low-power and long-range (LPLR) capabilities while remaining cost-effective, specifically Bluetooth Classic BR-EDR, Bluetooth LE, ZigBee, nRF24, and LoRa—for Vehicle-to-Infrastructure (V2I) and Vehicle-to-IoT (V2IoT) ecosystem interactions. During this research, several field tests were conducted employing different types of communication modules, across three distinct environments: an open-field inter-urban road, a forest inter-urban road, and an urban road. The modules were evaluated based on the communication range, messaging rate, error rate, and geographical data from GNSS (Global Navigation Satellite System) coordinates, using point-to-point communication between a roadside unit (RSU) and a moving vehicle equipped with an onboard unit (OBU). The results demonstrate the usability of these technologies for integrating vehicles into both public infrastructure (for V2I services) and private IoT systems, highlighting their potential for scalable, cost-effective deployment in smart transportation systems. Full article

This article presents CapFlow, an integrated approach to detailed image captioning and hashtag generation. Based on a thorough performance evaluation, the image captioning model utilizes a fine-tuned vision-language model with Low-Rank Adaptation (LoRA), while the hashtag generation employs the keyword extraction method. We evaluated the state-of-the-art image captioning models using both traditional metrics (BLEU, METEOR, ROUGE-L, and CIDEr) and the specialized CAPTURE metric for detailed captions. The hashtag generation models were assessed using precision, recall, and F1-score. The proposed method demonstrates competitive results against larger models while maintaining efficiency suitable for real-time applications. The image captioning model outperforms the base Florence-2 model and favorably compares with larger models. The KeyBERT implementation for hashtag generation surpasses other keyword extraction methods in both accuracy and speed. This work contributes to the field of AI-assisted content analysis and generation, offering insights into the practical implementation of advanced vision-language models for detailed image understanding and relevant tag generation. Full article

Efficiently reporting soil-specific information is of key importance for plant growth but can be quite demanding as well. Indeed, it may require expensive digitizers, subscriptions to services for communication links between each sensor and the cloud, and the incorporation of power-hungry elements. Added to this, soil sensors may vary drastically, e.g., in terms of power characteristics, response times, or interfacing options. The need for improved energy autonomy increases reporting complexity, as it presupposes that the participating components will enter a low-power (sleep) state when not in action. Furthermore, the IoT nodes hosting the sensing instruments should be able to work unattended for long periods under varying environmental conditions. In response to the aforementioned physical and technical challenges, this work highlights the details behind the cooperation of a cost-effective microprocessor equipped with a radio transceiver and some simple and widely available electronic components to form nodes that can host a diverse set of soil sensors and deliver reliable data in satisfactory ranges. The sensitivity and power efficiency of the LoRa protocol make it ideal for rural agri-field use; in the meantime, optimized action/sleep management, along with tiny solar panels, guarantee sustainable operation. The proposed system was tested utilizing various typical soil instruments, and its range coverage, consumption, and measurement quality were thoroughly evaluated under different installation settings, thus providing guidance for similar implementations and indicating its suitability for a wide set of monitoring applications. Full article

Recent studies on parameter-efficient fine-tuning (PEFT) have introduced effective and efficient methods for fine-tuning large language models (LLMs) on downstream tasks using fewer parameters than required by full fine-tuning. Low-rank decomposition adaptation (LoRA) significantly reduces the parameter count to 0.03% of that in full fine-tuning, maintaining satisfactory performance when training only two low-rank parameters. However, limitations remain due to the lack of task-specific parameters involved in training. To mitigate these issues, we propose the Lottery Rank-Pruning Adaptation (LoRPA) method, which utilizes the Lottery Ticket Hypothesis to prune less significant parameters based on their magnitudes following initial training. Initially, LoRPA trains with a relatively large rank size and then applies pruning to enhance performance in subsequent training with fewer parameters. We conducted experiments to compare LoRPA with LoRA baselines, including a setting with a relatively large rank size. Experimental results on the GLUE dataset with RoBERTa demonstrate that LoRPA achieves comparable results on the base scale while outperforming LoRA with various rank sizes by 0.04% to 0.74% on a large scale across multiple tasks. Additionally, on generative summarization tasks using BART-base on the CNN/DailyMail and XSum datasets, LoRPA outperformed LoRA at the standard rank size and other PEFT methods in most of the metrics. These results validate the efficacy of lottery pruning for LoRA in downstream natural-language understanding and generation tasks. Full article

Water pollution presents one of the biggest challenges in the world today, as the degradation of water quality of rivers in many instances is increasing so fast and poses a big danger to all forms of life, eventually causing many aquatic species and other species that depend on them to be endangered. Hence, with the development of Internet of Things (IoT) and Wireless Sensor Networks (WSNs), there arises a need to monitor river waters for a timely response in protecting the rivers, which is the aim of this paper. With respect to this project, we searched a little bit for some existing IoT technologies and other related work. In this paper, we propose a practical low-cost solution based on Long Range (LoRa) technology to obtain real-time observations of, with certain sensors, such water parameters as temperature, pH, conductivity and turbidity. Data gathered at a sensor node are transmitted via LoRa modulation to a gateway for processing and local storage on a Message Queuing Telemetry Transport (MQTT) server, visualization on a Node-RED interface, or transmission to the cloud. The prototype system created is employed in the actual field and demonstrates that the water quality monitoring in the river can be carried out effectively within a small scale of the area of roughly 20 km² depending on the location of the study site. Full article

With the advent of artificial intelligence digitization, intangible cultural heritage faces challenges in preservation and transmission. Utilizing modern technology to achieve digital protection and dissemination has become a crucial issue today. This study enhances the digital inheritance and development of Chinese intangible cultural heritage paper-cutting art through generative AI technologies, specifically Diffusion and LoRA models. The Analytic Hierarchy Process (AHP) was employed to categorize the cultural value of paper-cutting, selecting four core elements: “Spring Festival”, “Chinese Zodiac”, “Women”, and “Birds and Flowers”. Based on these, eight LoRA models were developed to generate paper-cutting-style patterns (using the FLUX.1-dev and Stable Diffusion 1.5 models). In the user satisfaction assessment, the Importance–Performance Analysis (IPA) method was used to analyze four dimensions of the model experience. The results indicate that the LoRA model excels in generating detailed paper-cutting patterns and accurately reproducing cultural elements, particularly in the generation of complex Chinese character designs. User feedback suggests that the LoRA model effectively enhances the digital representation and dissemination of paper-cutting art, though there is room for improvement in terms of generation speed and ease of operation. This study provides a new technological pathway for the digital preservation of intangible cultural heritage and promotes the modernization of paper-cutting art transmission. Full article

Wireless body area networks (WBANs), which continually gather and transmit patient health data in real time, are essential for improving healthcare administration. Patient outcomes can be improved by sending these data to medical professionals for prompt review and treatment. For the effective deployment of WBANs, communication solutions are necessary to maximize critical performance parameters, such as low power consumption, minimal delay, and acceptable data rates, while guaranteeing dependable transmission. Two prominent technologies in this field are LoRaWAN, which is renowned for its long-range capabilities and energy efficiency, and IEEE 802.15.6, which was created especially for short-range medical applications with high data throughput. This study provides a comparative evaluation of these two technologies to determine their suitability for diverse WBAN healthcare scenarios. By using the NS3, a simulation was performed to calculate six key performance metrics: throughput, arrival rate, delay, energy consumption, packet delivery ratio (PDR), and network lifetime. The study analyzed each technology’s performance under varying node counts. At a density of 50 nodes, IEEE 802.15.6 demonstrated superior throughput, with 45 kbps, compared to LoRaWAN, and a higher PDR of 30%. Additionally, IEEE 802.15.6 showed a higher arrival rate, of 0.33%, than LoRaWAN. On the other hand, LoRaWAN showed notable strengths in energy consumption, with only 42 J, compared to IEEE 802.15.6, and significantly lower delay, with a delay of 7 s. Additionally, LoRaWAN offered an extended network lifetime, of 18 h, compared to IEEE 802.15.6. Full article

We propose a story visualization technique that allows users to control the arrangement, poses, and styles of characters in a scene based on user-input doodle sketches. Our method utilizes a text encoder to process scene prompts and an image encoder to handle doodle sketches, generating inputs for a predefined scene generation model. Furthermore, we achieve efficient model training by fine-tuning the backbone network by applying a small dataset and employing a LoRA-based fine-tuning technique. We demonstrate that our method can generate characters with various poses and styles from doodle sketches, and it can validate the advantages of our approach by comparing it with the results from other story visualization studies. Full article

Cultured meat technology is a form of cellular agriculture where meat is produced from animal cells grown in a lab, instead of raising and slaughtering animals. This technology relies heavily on fetal bovine serum (FBS) in cell media; hence, production is costly and contributes significantly to ammonia and greenhouse gas emissions. Achieving the successful commercialization of cell-cultured food requires the critical resolution of manufacturing cost and safety concerns. Hence, our research efforts are focused on identifying commercially viable and ecologically sustainable alternatives to FBS. In this study, we evaluated the potential of twenty-six water-based algal and cyanobacterial extracts to stimulate cell growth for meat cultivation under 90% reduced serum conditions. The extracts were compared in viability, proliferation, and Trypan blue exclusion assays. In the first screening phase, the extracts were evaluated in a ZEM2S (zebrafish) cell culture in a 1% FBS regimen. Based on their ability to exhibit protein tolerance or promote cell proliferation, ten extracts were selected and further assayed in a QM7 cell culture. The QM7 cell line (myoblasts from Japanese quail) is highly relevant for meat cultivation because of its ability to differentiate into muscle fibers. Extracts derived from two microalgae species, Arthrospira platensis (Spirulina) and Dunaliella tertiolecta, demonstrated the highest tolerance in cell culture, above 10 μg/mL (expressed as total protein concentration). Tolerance at a 100 μg/mL concentration was demonstrated exclusively using an extract of blue spirulina (commercially purified Spirulina), which supported cell growth through multiple passages. Full article

Large language models (LLMs) have recently made significant advances, excelling in tasks like question answering, summarization, and machine translation. However, their enormous size and hardware requirements make them less accessible to many in the machine learning community. To address this, techniques such as quantization, prefix tuning, weak supervision, low-rank adaptation, and prompting have been developed to customize these models for specific applications. While these methods have mainly improved text generation, their implications for the text classification task are not thoroughly studied. Our research intends to bridge this gap by investigating how variations like model size, pre-training objectives, quantization, low-rank adaptation, prompting, and various hyperparameters influence text classification tasks. Our overall conclusions show the following: 1—even with synthetic labels, fine-tuning works better than prompting techniques, and increasing model size does not always improve classification performance; 2—discriminatively trained models generally perform better than generatively pre-trained models; and 3—fine-tuning models at 16-bit precision works much better than using 8-bit or 4-bit models, but the performance drop from 8-bit to 4-bit is smaller than from 16-bit to 8-bit. In another scale of our study, we conducted experiments with different settings for low-rank adaptation (LoRA) and quantization, finding that increasing LoRA dropout negatively affects classification performance. We did not find a clear link between the LoRA attention dimension (rank) and performance, observing only small differences between standard LoRA and its variants like rank-stabilized LoRA and weight-decomposed LoRA. Additional observations to support model setup for classification tasks are presented in our analyses. Full article

The urban road spatial structure is a crucial and complex component of urban design. Generative design models, such as the Stable Diffusion model, can rapidly and massively produce designs. However, the opacity of their internal architecture and the uncertainty of their outcomes mean that the results generated do not meet specific disciplinary assessment criteria, thus limiting their widespread application in planar design and planning. Additionally, traditional software processes targeting specific indicators are time-consuming and do not allow for rapid evaluation. To address these challenges, we utilized several areas of the road spatial structures in six cities and their corresponding four space-syntax parameters as training samples. We simultaneously trained two models: one is a LoRA Model based on the Stable Diffusion architecture used for generating road networks similar to those of various city road spatial structures; the other is a CoAtNet Model (Convolution + Transformer) used as an evaluation model to predict the space-syntax parameters of road structures and calculate the Mean Absolute Percentage Error (MAPE) relative to real urban samples. Subsequently, by linking these two models end-to-end, we were able to filter out generated samples with the smallest MAPE, thereby enhancing the structural similarity between the generated results and the actual urban road spatial structures. This process of rapid generation and swift evaluation of network configurations marks a critical advancement towards better performance and more customized design solutions. Full article

Landslides cause severe damage to life and property with a wide-ranging impact. Infiltration of rainfall is one of the significant factors leading to landslides. This paper reports on a phase creep landslide caused by long-term rainfall infiltration. A detailed geological survey of the landslide was conducted, and the deformation development pattern and mechanism of the landslide were analyzed in conjunction with climatic characteristics. Furthermore, reinforcement measures specific to the landslide area were proposed. To monitor the stability of the reinforced slope, a Beidou intelligent monitoring and warning system suitable for remote mountainous areas was developed. The system utilizes LoRa Internet of Things (IoT) technology to connect various monitoring components, integrating surface displacement, deep deformation, structural internal forces, and rainfall monitoring devices into a local IoT network. A data processing unit was established on site to achieve preliminary processing and automatic handling of monitoring data. The monitoring results indicate that the reinforced slope has generally stabilized, and the improved intelligent monitoring system has been able to continuously and accurately reflect the real-time working conditions of the slope. Over the two-year monitoring period, 13 early warnings were issued, with more than 90% of the warnings accurately corresponding to actual conditions, significantly improving the accuracy of early warnings. The research findings provide valuable experience and reference for the monitoring and warning of high slopes in mountainous areas. Full article

For both human health and the environment, air pollution is a serious concern. However, the available air quality monitoring networks have important limitations, such as the high implementation costs, limited portability, and considerable operational complexity. In this context, unmanned aerial systems (UASs) are emerging as a useful technological alternative due to their ability to cover large distances and access areas that are difficult or impossible for humans to reach. This article presents the development of an integrated platform that combines an unmanned aerial system (UAS) with specialized sensors to measure key parameters in relation to air quality, such as carbon monoxide (CO), ozone (O₃), and nitrogen dioxide (NO₂). In addition, a web application called PTECA is developed to visualize the data gathered by the wireless sensor array in real time. The platform incorporates a system that allows real-time tracking of the UAS route and measurement values during sample collection, employing the LoRa communication protocol. This solution represents a low-cost alternative that mitigates some of the limitations of traditional monitoring networks by offering greater portability and accessibility in terms of data collection. Preliminary tests successfully demonstrate the viability of the proposed system in a controlled airspace using geofencing. Full article

The Internet of Things (IoT) is a fast-growing field that has found a variety of applications, such as smart agriculture and industrial processing. In these applications, it is important for nodes to maximize the amount of useful information transmitted over a limited channel. This work seeks to improve the performance of low-powered sensor networks by developing an architecture that leverages existing techniques such as lossy compression and different queuing strategies in order to minimize their drawbacks and meet the performance needs of backend applications. The Age of Information (AoI) provides a useful metric for quantifying Quality of Service (QoS) in low-powered sensor networks and provides a method for measuring the freshness of data in the network. In this paper, we investigate QoS requirements and the effects of lossy compression and queue strategies on AoI. Furthermore, two important use cases for low-powered IoT sensor networks are studied, namely, real-time feedback control and image classification. The results highlight the relative importance of QoS metrics for applications with different needs. To this end, we introduce a QoS-aware architecture to optimize network performance for the QoS requirements of the studied applications. The proposed network architecture was tested with a mixture of application traffic settings and was shown to greatly improve network QoS compared to commonly used transmission architectures such as Slotted ALOHA. Full article

The long-range wide area network (LoRaWAN) protocol is one of the most effective technologies for internet of things (IoT) applications, offering long-distance connectivity with low power consumption. This paper presents a practical approach by implementing a LoRa-based measurement prototype across urban and rural environments in the city of Quito, with the aim of assessing the performance and applicability of the technology in manifold settings. Specifically, we develop the required data collection and transmission code in the underlying network, ensuring smooth network integration. Furthermore, test environments are thoroughly characterized for numerical results, highlighting the conditions in the cities of Quito. The results obtained in both scenarios were satisfactory, allowing the comparison of the system’s performance in different contexts and providing key aspects of its practical applications and effectiveness. As the main contribution, empirical data were obtained to understand how long-range low-energy connectivity behaves, providing valuable information for comparing system performance in high-altitude cities above sea level, identifying practical applications, and optimizing its use in real IoT implementations. Full article