Sensor-generated Data Research Articles

Cloud IoT-Oriented Secure College Physical Education Teaching Platform Vased on Deep Learning

In physical education (PE) teaching, the teaching platform is comprehensively applied to provide students with high-quality PE teaching resources to meet their learning needs in different forms. A human motion trajectory detection method is proposed based on deep learning and superpixels. The initial positioning of human body is performed through the attention mechanism and the YOLOv5 model, and the human target tracking is performed through superpixels. Aiming at the difficulty of cross-domain security management due to the lack of security infrastructure in multi-domain cloud IoT, a lightweight certificateless cross-domain authentication scheme is designed. Simulation results show that the proposed detection method has significantly improved accuracy and running time compared with traditional methods and can adapt to different detection scenarios. Furthermore, the proposed certificateless cross-domain authentication scheme can securely access sensor-generated data under the cloud IoT of the PE teaching platform in colleges.

Proposal of a Machine Learning Approach for Traffic Flow Prediction.

Rapid global urbanization has led to a growing urban population, posing challenges in transportation management. Persistent issues such as traffic congestion, environmental pollution, and safety risks persist despite attempts to mitigate them, hindering urban progress. This paper focuses on the critical need for accurate traffic flow forecasting, considered one of the main effective solutions for containing traffic congestion in urban scenarios. The challenge of predicting traffic flow is addressed by proposing a two-level machine learning approach. The first level uses an unsupervised clustering model to extract patterns from sensor-generated data, while the second level employs supervised machine learning models. Although the proposed approach requires the availability of data from traffic sensors to realize the training of the machine learning models, it allows traffic flow prediction in urban areas without sensors. In order to verify the prediction capability of the proposed approach, a real urban scenario is considered.

Sensors on Internet of Things Systems for the Sustainable Development of Smart Cities: A Systematic Literature Review.

The Internet of Things (IoT) is a critical component of smart cities and a key contributor to the achievement of the United Nations Sustainable Development Goal (UNSDG) 11: Sustainable Cities and Communities. The IoT is an infrastructure that enables devices to communicate with each other over the Internet, providing critical components for smart cities, such as data collection, generation, processing, analysis, and application handling. IoT-based applications can promote sustainable urban development. Many studies demonstrate how the IoT can improve smart cities' sustainable development. This systematic literature review provides valuable insights into the utilization of the IoT in the context of smart cities, with a particular focus on its implications for sustainable urban development. Based on an analysis of 73 publications, we discuss the role of IoT in the sustainable development of smart cities, focusing on smart communities, smart transportation, disaster management, privacy and security, and emerging applications. In each domain, we have detailed the attributes of IoT sensors. In addition, we have examined various communication technologies and protocols suitable for transmitting sensor-generated data. We have also presented the methods for analyzing and integrating these data within the IoT application layer. Finally, we identify research gaps in the literature, highlighting areas that require further investigation.

Human Activity Recognition through Smartphone Inertial Sensors with ML Approach

Human Activity Recognition (HAR) has several applications in healthcare, security, and assisted living systems used in smart homes. The main aim of these applications or systems is to classify body movement read from the built in sensors such as accelerometers and gyroscopes. Some actions could be performed in response to the output of these HAR systems. The number of smartphone users increases, whereas the sensors are widely available in different sizes and shapes (internal or external sensors). Recent advances in sensor technology and machine learning have led researchers to conduct studies on sensor technology such as HAR. HAR systems typically use a combination of sensors, such as accelerometers, gyroscopes, and cameras, to collect images or signal data that can be classified by machine learning algorithms. HAR research has focused on several key challenges including dealing with variability in sensor data, handling missing data or noise, and dealing with large amounts of sensor-generated data. In this work, several machine learning algorithms were tested in predefined settings using the KU-HAR dataset in a series of experiments. Subsequently, various performance metrics were calculated to assess the chosen algorithms’ performance. The experimental findings showed that the LightGBM classifier surpassed the other machine learning algorithms in performance metrics, such as accuracy, F1 score, precision, and recall. Although Gradient Boosting has lengthy training time, the other classifiers complete their training in an acceptable time period.

Using Large Language Models to Enhance the Reusability of Sensor Data.

The Internet of Things generates vast data volumes via diverse sensors, yet its potential remains unexploited for innovative data-driven products and services. Limitations arise from sensor-dependent data handling by manufacturers and user companies, hindering third-party access and comprehension. Initiatives like the European Data Act aim to enable high-quality access to sensor-generated data by regulating accuracy, completeness, and relevance while respecting intellectual property rights. Despite data availability, interoperability challenges impede sensor data reusability. For instance, sensor data shared in HTML formats requires an intricate, time-consuming processing to attain reusable formats like JSON or XML. This study introduces a methodology aimed at converting raw sensor data extracted from web portals into structured formats, thereby enhancing data reusability. The approach utilises large language models to derive structured formats from sensor data initially presented in non-interoperable formats. The effectiveness of these language models was assessed through quantitative and qualitative evaluations in a use case involving meteorological data. In the proposed experiments, GPT-4, the best performing LLM tested, demonstrated the feasibility of this methodology, achieving a precision of 93.51% and a recall of 85.33% in converting HTML to JSON/XML, thus confirming its potential in obtaining reusable sensor data.

Collaborative relay for achieving long-term and low-AoI data collection in UAV-aided IoT systems

In Internet of Things (IoT) systems, sensor nodes are frequently placed in remote and unattended locations to monitor environmental data. One significant challenge is ensuring the timely and efficient transmission of data generated by these sensor nodes back to the base station. The use of unmanned aerial vehicles (UAVs) can provide a practical solution to this challenge by acting as mobile relay nodes for facilitating data transmission. In most existing works, UAVs are typically restricted to collecting data within their designated areas and returning to the base station for data offloading, resulting in suboptimal data timeliness due to long-distance flights. A limited number of works have explored the utilization of relay collaboration by UAVs for data collection, enabling efficient and immediate transmission of sensor node data to the base station. Nevertheless, UAVs positioned at significant distances from the base station face challenges in obtaining timely energy replenishment. This makes them unable to effectively support long-duration data collection missions. In order to tackle these challenges, we develop a UAV-aided IoT collaborative data collection mechanism and propose a matching games-based data collection (MGDC) scheme. In this scheme, we begin by identifying convergence nodes within the ground sensor network, responsible for uploading sensor-generated data to passing UAVs. Furthermore, we divide the mission area into multiple subareas based on the number of available UAVs. Subsequently, using a matching game algorithm, we establish relay relationships between UAVs to enable efficient relay transmissions among paired UAVs. To achieve efficient data collection of UAVs, we employ an improved adaptive large neighborhood search (IALNS) algorithm for UAV flight path planning. Finally, we incorporate an alternating charging mode to ensure all UAVs have the opportunity to return to the base station for energy recharge. Through comprehensive experimentation, we confirm the significant enhancement provided by our proposed data collection scheme compared to existing schemes. This scheme effectively reduces system age of information (AoI) and extends the runtime of the system.

A Heart Image Segmentation Method Based on Position Attention Mechanism and Inverted Pyramid

In the realm of modern medicine, medical imaging stands as an irreplaceable pillar for accurate diagnostics. The significance of precise segmentation in medical images cannot be overstated, especially considering the variability introduced by different practitioners. With the escalating volume of medical imaging data, the demand for automated and efficient segmentation methods has become imperative. This study introduces an innovative approach to heart image segmentation, embedding a multi-scale feature and attention mechanism within an inverted pyramid framework. Recognizing the intricacies of extracting contextual information from low-resolution medical images, our method adopts an inverted pyramid architecture. Through training with multi-scale images and integrating prediction outcomes, we enhance the network’s contextual understanding. Acknowledging the consistent patterns in the relative positions of organs, we introduce an attention module enriched with positional encoding information. This module empowers the network to capture essential positional cues, thereby elevating segmentation accuracy. Our research resides at the intersection of medical imaging and sensor technology, emphasizing the foundational role of sensors in medical image analysis. The integration of sensor-generated data showcases the symbiotic relationship between sensor technology and advanced machine learning techniques. Evaluation on two heart datasets substantiates the superior performance of our approach. Metrics such as the Dice coefficient, Jaccard coefficient, recall, and F-measure demonstrate the method’s efficacy compared to state-of-the-art techniques. In conclusion, our proposed heart image segmentation method addresses the challenges posed by diverse medical images, offering a promising solution for efficiently processing 2D/3D sensor data in contemporary medical imaging.

Evaluation of Regional Road Transport Safety Service Level With Edge Computing in Scalable Sensor and Actuator Networks

In contemporary society, road transport has assumed a pivotal role in the daily lives of individuals, with virtually everyone being a road user in one capacity or another. Unfortunately, a significant portion of road accidents stem from user negligence and a lack of awareness regarding road safety principles and regulations. The majority of these accidents are fundamentally a result of human errors, errors that could potentially be averted through the comprehensive utilization of technology. The integration of edge computing within actuator networks, combined with the scalability of sensors, holds the potential to proactively detect and prevent impending collisions. This cutting-edge approach not only accumulates and processes sensor-generated data but also employs it to mitigate the occurrence of road accidents. This discussion delves into the potential advantages offered by smart sensors and advanced technology, focusing on the enhancement of safety within transportation networks. The core attributes of these intelligent technologies facilitate the establishment of robust connectivity and the effective implementation of processes, ushering in an innovative era in the realm of road safety through smart technology.

Predictive maintenance of industrial equipment using machine learning in industrial environment of Awka Metropolis, Nigeria

Failures of industrial equipment may lead to substantial operational interruptions and financial losses for enterprises. A viable solution to reduce such issues includes the deployment of machine learning-driven predictive maintenance. This study dives into the implementation of predictive maintenance in the particular industrial environment of Awka Metropolis, Nigeria. This research involves the design of a comprehensive approach, involving the collecting and preparation of data, along with the application of machine learning models. The core of our predictive maintenance approach resides in past equipment performance data, combined with sensor-generated data. Various machine learning methods, including decision trees, random forests, and recurrent neural networks, are used to anticipate future equipment faults. The study findings illustrate the usefulness of the predictive maintenance model in properly identifying approaching equipment problems, even under the specific circumstances of Awka Metropolis. Evaluation criteria such as precision, recall, and accuracy support the robustness of the model, underlining its trustworthiness. The paper also tackles the practical obstacles found during implementation, giving insights into their resolution, especially within parallel industrial situations. The findings underline the potential for cost reductions and heightened operational efficiency within regional industries by implementing proactive maintenance practices. Furthermore, the report suggests avenues for future inquiry and underlines the applicability of the model to varied businesses and geographical areas. The successful implementation of predictive maintenance in Awka Metropolis provides local firms a chance to boost equipment reliability while lowering downtime, so making major contributions to economic development and sustainability. As companies progressively embrace digital transformation, this study serves as a significant resource for practitioners and scholars alike, seeking to improve equipment maintenance in rising markets.

Sensor-Generated Data for Evaluation of Subclinical Mastitis Treatment Effectiveness with Garlic Extract (Allicin) in Dairy Cattle

The aim of this study was to determine the impact of subclinical mastitis treatment in dairy cattle on biomarkers registered with in-line sensors such as milk yield (MY), electric milk conductivity (EC), rumination time (RT), and somatic cell count (SCC). At the start of the experiment, all cows according to SCC level were divided into two groups: healthy cows (n = 30, with SCCs less than 200,000 per mL and without the growth of bacteria in the milk samples) and cows with subclinical mastitis (n = 32), with SCC levels greater than 200,000 per mL and with growth of bacteria. Streptococcus spp. was found in 15 samples, and Strep. uberis was found in 17 samples. Streptococcus spp. and Strep. uberis were sensitive to amoxicillin and calvulanic acid. According to these results, 32 cows with subclinical mastitis were treated with two treatment protocols: one 1 (n = 16) and two (n = 16). In the first protocol, we used SCC boluses and nonsteroidal anti-inflammatory drugs (SCCB and NSAID). The second protocol consists of intramammary antibiotics and anti-inflammatory medications (Synulox LC and NSAIDs). All parameters (MY, EC, RT, and SCC) were recorded with Lely Astronaut® A3 milking robots on the day of mastitis diagnosis (0 day) and 14 days after treatment began. All animal experimental procedures were approved by the ethical committee; the approval number is PK01696. On the basis of our findings, we may infer that SCC boluses and NSAIDs are effective in treating subclinical mastitis. After 14 days of treatment, the electrical conductivity of milk in cows treated with AB and NSAID was also higher in all quarters of the udder compared to cows treated with SCCB + NSAID. The RT of cows on disease diagnosis day of cows treated with AB and NSAID was 11.41% lower compared to cows treated with SCCB and NSAID, while the RT of cows after 14 days treated with AB and NSAID was 7.01% lower compared to cows treated with SCCB and NSAID. On the practical side, for treatment of subclinical mastitis, we recommend using a feed supplement SCC bolus (one per os) with a composition containing Meloxicam 20 mg with a single subcutaneous injection at a dosage of 2.5 mL per 100 kg body weight.

Novel IoT framework for event processing in healthcare applications

Data analysis depends heavily on the gathering, reasoning, and modeling of sensor-generated data. Applications for the Internet of Things (IoT) face difficulties in studying and decoding real-time data delivered through various wireless links. A data stream tracking technique called Event Process Healthcare (EPH) is used to extract relevant information from network results for use in immediate decision-making. For the data analysis of dependable healthcare applications, an event-driven IoT architecture with an event, context, and service layer is presented in this paper. In the proposed EPH method, a new algorithm known as Cloud-based Deep Learning (CDN) is introduced, which supports both patients and the healthcare industry utilizing a combination of machine learning techniques, an intelligent cloud system, and the deep learning norms serve as the foundation. Simulation is used to obtain empirical results, and it dramatically improves healthcare parameters furthermore, the EPH technique boosted precision, cut expenses, and improved health outcomes.

SEPARATE: A tightly coupled, seamless IoT infrastructure for deploying AI algorithms in smart agriculture environments

Precision agriculture generates large datasets from IoT infrastructures deployed for continuous crop monitoring. This data requires analysis to usefully transform this data deluge into insights that can deliver value-generating services to farmers in a timely manner. This paper introduces SEPARATE; a dynamic interoperable and decentralized infrastructure for executing both, training and inference stages of deep learning (DL) algorithms in smart agriculture scenarios. The presented infrastructure allows the execution of the inference stage at the edge, achieving a highly efficient and responsive local temperature prediction service to take actions based on the predictions generated. Moreover, the training stage is offloaded to the cloud along with the generated historical data, allowing the trained model to be periodically updated at the edge. On the one hand, our results show that the Convolutional Neural Network model together with the Long Short-Term Memory technique (CNNLSTM) obtains the best results in both prediction accuracy and computational time. On the other hand, an analysis has been carried out to determine how often the model must be retrained, obtaining results that indicate that from day 9–10, it would be necessary to retrain the model, although, until day 20, the precision is not greatly reduced. Moreover, the SEPARATE infrastructure enables the execution of real-time inference from sensor-generated data and seamless model retraining in an operational greenhouse for temperature forecast with satisfactory performance.

Online monitoring for error detection in vat photopolymerization

ABSTRACT Sensor implementation together with data processing in manufacturing provides continuous measurements enabling process monitoring, optimization, and automation. Additive manufacturing (AM) technologies are shifting from one-off and prototyping production to batch, mass, and continuous production. Current AM technologies are mostly focused on small production devices, without monitoring systems. This study aims to bring AM closer to automation by proposing an online monitor system for bottom-up photopolymerization AM (VPP) systems. The sensor-generated data is used to capture the detachment error of built parts from the build platform that otherwise cannot be observed physically by the machine operator. The detachment itself will not stop the build job, which results in lost material and operating time. The online monitoring procedure consists of two phases, an offline and an online one, respectively. The offline phase is used for training a prediction model to be used in connection with a control chart for online monitoring. The monitoring control chart is advantageous as only the detachment predictions need to be recorded. The research carried out in this article brings novelty in both the vat photopolymerization set-up, as well as in an online monitoring procedure that can be easily extended to similar technologies.

IoT and Wearables in Healthcare: Generating Insights from Sensor Data

The convergence of Internet of Things (IoT) devices and wearables within the healthcare sector signifies a paradigm shift, fundamentally altering the dynamics of patient data acquisition and interpretation. This article provides an in-depth exploration of the pivotal contributions of IoT and wearables to the healthcare domain, emphasizing their crucial role in extracting meaningful insights from sensor-generated data. In the ongoing digital evolution of healthcare, the acquisition and analysis of sensor data from diverse devices emerge as indispensable elements, driving the enhancement of patient care, facilitating advanced research methodologies and optimizing the management of various diseases.

Novel Smart Sensor Platform for Monitoring Multiple Pressure Injury Risk Factors: A Feasibility Study in a Post-Acute Care Facility.

Posture, temperature, and moisture have been identified as critical modifiable risk factors in pressure injury (PI) development. Microclimate is defined as temperature and humidity at the interface of the support surface and body. To our knowledge, no studies have used sensor technology to measure these parameters simultaneously in real time. Continuous monitoring of repositioning and microclimate provide real-time actionable insights to help deliver personalized care and measure the effectiveness of interventions. To evaluate the ability of a smart surface platform to collect and document clinical data on monitoring patients' movement and microclimate simultaneously and to compare data generated to nursing observations in order to construct an algorithm that is expected to evolve over time: (1) comparing the blinded data from nurses interacting with the patients and the system; and (2) data being collected is validating an algorithm that is expected to become more accurate over time. This prospective, descriptive single-site trial was conducted at a tertiary care facility in a large urban centre in Canada. Patients identified at risk of PIs received standard of care while placed on the smart surface for timed intervals. Nurses' assessment data were collected at three hourly timepoints using a comprehensive tool developed for the study. Sensors monitored patients' interface pressure moisture and temperature every four seconds. A comparative statistical analysis was conducted between the two datasets retrospectively. The study included a total of 104 participants; mean age of 59 years (range 21-92, ± 19.15). Sensor monitoring hours (1,407) generated 1,101,780 frames of surface data. A total of 511 nursing assessments were recorded during the study period. Sensor-generated data correlated strongly with nurse-collected data at cross-sectional intervals. There was a high level of agreement between information collected from sensors and nursing assessments: 94.7% for moisture (p<0.05), and 87.1% for temperature (p<0.05). Nurse-recorded posture assessments were compared to the smart surface platform interface pressure visualizations to determine the device's posture detection, resulting in a 92% accordance (matching 552 out of 600 nurse postures), with a binomial test determining the posture results to be statistically significant (p<0.05) (CI 95%). In addition, moisture events were matched to nurse assessments with 94.7% in accordance, identifying 39 bladder incontinence and 93 non-urinary moisture events (125 total events captured out of 132). The technology's ability to capture PI risk factors supports nursing practice. Supplementary data generated has the potential to improve efficiency of professional caregiver workflow and patient outcomes by informing targeted microclimate management strategies and decreasing unnecessary interventions. The large volume of data collected will be used as a basis for artificial intelligence applications with the potential to inform other clinical decision-making areas.

Reconsider the Art of Allergen Immunotherapy

This symposium took place during the European Academy of Allergy and Clinical Immunology (EAACI) Hybrid Congress, held in Prague, Czechia, in July 2022. Stephanie Dramburg, Medical Doctor at Charité Universitätsmedizin, Berlin, Germany, explained that the traditional diagnostic work-up for allergic rhinoconjunctivitis (AR) comprises retrospective clinical history and allergen exposure assessment, extract-based diagnostics, component-resolved diagnostics (CRD), which enables markers of genuine sensitisation to be distinguished from markers of cross-reactivity, and confirmation of clinical relevance. She highlighted that molecular IgE assessment supports the diagnostic work-up and personalised risk assessment in complex cases and that confirmation of clinical relevance of IgE results is key. Furthermore, broadening of a serum IgE response is indicative of ‘molecular spreading’. Dramburg considered that digital technologies have the potential to enhance medical decisions at the point-of-care via targeted patient information, guideline- and evidence-based clinical knowledge, and prospectively collected patient- and sensor-generated data. Marek Jutel, Medical Professor at Wroclaw Medical University, Poland, and the ALL-MED Medical Research Institute, Wroclaw, Poland, described that patients with allergies show different clinical pictures due to differing sensitisation profiles assessed at the molecular level. He noted that patients with allergies react differently to different allergen doses and allergen immunotherapy (AIT) preparations, and minor/intermediate allergens are necessary, in addition to major allergens, for optimal clinical efficacy. Jutel described that allergens that are decisive for AIT efficacy are defined in grass pollen but are not yet determined for other allergen sources such as birch and house dust mite (HDM). Petra Zieglmayer, Medical Professor at Karl Landsteiner University, Krems, Austria, and Head of Vienna Challenge Chamber, Austria, discussed that patients with allergies show complex molecular sensitisation profiles and that extract preparations from different manufacturers vary in terms of allergen composition, with major and intermediate allergens not always detectable. She clarified that optimal efficacy of AIT may only be expected from preparations containing all relevant allergen components in sufficient amounts. Zieglmayer proposed that the target should be to find a match between the patient molecular sensitisation profile and the allergen preparation and that this can be achieved.

Reconsider the Art of Allergen Immunotherapy

This symposium took place during the European Academy of Allergy and Clinical Immunology (EAACI) Hybrid Congress, held in Prague, Czechia, in July 2022. Stephanie Dramburg, Medical Doctor at Charité Universitätsmedizin, Berlin, Germany, explained that the traditional diagnostic work-up for allergic rhinoconjunctivitis (AR) comprises retrospective clinical history and allergen exposure assessment, extract-based diagnostics, component-resolved diagnostics (CRD), which enables markers of genuine sensitisation to be distinguished from markers of cross-reactivity, and confirmation of clinical relevance. She highlighted that molecular IgE assessment supports the diagnostic work-up and personalised risk assessment in complex cases and that confirmation of clinical relevance of IgE results is key. Furthermore, broadening of a serum IgE response is indicative of ‘molecular spreading’. Dramburg considered that digital technologies have the potential to enhance medical decisions at the point-of-care via targeted patient information, guideline- and evidence-based clinical knowledge, and prospectively collected patient- and sensor-generated data. Marek Jutel, Medical Professor at Wroclaw Medical University, Poland, and the ALL-MED Medical Research Institute, Wroclaw, Poland, described that patients with allergies show different clinical pictures due to differing sensitisation profiles assessed at the molecular level. He noted that patients with allergies react differently to different allergen doses and allergen immunotherapy (AIT) preparations, and minor/intermediate allergens are necessary, in addition to major allergens, for optimal clinical efficacy. Jutel described that allergens that are decisive for AIT efficacy are defined in grass pollen but are not yet determined for other allergen sources such as birch and house dust mite (HDM). Petra Zieglmayer, Medical Professor at Karl Landsteiner University, Krems, Austria, and Head of Vienna Challenge Chamber, Austria, discussed that patients with allergies show complex molecular sensitisation profiles and that extract preparations from different manufacturers vary in terms of allergen composition, with major and intermediate allergens not always detectable. She clarified that optimal efficacy of AIT may only be expected from preparations containing all relevant allergen components in sufficient amounts. Zieglmayer proposed that the target should be to find a match between the patient molecular sensitisation profile and the allergen preparation and that this can be achieved.

Type-2 fuzzy ontology-based semantic knowledge for indoor air quality assessment

Semantic web technology plays an increasing role in performing smart home applied programs and it has led to improve semantic interoperability among different systems. However, classical ontologies fail to illustrate ambiguous, incomplete, and uncertain knowledge often available in the real world. On the other hand, the air quality assessment carried out to determine “the degree of pollution” lacks accurately specified boundaries; therefore, the conventional approach based on classic ontology cannot extract real-valued memberships and consequently fails to support ambiguous, incomplete, and uncertain knowledge. Integrating semantic web of things technology (SWOT) and type-2 fuzzy logic improves the capability of semantic reasoning to retrieve query information. Annotation of sensor-generated data and the ability to infer and represent knowledge based on type-2 fuzzy logic are extremely essential when the available data are ambiguous and uncertain. Hence, in this paper, we have provided a framework to build an IoT-based home air quality assessment system by using type-2 fuzzy ontology so that smart home systems can make a decision and control appropriately based on predefined rules by employing the provided semantic reasoning.

Adaptive Model Verification for Modularized Industry 4.0 Applications

Cyber-Physical Systems (CPSs) are the core of Industry 4.0 applications, integrating advanced technologies such as sensing, data analytics, and artificial intelligence. This kind of combination typically consists of networked sensors and decision-making processes in which sensor-generated data drive the control decisions. Hence, the trustworthiness of the sensors is essential to guarantee performance, safety and quality during operation. Formal model verification techniques are a valuable tool allowing strong reasoning about the high-level design of CPSs. However, the uncertainty exhibited by the underlying sensor networks is often ignored. Manufacturing processes typically involve composition of various modular CPSs that work as a whole, such as multiple Collaborative Robots (cobots) working together as a production line, which improves the flexibility and resilience of the production process. It is still challenging to verify this class of compositional process while also considering uncertainty. We propose a novel verification framework for modular CPSs that combines sensor-level data-driven fault detection and system-level model-driven probabilistic model checking. The resulting framework can rigorously quantify sensor readings’ trustworthiness, enabling formal reasoning for system failure prediction and reliability analysis. We validated our approach on a cobots-based manufacturing process.

MuSDRI: Multi-Seasonal Decomposition Based Recurrent Imputation for Time Series

Missing values are a common problem of time series such as sensor-generated data due to issues of data collection, transmission, and storage. To deal with this problem, various data imputation methods have been proposed. Among these methods, variants of recurrent neural networks (RNNs) have attracted a lot of attention as they provide a natural representation of temporal dynamics. Though such methods have been shown to achieve higher imputation accuracy, they mostly focus on the modeling of temporal dynamics from a short-term perspective while ignoring the long-term features such as trend and seasonal information embedded in time series data. In this paper, we investigate different paradigms of time series imputation methods to capture temporal dynamics from both long-term and short-term perspectives. In particular, we propose a new learning paradigm called MuSDRI which combines the time series decomposition method Seasonal Trend decomposition using Loess (STL) with RNNs. Specifically, STL is used to decompose long sequences of time series into trend, seasonal, and remainder components such that long-term dynamics can be revealed; RNNs are used to capture local patterns from the remainder component extracted by STL such that short-term dynamics can be refined. Moreover, MuSDRI is able to handle time series with multiple seasonal patterns by learning their contributions with respect to the imputation task. We evaluate the proposed imputation methods on three real-world datasets. Experimental results show that the imputation methods considering both the long-term and short-term dynamics achieve higher imputation accuracy, and among them MuSDRI in general achieves the best performance.