A Cyber-Physical System (CPS)-enabled rehabilitation system framework for enhanced recovery rate in gait training systems is presented in this paper. Recent advancements in sensing and data analytics have paved the way for the transformation of healthcare systems from experience-based to evidence-based. To this end, this paper introduces a CPS-enabled rehabilitation system that collects, processes, and models the data from patient and rehabilitative training machines. This proposed system consists of a set of sensors to collect various physiological data as well as machine parameters. The sensors and data acquisition systems are connected to an edge computing unit that handles the data preprocessing, analytics, and results visualization. Advanced machine learning algorithms are used to analyze data from physiological data, machine parameters, and patients’ metadata to quantify each patient’s recovery progress, devise personalized treatment strategies, adjust machine parameters for optimized performance, and provide feedback regarding patient’s adherence to instructions. Moreover, the accumulation of the knowledge gathered by patients with different conditions can provide a powerful tool for better understanding the human-machine interaction and its impact on patient recovery. Such system can eventually serve as a ‘Virtual Doctor’, providing accurate feedback and personalized treatment strategies for patients.

Data-driven modeling and fault detection of multi-stage manufacturing processes remain challenging due to the increasing complexity of the manufacturing process, the lack of structural data, data multi-dimensionality, and the additional difficulty when dealing with large data sets. The implementation of add-on sensors and establishing data acquisition, transfer, storage and analysis has the potential to facilitate advanced data modeling techniques. However, besides the associated costs, dealing with high-volume multi-dimensional data sets can be a major challenge. This paper presents a novel methodology for early fault identification of multi-stage manufacturing processes using a statistical approach. The major advantage of the proposed methodology is its reliance on only the product quality measurements and basic product manufacturing records, given the presence of peer sets. This leads to a feasible fault identification solution in a sensor-less environment without investing costly data collection systems. The developed methodology transforms the end-of-process quality measurements to a process performance metric based on a density-based statistical approach and a peer-to-peer comparison of the machines at one stage of the process. This approach allows one to be more proactive and identify the problematic machines that could be affecting product quality. A case study in an actual multi-stage manufacturing process is used to demonstrate the effectiveness of the developed methodology.

The advent of Industry 4.0 technologies and in particular the Cyber-Physical Systems, Digital Twins and pervasive connected sensors is transforming many industries, among which smart scheduling is one of the most relevant. This paper contributes to the research on scheduling by proposing a framework to include equipment health predictions into the scheduling activity and embedding a field-synchronized Equipment Health Indicator module into the DT simulation. The metaheuristic approach to scheduling optimization is performed by a genetic algorithm, that is connected with the DT simulator and provides various generations of scheduling alternatives that are assessed through the simulator itself. The paper also proposes a practical Proof-of-Concept of the innovative framework, by developing an architecture to identify how the various framework modules are implemented and by applying the framework to a real application case, set in a laboratory assembly line environment.

In the rough machining stage, band saw machines are widely used to cut various raw materials into the required dimensions. The replacement of blade due to the blade degradation accounts for a large part of the total cost of band saw machine users. Therefore, mitigating the blade degradation by dynamically optimizing the cutting parameters can produce great economic benefits and is also a good exploration for smart manufacturing. To achieve this goal, the Convolutional Neural Network (CNN) model is proposed to map the complicated relationship between the cutting parameters and blade degradation. Then a simulation technique is used to search the optimal cutting parameters based on the blade degradation estimation from the model output, which would help alleviate the blade degradation. The proposed optimization method is validated on the data collected during the real manufacturing process of band saw machines. The comparative results demonstrate that the obtained optimal cutting parameters can effectively extend the service life of the blade for a band saw machine.

Bottle capping is an important manufacturing process in many production systems. Bottle capping defects, if left unattended, will cause leakage issues during packaging and transportation. To develop a bottle capping defect detection and diagnosis system with lower cost, this paper proposes a data-driven methodology. Our approach is based on the motor current signal from the automatic capping machine, and four different capping defects, such as loose cap, cocked cap, and no cap, are considered in the experiments. Based on the current signal collected from the bottle capping machine, the signal envelope is extracted and smoothed first, and the faulty signatures are discovered subsequently. By designing and selecting appropriate features to describe the discovered fault signature, bottle capping defects can be effectively detected and isolated with the help of machine learning models. The results in the case study demonstrate that all the bottle capping defects considered in this investigation can be effectively detected and isolated.

Power electronics modules such as inverters and rectifiers are crucial in industry and they are indispensable in various power conversion systems. There have been many studies on the fault diagnosis of power converters or power modules in the system but recently more attention has been paid on predicting failures. Most conventional techniques often rely on accurate physical models or high frequently sampled electric signals in simulation or experiment environments. In practice, however, the life and the degradation of power electrics devices are highly influenced by loads and operation regimes. The analysis needs considering various identical or similar devices in a networked power grid as well. Thus, it is not trivial to achieve the predictive analysis in such complex working conditions. This paper presents a systematic approach investigating the fault prediction of power converters in power conversion systems. Two data-driven methods with novel techniques, which take into account working condition variances and the data imbalance, have been developed and applied to an industry use case where only high level system heartbeat signals are available. These methods are validated to effectively predict the power converter failures.

The recent White House report on Artificial Intelligence (AI) (Lee, 2016) highlights the significance of AI and the necessity of a clear roadmap and strategic investment in this area. As AI emerges from science fiction to become the frontier of world-changing technologies, there is an urgent need for systematic development and implementation of AI to see its real impact in the next generation of industrial systems, namely Industry 4.0. Within the 5C architecture previously proposed in Lee et al. (2015), this paper provides an insight into the current state of AI technologies and the eco-system required to harness the power of AI in industrial applications.

As a critical mechanical component that converts rotary motion to linear motion with high precision, the ball screw has drawn a lot of attention in the field of Prognostics and Health Management (PHM). However, prognosis of the ball screw degradation has not been fully discussed yet in the current literature. This paper first justifies the prognosability of a ball screw via experimental studies, then proposes a systematic methodology for ball screw prognosis to implement the fault diagnosis, early diagnosis, health assessment and remaining useful life (RUL) prediction. Meanwhile, sensor-less and sensor-rich strategies are investigated and benchmarked in the experimental studies. The results demonstrate that the ball screw degradation behavior is available for prognosis and the proposed methodology can effectively help users to implement PHM analysis. Besides, the benchmark studies between sensor-less and sensor-rich strategies also achieve several practical conclusions that are valuable for real-world applications. (C) 2018 Elsevier Ltd. All rights reserved.

For implementing data analytic tools in real-world applications, researchers face major challenges such as the complexity of machines or processes, their dynamic operating regimes and the limitations on the availability, sufficiency and quality of the data measured by sensors. The limits on using sensors are often related to the costs associated with them and the inaccessibility of critical locations within machines or processes. Manufacturing processes, as a large group of applications in which data analytics can bring significant value to, are the focus of this study. As the cost of instrumenting the machines in a manufacturing process is significant, an alternative solution which relies solely on product quality measurements is greatly desirable in the manufacturing industry. In this paper, a minimal-sensing framework for machine anomaly detection in multistage manufacturing processes based on product quality measurements is introduced. This framework, which relies on product quality data along with products' manufacturing routes, allows the detection of variations in the quality of the products and is able to pinpoint the machine which is the cause of anomaly. A moving window is applied to the data, and a statistical metric is extracted by comparing the performance of a machine to its peers. This approach is expanded to work for multistage processes. The proposed method is validated using a dataset from a real-world manufacturing process and additional simulated datasets. Moreover, an alternative approach based on Bayesian Networks is provided and the performance of the two proposed methods is evaluated from an industrial implementation perspective. The results showed that the proposed similarity-based approach was able to successfully identify the root cause of the quality variations and pinpoint the machine that adversely impacted the product quality.

Preventing downtimes in machinery operation is becoming fundamental in industrial standards. The most common strategy to avoid costly production stoppages is the preventive maintenance, combining it with reactive maintenance in detected malfunctions. Preventive maintenance can reduce costs, increase uptime and help maintaining the quality of the produced goods. The fingerprint analysis concept will help in the implementation of preventive maintenance [[1]]. The asset in a good health state is monitored as a set of pre-defined operating conditions. The monitorization can be triggered during the whole asset’s life-time in the same pre-defined operating conditions. As a result, a reference value is taken which accounts for normality. This value will be compared with measurements made throughout the life. The goal is to be able to detect and determine the appearance of abnormalities. Similarly, the fixed cycle features test assesses the machine performance degradation in a fixed cycle[2]. In this case the concept is focused on machinery that works in close loops with comparable conditions. The machines parameters are measured during the working cycle, setting the baseline, and compared between them in the search of abnormalities that may point out any potential faults. Both concepts, fingerprint and fixed cycle feature test concept are applied to gearboxes. They are crucial elements in industrial machinery, conventionally monitored using accelerometers. Which have a significative cost and can be hard to install in place to provide useful information. Motor current signature analysis overcomes the inconveniences of accelerometers. This analysis technique provides a non-intrusive method, and it is based in readily available signals. Changes in the current signal are related with variations of the speed and/or external load of the electric motor. Thus, the health state of the gearbox connected to the motor can be examined through an exhaustive analysis of the input currents [[3]]. A specially designed test bench is used in which gears in different health status are tested. The measured signals are analyzed using discrete wavelet decomposition, in different decomposition levels and with different mother wavelets. Additionally, a dual-level time synchronous averaging analysis is performed on the same signal to compare the performance of the two methods. From both analyses, the relevant features of the signals are extracted, cataloged and classified using diverse methods. The results obtained allow to differentiate the different type of defects on the gears. Allowing to detect the different fault conditions and enabling the assessment of the health state of the gearbox using only the motor current signal. [1] Ferreiro, S., Konde, E., Fernández, S., & Prado, A. (2016, June). INDUSTRY 4.0: Predictive Intelligent Maintenance for Production Equipment. In European Conference of the Prognostics and Health Management Society, no (pp. 1-8). [2] Liao, L., & Lee, J. (2009). A novel method for machine performance degradation assessment based on fixed cycle features test. Journal of Sound and Vibration, 326(3-5), 894-908. [3] Arellano-Padilla, J., Sumner, M., Gerada, C., & Jing, L. (2009, September). A novel approach to gearbox condition monitoring by using drive rectifier input currents. In Power Electronics and Applications, 2009. EPE'09. 13th European Conference on (pp. 1-10). IEEE.

With the rapid advancement in wearable sensor technologies and predictive analytics, college and professional sports teams are facing new opportunities of leveraging such technologies and at the same time challenges of maintaining their competency. In professional sports today, the margin between winning and losing is narrow although its impact can be massive. Effective use of predictive analytics along with the implementation of advanced sensors can bring about a deeper understanding of players’ physical condition and performance potential throughout individual games or the entire season, and help sports medicine personnel get the most from available resources while keeping players healthy and minimizing their risks of injury. This paper presents a predictive analytics framework for analyzing and predicting soccer players’ performance data. The data consists of GPS and physiological measurements, collected from female soccer players during both practices and games using Zephyr Bioharness device. The proposed framework consists of data cleaning, filtering, visualizations and analytics modules to provide deeper insights into the data. The preprocessing modules automatically remove outliers using intelligent tools and determine first half, second half and potential overtime RISKS based on data patterns. Furthermore, comparison-based metrics have been developed to analyze the performance of players from different aspects including their activity level, fitness and consistency. For instance, Kolmogorov-Smirnov (KS) test was utilized to extract performance metrics based on players’ Heart Rate and Speed, or a Neural Network-based approach was utilized to analyze the Heart Rate recovery rate of the players and quantify their recovery rate, which is important for effective play. At the end, different visualization tools were used to combine players’

This paper focuses on analyzing motor current signature for fault diagnosis of gearboxes operating under transient speed regimes. Two different strategies are evaluated, extensively tested and compared to analyze the motor current signature in order to implement a condition monitoring system for gearboxes in industrial machinery. A specially designed test bench is used, thoroughly monitored to fully characterize the experiments, in which gears in different health status are tested. The measured signals are analyzed using discrete wavelet decomposition, in different decomposition levels using a range of mother wavelets. Moreover, a dual-level time synchronous averaging analysis is performed on the same signal to compare the performance of the two methods. From both analyses, the relevant features of the signals are extracted and cataloged using a self-organizing map, which allows for an easy detection and classification of the diverse health states of the gears. The results demonstrate the effectiveness of both methods for diagnosing gearbox faults. A slightly better performance was observed for dual-level time synchronous averaging method. Based on the obtained results, the proposed methods can used as effective and reliable condition monitoring procedures for gearbox condition monitoring using only motor current signature. (C) 2017 Elsevier Ltd. All rights reserved.

The running environment of high speed EMU (Electric Multiple Units) is complex and the electromagnetic interference is serious. Compared to electric type sensors, FBG (Fiber Bragg Grating) sensors have immunity to electromagnetic interference, less wiring and easy distributed networking features. FBG sensors are used for real-time temperature monitoring of EMU traction transformer and traction converter, and the abnormal temperature during operation cause alarm. Since the large temperature measuring points and the long monitoring time, the amount of recorded data is huge. Considering the real-time, integrity and maintainability of monitoring data, the data processing strategy for real-time storage, out of limit alarm, remote transmission and server data statistics and management is developed. ENLIGHT software is used for data acquisition, real-time storage and out of limit alarm of interrogator SM125. HUAWEI EM660 3G communication module and TeamViewer software are adopted to realize remote data transmission. In the receiving server, based on LabVIEW and Access database, the temperature history data analysis system is designed to realize the temperature data display, extreme values statistics and historical data query. The system is simple and reliable, which can be used in the long-term monitoring and data processing of FBG sensors.

Automatic disease diagnosis has been human concern for a long time. Since people are so busy and the doctors visit expenses areso expensive, a lot of different attempts have been done in the field of design of expert system for disease diagnosis. This paper describes aproject work aiming to develop a web based fuzzy expert system for human disease diagnosis. This program models the thinking pattern andhuman activity and leads to close the expert system and human action method. In this paper we have consulted with different physicians andanalyzed the diagnosis procedure and modeled them with a fuzzy expert system. This project is based on development of a web-based clinicaltool designed to improve the quality of the exchange of health information between health care professionals and patients . This system hasbeen tested on five diseases with sort throat symptom such as mononucleosis, scarlet fever, pharyngitis or tonsillitis, common cold and virusinfection and exhibited satisfactory results.

With the rapid advancement of Information and Communication Technologies (ICT) and the integration of advanced analytics into manufacturing, products and services, many industries are facing new opportunities and at the same time challenges of maintaining their competency and market needs. Such integration, which is called Cyber-physical Systems (CPS), is transforming the industry into the next level. CPS facilitates the systematic transformation of massive data into information, which makes the invisible patterns of degradations and inefficiencies visible and yields to optimal decision-making. This paper focuses on existing trends in the development of industrial big data analytics and CPS. Then it briefly discusses a systematic architecture for applying CPS in manufacturing called 5C. The 5C architecture includes necessary steps to fully integrate cyber-physical systems in the manufacturing industry. Finally, a case study for designing smart machines through the 5C CPS architecture is presented.

This paper proposes a comprehensive Prognostics and Health Management (PHM) framework for large fleets of geographically distributed assets. The objective of this research study is to optimize spare part inventory according to asset performance, ensuring efficient and consistent production and extended machine life. The concept of asset condition monitoring and performance prediction along with optimizing maintenance operation is proposed by leveraging existing fleet-level PHM and Decision Support Tools (DST). Dynamic clustering methodology is adopted to equip the prediction model with the ability to adaptive update. And the impact of performance degradation to production loss is evaluated through risk assessment to link asset performance with production.

The development of robust monitoring systems for assuring the consistency and stability of multistage manufacturing processes necessitates the use of add-on sensors and advanced data collection, storage, and analysis platforms to deal with the high-dimensional data collected from machines and products in multiple stages. In many cases, such an approach may not be feasible due to high implementation costs and the challenges of obtaining the process parameters and analyzing them effectively. This paper proposes an alternative approach for health monitoring and diagnosis of multistage manufacturing processes based on product quality measurements in a sensor-less environment. In the presented work, the available data consists of product quality parameters measured from multiple product types along with the manufacturing route associated with each product. A Gamma distribution is fit to the data for each parameter within a moving time window. Using the distribution fits, a metric is developed to represent the performance of each machine in a stage compared to its peers producing the same product. This metric is then aggregated across all the products produced by the machine to generate the final metric reflecting the overall performance of the machine. This performance metric is first calculated for the machines in the last stage. After flagging the underperforming machines in the last stage, the samples from those machines are removed from the data set and the remaining samples are used to calculate the similar metric for the prior stage. The suggested approach assumes the random distribution of products from one stage to the next to facilitate the implementation of a comparison-based approach. This approach is tested on a data set collected from a manufacturing plant. The results demonstrate the effectiveness of such approach for monitoring and diagnosis of multistage manufacturing processes when the data is not available from within the process.