Modern Cyber-physical Systems Research Articles

A Weighted Deep Representation Learning Model for Imbalanced Fault Diagnosis in Cyber-Physical Systems.

Predictive maintenance plays an important role in modern Cyber-Physical Systems (CPSs) and data-driven methods have been a worthwhile direction for Prognostics Health Management (PHM). However, two main challenges have significant influences on the traditional fault diagnostic models: one is that extracting hand-crafted features from multi-dimensional sensors with internal dependencies depends too much on expertise knowledge; the other is that imbalance pervasively exists among faulty and normal samples. As deep learning models have proved to be good methods for automatic feature extraction, the objective of this paper is to study an optimized deep learning model for imbalanced fault diagnosis for CPSs. Thus, this paper proposes a weighted Long Recurrent Convolutional LSTM model with sampling policy (wLRCL-D) to deal with these challenges. The model consists of 2-layer CNNs, 2-layer inner LSTMs and 2-Layer outer LSTMs, with under-sampling policy and weighted cost-sensitive loss function. Experiments are conducted on PHM 2015 challenge datasets, and the results show that wLRCL-D outperforms other baseline methods.

K-PdM: KPI-Oriented Machinery Deterioration Estimation Framework for Predictive Maintenance Using Cluster-Based Hidden Markov Model

Explosive increase of industrial data collected from sensors has brought increasing attractions to the data-driven predictive maintenance for industrial machines in cyber-physical systems (CPSs). Since machinery faults are always caused by performance deterioration of components, learning the deteriorating mode from observed sensor data facilitates the prognostics of impeding faults and predicting the remaining useful life (RUL). In modern CPSs, several key performance indicators (KPIs) are monitored to detect the corresponding fine-grained deteriorating modes of industrial machines. However, the overall deterioration estimation and RUL prediction based on these KPIs with various patterns have been a great challenge, especially without labels of deteriorating index or uninterpretable of root causes. In this paper, we proposed K-PdM , a cluster-based hidden Markov model for the machinery deterioration estimation and RUL prediction based on multiple KPIs. The method uncovers the fine-grained deteriorating modes of machines through each unlabeled KPI data and learns a mapping between each deteriorating KPI index and RULs. Accordingly, an overall deterioration estimation and RUL prediction of machine are able to be achieved based on the combination of each KPI’s deterioration estimation. Moreover, a set of interpretable semantic rules are setup to analyze the root cause of performance deterioration among KPIs. An experimental application is proposed to demonstrate its applicability based on the PHM08 data sets. The obtained results show their effectiveness to predict the RULs of machines.

Interdisciplinary Data Driven Production Process Analysis for the Internet of Production

Recent developments in the industrial field are strongly influenced by requirements of the fourth industrial revolution (I4.0) for modern Cyber-Physical Production Systems (CPPS) and the coherent phenomenon of industrial big data (IBD). I4.0 is characterized by a growing amount of interdisciplinary work and cross-domain exchange of methods and knowledge. Similar to the development of the Internet of Things (IoT) for the consumer market, the emergence of an Internet of Production (IoP) in the industrial field is imminent. The future vision for an IoP is based on aggregated, multi-perspective and persistent data sets that can be seamlessly and semantically integrated to allow diagnosis and prediction in domain-specific real-time. In this paper, we demonstrate an exemplary scenario of collaborative cross-domain work, in which domain-experts from largely different fields of expertise, i.e. heavy plate rolling (HPR), injection molding (IM) and machine learning (ML), generate insights through data driven process analysis in two use cases. Specifically, in the HPR use case, reinforcement-learning was utilized to support the planning phase of the process aiming to reduce manual work load and to ultimately generate process plans that serve as a foundation for a simulation to calculate process results. On the contrary, in the IM use case, supervised-learning was utilized to learn a complex and computationally demanding finite element simulation model in order to predict process results for unknown process configurations, which can be used to optimize the process planning phase. While both use cases had the overall goal to utilize ML to gain new insights about the respective process, the actual ML application was utilized with reversed purpose. Particularly, in the HPR use case, ML was used to learn the process planning in order to calculate process results while in the IM use case, ML was used to predict process results in order to improve the process planning. We facilitate the communication between physically separated domain experts and the exchange of gained insights in the respective use cases by a framework that addresses the specific needs of cross-domain collaboration. We show that the insights gained from two largely different use cases are valuable to the domain experts of the other respective use case, facilitating cross-domain data driven production process analysis for future IoP scenarios.

Автоматизация в литейном производстве: современные информационные и киберфизические системы

Автоматизация в литейном производстве: современные информационные и киберфизические системы

Accurate Sample Time Reconstruction of Inertial FIFO Data.

In the context of modern cyber-physical systems, the accuracy of underlying sensor data plays an increasingly important role in sensor data fusion and feature extraction. The raw events of multiple sensors have to be aligned in time to enable high quality sensor fusion results. However, the growing number of simultaneously connected sensor devices make the energy saving data acquisition and processing more and more difficult. Hence, most of the modern sensors offer a first-in-first-out (FIFO) interface to store multiple data samples and to relax timing constraints, when handling multiple sensor devices. However, using the FIFO interface increases the negative influence of individual clock drifts—introduced by fabrication inaccuracies, temperature changes and wear-out effects—onto the sampling data reconstruction. Furthermore, additional timing offset errors due to communication and software latencies increases with a growing number of sensor devices. In this article, we present an approach for an accurate sample time reconstruction independent of the actual clock drift with the help of an internal sensor timer. Such timers are already available in modern sensors, manufactured in micro-electromechanical systems (MEMS) technology. The presented approach focuses on calculating accurate time stamps using the sensor FIFO interface in a forward-only processing manner as a robust and energy saving solution. The proposed algorithm is able to lower the overall standard deviation of reconstructed sampling periods below 40 s, while run-time savings of up to 42% are achieved, compared to single sample acquisition.

Security of Cyber-Physical Systems in the Presence of Transient Sensor Faults

This article is concerned with the security of modern Cyber-Physical Systems in the presence of transient sensor faults. We consider a system with multiple sensors measuring the same physical variable, where each sensor provides an interval with all possible values of the true state. We note that some sensors might output faulty readings and others may be controlled by a malicious attacker. Differing from previous works, in this article, we aim to distinguish between faults and attacks and develop an attack detection algorithm for the latter only. To do this, we note that there are two kinds of faults—transient and permanent; the former are benign and short-lived, whereas the latter may have dangerous consequences on system performance. We argue that sensors have an underlying transient fault model that quantifies the amount of time in which transient faults can occur. In addition, we provide a framework for developing such a model if it is not provided by manufacturers. Attacks can manifest as either transient or permanent faults depending on the attacker’s goal. We provide different techniques for handling each kind. For the former, we analyze the worst-case performance of sensor fusion over time given each sensor’s transient fault model and develop a filtered fusion interval that is guaranteed to contain the true value and is bounded in size. To deal with attacks that do not comply with sensors’ transient fault models, we propose a sound attack detection algorithm based on pairwise inconsistencies between sensor measurements. Finally, we provide a real-data case study on an unmanned ground vehicle to evaluate the various aspects of this article.

User-Guided Visual Analysis of Cyber-Physical Production Systems

Modern cyber-physical production systems (CPPS) connect different elements like machine tools and workpieces. The constituent elements are often equipped with high-performance sensors as well as information and communication technology, enabling them to interact with each other. This leads to an increasing amount and complexity of data that requires better analysis tools to support system refinement and revision performed by an expert. This paper presents a user-guided visual analysis approach that can answer relevant questions concerning the behavior of cyber-physical systems. The approach generates visualizations of aggregated views that capture an entire production system as well as specific characteristics of individual data features. To show the applicability of the presented methodologies, an exemplary production system is simulated and analyzed.

A wireless communication system for energy and environmental monitoring

Abstract. Custom-fit communication systems are key elements in modern cyber-physical sensor systems. Therefore a wireless communication system (WCS) for sensor/actuator communication has been developed to facilitate energy and environmental monitoring on the shop floor of industrial production sites. Initially, the distinct demands and requirements are described. As the WCS has been designed for new installations as well as for retrofitting already installed facilities, the WCS has to be able to coexist with other wireless communication systems already allocated in the same frequency band. The WCS handles measurement data from both, energy-autarkic sensors and fast line-powered sensors. Mobile users in the field equipped with mobile devices are served by the system, too. A modular hardware concept has been chosen for easy system modification or for the integration of new wireless standards. Finally, measured results for the coexistence capability are presented.

Context-aware wireless mobile autonomic computing and communications: research trends and emerging applications

Growing attention has recently been devoted to context-aware computing and communication systems, in particular concerning their evolution toward the new paradigm of context-aware autonomic computing and communications. Indeed, context awareness and autonomicity appear to be the indispensable glue technologies to accomplish efficient integration of modern software-intensive cyber-physical systems, operating in open and non-deterministic environments, and to master their complexity. This article provides an update on the latest developments in this field. Proposed methodologies and techniques are discussed and framed in the overall research problem. Moreover, future trends and research challenges are outlined by relating the basic concepts of context-aware autonomic computing and communications to the emerging paradigm of the self-aware Internet of Things, applications of which appear to be very promising examples of next-generation computing and communication systems.

Towards Knowledge Capturing and Innovative Human-system Interface in an Open-source Factory Modelling and Simulation Environment

The capital value of knowledge is fundamental in modern cyber-physical systems. The user is provided with information from multiple and sometimes conflicting sources. This large amount of information must then be analyzed and interpreted by the human brain. Information must be used in a profitable way, and transformed into useable knowledge, which can then be effectively used in decision making activities. It is therefore critical to provide the required information and knowledge in order to support digital factory activities such as modelling and simulation of manufacturing systems. The advent and implementation of Industry 4.0 will make it a requirement for virtual planning and real systems within smart factories, to interact and share large quantities of information with each other. In this respect, one of the major challenges is making sure that the right people have the right information, at the right time to make the right decisions. The aim of this research is to provide a multi-level just-in-time simulation tool to support decision making in digital factory planning. The tool makes use of a suitable means for the capturing and representation of manufacturing system knowledge in several types of industrial sectors and types of companies (large, SME). Furthermore, to support the modelling activity, the human-system interface of the simulation tool makes use of just-in-time information retrieval (JITIR). The aim of JITIR is to proactively yet non-intrusively provide the required information at the right time based on the users’ context during the modeling and simulation activity.

A survey of intrusion detection techniques for cyber-physical systems

Pervasive healthcare systems, smart grids, and unmanned aircraft systems are examples of Cyber-Physical Systems (CPSs) that have become highly integrated in the modern world. As this integration deepens, the importance of securing these systems increases. In order to identify gaps and propose research directions in CPS intrusion detection research, we survey the literature of this area. Our approach is to classify modern CPS Intrusion Detection System (IDS) techniques based on two design dimensions: detection technique and audit material. We summarize advantages and drawbacks of each dimension’s options. We also summarize the most and least studied CPS IDS techniques in the literature and provide insight on the effectiveness of IDS techniques as they apply to CPSs. Finally, we identify gaps in CPS IDS research and suggest future research areas.

High performance dynamic voltage/frequency scaling algorithm for real-time dynamic load management

Modern cyber-physical systems assume a complex and dynamic interaction between the real world and the computing system in real-time. In this context, changes in the physical environment trigger changes in the computational load to execute. On the other hand, task migration services offered by networked control systems require also management of dynamic real-time computing load in nodes. In such systems it would be difficult, if not impossible, to analyse off-line all the possible combinations of processor loads. For this reason, it is worthwhile attempting to define new flexible architectures that enable computing systems to adapt to potential changes in the environment.We assume a system composed by three main components: the first one is responsible of the management of the requests arisen when new tasks require to be executed. This management component asks to the second component about the resources available to accept the new tasks. The second component performs a feasibility analysis to determine if the new tasks can be accepted coping with its real-time constraints. A new processor speed is also computed. A third component monitors the execution of tasks applying a fixed priority scheduling policy and additionally controlling the frequency of the processor.This paper focus on the second component providing a “correct” (a task never is accepted if it is not schedulable) and “near-exact” (a task is rarely rejected if it is schedulable) algorithm that can be applicable in practice because its low/medium and predictable computational cost. The algorithm analyses task admission in terms of processor frequency scaling. The paper presents the details of a novel algorithm to analyse tasks admission and processor frequency assignment. Additionally, we perform several simulations to evaluate the comparative performance of the proposed approach. This evaluation is made in terms of energy consumption, task rejection ratios, and real computing costs. The results of simulations show that from the cost, execution predictability, and task acceptance points of view, the proposed algorithm mostly outperforms other constant voltage scaling algorithms.

Challenges in transformation of existing real-time embedded systems to cyber-physical systems

Most research directions in the context of cyber-physical systems have focused on developing computing foundations for designing, analyzing, and reasoning about systems being constructed from scratch. However, an equally important issue is to develop sound methods for transforming existing real-time embedded systems to ones that conform with standards of modern cyber-physical systems. This essay proposes several research directions for achieving such transformation techniques.