Search

This paper tackles the task of singulating and grasping paper-like deformable objects. We refer to such tasks as paper-flipping. In contrast to manipulating deformable objects that lack compression strength (such as shirts and ropes), minor variations in the physical properties of the paper-like deformable objects significantly impact the results, making manipulation highly challenging. Here, we present Flipbot, a novel solution for flipping paper-like deformable objects. Flipbot allows the robot to capture object physical properties by integrating exteroceptive and proprioceptive perceptions that are indispensable for manipulating deformable objects. Furthermore, by incorporating a proposed coarse-to-fine exploration process, the system is capable of learning the optimal control parameters for effective paper-flipping through proprioceptive and exteroceptive inputs. We deploy our method on a real-world robot with a soft gripper and learn in a self-supervised manner. The resulting policy demonstrates the effectiveness of Flipbot on paper-flipping tasks with various settings beyond the reach of prior studies, including but not limited to flipping pages throughout a book and emptying paper sheets in a box. The code is available here: https://robotll.github.io/Flipbot/. © 2023 IEEE.

Pencil (graphite) on paper (PoP) electronics has gained significant attention due to its low cost and versatility in developing customizable disposable electronics. Despite the convenience of the PoP methodology in device fabrication its application in developing electronics is limited due to the absence of a bandgap in graphite. The present work is the first demonstration of the direct writing of 2D MoS2 trace on paper which was utilized as a pressure and strain sensor. Detailed characterization studies revealed the formation of rhombohedral and hexagonal phased MoS2 micro-flowers. XPS studies revealed the formation of mixed 1T (metallic) and 2H (semiconducting) phases of MoS2. With the increase in the number of traces of 2D MoS2 on paper, the resistance of the channel decreased owing to the increase in conducting channels. The as-fabricated pressure sensor displayed a sensitivity of 2.21 kPa-1 and strain sensor displayed a GF of 13. Also, the sensor displayed an excellent durability for 1600 cycles. The superior response of the sensor can be attributed to the synergistic properties of micro fibrous paper substrate and piezo-dependent electron transfer of 2D-MoS2. The presented technique enables 'make it yourself' multifunctional device that have numerous potential applications in the field of consumer electronics and personal healthcare monitoring.

Pencil (graphite) on paper (PoP) electronics has gained significant attention due to its low cost and versatility in developing customizable disposable electronics. Despite the convenience of the PoP methodology in device fabrication its application in developing electronics is limited due to the absence of a bandgap in graphite. The present work is the first demonstration of the direct writing of 2D MoS2 trace on paper which was utilized as a pressure and strain sensor. Detailed characterization studies revealed the formation of rhombohedral and hexagonal phased MoS2 micro-flowers. XPS studies revealed the formation of mixed 1T (metallic) and 2H (semiconducting) phases of MoS2. With the increase in the number of traces of 2D MoS2 on paper, the resistance of the channel decreased owing to the increase in conducting channels. The as-fabricated pressure sensor displayed a sensitivity of 2.21 kPa-1 and strain sensor displayed a GF of 13. Also, the sensor displayed an excellent durability for 1600 cycles. The superior response of the sensor can be attributed to the synergistic properties of micro fibrous paper substrate and piezo-dependent electron transfer of 2D-MoS2. The presented technique enables 'make it yourself' multifunctional device that have numerous potential applications in the field of consumer electronics and personal healthcare monitoring.

Data Analytic (DA) is currently a very important approach to obtain the best possible benefit of the data, but its degree of difficulty compared to other computational processes limits its implementation in organizations. In addition, it remains a term highly confused with respect to other areas of data science, such as data mining, and Big Data. There are few methodologies for its implementation, in addition, its complexity makes it non-participatory, with a certain degree of resistance on the part of the users. The knowledge and skills of users must be exploited when analyzing and producing the business intelligence necessary for the Data Analytic execution. In this position paper, we explore the possibilities of contribution through a Methodology for Data Analytic, which is extended with the incorporation of features extracted from a user-centered design (UCD) approach inspired by the ISEA methodology. This existing DA methodology, called MIDANO, will be extended using gamification techniques, to facilitate the applicability and understanding, in order to guarantee a participation of organizational actors, resulting in a Participatory Data Analytic Methodology. © 2020 IEEE.

Pencil (graphite) on paper (PoP) electronics has gained significant attention due to its low cost and versatility in developing customizable disposable electronics. Despite the convenience of the PoP methodology in device fabrication its application in developing electronics is limited due to the absence of a bandgap in graphite. The present work is the first demonstration of the direct writing of 2D MoS2 trace on paper which was utilized as a pressure and strain sensor. Detailed characterization studies revealed the formation of rhombohedral and hexagonal phased MoS2 micro-flowers. XPS studies revealed the formation of mixed 1T (metallic) and 2H (semiconducting) phases of MoS2. With the increase in the number of traces of 2D MoS2 on paper, the resistance of the channel decreased owing to the increase in conducting channels. The as-fabricated pressure sensor displayed a sensitivity of 2.21 kPa-1 and strain sensor displayed a GF of 13. Also, the sensor displayed an excellent durability for 1600 cycles. The superior response of the sensor can be attributed to the synergistic properties of micro fibrous paper substrate and piezo-dependent electron transfer of 2D-MoS2. The presented technique enables 'make it yourself' multifunctional device that have numerous potential applications in the field of consumer electronics and personal healthcare monitoring.

Currently, we are witnessing an exponential growth in the amount of data being generated and captured at multiple locations. This trend will continue over the next years. Hence, we have envisioned a scenario in which many objects will be referencing to or generating location information. Thus, the need for appropriately managing geospatial data is evident. In this paper, we present our vision for an integral Geo Linked Data platform; pointing out the current limitations and challenges in the GeoRDFization, Storage, Query Federation, and Visualization of data with an inherent spatial context.

The paper discusses a framework for managing the ICT-enabled planning of the service production process in a Public Administration (PA). PAs deliver a vast amount of services to citizens, each characterized by a production cost and a customer value. The production cost is the effort needed for the service production, it is influenced by several factors, e.g. By the degree of reuse in the service production process among services that are related by similarities. The customer value of services is usually measured as the trade off between benefits the service provides to customers and sacrifices customers have to deal with. In this paper we deal with a typical problem a PA must face in planning the service production process, namely to decide whether to privilege for a given budget efficiency (maximization of the number of services produced within the budget) or else efficacy (maximization of the overall customer value of services produced within the budget). The framework proposed in the paper is made up of (i) a model for the representation of a repository of services, (ii) a model for the definition of a service portfolio, and (iii) different approaches to be used in decision making for the achievement of efficiency vs. efficacy service production objectives. Two real life case studies are discussed to provide evidence of the impacts of the proposed framework.

A new dynamic simulator for pulp and paper applications has been developed. The simulator is useful for process optimization and operator training, especially when constructing a new paper machine, or rebuilding an existing one, or introducing new employees. The main advantage of the simulator is that the same control system (Distributed Control System source code) and the same displays in the Operator stations are used in the simulator as in the real plant. In this way process operators get very realistic training of various events. Optimization of an existing process line can be done by testing different ways of operation, and then selecting the best way of operation. A suitable rebuild of an existing mill could be found in a similar manner, i.e. by simulating different rebuild possibilities. A DCS functionality checkout can be done before startup of a real mill, since identical DCS code is used in the simulator as in the real mill., Export Date: 7 October 2020; Conference Paper

Dielectric response measurements are an important maintenance and diagnostic tool for large oil filled power transformers. Regardless of if the measurements being performed in the time domain or in the frequency domain, the developed tools rely on the dielectric response being linear with excitation voltage magnitude. This paper details work done in an initial investigation into the measurement of the non-linear dielectric response in oil-paper insulated power transformers using frequency domain measurements. Measurements were performed on two transformers, with non-linear response currents measured at the higher excitation voltage levels. It was found that non-linearity in the dielectric response will result in the generation of harmonics currents particularly 3rd and 5th harmonics. There will also be effects on the fundamental current magnitude due to the higher order dielectric susceptibilities. The higher order susceptibilities will thus be evident in the measured fundamental complex admittance of the dielectric.

Dielectric response measurements are an important maintenance and diagnostic tool for large oil filled power transformers. There are numerous operational disadvantages to performing dielectric response measurements offline, such as the need to disconnect the transformer from the grid, the time taken for the tests and the risk posed to the network. There is therefore a need to develop techniques to perform dielectric response measurements online. This paper presents the results from experimental work conducted in developing a method for performing Frequency Domain Spectroscopy (FDS) measurements on an energized three phase transformer in the laboratory. A chirp excitation voltage signal is used in this work to reduce the measurement time. Extraction of the magnitude and phase information, from the chirp signal, is accomplished using a chirp-matched fitter. Signal injection onto and out of the transformer under test is achieved via neutral connections or “virtual” neutral connections in the set-up. The experimental work establishes the feasibility of using the FDS test on an energized transformer.

Energy audit is a primary step towards a more rational use of primary energy sources in the energy-intensive process such as the production of packaging paper. By the adoption of the Law on efficient use of energy in 2013, energy audits have become mandatory for all industrial companies in Serbia. This paper presents the methodology and results of the one-day energy audit which was conducted in the Paper Factory - Belgrade (FHB). The measures to increase energy efficiency of the production process, together with the quantification of the savings in the primary energy consumption, are proposed based on the results of the study.

Frequency Domain Spectroscopy (FDS) is used to assess the insulation condition of oil-paper power transformers. Dissipation factor is one of the conventional indicators to analyze insulation ageing status. In this paper, the imaginary admittance of the transformers insulation, after removal of the geometric capacitance, is proposed as an alternative indicator to assist in the interpretation of ageing status. Ageing effects on the imaginary admittance are investigated both through simulation results and experimental results.

Vehicle manoeuvre analysis is vital for road safety as it helps understand driver behaviour, traffic flow, and road conditions. However, classifying data from in-vehicle acquisition systems or simulators for manoeuvre recognition is complex, requiring data fusion and machine learning (ML) algorithms. This paper proposes a hybrid approach that combines multivariate multiscale entropy (MMSE) and one-dimensional convolutional neural networks (1D-CNNs). MMSE is utilised for early feature extraction and data fusion, and the extracted features are classified using 1D-CNNs, achieving an impressive 87% test accuracy in multiclass classification. This paper provides insights into improving vehicle manoeuvre classification using advanced ML techniques and data fusion methods to handle complex data sets effectively. Ultimately, this approach can enhance the understanding of driver behaviour, inform policy decisions, and develop more effective strategies to enhance road safety.

On-device transfer learning suggests fine-tuning pretrained neural networks on new input data directly on edge devices. The memory limitation of edge devices necessitates using memory-efficient fine-tuning methods. Fine-tuning involves two primary phases: the forward-pass phase and the backwardpass phase. The forward-pass phase generates output activations, and the backward-pass phase computes gradients and updates the parameters accordingly. Although the forward-pass phase demands a temporary memory to store a layer’s input and output activations, the backward-pass phase may require storing the output activations from all layers to compute gradients. This fact introduces the memory cost of the backward-pass phase as the main contributor to the huge training memory demands of deep neural networks (DNNs), which has been the focus of many studies. However, little attention has been made to how the temporary activation memory involved in the forward-pass phase may also act as the memory bottleneck, which is the main focus of this paper. This paper aims to mitigate this memory bottleneck by pruning unimportant channels from layers that require significant temporary activation memory. Experimental results demonstrate how the proposed method effectively reduces peak activation memory and total memory costs of MobileNetV2 by 65% and 59%, respectively, at the cost of 3% accuracy drop.

Artificial intelligence (AI) has recently been used as a tool for various visual storytelling, but text-to-image models are a stochastic machine learning process that requires human intervention to assist creation better. As we all know, pre-visualization is an important stage in film production, involving subjective choices by the creators. This paper investigates how AI can assist filmmakers during the pre-production stage by generating mood boards from text. We propose a novel preproduction pipeline and guidelines that leverage text-to-image models to create visual previews of film projects. We also conduct a case study to validate and evaluate our approach's effectiveness. Our case study suggests that following the guidelines we have developed can assist filmmakers in generating mood boards that effectively convey the desired atmosphere of their projects and potentially contribute to enhancing the creative process. Our paper aims to contribute to the field of AI art and the film industry. © 2023 IEEE.

The metabolic rate has attracted widespread interest because it is the most important parameter in thermal comfort assessment. Nevertheless, it is often determined inaccurately. This paper establishes a novel metabolic rate model, which we believe is the first of its kind to achieve straightforward metabolism evaluation via wearable technology. The theory starts from the root of the metabolic rate, combining a new heat-loss calculation approach, which, together with heart rate, skin resistance, and body muscle rate, finally results in a linear model. In addition, a novel metabolic rate sensor is designed and invented based on the proposed theory to achieve a miniature, real-time, user-friendly, comfortable measuring device. Three environmental temperature conditions (22, 25 & 28 ) under four physical activity stages are involved in this paper's experiments. The remarkable results demonstrate that a linear relationship is established and verified between the metabolic rate (0 ~ 3.5 Mets) tested by the instrument and the newly proposed model. A high coefficient of determination exists at all temperature conditions, showing at least 95% accuracy and less than 2% uncertainty. The most notable finding is that the proposed model is deterministic for fixed objects, regardless of the experimental temperature conditions. This research focuses on wearable metabolic rate sensor measurement. Further still, the innovative approach and the designed sensor can play a constructive role in thermal comfort evaluation. © 2023 University of Split, FESB.

In an erasure-coded distributed storage system (DSS), requesting a file requires downloading information from multiple storage nodes, called servers, which leads to cross-server network traffic. The cross-server transmission cost can be reduced if these servers are equipped with extra memory to cache some information about the files. This paper considers the so-called cache-aided DSS (CADSS), where each server is connected to several caching proxies through a shared link, and studies the transmission cost incurred by file requests. For simplicity, we focus on a CADSS in which the servers are connected via a one-hop link, and each server is connected to the same number of caching proxies. When the caching proxies receive file requests, a server first downloads some symbols from the other servers, called the helper servers, and then broadcasts some symbols to the caching proxies. For a single server, the maximum number of symbols downloaded from the helper nodes (respectively broadcast to its caching proxies) normalized by the file size is called the cross-server (respectively local) reads. This paper first optimizes the cross-server and local reads separately. Whether from the perspective of optimizing cross-server or local reads, a CADSS can be interpreted as an equivalent single-server caching system but with different system parameters. This paper analyzes the optimal tradeoff between the cross-server and local reads. It is shown that the optimal cross-server and local reads can be achieved simultaneously for some parameters, while a tradeoff exists for some other parameters. We characterize the two extreme points of the optimal tradeoff curve and derive the optimal tradeoff for some specific parameters. © 2023 IEEE.

This paper aims to efficiently save energy consumption of multiple-input-single-output (MISO) networks. To this end, an optimization problem is formulated to minimize the transmit power of the base station (BS) while ensuring the minimal required information rate of each user. Reinforcement learning (RL) is a significant solution to tackle this non-convex problems. Nevertheless, designing efficient reward functions that accurately describe the involved optimization problems for RL is a challenging problem, which may result in low learning efficiency and even inaccurate learning. To cope with such an issue, this paper proposes an inverse reinforcement learning (IRL)-based approach to minimize the transmit power of MISO networks, in which the generative adversarial imitation learning (GAIL) is employed as the IRL framework and the reward function is represented by an artificial neural network (ANN) model, whose parameters are learned automatically in terms of an expert policy based on the semidefinite relaxation (SDR)-based approach. Since SDR is able to achieve the near optimal solution, the learned reward function is more accurate than traditional artificially designed one. Simulations show that, the proposed IRL-based approach reaches similar energy consumption to the SDR-based approach, and reduces about 50% transmit power compared to traditional RL-based one.

Understanding the needs and constraints of systems in general and a system-of-systems in particular can be challenging, yet crucial. Relying only on upfront activities will not be sufficient. Important information can be gathered around the performance and behavior of the system as well as stakeholder needs in operation. A digital twin is a way to model and understand the operation of a system. To understand the challenges and enablers related to digital twins in a system-of-systems context, we performed a literature study. In total, only 10 papers were identified that explicitly address this topic, all from the last five years, indicating that this is an active field of research. The papers revealed that definitions and terminology are unclear and that similar challenges as for systems-of-systems also exist for systems-of-digital twins. The complexity and dynamic nature of systems-of-systems motivate further study of digital twins to understand needs and constraints. However, key challenges such as concepts and principles of digital twins for systems-of-systems, cost and benefits, and evolution needs to be better understood.

Technical advances as well as continuously evolving business demands are reshaping the need for flexible connectivity in industrial control systems. A way to achieve such needs is by using a service-oriented approach, where a connectivity service middleware provides controller as well as protocol-specific interfaces. The Message Queuing Telemetry Transport (MQTT) protocol is a widely used protocol for device-to-device communication in the Internet of Things (IoT). However it is not commonly integrated in industrial control systems. To address this gap, this paper describes the development and implementation of a prototype of a connectivity service middleware for MQTT within an industrial private control network. The prototype implementation is done in the context of an industrial controller, and used in a simulated modular automation system. Furthermore, various deployment scenarios are evaluated with respect to response time and scalability of the connectivity service., Conference paper; Export Date: 26 July 2023; Cited By: 0; Conference name: 12th Mediterranean Conference on Embedded Computing, MECO 2023; Conference date: 6 June 2023 through 10 June 2023; Conference code: 190017

Cybersecurity is an important concern in systems-of-systems (SoS), where the effects of cyber incidents, whether deliberate attacks or unintentional mistakes, can propagate from an individual constituent system (CS) throughout the entire SoS. Unfortunately, the security of an SoS cannot be guaranteed by separately addressing the security of each CS. Security must also be addressed at the SoS level. This paper reviews some of the most prominent cybersecurity risks within the SoS research field and combines this with the cyber and information security economics perspective. This sets the scene for a structured assessment of how various cyber risks can be addressed in different SoS architectures. More precisely, the paper discusses the effectiveness and appropriateness of five cybersecurity policy options in each of the four assessed SoS archetypes and concludes that cybersecurity risks should be addressed using both traditional design-focused and more novel policy-oriented tools.

Multi-core computing platforms have recently been introduced in the vehicle domain to enable novel data-intensive and computation-demanding features in advanced real-time embedded system applications (e.g., a modern car) that could not be supported by traditional single-core computing platforms. Although the existing model-and component-based software development techniques can efficiently deal with the challenge of managing the software complexity in these systems, the core challenge of ensuring their timing predictability comes at the cost of performance loss and underutilization of the systems' resources. This paper identifies a sheer standoff between ensuring timing predictability and performance when the traditional model-and component-based software development techniques are used to develop software architectures of these systems. Note that ensuring timing predictability of a real-time system during its development is as important as ensuring its correct logical functionality. This paper takes the position that the traditional techniques fail to resolve this standoff mainly because they consider the following three viewpoints either in isolation or in pairs, but not altogether: (i) viewpoint of the software architecture modeling, (ii) viewpoint of the execution environment modeling, and (iii) viewpoint of the timing predictability verification. To mitigate this standoff, the paper discusses several approaches and points out opportunities for further research in developing a novel model-and component-based software development framework that jointly considers the three viewpoints during the development of these systems.

Employing an agile development methodology, particularly Scrum, in a distributed student project setting is challenging for both teachers and involved students. Allowing distributed student teams to self-organize and assign key Scrum roles using various strategies, specifically regarding the locations of students taking on key roles, increases the complexity of such projects. In addition, the interaction of the Project Owner role with the project customer, which occurs outside the distributed student team, adds a new dimension to this problem. This paper investigates the impact of various key role assignment strategies, and their interactions, on the performance of distributed student projects. Furthermore, we investigate the intensity of collaboration within the distributed team and between key project roles, as well as their impact on project performance. We analyzed data collected on 37 distributed student projects conducted over the course of eight academic years. The results reveal that letting students assign key project roles regardless of their location in the distributed team has no significant impact on the quality of project outcomes. However, a deeper analysis uncovers that more educationally desirable assignments of those roles exist; favoring increased collaboration intensity within distributed student teams.

This paper performs a comparative evaluation of various generations of Controller Area Network (CAN), including the classical CAN, CAN Flexible Data-Rate (FD), and CAN Extra Long (XL). We utilize response-time analysis for the evaluation. In this regard, we identify that the state of the art lacks the response-time analysis for CAN XL. Hence, we discuss the worst-case transmission times calculations for CAN XL frames and incorporate them to the existing analysis for CAN to support response-time analysis of CAN XL frames. Using the extended analysis, we perform a comparative evaluation of the three generations of CAN by analyzing an automotive industrial use case. In crux, we show that using CAN FD is more advantageous than the classical CAN and CAN XL when using frames with payloads of up to 8 bytes, despite the fact that CAN XL supports higher bit rates. For frames with 12-64 bytes payloads, CAN FD performs better than CAN XL when running at the same bit rate, but CAN XL performs better when running at a higher bit rate. Additionally, we discovered that CAN XL performs better than the classical CAN and CAN FD when the frame payload is over 64 bytes, even if it runs at the same or higher bit rates than CAN FD.

The traditional approach of categorizing test results as 'Pass' or 'Fail' based on fixed thresholds can be labor-intensive and lead to dropping test data. This paper presents a framework to enhance the semi-automated software testing process by detecting deviations in executed data and alerting when anomalous inputs fall outside data-driven thresholds. In detail, the proposed solution utilizes classification with convolutional neural networks and prediction modeling using linear regression, Ridge regression, Lasso regression, and XGBoost. The study also explores transfer learning in a highly correlated use case. Empirical evaluation at a leading Telecom company validates the effectiveness of the approach, showcasing its potential to improve testing efficiency and accuracy. Despite its significance, limitations include the need for further research in different domains and industries to generalize the findings, as well as the potential biases introduced by the selected machine learning models. Overall, this study contributes to the field of semi-automated software testing and highlights the benefits of leveraging data-driven thresholds and machine learning techniques for enhanced software quality assurance processes.

Today, a plethora of programming languages exists, each better suited for a particular concern. For example, Python is suited for data analysis but not web development, whereas JavaScript is the inverse. As software complexity grows and needs to address multiple concerns, different programming languages are often used in combination, despite the burden of bridging them (e.g., using Java Native Interface). Polyglot programming emerged as a solution allowing the seamless mixing of multiple programming languages. GraalVM and PolyNote are examples of runtimes allowing polyglot programming. However, there is a striking lack of support at design time for building and analyzing polyglot code. To the best of our knowledge, there is no uniform language-agnostic way of reasoning over multiple languages to provide seamless code analysis, since each language comes with its own form of Abstract Syntax Trees (AST). In this paper, we present an approach to build a uniform yet polyglot AST over polyglot code, so that it is easier to perform global analysis. We first motivate this challenge and identify the main requirements for building a polyglot AST. We then propose a proof of concept implementation of our solutions on GraalVM's polyglot API. On top of the polyglot AST, we demonstrate the ability to implement several polyglot-specific analysis services, namely auto-completion, consistency checking, type inference, and rename refactoring. Our evaluation on three polyglot projects taken from GitHub, and involving JavaScript and Python code, shows that we can build a polyglot AST without significant overhead. We also demonstrate the usefulness of the polyglot analysis services through the provided automation, as well as their scalability.

The reliability of Artificial Neural Networks (ANNs) has emerged as a prominent research topic due to their increasing utilization in safety-critical applications. Long Short-Term Memory (LSTM) ANNs have demonstrated significant advantages in healthcare applications, primarily attributed to their robust processing of time-series data and memory-facilitated capabilities. This paper, for the first time, presents a comprehensive and fine-grain analysis of the resilience of LSTM-based ANNs in the context of gait analysis using fault injection into weights. Additionally, we improve their resilience by replacing faulty weights with zero, enabling ANNs to withstand environments that are up to 20 times harsher while experiencing up to 7 times fewer critical faults than an unprotected ANN.

Novel production paradigms aim at increasing the efficiency and flexibility of production systems. Nonetheless, traditional industrial infrastructures lack the mechanisms needed to support these new paradigms. One of the main limiting factors is the architecture, which follows the automation pyramid in which subsystems are divided in layers depending on their functionalities. This allowed to meet the timing and dependability requirements of the production subsystems, however at the cost of limiting the exchange of information required to provide increased flexibility to the system. For this reason, in this paper we propose a new industrial architecture with a single network infrastructure to connect all the devices that conform to the industrial systems. On top of that, we design an Automatic Network Configurator to support the automatic configuration of the system. To assess the feasibility of our design and evaluate its performance, we implement the first instance of the architecture capable of supporting changes in the traffic requirements during run-time, i.e., without stopping or disrupting the system's operation. Furthermore, we use the implemented instance to measure the time required for reconfigurations.

The construction industry is increasingly equipping its machinery with sophisticated embedded systems and modern connectivity. Technology advancements in connected safety-critical systems are complex, with cyber-security becoming a more critical factor. Due to interdependencies and network connectivity, attack surfaces and vulnerabilities have increased significantly. Consequently, it is imperative to perform a risk assessment and implement robust security testing methods in order to prevent cyber-attacks on machinery segments. This paper presents a method for identifying potential security threats that also affect machine functional safety, facilitated by identifying threats in the threat modeling process and analyzing safety-security synergies. By identifying such risks, attack scenarios are created to simulate cyber-attacks and create test cases for validation. This approach integrates security testing into the current testing process by using penetration testing tools and utilizing a Hardware-in-the-Loop(HIL) test setup and it is verified with a simulated Denial of Service attack over a CAN network.

In this paper, we investigate the problem of modeling time-series as a process generated through (i) switching between several independent sub-models; (ii) where each sub-model has heteroskedastic noise, and (iii) a polynomial bias, describing nonlinear dependency on system input. First, we propose a generic nonlinear and heteroskedastic statistical model for the process. Then, we design Maximum Likelihood (ML) parameters estimation method capable of handling heteroscedasticity and exploiting constraints on model structure. We investigate solving the intractable ML optimization using population-based stochastic numerical methods. We then find possible model change-points that maximize the likelihood without over-fitting measurement noise. Finally, we verify the usefulness of the proposed technique in a practically relevant case study, the execution-time of odometry estimation for a robot operating radar sensor, and evaluate the different proposed procedures using both simulations and field data.

Modern manufacturing systems collect a huge amount of data which gives an opportunity to apply various Machine Learning (ML) techniques. The focus of this paper is on the detection of anomalous behavior in industrial manufacturing systems by considering the temporal nature of the manufacturing process. Long Short-Term Memory (LSTM) networks are applied on a publicly available dataset called Modular Ice-cream factory Dataset on Anomalies in Sensors (MIDAS), which is created using a simulation of a modular manufacturing system for ice cream production. Two different problems are addressed: anomaly detection and anomaly classification. LSTM performance is analysed in terms of accuracy, execution time, and memory consumption and compared with non-time-series ML algorithms including Logistic Regression, Decision Tree, Random Forest, and Multi-Layer Perceptron. The experiments demonstrate the importance of considering the temporal nature of the manufacturing process in detecting anomalous behavior and the superiority in accuracy of LSTM over non-time-series ML algorithms. Additionally, runtime adaptation of the predictions produced by LSTM is proposed to enhance its applicability in a real system.

Society 5.0 is a visionary project aimed at creating sustainable smart societies, which relies on advanced computer science and software engineering techniques. Hence, software engineering plays a critical role in enabling Society 5.0 by providing the necessary technological advances and practices to address societal issues. This paper presents a systematic mapping study that focuses on identifying the key software engineering technologies and challenges that are critical for achieving the goals of Society 5.0. The study begins with an initial set of 1646 peer-reviewed publications, from which a final set of 29 primary studies was selected through a rigorous selection process. The selected studies were subjected to meticulous data extraction, analysis, and synthesis. The primary studies identified 13 software engineering technologies that act as enablers of Society 5.0. However, the application of these technologies also comes with technical challenges. The selected primary studies identified ten software engineering open challenges that hamper the realisation of Society 5.0. The findings of this study have important implications for software engineering researchers and practitioners involved in developing sustainable smart societies. The identification of the key technologies and challenges in Society 5.0 can help guide future research and development efforts in this field.

The production industry is looking for new solutions to improve the reliability, safety and efficiency of traditional processes. Current developments in artificial intelligence and machine learning have enabled a high level of autonomy in smart-manufacturing and production systems within Industry 4.0, thus paving the way towards fully Autonomous Cyber-Physical Production Systems (ACPPS). Although ACPPS can have many advantages, there still remains a concern regarding how much we can trust those systems, due to limited predictability, transparency, and explainability, as well as emerging vulnerabilities related to machine learning systems. In this paper, we present the findings of a study conducted on the possible risks related to the trustworthiness of ACPPS, and the consequences they have on the system and its environment.

Wireless Time-Sensitive Networks (TSN) are needed to fulfill the requirements of real-time applications in areas where wired connections are not feasible. Wireless TSN combines the real-time capabilities of TSN with the flexibility of wireless connectivity opening a path for new use cases while providing determinism to time-critical scenarios such as autonomous vehicles. Industrial automation is integrating TSN with various wireless technologies such as WiFi, 4G, and 5G. This paper presents an ongoing work which aim is to experimentally analyze the performance of TSN when integrated with 4G, 5G, and WiFi in real-world scenarios. This will help both the researchers and industries to have a clear view of the network performance regarding the end-to-end latency requirements when designing their applications and use cases.

Container-based virtualization has become a promising deployment model for industrial applications mainly due to its benefits, such as providing support for co-located applications in heterogeneous environments. However, such facilitation brings challenges, including full temporal isolation among real-time applications and support for time-critical applications. In this paper, we tackle such challenges, in particular when the applications are time-sensitive Control Applications. The literature suggests that flexible timing constraints for Control Applications are beneficial in responding to disturbances and minimizing response deviation. Therefore, we propose a mechanism to support such a runtime adaptation in container-based virtualization. To show the performance of the proposed mechanism, we implement our approach on a Linux-based hierarchical scheduling platform, and we evaluate it for a Control application.

Offline scheduling of Scheduled Traffic (ST) in Time-Sensitive Networks (TSN) without taking into account the quality of service of non-ST traffic, e.g., time-sensitive traffic such as Audio-Video Bridging (AVB) traffic, can potentially cause deadline misses for non-ST traffic. In this paper, we report our ongoing work to propose a solution that, regardless of the ST scheduling algorithm being used, can ensure meeting timing requirements for non-ST traffic. To do this, we define a frame called Guard Frame (GF) that will be scheduled together with all ST frames. We show that a proper design for the GFs will leave necessary porosity in the ST schedules to ensure that all non-ST traffic will meet their timing requirements.

The surge in the development and adoption of Electric Vehicles (EVs) globally is a trend many countries are paying close attention to. This inevitably means that a significant number of EV batteries will soon reach their End-of-Life (EoL). This looming issue reveals a notable challenge: there's currently a lack of sustainable strategies for managing Lithium-ion Batteries (LiBs) when they reach their EoL stage. The process of disassembling these battery packs is challenging due to their intricate design, involving several different materials and components integrated tightly for performance and safety. Consequently, effective disassembly and subsequent recycling procedures require highly specialized methods and equipment, and involve significant safety and health risks. Moreover, existing recycling technologies often fail to recover all valuable and potentially hazardous materials, leading to both economic and environmental loss. This paper provides an overview and analysis of possible challenges arising in the domain of automated battery disassembly and recycling of EV batteries that reached their EoL. We provide insight into the disassembly process as well as optimization of the disassembly sequence with the goal of minimizing the overall cost and environmental footprint.

Robot operating system 2 (ROS 2) is increasingly popular both in research and commercial robotic systems. ROS 2 is designed to allow real-time execution and data communication, enabling rapid prototyping and deployment of robotic systems. In order to predict and calculate execution times in ROS 2, one needs to analyze its internal scheduler, called executor. The executor has been updated in various distributions of ROS 2, which is shown to impact significantly the periodic execution invoked by the underlying operating system's timers, potentially causing unexpected latencies. To expose the mentioned impact due to executor differences, in this paper, we present an experimental evaluation of the execution behavior of ROS 2's schedulable entities, namely callbacks, among the existing versions of the executor. We visualize the differences of callback execution order via simulation, and we create design-level scenarios that impact the execution of periodically scheduled callbacks, negatively. Moreover, we show how such negative impact can be mitigated by using multi-threaded executors. Finally, we illustrate the observed behavior on a real-world centralized multi-agent robot system. Our work aims to raise awareness within the ROS 2 developer community, regarding possible problems of timer blocking, and propose a mitigation solution of the latter.

Tight end-to-end worst-case delay bounds for periodic traffic streams are essential for time sensitive networks. In this paper, we provide an algorithm to compute a tight (and accurate) end-to-end worst-case bound by considering distinct topological patterns and the manner in which streams enter and leave switches. This refined analysis uses non-preemptive, strict-priority arbitration mechanism commonly deployed in Ethernet switches. Compared to the state-of-the-art that considers all higher and equal priority interference as contributing to the worst-case bound, we present an analytical approach for computing a tighter worst-case delay bound and prove through discrete event simulations that only a certain number of equal-priority interference streams can actually affect the worst-case case. Our results enable efficient resource allocation and have implications for online re-configuration mechanisms for time-sensitive factory communication systems.

Tiny machine learning (TinyML) represents an emerging research direction that aims to realize machine learning on Internet of Things (IoT) devices. The current TinyML research seems to focus on supporting the deployment of deep learning models on microprocessors, while the models themselves are trained on high performance computers or clouds. However, in the resource/time constrained IoT contexts, it is more desirable to perform data analytics and learning tasks directly on edge devices for crucial benefits such as increased energy efficiency, reduced latency as well as lower communication cost.To address the above challenge, this paper proposes a tiny federated learning algorithm for enabling learning of Bayesian classifiers based on distributed tiny data storage, referred to as TFL-BC. In TFL-BC, Bayesian learning is executed in parallel across multiple edge devices using local (tiny) training data and subsequently the learning results from local devices are aggregated via a central node to obtain the final classification model. The results of experiments conducted on a set of benchmark datasets demonstrate that our algorithm can produce final aggregated models that outperform single tiny Bayesian classifiers and that the result of tiny federated learning (of Bayesian classifier) is independent of the number of data partitions used for generating the distributed local training data.

Power control for the device-to-device interference channel with single-antenna transceivers has been widely analyzed with both model-based methods and learning-based approaches. Although the learning-based approaches, i.e., data-driven and model-driven, offer performance improvement, the widely adopted graph neural network suffers from learning the heterophilous power distribution of the interference channel. In this paper, we propose a deep learning architecture in the family of graph transformers to circumvent the issue. Experiment results show that the proposed methods achieve the state-of-the-art performance across a wide range of untrained network configurations. Furthermore, we show there is a trade-off between model complexity and generality. © 2023 IEEE.

To overcome the scaling and performance limitations, the Directed Acyclic Graph (DAG) is utilized as the underlying storage model of blockchain systems, which enables concurrent transaction processing and confirmation. However, accompanied by high performance, DAG-based blockchains still suffer from the severe challenge of constrained storage scalability, i.e., expensive storage overhead. Based on an in-depth analysis of the data, we discover that the root cause of storage overhead stems from the considerable data redundancy in the DAG-based blockchains. In this paper, we propose GeckoDAG, a lightweight DAG-based blockchain, whose design consists of two steps. First, we abstract a storage model named Basic from the existing DAG-based blockchain systems, which offers both high performance and security. On top of Basic, we then devise GeckoDAG, which merges previous transactions into Transaction Union (TU) and reduces the data redundancy in TU, thus lowering the storage overhead. To evaluate our design, we implement a prototype of GeckoDAG and conduct various experiments on it. The experimental results demonstrate that GeckoDAG can offer storage scalability while maintaining the security and efficiency of DAG-based blockchains. © 2023 IEEE.

Integrated sensing and communications (ISAC) has been considered as a key technology for future wireless networks. Orthogonal time frequency space (OTFS) modulation is considered as an effective technology to achieve ISAC in high-mobility scenarios due to its ability to combat high delay and Doppler shifts. However, although the performance analysis of OTFS-based ISAC system has been investigated, resource allocation and joint optimization for communication and sensing are still in its fancy. In this paper, we consider resource allocation in OTFS-based ISAC systems. We first introduce the system model and formulate the resource allocation problems. Then, several related works are discussed in detail. Next, an OTFS-based multiple-user multiple-input multiple-output (MIMO) ISAC system model is investigated, where the base station is considered as an ISAC transmitter and sends the message to different vehicles while simultaneously estimating their relative range and velocity via output feedback, respectively. We then formulate the sensing and communication trade-off optimization problem considering the delay Doppler resource block (DDRB) allocation and provide some sample results. Finally, we show several future directions for resource allocation in OTFS-based ISAC systems. © 2023 IEEE.

This paper proposes a new topology of a high-temperature superconducting linear flux switching motor (HTS-LFSM). The electromagnetic characteristics of this novel HTS-LFSM, including flux density, back-EMF, thrust force, and ripple, are analyzed. Key topology parameters of the HTS-LFSM are optimized, and its performance is compared with that of the traditional LFSPM. The comparison shows that the proposed HTS-LFSM can generate higher thrust force and smaller thrust ripple. The results indicate that the proposed machine offers more performance advantages for modern traffic. © 2023 IEEE.

Digital Computing-in-Memory (DCIM) is an emerging architecture that integrates digital logic into memory for efficient AI computing. However, current DCIM designs heavily rely on manual efforts. This increases DCIM design time and limits the optimization space, making it challenging to satisfy the user specifications of diverse AI applications. This paper presents AutoDCIM, the first automated DCIM compiler. Au-toDCIM takes the user specifications as inputs and generates a DCIM macro architecture with an optimized layout. AutoDCIM's template-based generation balances handcrafted cell design and agile macro development. AutoDCIM's layout exploration loop analyzes diverse DCIM array partitioning schemes to satisfy user specifications. The auto-generated DCIM macros present competitive efficiency results in comparison with state-of-the-art silicon-verified DCIM macros. © 2023 IEEE.

This paper presents the Instruction-Level Abstraction (ILA) as a formal software/hardware interface for accelerator-rich platforms. The ILA provides a common framework for formal functional specification of processors and accelerator behavior, verifying their implementations, and reasoning about software-hardware interactions of programs with accelerators. The ILA-MCM specification extends the ILA to enable reasoning about interactions of accelerators with other compute engines through shared memory. The 3LA compilation flow for accelerators uses the ILA model to bridge the gap between fine-grain compiler intrinsics and coarse-grain accelerator operations. © 2023 IEEE.

On-chip power meters play a critical role in power management by generating timely and accurate power traces at runtime. However, both performance-counter-based and existing RTL-based on-chip power meters have difficulty in providing sufficient response time for fast power and voltage management scenarios. Additionally, they can be costly to implement for large-scale DNN accelerators with many homogeneous process elements. To address these limitations, this paper proposes PROPHET, a data-pattern-based predictive on-chip power meter targeting multiply-accumulate-based DNN accelerators. By sampling pre-defined data patterns during memory access, PROPHET can predict power consumption before it actually happens. In our experiments, PROPHET predicts power consumption dozens of clock cycles in advance, with a temporal resolution of 4 clock cycles and NMAE < 7% and area overhead < 2% for various systolic-array-based DNN accelerators. PROPHET has the potential to enable fine-grained power management and optimization for large-scale DNN accelerators, improving their energy efficiency. © 2023 IEEE.

Processing-in-memory (PIM) architectures have shown great abilities for neural network (NN) acceleration on edge devices that demand low latency under severe area constraints. Heterogeneous PIM architectures with different PIM implementation approaches such as RRAM-based PIM and SRAM-based PIM can further improve the performance. However, the automatic generation of heterogeneous PIM architectures faces the following two unresolved problems. First, existing work has not considered the design for heterogeneous PIM-based NN accelerators with multiple memory technologies. Second, for PIM with insufficient memory on edge devices, it is challenging to find the optimal runtime weight scheduling strategy in an O(L!) optimization space for the NN with L layers.In this paper, we propose PIM-HLS, an automatic hardware generation tool for heterogeneous PIM-based NN accelerators. Aiming at the problems above, we first point out that heterogeneous PIM can improve the performance under severe area constraints. Then we optimize the architectures for each NN layer by taking the advantage of different memory technologies. We also define the optimization problem of runtime weight scheduling and mapping for the first time, and propose a dynamic-programming-based weight scheduling algorithm to reduce the optimization space to O(L2). We implement PIM-HLS to automatically generate the hardware code and the instructions. Results show that we achieve an averagely 5.9× speedup with 72.8% less area compared with state-of-the-art PIM designs. © 2023 IEEE.

With the rapid development of computer science, neural network (NN) has received widespread attention. However, it is mainly used as a black box in identification and prediction, which is highly complicated. Its explainability should be mentioned when it is used in real applications. So, this paper derives the explainable control scheme to make sure its theoretical rigorousness. Then a straightforward explanation of the neural network controller using a block diagram is given to help researchers to reproduce the method. Since the detailed design of the neural network should concern the principle of permanent magnet synchronous motor (PMSM), this analyses the real PMSM system and finds the control performance characteristic, especially in spontaneous disturbances. Unlike the black-box neural network, the recurrent radial basis function neural network (RRBF) controller method is proposed, guided by overshoot. Because various PMSMs have various performances of speed changing, three key functions of RRBF are proposed to check which variant is the most suitable for reducing the overshoot. Finally, the RRBF method is implemented in STM32, and its effectiveness in the experiment is checked. © 2023 IEEE.

With nonlinear effects in the electromechanical oscillators, subharmonic oscillations can be generated, whose period is multiple that of the driving signal. Such oscillations possess multiple states with identical amplitude and evenly offset phases. Their phases can be used to encode information which was proposed early to half a century ago with the name 'Parametron.' Until recently, the oscillations with a period tripled that of the drive just realized through enhanced energy transfer between the modes with a ratio of eigenfrequencies close to 1: 3 in the micro-scale mechanical resonator. In this paper, we demonstrate a well-designed structure of a nano-scale GaAs beam resonator that can generate period-tripled subharmonic oscillations through the mode coupling at 1: 3 internal resonances using numerical simulation and theoretical analysis. The generated states can be used to encode and store information on a base-3 basis. Similar to the micro-scale counterpart, for encoding, an extra excitation pulse is required. By changing its phase, the resonator can switch among the stable states. Furthermore, multiple sets of period-tripled states can be generated via different excitation schemes or non-linear effects. This design has a smaller size, lower energy consumption, and higher operating frequency in comparison with the micro-scale resonators, which favors the practical applications of the mechanical-based many-valued logic elements. © 2023 IEEE.