Data-Driven Insights through Industrial Retrofitting: An Anonymized Dataset with Machine Learning Use Cases.

Small and medium-sized enterprises (SMEs) often encounter practical challenges and limitations when extracting valuable insights from the data of retrofitted or brownfield equipment. The existing literature fails to reflect the full reality and potential of data-driven analysis in current SME environments. In this paper, we provide an anonymized dataset obtained from two medium-sized companies leveraging a non-invasive and scalable data-collection procedure. The dataset comprises mainly power consumption machine data collected over a period of 7 months and 1 year from two medium-sized companies. Using this dataset, we demonstrate how machine learning (ML) techniques can enable SMEs to extract useful information even in the short term, even from a small variety of data types. We develop several ML models to address various tasks, such as power consumption forecasting, item classification, next machine state prediction, and item production count forecasting. By providing this anonymized dataset and showcasing its application through various ML use cases, our paper aims to provide practical insights for SMEs seeking to leverage ML techniques with their limited data resources. The findings contribute to a better understanding of how ML can be effectively utilized in extracting actionable insights from limited datasets, offering valuable implications for SMEs in practical settings.

Machine Learning for Industry 4.0: A Systematic Review Using Deep Learning-Based Topic Modelling.

Machine learning (ML) has a well-established reputation for successfully enabling automation through its scalable predictive power. Industry 4.0 encapsulates a new stage of industrial processes and value chains driven by smart connection and automation. Large-scale problems within these industrial settings are a prime example of an environment that can benefit from ML. However, a clear view of how ML currently intersects with industry 4.0 is difficult to grasp without reading an infeasible number of papers. This systematic review strives to provide such a view by gathering a collection of 45,783 relevant papers from Scopus and Web of Science and analysing it with BERTopic. We analyse the key topics to understand what industry applications receive the most attention and which ML methods are used the most. Moreover, we manually reviewed 17 white papers of consulting firms to compare the academic landscape to an industry perspective. We found that security and predictive maintenance were the most common topics, CNNs were the most used ML method and industry companies, at the moment, generally focus more on enabling successful adoption rather than building better ML models. The academic topics are meaningful and relevant but technology focused on making ML adoption easier deserves more attention.

A Systematic Review of Wi-Fi and Machine Learning Integration with Topic Modeling Techniques.

Wireless networks have drastically influenced our lifestyle, changing our workplaces and society. Among the variety of wireless technology, Wi-Fi surely plays a leading role, especially in local area networks. The spread of mobiles and tablets, and more recently, the advent of Internet of Things, have resulted in a multitude of Wi-Fi-enabled devices continuously sending data to the Internet and between each other. At the same time, Machine Learning has proven to be one of the most effective and versatile tools for the analysis of fast streaming data. This systematic review aims at studying the interaction between these technologies and how it has developed throughout their lifetimes. We used Scopus, Web of Science, and IEEE Xplore databases to retrieve paper abstracts and leveraged a topic modeling technique, namely, BERTopic, to analyze the resulting document corpus. After these steps, we inspected the obtained clusters and computed statistics to characterize and interpret the topics they refer to. Our results include both the applications of Wi-Fi sensing and the variety of Machine Learning algorithms used to tackle them. We also report how the Wi-Fi advances have affected sensing applications and the choice of the most suitable Machine Learning models.

Industrial Needs in the Fields of Artificial Intelligence, Internet of Things and Edge Computing.

Industry 4.0 corresponds to the Fourth Industrial Revolution, resulting from technological innovation and research multidisciplinary advances. Researchers aim to contribute to the digital transformation of the manufacturing ecosystem both in theory and mainly in practice by identifying the real problems that the industry faces. Researchers focus on providing practical solutions using technologies such as the Industrial Internet of Things (IoT), Artificial Intelligence (AI), and Edge Computing (EC). On the other hand, universities educate young engineers and researchers by formulating a curriculum that prepares graduates for the industrial market. This research aimed to investigate and identify the industry's current problems and needs from an educational perspective. The research methodology is based on preparing a focused questionnaire resulting from an extensive recent literature review used to interview representatives from 70 enterprises operating in 25 countries. The produced empirical data revealed (1) the kind of data and business management systems that companies have implemented to advance the digitalization of their processes, (2) the industries' main problems and what technologies (could be) implemented to address them, and (3) what are the primary industrial needs and how they can be met to facilitate their digitization. The main conclusion is that there is a need to develop a taxonomy that shall include industrial problems and their technological solutions. Moreover, the educational needs of engineers and researchers with current knowledge and advanced skills were underlined.

Role of Academics in Transferring Knowledge and Skills on Artificial Intelligence, Internet of Things and Edge Computing.

Universities play an essential role in preparing human resources for the industry of the future. By providing the proper knowledge, they can ensure that graduates will be able to adapt to the ever-changing industrial sector. However, to achieve this, the courses provided by academia must cover the current and future industrial needs by considering the trends in scientific research and emerging technologies such as Artificial Intelligence (AI), Internet of Things (IoT), and Edge Computing (EC). This work presents the survey results conducted among academics to assess the current state of university courses, regarding the level of knowledge and skills provided to students about the Internet of Things, Artificial Intelligence, and Edge Computing. The novelty of the work is that (a) the research was carried out in several European countries, (b) the current curricula of universities from different countries were analyzed, and (c) the results present the teachers' perspective. To conduct the research, the analysis of the relevant literature took place initially to explore the issues of the presented subject, which will increasingly concern the industry in the near future. Based on the literature review results and analysis of the universities' curricula involved in this study, a questionnaire was prepared and shared with academics. The outcomes of the analysis reveal the areas that require more attention from scholars and possibly modernization of curricula.

Compacting oblivious agents on dynamic rings.

In this paper we investigate dynamic networks populated by autonomous mobile agents. Dynamic networks are networks whose topology can change continuously, at unpredictable locations and at unpredictable times. These changes are not considered to be faults, but rather an integral part of the nature of the system. The agents can autonomously move on the network, with the goal of solving cooperatively an assigned common task. Here, we focus on a specific network: the unoriented ring. More specifically, we study 1-interval connected dynamic rings (i.e., at any time, at most one of the edges might be missing). The agents move according to the widely used Look-Compute-Move life cycle, and can be homogenous (thus identical) or heterogenous (agents are assigned colors from a set of c > 1 colors). For identical agents, their goal is to form a compact segment, where agents occupy a continuous part of the ring and no two agents occupy the same node: we call this the Compact Configuration Problem. In the case of agents with colors, called the Colored Compact Configuration Problem, the goal is to group agents such that each group is formed by all agents having the same color, it occupies a continuous segment of the network, and groups of agents having different colors occupy distinct areas of the network. In this paper we determine the necessary conditions to solve both proposed problems. For all solvable cases, we provide algorithms for both the monochromatic and the colored version of the compact configuration problem. All our algorithms work even for the simplest model where agents have no persistent memory, no communication capabilities and do not agree on a common orientation within the network. To the best of our knowledge this is the first work on the compaction problem in a dynamic network.

PortWeather: A Lightweight Onboard Solution for Real-Time Weather Prediction.

Maritime journeys significantly depend on weather conditions, and so meteorology has always had a key role in maritime businesses. Nowadays, the new era of innovative machine learning approaches along with the availability of a wide range of sensors and microcontrollers creates increasing perspectives for providing on-board reliable short-range forecasting of main meteorological variables. The main goal of this study is to propose a lightweight on-board solution for real-time weather prediction. The system is composed of a commercial weather station integrated with an industrial IOT-edge data processing module that computes the wind direction and speed forecasts without the need of an Internet connection. A regression machine learning algorithm was chosen so as to require the smallest amount of resources (memory, CPU) and be able to run in a microcontroller. The algorithm has been designed and coded following specific conditions and specifications. The system has been tested on real weather data gathered from static weather stations and onboard during a test trip. The efficiency of the system has been proven through various error metrics.

SEAI: Social Emotional Artificial Intelligence Based on Damasio's Theory of Mind.

A socially intelligent robot must be capable to extract meaningful information in real time from the social environment and react accordingly with coherent human-like behavior. Moreover, it should be able to internalize this information, to reason on it at a higher level, build its own opinions independently, and then automatically bias the decision-making according to its unique experience. In the last decades, neuroscience research highlighted the link between the evolution of such complex behavior and the evolution of a certain level of consciousness, which cannot leave out of a body that feels emotions as discriminants and prompters. In order to develop cognitive systems for social robotics with greater human-likeliness, we used an "understanding by building" approach to model and implement a well-known theory of mind in the form of an artificial intelligence, and we tested it on a sophisticated robotic platform. The name of the presented system is SEAI (Social Emotional Artificial Intelligence), a cognitive system specifically conceived for social and emotional robots. It is designed as a bio-inspired, highly modular, hybrid system with emotion modeling and high-level reasoning capabilities. It follows the deliberative/reactive paradigm where a knowledge-based expert system is aimed at dealing with the high-level symbolic reasoning, while a more conventional reactive paradigm is deputed to the low-level processing and control. The SEAI system is also enriched by a model that simulates the Damasio's theory of consciousness and the theory of Somatic Markers. After a review of similar bio-inspired cognitive systems, we present the scientific foundations and their computational formalization at the basis of the SEAI framework. Then, a deeper technical description of the architecture is disclosed underlining the numerous parallelisms with the human cognitive system. Finally, the influence of artificial emotions and feelings, and their link with the robot's beliefs and decisions have been tested in a physical humanoid involved in Human-Robot Interaction (HRI).

Can a Humanoid Face be Expressive? A Psychophysiological Investigation.

Non-verbal signals expressed through body language play a crucial role in multi-modal human communication during social relations. Indeed, in all cultures, facial expressions are the most universal and direct signs to express innate emotional cues. A human face conveys important information in social interactions and helps us to better understand our social partners and establish empathic links. Latest researches show that humanoid and social robots are becoming increasingly similar to humans, both esthetically and expressively. However, their visual expressiveness is a crucial issue that must be improved to make these robots more realistic and intuitively perceivable by humans as not different from them. This study concerns the capability of a humanoid robot to exhibit emotions through facial expressions. More specifically, emotional signs performed by a humanoid robot have been compared with corresponding human facial expressions in terms of recognition rate and response time. The set of stimuli included standardized human expressions taken from an Ekman-based database and the same facial expressions performed by the robot. Furthermore, participants' psychophysiological responses have been explored to investigate whether there could be differences induced by interpreting robot or human emotional stimuli. Preliminary results show a trend to better recognize expressions performed by the robot than 2D photos or 3D models. Moreover, no significant differences in the subjects' psychophysiological state have been found during the discrimination of facial expressions performed by the robot in comparison with the same task performed with 2D photos and 3D models.

Open Biomedical Engineering education in Africa.

Despite the virtual revolution, the mainstream academic community in most countries remains largely ignorant of the potential of web-based teaching resources and of the expansion of open source software, hardware and rapid prototyping. In the context of Biomedical Engineering (BME), where human safety and wellbeing is paramount, a high level of supervision and quality control is required before open source concepts can be embraced by universities and integrated into the curriculum. In the meantime, students, more than their teachers, have become attuned to continuous streams of digital information, and teaching methods need to adapt rapidly by giving them the skills to filter meaningful information and by supporting collaboration and co-construction of knowledge using open, cloud and crowd based technology. In this paper we present our experience in bringing these concepts to university education in Africa, as a way of enabling rapid development and self-sufficiency in health care. We describe the three summer schools held in sub-Saharan Africa where both students and teachers embraced the philosophy of open BME education with enthusiasm, and discuss the advantages and disadvantages of opening education in this way in the developing and developed world.

Inference of human affective states from psychophysiological measurements extracted under ecologically valid conditions.

Compared to standard laboratory protocols, the measurement of psychophysiological signals in real world experiments poses technical and methodological challenges due to external factors that cannot be directly controlled. To address this problem, we propose a hybrid approach based on an immersive and human accessible space called the eXperience Induction Machine (XIM), that incorporates the advantages of a laboratory within a life-like setting. The XIM integrates unobtrusive wearable sensors for the acquisition of psychophysiological signals suitable for ambulatory emotion research. In this paper, we present results from two different studies conducted to validate the XIM as a general-purpose sensing infrastructure for the study of human affective states under ecologically valid conditions. In the first investigation, we recorded and classified signals from subjects exposed to pictorial stimuli corresponding to a range of arousal levels, while they were free to walk and gesticulate. In the second study, we designed an experiment that follows the classical conditioning paradigm, a well-known procedure in the behavioral sciences, with the additional feature that participants were free to move in the physical space, as opposed to similar studies measuring physiological signals in constrained laboratory settings. Our results indicate that, by using our sensing infrastructure, it is indeed possible to infer human event-elicited affective states through measurements of psychophysiological signals under ecological conditions.

Wearable wireless tactile display for virtual interactions with soft bodies.

We describe here a wearable, wireless, compact, and lightweight tactile display, able to mechanically stimulate the fingertip of users, so as to simulate contact with soft bodies in virtual environments. The device was based on dielectric elastomer actuators, as high-performance electromechanically active polymers. The actuator was arranged at the user's fingertip, integrated within a plastic case, which also hosted a compact high-voltage circuitry. A custom-made wireless control unit was arranged on the forearm and connected to the display via low-voltage leads. We present the structure of the device and a characterization of it, in terms of electromechanical response and stress relaxation. Furthermore, we present results of a psychophysical test aimed at assessing the ability of the system to generate different levels of force that can be perceived by users.

A flexible bioreactor system for constructing in vitro tissue and organ models.

To develop in vitro models of cells, tissues and organs we have designed and realized a series of cell culture chambers. Each chamber is purpose designed to simulate a particular feature of the in vivo environment. The bioreactor system is user friendly, and the chambers are easy to produce, sterilize and assemble. In addition they can be connected together to simulate inter-organ or tissue cross-talk. Here we discuss the design philosophy of the bioreactor system and then describe its construction. Preliminary results of validation tests obtained with hepatocytes and endothelial cells are also reported. The results show that endothelial cells are extremely sensitive to small levels of shear stress and that the presence of heterotypic signals from endothelial cells enhances the endogenous metabolic function of hepatocytes.

Squeeze pressure bioreactor: a hydrodynamic bioreactor for noncontact stimulation of cartilage constructs.

A novel squeeze pressure bioreactor for noncontact hydrodynamic stimulation of cartilage is described. The bioreactor is based on a small piston that moves up and down, perpendicular to a tissue construct, in a fluid-filled chamber. Fluid displaced by the piston generates a pressure wave and shear stress as it moves across the sample, simulating the dynamic environment of a mobile joint. The fluid dynamics inside the squeeze pressure bioreactor was modeled using analytical and computational methods to simulate the mechanical stimuli imposed on a construct. In particular, the pressure, velocity field, and wall shear stress generated on the surface of the construct were analyzed using the theory of hydrodynamic lubrication, which describes the flow of an incompressible fluid between two surfaces in relative motion. Both the models and in-situ pressure measurements in the bioreactor demonstrate that controlled cyclic stresses of up to 10 kPa can be applied to tissue constructs. Initial tests on three-dimensional scaffolds seeded with chondrocytes show that glycosaminoglycan production is increased with regard to controls after 24 and 48 h of cyclic noncontact stimulation in the bioreactor.

Development and evaluation of a social robot platform for therapy in autism.

People with ASD (Autism Spectrum Disorders) have difficulty in managing interpersonal relationships and common life social situations. A modular platform for Human Robot Interaction and Human Machine Interaction studies has been developed to manage and analyze therapeutic sessions in which subjects are driven by a psychologist through simulated social scenarios. This innovative therapeutic approach uses a humanoid robot called FACE capable of expressing and conveying emotions and empathy. Using FACE as a social interlocutor the psychologist can emulate real life scenarios where the emotional state of the interlocutor is adaptively adjusted through a semi closed loop control algorithm which uses the ASD subject's inferred "affective" state as input. Preliminary results demonstrate that the platform is well accepted by ASDs and can be consequently used as novel therapy for social skills training.

Finite element modelling and design of a concentration gradient generating bioreactor: application to biological pattern formation and toxicology.

This paper describes the use of a microfluidic gradient maker for the toxicological analysis of some conventional biomolecules such as hydrogen peroxide and a local anaesthetic, lidocaine on different cell cultures, human endothelial cells and myoblasts, respectively. The microfluidic device was designed and simulated using COMSOL Multiphysics and the concentration gradient in the microfluidic network was analysed through a fluid-dynamic and mass-transport study. Subsequently the device was fabricated with soft lithography, casting PDMS in a master to obtain channels about 250 microm deep. Hydrogen peroxide was tested on human endothelial cells, while lidocaine was tested on C2C12 myoblasts and an analysis was performed using propidium iodide staining followed by an imaging processing routine to obtain quantitative dose-response profiles in the gradient maker. The results show that the Gradient Maker (GM) bioreactor is a more sensitive method for detection of cell toxicity, and compared with testing of drug toxicity using microwells with individual cell cultures, allows one shot testing with a single cell culture exposed to a large number of concentrations. Moreover, the Gradient Maker was also modelled in order to realise biological pattern formation using two morphogenes acting as activator and inhibitor with varying diffusion rates.

In vitro liver model using microfabricated scaffolds in a modular bioreactor.

Hepatocyte function on 3-D microfabricated polymer scaffolds realised with the pressure-activated microsyringe was tested under static and dynamic conditions. The dynamic cell culture was obtained using the multicompartment modular bioreactor system. Hepatocyte cell density, glucose consumption, and albumin secretion rate were measured daily over a week. Cells seeded on scaffolds showed an increase in cell density compared with monolayer controls. Moreover, in dynamic culture, cell metabolic function increased three times in comparison with static monolayer cultures. These results suggest that cell density and cell-cell interactions are mediated by the architecture of the substrate, while the endogenous biochemical functions are regulated by a sustainable supply of nutrients and interstitial-like flow. Thus, a combination of 3-D scaffolds and dynamic flow conditions are both important for the development of a hepatic tissue model for applications in drug testing and regenerative medicine.