The integration of the Internet of Things (IoT) and Machine Learning (ML) technologies has opened up for the development of novel types of systems and services. Federated Learning (FL) has enabled the systems to collaboratively train their ML models while preserving the privacy of the data collected by their IoT devices and objects. Several FL frameworks have been developed, however, they do not enable FL in open, distributed, and heterogeneous IoT environments. Specifically, they do not support systems that collect similar data to dynamically discover each other, communicate, and negotiate about the training terms (e.g., accuracy, communication latency, and cost). Towards bridging this gap, we propose ART4FL, an end-to-end framework that enables FL in open IoT settings. The framework enables systems' users to configure agents that participate in FL on their behalf. Those agents negotiate and make commitments (i.e., contractual agreements) to dynamically form federations. To perform FL, the framework deploys the needed services dynamically, monitors the training rounds, and calculates agents' trust scores based on the established commitments. ART4FL exploits a blockchain network to maintain the trust scores, and it provides those scores to negotiating agents' during the federations' formation phase.

Interactive machine learning (ML) adds a human-in-the-loop aspect to a ML system. Even though the input from human users to the system is a central part of the concept, the uncertainty caused by the human feedback is often not considered in interactive ML. The assumption that the human user is expected to always provide correct feedback, typically does not hold in real-world scenarios. This is especially important for when the cognitive workload of the human is high, for instance in online learning from streaming data where there are time constraints for providing the feedback. We present experiments of interactive online ML with human participants, and compare the results to simulated experiments where humans are always correct. We found combining the two interactive learning paradigms, active learning and machine teaching, resulted in better performance compared to machine teaching alone. The results also showed an increased discrepancy between the experiments with human participants and the simulated experiments when the cognitive workload was increased. The findings suggest the importance of taking uncertainty caused by human factors into consideration in interactive ML, especially in situations which requires a high cognitive workload for the human.

On-demand transport services are often envisioned as stand-alone modes or as a replacement for conventional public transport modes. This leads to a comparison of service efficiencies, or direct competition for passengers between them. The results of this work point to the positive effects of the inclusion of DRT into the public transport network. We simulate a day of operation of a DRT service in a rural area and demonstrate that a DRT system that focuses on increasing accessibility for travellers with poor public transport access can be quite efficient, especially for reducing environmental impact. We show that DRT, while it produces more vehicle kilometres than private cars would inside the DRT operating zone, can help to reduce the vehicle kilometres travelled for long-distance trips. The results of this study indicate the need for a more systemic evaluation of the impact of the new mobility modes.

This paper concerns the novel concept of an Interactive Dynamic Intelligent Virtual Sensor (IDIVS), which extends virtual/soft sensors towards making use of user input through interactive learning (IML) and transfer learning. In research, many studies can be found on using machine learning in this domain, but not much on using IML. This paper contributes by highlighting how this can be done and the associated positive potential effects and challenges. An IDIVS provides a sensor-like output and achieves the output through the data fusion of sensor values or from the output values of other IDIVSs. We focus on settings where people are present in different roles: from basic service users in the environment being sensed to interactive service users supporting the learning of the IDIVS, as well as configurators of the IDIVS and explicit IDIVS teachers. The IDIVS aims at managing situations where sensors may disappear and reappear and be of heterogeneous types. We refer to and recap the major findings from related experiments and validation in complementing work. Further, we point at several application areas: smart building, smart mobility, smart learning, and smart health. The information properties and capabilities needed in the IDIVS, with extensions towards information security, are introduced and discussed.

In rural areas with low demand, demand responsive transport (DRT) can provide an alternative to the regular public transport bus lines, which are expensive to operate in such conditions. With simulation, we explore the potential effects of introducing a DRT service that replaces existing bus lines in Lolland municipality in Denmark, assuming that the existing demand remains unchanged. We set up the DRT service in such a way that its service quality (in terms of waiting time and in-vehicle time) is comparable to the replaced buses. The results show that a DRT service can be more cost efficient than regular buses and can produce significantly less CO2 emissions when the demand level is low. Additionally, we analyse the demand density at which regular buses become more cost efficient and explore how the target service quality of a DRT service can affect operational characteristics. Overall, we argue that DRT could be a more sustainable mode of public transport in low demand areas.

Multimodal journey planners are used worldwide to support travelers in planning and executing their journeys. Generated travel plans usually involve local mobility service providers, consider some travelers' preferences, and provide travelers information about the routes' current status and expected delays. However, those planners cannot fully consider the special situations of individual cities when providing travel planning services. Specifically, authorities of different cities might define customizable regulations or constraints of movements in the cities (e.g., due to construction works or pandemics). Moreover, with the transformation of traditional cities into smart cities, travel planners could leverage advanced monitoring features. Finally, most planners do not consider relevant information impacting travel plans, for instance, information that might be provided by travelers (e.g., a crowded square) or by mobility service providers (e.g., changing the timetable of a bus). To address the aforementioned shortcomings, in this paper, we propose ROUTE, a framework for customizable smart mobility planners that better serve the needs of travelers, local authorities, and mobility service providers in the dynamic ecosystem of smart cities. ROUTE is composed of an architecture, a process, and a prototype developed to validate the feasibility of the framework. Experiments' results show that the framework scales well in both centralized and distributed deployment settings.

The increasing popularity of e-commerce requires efficient solutions for the provision of last mile logistics. There are different approaches for delivering parcels, e.g., home delivery, service points, or parcel lockers, which have different advantages and disadvantages for customers and logistics providers in terms of flexibility, accessibility, and operating costs. We have studied a novel transportation solution where electric vehicles dynamically set up smart cargo boxes, from which customers can fetch their delivery at any time of the day. This provides customers with a more flexible access to their packages and allows the service provider to deliver the parcels more efficiently. In this article, we present the results of a feasibility study conducted in Västra Hamnen, Malmö (Sweden). The developed simulation model shows that smart boxes not only are a viable approach for efficient last mile deliveries, but also result in considerably smaller travel distances compared to conventional package delivery.

Interactive Machine Learning (ML) has the potential to lower the manual labelling effort needed, as well as increase classification performance by incorporating a human-in-the loop component. However, the assumptions made regarding the interactive behaviour of the human in experiments are often not realistic. Active learning typically treats the human as a passive, but always correct, participant. Machine teaching provides a more proactive role for the human, but generally assumes that the human is constantly monitoring the learning process. In this paper, we present an interactive online framework and perform experiments to compare active learning, machine teaching and combined approaches. We study not only the classification performance, but also the effort (to label samples) and attention (to monitor the ML system) required of the human. Results from experiments show that a combined approach generally performs better with less effort compared to active learning and machine teaching. With regards to attention, the best performing strategy varied depending on the problem setup.

When planning interventions to limit the spread of Covid-19, the current state of knowledge about the disease and specific characteristics of the population need to be considered. Simulations can facilitate policy making as they take prevailing circumstances into account. Moreover, they allow for the investigation of the potential effects of different interventions using an artificial population. Agent-based Social Simulation (ABSS) is argued to be particularly useful as it can capture the behavior of and interactions between individuals. We performed a systematic literature reviewand identified 126 articles that describe ABSS of Covid-19 transmission processes. Our reviewshowed that ABSS is widely used for investigating the spread of Covid-19. Existing models are very heterogeneous with respect to their purpose, the number of simulated individuals, and the modeled geographical region, as well as how they model transmission dynamics, disease states, human behavior, and interventions. To this end, a discrepancy can be identified between the needs of policy makers and what is implemented by the simulation models. This also includes how thoroughly the models consider and represent the real world, e.g. in terms of factors that affect the transmission probability or how humans make decisions. Shortcomingswere also identified in the transparency of the presented models, e.g. in terms of documentation or availability, as well as in their validation, which might limit their suitability for supporting decision-making processes. We discuss how these issues can be mitigated to further establish ABSS as a powerful tool for crisis management.

This article demonstrates an approach to the simulation of Demand Responsive Transport (DRT) – a flexible transport mode that typically operates as a combination of taxi and bus modes. Travellers request individual trips and DRT is capable of adjusting its routes or schedule to the needs of travellers. It has been seen as a part of the public transport network, which has the potential to reduce operational costs of public transport services, to provide better service quality for population groups with limited mobility and to improve transport fairness. However, a DRT service needs to be thoroughly planned to target the intended user groups, attract a sufficient demand level and maintain reasonable operational costs. As the demand for DRT is dynamic and heterogeneous, it is difficult to simulate it with a macro approach. To address this problem, we develop and evaluate an individual-based simulation comprising models of traveller behaviour for both supply and demand sides. Travellers choose a trip alternative with a mode choice model and DRT vehicle routing utilises a model of travellers’ mode choice behaviour to optimise routes. This allows capturing supply-side operational costs and demand-side service quality for every individual, what allows for designing a personalised service that can prioritise needy groups of travellers improving transport fairness. By simulating different setups of DRT services, the simulator can be used as a decision support tool.

When simulation models shall be used to support decision-making, the trustworthiness of the results need to be ensured. Usually, models are validated against real-world data. Yet, in the ongoing pandemic, there is a lack of respective data that can be used to validate the model’s behaviour. To overcome this issue, this chapter discusses the validation of simulation models for the Covid-19 pandemic by comparing their results among each other. To this end, we present a formal comparison between the existing behaviour-based epidemiological model that was developed at the University of Oxford and the ASSOCC model.

The Internet of Things (IoT) is revolutionizing our environments with novel types of services and applications by exploiting the large number of diverse connected things. One of the main challenges in the IoT is to engineer systems to support human users to achieve their goals in dynamic and uncertain environments. For instance, the mobility of both users and devices makes it infeasible to always foresee the available things in the users’ current environments. Moreover, users’ activities and/or goals might change suddenly. To support users in such environments, we developed an initial approach that exploits the notion of Emergent Configurations (ECs) and mixed initiative techniques to engineer self-configuring IoT systems. An EC is a goal-driven IoT system composed of a dynamic set of temporarily connecting and cooperating things. ECs are more flexible and usable than IoT systems whose constituents and interfaces are fully specified at design time

The introduction of this book sets the stage of performing social simulations in a crisis. The contents of the book are based on the experience of creating a large scale and complex social simulation for the Covid-19 crisis. However, the contents are reaching much further than just this experience. We will show the general contribution that social simulations based on fundamental social-psychological principles can have in times of crises. In times of big societal changes due to a pandemic or other disaster, these simulations can give handles to support decision makers in their difficult task to act in a very short time with many uncertainties. Besides giving our results, we also will indicate why the results are trustworthy and interesting. Finally we also look what challenges should be picked up to convert the successful project into a sustainable research area.

Smart homes promise to improve the quality of life of residents. However, they collect vasts amounts of personal and sensitive data, making privacy protection critically important. We propose a framework, called PRASH, for modeling and analyzing the privacy risks of smart homes. It is composed of three modules: a system model, a threat model, and a set of privacy metrics, which together are used for calculating the privacy risk exposure of a smart home system. By representing a smart home through a formal specification, PRASH allows for early identification of threats, better planning for risk management scenarios, and mitigation of potential impacts caused by attacks before they compromise the lives of residents. To demonstrate the capabilities of PRASH, an executable version of the smart home system configuration was generated using the proposed formal specification, which was then analyzed to find potential attack paths while also mitigating the impacts of those attacks. Thereby, we add important contributions to the body of knowledge on the mitigations of threat agents violating the privacy of users in their homes. Overall, the use of PRASH will help residents to preserve their right to privacy in the face of the emerging challenges affecting smart homes.

Many Internet of Things (IoT) systems generate a massive amount of data needing to be processed and stored efficiently. Cloud computing solutions are often used to handle these tasks. However, the increasing availability of computational resources close to the edge has prompted the idea of using these for distributed computing and storage. Edge computing may help to improve IoT systems regarding important quality attributes like latency, energy consumption, privacy, and bandwidth utilization. However, deciding where to deploy the various application components is not a straightforward task. This is largely due to the trade-offs between the quality attributes relevant for the application. We have performed a systematic mapping study of 98 articles to investigate which quality attributes have been used in the literature for assessing IoT systems using edge computing. The analysis shows that time behavior and resource utilization are the most frequently used quality attributes; further, response time, turnaround time, and energy consumption are the most used metrics for quantifying these quality attributes. Moreover, simulation is the main tool used for the assessments, and the studied trade-offs are mainly between only two qualities. Finally, we identified a number of research gaps that need further study.

The international community has largely recognized that the Earth's climate is changing. Mitigating its global effects requires international actions. The European Union (EU) is leading several initiatives focused on reducing the problems. Specifically, the Climate Action tries to both decrease EU greenhouse gas emissions and improve energy efficiency by reducing the amount of primary energy consumed, and it has pointed to the development of efficient building energy management systems as key. In traditional buildings, households are responsible for continuously monitoring and controlling the installed Heating, Ventilation, and Air Conditioning (HVAC) system. Unnecessary energy consumption might occur due to, for example, forgetting devices turned on, which overwhelms users due to the need to tune the devices manually. Nowadays, smart buildings are automating this process by automatically tuning HVAC systems according to user preferences in order to improve user satisfaction and optimize energy consumption. Towards achieving this goal, in this paper, we compare 36 Machine Learning algorithms that could be used to forecast indoor temperature in a smart building. More specifically, we run experiments using real data to compare their accuracy in terms of R-coefficient and Root Mean Squared Error and their performance in terms of Friedman rank. The results reveal that the ExtraTrees regressor has obtained the highest average accuracy (0.97%) and performance (0,058%) over all horizons.

Contemporary Machine Learning (ML) often focuseson large existing and labeled datasets and metrics aroundaccuracy and performance. In pervasive online systems, conditionschange constantly and there is a need for systems thatcan adapt. In Machine Teaching (MT) a human domain expertis responsible for the knowledge transfer and can thus addressthis. In my work, I focus on domain experts and the importanceof, for the ML system, available features and the space they span.This space confines the, to the ML systems, observable fragmentof the physical world. My investigation of the feature space isgrounded in a conducted study and related theories. The resultof this work is applicable when designing systems where domainexperts have a key role as teachers.

Engineering Internet of Things (IoT) systems is a challenging task partly due to the dynamicity and uncertainty of the environment including the involvement of the human in the loop. Users should be able to achieve their goals seamlessly in different environments, and IoT systems should be able to cope with dynamic changes. Several approaches have been proposed to enable the automated formation, enactment, and self-adaptation of goal-driven IoT systems. However, they do not address deployment issues. In this paper, we propose a goal-driven approach for deploying self-adaptive IoT systems in the Edge-Cloud continuum. Our approach supports the systems to cope with the dynamicity and uncertainty of the environment including changes in their deployment topologies, i.e., the deployment nodes and their interconnections. We describe the architecture and processes of the approach and the simulations that we conducted to validate its feasibility. The results of the simulations show that the approach scales well when generating and adapting the deployment topologies of goal-driven IoT systems in smart homes and smart buildings.

Smart connected homes are integrated with heterogeneous Internet-connected devices interacting with the physical environment and human users. While they have become an established research area, there is no common understanding of what composes such a pervasive environment making it challenging to perform a scientific analysis of the domain. This is especially evident when it comes to discourse about privacy threats. Recognizing this, we aim to describe a generic smart connected home, including the data it deals with in a novel privacy-centered system model. Such is done using concepts borrowed from the theory of Contextual Integrity. Furthermore, we represent privacy threats formally using the proposed model. To illustrate the usage of the model, we apply it to the design of an ambient-assisted living use-case and demonstrate how it can be used for identifying and analyzing the privacy threats directed to smart connected homes.

In interactive machine learning, human users and learning algorithms work together in order to solve challenging learning problems, e.g. with limited or no annotated data or trust issues. As annotating data can be costly, it is important to minimize the amount of annotated data needed for training while still getting a high classification accuracy. This is done by attempting to select the most informative data instances for training, where the amount of instances is limited by a labelling budget. In an online learning setting, the decision of whether or not to select an instance for labelling has to be done on-the-fly, as the data arrives in a sequential order and is only valid for a limited time period. We present a taxonomy of interactive online machine learning strategies. An interactive learning strategy determines which instances to label in an unlabelled dataset. In the taxonomy we differentiate between interactive learning strategies when the computer controls the learning process (active learning) and those when human users control the learning process (machine teaching). We then make a distinction between what triggers the learning: active learning could be triggered by uncertainty, time, or randomly, whereas machine teaching could be triggered by errors, state changes, time, or factors related to the user. We also illustrate the taxonomy by implementing versions of the different strategies and performing experiments on a benchmark dataset as well as on a synthetically generated dataset. The results show that the choice of interactive learning strategy affects performance, especially in the beginning of the online learning process, when there is a limited amount of labelled data.

Internet of Things (IoT) environments encompass different types of devices and objects that offer a wide range of services. The dynamicity and uncertainty of those environments, including the mobility of users and devices, make it hard to foresee at design time available devices, objects, and services. For the users to benefit from such environments, they should be proposed services that are relevant to the specific context and can be provided by available things. Moreover, environments should be configured automatically based on users' preferences. To address these challenges, we propose an approach that leverages Artificial Intelligence techniques to recognize users' activities and provides relevant services to support users to perform their activities. Moreover, our approach learns users' preferences and configures their environments accordingly by dynamically forming, enacting, and adapting goal-driven IoT systems. In this paper, we present a conceptual model, a multi-tier architecture, and processes of our approach. Moreover, we report about how we validated the feasibility and evaluated the scalability of the approach through a prototype that we developed and used.

The advances in Internet of things lead to an increased number of devices generating and streaming data. These devices can be useful data sources for activity recognition by using machine learning. However, the set of available sensors may vary over time, e.g. due to mobility of the sensors and technical failures. Since the machine learning model uses the data streams from the sensors as input, it must be able to handle a varying number of input variables, i.e. that the feature space might change over time. Moreover, the labelled data necessary for the training is often costly to acquire. In active learning, the model is given a budget for requesting labels from an oracle, and aims to maximize accuracy by careful selection of what data instances to label. It is generally assumed that the role of the oracle only is to respond to queries and that it will always do so. In many real-world scenarios however, the oracle is a human user and the assumptions are simplifications that might not give a proper depiction of the setting. In this work we investigate different interactive machine learning strategies, out of which active learning is one, which explore the effects of an oracle that can be more proactive and factors that might influence a user to provide or withhold labels. We implement five interactive machine learning strategies as well as hybrid versions of them and evaluate them on two datasets. The results show that a more proactive user can improve the performance, especially when the user is influenced by the accuracy of earlier predictions. The experiments also highlight challenges related to evaluating performance when the set of classes is changing over time.

The Internet of Things (IoT) has enabled physical objects and devices, often referred to as things, to connect and communicate. This has opened up for the development of novel types of services that improve the quality of our daily lives. The dynamicity and uncertainty of IoT environments, including the mobility of users and devices, make it hard to foresee at design time available things and services. Further, users should be able to achieve their goals seamlessly in arbitrary environments. To address these challenges, we exploit Artificial Intelligence (AI) to engineer smart IoT systems that can achieve user goals and cope with the dynamicity and uncertainty of their environments. More specifically, the main contribution of this paper is an approach that leverages the notion of Belief-Desire-Intention agents and Machine Learning (ML) techniques to realize Emergent Configurations (ECs) in the IoT. An EC is an IoT system composed of a dynamic set of things that connect and cooperate temporarily to achieve a user goal. The approach enables the distributed formation, enactment, adaptation of ECs, and conflict resolution among them. We present a conceptual model of the entities of the approach, its underlying processes, and the guidelines for using it. Moreover, we report about the simulations conducted to validate the feasibility of the approach and evaluate its scalability. View Full-Text

During the COVID-19 crisis there have been many difficult decisions governments and other decision makers had to make. E.g. do we go for a total lock down or keep schools open? How many people and which people should be tested? Although there are many good models from e.g. epidemiologists on the spread of the virus under certain conditions, these models do not directly translate into the interventions that can be taken by government. Neither can these models contribute to understand the economic and/or social consequences of the interventions. However, effective and sustainable solutions need to take into account this combination of factors. In this paper, we propose an agent-based social simulation tool, ASSOCC, that supports decision makers understand possible consequences of policy interventions, but exploring the combined social, health and economic consequences of these interventions.

For the efficient execution of Deep Neural Networks (DNN) in the Internet of Things, computation tasks can be distributed and deployed on edge nodes. In contrast to deploying all computation to the cloud, the use of Distributed DNN (DDNN) often results in a reduced amount of data that is sent through the network and thus might increase the overall performance of the system. However, finding an appropriate deployment scenario is often a complex task and requires considering several criteria. In this paper, we introduce a multi-criteria decision-making method based on the Analytical Hierarchy Process for the comparison and selection of deployment alternatives. We use the RECAP simulation framework to model and simulate DDNN deployments on different scales to provide a comprehensive assessment of deployments to system designers. In a case study, we apply the method to a smart city scenario where different distributions and deployments of a DNN are analyzed and compared.

Deep neural networks have the ability to generalize beyond observed training data. However, for some applications they may produce output that apriori is known to be invalid. If prior knowledge of valid output regions is available, one way of imposing constraints on deep neural networks is by introducing these priors in a loss function. In this paper, we introduce a novel way of constraining neural network output by using encoded regions with a loss function based on gradient interpolation. We evaluate our method in a positioning task where a region map is used in order to reduce invalid position estimates. Results show that our approach is effective in decreasing invalid outputs for several geometrically complex environments.

Machine learning research today is dominated by atechnocentric perspective and in many cases disconnected fromthe users of the technology. The machine teaching paradigm insteadshifts the focus from machine learning experts towards thedomain experts and users of machine learning technology. Thisshift opens up for new perspectives on the current use of machinelearning as well as new usage areas to explore. In this study,we apply and map existing machine teaching principles ontoa contextual machine teaching implementation in a commutingsetting. The aim is to highlight areas in machine teaching theorythat requires more attention. The main contribution of this workis an increased focus on available features, the features space andthe potential to transfer some of the domain expert’s explanatorypowers to the machine learning system.

Building interpretable machine learning agents is a challenge that needs to be addressed to make the agents trustworthy and align the usage of the technology with human values. In this work, we focus on how to evaluate interpretability in a machine teaching setting, a settingthat involves a human domain expert as a teacher in relation to a machine learning agent. By using a prototype in a study, we discuss theinterpretability denition and show how interpretability can be evaluatedon a functional-, human- and application level. We end the paperby discussing open questions and suggestions on how our results can be transferable to other domains.

The Internet of Things has become a central and exciting research area encompassing many fields in information and communication technologies and adjacent domains. IoT systems involve interactions with heterogeneous, distributed, and intelligent things, both from the digital and physical worlds including the human in the loop. Thanks to the increasingly wide spectrum of applications and cheap availability of both network connectivity and devices, a number of different stakeholders from industry, academia, society and government are part of the IoT ecosystem.

Smart connected home systems bring different privacy challenges to residents. The contribution of this paper is a novel privacy grounded classification of smart connected home systems that is focused on personal data exposure. This classification is built empirically through k-means cluster analysis from the technical specification of 81 commercial Internet of Things (IoT) systems as featured in PrivacyNotIncluded – an online database of consumer IoT systems. The attained classification helps us better understand the privacy implications and what is at stake with different smart connected home systems. Furthermore, we survey the entire spectrum of analyzed systems for their data collection capabilities. Systems were classified into four tiers: app-based accessors, watchers, location harvesters, and listeners, based on the sensing data the systems collect. Our findings indicate that being surveilled inside your home is a realistic threat, particularly, as the majority of the surveyed in-home IoT systems are installed with cameras, microphones, and location trackers. Finally, we identify research directions and suggest some best practices to mitigate the threat of in-house surveillance.

For the provision of efficient and high-quality public transport services in rural areas with a low population density, the introduction of Demand Responsive Transport (DRT) services is reasonable. The optimal design of such services depends on various socio-demographical and environmental factors, which is why the use of simulation is feasible to support planning and decision-making processes. A key challenge for sound simulation results is the generation of realistic demand, i.e., requests for DRT journeys. In this paper, a method for modelling and simulating commuting activities is presented, which is based on statistical real-world data. It is applied to Sjöbo and Tomelilla, two rural municipalities in southern Sweden.

Deep neural networks for positioning can improve accuracy by adapting to inhomogeneous environments. However, they are still susceptible to noisy data, often resulting in invalid positions. A related task, map matching, can be used for reducing geographical invalid positions by aligning observations to a model of the real world. In this paper, we propose an approach for positioning, enhanced with map matching, within a single deep neural network model. We introduce a novel way of reducing the number of invalid position estimates by adding map information to the input of the model and using a map-based loss function. Evaluating on real-world Received Signal Strength Indicator data from an asset tracking application, we show that our approach gives both increased position accuracy and a decrease of one order of magnitude in the number of invalid positions.

In interactive machine learning it is important to select the most informative data instances to label in order to minimize the effort of the human user. There are basically two categories of interactive machine learning. In the first category, active learning, it is the computational learner that selects which data to be labelled by the human user, whereas in the second one, machine teaching, the selection is done by the human teacher. It is often assumed that the human user is a perfect oracle, i.e., a label will always be provided in accordance with the interactive learning strategy and that this label will always be correct. In real-world scenarios however, these assumptions typically do not hold. In this work, we investigate how the reliability of the user providing labels affects the performance of online machine learning. Specifically, we study reluctance, i.e., to what extent the user does not provide labels in accordance with the strategy, and fallibility, i.e., to what extent the provided labels are incorrect. We show results of experiments on a benchmark dataset as well as a synthetically created dataset. By varying the degree of reluctance and fallibility of the user, the robustness of the different interactive learning strategies and machine learning algorithms is explored. The experiments show that there is a varying robustness of the strategies and algorithms. Moreover, certain machine learning algorithms are more robust towards reluctance compared to fallibility, while the opposite is true for others

Today, the Internet of Things (IoT) is mainly associated with vertically integrated systems that often are closed and fragmented in their applicability. To build a better IoT ecosystem, the open IoT platform has become a popular term in the recent years. However, this term is usually used in an intuitive way without clarifying the openness aspects of the platforms. The goal of this paper is to characterize the openness types of IoT platforms and investigate what makes them open. We conducted a systematic mapping study by retrieving data from 718 papers. As a result of applying the inclusion and exclusion criteria, 221 papers were selected for review. We discovered 46 IoT platforms that have been characterized as open, whereas 25 platforms are referred as open by some studies rather than the platforms themselves. We found that the most widely accepted and used open IoT platforms are NodeMCU and ThingSpeak that together hold a share of more than 70% of the declared open IoT platforms in the selected papers. The openness of an IoT platform is interpreted into different openness types. Our study results show that the most common openness type encountered in open IoT platforms is open-source, but also open standards, open APIs, open data and open layers are used in the literature. Finally, we propose a new perspective on how to define openness in the context of IoT platforms by providing several insights from the different stakeholder viewpoints.

This paper presents a computational componentdesigned to improve and evaluate emergency handlingplans.In real-time, the component operates as thecore of an Internet of Things (IoT) infrastructureaimed at crowd monitoring and optimum evacuationpaths planning. In this case, a software architecturefacilitates achieving the minimum time necessary toevacuate people from a building.In design-time,the component helps discovering the optimal buildingdimensions for a safe emergency evacuation, evenbefore (re-) construction of a building. The space andtime dimension are discretized according to metrics andmodels in literature. The component formulates andsolves a linearized, time-indexed flow problem on anetwork that represents feasible movements of people ata suitable frequency. The CPU time to solve the modelis compliant with real-time use. The application of themodel to a real location with real data testifies the modelcapability to optimize the safety standards by smallchanges in the building dimensions, and guarantees anoptimal emergency evacuation performance.

During the last decade, a large number of different definitions and taxonomies of Internet of Things (IoT) systems have been proposed. This has resulted in a fragmented picture and a lack of consensus about IoT systems and their constituents. To provide a better understanding of this issue and a way forward, we have conducted a Systematic Mapping Study (SMS) of existing IoT System taxonomies. In addition, we propose a characterization of IoT systems synthesized from the existing taxonomies, which provides a more holistic view of IoT systems than previous taxonomies. It includes seventeen characteristics, divided into two groups: elements and quality aspects. Finally, by analyzing the results of the SMS, we draw future research directions.

Although the availability of sensor data is becoming prevalent across many domains, it still remains a challenge to make sense of the sensor data in an efficient and effective manner in order to provide users with relevant services. The concept of virtual sensors provides a step towards this goal, however they are often used to denote homogeneous types of data, generally retrieved from a predetermined group of sensors. The DIVS (Dynamic Intelligent Virtual Sensors) concept was introduced in previous work to extend and generalize the notion of a virtual sensor to a dynamic setting with heterogenous sensors. This paper introduces a refined version of the DIVS concept by integrating an interactive machine learning mechanism, which enables the system to take input from both the user and the physical world. The paper empirically validates some of the properties of the DIVS concept. In particular, we are concerned with the distribution of different budget allocations for labelled data, as well as proactive labelling user strategies. We report on results suggesting that a relatively good accuracy can be achieved despite a limited budget in an environment with dynamic sensor availability, while proactive labeling ensures further improvements in performance.

The deployment of Internet of Things (IoT) applications is complex since many quality characteristics should be taken into account, for example, performance, reliability, and security. In this study, we investigate to what extent the current edge computing simulators support the analysis of qualities that are relevant to IoT architects who are designing an IoT system. We first identify the quality characteristics and metrics that can be evaluated through simulation. Then, we study the available simulators in order to assess which of the identified qualities they support. The results show that while several simulation tools for edge computing have been proposed, they focus on a few qualities, such as time behavior and resource utilization. Most of the identified qualities are not considered and we suggest future directions for further investigation to provide appropriate support for IoT architects.

The recent years have witnessed an enormous growth of mobile services for energy management in buildings. However, these solutions are often proprietary, non-interoperable, and handle only a limited function, such as lighting, ventilation, or heating. To address these issues, we have developed an open platform that is an integrated energy management solution for buildings. It includes an ecosystem of mobile services and open APIs as well as protocols for the development of new services and products. Moreover, it has an adapter layer that enables the platform to interoperate with any building management system (BMS) or individual device. Thus, the platform makes it possible for third-party developers to produce mobile energy efficiency applications that will work independently of which BMS and devices are used in the building. To validate the platform, a number of services have been implemented and evaluated in existing buildings. This has been done in cooperation with energy companies and property owners, together with the residents and other users of the buildings. The platform, which we call Elis, has been made available as open source software under an MIT license. View Full-Text

The advances in Internet of Things lead to an increased number of devices generating and streaming data. These devices can be useful data sources for Activity Recognition by using Machine Learning. However, as the set of available sensors may vary over time, e.g. due to mobility of the sensors and technical failures, the feature space might also change over time. Moreover, the labelled data necessary for the training is often costly to acquire. Active Learning is a type of Interactive Machine Learning where the model is given a budget for requesting labels from an oracle, and aims to maximize accuracy by careful selection of what data points to label. It is generally assumed that a query always gets a correct response, but in many real-world scenarios this is not a realistic assumption. In this work we investigate different Proactive Learning strategies, which explore the human factors of the oracle and aspects that might influence a user to provide or withhold labels. We implemented four proactive strategies and hybrid versions of them. They were evaluated on two datasets to examine how a more proactive, or reluctant, user affects performance. The results show that a more proactive user can improve the performance, especially when the user is influenced by the accuracy of earlier predictions. The experiments also highlight challenges related to evaluating performance when the set of classes is changing over time.

Different storage and handling conditions in cold supply chains often cause variations in the remaining shelf life of perishable foods. In particular, the actual shelf life may differ from the expiration date printed on the primary package. Based on temperature sensors placed on or close to the food products, a remaining shelf-life prediction (RSLP) service can be developed, which estimates the remaining shelf life of individual products, in real-time. This type of service may lead to decreased food waste and is used for discovering supply chain inefficiencies and ensuring food quality. Depending on the system architecture, different service qualities can be obtained in terms of usability, accuracy, security, etc. This article presents a novel approach for how to identify and select the most suitable system architectures for RSLP services. The approach is illustrated by ranking different architectures for a RSLP service directed towards the supply chain managers. As a proof of concept, some of the most highly ranked architectures have been implemented and tested in food cold supply chains.

With advances in Internet of Things many opportunities arise if the challenges of continual learning in a multimodal setting can be tackled. One common issue in Online Learning is to obtain labelled data, as this generally is costly. Active Learning is a popular approach to collect labelled data efficiently, but in general includes unrealistic assumptions. In this work we present a first step towards a taxonomy of Interactive Learning strategies in a multimodal and dynamic setting. By relaxing assumptions of standard Active Learning, the strategies become better suited for real-world settings and can achieve better performance.

A set of important criteria to consider when evaluating potential road user charging system (RUCS) are identified. These criteria are grouped into five categories: charging precision, system costs & societal benefits, flexibility & modifiability, operational aspects, and security & privacy. The criteria are then used in a comparative analysis of five RUCS candidates for heavy goods vehicles. Two solutions are position-based systems and one is based on tachographs. The two remaining solutions are based on fuel taxes. For each of the solutions we estimate how well it fulfils each of the criteria. One way of making general comparisons of the approaches is to give each of the criteria a specific weight corresponding to how important it is. We show that these weights heavily influence the outcome of the comparison. We conclude by pointing out a number of important issues needing attention in the process of developing RUCS.

Recent proliferation of surveillance systems is mostly attributed to advances in both image-processing techniques and hardware enhancement of smart cameras, as well as the ubiquity of sensor-driven architectures. Owing to these capabilities, new aspects are coming to the forefront. This paper addresses the current state-of-the-art and provides researchers with an overview of existing surveillance solutions, analyzing their properties as a system and drawing attention to relevant challenges when developing, deploying and managing them. Also, some of the more prominent application domains are highlighted here. In an effort to understand the development of the advanced solutions, based on their most distinctive characteristics, we propose a taxonomy for surveillance systems to help classify them and reveal gaps in existing research. We conclude by identifying promising future research lines.

The increasing presence of heterogeneous Internet of Things devices inside the home brings with it added convenience and value to the householders. At the same time, these devices tend to be Internet-connected and continuously monitor and collect data about the residents and their daily lifestyle activities. Such data can be of a sensitive nature, given that the house is the place where privacy is naturally expected. To gain insight into this state of affairs, we empirically investigate the privacy policies of 87 different categories of commercial smart home devices in terms of data being collected. This is done using a combination of manual and data mining techniques. The overall contribution of this work is a model that identifies and categorizes smart connected home data in terms of its collection mode, collection method, and collection phase. Our findings bring up several implications for smart connected home privacy, which include the need for better security controls to safeguard the privacy of the householders.

The rapidly evolving Internet of Things (IoT) includes applications which might generate a huge amount of data, this requires appropriate platforms and support methods. Cloud computing offers attractive computational and storage solutions to cope with these issues. However, sending to centralized servers all the data generated at the edge of the network causes latency, energy consumption, and high bandwidth demand. Performing some computations at the edge of the network, known as Edge computing, and using a hybrid Edge-Cloud architecture can help addressing these challenges. While such architecture may provide new opportunities to distribute IoT applications, making optimal decisions regarding where to deploy the different application components is not an easy and straightforward task for designers. Supporting designers’ decisions by considering key quality attributes impacting them in an Edge-Cloud architecture has not been investigated yet. In this paper, we: explore the importance of decision support for the designers, discuss how different attributes impact the decisions, and describe the required steps toward a decision support framework for IoT application designers.