Ebook: Intelligent Environments 2021

Intelligent Environments (IEs) combine physical spaces with ICT and pervasive technology to increase the users’ awareness of their surroundings, empower them to carry out their tasks, enrich their experience, and enhance their ability to manage such environments. IEs’ growing community, from academia to practitioners, is working on bringing IEs to life. Their work is driven by innovative ideas and technological progress, which makes it possible to place an increasing number of sensors and computing devices in IE in an affordable and energy-efficient manner.

The 17th International Conference on Intelligent Environments is taking place in the challenging times of COVID-19. ICT researchers and engineers have mostly been fortunate in that the pandemic has not significantly impacted their ability to work. Due to social distancing measures, a lot of human activity has moved online and to virtual environments, which is actually stimulating for ICT development. On one hand, this makes conferences such as ours more accessible as it lowers the cost and removes the need to travel. On the other hand, it diminishes the experience. An important aspect of conferences is engaging in informal discussions and sharing ideas in corridors, over a coffee or meal. Despite some efforts, this does not (yet) translate well to virtual environments. This was particularly felt by the workshops at this year’s conference, which are unfortunately fewer than in previous years. Nevertheless, we are pleased to include the proceedings of the following workshops, which emphasize multi-disciplinary and transversal aspects of IEs, as well as cutting-edge topics:

• 10th International Workshop on the Reliability of Intelligent Environments (WoRIE’21)

• 3rd International Workshop on Intelligent Environments and Buildings (IEB’21)

• 1st International Workshop on Self- Learning in Intelligent Environments (SeLIE ’21)

• 1st International Workshop on Artificial Intelligence and Machine Learning for Emerging Topics (ALLEGET ’21)

The proceedings contain contributions reflecting the latest research development in IEs and related areas, focusing on pushing the boundaries of the current state of the art and contributing to an establishment of IEs in the real world. We would like to thank all the contributing authors, as well as the members of the organizing committees and program committees of the workshops for their highly valuable work, which contributed to the success of the Intelligent Environments 2021 event. We are grateful to our technical sponsors, the IEEE Systems Man & Cybernetics Society, IOS Press and the Association for the Advancement of Artificial Intelligence.

The Editors

April 2021

In many experimental research projects on e-health that make use of smart technologies (like Internet of Things and Artificial Intelligence), it is common (and good) practice to organize experimental pilot studies aimed at validating the proposed protocols and solutions. Such an experimentation however, is at the boundary of different disciplines, involve humans and it typically addresses a wide range of stakeholders (including medical specialists and patients). Furthermore, it is also dependent on several user-related aspects like acceptance, usability, ethics and on the specific common practices of the applicative domain.

It is clear that this scenario poses several challenges that go beyond the typical challenges of experimentations that just focus on technology. The collection of requirements, for example, needs to address a variety of stakeholders, and in some cases the interaction with some of them must be mediated. The development of the required technology itself may be constrained by ethical requirements, and the planning of the pilot should also address the validation of the medical practice in the use of the developed technology and not just the validation of the technology itself, a fact that makes the technological design blurred with the development of the medical practice. Finally, the pilot themselves are exposed to human factors due to the involvement of end users. All these aspects may, in the end, impact on the reliability of the entire pilot.

Taking inspiration from a recently concluded project comprising a relatively long experimental pilot that involved technologies developed on-purpose, humans and medical protocols, this keynote discusses the reliability issues arose during the pilots and their impact on the final evaluation of the project results.

It is very hard (or ineffective) to take an old system and add to it security features like plug-ins. Therefore, a computer system is much more reliable designed with the approach of security-by-design. Nowadays, there are several tools, middlewares, and platforms designed with this concept in mind, but they must be appropriately used to guarantee a suitable level of reliability and safety. A security-by-design approach is fundamental when creating a distributed application in the IoT field, composed of sensors, actuators, and cloud services. The IoT usually requires handling different programming languages and technologies in which a developer might not be very expert.

Through a use case, we analyzed the security of some IoT components of Amazon Web Services (AWS) from a novice programmer’s point of view. Even if such a platform could be secure by itself, a novice programmer could do something wrong and leave some possible attack points to a malicious user.

To this end, we also surveyed a small pool of novice IoT programmers from a consulting engineering company. Even if we discovered that AWS seems quite robust, we noticed that some common security concepts are often not clear or applied, leaving the door open to possible issues.

It is important to be able to judge the performance or dependability metrics of a system and often we do so by using abstract models even when the system is in the conceptual phase. Evaluating a system by performing measurements can have a high temporal and/or financial cost, which may not be feasible. Mathematical models can provide estimates about system behavior and we need tools supporting different types of formalisms in order to compute desired metrics. The Mercury tool enables a range of models to be created and evaluated for supporting performance and dependability evaluations, such as reliability block diagrams (RBDs), dynamic RBDs (DRBDs), fault trees (FTs), stochastic Petri nets (SPNs), continuous and discrete-time Markov chains (CTMCs and DTMCs), as well as energy flow models (EFMs). In this paper, we introduce recent enhancements to Mercury, namely new SPN simulators, support to prioritized timed transitions, sensitivity analysis evaluation, several improvements to the usability of the tool, and support to DTMC and FT formalisms.

Fuzzy control systems are widely used to carry out control actions where variables with a certain degree of ambiguity are involved. Such control actions are modelled based on fuzzy rules that are expressed in a language analogous to that used by humans to conceive their reality. In this sense, fuzzy controllers can be applied in emerging technologies such as the Internet of Things (IoT) aimed at controlling the real world through a set of objects interconnected to the Internet. This article presents a fuzzy controller of Mamdani-type developed in Matlab. This controller is integrated into a smart home– created by the OpenHAB tool– using RESTful services. The results of the evaluation reveal that the integration of the fuzzy controller with smart home can provide an accurate control of comfort specifically in this case. In addition, a better organization of the control rules was also possible using the proposed fuzzy controller especially when large systems are developed.

As the complexity of intelligent environments grows, there is a need for more sophisticated and flexible interfaces. Conversational systems constitute a very interesting alternative to ease the users’ workload when interacting with such environments, as they can operate them in natural language. A number of commercial toolkits for their implementation have appeared recently. However, these are usually tailored to specific implementations of the processes involved for processing the user’s utterance and generate the system response. In this paper, we present a modular architecture to develop conversational systems by means of a plug-and-play paradigm that allows the integration of developers’ specific implementations and commercial utilities under different configurations that can be adapted to the specific requirements for each system.

Buildings consume a significant amount of energy worldwide in maintaining comfort for occupants. Building energy management systems (BEMS) are employed to ensure that the energy consumed is used efficiently. However these systems often do not adequately perform in minimising energy use. This is due to a number of reasons, including poor configuration or a lack of information such as being able to anticipate changes in weather conditions. We are now at the stage that building behaviour can be simulated, whereby simulation tools can be used to predict building conditions, and therefore enable buildings to use energy more efficiently, when integrated with BEMS. What is required though, is an accurate model of the building which can effectively represent the building processes, for building simulation. Building information modelling (BIM) is a relatively new method of representing building models, however there still remains the issue of data translation between a BIM and simulation model, which requires calibration with a measured set of data. If there a lack of information or a poor translation, a level of uncertaintly is introduced which can affect the simulation’s ability to accurate predict control strategies for BEMS. This paper explores effects of uncertainty, by making assumptions on a building model due to a lack of information. It will be shown that building model calibration as a method of addressing uncertainty is no substitute for a well defined model.

We are drowning in information while starving for wisdom said the two-time winner of the Pulitzer Prize Edward Osborne Wilson. We are attached 24 hours to our “smart” phone that gets us closer to those who are far away and keeps us apart from those who are near, and we want to live in “smart” buildings and “smart” cities where systems are used and data can be gathered to save energy, create comfortable ambiances, regulate traffic, or deal with our waste products, but we surely need to reconsider what “Smart” living is all about. From the roman empire to the actual high-rise buildings, through the modern movement leaded by Le Corbusier, each time has used the technology available, but neither of the great master pieces of architecture and urbanism is remembered by the technology that made it possible but by Its ability to seduce the minds, hearts and souls of its habitants and the generations that came after them. It is essential to know that we truly need, to clarify WHY and WHAT FOR we do things before we solve or engage ourselves in HOW we make it happen.

In this paper, we introduce a human-centric lighting control system optimized to support sleep and circadian coordination. We present an approach to optimize lighting that combines a “digital siblings” approach, i.e., a stochastic extension of a digital twin. It estimates and optimizes parameters in experimentally validated models of circadian and sleep regulation with a novel optimization algorithm for optimal timing of light exposure. We acknowledge that people have varying preferences for the lighting levels throughout the day based on their personal preferences and schedules and explicitly include this as a key parameter in our optimization strategy. Our results show that with the suggested lighting schedules, alignment of circadian rhythmicity to the desired sleep-wake schedule can be achieved with minimal disruption to people’s daily lives.

We highlight a lack of models and theories associated with the Smart Campus concept and also an absence of processes to support its design and development. This paper provides a first approach to a theory and a set of design principles to guide their development. The theory and principles are flexible enough to be easily adapted and adopted by any organization interested in developing a Smart Campus.

The SSHARE project will develop innovative self-sufficient envelope for buildings aimed at net zero energy, thereby contributing to the European technology. Envelope is a combination of two breaking through technologies: HUNTER-Humidity to Electricity Convertor and Advanced Radiant Panel for Buildings that will cool or heat the building, depending on the time of year, imitating perspiration of living beings and using only water as both thermal and electric energy supply. Successful realization of the project is assured by implementing a coordinated network of knowledge sharing in materials science, chemistry and mechanical engineering; by solidifying the state-of-the-art understanding in nanoelectronics and energy efficiency, and by applying bottom-up nanoengineering approaches via an international and inter-sector collaboration of highly qualified researchers from Portugal, Spain, Ukraine, Belarus, Tajikistan, Uzbekistan, Azerbaijan and the Joint Institute for Nuclear Research Russian Federation. Technological (panels fabrication) as well as fundamental (renewable energy) issues will be assessed by this multidisciplinary consortium. This paper explains the basis and principles for the development of a new generation of building materials and hence the creation of net zero building. Sharing the culture of research and innovation, the SSHARE project will allow applying recent advancements in nanotechnology science and mechanical engineering to address ““Plus Energy Houses”” EU 2050 concept.”.

This paper focuses on the use of computational tools to develop a data driven approach for an analytical study about different urban systems. This “framework” examines urban Big Data in the old medina of Sale in Morocco. The computational tools are more effective to provide insights within complexity, becoming a key to generate more efficient solutions throughout the design process. The findings of this study highlight the potential of a data driven approach to explore analytical aspects and move further to generative design using algorithms.

Managing mobility, both of people and goods, in cities is a thorny issue. The travel needs of urban populations are increasing and put pressure on transport infrastructure. The Moroccan cities are no exception and will struggle, in the short term, to respond to the challenges of the acceleration of the phenomenon of urbanization and the increase in demand for mobility. This will inevitably prevent them from turning into smart cities. The term smart certainly alludes to better use of technologies, but smart mobility is also defined as “a set of coordinated actions intended to improve the efficiency, effectiveness and environmental sustainability of cities” [1]. The term mobility highlights the preponderance of humans over infrastructure and vehicles. Faced with traffic congestion, the solutions currently adopted which consist of fitting out and widening the infrastructures, only encourage more trips and report the problem with more critical consequences. It is true that beyond a certain density of traffic, even Intelligent Transport Systems (ITS) are not useful. The concept of dynamic lane management or Advanced Traffic Management (ATM) opens up new perspectives. Its objective is to manage and optimize road traffic in a variable manner, in space and in time. This article is a summary of the development of a road infrastructure dedicated to Heavy Goods Vehicles (HGV), the first of its kind in Morocco. It aims to avoid the discomfort caused by trucks in the urban road network of the city of Casablanca. This research work is an opportunity to reflect on the introduction of ITS and ATM to ensure optimal use of existing infrastructure before embarking on heavy and irreversible infrastructure projects.

The urban, social, industrial and technological evolution in recent years has forced designers, entrepreneurs and public entities to rethink the conceptual and operational logic for the construction of new urban settlements, according to the authentic principles of mutual respect between man and nature, eliminating all thought extremism and purely conceptual theories. In order to make their respective professional and life skills available to future generations, we have set up two international working groups, the ISCW and IS.Smart, in collaboration with the UFC universities of Fortaleza (Brazil), UIR of Rabat (Morocco) and UNITO of Turin (Italy), which for some years now, have been working to transform the notions of smart cities and smart buildings, combined with the new industrial revolution 4.0 and tokenomics, into concrete, multidisciplinary and educational activities. This publication, the result of the work developed in the last three years, intends to be an easy-to-read and application tool for all operators who wish to approach the “smart world” in a professional manner, especially in situations with serious housing shortage and still in the process of social and economic development. By reporting experiences and methodologies already field tested by the author in some working circumstances and that have given rise to case studies, similar to the “Polo Multimodal Pecem” (Brazil), the reader will find useful insights such as the definition of the smart concept and “smart” objectives, the approach to the project, the definition of evaluation parameters, operational examples of management, capital raising and crowdfunding, tokenomics and e- conomy. I do believe that the future of the city will be increasingly linked to these issues, which have now become fundamental and necessary in the projection and planning processes of urban settlements.

Past research has mainly applied linear cointegration analysis to study the relationship of crude oil prices with the prices of other commodities. However, recent methodological innovations in cointegration analysis allow for a more thorough analysis of the co-movement of commodity prices and detect asymmetric and thresholds co-movements. Following Enders and Siklos [1] and Hansen and Seo [2], we apply threshold cointegration analysis, detecting co-movements that earlier studies based on linear cointegration analysis could not detect. We find that adjustments to positive and negative deviations from the long-run equilibrium are asymmetric for copper, food and agricultural raw materials in the short-run. Moreover, the adjustments for aluminum and nickel are symmetric. The price Granger causalities behave as expected for metals and agricultural raw material prices. Food prices, however, behave differently. In sum, the results of this paper underscore the importance of consistently testing nonlinear cointegration and point out the complex interactions that take place between the markets of oil and other commodities.

In our growing cities, climate change and energy related uncertainties are of great concern. The impact of the Urban Heat Island on comfort, health and the way we use energy still requires further clarification. The outdoor-indoor energy balance model (3D-City Irradiance) presented in this article was developed so as to address these issues. The effects of view factors between urban surfaces on three-dimensional radiation and the effects of fully integrated outdoor-indoor energy balance schemes on heat islands and building indoor thermal loads could be included within different building blocks at a resolution of several metres. The model operated under the ‘stand alone’ mode. It was tested using the Building Energy Simulation Test (BESTest) which demonstrated good levels of agreement for diurnal and seasonal simulations.

This paper proposes an algorithm for learning to move the desired object by humanoid robots. In this algorithm, the semantic segmentation algorithm and Deep Reinforcement Learning (DRL) algorithms are combined. The semantic segmentation algorithm is used to detect and recognize the object be moved. DRL algorithms are used at the walking and grasping steps. Deep Q Network (DQN) is used to walk towards the target object by means of the previously defined actions at the gate manager and the different head positions of the robot. Deep Deterministic Policy Gradient (DDPG) network is used for grasping by means of the continuous actions. The previously defined commands are finally assigned for the robot to stand up, turn left side and move forward together with the object. In the experimental setup, the Robotis-Op3 humanoid robot is used. The obtained results show that the proposed algorithm has successfully worked.

The treatment process at home after hospitalization may become challenging for elders and people having any physical or cognitive disability. Such patients can, nowadays, be supported by Autonomous and Intelligent Monitoring Systems (AIMSs) that may get new levels of functionalities thanks to technologies like Reinforcement Learning, Deep Learning and Internet of Things.

We present an AIMS that can assist impaired patients in taking medicines in accordance with their treatment plans. The demonstration of the AIMS via mobile app shows promising results and can improve the quality of healthcare at home.

Numerous e-news channels publish the daily happenings in the world from different sources. These huge amounts of news articles have lamentably conceived the information overload issue among the users. Hence text mining, which aims in extracting previously unknown information from unstructured text, has been widely used by several researchers to segregate full news articles however, the news headlines categorization is still specifically limited. Therefore, considering this limitation, the current research aims to propose a framework that will self-learn and automatically classify any given news headline into its corresponding news category using artificial intelligence methods i.e. text mining and machine learning algorithms. The proposed framework consists of three stages: Exploratory Data Analysis, Text Pre-processing, and Text Classification. For exploratory data analysis, the top 10 most frequent balanced news categories are chosen so that further processing of data can be done on a more balanced version of the dataset. After exploring the data, text pre-processing techniques are applied to make the data transformed, normalized, and structured. Finally, text classification is carried out with two approaches: unsupervised classification using Mean Shift and K-means algorithms and supervised classification using Logistic Regression with Bag of Words and TF-IDF algorithm. To depict the working of the proposed framework, a case study is presented on a news headlines dataset which accurately performed news headlines classification.