Generalized global solar radiation forecasting model via cyber-secure deep federated learning.

Recently, the increasing prevalence of solar energy in power and energy systems around the world has dramatically increased the importance of accurately predicting solar irradiance. However, the lack of access to data in many regions and the privacy concerns that can arise when collecting and transmitting data from distributed points to a central server pose challenges to current predictive techniques. This study proposes a global solar radiation forecasting approach based on federated learning (FL) and convolutional neural network (CNN). In addition to maintaining input data privacy, the proposed procedure can also be used as a global supermodel. In this paper, data related to eight regions of Iran with different climatic features are considered as CNN input for network training in each client. To test the effectiveness of the global supermodel, data related to three new regions of Iran named Abadeh, Jarqavieh, and Arak are used. It can be seen that the global forecasting supermodel was able to forecast solar radiation for Abadeh, Jarqavieh, and Arak regions with 95%, 92%, and 90% accuracy coefficients, respectively. Finally, in a comparative scenario, various conventional machine learning and deep learning models are employed to forecast solar radiation in each of the study regions. The results of the above approaches are compared and evaluated with the results of the proposed FL-based method. The results show that, since no training data were available from regions of Abadeh, Jarqavieh, and Arak, the conventional methods were not able to forecast solar radiation in these regions. This evaluation confirms the high ability of the presented FL approach to make acceptable predictions while preserving privacy and eliminating model reliance on training data.

A hover view over effectual approaches on pandemic management for sustainable cities - The endowment of prospective technologies with revitalization strategies.

The COVID-19 pandemic affects all of society and hinders day-to-day activities from a straightforward perspective. The pandemic has an influential impact on almost everything and the characteristics of the pandemic remain unclear. This ultimately leads to ineffective strategic planning to manage the pandemic. This study aims to elucidate the typical pandemic characteristics in line with various temporal phases and its associated measures that proved effective in controlling the pandemic. Besides, an insight into diverse country's approaches towards pandemic and their consequences is provided in brief. Understanding the role of technologies in supporting humanity gives new perspectives to effectively manage the pandemic. Such role of technologies is expressed from the viewpoint of seamless connectivity, rapid communication, mobility, technological influence in healthcare, digitalization influence, surveillance and security, Artificial Intelligence (AI), and Internet of Things (IoT). Furthermore, some insightful scenarios are framed where the full-fledged implementation of technologies is assumed, and the reflected pandemic impacts in such scenarios are analyzed. The framed scenarios revolve around the digitalized energy sector, an enhanced supply chain system with effective customer-retailer relationships to support the city during the pandemic scenario, and an advanced tracking system for containing virus spread. The study is further extended to frame revitalization strategies to highlight the expertise where significant attention needs to be provided in the post-pandemic period as well as to nurture sustainable development. Finally, the current pandemic scenario is analyzed in terms of occurred changes and is mapped into SWOT factors. Using Fuzzy Technique for Order of Preference by Similarity to Ideal Solution based Multi-Criteria Decision Analysis, these SWOT factors are analyzed to determine where prioritized efforts are needed to focus so as to traverse towards sustainable cities. The results indicate that the enhanced crisis management ability and situational need to restructure the economic model emerges to be the most-significant SWOT factor that can ultimately support humanity for making the cities sustainable.

Prediction of energy expenditure during activities of daily living by a wearable set of inertial sensors.

Physical inactivity is responsible for 7-10% of all premature deaths worldwide. Thus, valid, reliable and unobtrusive methods for monitoring activities of daily living (ADL) to predict total energy expenditure (TEE) is desired. Multiple methods exist to quantify TEE, but microelectromechanical systems (MEMSs) are the only method, which has shown promising results and are applicable for long-term monitoring in the field. However, no perfect method exists for predicting TEE on a daily basis. The present study evaluates TEE estimation based on a MEMS (Xsens Link system) taking gender and heart rate into account. Fifteen individuals performed seven ADL wearing the Xsens Link system, a heart rate belt and an oxygen mask. Multiple linear regression models were established for sedentary and dynamic activities and evaluated by leave-one-out cross-validation and compared with indirect calorimetry. The linear regression model showed better prediction for dynamic activities (adjusted R2 0.95±0.16) compared to sedentary activities (adjusted R2 0.61±0.19). The root-mean-square error for the TEE estimation ranged between 0.02 and 0.08 kJ/min/kg for the sedentary and dynamic models, respectively. The study showed a viable approach to predict TEE in ADL compared to previously published results. Further studies are warranted to reduce the number of sensors in the estimation of TEE.

Integrated Management of Energy, Wellbeing and Health in the Next Generation of Smart Homes.

This contribution proposes an implementation for next generation smart homes, where heterogeneous data, coming from multiple sensors (medical, wellbeing, energy, contextual, etc.) and house equipment (smart fridge, smart TV, etc.), need to be managed, secured and visualized. As a first step, it focuses only on energy and health data. However, it aims to lay the foundations to manage any type of information towards the development of smart interactions with the house, which might include artificial intelligence and machine learning. These data are securely collected using a central Web of Things gateway, located inside the smart home. For the e-health part, a set of possible use-cases is provided, along with the current progress of the implantation. In this regard, the main idea is to link the next generation smart homes with external medical entities in order to provide, first, quick intervention in the event of an abnormality being detected, and to be able to provide basic medical services such as remote consultations with a doctor for a particular health issue. This vision can be very promising, particularly in rural areas, where access to medical services is difficult. As for the energy part, the aim is to collect users' energy consumption inside the smart home, which can be supplied from different sources (heat, water, gas, or electricity), and to enable the use of advanced algorithms to predict and manage local energy consumption and production (if any). This approach combines data collected from smart meters, operational information of the smart energy devices (the status of smart plugs), user's requests and external network signals such as energy prices. By using a home energy management system that accepts such input parameters, the operation of in-home devices and appliances can be optimally controlled according to different objectives (e.g., minimizing energy costs and maximizing user's comfort level).

Multi-Residential Activity Labelling in Smart Homes with Wearable Tags Using BLE Technology.

Smart home platforms show promising outcomes to provide a better quality of life for residents in their homes. One of the main challenges that exists with these platforms in multi-residential houses is activity labeling. As most of the activity sensors do not provide any information regarding the identity of the person who triggers them, it is difficult to label the sensor events in multi-residential smart homes. To deal with this challenge, individual localization in different areas can be a promising solution. The localization information can be used to automatically label the activity sensor data to individuals. Bluetooth low energy (BLE) is a promising technology for this application due to how easy it is to implement and its low energy footprint. In this approach, individuals wear a tag that broadcasts its unique identity (ID) in certain time intervals, while fixed scanners listen to the broadcasting packet to localize the tag and the individual. However, the localization accuracy of this method depends greatly on different settings of broadcasting signal strength, and the time interval of BLE tags. To achieve the best localization accuracy, this paper studies the impacts of different advertising time intervals and power levels, and proposes an efficient and applicable algorithm to select optimal value settings of BLE sensors. Moreover, it proposes an automatic activity labeling method, through integrating BLE localization information and ambient sensor data. The applicability and effectiveness of the proposed structure is also demonstrated in a real multi-resident smart home scenario.