Magnus Karlsson

This study investigates federated deep learning for multi-horizon indoor climate forecasting in historic buildings. Unlike traditional centralized or isolated local learning approaches, this work explores federated learning (FL) as a solution that balances prediction accuracy with data privacy. By combining FL with deep learning, the study addresses challenges in distributed environments, including data heterogeneity and communication efficiency. Eight deep learning architectures and six FL algorithms were evaluated across three learning cases: local learning, centralized learning, and FL. The case study used several years of hourly data on indoor temperature, relative humidity, and CO₂ concentration from three public historic buildings, along with outdoor weather conditions. Time-series dense encoder (TiDE) and gated recurrent units (GRU) consistently outperformed other architectures, capturing temporal dependencies effectively across all cases. Among FL algorithms, federated averaging (FedAvg) and federated median (FedMedian) demonstrated the best overall performance, with fast convergence and low validation loss. A comparison of the three learning cases revealed distinct trade-offs. Centralized learning typically achieved the highest predictive accuracy but required full data aggregation, raising privacy concerns. Local learning ensured data privacy but performed worse due to limited data. FL offered a balance, enabling collaborative model training without sharing raw data. TiDE combined with FedAvg or FedMedian achieved accuracy within 5–10% of the centralized models, measured by the coefficient of variation of the root mean squared error (CV-RMSE), making it a promising approach for privacy-sensitive scenarios. The results highlight the potential of FL for accurate, privacy-preserving indoor climate forecasting. The proposed approach supports scalable, institution-level collaboration without compromising data confidentiality. It is especially relevant for facility managers and heritage conservationists who are seeking data-driven indoor climate control in sensitive heritage buildings.

This study showcases the digitalization of Löfstad Castle in Sweden to contribute to preserving its heritage values. The castle and its collections are deteriorating due to an inappropriate indoor climate. To address this, thirteen cloud-connected sensor boxes, equipped with 84 sensors, were installed throughout the main building, from the basement to the attic, to continuously monitor various indoor environmental parameters. The collected extensive multi-parametric data form the basis for creating a parametric digital twin of the building. The digital twin and detailed data analytics offer a deeper understanding of indoor climate and guide the adoption of appropriate heating and ventilation strategies. The results revealed the need to address high humidity problems in the basement and on the ground floor, such as installing vapor barriers. Opportunities for adopting energy-efficient heating and ventilation strategies on the upper floors were also highlighted. The digitalization solution and findings are not only applicable to Löfstad Castle but also provide valuable guidance for the conservation of other historic buildings facing similar challenges.

Digital transformation in the built environment generates vast data for developing data-driven models to optimize building operations. This study presents an integrated solution utilizing edge computing, digital twins, and deep learning to enhance the understanding of climate in buildings. Parametric digital twins, created using an ontology, ensure consistent data representation across diverse service systems equipped by different buildings. Based on created digital twins and collected data, deep learning methods are employed to develop predictive models for identifying patterns in indoor climate and providing insights. Both the parametric digital twin and deep learning models are deployed on edge for low latency and privacy compliance. As a demonstration, a case study was conducted in a historic building in Östergötland, Sweden, to compare the performance of five deep learning architectures. The results indicate that the time-series dense encoder model exhibited strong competitiveness in performing multi-horizon forecasts of indoor temperature and relative humidity with low computational costs.

Digital transformation in buildings accumulates massive operational data, which calls for smart solutions to utilize these data to improve energy performance. This study has proposed a solution, namely Deep Energy Twin, for integrating deep learning and digital twins to better understand building energy use and identify the potential for improving energy efficiency. Ontology was adopted to create parametric digital twins to provide consistency of data format across different systems in a building. Based on created digital twins and collected data, deep learning methods were used for performing data analytics to identify patterns and provide insights for energy optimization. As a demonstration, a case study was conducted in a public historic building in Norrköping, Sweden, to compare the performance of state-of-the-art deep learning architectures in building energy forecasting.

Historic buildings require good maintenance to sustain their function and preserve embodied heritage values. Previous studies have demonstrated the benefits of digitalization techniques in improving maintenance and managing threats to historic buildings. However, there still lacks a solution that can consistently organize data collected from historic buildings to reveal operating conditions of historic buildings in real-time and to facilitate various data analytics and simulations. This study aims to provide such a solution to help achieve preventive conservation. The proposed solution integrates Internet of Things and ontology to create digital twins of historic buildings. Internet of Things enables revealing the latest status of historic buildings, while ontology provides a consistent data schema for representing historic buildings. This study also gives a reference implementation by using public cloud services and open-source software libraries, which make it easier to be reused in other historic buildings. To verify the feasibility of the solution, we conducted a case study in the City Theatre, Norrköping, Sweden. The obtained results demonstrate the advantages of digital twins in providing maintenance knowledge and identifying potential risks caused by fluctuations of relative humidity.

Plant Factory is a newly emerging industry aiming at transforming crop production to an unprecedented model by leveraging industrial automation and informatics. However, today’s plant factory and vertical farming industry are still in a primitive phase, and existing industrial cyber-physical systems are not optimal for a plant factory due to diverse application requirements on communication, computing and artificial intelligence. In this paper, we review use cases and requirements for future plant factories, and then dedicate an architecture that incorporates the communication and computing domains to plant factories with a preliminary proof-of-concept, which has been validated by both academic and industrial practices. We also call for a holistic co-design methodology that crosses the boundaries of communication, computing and artificial intelligence disciplines to guarantee the completeness of solution design and to speed up engineering implementation of plant factories and other industries sharing the same demands.

Researches on the Internet of Things (IoT) and cloud computing have been pervasive in both the academic and industrial world. IoT and cloud computing are seen as cornerstones to digital transformation in the industry. However, restricted by limited resources and the lack of expertise in information and communication technologies, small- and medium-sized enterprises (SMEs) have difficulty in achieving digitalization of their business. In this paper, we propose a reference framework for SMEs to follow as a guideline in the journey of digital transformation. The framework features a three-stage procedure that covers business, technology, and innovation, which can be iterated to drive product and business development. A case study about digital transformation taking place in the vertical plant wall industry is detailed. Furthermore, some solution design principles that are concluded from real industrial practice are presented. This paper reviews the digital transformation practice in the vertical plant wall industry and aims to accelerate the pace of SMEs in the journey of digital transformation.

Monitoring the indoor environment of historic buildings helps to identify potential risks, provide guidelines for improving regular maintenance, and preserve cultural artifacts. However, most of the existing monitoring systems proposed for historic buildings are not for general digitization purposes that provide data for smart services employing, e.g., artificial intelligence with machine learning. In addition, considering that preserving historic buildings is a long-term process that demands preventive maintenance, a monitoring system requires stable and scalable storage and computing resources. In this paper, a digitalization framework is proposed for smart preservation of historic buildings. A sensing system following the architecture of this framework is implemented by integrating various advanced digitalization techniques, such as Internet of Things, Edge computing, and Cloud computing. The sensing system realizes remote data collection, enables viewing real-time and historical data, and provides the capability for performing real-time analysis to achieve preventive maintenance of historic buildings in future research. Field testing results show that the implemented sensing system has a 2% end-to-end loss rate for collecting data samples and the loss rate can be decreased to 0.3%. The low loss rate indicates that the proposed sensing system has high stability and meets the requirements for long-term monitoring of historic buildings.

Today, the edge-cloud computing paradigm starts to gain increasing popularity, aiming to enable short latency, fast decision-making and intelligence at the network edge, especially for industrial applications. The container-based virtualization technology has been put on the roadmap by the industry to implement edge-cloud computing infrastructures. Has the performance of the container-based edge-cloud computing stacks reached industry requirement? In this paper, from the industrial client perspective, we provide a performance evaluation methodology and apply it to the state-of-the-art containerization-based edge-cloud computing infrastructures. The influences of the message sending interval, payload, network bandwidth and concurrent devices on full stack latency are measured, and the processing capability of executing machine learning tasks are benchmarked. The results show that containerization on the edge does not introduce noticeable performance degradation in terms of communication, computing and intelligence capabilities, making it a promising technology for the edge-cloud computing paradigm. However, there is a large room for performance improvement between current implementation of the edge-cloud infrastructure and the demanding requirements anticipated by time-critical industrial applications. We also emphasize and showcase that partitioning of an industrial application into microservices throughout the whole stack can be considered during solution design. The proposed evaluation methodology can be a reference to users of edge-cloud computing as well as developers to get a client perspective overview of system performance.