Browsing by Author "Främling, Kary, Prof., Aalto University, Department of Computer Science, Finland"

Filter results by typing the first few letters
Now showing 1 - 2 of 2
  • Moving towards data-driven decision-making in maintenance
    (2019) Madhikermi, Manik
     School of Science | Doctoral dissertation (article-based)
    Traditionally, in the maintenance industry, maintenance efficiency is limited by the capability of the experts making the decision. However, the advancement of digital technologies made it possible to improve the effectiveness and efficiency of maintenance activities by adding insight from the data to expert assessment. The opportunity provided by data for decision making made the companies to shift towards a new type of maintenance strategy called data-driven maintenance. Despite of opportunities, data and analytical tools' companies are still struggling to fully harness data asset to improve maintenance activities because of data-centric challenges. Hence, the main objective of this dissertation is to identify and mitigate those challenges that limit organizational decision-making capabilities to improve maintenance effectiveness. In this dissertation, firstly, quantitative and descriptive analyses of case studies in Finnish Multinational Manufacturing Companies have been carried out to identify key data-centric challenges. The study identified Data Quality, Interoperability, and Data extraction as key challenges. Furthermore, each of the identified challenges have been investigated through one or more original publications. The main results achieved in this dissertation are methods and frameworks to i) assess and compare data quality of maintenance  reporting procedure ii) two-level interoperability framework for inter-system interoperability iii) data discovery methodology to extract data for Extract, Transform and Load process. The applicability and validity of each of the proposed methodologies and framework has been validated through one or multiple use cases. For validation, three different tools namely, MRQA Dashboard, Open-messaging Middleware, and Data Model Logger have been developed to tackle each of the identified data-centric challenges.
  • A Scalable and Fault-Tolerant IoT Architecture for Smart City Environments
    (2022) Javed, Asad
     School of Science | Doctoral dissertation (article-based)
    The Internet of Things (IoT) technology is becoming a promising computing infrastructure for the future developments of smart, connected environments. IoT has provided a digital world in which real-life objects are linked with other smart systems to facilitate rapid communication. This vision is spreading rapidly to various domains, ranging from smart homes to smart cities and industrial automation. One of the core aspects of IoT resides in the seamless connectivity and accessibility of heterogeneous devices over the network. Such an interaction is made possible by the adoption of much needed IoT protocols that offer open and standardized interfaces. A smart city is one of the examples in which devices, gateways, stakeholders, service providers, and computing platforms are fully connected to provide a ubiquitous environment. This communication involves the integration of two computing paradigms, namely edge and cloud, in order to achieve the full potential of IoT in smart cities. Their integration also demands fault tolerance, which might be possible with different software stacks on both the edge and cloud. However, it is more convenient to employ the same software stacks to ensure unified fault-tolerant management. Also, the large-scale data processing in smart cities has led to an increase in data volume, variety, and velocity. As a consequence, IoT-based systems need to process, manage, and store a large amount of real-time data on the edge, i.e., closer to the data sources, to minimize latency and save network bandwidth. This research investigates the scalability and fault tolerance of IoT applications in smart cities. The overall objective is to ensure that the smart city applications are resilient to failures and scale based on the increasing demands of users. The objective is pursued through three research questions, which identify some of the most significant challenges in smart cities: (i) How can IoT messaging standards enable real-time device-generated data processing and discovery? (ii) What is the role of edge computing in enabling scalable computation in dynamically changing environments? (iii) How can edge and cloud computing collectively provide fault tolerance capabilities? Each of the identified barriers is addressed through novel techniques presented in the research publications. Finally, the proposed solutions are validated by implementing them in various real-life case studies. The results indicate that, by adopting edge and cloud computing technologies, the proposed solutions are able to provide scalability, optimize network latency, and handle hardware- and network-based failures.