Linkedin

Machine learning tools for improving data quality and dependability in IoT applications

Date: 05/11/2019

1. Project Title/ Job position title: Machine learning tools for improving data quality and dependability in IoT applications

2. Area of knowledge: Physical Sciences, Mathematics and Engineering Panel

3. Group of disciplines: Theoretical and Applied Mathematics, Computer Sciences and IT

4. Research project/ Research Group description:
Current scalable systems and processes produce increasing amounts of data generated by an ever-growing number of sensors and activities. To handle these large amounts of data, Internet-of-Things (IoT) platforms can be used to efficiently take care of automating several processes, from collecting to storing and providing access to these data.
However, these platforms neglect quality assurance mechanisms to avoid data quality degradation, e.g., due to sensor faults causing drift or outliers, time-variability of processes, communication failures, to mention a few. Additionally, they do not leverage from statistical and machine learning tools to go beyond the provision of raw data to provide meaningful insights on the system or process features, e.g., forecasts, pattern matching, or event classification, thus benefitting decision-making procedures and services that depend on the data.
This proposal aims to develop a scalable generic framework and configurable platform for data dependability and knowledge extraction on IoT contexts, clearly separating generalizable methodologies from mechanisms to ease configuration and adaptation to specific application fields. Case studies on energy efficiency and flexibility management in buildings and environmental monitoring will demonstrate the generic nature and the adaptability of the proposed framework and platform.

5. Job position description:
The student will be involved in the various tasks required for developing the generic framework, including the design of the underlying machine-learning-based solutions for failure detection and data processing, the definition of an architectural solution enabling the deployment of these solutions in multiple application scenarios, the definition of methods to ease configuration tasks, and the implementation and validation of these solutions and methods in the scope of multiple use cases.
The project is developed with members of the Navigators group of the LASIGE research lab. Several members of the group (and lab) are involved in research activities aimed at achieving increased dependability, adaptability and performance, with fruitful and outstanding results in the past. In particular, the project is aligned with the goals of the AQUAMON project (PTDC/CCI-COM/30142/2017 from Fundação para a Ciência e a Tecnologia), aimed at developing a platform for dependable monitoring with WSNs in water environments, and the CSESI Hub, a collaborative laboratory on Smart Energy Services Innovation Hub, also financed by the Fundação para a Ciência e a Tecnologia.

6. Group leader:
* title: Professors
* full name: António Casimiro and Pedro Ferreira
* email: casim@ciencias.ulisboa.pt and pmf@ciencias.ulisboa.pt
* research project/ research group website: http://navigators.di.fc.ul.pt
* website description: the site belongs to one of the groups of the LASIGE Lab (www.lasige.pt)