MARVEL: Multimodal Extreme-Scale Data Analytics for Smart Cities Environments
A virtual MARVEL kick-off meeting was held on 18-19 January 2021. Fifty participants from seventeen partner organisations and the Project Officer representing the EC discussed roles, ways of working, quality assurance process, ethics compliance and financial reporting. The MARVEL project is specifically aimed at utilizing multimodal perception and intelligence for smart cities environment. The project is coordinated by the Foundation for Research and Technology Hellas (FORTH).
Multimodal Perception & Intelligence for Smart Cities
As the urbanization is at rampant growth, cities are faced with the challenges of improving air and water quality, reducing noise pollution, waste management, scarcity of resources, traffic congestions, coping with extreme weathers, and other issues related to growth, performance, competitiveness and citizens’ livelihood. To address these issues, city managements are becoming increasingly interested in smart city environments as one of the possible solutions.
A smart city environment makes use of the inter-connectivity of technological devices to improve quality of life by taking better decisions pertaining to citizens’ health, safety, time and living standards. With the growth of Internet of Things, a large number of variants of sensing devices are becoming ubiquitous, producing multi-source and multi-modal data.
The smart city environment comprises a complex and dynamic system in which the inter-connected devices do not necessarily share common features in terms of hardware/software specifications, connectivity and the nature of data generated. This multimodal perception of the environment produces heterogeneous datasets with the potential of multiple aspects of analytics using artificial intelligence including audio-visual scene detection, event recognition (such as fire, gunshot, car accident, anomaly, etc.), automated surveillance, healthcare services, smart utility management, traffic management, law enforcement, tracking suspected individuals, and providing contextual awareness in a smart city environment to improve the quality of life and services to citizens.
Issues & Challenges
- Collecting, fusing, analysing and data mining datasets with varying features produced by the heterogeneous devices
- Providing accurate, real-time predictions with the constraints of sheer volume of data and the high complexity and dynamicity of the environment
- Information security, management and processing of personal data while preserving privacy, and ensuring ethical limits
- Coping the individual issues of edge, fog and cloud computing in the form an integrated edge-fog-cloud computing paradigm
What MARVEL can do
MARVEL stands apart from the traditional big data architectures that mainly rely on distributed resources and heterogeneous data sources in smart city environments. Contrary to the existing computing and storage frameworks that rely on a single paradigm (edge, fog or cloud), MARVEL is aimed at introducing an integrated edge-fog-cloud pervasive computing framework for enabling the aforementioned tasks in a smart city environment.
The data collected from multimodal data sources and analysed through machine learning algorithms will be used for real-time decision making and triggering appropriate actions to deliver enhanced services to the citizens. The extreme-scale analytics of the big data will be used to capture the vibe of the city and provide insights and predictive capability of major outdoor events. An important aspect of MARVEL will be to ensure privacy, ethical and legal limits while delivering the intended performance.
Another aspect of MARVEL is to cope with the contemporary challenges in big data pertaining to data acquisition, analysis, processing, storage and visualization. This will help generate new knowledge and lead to technological advancements. At the same time, it will pave the way for empowering smart city authorities to better support their societies.
Role of Privanova
A smart city environment is exposed to threats originating from information security, data privacy and cyber-related factors where unauthorised access to information can cause undesired consequences. At the same time, there are also social and environmental aspects which create concerns regarding ethics.
Privanova supports the project by devising efficient, innovative privacy preserving mechanisms, Data Protection Management strategies and ethical compliance framework. Within MARVEL, Privanova assists the coordinator in achieving and maintaining the ethics compliance by addressing the Conditional Clearance and Ethics Requirements the project received following a successful scientific evaluation.
In addition, Privanova will develop a privacy-by-design framework for data protection by transposing the GDPR rules into meaningful, actionable items which all technical partners can understand and act upon. This will further help to protect the privacy of the individuals involved in research activities in the project, especially during its experimentation and piloting phase.
At the same time, relying on our experience from other EU-funded, IoT-related projects where Privanova takes part such as the IoT-NGIN or CYRENE, Privanova will generate useful insights for technology developers to help implement relevant privacy safeguards from the outset. Finally, Privanova will use its connections to several EU flagship initiatives such as Living-in.EU which focuses on digital transformation in the European cities to promote the project outcomes.