CoE in Networked Sensing, Inference and Learning (NETSENSE)

Extracting information through sensing the physical world is of high and rapidly increasing importance in various every-day and also more specialized applications building the Internet-of-Things (IoT).  Those include intelligent transport systems, robotics and autonomous machines, health and well-being, human and gesture detection, gas and material detection, environment imaging, surveillance and security. The current paradigm for sensing and data processing is to collect raw data from sensors and then process data in the cloud using artificial intelligence (AI) tools, especially machine learning (ML) as well as statistical inference or data analytics. While the AI/ML methods are general purpose and easy to use, they have severe limitations when it comes to scalability, predictability , energy efficiency and sustainability, security, and privacy. Furthermore, existing approaches put large burden on the utilized communications infrastructure.

 

In NETSENSE, Tampere University, Aalto University, and University of Oulu combine their forces to form a world-class multi-disciplinary research entity, where physical world meets information. We will overcome the state-of-the-art limitations by conducting break-through research in networked sensing, inference, and learning. We will develop methods that take the physical world limitations and hardware constraints into account; they will scale from small data to big data and use sensing, network, and computational resources in an energy efficient and sustainable way. They will have provable performance, provide security, and maintain privacy. We apply the rigorous research tradition of electrical engineering, together with computational sciences, and combine the wide expertise of the participating principal investigators that spans from sensor hardware to data science, comprising signal processing, AI/ML, wireless networks and radio frequency circuits and antennas, with significant synergy benefits.

 

Our broad research vision and objective is to create new technological and scientific breakthroughs to facilitate and enable future sustainable smart industries and society through networked intelligence, sensing, inference, and trust. Direct applications include, but are not limited to: smart traffic and intelligent transport systems, elderly care, proximity-based disease control and health monitoring; energy and resource efficiency in industry, asset tracking and autonomous robots, public safety and security via privacy-preserving surveillance and monitoring of rapid changes, anomalies and adversarial events.