Research topics and application areas
The Pervasive Systems research group has strong expertise on various themes of Internet of Things and Pervasive Computing, from wireless networking to sensor data analytics, distributed services, sensor data analytics. The research themes are focused on the design and analysis of the following topics and their interaction:
- Massive Smart Sensing: Collaborative embedded and opportunistic sensing
- Edge networking: Wireless and opportunistic networking mainly at the edge of the system
- Sensor data analytics: Spatio-temporal sensor data analytics and AI
Most important enabling technologies that have emerged to make pervasive computing and the Internet of Things vision a reality are: wireless networking, localization, distributed systems, mobile computing, smart networked sensors, embedded platforms, sensor data analytics and reasoning, etc. They have to deal with ever-changing user requirements, environmental situations and system resources. They have to be privacy aware and secure to encourage wide use and to provide the most optimal services.
Within the scope of Pervasive Systems research, the following topics are part of our work.
- Joint communication and sensing is a research domain within the PS research program, in which we are exploring the signal processing and machine learning algorithms to manage health (e.g., monitoring vital signs), improve personal awareness and enhance human sensing. New techniques are designed through which data from off-body sensors are collected and analyzed through optimized AI and machine learning techniques to support people in their daily lives. This research theme combines cross-layer sensing, communication, and processing and can be used in various applications ranging from safety, sports, health-care, and animal monitoring.
- Participatory and opportunistic sensing, in which people carrying mobile phones form a network of mobile sensor nodes and through an interactive process participate in data gathering, analysis, and sharing. The focus is on the design and development of (i) mechanisms and protocols for opportunistic networking and sensing and (ii) distributed data processing and reasoning techniques based on machine learning and AI for analyzing large amount of data generated by public mobile phones.
- Extreme networking in harsh environments is another research theme within PS. Networking in extreme environments such as deep underground and underwater, in space, as well within body, in which traditional communication mechanisms will not suffice and in presence of high interference, unreliability, and unpredictability introduce many scientific and technological challenges. The Industrial Internet of Things (IIoT) is a domain in which wireless connectivity is a key enabler for bringing the expected evolution in remote monitoring, intelligent analytics, and control of industrial processes. Low-latency, real-time, scalability and robustness are key challenges addressed.
- Sensor data analytics on embedded devices and optimization of machine learning and artificial intelligence to meet requirements of the embedded systems platform and applications has always been in the core of PS research lines. Recently, we have extended this line of research through optimization of deep learning and big data analysis techniques for embedded platforms and are frequently implementing AI on the edge (embedded AI).
- Trust, security, and privacy are essential aspects for pervasive systems, as these systems are integrated in daily live, and the operations and people depend on those. Moreover, smart sensing systems together with machine learning has the potential to get deep insights into operations and personal aspects. Although getting such insights is a major objective, misuse should be avoided. We deal with these challenges using distributed algorithms such as outlier detection, secure data sharing, hybrid AI and federated learning.
- Novel applications: pervasive technologies for healthcare, sport, smart homes and buildings, smart manufacturing, smart cities, autonomous driving.