Autonomous Interactive Systems and Machine Intelligence

The department of Autonomous Interactive Systems and Machine Intelligence conducts research on theoretical, computational and experimental aspects of autonomous, interactive, intelligent and connected systems. The conducted research outcomes in signal processing, machine learning, continual active learning, and vision which are used to develop novel methods that act as fundamental technology enablers of autonomous and interactive systems. The mission of the group is to explore and tackle challenges that emerge both in robot-people interactions (e.g., collaborative decision making, explainable AI, affective robots) and in connected robot-environment interactions (e.g., sensing, perception, mapping, swarms). The group is concerned with a broad variety of real-world problems autonomous and interactive systems can contribute to, from connected and autonomous vehicles,
connected robots in Industrial Environments to precision agriculture and swarm intelligence in m-health, from emergency response to advanced decision-making. The group envisions a world where autonomous and interactive systems operate safely and effectively alongside humans and form trusting partnerships to improve the well-being of individuals and societies. This vision drives our research in developing theoretical foundations and computational frameworks that enable reliable and intelligent autonomy.

The main research activities, which are conducted by the member of the group are summarized below:

  • Design and development of robust and reliable continual active learning and explainable AI based perception systems, for trustworthy knowledge acquisition & enhanced situational awareness
  • AI based methods for enabling trustworthy decision-making in connected autonomous interactive systems through multi-agent cooperative processing, learning that can be directly applied in various applications including indicatively cooperative awareness, localization, mapping and motion planning solutions
  • Training and Validation of the perception solutions in commercial, open source & in-house integrated cross layer simulators
  • Development of adaptive, scalable, social, intuitive, AI-powered information rendering front-end of HMIs accommodating different user types, profiles, devices (e.g. HUDs, eHMIs, smartwatch) that will seamlessly operate and provide information to the user.

The group aims to research and develop machine learning, machine vision, computational intelligence and human-machine interaction approaches that can be effectively applied in autonomous, interactive and connected intelligent systems. The department's research includes: i) state-of-the-art technologies of distributed machine learning systems and artificial intelligence for the recognition, isolation, extraction and visualization of important information features / patterns using deep/federated learning algorithms, multimodal and distributed optimization methods, cooperative processing, analysis and extraction techniques, etc.) ii) 3D / 4D information processing technologies, image & video processing, computer vision, iii) enhanced immersive visual interfaces that support interactive display, interactive learning and realtime collaborative control for seamless and faster data analysis iv) integrated decision support solutions for human-machine and machine-machine interaction systems. The research activity of the department includes the applications of the above technologies in digital twins and in cyber-physical systems, in autonomous transport systems, in the Visualization of Industrial Operations, in the Interaction of operators with machines in Industrial environments, in Precision Agriculture, as well as in m-health systems.