In this thematic line, a set of project-teams sharing a set of common goals, methodologies and scientific tools has been working for several years. The focus is to research on fundamental and open issues for human-centered robots, operating in long-term, in the strands of medical robotics and mobility solutions for human-made environments, including wheeled assistance mobile robots and driverless vehicles.
The multi-disciplinary research team working on this thematic line includes experts in robotics, computer vision, pattern recognition, machine learning, brain computer interfaces, intelligent control, and embedded systems. The research is focused on three major axes: Mobility (indoor and outdoor), Medical Robotics, and Computational Intelligence.
Indoor Mobility: the team is very active in the research of assistive mobile robots to increase autonomy of people with disabilities. Navigation and SLAM, local planning, and the use of human-machine interfaces (HMI), like brain computer interface (BCI), have been developed. The team is evolving toward the research on multi-assistive collaborative robots with a strong focus on collaboration with users, and cooperation among mobile robots and the infrastructure.
Outdoor Mobility: the team has a significant track record on the development of navigation and perception technologies for autonomous vehicles, such as pedestrian detection, sensor fusion for navigation, and path following. More recently, the team has started to research on multi-vehicle systems for new urban mobility/transportation systems based on automated and cooperative electric vehicles. Intelligent intersection management, obstacle detection and avoidance, electric vehicle technologies, V2X communications and route planning are otherkey research topics.
Medical Robotics: A key goal of this research axis is to design telerobotic control architectures with high-quality haptic feedback for minimally invasive surgeries and for beating-heart surgeries with autonomous motion compensation. Robotic-assisted tele-diagnosis through ultrasound imaging is a non-surgical application with specific challenges (e.g., large and random time delays) that we plan to tackle too. For the rehabilitation area, we plan to create an advanced patient-robot setup for upper and lower limb rehabilitation programs. Lightweight robots with compliant control architectures, biosignals detecting muscle activity and gesture dynamics, and advanced interfaces for online monitorization and assessment will be coherently merged into a single setup, to boost upper and lower limb rehabilitation programs.
Computational Intelligence: Targeting industrial problems, research will be performed on: 1) soft sensor methodologies able to tackle processes with multiple operating modes, dynamic optimization and on-line learning in changing environments, ultra high dimensional variable selection and modeling, transductive learning making use of both labeled and unlabeled data for training in data stream mining with improved prediction performance; 2) robust control methodologies with safety working region to deal with process uncertainties.