The STREAM research group was created by combining expertise in Guidance, Navigation and Control (GNC) applied to autonomous terrestrial, aerial and space platforms. The group's activity focuses on the development of innovative methodologies, both theoretical and applied, for the control and planning of autonomous operations in complex scenarios. In particular, the group is active in the development of advanced distributed control and cooperative planning algorithms for the management of multi-agent systems.

STREAM studies and implements solutions for a wide range of robotic applications, including precision agriculture, critical infrastructure inspection and drone fleet management. Methods drawn from consensus theory, predictive control and decentralised filters are used to provide an adequate level of robustness, scalability and adaptability to the mission.

STREAM's activities also extend to the study of GNC techniques for spacecraft, addressing issues such as precision attitude control, trajectory planning, formation flight and proximity operations. The methods used include artificial potential field, sliding mode and optimal control, to combine the level of precision required in space with the available computational power.

The group is also exploring Machine Learning techniques, such as reinforcement learning and neural network-based supervised learning, for the enhancement of classical filter-based navigation, and for the development of autonomous behaviour in dynamic and unknown environments.

The proposed solutions are validated through a progressive testing process that includes Software-In-The-Loop and Hardware-In-The-Loop simulations in realistic environments using ROS2, Matlab/Simulink and Python. STREAM uses an experimental infrastructure that includes terrestrial robotic platforms, multi-rotor drones and an indoor cage equipped with a motion capture system.