We investigate issues related to coordination mechanisms and automated negotiation in multi-agent systems, agent learning, swarm formation and coordination of our own robotic swarm.

We are studying different models of deep neural networks and deep reinforcement learning for recognition and tracking tasks, and for autonomous driving.

We use computer vision and planning techniques to develop specialized software for social robots that interact with people, can interpret people’s action and respond appropriately. We develop software for assistive robots that are able to ensure multi-modal interaction and perform actions for the well-being of people with special needs, thus integrating these robots in our Active and Assisted Living applications to offer a better and user friendly environment.

Recently, we have started research in explainable models of deep neural network recognition tasks.