A key research goal for the scientific community is to use robotic systems effectively not only in industrial production lines, but also in service applications such as healthcare, logistics, and domestic. When robots are required to perform complex tasks  in unstructured and dynamic environments, the computational power required is typically not fully available onboard. Therefore, complex planning and control algorithms can run on an edge cloud, which is a high-performance computer installed in the same local network as the robot, provided with a wireless connection.  

The first objective of the master thesis project is to develop a simulation environment (e.g., based on robosuite or similar approaches) that includes a robotic mobile manipulator connected via a 5G or WiFi6 communication channel to an edge cloud.   The  secondo objective is to develop a simple reinforcement learning system that take into account network delay.