The Office of Naval Research (ONR) has awarded PI Darinka Dentcheva, professor and chair of the Department of Mathematics, and co-PI Brendan Englot, assistant professor in the Department of Mechanical Engineering, and a grant of $900,055 for their three-year project “Risk-Averse Learning and Control for Distributed Dynamical Systems with Partial Information.” This project is in collaboration with co-PI Andrzej Ruszczynski, distinguished professor at Rutgers University.
When information about the state of a multi-robot system and its environment is incomplete and distributed across the members of the robot team, there are open challenges in risk evaluation and control. Each robot has only partial knowledge of the other robots' states, and only local information about the environment. Controlling the system must involve learning the environment and system's uncertain state, exchanging information between robots, and risk evaluation.
In this project, the team will develop a general methodology for dynamic risk allocation and distributed risk-averse control. Three sources of risk will be addressed: the risk due to an uncertain environment, the risk associated with not knowing the full state of the system (due to information limitations), and the risk associated with uncertain future evolution of the system and its environment. The risk we are concerned with pertains both to the integrity of the system as well as to its performance, that is, completing the mission successfully.
A team of mobile robots operating in an underwater environment faces all of these aforementioned challenges. Environmental processes, such as winds, waves, and currents, perturb the robots. The extent of uncertainty in the robots' sensor observations also depends on environmental factors. Underwater robots are also constrained by the bandwidth with which they can communicate wirelessly, and the range at which they are visible to one another.
This project will support the development and experimental testing of this framework using teams of ground-based and underwater robots.
“This project will be an exciting opportunity to address the many sources of uncertainty plaguing underwater robots, allowing them to learn how to perform more effectively as a team despite the natural constraints and limitations of their environment,” said Englot. “Our unique multi-disciplinary team will allow us to develop new theory that can also be put into practice and provide new capabilities for autonomous systems.
Dentcheva said, "The mathematical foundations of risk-averse control of decentralized partially observable dynamical systems and the associated methods of statistical learning are the focus of our research agenda. Operation of underwater robots integrates all challenges such dynamical systems face. We hope to impact the way complex problems of this type are treated."