SBIR-STTR Award

Global Optimization of Graph Learning Structures (GOGLS) is a platform- agnostic method of globally optimizing search graphs from individual autonomous agents using techniques derived from federated learning. The GOGLS mission and focus will be to provide a learning framework for robots to share search information and in essence, see through the eyes of their neighbors. The GOGLS system relies on passing and fusing the latest high level model of the environment between robotic autonomous systems instead of the low level data. This allows the group to explore and learn about the environment in a federated learning inspired manner without the need for large online data transfers.