Roboticists have developed several systems that might ultimately operate in real-world environments. Because most real-world settings, particularly public areas, are dynamic and unpredictable, robots must be able to develop a thorough awareness of their surroundings to navigate them efficiently. Researchers at the University of Pennsylvania’s GRASP Laboratory recently researched to investigate how features associated with a specific environment might improve a robot’s awareness and ability to traverse its surroundings.
The new research employs several strategies that integrate global dynamics approaches with machine learning. The phrase ‘global dynamics’ in this context refers to the overall dynamics of a given environment and the characteristics that define these dynamics. In the future, the framework developed by this team of researchers might help to improve the capacity of existing and emerging robots to navigate new and dynamic environments.