Building Better Bots with Bayes: Probabilistic Human-Machine Reasoning for Collaborative Autonomy

From the development of foundational state space estimation tools like the Kalman filter to state of the art machine learning techniques for sensor fusion and decision making, probabilistic models and reasoning algorithms are the “lingua franca” for modern robotics and autonomous systems. The COHRINT Lab at CU develops and leverages probabilistic AI in new and unique ways to tackle fundamental research questions for current and futuristic systems. Dr. Nisar Ahmed will highlight his lab’s recent work on human-machine/robot interaction for collaborative information gathering and reasoning, using probabilistic Bayesian state estimation and decision-making algorithms. These methods not only plug in seamlessly to existing autonomy architectures, but also exploit the ability of human collaborators to provide semantic data (via user-friendly interfaces) that is rich with useful “out of band” information for autonomous platforms. In essence, these methods open the door to “soft re-programming” of autonomous reasoning from the outside by end-users (who are not robotics experts or computer scientists). Aerospace applications such as integrated UAS surveillance/reconnaissance, UAS-enabled wilderness search and rescue, and remote robotic space exploration will be demonstrated and discussed.

Speaker(s): Dr. Nisar Ahmed

Room: 426, Bldg: Beatty Hall, 550 Huntington Ave, Boston, Massachusetts, United States, 02115