Autonomous systems, such as self-driving cars, are becoming tangible technologies that will soon impact the human experience. However, the desirable impacts of autonomy are only achievable if the underlying algorithms can handle the unique challenges humans present: People tend to defy expected behaviors and do not conform to many of the standard assumptions made in robotics. To design safe, trustworthy autonomy, we must transform how intelligent systems interact, influence, and predict human agents. In this talk, we'll focus on the interactions between vehicles and vulnerable agents (i.e., pedestrians). We'll discuss recent methods for inferring pedestrian intent to predict future motion, using model-based nominal predictions that are "warped" by residual learning. We'll also present methods for obtaining inverse models of human behavior to infer the presence of occluded pedestrians, effectively capturing intuition similar to humans at areas of interest like crosswalks. These methods are used to generate safe interactions between vehicles and pedestrians (sometimes with guarantees), which are demonstrated on fully equipped test vehicles.
Copyright. All Rights Reserved.