Waterloo.AI Seminar: Boris Ivanovic on "Effectively Integrating Behavior Prediction within the Modern Robotic Autonomy Stack"

Thursday, November 10, 2022 2:00 pm - 3:30 pm EST
Boris Ivanovic

Title: Effectively Integrating Behavior Prediction within the Modern Robotic Autonomy Stack

Abstract: Merging into traffic is one of the most common day-to-day maneuvers we perform as drivers, yet it still poses a major problem for self-driving vehicles. The reason that humans can naturally navigate through such social interaction scenarios is that we have an intrinsic capacity to reason about other people’s intents, beliefs, and desires, using this reasoning to predict what might happen in the future and make corresponding decisions. In this talk, computational techniques for enabling self-driving cars to make predictions of the social world around them will be discussed, focusing on the rapidly-evolving field of trajectory forecasting. The talk will feature three segments, beginning with a discussion about the problem of trajectory forecasting and methods for solving it. Then, we’ll discuss how, even with the ability to predict the future, it is still unclear how best to integrate such trajectory forecasting models within the autonomy stack. For instance, what information is required from upstream perception modules? How can future predictions be efficiently incorporated in downstream planning and control algorithms? Finally, we’ll conclude with a view towards the future, discussing recent advancements in generalizing trajectory forecasting approaches to diverse environments and other exciting open research problems.

Speaker Bio: Boris Ivanovic is a Research Scientist in NVIDIA’s Autonomous Vehicle Research Group. Prior to joining NVIDIA, he received his Ph.D. in Aeronautics and Astronautics in 2021 and an M.S. in Computer Science in 2018, both from Stanford University. He received his B.A.Sc. in Engineering Science from the University of Toronto in 2016. Boris' research interests are rooted in trajectory forecasting and its interactions with the rest of the autonomy stack. This usually includes a mix of improving raw prediction performance, integrating prediction with perception and planning, holistically evaluating autonomy stack performance, and leveraging behaviour models for simulation. He has also previously conducted research in the fields of computer vision, natural language processing, and data science.

Date and Time: 
Thursday November 10th, 2022
2:00 PM - 3:30 PM EST

Virtual Seminar: