IEEE Kingston Section

IEEE

Ingenuity Labs and IEEE Joint Invited Lecture – Uncertainty Assessment for Deep Networks: Making Autonomous Driving Perception Aware of Its Own Limitations

The EMB/RA/CS Societies Joint Chapter of IEEE Kingston and Queen’s Ingenuity Labs Research Institute are proud to present the following invited lecture:

 

UNCERTAINTY ASSESSMENT FOR DEEP NETWORKS: MAKING AUTONOMOUS DRIVING PERCEPTION AWARE OF ITS OWN LIMITATIONS

 

Date:  Wednesday November 20th, 2019.

Time:  12:30 – 1:30 PM

Location: Mitchell Hall, Room 395, Queen’s University 

Speaker:  Prof. Steve Waslander, University of Toronto Institute for Aerospace Studies (UTIAS). Director, Toronto Robotics and Artificial Intelligence Laboratory (TRAILab).

Light Refreshments: 1:30 – 2:00PM, Mitchell Hall, Room 395, Queen’s University

 

Abstract

Most autonomous vehicle perception approaches are primarily reliant on modern deep neural networks (DNNs).   DNNs have shown breakthrough performance or object detection, tracking and prediction, scene segmentation, vehicle localization and mapping, providing accurate bounding boxes for vehicles and pedestrians, lane boundaries and signage over extensive datasets and on-road testing. Yet, these networks are not uniformly consistent in the quality of their perception outputs, and much can be gained by accumulating evidence over time.  In this talk, I will lay out our progress in 3D object detection to improve detection accuracy for a range of sensor configurations, and demonstrate the effects of adverse weather on these approaches.  Further, I will describe our approach to providing reliable uncertainty estimates for network outputs that enable proper Bayesian inference when incorporating prior information and tracking object motion through time.

 

Speaker Bio:

Prof. Steven Waslander is a leading authority on autonomous aerial and ground vehicles, including multirotor drones and self-driving cars.  He received his B.Sc.E.in 1998 from Queen’s University, his M.S. in 2002 and his Ph.D. in 2007, both from Stanford University in Aeronautics and Astronautics, where as a graduate student he created the Stanford Testbed of Autonomous Rotorcraft for Multi-Agent Control (STARMAC), the world’s most capable outdoor multi-vehicle quadrotor platform at the time. He was recruited to Waterloo from Stanford in 2008, where he founded and directs the Waterloo Autonomous Vehicle Laboratory (WAVELab), extending the state of the art in autonomous drones and autonomous driving through advances in localization and mapping, object detection and tracking, integrated planning and control methods and multi-robot coordination. His work on autonomous vehicles has resulted in the Autonomoose, the first autonomous vehicle created at a Canadian University to drive on public roads. His insights into autonomous driving have been featured in the Globe and Mail, Toronto Star, National Post, the Rick Mercer Report, and on national CBC Radio.  In 2018, he joined the University of Toronto Institute for Aerospace Studies (UTIAS), and founded the Toronto Robotics and Artificial Intelligence Laboratory (TRAILab).

To attend this seminar, RSVP by clicking this Link

For more information, please contact Dr. Joshua Marshall or Dr. Keyvan Hashtrudi-Zaad

 

 

 

 

 

Comments are closed.