Task 010: Self-flying Drones and Visual Analytics

Event Date: April 8, 2021
Time: 11:00 am (ET) / 8:00am (PT)
Priority: No
College Calendar: Show
Kendall J. Queen, University of Pennsylvania
Event-Based Lane Detection for Autonomous Ground Vehicles
ABSTRACT: Event cameras record changes in the logarithm of the light intensity at high frequency, low latency, and low power. Methods using event cameras have demonstrated the ability to extract optical flow, depth, ego-motion, and recognize objects. However, limited work has studied the ability to use events for control. In this presentation, I will propose a method for lane following using event cameras composed of three key components: 1) a lane detection algorithm using a Hough transform applied on event images, 2) a Kalman filter for tracking the lane, and 3) a PD controller minimizing cross track error.  We study the accuracy of our lane detection, and we successfully demonstrate success in lane following using our approach in the CARLA simulator.
Bio: Kendall Queen is an Electrical & Systems Engineering PhD student at the University of Pennsylvania under the advisement of Kostas Daniilidis, PhD. He received his B.S. degree in Computer Engineering from the University of Maryland, Baltimore County (UMBC) as a Meyerhoff Scholar in 2016. He then earned his M.S. degree with a focus in Robotics from Penn. Kendall’s interests lie at the intersection of robotics and computer vision. Currently Kendall is investigating event cameras’ applications to robotic systems specifically autonomous ground vehicles.