Title: Collaborative Perception and Planning for Multi-View and Multi-Robot Systems
Date: Monday, April 24th, 2023
Time: 11:00 AM – 12:30 PM EST
Location: CODA C1215, or Zoom Link
Join our Cloud HD Video Meeting
Zoom is the leader in modern enterprise video communications, with an easy, reliable cloud platform for video and audio conferencing, chat, and webinars across mobile, desktop, and room systems. Zoom Rooms is the original software-based conference room solution used around the world in board, conference, huddle, and training rooms, as well as executive offices and classrooms. Founded in 2011, Zoom helps businesses and organizations bring their teams together in a frictionless environment to get more done. Zoom is a publicly traded company headquartered in San Jose, CA.
gatech.zoom.us
Nathan Glaser
Robotics Ph.D. Student
School of Electrical and Computer Engineering
Georgia Institute of Technology
Committee:
Dr. Zsolt Kira (Advisor) – School of Interactive Computing, Georgia Institute of Technology
Dr. James Hays – School of Interactive Computing, Georgia Institute of Technology
Dr. Patricio Vela – School of Electrical and Computer Engineering, Georgia Institute of Technology
Dr. Pratap Tokekar – Department of Computer Science, University of Maryland
Dr. Milutin Pajovic – Senior Research Scientist, Analog Devices
Abstract:
The field of robotics has historically focused on egocentric, single-agent systems. However, robots in such systems are susceptible to single points of failure. For instance, a single sensor failure or adverse environmental condition can render an isolated robot
"blind". On the other hand, robots in multi-agent systems have the opportunity to overcome potentially dangerous blind spots, via communication and collaboration with their peers. In this proposal, we address these communication-critical settings with Collaborative Perception and Planning for Multi-View and Multi-Robot Systems.
First, we develop several learned communication and spatial registration schemes for collaboration. These schemes allow us to efficiently communicate and align visual observations between moving agents. We demonstrate improved egocentric semantic segmentation accuracy for a swarm of obstruction-prone aerial quadrotors. Second, we develop a distributed multi-agent SLAM algorithm that efficiently maps a shared scene, especially when robots are only allowed limited communication during rendezvous. We additionally develop a distributed multi-agent trajectory exchange method that validates and scores trajectories for self-driving vehicles. Using our method, we demonstrate reduced collision rates as compared to single-agent and multi-agent baselines. Third, we propose to apply these collaboration techniques to multi-view perception for robotic agriculture.