Vision-based Drones: What's Next?

Full Day Workshop at IROS 2018

October 5, Madrid Municipal Conference Centre, Room 1.L5, Madrid, Spain

Program

The program includes invited talks, contributed papers and real-time demos. The final schedule might slightly change.

    Session 1
    • 08:50 - 09:00 Welcome & Introduction
    • 09:00 - 09:30 Talk 1 – N. Roy, MIT, “”
    • 9:30 - 10:00 Talk 2 – L. Benini, University of Bologna and ETH Zurich, "Ultra Low Power Deep-Learning-powered Autonomous Nano Drones"
    • 10:00 - 10:50 Contributed papers pitch 6 minutes per paper
    • 10:50 - 11:00 Talk 3 – L. Seneviratne, Khalifa University, "MBZIRC 2020 Challenges Related to Vision Based Drones"
    • 11:00 - 11:30 Coffee Break and paper poster presentation
    Session 2
    • 11:30 - 12:00 Talk 4 - Guido De Croon, Tu Delft, “Self-supervised learning for improved vision-based navigation”
    • 12:00 - 12:30 Talk 5 - Giuseppe Loianno, New York University, “Autonomous Agile Human-Friendly Drones”
    • 12:30 - 13:00 Talk 6 - Larry Matthies, NASA JPL, “Vision Systems for Planetary Landers and Drones: Progress and Challenges”
    • 13:00 - 13:30 Talk 7 - Davide Scaramuzza, University of Zurich, “Vision Systems for Planetary Landers and Drones: Progress and Challenges”
    • 13:30 - 14:30 Lunch Break
    Session 3
    • 14:30 - 14:50 Talk 8 - Andrey Kolobov, Microsoft, "Vision for Longer Drone Missions"
    • 14:50 - 15:10 Talk 9 - Max Ruffo, Terabee, “What market is ready to pull out the next challenge for direct implementation?”
    • 15:10 - 15:30 Talk 10 - Zhyiuan Li, DJI, “”
    • 15:30 - 15:50 Talk 11 - Nikolai Smolyanski, NVIDIA, “Stereo Cameras on Drones: End-to-end Deep Neural Networks for Reliable Depth Perception”
    • 15:50 - 16:30 Introduction to Autronomous Racing Drone Competition. 15 minutes, Pitch from 3 best teams, 5 minutes each
    • 16:30 - 17:00 Coffee break, demo sessions and poster presentation from Drone Competition
    Session 4
    • 17:00 - 17:30 Talk 12 - Juan Nieto, and Roland Siegwart, ETH Zurich, “Getting close to structure: Autonomous MAV flight in cluttered environments”
    • 17:30 - 18:00 Talk 13 - Pascual Campoy, UPM, “”
    • 18:00 - 18:30 Talk 14 - Xuning Yang and Nathan Michael, CMU, “Toward intuitive human controlled MAVs: motion primitives based teleoperation”
    • 18:30 - 18:45 Panel Discussion and Closing Remarks

Updates

  • 09/06/2018 The program is out and the digest is also available.
  • 09/06/2018 The notifications of acceptance/rejection for contributed papers have been sent out. If you didn't receive it, please contact the organizers.
  • 08/01/2018 A Journal of Field Robotics special issue on workshop related topics will be available shortly after the workshop. Contributors to the workshop will be highly encouraged to submit an extended version of their work.
  • 07/15/2018 The call for submission is open. Please go to the submission page.
  • 06/08/2018 The website is available

Overview

The workshop will feature peer-reviewed paper presentations and live demos.

Motivation

Autonomous micro helicopters are starting to play a major role in tasks like search and rescue, environment monitoring, security surveillance, transportation and inspection. However, for such operations, two main challenges arise. The use GPS based navigation is not sufficient. Fully autonomous operation in cities or other dense indoor and outdoor environments requires micro helicopters to fly at low altitudes, where GPS signals are often shadowed or absent. In addition, during the previous mentioned tasks, agile motions are still not possible, compromising the execution of critical missions. These should be typically accomplished in a fast and agile manner and within a limited amount of time. Thus, several perception and control challenges have still to be addressed and solved. Unmanned Aerial Vehicles (UAVs) should be able to fly autonomously with agility in extreme navigation conditions guaranteeing robust high rate state estimation for closed loop control. On the other hand, multiple MAVs have been endowed with manipulation and transportation capabilities. Although the complexity of such systems increases with the number of agents, MAVs can perform tasks in a collaborative manner and exchange information between each other to make better decisions and optimize tasks.

Goal

This workshop will focus on the future challenges, visionary, cutting-edge and technological ideas in the area of vision-based drones. There are still several open research and scientific questions related to the best and efficient environment representations for navigation and toward unified solutions for manipulation, transportation, locomotion, human-robot interaction, and heterogeneity. How can drones autonomy change the human mobility? How can these machines interact with humans during a task predicting his future behavior and provide situational awareness relaxing communication constraints? What algorithms can automatically establish the best heterogeneity degree in terms of vehicles and sensors in a mission? How do we co-design perception and action loops for fast navigation of nano scale aerial platforms? What role should machine learning play for autonomy? The main goal is to encourage discussion between experts in the field on the previous topics and set upcoming new research topics and ideas, and roadmap for the upcoming 5-10 years. The workshop will feature, in addition to contributed and invited talks, real-time demos provided both from academia and industry actively working in this field.

Topics of Interest

Topics of interest to this workshop include, but are not necessarily limited to:
  • Visionary ideas for autonomy of vision-based UAVs
  • Agile autonomous navigation, transportation and manipulation with UAVs
  • High-speed visual control and state estimation of aerial vehicles
  • Long term and range perception for UAVs without GPS
  • Sensor fusion for autonomous navigation in unstructured environment
  • System software and hardware architectures
  • Mapping and Obstacle avoidance
  • Perception in challenging and dynamic environments
  • Modeling and benchmarking of performances for three-dimensional navigation
  • Dynamic visual servo control of aerial vehicles
  • Cooperative estimation and control with multiple aerial vehicles
  • Resource constrained navigation
  • Field robotics
  • Search and rescue robotics

We encourage the submission of papers describing work in progress, preliminary results, novel concepts, and industry experiences. Contributions are expected from researchers working in different fields related to autonomous aerial vehicles.

Call for papers

All submitted papers will be reviewed by at least two international experts (see Program Committee Member list below) on the basis of technical quality, relevance, significance, and clarity. We accept extended abstracts (2-4 pages), experiences' reports (2-4 pages), or full research papers (up to 6 pages). We also encourage submission of live demos and working systems (up to 2 pages). All accepted papers will appear on the workshop website. We are currently exploring the possibility of a journal special issue for the best contributions at the workshop.

Important Dates

  • June 15: Call for submission opens
  • July 31 August 21: Deadline of full-length papers
  • August 31: Notification of acceptance
  • October 5: Workshop at IROS 2018 Madrid Municipal Conference Centre, Madrid, Spain

Organizers

Giuseppe Loianno
loiannog@nyu.edu
New York University

Davide Scaramuzza
sdavide@ifi.uzh.ch
University of Zurich

Vijay Kumar
kumar@seas.upenn.edu
University of Pennsylvania

Program Committee

  • Dr. Shaojie Shen, HKUST
  • Dr. Gary McGrath, Qualcomm Research
  • Dr. Nikolai Smolyanskiy, NVIDIA
  • Dr. Tarek Taha, Khalifa University
  • Dr. Debadeepta Dey, Microsoft
  • Dr. Juan Nieto, ETH Zurich
  • Dr. Nathan Michael, CMU
  • Dr. Martin Saska, CTU