ICRA 2022 Workshop on Releasing Robots into the Wild: Simulations, Benchmarks, and Deployment

The workshop will take place on Friday 27 May 2022 @ 08:30 ET – 17:15 ET in Room 113C.

Overview

Machine learning and learning-based control promise to lead robot capabilities and performance beyond what human designs can attain. However, the results from recent years of fast-paced progress in reinforcement learning and learning-based control have proven challenging to compare. While acknowledging that learning robots carry a plethora of fascinating open research questions, in this workshop, we specifically eye physics-based simulations and learning for decision making. We argue that two important roadblocks hamper the transfer of robot learning research into real-world applications: (i) the scarcity of sufficiently realistic simulation tools, tasks, and datasets to reliably compare algorithmic progress; and (ii) the lack of reliable and repeatable processes to transfer those simulation results to the real-world.

Discussion Themes | Program | Speakers | Accepted Papers | Organizers

Discussion Themes

Theme A: Realistic Simulation Tools, Tasks, and Datasets to Reliably Compare Algorithmic Progress

  • Many simulators exist, but yet none has really become a common benchmark. What is missing?
  • What constitutes a fair benchmark, how do we design for them, and how can we avoid overfitting to the specific benchmarks/datasets?
  • What benchmark tasks and metrics should we prioritize, do they have sufficient coverage over our desired robot performance, e.g. in terms of efficiency, safety and robustness?
  • How does/can simulation help benchmarking and robotics research, what are the desirable robot simulator properties, what is hampering their implementations and what are their limitations?

Theme B: Reliable and Repeatable Processes to Transfer Simulation Results to the Real World

  • The Turing test for robotics: what should be tested in an autonomous system to guarantee its functionality in unstructured environments?
  • How can we guarantee that the conclusions we draw in the simulation are also valid in the physical world? Shall we rely on a (possibly utopic) algorithm to perfectly transfer knowledge between domains or can we do something better?
  • What are the biggest issues when transferring an algorithm from simulation to the real world, how these issues can be addressed, and can the solutions to these issues be made reliable and scalable, leveraging standard engineering practices, for real-life deployment?
  • We can never exactly simulate the real world, thus, we cannot expect perfect transfer. However, can we define a bound/set an expectation of how well we theoretically expect a robot, trained in simulation, to perform in the real world?

Program

Below is the schedule of our workshop. Times are in ET.

Theme A: Realistic Simulation Tools, Tasks, and Datasets to Reliably Compare Algorithmic Progress

08:30-08:40 Opening Remarks
08:40-08:50 Introductory Talk by Angela Schoellig
08:50-09:00 Theme A Introduction
09:00-09:20 Invited Talk A.1 Liila Torabi
09:20-09:40 Invited Talk A.2 Joelle Pineau
09:40-10:00 Invited Talk A.3 Erwin Coumans
10:00-10:20 Coffee Break and Poster Session
10:20-10:40 Invited Talk A.4 Karime Pereida
10:40-11:40 Theme A Panel Discussion
11:40-12:00 Spotlight Talk from the Winner of the ICRA 2022 DodgeDrone Challenge: Vision-based Agile Drone Flight
12:00-12:30 Spotlight Talks
12:30-13:45 Lunch and Poster Session

Theme B: Reliable and Repeatable Processes to Transfer Simulation Results to the Real World

13:45-13:50 Theme B Introduction
13:50-14:00 Introductory Talk by Davide Scaramuzza
14:00-14:20 Invited Talk B.1 Ingmar Posner
14:20-14:40 Invited Talk B.2 Jens Kober
14:40-15:00 Invited Talk B.3 David Hsu
15:00-15:20 Coffee Break and Poster Session
15:20-15:40 Invited Talk B.4 Animesh Garg
15:40-16:00 Invited Talk B.5 Raquel Urtasun
16:00-17:00 Theme B Panel Discussion
17:00-17:05 Concluding Remarks

Invited Speakers

Theme A: Realistic Simulation Tools, Tasks, and Datasets to Reliably Compare Algorithmic Progress

  • Joelle Pineau from McGill and Facebook Research. Talk title: “Building Reproducible, Reusable, and Robust Deep Learning Systems”
  • Erwin Coumans from Google Brain and the PyBullet development team . Talk title: “Simulation and Sim-to-Real for Quadruped Robot Locomotion”
  • Liila Torabi from Nvidia and the Isaac team. Talk title: “NVIDIA Isaac Sim: A Platform for Developing and Training Smarter Robots”
  • Karime Pereida from Ocado Technology. Talk title: “From Simulation to Production: Reinforcement Learning in Industry”

Theme B: Reliable and Repeatable Processes to Transfer Simulation Results to the Real World

  • Jens Kober from TU Delft. Talk title: “Interactions: A Blessing or a Curse?”
  • Ingmar Posner from Oxford. Talk title: “Learning to Simulate: Generating Data for Perception and Action”
  • Raquel Urtasun from University of Toronto and Waabi. Talk title TBD
  • David Hsu from the National University of Singapore. Talk title: “Closing the Planning-Learning Loop: Autonomous Driving, Object Manipulation,…”
  • Animesh Garg from the University of Toronto and Nvidia. Talk title: “Paving the Path to Robot Autonomy with Simulation

Accepted Papers

The accepted papers are listed as follows.

  • “Adaptive Nonlinear MPC for Quadrotors”, Drew Hanover and Davide Scaramuzza [PDF]
  • “MIDGARD: A Simulation Platform for Autonomous Navigation in Unstructured Environments”, Giuseppe Vecchio, Simone Palazzo, Dario C. Guastella, Ignacio Carlucho, Stefano V. Albrecht, Giovanni Muscato, and Concetto Spampinato [PDF]
  • “Interactive OAISYS: A photorealistic terrain simulation for robotics research”, Marcus G. Müller, Jaeyoung Lim, Lukas Schmid, Hermann Blum, Wolfgang Stürzl, Abel Gawel, Roland Siegwart, and Rudolph Triebel [PDF]
  • “The Stochastic Road Network Environment for Robust Reinforcement Learning”, John D. Martin, Paul Szenher, Xi Lin, and Brendan Englot [PDF]
  • “Quantifying and Using System Uncertainty in UAV Navigation”, Fabio Arnez, Ansgar Radermacher, and Huascar Espinoza [PDF]
  • “SWARM Simulation Platform: Algorithm Benchmarking and Development Suite for Multi-Agent Systems”, Tyler Fedrizzi, Tong Yao, and Shreyas Sundaram [PDF]
  • “BURG-Toolkit: Robot Grasping Experiments in Simulation and the Real World”, Martin Rudorfer, Markus Suchi, Mohan Sridharan, Markus Vincze, and Ales Leonardis [PDF]
  • “Sim-to-Real Transfer for High-Speed Quadrotor Flight”, Elia Kaufmann and Davide Scaramuzza [PDF]
  • “Sim-to-Real Strategy for Spatially Aware Robot Navigation in Uneven Outdoor Environments”, Kasun Weerakoon, Adarsh Jagan Sathyamoorthy, and Dinesh Manocha [PDF]

Organizers

Angela P. Schoellig, Associate Professor, Technical University of Munich and University of Toronto Institute for Aerospace Studies
Davide Scaramuzza, Professor, Department of Informatics, University of Zurich
Adam Hall, Ph.D. Student, University of Toronto Institute for Aerospace Studies
Zhaocong Yuan, MASc Student, University of Toronto Institute for Aerospace Studies
Jacopo Panerati, Postdoctoral Fellow, University of Toronto Institute for Aerospace Studies
SiQi Zhou, Ph.D. Candidate, University of Toronto Institute for Aerospace Studies
Lukas Brunke, Ph.D. Student, University of Toronto Institute for Aerospace Studies
Melissa Greeff, Ph.D. Candidate, University of Toronto Institute for Aerospace Studies

University of Toronto Institute for Aerospace Studies