[agents] [meetings] [news] Online Submission Open: Benchmark for Autonomous Robot Navigation (BARN) Challenge -- ICRA 2022 Competition

[Xuesu Xiao]

Submission Form:
https://docs.google.com/forms/d/e/1FAIpQLScGCMwVm-Kzg2c3kkXd_IUNwHw8D3s06ydC
g4lPgOJYkEy8aQ/viewform

 

Competition Website:
https://www.cs.utexas.edu/~xiao/BARN_Challenge/BARN_Challenge.html

 

Participation Instructions:
https://github.com/Daffan/nav-competition-icra2022

 

 

Dear roboticists, 

 

are you interested in agile robot navigation in highly constrained spaces
with a lot of obstacles around, e.g., cluttered households or after-disaster
scenarios? Do you think mobile robot navigation is mostly a solved problem?
Are you looking for a hands-on project for your robotics class, but may not
have (sufficient) robot platforms for your students? 

 

If your answer is yes to any of the above questions, we sincerely invite you
to participate in our BARN challenge
(https://www.cs.utexas.edu/~xiao/BARN_Challenge/BARN_Challenge.html)! The
BARN Challenge aims at evaluating state-of-the-art autonomous navigation
systems to move robots through highly constrained environments in a safe and
efficient manner. The task is to navigate a standardized Clearpath Jackal
robot from a predefined start to a goal location as quickly as possible
without any collision. The challenge will take place both in the simulated
BARN dataset and in physical obstacle courses at ICRA2022. 

 

1. The competition task is designing ground navigation systems to navigate
through all 300 BARN environments
(https://www.cs.utexas.edu/~xiao/BARN/BARN.html) and physical obstacle
courses constructed at ICRA2022 as fast as possible without collision. 

 

2. The 300 BARN environments can be the training set for learning-based
methods, or to design classical approaches in. During the simulation
competition, we will generate another 50 unseen environments unavailable to
the participants before the competition. 

 

3. We will standardize a Jackal robot in the Gazebo simulation, including a
2D Hokuyo with 720-dim 270-degree field-of-view 2D LiDAR, max speed of 2m/s,
etc. 

 

4. Participants can use any approaches to tackle the navigation problem,
such as using classical sampling-based or optimization-based planners,
end-to-end learning, or hybrid approaches. We will provide baselines for
reference.  

 

5. Standardized metrics/scoring system is provided on the website. 

 

6. Clearpath Robotics will provide a physical Jackal with the specified
sensor and actuator at Philadelphia and we will set up physical obstacle
courses in the venue. We will invite the top teams in simulation to compete
in the real-world. The team who achieves the fastest navigation in the
physical obstacle courses wins. 

 

If you are interested in participating, please sign up at
https://docs.google.com/forms/d/e/1FAIpQLSdJ6cUMHn8tQDNNkOistlpSmkS5jFt3-Xz6
oh1FCMzRgxpX_g/viewform?usp=sf_link 

 

Co-Organizers: 

Xuesu Xiao (UT Austin/Everyday Robots/GMU)

Zifan Xu (UT Austin)

Zizhao Wang (UT Austin)

Yunlong Song (University of Zurich/ETH Zurich)

Garrett Warnell (US Army Research Lab/UT Austin)

Peter Stone (UT Austin/Sony AI)

Tingnan Zhang (Robotics at Google)

 

Sponsor:

Clearpath Robotics, https://clearpathrobotics.com/

 

 

Thanks

Xuesu

 

-----------------------

Xuesu Xiao, Ph.D.

--

Incoming Assistant Professor (Fall 2022)

Department of Computer Science

George Mason University

--

Roboticist, The Everyday Robot Project

X, The Moonshot Factory

xuesuxiao at google.com <mailto:xuesuxiao at google.com> 

https://x.company/projects/everyday-robots/

--

Research Affiliate

Department of Computer Science

The University of Texas at Austin

xiao at cs.utexas.edu <mailto:xiao at cs.utexas.edu> 

https://www.cs.utexas.edu/~xiao/

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cs.umbc.edu/pipermail/agents/attachments/20220329/621160d0/attachment.html>