IROS 2024 Cooperative Aerial Robots Inspection Challenge (CARIC@IROS24)

• 1. Introduction
• 2. CARIC at IROS 2024
  • 2.1. How to participate?
  • 2.2. Important dates
  • 2.3. Prize

1. Introduction

    Inspection, repair and maintenance is a multi-billion business that is projected to grow substaintially in the next decade, with integration of automation and AI being the driving force. At its core, the primary goal in an inspection mission is to capture images on the surface of some structures at the best possible quality. However, exploration is also a secondary objective that needs to be addressed to identify the structure and its surface. Oftentimes, bounding box(es) can be set around the target of interest to narrow the area of exploration.

    Thanks to their unique mobility, aerial robots have become widely adopted for inspection tasks. We believe that the next breakthrough in this industry shall be delivered by Cooperative Aerial Robots Inspection (CARI) systems. Moreover, in the same spirit of economic specialization, heterogeneous CARI systems has the potential to acheive greater efficiency, quality and versatility compared to single-UAV or homogenous systems. Nevertheless, to accomplish this vision, novel cooperative strategies that can optimally coordinate the operation of multiple robots remain an open research problem.

    To accelerate this development, we introduce the Cooperative Aerial Robots Inspection Challenge Benchmark, a software stack based on Gazebo, RotorS and other open-source packages. The objective of CARIC is twofold. First, it aims to faithfully simulate multi-UAV systems operating in typical real-world inspection missions. Second, based on this tool, different cooperative inspection schemes can be benchmarked based on a common metric. The software stack is made public to benefit the community and we would like to welcome all who are interested to participate in the challenge to be held at IROS 2024.

    CARIC software package can simulate UAV dynamics, physical collisions, camera-FOV-based instance detection, and line-of-sight-only communications

2. CARIC at IROS 2024

2.1. How to participate?

    The challenge’s procedure is as follows:

• Sign up via the following Form.
• Read through the description of CARIC software stack in the Guideline.
• Participants develop their CARI schemes based on CARIC. Do notice the ground rules. The implementation can be in python, C++ or docker executable.
• Send your code / executable to Dr. Thien-Minh Nguyen via thienminh.npn@ieee.org and cc xu0021ng@e.ntu.edu.sg.
• The submitted method will be evaluated on three out of the following five scenarios:

     The scale and position of the building will not be the same as in the released models. Only one big bounding box will be defined in the test scenarios.

• The methods will be ranked based on the **total mission scores obtained in all three chosen scenarios.
• If you have any technical inquiry, please raise an issue on our github. You may also join our Discord group and chat with our developers!

2.2. Important dates

    (All dates and time are in GMT+7)

• Last day to register for the challenge 24th September 2024.
• Last day to update your code 29th September 2024.
• Annoucement of results 4th October 2024.
• Workshop date 13th October 2024.

2.3. Prize

    The winner will receive a material or monetary prize of SGD1000 and a certificate. If you are interested in being a sponsor of the challenge, please contact thienminh.npn@ieee.org and cc xu0021ng@e.ntu.edu.sg.

Organizers:

Thien-Minh Nguyen

Research Assistant
Professor
School of Electrical & Electronic Engineering
Nanyang Technological University

Muqing Cao

Research Fellow
School of Electrical & Electronic Engineering
Nanyang Technological University

Shenghai Yuan

Research Fellow
School of Electrical & Electronic Engineering
Nanyang Technological University

Xinhang Xu

Research Assistant
School of Electrical & Electronic Engineering
Nanyang Technological University

Lihua Xie

Professor
Director for Centre for Advanced Robotics Technology Innovation (CARTIN)
President's Chair in Control Engineering
School of Electrical & Electronic Engineering
Nanyang Technological University

Ben M. Chen

Professor
Vice-Chairman (Graduate)
Department of Mechanical and Automation Engineering
The Chinese University of Hong Kong