New Autonomous Aerial Manipulation Platform Unveiled at CoRL 2024

We’re excited to share our latest research, presented at CoRL 2024, on an open-source autonomous aerial manipulation platform capable of grasping objects using onboard perception only. This innovative system, designed for real-world applications, demonstrates high success rates and adaptability without relying on external tracking.

raptor-outdoor-drone-grasping-object

We are thrilled to announce our latest paper, “An Open-Source Soft Robotic Platform for Autonomous Aerial Manipulation in the Wild,” presented at the Conference on Robot Learning (CoRL) 2024. Join us on Thursday at 9:30 am for the Poster Spotlight and again at 11 am for the Poster Session!

Our new aerial platform enables autonomous grasping and manipulation outdoors, powered entirely by onboard perception. This groundbreaking system performs with an 85% success rate across 144 grasping attempts, eliminating the need for external tracking systems.

Key Innovations:

• Autonomous Outdoor Grasping: Uses onboard perception only, allowing for reliable operation in real-world environments.
• High Success Rate: Achieved 85% success in testing with multiple objects.
• Integrated Systems: Combines SLAM-based localization and learning-based object detection.
• Adaptive Grasping: Features a soft robotic gripper capable of handling various objects.
• Open-Source Access: ROS 2 software stack and hardware design are available for the research community.

Learn more:

• Video: external page Watch the platform in action
• PDF: external page Read the full paper on arXiv
• GitHub: external page Explore the open-source project

Congratulations to our team—Erik Bauer, Marc Blöchlinger, Pascal Strauch, Arman Raayatsanati, Curdin Cavelti, and Prof. Robert Katzschmann—for their work on this transformative aerial manipulation platform. This project is a significant step forward for real-world aerial robots, expanding their potential in various outdoor applications.

JavaScript has been disabled in your browser