TubePicker

Detect and retrieve sample tubes on Mars.

If first html tag is indented, and this include is called after a list, the html tag is considered a list element, and things go wrong. Leaving a hidden unindented line here
Period of Performance: 2021–2022
My role: PI
Additional Contributors: Publication authors (see References below).
sponsor Funded by the European Space Agency (ESA).
project thumbnail

Detect and retrieve sample tubes on Mars.

If first html tag is indented, and this include is called after a list, the html tag is considered a list element, and things go wrong. Leaving a hidden unindented line here
Period of Performance: 2021–2022
My role: PI
Additional Contributors: Publication authors (see References below).
sponsor Funded by the European Space Agency (ESA).
project thumbnail

In this project, we developed a two-stage approach to address the complex challenge of accurately estimating the pose of sample tubes on the surface of Mars (Tiñini Alvarez et al., 2022). These tubes, deposited by the Perseverance rover, are essential to the Mars Sample Return campaign, a mission that aims to bring back Martian soil and rock samples to Earth. Our work focuses on supporting the Sample Fetch Rover (SFR), which will autonomously locate and retrieve these tubes.

The first stage of our system uses a modified object detection model to identify the sample tubes in images captured by the rover. Once localized, the second stage employs a keypoint estimation technique to accurately predict the tube’s position and orientation using monocular images. By combining cutting-edge deep learning methods with traditional computer vision algorithms, we ensured that the system remains lightweight and capable of operating under the mission’s tight constraints, such as limited computational power and memory.

The developed solution was then integrated and validated on a testbed at ESA’s ESTEC facility. This testbed consisted of an ESA-designed mobile manipulation platform, called RABBIT, which mimics the capabilities of a sample tube-fetching rover. The RABBIT platform allowed us to test and refine our approach in realistic conditions, ensuring that the system can perform efficiently in a Mars-like environment.

If first html tag is indented, and this include is called after a list, the html tag is considered a list element, and things go wrong. Leaving a hidden unindented line here
The RABBIT mobile manipulation platform at ESA's ESTEC facility, where our work was integrated and validated.

References

  1. Sample-Tube Pose Estimation Based on Two- Stage Approach for Fetching on Mars Surface.
    Israel Raul Tiñini Alvarez, Ignacio Perez, Tim Wiese, Laura Bielenberg, and Renaud Detry.
    In 16th Symposium on Advanced Space Technologies in Robotics and Automation, 2022.