Ricerca CERCA

Vision-based pose estimation of an unknown and uncooperative object for on-orbit servicing and debris removal applications

azienda Thesis in external company    


Reference persons MARTINA MAMMARELLA

External reference persons Fabrizio Dabbene (CNR-IEIIT); Federica Paganelli (AIKO); Luca Romanelli (AIKO)

Research Groups 04-Automazione e Robotica


Description Nowadays all the major space agencies and industries are undertaking the enhancement of the automation capabilities of spacecraft. Autonomous operations are extremely advantageous for space missions, with a wide range of possible applications. For example, autonomous relative navigation around unknown and uncooperative objects is particularly appealing. Precise pose and motion estimation of an uncooperative object, such as a Resident Space Object (RSO) has a potential utilization in the domain of space debris removal and on-orbit servicing. This thesis aims to explore the potential of combining vision-based pose estimation with filtering methods. In today's context, vision-based systems offer a promising approach to achieve highly accurate relative trajectory estimation for satellites. By equipping satellites with cameras, we not only enhance their ability to investigate Resident Space Objects (RSOs) but also expand their sensor capabilities for autonomous navigation tasks. The initial pose estimate obtained from the camera, which includes information about the relative position and attitude between the observing satellite and the RSO, can be integrated into the navigation filter. This data can then be fused with signals from other sources such as GPS, Sun sensors, Star trackers, IMU, and more.
The candidate will be responsible for developing a navigation system that integrates information from the camera and the onboard sensors. The main objective is to address a specific use case involving a Low Earth Orbit RSO, which refers to debris that needs to be captured and guided toward a re-entry trajectory.
The project will involve a range of different activities, starting from modeling the sensors installed on board to designing a robust filtering technique for accurately estimating the pose of the RSO throughout its orbit.
The candidate will review and explore different filtering techniques, considering both classical and Artificial Intelligence approaches, and select the most suitable one based on the mission scenario, computational limitations, and robustness requirements.

See also  cnr+aiko - thesis proposal.pdf 

Required skills  Estimation theory sensor fusion, computer vision, robotics
MATLAB/Simulink and Python coding skills; C++ preferred but not required
Spoken/written English (excellent)

Deadline 12/07/2024      PROPONI LA TUA CANDIDATURA

© Politecnico di Torino
Corso Duca degli Abruzzi, 24 - 10129 Torino, ITALY