Hardware in the Loop Performance Assessment of LIDAR-Based Spacecraft Pose Determination

Area of Science:

• Robotics and Spacecraft Systems Engineering
• Computer Vision and Sensor Fusion
• Geospatial Information Science

Background:

• Accurate pose determination is critical for autonomous spacecraft operations, especially during close-proximity maneuvers like on-orbit servicing.
• Existing methods for calibrating Light Detection and Ranging (LIDAR) and camera systems often rely on specific, easily recognizable targets, which may not be practical in space.
• Evaluating the performance of LIDAR-based pose estimation algorithms requires a reliable benchmark, typically derived from high-accuracy camera-based measurements.

Purpose of the Study:

• To develop and validate an original, semi-analytic calibration approach for hardware-in-the-loop performance assessment of pose determination algorithms.
• To enable accurate extrinsic calibration between LIDAR and monocular camera systems without requiring ad-hoc homologous targets.
• To establish a benchmark for evaluating LIDAR-based pose estimation accuracy using camera-derived pose data.

Main Methods:

• A laboratory setup was created using a scanning LIDAR, a monocular camera, and a scaled satellite-like target replica.
• Point cloud data from the LIDAR was processed using model-based algorithms to estimate target pose.
• Simultaneously acquired camera images were processed using Perspective-n-Points solutions to obtain benchmark pose estimates.
• An original calibration method was implemented to precisely determine the extrinsic relative calibration between the camera and LIDAR.

Main Results:

• The developed semi-analytic calibration approach successfully enabled hardware-in-the-loop performance assessment of LIDAR-based pose determination algorithms.
• The calibration method proved effective without the need for specialized, sensor-specific targets.
• High-accuracy pose estimates from the camera provided a reliable benchmark for evaluating LIDAR-derived pose accuracy.

Conclusions:

• The proposed calibration technique offers an easy-to-reproduce and effective solution for sensor calibration in pose determination tasks.
• This method is highly relevant for space applications requiring precise relative navigation between non-cooperative platforms.
• The findings contribute to advancing capabilities for on-orbit servicing, active debris removal, and other close-proximity space operations.