Abstract

In this paper an original, easy to reproduce, semi-analytic calibration approach is developed for hardware-in-the-loop performance assessment of pose determination algorithms processing point cloud data, collected by imaging a non-cooperative target with LIDARs. The laboratory setup includes a scanning LIDAR, a monocular camera, a scaled-replica of a satellite-like target, and a set of calibration tools. The point clouds are processed by uncooperative model-based algorithms to estimate the target relative position and attitude with respect to the LIDAR. Target images, acquired by a monocular camera operated simultaneously with the LIDAR, are processed applying standard solutions to the Perspective-n-Points problem to get high-accuracy pose estimates which can be used as a benchmark to evaluate the accuracy attained by the LIDAR-based techniques. To this aim, a precise knowledge of the extrinsic relative calibration between the camera and the LIDAR is essential, and it is obtained by implementing an original calibration approach which does not need ad-hoc homologous targets (e.g., retro-reflectors) easily recognizable by the two sensors. The pose determination techniques investigated by this work are of interest to space applications involving close-proximity maneuvers between non-cooperative platforms, e.g., on-orbit servicing and active debris removal.

Highlights

  • Pose determination is a very important task for space activities like On-Orbit Servicing (OOS)and Active Debris Removal (ADR), in which a main spacecraft has to approach a man-made target to carry out inspection, repair, maintenance or refueling operations [1], as well as safe disposal [2].Among these activities, autonomous refueling is receiving increasing attention due to its promising economic return [3]

  • With regards to the pose determination sensors, a 3D scanning LIDAR, namely the VLP-16 produced by Velodyne (San Jose, CA, USA) [36], and a monocular camera, namely the AVT Marlin F-145C2 produced by Allied Vision

  • This paper presented a strategy for hardware-in-the-loop performance assessment of algorithms designed to estimate the relative position and attitude of uncooperative, known targets by processing range measurements collected using active LIDAR systems

Read more

Summary

Introduction

Active Debris Removal (ADR), in which a main spacecraft (commonly called chaser) has to approach a man-made target (e.g., operative/inoperative satellites or abandoned rocket bodies) to carry out inspection, repair, maintenance or refueling operations [1], as well as safe disposal [2]. Among these activities, autonomous refueling is receiving increasing attention due to its promising economic return [3]. Autonomous landing on unknown environment must rely on robust terrain-relative navigation algorithms In this respect, LIDAR represent a promising solution [5]

Objectives
Methods
Conclusion
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call