Characterization and Implementation of a Vision-Based 6-DOF Localization System

摘要

Testing and validation of flight hardware in ground-based facilities ca n result in significant cost savings and risk reduction. We focus particular attention on ground validation for small satellite proximity operations. We designed a relative motion emulation system for aerospace vehicles using Relative Motion Vehicles (RMVs) capable of large planar motion and limited 6 degree-of-freedom motion. Though the state of the RMV is approximated through internal sensors, an external measurement system is required to correct for secular drift. The commercial NorthStar system from Evolution Robotics provides a vision-based 2-D localization solution using active infrared (IR) beacons. This paper addresses characterization and implementation of the NorthStar system as a 6-DOF state measurement system. The noise characteristics of the NorthStar Detector were determined, as were the output characteristics of the IR light emitting diode (LED) beacons. These characteristics were used to produce a high fidelity model of the measurement system in which multiple Detectors are placed around the perimeter of the workspace and multiple beacons are installed on the RMV. Simulation and hardware test results demonstrate the efficacy of this approac h. Nomenclature