Variable-magnification optical stimulator for training and validation of spaceborne vision-based navigation

Abstract

This paper addresses the design, calibration, and utilization of a hardware-in-the loop optical stimulator for the training and validation of vision-based sensors and algorithms used for spaceborne navigation. This testbed overcomes deficiencies of existing approaches in literature, which mainly focus on emulating single operational zones such as inertial, far-range, or close-range relative navigation. Consequently, the behavior of vision-based sensors and algorithms operating in mixed or transitioning navigation regimes have remained poorly assessed before launch. To address these issues, this novel optical stimulator reproduces the large geometric and radiometric dynamic range of the space environment with a pair of lenses to achieve variable magnification of an organic light emitting diode monitor. This approach is shown to support vision-based sensors of both narrow and wide field of view, increase the radiometric dynamic range of the testbed, and emulate close-range scenarios when light from an object is not collimated. A calibration procedure is presented that places sources of light at the intended angular location over the magnification range in the presence of varying distortions. The calibrated testbed is then used to simulate dynamic inertial and relative navigation scenarios of operational relevance. These hardware-in-the-loop simulations allow for the training and validation of optical hardware, software, and algorithms for future distributed space systems.