Overview
The project investigates the suitability of a development and testing approach based on synthetic imagery for vision-based navigation (VBN) in space. To this end, the transferability of AI-based pose-estimation models developed in the Visual Servoing Testbed (VSTB) to physical mock-up environments is systematically analyzed.
The results are intended to enable a more efficient development process in which synthetic environments are used for early-stage development and optimization, while physical test environments are used for validation under realistic domain-gap conditions. This provides the basis for further hardware-oriented development of the Smart Space Camera (SSC) in future ESA programmes.
