Virtual Simulation for Testing Robotic Control Architectures for Autonomous Rendezvous Docking

John Karasinski, Christopher Lorenzen, and Melanie Stich

Department of Mechanical & Aerospace Engineering, University of California, Davis

The viability of a small inspection satellite with deliberative or reactive robotic architectures are discussed and methods for testing their robustness are developed. Some past docking and computer vision examples are presented as well as their applicable contributions. A navigation system primarily based on computer vision or laser based position determination is used to compare the success of two different control system architectures. Each architecture is applied to the same environment, and varying initial states are used to test their ability to adapt to different situations. The fuel burned and time to target are used to quantify the relative performance of these architectures. The proposed test environment and simulation methods are introduced to serve as a means for acquiring data to develop empirical cause and effect relationships between the different control design parameters.