Adaptive Relative Orbit Control for Spacecraft Using On/Off Thrusters Under Uncertain Disturbances

This paper proposes an adaptive relative orbit control law using ON/OFF thrusters for active debris removal (ADR). Relative orbit control with non-cooperative targets is essential for ADR missions. However, uncertain disturbances, such as atmospheric drag, exist in low Earth orbit. In addition, the ADR mission is expected to be a long-term mission. Thus, a highly autonomous control law that accounts for these uncertainties is required. This paper employs deep model-reference adaptive control (DMRAC) to achieve high-precision relative orbit control without measuring relative acceleration. DMRAC incorporates deep neural networks (DNNs) and a generative network into the model-reference adaptive control (MRAC) framework. Furthermore, the proposed controller accounts for the constraints of ON/OFF thrusters. The performance of the proposed controller is examined through numerical simulations.