Thruster-based ADCS Momentum Management for Deep-Space CubeSats Arcsec Space

Abstract

The CubeSat platform has revolutionized access to space, enabling a new era of low-cost, rapid-development missions. While these nanosatellites have demonstrated widespread utility in Low Earth Orbit (LEO) applications, there is growing interest from space agencies and academia in deploying them in deep-space environments. However, traditional LEO actuators, such as magnetorquers, rely on an ambient magnetic field and are consequently ineffective in deep space. To address this hardware gap, this thesis investigates a thruster-based approach for deep-space momentum management, specifically focusing on reaction wheel desaturation and post-deployment detumbling maneuvers to define the capability of an universal deep-space AOCS (based on an ESA future commercial call). The primary objectives are to investigate thruster control in CubeSats and to evaluate the impact of different control strategies on overall maneuver performance. Because thrusters are discrete (on/off), unipolar actuators, they introduce allocation challenges compared to continuous torque providers. This work explores thruster control by comparing heuristic bang-bang strategies against pulse-modulated continuous approximation architectures. For the latter, the software pipeline is divided into three distinct stages: a continuous control law, a thruster allocation logic (evaluating pseudo-inverse versus direct-table mapping, and clipping versus proportional scaling), and a discrete pulse modulator (PWM versus impulse accumulation). The algorithms are validated within a Simulink environment, adapting an existing magnetorquer-based ADCS model provided by arcsec. The simulated plant features a fixed, four-nozzle cold-gas propulsion system, representing a highly constrained architecture at the boundary of 3-DOF rotational control capability. To manage the coupled dynamics during desaturation, a dual-loop control architecture featuring a feed-forward path is implemented. Simulation results demonstrate that while heuristic bang-bang control provides highly robust and rapid actuation, continuous architectures offer superior mission extendibility. Crucially, the comparative analysis reveals that continuous controllers demand strict direction-preserving allocation (proportional scaling) and memoryless modulation (PWM) to prevent cross-axis lock-outs and post-maneuver pointing spikes. Ultimately, this master thesis establishes an algorithmic and design framework for constrained momentum management, laying a foundation for the development of a versatile AOCS adequate deep-space CubeSat missions.