|
Original research
Pages 171-176
10.61552/JAI.2026.03.006
Mohammadparsa Sadoughifar
Abstract
Small satellites, particularly CubeSats, have become a cornerstone of modern space missions due to their low cost and versatile applications in Earth observation, communications, and scientific research. However, their size, weight, and power (SWaP) limitations make energy management one of the most critical challenges for mission success. Traditional methods—such as fixed solar panel operation, deterministic battery scheduling, and passive thermal control—are reliable and well established but often lack adaptability in dynamic orbital conditions.
Recent advances in artificial intelligence (AI) and machine learning (ML) have enabled intelligent energy management strategies capable of forecasting demand, detecting anomalies, and autonomously optimizing subsystem operations. Techniques including neural networks, support vector machines, decision trees, deep reinforcement learning, and edge computing demonstrate significant potential for extending mission lifetime and increasing autonomy. Emerging paradigms such as federated learning and digital twins further highlight future opportunities for predictive and collaborative energy management.
This review provides a comparative analysis of traditional and intelligent methods, emphasizing their respective strengths, limitations, and applicability. The study concludes that hybrid approaches—integrating robust traditional techniques with adaptive AI-driven strategies—represent the most promising path for next-generation CubeSat missions.
The findings are particularly relevant for CubeSat missions in Earth observation and communication, where real-time autonomy is increasingly critical.
Keywords: CubeSat, Energy Management, Artificial Intelligence, Machine Learning, Deep Reinforcement Learning.
Recieved: 22.08.2025. Revised: 29.09.2025. Accepted: 30.10.2025.
|
PDF 