Real-Time Estimation and Control for Orbital Pursuit–Evasion with Angles-Only Navigation

The space environment is vital for civil and military applications, yet it is increasingly vulnerable to threats that could compromise key assets and services. This work presents different autonomous evasive strategies integrating angles-only navigation with real-time onboard control, forming a foundation for autonomous space defense. Rooted in pursuit–evasion games, the proposed framework addresses the limited observability brought by angular measurements through appropriate relative dynamic model selection. Initially, zero-sum game theory is used to formalize the problem, leveraging state-dependent Riccati equation control to optimize evasion with onboard power constraints. Then, the full non-zero-sum game problem is introduced. This framework allows the introduction of realistic interactions where adversaries do not share a common objective and have uncertain and incomplete information about each other. Real-time state and parameter simultaneous estimation is executed through a tailor-designed unscented Kalman filter, improving evasion optimality by enhancing system observability. Various simulations validate the proposed strategy, demonstrating robustness in diverse encounter scenarios. This research advances space defense by improving evasive control and tactics, enhancing the survivability of satellite operations in hostile, uncertain environments.