I Can Hear You Coming: RF Sensing for Uncooperative Satellite Evasion

TL;DR

The paper tackles robust, autonomous adversary avoidance for satellites in contested environments by leveraging intercepted RF signals to inform a constrained reinforcement learning policy. It introduces a Cat & Mouse scenario where a mouse satellite uses RF-based localization of a non-cooperative cat and a constrained SAC-based RL policy, augmented by an MPC thrust controller, to minimize deviation from its origin while maintaining a safe distance. The framework is validated against two baselines (Delta-V optimization and Greedy Recursive Search) using both simulated data and real-world SSN trajectories, demonstrating that RF-informed CRL provides smoother, fuel-efficient evasion with superior performance under noisy observations. This work highlights RF sensing as a scalable, onboard sensing modality that can enhance space-domain awareness and autonomous maneuvering in future congested and contested space operations.

Abstract

This work presents a novel method for leveraging intercepted Radio Frequency (RF) signals to inform a constrained Reinforcement Learning (RL) policy for robust control of a satellite operating in contested environments. Uncooperative satellite engagements with nation-state actors prompts the need for enhanced maneuverability and agility on-orbit. However, robust, autonomous and rapid adversary avoidance capabilities for the space environment is seldom studied. Further, the capability constrained nature of many space vehicles does not afford robust space situational awareness capabilities that can be used for well informed maneuvering. We present a "Cat & Mouse" system for training optimal adversary avoidance algorithms using RL. We propose the novel approach of utilizing intercepted radio frequency communication and dynamic spacecraft state as multi-modal input that could inform paths for a mouse to outmaneuver the cat satellite. Given the current ubiquitous use of RF communications, our proposed system can be applicable to a diverse array of satellites. In addition to providing a comprehensive framework for training and implementing a constrained RL policy capable of providing control for robust adversary avoidance, we also explore several optimization based methods for adversarial avoidance. These methods were then tested on real-world data obtained from the Space Surveillance Network (SSN) to analyze the benefits and limitations of different avoidance methods.

Figures (10)

• Figure 1: Our proposed sensing and control architecture for a Cat and Mouse scenario.
• Figure 2: The ECEF (black), orbital (red), and Hill (yellow) reference frames, as well as orbital angles ($\Omega, i, \omega$) and mean anomaly ($M$) where $\boldsymbol{p}_o$ is the point on the orbit that serves as the origin to the Hill frame.
• Figure 3: An image depicting the constellation configuration used, where the red dots resemble a single satellite in the constellation.
• Figure 4: The simulated error between the cat position estimated by a TDOA localization algorithm and its true position during a single run of the simulation.
• Figure 5: The trajectories the cat spacecraft during several randomly generated scenarios in simulation.