Patterns of Life and Maneuver Detection for Cislunar Trajectory Maintenance

Karina Rivera, University of Colorado Boulder; Anthony Zara, University of Colorado Boulder; Daniel Aguilar Marsillach, University of Colorado Boulder; Marcus Holzinger, University of Colorado Boulder; Ian Elliott, University of Colorado Boulder; Natasha Bosanac, University of Colorado Boulder

Keywords: Cislunar Space Domain Awareness, Patterns of Life, Receding Horizon Control, Cislunar Space, Maneuver Detection, Optimal Control-Based Estimator

Abstract:

In the past decades, interest in cislunar space exploration has increased predominantly due to its value for testing novel technologies, interplanetary mission staging, lunar exploration, and Earth-to-Moon communication. With missions such as NASA’s Lunar Orbital Platform-Gateway, China’s Chang’e-5 mission, and India’s Chandrayaan-2 orbiter, to name a few, the need for space domain awareness (SDA) capabilities within cislunar space is becoming essential to ensure the safety of future cislunar operations. SDA capabilities in cislunar space allow operators to gather, transmit and analyze space objects’ information accurately; this serves as a decision-making tool for future and current mission planning. The safety and security of cislunar activities could be affected if there is not an effective understanding of the space domain. Currently, there are efforts that begin to explore methodologies and frameworks to support SDA; however, there are still a number of variables in cislunar space yet to be understood to have an accurate prediction, detection, and tracking of space vehicles to support cislunar SDA.        

An important element of cislunar SDA is the understanding of the space environment and how it affects the patterns of life of missions. Patterns of life establish a set of expected actions and movements which are used to determine a space vehicle’s current behavior and predict its future behavior. To support SDA operations, it is useful to possess the ability to predict and characterize space objects’ nominal behavior in cislunar space and to detect when this behavior changes. Since the dynamics of cislunar space are unstable and highly sensitive to disturbances, in this work “nominal behavior” refers to vehicles in orbits that possess stationkeeping control. Hence, the contribution of this paper is the identification and characterization of space vehicles’ patterns of life. This necessitates the ability to recognize nominal and anomalous maneuver sequences, to aid the development of cislunar SDA architectures.  

Maneuvers can be detected and estimated using the optimal control-based estimator (OCBE), which provides a filtered state, adjoint, and control estimate. In this work, the ballistic form of the OCBE is used, which assumes the nominal space vehicle’s trajectory uses no control i.e. is coasting or evolving under natural dynamics, since most often, space vehicles are not thrusting. This assumption limits the current implementation of the OCBE to modeling impulsive maneuvers only. The resultant state and control estimates are used to train a one-class support vector classifier (SVC) to build a model of the observed space vehicle’s pattern of life. This model is then evaluated on incoming data in real time to detect anomalous behavior with respect to the space vehicle’s expected station-keeping strategy without a-priori knowledge of possible changes. As such, the OCBE is used for both maneuver and behavior change detection.

There are an infinite number of candidate orbits in cislunar space to analyze space vehicles’ nominal maneuver sequences and obtain the respective patterns of life. Therefore, to simplify the problem, this paper focuses on a variety of trajectories that exist near natural repeating orbits around the Earth-Moon L1 and L2 Lagrange equilibrium points. Since a majority of the selected orbits have unstable behavior, this paper studies L2 power-optimal (i.e., electric propulsion) and impulsive (i.e., chemical propulsion) trajectories designed through a receding horizon stationkeeping control approach. The patterns of life for the selected trajectories are first studied in simplified dynamical model known as the circular restricted three-body problem (CR3BP). However, a higher fidelity full ephemeris model is used to provide a more accurate representation of the expected behavior of these orbits in cislunar space.

Date of Conference: September 14-17, 2021

Track: Cislunar SSA

View Paper