Sam Wishnek, BAE Systems
Keywords: maneuver, maneuver estimation, optical,
Abstract:
Maneuvering objects introduce significant additional challenges to the maintenance of object custody and space situational awareness. When an object with a known orbit undergoes an unknown maneuver, the maneuver introduces four new degrees of freedom. These are the time of the maneuver and the three elements of the delta-v vector. For an approximately instantaneous maneuver, the only remaining constraint is that the new and old orbits intersect at the point where the object maneuvered. With these constraints and two additional post-maneuver electro-optical observations, a unique time and delta-v vector can be found to describe the maneuver and post-maneuver orbit. Inspired by initial orbit determination algorithms, this research explains and assesses a new algorithm that uses this minimal amount of data to provide an initial maneuver estimate. Key metrics include the speed, accuracy, and stability of the solution under realistic dynamics and errors.
Much of the existing work on maneuver estimation tends to focus on either comparing pre-maneuver and post-maneuver orbits or running a filter over data that encompasses the maneuver time. Both methods assume a relatively large amount of data to perform the estimation and determine a precise solution. An important difference between prior work and the proposed work is that the proposed work can estimate the maneuver with the bare minimum amount of data. Three electro-optical observations are required for initial orbit determination when there is no prior information, so the initial maneuver determination algorithm necessarily leverages the pre-maneuver state to determine the post-maneuver state. The new algorithm provides a new capability to perform this estimation with too little data for existing algorithms to estimate the maneuver or post-maneuver state.
This approach is inspired by approaches taken by existing algorithms for initial orbit determination. The Gooding algorithm for angles-only initial orbit determination takes three angles-only observations of a space object, assumes range values for the first and third observation, and finds the orbit that fits the first and third position states. The first and third range values are iteratively corrected to minimize the error with the middle observation. The proposed maneuver determination approach uses a similar initial setup by applying trial range values to the latter angles-only observation and then replaces the other range guess with a trial value of the time or eccentric anomaly of the maneuver and then corrects based on the error between the predicted and observed first observation.
The research covers the method of the algorithms and their relative performance in time and accuracy running inputs with realistic dynamics and errors. The accuracy of the solution under measurement error is an essential metric for demonstrating the real-life utility of the algorithms. Analysis of the algorithm includes a build-up to progressively more realistic scenarios. The full analysis shows acceptable performance with realistic initial state error, measurement error, high precision orbital dynamics, and finite maneuvers. This is despite the proposed algorithm using simple two-body dynamics and an assumption of instantaneous maneuvers to optimize the algorithm for speed. Additional results include performance metrics for various maneuver magnitudes, observation geometries, and relative times between observations.
Date of Conference: September 17-20, 2024
Best Paper Award Winner 2024
Track: Astrodynamics