Thibault Lebeke, Exotrail; Andrea Fiorentino, Exotrail
Keywords: low-thrust, electric propulsion, optimal control, operations, rendezvous proximity operations
Abstract:
Rendezvous and proximity operations once limited to the servicing of the International Space Station might be on the point of becoming the key technology for several commercial applications such as in-orbit servicing and tugging. Many potential customers, from commercial to military application, are eager to find a provider for this kind of space services. Once the ability to perform rendezvous and proximity operations is manage and demonstrate throw a dedicated mission a truly commercial services can be set.
The goal of this work is to build an environment of controller and simulator for rendezvous and proximity operations with low-thrust propulsion system. These elements aim to be deployed in commercial mission design and operations software with dedicated constraints:
Generic to fit with a variety of mission. This led to a modular approach where dynamics model and optimizer can be fit in a specific object in order to switch easily from one to another depending on the mission needs.
Robust to integrate operations constraint and manage safety with space agency guideline for rendezvous and proximity operations.
Realistic to compute a control achievable by the spacecraft.
Integrated in a specific simulation and space dynamics environment call by both mission design and operations software.
The main novelty of our approach is to cover the entire project cycle from early phases and simulations all the way to the generation of operationally viable maneuver plans.
Three kinds of proximity operations are planned to be integrated in this module:
Formation flying.
Inspection operations mission, object of this abstract.
Rendezvous mission, intended for future development.
Inspection operations mission is defined as the transfer of the chaser from his initial state to a dedicated stable circular relative orbit around the target. Here the initial state is considered close (about 20km here) to target.
At the heart of this module, a model-based controller that solves a convex program with two opposite objectives:
The minimum fuel maneuver plan.
The minimum error whereas mission goal.
To enhance the operational readiness of the proposed control plan, realistic characteristics of commercial low-thrust propulsion system are modelled:
Operation at a constant thrust value and limited duty cycle.
Attitude limitations can be enforced by only allowing thrusting in certain directions.
Due to the nature of rendezvous and proximity operations, safety and security constraints as to be taken in account. Guideline on safe close proximity operations from spatial agency is used as constraint to solve the convex problem. The controller ensures that the trajectory does not encounter a keep-out zone around the target.
The Clohessy-Wiltshire or the linearized Gauss variational equations can be used to predict the orbital motion during the computation of the control. Perturbations due to differential drag and differential gravity can be modelled as constant disturbances over the control horizon in order to further optimize the proposed maneuver plan
The controller is triggered based on more conservative bounds with respect to the hard mission-level constraints, thus allowing to account for unmodelled dynamics and other sources of error. Finally, as it is customary in spacecraft operations, numerically smoothed orbital states are fed to the controller to mitigate the undesired high-frequency oscillations caused by the asphericity of the Earth. With these smoothed orbital states, disturbance acceleration from unmodelled dynamics can be interpreted in the controller to find the best maneuver plan for our missions.
The controller is part of a complex simulation environment that has been developed and that will be described in full details. The inertial states of the different spacecrafts are propagated using state of the art mathematical routines and high-fidelity force models made available via the open-source astrodynamics library Orekit. The modularity of our approach allows to perform simulations starting from noisy observations which are filtered through an Unscented Kalman Filter to simulate a realistic operational workflow.
Date of Conference: September 27-20, 2022
Track: Conjunction/RPO