- Partnership Projects
- Core Competitiveness
- Future Preparation
- Space Solutions
- How to Apply
- Our Projects
The use of electric propulsion for orbit-raising is a key factor to reduce the cost of access to space. The counterpart is a longer Launch and Early Operations Phase (LEOP) that induces additional cost for the ground segment and the operators. Introducing on-board autonomy would allow reducing the operational cost of such long EOR.
The advent of GNSS-based navigation solutions is a major enabler of autonomous orbit determination for telecom satellites. By adding an autonomous guidance function, it is then possible to significantly enhance autonomy during transfer.
In the reference process, all the guidance tasks are performed by ground and uploaded on the spacecraft: the main objective of the activity is to design and validate innovative guidance optimization algorithms that can replace the ground-based optimization tools. The algorithms must be simpler, as the available on-board CPU is much more constrained than for a ground-based solution.
The other objectives of the study are to establish that operations costs are reduced by at least 80%, that the algorithms can run in real-time on a real processor, and that the marginal loss in optimality introduced by the on-board CPU constraint is acceptable and compensated by the expected LEOP cost reductions.
The first main challenge is the selection and improvement of a most promising algorithmic approach among many existing optimization techniques, compatible with strict CPU constraints on one hand and complex attitude and operational constraints on the other hand.
The second main challenge is to validate the algorithm prototype in real-time in a realistic hardware environment, relying on autocoding processes to minimize the effort from concept to implementation.
In a context where electrical orbit raising represents a growing share of LEOP situations, the chief benefit is a reduction in operations costs (the project’s target is a five-fold reduction). In addition to the reduction in cost in itself, this allows to optimize the allocation of resources over the long duration of the transfer and avoid bottlenecks in case of multiple simultaneous launches (e.g. LEO-to-LEO orbit raising for large constellations).
The major output of the study is the detailed definition of an on-board optimization algorithm capable of computing a complete attitude and thrust profile for low-thrust electrical orbit-raising, for both LEO and GEO transfers. This algorithm is also implemented as a real-time software prototype, running in a processor-in-the-loop setup.
The optimum trajectory (both thrust direction and attitude profile compatible with AOCS and power constraints) found by the autonomous on-board guidance function was demonstrated to be within less than 1% of the true optimum in terms of transfer time.
The complete architecture of a system based on autonomous guidance for EOR involves the following major modules:
The first task involves the review of candidate algorithms, as well as the definition of a reference mission. The second task is dedicated to establishing the detailed requirements (functional, performance, software). In task 3, the candidate algorithms are selected and investigated in more detail, in order to assess feasibility and relative merit. Task 4 deals with fast prototyping of the candidate solution, based on automatic S/W generation and implementation. Task 5 is the verification and validation on the real-time test-bench, with a processor in the loop.
The study is complete. This study has solidly established the possibility to reduce operations costs substantially for long electric-propulsion transfers, taking advantage of future GNSS capabilities. Two strategies are proposed:
The latter strategy was developed in detail, implemented in a representative simulation environment, for demonstrating performance and robustness. As this on-Board Guidance Function Software (OBGF-SW) prototype was developed following formal coding standards, we could then produce the flight-ready code via automatic code-generation, and verify compatibility with on-board hardware during Processor-In-the-Loop tests on a LEON3 board.