A New Mixed Iterative Algorithm to Solve the Fuel-Optimal Linear Impulsive Rendezvous Problem

Denis Arzelier, Christophe Louembet, Aude Rondepierre, C. Kara‐ZaïtriHAL 2013

TL;DR

This paper introduces a novel "mixed iterative algorithm" to efficiently solve the complex problem of fuel-optimal spacecraft rendezvous. By combining algebraic geometry with calculus of variations, the algorithm can determine the optimal number and timing of thruster firings, overcoming limitations of previous methods that required pre-specifying these parameters.

What Problem Does This Paper Address?

Planning rendezvous missions for spacecraft, such as docking with the International Space Station or forming satellite constellations, is a critical but computationally challenging task. A key objective is to minimize fuel consumption while ensuring the spacecraft arrive at the same location at the same time. This problem, known as the "fuel-optimal impulsive rendezvous problem," becomes particularly difficult when the number of required thruster firings (impulses) is not known beforehand. Existing methods often struggle to optimize this number of impulses or guarantee global optimality for a fixed number of impulses. This paper aims to address these limitations by developing a more robust and efficient algorithm.

Methods

The core of this paper is a "mixed iterative algorithm" that combines two main approaches:

  • Algebraic Formulation of Optimality Conditions: The authors leverage Carter's necessary and sufficient conditions for optimality, which are expressed in terms of polynomial equations. This allows them to utilize powerful numerical tools from algebraic geometry.
  • Calculus of Variations (Lion-Handelsman's Conditions): The algorithm builds upon the work of Lion and Handelsman, which uses variational principles to iteratively improve trajectories by adding or shifting impulses.

The algorithm starts with an initial two-impulse trajectory and iteratively refines it. It intelligently adds new impulses or adjusts existing ones based on conditions derived from the primer vector theory and algebraic geometry. A key innovation is that this approach avoids the need for local optimization schemes, which can lead to issues like "cusps" in the primer vector norm and potential sub-optimality. The algorithm is designed to handle both circular and elliptic reference orbits.

Datasets

Not discussed in the paper. The paper focuses on developing and testing an algorithm using established theoretical examples and a realistic mission scenario.

Key Insights and Ideas

  • Bridging the Gap: The paper successfully integrates algebraic geometry tools with variational calculus to tackle a long-standing problem in astrodynamics. This interdisciplinary approach is a significant contribution.
  • Optimizing the Number of Impulses: A major breakthrough is the algorithm's ability to implicitly optimize the number of impulses, rather than requiring it to be pre-specified. This is crucial for finding truly optimal solutions.
  • Avoiding Local Minima: By avoiding local optimization steps, the algorithm is less prone to getting stuck in sub-optimal solutions and can achieve smoother primer vector norms, which are indicative of optimality.
  • Efficiency and Robustness: The algorithm demonstrates efficiency in practice, finding optimal solutions within reasonable computation times and showing robustness across various orbital scenarios (circular, elliptic, and a realistic mission).

Key Results

The paper presents several numerical examples to validate the algorithm:

  • Carter's Examples: The algorithm successfully reproduces and, in some cases, improves upon results from previous studies (e.g., Carter [8], [20]) for academic rendezvous problems. For Carter's first example, the algorithm converges to a four-impulse solution that is a tight approximation of the optimal two-impulse solution, outperforming Carter's original two-impulse solution.
  • PRISMA Mission Case Study: A more realistic scenario based on the PRISMA mission demonstrates the algorithm's applicability to complex, real-world problems. The algorithm achieves a fuel-optimal three-impulse solution with a cost that is 27% lower than a suboptimal two-impulse strategy.
  • Computational Efficiency: The algorithm is shown to be computationally efficient, finding optimal solutions within seconds or minutes, even for complex scenarios. For instance, the PRISMA case study was solved in 13 seconds with 3 iterations.

Limitations and Future Work

  • Analytical Proof of Convergence: While the algorithm performs well in practice, a formal analytical proof of convergence is not provided.
  • Simplified Dynamics: The current model uses linearized relative motion and assumes impulsive maneuvers. Future work could incorporate more sophisticated transition matrices that account for orbital perturbations like atmospheric drag or gravity harmonics (e.g., J2 term).
  • Collision Avoidance: The authors suggest extending the algorithm to include collision avoidance constraints for optimal trajectory planning.
  • Local Optimization Avoidance: While the algorithm avoids local optimization schemes, further research could explore how to eliminate cusp occurrences in the primer vector norm graph more rigorously.

Summary updated

BibTeX
@inproceedings{arzelier2013new,
  author = {Denis Arzelier and Christophe Louembet and Aude Rondepierre and C. Kara‐Zaïtri},
  title = {{A New Mixed Iterative Algorithm to Solve the Fuel-Optimal Linear Impulsive Rendezvous Problem}},
  booktitle = {HAL},
  year = {2013},
}
Figure