Differential Correction

Author: CislunarSpace
Website: https://cislunarspace.cn

Definition

Differential Correction is a core iterative numerical method in orbital mechanics that employs the State Transition Matrix (STM) to perform linearized corrections on orbital initial conditions, progressively converging the orbit toward a solution satisfying specific constraints (such as periodicity conditions, perilune altitude constraints, or passage point constraints). It forms the mathematical foundation of the Shooting Method and is an indispensable tool in virtually all numerical orbit design workflows.

In cislunar space dynamics research, differential correction is widely used to solve for the initial conditions of periodic orbits near libration points (Halo orbits, Lyapunov orbits, DRO, etc.), as well as for state matching at connection points in the patched method.

Core Elements

State Transition Matrix

The State Transition Matrix $\boldsymbol{\Phi}(t, t_0)$ describes the linear sensitivity of the orbital state to changes in initial conditions. Given the state vector $\mathbf{x} = [\mathbf{r}^T, \mathbf{v}^T]^T \in \mathbb{R}^6$ :

\delta \mathbf{x}(t) = \boldsymbol{\Phi}(t, t_0) \cdot \delta \mathbf{x}(t_0)

where $\delta \mathbf{x}$ is the state deviation. The STM satisfies the variational differential equation:

\dot{\boldsymbol{\Phi}}(t, t_0) = \mathbf{A}(t) \cdot \boldsymbol{\Phi}(t, t_0), \quad \boldsymbol{\Phi}(t_0, t_0) = \mathbf{I}_6

where $\mathbf{A}(t) = \frac{\partial \mathbf{f}}{\partial \mathbf{x}}\bigg|_{\mathbf{x}(t)}$ is the Jacobian matrix of the dynamical equations. In numerical propagation, the STM equation is typically integrated simultaneously with the orbital equations.

Basic Correction Algorithm

Let the target constraint be $\mathbf{g}(\mathbf{x}_f) = \mathbf{0}$ (e.g., periodic conditions requiring a specific relationship between terminal and initial states), and the current terminal state deviation be $\Delta \mathbf{x}_f$ . Using STM linearization:

\Delta \mathbf{x}_f \approx \frac{\partial \mathbf{g}}{\partial \mathbf{x}_f} \cdot \boldsymbol{\Phi}(t_f, t_0) \cdot \Delta \mathbf{x}_0

Defining the constraint sensitivity matrix to initial conditions as $\mathbf{M} = \frac{\partial \mathbf{g}}{\partial \mathbf{x}_f} \cdot \boldsymbol{\Phi}(t_f, t_0)$ , the initial condition correction is:

\Delta \mathbf{x}_0 = -\mathbf{M}^{\dagger} \cdot \mathbf{g}(\mathbf{x}_f)

where $\mathbf{M}^{\dagger}$ is the pseudo-inverse of $\mathbf{M}$ (reducing to the regular inverse when $\mathbf{M}$ is square). Iterate until $\|\mathbf{g}(\mathbf{x}_f)\| < \varepsilon$ ( $\varepsilon$ is the convergence tolerance).

Periodic Orbit Differential Correction

For symmetric periodic orbits in the CR3BP (such as DRO), the symmetry properties can simplify differential correction:

DRO Shooting Conditions ( $x$ -axis symmetric):

DRO orbits are symmetric about the $x$ -axis, so only a half-period needs to be integrated. At the $x$ -axis crossing point, the constraints are:

y(t_f) = 0, \quad \dot{x}(t_f) = 0, \quad \dot{z}(t_f) = 0

The free variables are $\dot{y}_0$ from the initial conditions and the orbital period $T$ (or half-period $T/2$ ). The correction equations are:

\begin{bmatrix} \delta \dot{y}_0 \\ \delta T/2 \end{bmatrix} = -\mathbf{M}^{-1} \begin{bmatrix} y_f \\ \dot{x}_f \end{bmatrix}

where $\mathbf{M}$ is the $2 \times 2$ sensitivity matrix formed from the relevant components of the STM.

Application in the Patched Method

Wei et al. (2026) used differential correction in their patched method for DRO transfer trajectory design in three ways:

LEO-to-perilune arc correction: Adjusting the LEO departure velocity so that the trajectory satisfies perilune altitude constraints upon reaching the Moon
Perilune-to-DRO arc correction: Correcting initial conditions from the perilune so that the terminal state converges to the target DRO
Patching point matching correction: When the velocities of two arcs at the perilune do not match, correcting parameters of each arc to reduce the velocity difference $\|\Delta \mathbf{v}_{\text{PLF}}\|$

Single-Shooting vs. Multi-Shooting

Method	Description	Applicable Scenario
Single-shooting	Integrate from initial point to terminal point in one pass, correct initial conditions using STM	Short arcs, simple constraints
Multi-shooting	Divide the orbit into multiple segments, independently integrate each and impose continuity constraints at nodes	Long arcs, complex constraints, high precision

Multi-shooting uses separate STMs for each segment, coupling segments through state matching constraints at nodes. Compared to single-shooting, multi-shooting offers better numerical stability and faster convergence.

Convergence Analysis

The convergence of differential correction is influenced by:

Quality of the initial guess: Closer to the true solution means more accurate linearization and faster convergence
Nonlinearity of constraints: Strongly nonlinear constraints may require more iterations or damping strategies
Numerical accuracy of the STM: Long-duration propagation may accumulate STM errors, affecting correction direction accuracy
Matching of free variables to constraints: The number of free variables should equal the number of constraints; otherwise the problem is under- or over-determined

Typical convergence is quadratic (Newton method property), with 3-5 iterations typically sufficient to reach $10^{-12}$ precision when the initial guess is reasonably good.

Application Value

The central role of differential correction in cislunar space orbit design is reflected in:

Periodic Orbit Solving: The initial conditions of all periodic orbit families — Halo, Lyapunov, DRO, and others — depend on differential correction
Transfer Orbit Optimization: In both patched and direct methods, differential correction adjusts design parameters to satisfy terminal constraints
Orbit Keeping Strategy: Orbit keeping maneuver design is fundamentally a differential correction problem — determining the minimum velocity increment to return the orbit to its nominal trajectory
Mission Feasibility Assessment: Rapid differential correction can determine whether a transfer scheme satisfies all constraint conditions

References

Wei Z, et al. Research on powered lunar flyby transfer injection to cislunar distant retrograde orbit families[J]. Journal of Beijing University of Aeronautics and Astronautics, 2026.
Zimovan E M. Characteristics and design strategies for near rectilinear halo orbits within the Earth-Moon system[D]. Purdue University, 2017.
Howell K C, Pernicka H J. Numerical determination of Lissajous trajectories in the restricted three-body problem[J]. Celestial Mechanics, 1987, 41(1-4): 107-124.
Pavlak T A, Howell K C. Trajectory design in the planar circular restricted three-body problem using polynomial target maps[C]. AAS/AIAA Astrodynamics Specialist Conference, 2010.