Kinematic Differential Equations

1.4 Kinematic Differential Equations

Kinematic differential equations describe how the orientation of a rigid body evolves with time when its angular velocity is known. These equations link the body angular velocity vector to the time derivatives of the commonly used attitude representations:

• Euler angles
• Direction Cosine Matrix (DCM)
• Quaternions

Together they form the backbone of spacecraft and aircraft attitude propagation.

1.4.1 Euler Angle Rate Equations

Euler angles provide a three-parameter description of orientation, but their time derivatives are not simply the components of the angular velocity vector. Instead, a nonlinear mapping relates the body angular velocity to the Euler angle rates.

For the widely used intrinsic 3–2–1 (yaw–pitch–roll) sequence, let the Euler angles be \(\phi\) (roll), \(\theta\) (pitch), \(\psi\) (yaw), and let the body angular velocity in the body frame be \(\boldsymbol{\omega} = [\,p,\ q,\ r\,]^T\). The relation between body rates and Euler angle rates is

\[ \begin{bmatrix} p \\[2pt] q \\[2pt] r \end{bmatrix} = \begin{bmatrix} 1 & 0 & -\sin\theta \\ 0 & \cos\phi & \sin\phi\cos\theta \\ 0 & -\sin\phi & \cos\phi\cos\theta \end{bmatrix} \begin{bmatrix} \dot\phi \\[2pt] \dot\theta \\[2pt] \dot\psi \end{bmatrix}. \]

The coefficient matrix becomes singular when \(\theta = \pm 90^\circ\), reflecting the well-known gimbal lock of Euler angles. To obtain \(\dot\phi, \dot\theta, \dot\psi\), we must invert this system, which is only possible away from the singular attitudes.

Key points

• The mapping contains trigonometric functions and is numerically sensitive.
• Singularities occur when the second angle aligns two of the axes.
• Euler angles are good for human interpretation, but not ideal for long-duration propagation.

1.4.2 DCM Kinematic Equation

The Direction Cosine Matrix \(C\) transforms vector components between two reference frames. Its time derivative is directly related to the angular velocity through a skew-symmetric matrix.

Define the skew-symmetric matrix of angular velocity \(\boldsymbol\omega = [\,\omega_1,\ \omega_2,\ \omega_3\,]^T\):

\[ \Omega = \begin{bmatrix} 0 & -\omega_3 & \omega_2 \\ \omega_3 & 0 & -\omega_1 \\ -\omega_2 & \omega_1 & 0 \end{bmatrix}. \]

The kinematic differential equation for the DCM is then

\[ \dot{C} = C\,\Omega. \]

Interpretation

• Each row of \(C\) evolves according to the cross product between the corresponding basis vector and the angular velocity.
• In exact arithmetic, the orthonormality of \(C\) is preserved by this equation.
• In numerical work, accumulated errors require occasional re-orthonormalisation of \(C\).

Element-wise, each component \(C_{ij}\) satisfies a first-order ODE obtained from the matrix equation. The formulation is compact and free of singularities.

1.4.3 Quaternion Kinematic Equation

Quaternions provide a four-parameter, globally nonsingular representation of orientation. Let the unit quaternion be

\[ \mathbf q = \begin{bmatrix} q_0 \\ q_1 \\ q_2 \\ q_3 \end{bmatrix}, \qquad q_0 \text{ (scalar part)}. \]

Define the \(4\times4\) matrix built from \(\boldsymbol\omega\):

\[ \Omega_q = \begin{bmatrix} 0 & -\omega_1 & -\omega_2 & -\omega_3 \\ \omega_1 & 0 & \omega_3 & -\omega_2 \\ \omega_2 & -\omega_3 & 0 & \omega_1 \\ \omega_3 & \omega_2 & -\omega_1 & 0 \end{bmatrix}. \]

The quaternion kinematic equation is

\[ \dot{\mathbf q} = \frac{1}{2}\,\Omega_q\,\mathbf q. \]

Splitting into scalar and vector parts with \(\boldsymbol\varepsilon = [\,q_1,\ q_2,\ q_3\,]^T\), we obtain

\[ \dot q_0 = -\tfrac12\left(q_1\omega_1 + q_2\omega_2 + q_3\omega_3\right), \] \[ \dot{\boldsymbol\varepsilon} = \tfrac12\bigl(q_0 I_3 + [\boldsymbol\varepsilon]_\times\bigr)\boldsymbol\omega, \]

where \([\boldsymbol\varepsilon]_\times\) is the skew-symmetric matrix of \(\boldsymbol\varepsilon\).

Why quaternions are preferred

• No attitude singularities.
• Linear in \(\boldsymbol\omega\) when angular velocity is known, which simplifies integration.
• Easy to renormalise: \(\mathbf q \leftarrow \dfrac{\mathbf q}{\|\mathbf q\|}\).
• Standard choice for spacecraft EKF/UKF attitude filters, star trackers, and flight software.

1.4.4 Numerical Integration Concepts

Regardless of which attitude representation is used, numerical integration is required when the angular velocity \(\boldsymbol\omega(t)\) varies with time.

A. Euler Angles

• Highly sensitive to singularities in the chosen sequence.
• Require trigonometric evaluations at every step.
• Not ideal for high-fidelity, long-duration propagation.

B. Direction Cosine Matrix

• Simple linear ODE: \(\dot{C} = C\Omega\).
• Good accuracy and conceptually clean.
• Requires periodic re-orthonormalisation (e.g. SVD or Gram–Schmidt) to control numerical drift.

C. Quaternions

• Most robust representation for numerical propagation.
• Integration is stable and avoids gimbal lock.
• Only one scalar constraint (unit norm) to enforce.
• Widely used in modern spacecraft and UAV systems.

D. Integration Methods

• Forward Euler – conceptually simple but low order.
• Runge–Kutta (e.g. RK2, RK4) – common for onboard propagation.
• Embedded RK schemes – allow adaptive step-size control.
• Lie group / geometric integrators – preserve the SO(3) structure (advanced).

E. Practical Considerations

• Renormalise quaternions regularly to maintain unit norm.
• Choose sufficiently small timesteps for numerical stability.
• Use consistent conventions (frame definitions, rotation direction) across the whole pipeline.

1.4 Summary

Kinematic differential equations provide the mathematical bridge between angular velocity and attitude change. Euler angles are intuitive but suffer from singularities; the DCM is globally valid but requires orthonormality maintenance; quaternions offer a compact, nonsingular, and numerically stable representation that has become the standard for modern GNC applications.