DCM kinematics — Ċ = −[ω×]C

WHAT is $C$?

Because the entries are cosines of angles between axes, $C_{ij} = \hat{b}_i \cdot \hat{n}_j$.

WHY does $C$ change, and HOW fast?

Step 1 — The key constraint: orthonormality is preserved forever

Basis vectors stay unit-length and mutually perpendicular no matter how the body spins. So at all times: $C(t)\,C(t)^{T} = I$

Why this step? This single fact is everything. It forces the derivative to have a special (skew) structure.

Step 2 — Differentiate the constraint

Differentiate $C C^{T} = I$ using the product rule: $\dot{C}\,C^{T} + C\,\dot{C}^{T} = 0$

Notice $C\dot{C}^{T} = (\dot{C}C^{T})^{T}$. Let $\Omega \equiv \dot{C}\,C^{T}$. Then: $\Omega + \Omega^{T} = 0 \quad\Rightarrow\quad \Omega = -\Omega^{T}$

Why this step? We just proved $\Omega = \dot{C}C^{T}$ is skew-symmetric. A $3\times3$ skew matrix has only 3 independent numbers — exactly the components of an angular velocity vector.

Step 3 — Identify $\Omega$ physically

Any skew-symmetric $3\times3$ matrix can be written as the cross-product operator of some vector $\vec{w}$: $[\vec{w}\times] = \begin{bmatrix} 0 & -w_3 & w_2 \\ w_3 & 0 & -w_1 \\ -w_2 & w_1 & 0 \end{bmatrix},\qquad [\vec{w}\times]\vec{a} = \vec{w}\times\vec{a}$

Which vector is $\vec{w}$? Take a body-fixed vector $\vec{p}$ (constant in the body frame). In inertial coordinates its rate of change is $\dot{\vec p}_{\mathcal N} = \vec\omega_{\mathcal N} \times \vec p_{\mathcal N}$ (rotating vector). Working the transport through $C$ (see box below) shows $\vec w$ is the angular velocity expressed in the body frame, $\vec\omega_{\mathcal B}$, with a minus sign.

Why the minus sign? $C$ maps $\mathcal N \to \mathcal B$. As the body spins forward by $\vec\omega$, the inertial vectors appear to spin backward in the body frame — hence $-[\omega\times]$. (If instead you track $C_{\mathcal N/\mathcal B} = C^T$, you get $\dot{C}^{T} = [\omega\times]C^{T}$, no minus. The sign is a bookkeeping choice of direction.)

Deriving the sign cleanly (Feynman-style transport)

Let $\vec p$ be fixed in body ($\vec p_{\mathcal B}=$ const). Then $\vec p_{\mathcal B} = C\vec p_{\mathcal N}$, so $0 = \dot{\vec p}_{\mathcal B} = \dot C\,\vec p_{\mathcal N} + C\,\dot{\vec p}_{\mathcal N}.$ Using $\dot{\vec p}_{\mathcal N} = \vec\omega_{\mathcal N}\times\vec p_{\mathcal N}$ and $\vec\omega_{\mathcal B}=C\vec\omega_{\mathcal N}$: $\dot C\,\vec p_{\mathcal N} = -C(\vec\omega_{\mathcal N}\times \vec p_{\mathcal N}) = -[\vec\omega_{\mathcal B}\times]\,C\,\vec p_{\mathcal N}.$ True for all $\vec p_{\mathcal N}$ ⟹ $\dot C = -[\vec\omega_{\mathcal B}\times]C$. Why valid? We used the identity $C(\vec a\times\vec b) = (C\vec a)\times(C\vec b)$ for proper rotations.

![[3.5.04-DCM-kinematics-—-Ċ-=-−[ω×]C.png]]

WHY it matters (the 20% that gives 80%)

• It's the attitude propagation ODE: integrate $\dot C = -[\omega\times]C$ to know where the spacecraft points.
• It guarantees orthonormality by construction — the derivative lives in the tangent space of the rotation group $SO(3)$, so exact solutions never "lose" orthogonality (numerically you must renormalize).
• It's the parent of quaternion/Euler-angle kinematics: all are re-parametrizations of this same equation.

Worked Examples

Common Mistakes

Flashcards

Connections

• Skew-symmetric matrices & cross-product operator
• Rotation group SO(3) and Lie algebra so(3)
• Quaternion kinematics — q̇ = ½ Ω(ω) q
• Euler angle kinematics & gimbal lock
• Matrix exponential and rotation about a fixed axis
• Attitude propagation & determination (TRIAD, QUEST)
• Poisson's equation for rotating frames

![[3.5.04-DCM-kinematics-—-Ċ-=-−[ω×]C.png]]

Concept Map

Hinglish (regional understanding)

![[audio/3.5.04-DCM-kinematics-—-Ċ-=-−[ω×]C.mp3]]