Intuition The ONE core ideaTwo observers looking at the same arrow in space write down different lists of three numbers for it, because they aligned their measuring axes differently. A Direction Cosine Matrix (DCM) is the little table of numbers that converts one observer's list into the other's — and building it from scratch needs only three tools: unit vectors, the dot product, and the 2D rotation of a pair of axes.
This page assumes nothing . Before you can read the parent note the DCM topic , you must own every symbol it silently uses. We build them in order, each on top of the last.
A vector is an arrow: it has a length and a direction. Physically it might be "the direction the star tracker is looking" or "where gravity pulls". We draw it as an arrow with a tail and a head.
Definition Reference frameA reference frame is a set of three mutually perpendicular measuring directions (axes) glued to some observer. Point along each axis, ask "how far does the arrow reach along this direction?", and you get three numbers — the arrow's components in that frame.
The whole topic exists because there are two frames:
N inertial (navigation) frame. Think: fixed to the stars / the sky. Guidance and orbits are described here.B body frame. Glued to the spacecraft itself. Sensors and thrusters live here.
Intuition Why two frames at all?A gyroscope measures rotation in the body . But you want to know where you point relative to the stars . The same physical arrow has one list of numbers in B N
A unit vector is an arrow whose length is exactly 1 n ^ n point ; its length carries no information.
The parent note writes the frame axes as n ^ 1 , n ^ 2 , n ^ 3 b ^ 1 , b ^ 2 , b ^ 3 1 , 2 , 3
Three arrows of length 1
Why we need it: the DCM's entries are built entirely from how these hatted arrows line up with each other.
This is the single most important tool, so we build it carefully.
For two arrows a b dot product a ⋅ b a ⋅ b = ∣ a ∣ ∣ b ∣ cos θ ∣ a ∣ a θ between them.
Intuition Why does a dot product answer "how aligned?"cos θ 1 same way, θ = 0 0 θ = 9 0 ∘ − 1 θ = 18 0 ∘ measures agreement .
Now the key special case, which is the whole reason the topic is called "direction cosine ":
Intuition Tie-in to the DCMThe parent says entry C ij = b ^ i ⋅ n ^ j = cos θ ij i j why every entry is a "direction cosine" .
Common mistake "Dot product needs coordinates"Wrong feeling: you can only compute a ⋅ b a 1 b 1 + a 2 b 2 + a 3 b 3 Fix: that component formula is equal to ∣ a ∣∣ b ∣ cos θ cos θ
Every rotation is measured by an angle , and the machinery needs sin cos
Definition Sine and cosine on a right triangleDraw a right triangle. Pick one non-right corner and call its angle θ cos θ = hypotenuse adjacent side , sin θ = hypotenuse opposite side .
Intuition The picture that makes them one ideaStand a point on a circle of radius 1 θ cos θ sin θ θ cos 2 θ + sin 2 θ = 1 1
Why the topic needs this: the elementary rotation matrices R 1 , R 2 , R 3 entirely out of cos sin
The parent's derivation uses cos ( ϕ − θ ) sin ( ϕ − θ )
Intuition Why these appearRotating the axes by + θ point by − θ ϕ ϕ − θ cos ( ϕ − θ ) sin ( ϕ − θ ) entire origin of R 3 ( θ )
The DCM is a 3 × 3 matrix . Here is the bare minimum, from zero.
Definition Matrix and matrix–vector productA 3 × 3 matrix is a grid of 9 3 3 C ij row i j .
To multiply a matrix C v = [ v 1 , v 2 , v 3 ] T dot product of one row with v .
( C v ) i = C i 1 v 1 + C i 2 v 2 + C i 3 v 3 .
T v T C T flip across the diagonal : row i j j i T C − 1 = C T free (just re-read the grid the other way), which is why it's such a computational win.
Definition Matrix productA B i j A B row i A with column j B . This is not commutative: A B = B A order of the three Euler rotations matters.
Why the topic needs this: composing three rotations (R 1 R 2 R 3 is matrix multiplication, and transposing gives the inverse rotation.
The parent extracts angles back out of the DCM with these.
arcsin ( x ) "which angle has this sine?" It undoes sin − 9 0 ∘ + 9 0 ∘ 15 0 ∘ 3 0 ∘
atan2 ( y , x ) "which angle has vertical part y x Crucially it uses the signs of both y x atan2 arctan atan2
Common mistake Using plainarctan ( y / x ) Wrong feeling: "arctan Why it fails: the ratio y / x arctan atan2 atan2
The parent's small-angle example writes I − [ α × ]
Definition Skew-symmetric matrixI identity matrix (1 0 [ α × ] 3 × 3 α = ( α 1 , α 2 , α 3 ) anti-symmetric : flipping it across the diagonal negates every entry ([ α × ] T = − [ α × ] [ α × ] = 0 α 3 − α 2 − α 3 0 α 1 α 2 − α 1 0
You don't need to master this yet — its own note does — just recognize the shape when the parent's linearization appears.
Angle difference identity
Matrix product order matters
Test yourself — you're ready for the parent note when you can answer each without hesitation.
What does a hat, as in n ^ Its length is exactly 1
Why is the dot product of two unit vectors just cos θ Because a ^ ⋅ b ^ = ∣ a ^ ∣∣ b ^ ∣ cos θ 1
On a right triangle, what is cos θ Adjacent side divided by hypotenuse.
In which quadrants is cos θ Quadrants II and III (angles between 9 0 ∘ 27 0 ∘
Why does axis-rotation produce cos ( ϕ − θ ) Rotating axes by + θ ϕ ϕ − θ
How do you compute entry i C v Dot product of row i C v
What does the transpose C T C ij Moves it to row j i
Why must atan2 ( y , x ) arctan ( y / x ) It uses the signs of both y x
What is the identity matrix I Ones on the diagonal, zeros elsewhere; multiplying by it changes nothing.
Why does the order of the three Euler rotations matter? Matrix multiplication is not commutative: A B = B A
