Topic 5: Relative Motion
A vector can be defined in one reference frame, differentiated in another and expressed in yet another basis (Slides 17 and 20). Position vectors simply add (Notes Eq. 3.16), but time derivatives depend on the observer. The Coriolis theorem (Notes Eq. 3.8) converts a derivative taken in a moving frame into one taken in a fixed frame, and everything else in the topic (relative velocity, relative acceleration, inertia forces) follows from applying it.
1. Introduction: Observers and Reference Frames
1.1 Different observers see different motions
When we drive a car or travel on an airplane, our reference frame moves with respect to the frame of an observer on the ground, and the velocities and accelerations we measure differ from those measured on the ground (Notes Ch. 3, p. 36). For a particle $P$ observed from two frames $S_0$ and $S_1$ (Slide 3):
1.2 Why we need relative motion
- An observer is the same thing as a reference frame (Slide 3).
- Quite often one of the two descriptions is easy to compute and the other is the one we need (Notes p. 36).
- Newton's laws apply only in inertial frames; any accelerating or rotating frame is non-inertial. Observations made in a moving frame must be translated to an inertial frame, or inertia (fictitious) forces must be added (Section 11).
1.3 Absolute and relative quantities
Quantities measured in the frame called "fixed" are absolute; those measured in a "moving" frame are relative (Notes p. 36). The naming is a convention: calling a frame "fixed" or "moving", and a position "absolute" or "relative", is completely arbitrary and subjective (Slide 23).
1.4 First step: the motion of the frames themselves
To understand relative motion we must first describe how one frame moves with respect to another (Slide 3). A frame $S_1 = \{A; \mathcal{B}_1\}$ has two independent ingredients:
- Position and translational motion: the motion of its origin $A$ (Section 2).
- Attitude and rotation: the orientation of its vector basis $\mathcal{B}_1$ (Sections 3โ4) and how it changes in time (Sections 5โ8). Attitude means orientation.
2. Position and Translational Motion of a Reference Frame
Consider $S_0 = \{O; \mathcal{B}_0\}$ and $S_1 = \{A; \mathcal{B}_1\}$. The translational motion of $S_1$ with respect to $S_0$ is the motion of the origin $A$ as seen from $S_0$ (Slide 4):
Alternatively, the motion of $O$ can be described from $S_1$ with $\mathbf{r}_1^O, \mathbf{v}_1^O, \mathbf{a}_1^O$. Since $\mathbf{r}_1^O = \mathbf{AO} = -\mathbf{OA}$, the positions are opposite, $\mathbf{r}_1^O = -\mathbf{r}_0^A$, but in general $\mathbf{v}_1^O \neq -\mathbf{v}_0^A$ and $\mathbf{a}_1^O \neq -\mathbf{a}_0^A$ (Slide 4).
$\mathbf{v}_1^O$ is the derivative of $\mathbf{r}_1^O$ taken in $S_1$, while $\mathbf{v}_0^A$ is the derivative of $\mathbf{r}_0^A$ taken in $S_0$. If $\mathcal{B}_1$ rotates, the two derivatives of the same vector differ. The Coriolis theorem of Section 6 gives $\mathbf{v}_1^O = -\left.\frac{d\mathbf{r}_0^A}{dt}\right|_1 = -\left(\mathbf{v}_0^A - \boldsymbol{\omega}_{10}\times\mathbf{r}_0^A\right)$, which equals $-\mathbf{v}_0^A$ only when $\boldsymbol{\omega}_{10}\times\mathbf{r}_0^A = \mathbf{0}$ (for instance, when $S_1$ does not rotate).
3. Attitude of a Reference Frame: the Change of Basis
3.1 Definition of attitude
The attitude of $S_1$ with respect to $S_0$ is its instantaneous orientation in space. The position of the origin does not matter: only the directions of the vectors of the bases (Slide 5).
3.2 Change-of-basis relations
The simplest description writes the vectors of $\mathcal{B}_1$ in terms of those of $\mathcal{B}_0$ (Slide 5, Notes Eq. 1.8):
3.3 Components of a vector in either basis
Any vector $\mathbf{b}$ can be written in either basis; the vector is the same physical entity (Slide 5, Notes Eq. 1.6):
Derivation (Notes Eqs. 1.9โ1.12). Substitute the change-of-basis relations into the second expression:
Collect the coefficients of $\mathbf{i}_0$, $\mathbf{j}_0$, $\mathbf{k}_0$ and compare with $b_{x0}\mathbf{i}_0 + b_{y0}\mathbf{j}_0 + b_{z0}\mathbf{k}_0$ (components in a basis are unique):
In matrix form (Slide 5, Notes Eqs. 1.12โ1.13):
When the components of $\mathbf{b}$ change from one basis to another, the vector is not moved and not rotated: it is only described from a different basis, so the numbers that represent it change (Notes Sec. 1.5).
3.4 Properties of $[{}_0R_1]$
For two right-handed orthonormal bases (Slide 6, Notes Eqs. 1.14โ1.15):
- Orthogonal: the inverse equals the transpose, $[{}_1R_0] = [{}_0R_1]^{-1} = [{}_0R_1]^T$, so $[{}_0R_1]^T[{}_0R_1] = [U]$. Writing $\mathbf{i}_0 = \ell_{x0,x1}\mathbf{i}_1 + \ell_{x0,y1}\mathbf{j}_1 + \ell_{x0,z1}\mathbf{k}_1$, etc., this means $\ell_{x0,x1} = \ell_{x1,x0}$, $\ell_{x0,y1} = \ell_{y1,x0}$, and so on; consistent with $\mathbf{i}_0\cdot\mathbf{j}_1 = \mathbf{j}_1\cdot\mathbf{i}_0$.
- Unit determinant: $|[{}_0R_1]| = 1$, because both bases are right-handed.
- Columns = components of $\mathbf{i}_1, \mathbf{j}_1, \mathbf{k}_1$ in $\mathcal{B}_0$. This is the practical way to write the matrix.
- Rows = components of $\mathbf{i}_0, \mathbf{j}_0, \mathbf{k}_0$ in $\mathcal{B}_1$.
- Eigenvalues: $\lambda = 1,\ e^{i\beta},\ e^{-i\beta}$.
3.5 Axis and angle of rotation
A result due to Euler states that any two right-handed bases are related by a single rotation about some axis $\mathbf{e}$ by some angle (Notes Sec. 1.5.1). In terms of the matrix (Slide 6): the axis of rotation is the eigenspace of $\lambda = 1$ (a vector on it has identical components in $\mathcal{B}_1$ and $\mathcal{B}_0$, since $[{}_0R_1]\mathbf{e} = \mathbf{e}$), and the rotation angle is $\beta$. The same matrix can be read as the static relation between components in two bases, or as the rotation that, seen from $\mathcal{B}_0$, carries $\mathcal{B}_0$ onto $\mathcal{B}_1$; for this reason it is called a rotation matrix (Notes Sec. 1.5.1).
3.6 Composition rules
With three bases $\mathcal{B}_0, \mathcal{B}_1, \mathcal{B}_2$, substituting the expressions of $\mathbf{i}_1, \mathbf{j}_1, \mathbf{k}_1$ into $\mathbf{i}_2 = \ell_{x2,x1}\mathbf{i}_1 + \ell_{x2,y1}\mathbf{j}_1 + \ell_{x2,z1}\mathbf{k}_1$ gives $\mathbf{i}_2$ directly in $\mathcal{B}_0$ (Slide 7):
Equivalently, from $[\mathbf{b}]_0 = [{}_0R_1][\mathbf{b}]_1$ and $[\mathbf{b}]_1 = [{}_1R_2][\mathbf{b}]_2$ (Notes Eq. 1.16):
4. Simple Rotations and Euler Angles
4.1 Simple rotations
A simple rotation has its axis along one of the basis vectors. That vector is unchanged (e.g. $\mathbf{i}_1 \equiv \mathbf{i}_0$) and the rest of space turns by an angle that is positive by the right-hand rule (Slide 8, Notes Sec. 1.5.2). Writing the new vectors in the old ones and placing them as columns (Notes Eqs. 1.18โ1.20):
| Axis | Relations | Matrix |
|---|---|---|
| $\mathbf{i}_0$, angle $\theta_x$ | $\mathbf{i}_1 = \mathbf{i}_0$ $\mathbf{j}_1 = \cos\theta_x\,\mathbf{j}_0 + \sin\theta_x\,\mathbf{k}_0$ $\mathbf{k}_1 = -\sin\theta_x\,\mathbf{j}_0 + \cos\theta_x\,\mathbf{k}_0$ | $\begin{bmatrix} 1 & 0 & 0 \\ 0 & \cos\theta_x & -\sin\theta_x \\ 0 & \sin\theta_x & \cos\theta_x \end{bmatrix}$ |
| $\mathbf{j}_0$, angle $\theta_y$ | $\mathbf{i}_1 = \cos\theta_y\,\mathbf{i}_0 - \sin\theta_y\,\mathbf{k}_0$ $\mathbf{j}_1 = \mathbf{j}_0$ $\mathbf{k}_1 = \sin\theta_y\,\mathbf{i}_0 + \cos\theta_y\,\mathbf{k}_0$ | $\begin{bmatrix} \cos\theta_y & 0 & \sin\theta_y \\ 0 & 1 & 0 \\ -\sin\theta_y & 0 & \cos\theta_y \end{bmatrix}$ |
| $\mathbf{k}_0$, angle $\theta_z$ | $\mathbf{i}_1 = \cos\theta_z\,\mathbf{i}_0 + \sin\theta_z\,\mathbf{j}_0$ $\mathbf{j}_1 = -\sin\theta_z\,\mathbf{i}_0 + \cos\theta_z\,\mathbf{j}_0$ $\mathbf{k}_1 = \mathbf{k}_0$ | $\begin{bmatrix} \cos\theta_z & -\sin\theta_z & 0 \\ \sin\theta_z & \cos\theta_z & 0 \\ 0 & 0 & 1 \end{bmatrix}$ |
Check (for $z$): $\det = \cos^2\theta_z + \sin^2\theta_z = 1$, and the first two columns are orthogonal: $(\cos\theta_z)(-\sin\theta_z) + (\sin\theta_z)(\cos\theta_z) + 0 = 0$. The inverse matrices are the same with $\theta \to -\theta$: only the position of the minus sign before the sines changes. To avoid sign errors, always draw the rotation looking down from the stationary axis (Notes Sec. 1.5.2). The $y$ case looks different only because, seen from $\mathbf{j}_0$, the plane is $Ozx$ in that cyclic order.
4.2 Euler angles
Any attitude is the composition of three simple rotations about different axes. To relate $\mathcal{B}_0$ to $\mathcal{B}_3$ we introduce the auxiliary bases $\mathcal{B}_1$ and $\mathcal{B}_2$ (Slide 9):
- $\mathcal{B}_0 \to \mathcal{B}_1$: rotation $\psi$ about $\mathbf{k}_0 \equiv \mathbf{k}_1$;
- $\mathcal{B}_1 \to \mathcal{B}_2$: rotation $\theta$ about $\mathbf{i}_1 \equiv \mathbf{i}_2$;
- $\mathcal{B}_2 \to \mathcal{B}_3$: rotation $\varphi$ about $\mathbf{k}_2 \equiv \mathbf{k}_3$.
$\psi$ (precession), $\theta$ (nutation) and $\varphi$ (spin) are the standard Euler angles, also called 3-1-3 angles after the order of the axes: third, first, third again. Many other choices exist; a common one is 3-2-1 (yaw, pitch, roll). The only rule is that the same axis cannot be repeated in two consecutive rotations (Slide 9). The Notes also use the 1-2-3 sequence (Notes Eq. 1.23).
4.3 Example: 3-2-1 attitude of the Space Shuttle
The attitude of the red frame $S_r$ (attached to the shuttle) relative to the blue frame $S_b$ is described with 3-2-1 angles, the convention of flight mechanics; both frames start coincident (Slides 10โ13).
- First rotation, $\psi$ about $\mathbf{k}_b \equiv \mathbf{k}_1$ (Slide 11): $\ \mathbf{i}_1 = \cos\psi\,\mathbf{i}_b + \sin\psi\,\mathbf{j}_b$, $\ \mathbf{j}_1 = -\sin\psi\,\mathbf{i}_b + \cos\psi\,\mathbf{j}_b$, $\ \mathbf{k}_1 = \mathbf{k}_b$.
- Second rotation, $\theta$ about $\mathbf{j}_1 \equiv \mathbf{j}_2$ (Slide 12): $\ \mathbf{i}_2 = \cos\theta\,\mathbf{i}_1 - \sin\theta\,\mathbf{k}_1$, $\ \mathbf{j}_2 = \mathbf{j}_1$, $\ \mathbf{k}_2 = \sin\theta\,\mathbf{i}_1 + \cos\theta\,\mathbf{k}_1$.
- Third rotation, $\varphi$ about $\mathbf{i}_2 \equiv \mathbf{i}_r$ (Slide 13): $\ \mathbf{i}_r = \mathbf{i}_2$, $\ \mathbf{j}_r = \cos\varphi\,\mathbf{j}_2 + \sin\varphi\,\mathbf{k}_2$, $\ \mathbf{k}_r = -\sin\varphi\,\mathbf{j}_2 + \cos\varphi\,\mathbf{k}_2$.
Composing the three matrices, $[{}_bR_r] = [{}_bR_1(z,\psi)]\cdot[{}_1R_2(y,\theta)]\cdot[{}_2R_r(x,\varphi)]$:
5. Angular Motion of a Reference Frame
5.1 Angular velocity vector
So far the change of basis was stationary, but bases can rotate with respect to each other in time. The rotation of $\mathcal{B}_1$ relative to $\mathcal{B}_0$ is represented at each instant by the angular velocity vector $\boldsymbol{\omega}_{10}$ (Slide 14):
- its direction is the instantaneous axis of rotation of $\mathcal{B}_1$ (right-hand rule);
- its magnitude is the angular rate.
$\boldsymbol{\omega}_{10}$ and the motion of the origin of $S_1$ are unrelated: two completely different things (Slide 14).
A vector $\mathbf{A}$ of constant magnitude rotating about an axis $\mathbf{e}_\omega$ at rate $\dot\theta$ keeps its component along $\mathbf{e}_\omega$, while its perpendicular part describes a circle. In a time $dt$ it changes by $|d\mathbf{A}| = A_\perp\,d\theta = \dot\theta\,|\mathbf{e}_\omega\times\mathbf{A}|\,dt$, in the direction of $\mathbf{e}_\omega\times\mathbf{A}$ (Notes Sec. 1.7.1). Hence $\dfrac{d\mathbf{A}}{dt} = \dot\theta\,\mathbf{e}_\omega\times\mathbf{A} = \boldsymbol{\omega}\times\mathbf{A}$ with $\boldsymbol{\omega} = \dot\theta\,\mathbf{e}_\omega$ (Notes Eq. 1.62).
5.2 Basis vectors of $\mathcal{B}_1$ seen from $S_0$
The vectors of $\mathcal{B}_1$ are unit vectors, so they change only in direction (Slide 14, Notes Eqs. 3.3โ3.5):
5.3 A vector that is constant in $S_1$
Let $\mathbf{c} = c_{x1}\mathbf{i}_1 + c_{y1}\mathbf{j}_1 + c_{z1}\mathbf{k}_1$ with constant components. In $S_0$, the product rule gives $\dot c_{x1}\mathbf{i}_1 + c_{x1}\,d\mathbf{i}_1/dt|_0$ for each term, with $\dot c_{x1} = 0$ (Slide 14):
5.4 Angular acceleration vector
The rotation can be accelerated. The angular acceleration vector is (Slide 15)
- Its direction may differ from that of $\boldsymbol{\omega}_{10}$.
- $\boldsymbol{\omega}_{10}$ does not describe the current attitude of $S_1$, only how it is changing in time.
- The component of $\boldsymbol{\alpha}_{10}$ parallel to $\boldsymbol{\omega}_{10}$ describes how the magnitude of $\boldsymbol{\omega}_{10}$ changes; the perpendicular component describes how its direction changes. With $\boldsymbol{\omega}_{10} = \omega_{10}\mathbf{e}_\omega$, the product rule gives $\boldsymbol{\alpha}_{10} = \dot\omega_{10}\,\mathbf{e}_\omega + \omega_{10}\,d\mathbf{e}_\omega/dt|_0$, and $d\mathbf{e}_\omega/dt \perp \mathbf{e}_\omega$ because $\mathbf{e}_\omega$ has unit length (Notes Eq. 1.53).
5.5 Example: simple rotation about $\mathbf{k}_0$ (Slide 16)
At $t = 0$ the frames coincide; then $S_1$ rotates about $\mathbf{k}_0$ by $\psi = \psi(t)$: $\ \mathbf{i}_1 = \cos\psi\,\mathbf{i}_0 + \sin\psi\,\mathbf{j}_0$, $\ \mathbf{j}_1 = -\sin\psi\,\mathbf{i}_0 + \cos\psi\,\mathbf{j}_0$, $\ \mathbf{k}_1 = \mathbf{k}_0$.
Direct differentiation in $S_0$ ($\mathbf{i}_0, \mathbf{j}_0, \mathbf{k}_0$ are constant there). Chain rule: $\frac{d}{dt}\cos\psi(t) = -\sin\psi\,\dot\psi$, $\ \frac{d}{dt}\sin\psi(t) = \cos\psi\,\dot\psi$:
Identification. With $\boldsymbol{\omega}_{10} = \dot\psi\,\mathbf{k}_1$: $\ \dot\psi\,\mathbf{k}_1\times\mathbf{i}_1 = \dot\psi\,\mathbf{j}_1$, $\ \dot\psi\,\mathbf{k}_1\times\mathbf{j}_1 = -\dot\psi\,\mathbf{i}_1$, $\ \dot\psi\,\mathbf{k}_1\times\mathbf{k}_1 = \mathbf{0}$, which reproduce the three derivatives. Therefore
6. The Coriolis Theorem
Positions, velocities and forces are often easiest to write in the basis of a moving frame (the wire, the ring, the aircraft). Newton's law, however, needs derivatives taken in the inertial frame. The Coriolis theorem is the bridge: it gives the derivative in $S_0$ of a vector whose components are known in $\mathcal{B}_1$ (Slide 17, Notes Sec. 3.1).
6.1 Derivation
Consider $\mathbf{b} = b_{x1}(t)\,\mathbf{i}_1 + b_{y1}(t)\,\mathbf{j}_1 + b_{z1}(t)\,\mathbf{k}_1$, not constant in $S_1$ (Slide 17; the Notes call it $\mathbf{A}$, Notes Eq. 3.1).
Derivative in $S_1$. In $S_1$ the basis vectors are constant, so only the components change (Notes Eq. 3.7):
Derivative in $S_0$. The basis vectors of $\mathcal{B}_1$ are not constant in $S_0$, so the product rule applies to each term, $\frac{d}{dt}(b_{x1}\mathbf{i}_1) = \dot b_{x1}\mathbf{i}_1 + b_{x1}\,\frac{d\mathbf{i}_1}{dt}\big|_0$ (Notes Eq. 3.2):
6.2 Meaning and remarks
- The derivative of $\mathbf{b}$ in $S_0$ equals its derivative in $S_1$ plus the rotation of $\mathbf{b}$ due to the angular motion of $S_1$ relative to $S_0$ (Slide 17).
- For vectors, the frame of each derivative must always be written. Derivatives of scalars are the same in every frame (Notes p. 37).
- The result holds unchanged when $S_1$ also translates: a translation does not change the orientation of $\{\mathbf{i}_1, \mathbf{j}_1, \mathbf{k}_1\}$ (Notes p. 37).
- Remember: a vector can be defined in one frame, differentiated in another and expressed in yet another (Slide 17).
6.3 Consequence: $\boldsymbol{\alpha}_{10}$ is the same in $S_0$ and $S_1$
Example (Slide 18): derivative of $\mathbf{b} = t\,\mathbf{i}_1 + t^2\,\mathbf{j}_1 + t^3\,\mathbf{k}_1$
$S_1$ rotates about $\mathbf{k}_0 \equiv \mathbf{k}_1$ by $\psi(t)$, so $\boldsymbol{\omega}_{10} = \dot\psi\,\mathbf{k}_1$- In $S_1$: $\frac{d}{dt}t = 1$, $\frac{d}{dt}t^2 = 2t$, $\frac{d}{dt}t^3 = 3t^2$, so $\left.\frac{d\mathbf{b}}{dt}\right|_1 = \mathbf{i}_1 + 2t\,\mathbf{j}_1 + 3t^2\,\mathbf{k}_1$.
- Rotation term, with $\mathbf{k}_1\times\mathbf{i}_1 = \mathbf{j}_1$, $\mathbf{k}_1\times\mathbf{j}_1 = -\mathbf{i}_1$, $\mathbf{k}_1\times\mathbf{k}_1 = \mathbf{0}$: $\ \boldsymbol{\omega}_{10}\times\mathbf{b} = \dot\psi\,t\,\mathbf{j}_1 - \dot\psi\,t^2\,\mathbf{i}_1$.
- In $S_0$: $\left.\frac{d\mathbf{b}}{dt}\right|_0 = \left(1 - \dot\psi\,t^2\right)\mathbf{i}_1 + \left(2t + \dot\psi\,t\right)\mathbf{j}_1 + 3t^2\,\mathbf{k}_1$. This derivative taken in $S_0$ is expressed in $\mathcal{B}_1$.
- Cross-check: writing $\mathbf{b}$ in $\mathcal{B}_0$ with the relations of Section 5.5 and differentiating component by component gives the same vector (checked symbolically).
7. Composition of Angular Motions
7.1 Angular velocities add
If $S_2$ rotates relative to $S_1$ with $\boldsymbol{\omega}_{21}$ and $S_1$ relative to $S_0$ with $\boldsymbol{\omega}_{10}$ (Slide 19, Notes Eq. 3.10):
7.2 Angular accelerations do not simply add
Differentiate in $S_0$ (Slide 19, Notes Eq. 3.11). The last term is $\boldsymbol{\alpha}_{10}$; the first is the $S_0$-derivative of a vector naturally defined in $S_1$, so by Coriolis $\left.\frac{d\boldsymbol{\omega}_{21}}{dt}\right|_0 = \left.\frac{d\boldsymbol{\omega}_{21}}{dt}\right|_1 + \boldsymbol{\omega}_{10}\times\boldsymbol{\omega}_{21}$:
7.3 Inverse motion
In general $\boldsymbol{\omega}_{lm}$ is the angular velocity of frame $l$ as seen from frame $m$. When composing in components, express every vector in the same basis first (Notes p. 38).
8. Angular Velocity from Simple Rotations
8.1 3-1-3 Euler angles as functions of time
For $\psi(t), \theta(t), \varphi(t)$, $S_0 \to S_1 \to S_2 \to S_3$ (Slide 20). Each simple rotation has the angular velocity "rate of its angle times its axis" (Section 5.5):
8.2 Expressed in $\mathcal{B}_0$ (Slide 21)
From $\mathbf{j}_1 = -\sin\psi\,\mathbf{i}_0 + \cos\psi\,\mathbf{j}_0$ and $\mathbf{k}_2 = -\sin\theta\,\mathbf{j}_1 + \cos\theta\,\mathbf{k}_1$:
8.3 Expressed in $\mathcal{B}_3$ (Slide 21)
Inverting the last rotation, $\mathbf{i}_2 = \cos\varphi\,\mathbf{i}_3 - \sin\varphi\,\mathbf{j}_3$, $\mathbf{j}_2 = \sin\varphi\,\mathbf{i}_3 + \cos\varphi\,\mathbf{j}_3$; inverting the second, $\mathbf{k}_1 = \sin\theta\,\mathbf{j}_2 + \cos\theta\,\mathbf{k}_2$. Then
8.4 Exercise of Slide 21: 3-2-1 (yaw-pitch-roll) angles
With the rotations of Section 4.3: $\ \boldsymbol{\omega}_{rb} = \dot\psi\,\mathbf{k}_b + \dot\theta\,\mathbf{j}_1 + \dot\varphi\,\mathbf{i}_2$.
In $\mathcal{B}_b$, with $\mathbf{j}_1 = -\sin\psi\,\mathbf{i}_b + \cos\psi\,\mathbf{j}_b$ and $\mathbf{i}_2 = \cos\theta\cos\psi\,\mathbf{i}_b + \cos\theta\sin\psi\,\mathbf{j}_b - \sin\theta\,\mathbf{k}_b$:
In $\mathcal{B}_r$: $\mathbf{j}_1 = \mathbf{j}_2 = \cos\varphi\,\mathbf{j}_r - \sin\varphi\,\mathbf{k}_r$ and $\mathbf{k}_b = \mathbf{k}_1 = -\sin\theta\,\mathbf{i}_r + \cos\theta\sin\varphi\,\mathbf{j}_r + \cos\theta\cos\varphi\,\mathbf{k}_r$ (inverting the second and third rotations), so
9. Configuration and Motion of a Reference Frame
The configuration and motion of $S_1 = \{A; \mathcal{B}_1\}$ relative to $S_0$ are completely specified by (Slide 22):
- the position of its origin as a function of time, $\mathbf{r}_0^A(t)$;
- the attitude of its basis as a function of time, $\mathbf{i}_1(t), \mathbf{j}_1(t), \mathbf{k}_1(t)$, for example through Euler angles $\psi(t), \theta(t), \varphi(t)$.
From these we compute $\mathbf{v}_0^A(t)$, $\mathbf{a}_0^A(t)$, $\boldsymbol{\omega}_{10}(t)$ and $\boldsymbol{\alpha}_{10}(t)$. The relative-motion expressions are especially useful when the motion of $S_1$ is known a priori: then none of these functions of time are unknowns (Slide 22, Notes Sec. 3.2).
10. Relative Kinematics of the Point Particle
10.1 Motivation
- We often have several observers and need to relate the position, velocity and acceleration of $P$ that each sees. Example: sensors on an experimental aircraft measure a hurricane; the motion of the aircraft must be taken into account to "translate" the measurements to a ground-based frame (Slide 23).
- In other cases it is simpler to obtain these vectors in a moving frame than in the fixed one (Slide 23).
10.2 Position
10.3 Velocity
Step 1. The absolute velocity is the derivative of $\mathbf{r}_0^P$ in $S_0$; the derivative of a sum is the sum of the derivatives (Notes Eq. 3.17):
Step 2. $\mathbf{r}_1^P$ is naturally described in $S_1$, so apply the Coriolis theorem (Notes Eqs. 3.18โ3.19), with $\left.d\mathbf{r}_1^P/dt\right|_1 \equiv \mathbf{v}_1^P$ the relative velocity:
10.4 Acceleration
Step 1. Differentiate the velocity in $S_0$ (Notes Eq. 3.21):
Step 2. Coriolis on the relative velocity, with $\left.d\mathbf{v}_1^P/dt\right|_1 \equiv \mathbf{a}_1^P$ the relative acceleration (Notes Eqs. 3.22 and 3.24):
Step 3. Product rule for the cross product (Notes Eq. 1.67), then Coriolis on $\mathbf{r}_1^P$ as in Section 10.3 (Notes Eq. 3.23):
Step 4. Add the three terms. $\boldsymbol{\omega}_{10}\times\mathbf{v}_1^P$ appears twice (Steps 2 and 3), which gives the factor 2 (Notes Eq. 3.25):
| Term | Name (Slide 25) |
|---|---|
| $\mathbf{a}_0^A$ | acceleration of the origin of $S_1$ |
| $\mathbf{a}_1^P$ | relative acceleration |
| $\boldsymbol{\alpha}_{10}\times\mathbf{r}_1^P$ | Euler acceleration |
| $\boldsymbol{\omega}_{10}\times(\boldsymbol{\omega}_{10}\times\mathbf{r}_1^P)$ | centripetal acceleration |
| $2\boldsymbol{\omega}_{10}\times\mathbf{v}_1^P$ | Coriolis acceleration |
10.5 Summary
10.6 Remarks on each term
- The translation of $A$ appears only through $\mathbf{v}_0^A$ and $\mathbf{a}_0^A$; all other terms come from the relative motion of $P$ and/or the angular motion of $S_1$ (Notes p. 39).
- The direction of $\boldsymbol{\omega}_{10}$ is the axis of rotation of $S_1$ relative to $S_0$.
- $\boldsymbol{\omega}_{10}\times\mathbf{r}_1^P$ is perpendicular to $\boldsymbol{\omega}_{10}$ and $\mathbf{r}_1^P$, with magnitude $\omega_{10}\,r_{1,\perp}^P$, where $r_{1,\perp}^P$ is the component of $\mathbf{r}_1^P$ perpendicular to $\boldsymbol{\omega}_{10}$.
- The same holds for $\boldsymbol{\alpha}_{10}\times\mathbf{r}_1^P$ with respect to $\boldsymbol{\alpha}_{10}$; since $\boldsymbol{\alpha}_{10}$ and $\boldsymbol{\omega}_{10}$ are in general not parallel, it can have a component along $\boldsymbol{\omega}_{10}$.
- The centripetal term has magnitude $r_{1,\perp}^P\,\omega_{10}^2$ and always points toward the line through the origin along $\boldsymbol{\omega}_{10}$.
- The factor 2 of the Coriolis term comes from two effects of equal importance: the rotation of the relative velocity (Step 2) and the change of $\boldsymbol{\omega}_{10}\times\mathbf{r}_1^P$ as $P$ moves (Step 3). It exists only in a rotating frame and only when $P$ has relative velocity.
10.7 Translating frame without rotation
11. Relative Dynamics: Inertia Forces
Newton's laws are valid only in an inertial frame. Assuming $S_0$ is inertial, $m^P\mathbf{a}_0^P = \mathbf{F}$ (Notes Eq. 4.59). If $S_1$ accelerates or rotates relative to $S_0$ it is non-inertial. The relative acceleration formula shows how to fix Newton's law so that it can be applied in $S_1$ (Slide 27, Notes Sec. 4.9).
11.1 Derivation
Step 1. Substitute Notes Eq. 3.27 into Newton's law (Notes Eq. 4.60):
Step 2. Keep the relative acceleration on the left (Notes Eq. 4.61):
Step 3. Define the fictitious or inertia force (Notes Eq. 4.62):
Eq. 4.63 contains the same information as Eq. 4.60: terms have only been moved across. The new reading is that an observer in $S_1$ may apply Newton's second law if the force $\mathbf{F}_i$ is added to the actual forces. Every term of $\mathbf{F}_i$ is computed from the relative position and velocity of $P$ once $\mathbf{a}_0^A$, $\boldsymbol{\omega}_{10}$, $\boldsymbol{\alpha}_{10}$ are known (Notes p. 53).
11.2 Names of the four terms
| Term | Name (Slide 28, Notes p. 53) | Symbol |
|---|---|---|
| $-m^P\mathbf{a}_0^A$ | force due to the acceleration of the origin | $\mathbf{F}_A$ |
| $-m^P\boldsymbol{\alpha}_{10}\times\mathbf{r}_1^P$ | Euler's force | $\mathbf{F}_E$ |
| $-m^P\boldsymbol{\omega}_{10}\times(\boldsymbol{\omega}_{10}\times\mathbf{r}_1^P)$ | centrifugal force due to the rotation of $S_1$ | $\mathbf{F}_C$ |
| $-2m^P\boldsymbol{\omega}_{10}\times\mathbf{v}_1^P$ | Coriolis force | $\mathbf{F}_{COR}$ |
11.3 General properties (Slide 28)
- All inertia forces are proportional to the mass of the particle.
- Forces 1 to 3 are locally indistinguishable from a gravitational field: an observer in $S_1$ cannot tell them apart from actual gravity.
- Force 4 (Coriolis) is different: it is zero when $P$ is at rest in $S_1$ and appears only when $P$ has relative velocity with a component perpendicular to $\boldsymbol{\omega}_{10}$.
- We can obtain the equations of motion in inertial and non-inertial frames; choose the simplest option for each problem.
11.4 Physical intuition (Notes p. 53)
Although fictitious, these forces may feel very real to an observer on a non-inertial frame; they are due solely to the motion of that frame relative to an inertial one.
- Braking vehicle. We tend to go forward when a vehicle brakes, and it is common to say the inertia force $-m\mathbf{a}_0^A$ pushes us. For a ground observer no such force exists: the vehicle decelerates through the friction of the wheels and the passenger's mass tends to continue forward.
- Turning vehicle. The road gives the vehicle a transverse friction force toward the centre of curvature. The passenger tends to continue in uniform rectilinear motion and feels pushed outward, until the seat and the floor transmit the necessary force.
- Accelerating merry-go-round. Besides the outward centrifugal force, a child feels a force pointing backward with respect to the direction of travel: the Euler term $-m\boldsymbol{\alpha}_{10}\times\mathbf{r}_1^P$.
- Ball on the merry-go-round. A ball passed from the outer to the inner part turns away from the expected straight line toward the direction of travel: the Coriolis force. It also explains the sense of rotation of hurricanes in each hemisphere, and why a cannonball shot toward the East lands farther than one shot equally toward the West.
12. Energy and Potentials in a Non-Inertial Frame
12.1 Energy and other quantities in $S_1$
The kinetic energy $T_1^P = \tfrac12 m^P\,\mathbf{v}_1^P\cdot\mathbf{v}_1^P$ and the mechanical energy $E_1^P$ of $P$ can always be defined in $S_1$, even if non-inertial. If the necessary conditions are met (the same as before, with the inertia forces counted among the forces), $E_1^P$ is a conserved quantity. The angular momentum $\mathbf{H}_{A1}^P$ or any other quantity used so far can also be defined in $S_1$ (Slide 29).
12.2 Force due to the acceleration of the origin
12.3 Centrifugal force
Decompose $\mathbf{r}_1^P = \mathbf{r}_\parallel + \mathbf{r}_\perp$ along and across the axis, with $\boldsymbol{\omega}_{10} = \omega_{10}\mathbf{e}_\omega$ and $\mathbf{r}_\parallel = r_\parallel\mathbf{e}_\omega$ (Slide 30). Using $\mathbf{a}\times(\mathbf{b}\times\mathbf{c}) = (\mathbf{a}\cdot\mathbf{c})\mathbf{b} - (\mathbf{a}\cdot\mathbf{b})\mathbf{c}$ and $\boldsymbol{\omega}_{10}\cdot\mathbf{r}_1^P = \omega_{10}r_\parallel$:
12.4 Euler force
$\mathbf{F}_E = -m^P\boldsymbol{\alpha}_{10}\times\mathbf{r}_1^P$ is always perpendicular to $\boldsymbol{\alpha}_{10}$ and $\mathbf{r}_1^P$, and it does not admit a potential: $\nabla\times\mathbf{F}_E \neq \mathbf{0}$ (Slide 31). For a given $\boldsymbol{\alpha}_{10}$, $\nabla\times(\boldsymbol{\alpha}_{10}\times\mathbf{r}) = 2\boldsymbol{\alpha}_{10}$, so $\nabla\times\mathbf{F}_E = -2m^P\boldsymbol{\alpha}_{10}$.
12.5 Coriolis force
$\mathbf{F}_{COR} = -2m^P\boldsymbol{\omega}_{10}\times\mathbf{v}_1^P$ is always perpendicular to $\boldsymbol{\omega}_{10}$ and $\mathbf{v}_1^P$, and proportional to $v_\perp$, the component of $\mathbf{v}_1^P$ perpendicular to $\boldsymbol{\omega}_{10}$ (Slide 31). Its power vanishes:
Example (Slide 31): equatorial launch
$\boldsymbol{\omega}_{10} = \Omega\,\mathbf{e}_N$ is horizontal at the equator- Local right-handed unit vectors East $\mathbf{e}_E$, North $\mathbf{e}_N$, Up $\mathbf{e}_U$: $\mathbf{e}_N\times\mathbf{e}_E = -\mathbf{e}_U$.
- Toward the East, $\mathbf{v}_1^P$ has a component $v\,\mathbf{e}_E$: $\ -2m\Omega\,\mathbf{e}_N\times(v\,\mathbf{e}_E) = 2m\Omega v\,\mathbf{e}_U$, upward. The trajectory is higher and longer.
- Toward the West ($v\,\mathbf{e}_E$ with $v \lt 0$) the force points downward: lower and shorter trajectory. This is the cannonball effect of Section 11.4.
12.6 Conservation of $E_1^P$ in a frame rotating at constant rate
With a fixed origin ($\mathbf{a}_0^A = \mathbf{0}$) and constant $\boldsymbol{\omega}_{10}$ ($\boldsymbol{\alpha}_{10} = \mathbf{0}$), the Euler force vanishes, the centrifugal force has a potential and the Coriolis force does no work. If the actual forces that work are conservative and the constraints are fixed in $S_1$:
13. Method for Relative-Motion Problems
This checklist condenses Slides 22, 26 and 28 into the order used in the solved problems:
- Choose the frames. $S_0$ inertial; $S_1$ attached to the moving body (wire, ring, platform, spacecraft) so that the constraint or the motion is simple in $S_1$.
- Describe the motion of $S_1$: $\mathbf{r}_0^A, \mathbf{v}_0^A, \mathbf{a}_0^A$, $[{}_0R_1]$, $\boldsymbol{\omega}_{10}, \boldsymbol{\alpha}_{10}$ (Section 9).
- Relative kinematics of $P$: $\mathbf{r}_1^P, \mathbf{v}_1^P, \mathbf{a}_1^P$ with the coordinates of the problem.
- Either compute $\mathbf{a}_0^P$ with Notes Eq. 3.27 and write $m\mathbf{a}_0^P = \mathbf{F}$ in $S_0$, or write $m\mathbf{a}_1^P = \mathbf{F} + \mathbf{F}_i$ in $S_1$ (Notes Eq. 4.63). Both give the same equations.
- Project on the directions where the reactions vanish to get the equations of motion; project on the others to get the reactions.
- First integrals: $E_1^P$ if $\mathbf{a}_0^A$ and $\boldsymbol{\omega}_{10}$ are constant (or zero) and the constraints are fixed in $S_1$.
14. Source Map
| Content | Notes.pdf (printed pages) | Slides |
|---|---|---|
| Observers, absolute and relative | Ch. 3 introduction (p. 36) | 3, 23 |
| Translational motion of a frame | Sec. 3.2 (p. 38) | 4 |
| Change of basis, properties, composition | Sec. 1.5, Eqs. 1.6โ1.17 (pp. 13โ14) | 5โ7 |
| Axis and angle of rotation | Sec. 1.5.1 (pp. 14โ15) | 6 |
| Simple rotations | Sec. 1.5.2, Eqs. 1.18โ1.20 (pp. 15โ16) | 8 |
| Euler angles, 3-2-1 example | Sec. 1.5.3, Eqs. 1.21โ1.23 (pp. 16โ17) | 9โ13 |
| Angular velocity and acceleration | Sec. 1.7, Eqs. 1.62, 1.73; Eqs. 3.3โ3.5, 3.9 (pp. 21โ25, 37) | 14โ16 |
| Coriolis theorem | Sec. 3.1, Eqs. 3.1โ3.8 (pp. 36โ37) | 17โ18 |
| Composition of angular motions | Eqs. 3.10โ3.15 (pp. 37โ38) | 19 |
| Angular velocity from Euler angles | none | 20โ21 |
| Configuration and motion of a frame | Sec. 3.2 (p. 38) | 22 |
| Relative kinematics | Sec. 3.2, Eqs. 3.16โ3.29 (pp. 38โ40) | 23โ26 |
| Relative dynamics, inertia forces | Sec. 4.9, Eqs. 4.59โ4.63 (pp. 52โ53) | 27โ28 |
| Potentials, Euler and Coriolis forces | none | 29โ31 |
Worked problems (Problems 7, 9, 20, 21, 25, 26, 27, 28, 30, 33, 34, 36, 37, 40, 41, 47, 49, 51) are solved step by step, in the order of the course, on the Topic 5 problems page.