In variational mechanics, an action integral $\mathcal{S}$ is constructed from:
$
\begin{aligned}
\mathcal{S} = \int _{{t_{0}}} ^{t_{f}} \mathcal{L} \, dt
\end{aligned}
\tag{1}
$
where $\mathcal{L}$ is the Lagrangian. For a constant mass system, we can write:
$
\mathcal{L} = \frac{1}{2} \hat{\boldsymbol \omega} \cdot J \cdot \hat{\boldsymbol \omega}
\tag{2}
$
As in Manchester and Peck's paper, $\hat{\boldsymbol \omega}$ is indicates the formation of a quaternion with zero scalar part from the vector ${\boldsymbol \omega}$:
$
\hat{\boldsymbol \omega} = \begin{bmatrix}
0 \\
{\boldsymbol \omega}
\end{bmatrix}
$
$^{\epsilon} \hat{\boldsymbol \omega}$ is the variation of the angular velocity and is given by:
$
^{\epsilon} \hat{\boldsymbol \omega}
= 2 {^\epsilon q^{\dagger}} {^\epsilon {\dot{q}}}
\approx \hat{\boldsymbol \omega} + \epsilon (\hat{\boldsymbol \omega}\hat{\boldsymbol \eta} - \hat{\boldsymbol \eta} \hat{\boldsymbol \omega} + 2 \dot{\hat{\boldsymbol \eta}})
\tag{3}
$
```ad-important
Please see Equation ($16$) of Manchester's paper, which I believe has a typo; Equation ($2$) above is from your handwritten result.
```
Now, if Equation ($3$) is substituted into Equation ($2$), we should be getting what I think is called the varied Lagrangian $^\epsilon \mathcal{L}$.
Further, if you substitute the varied Lagrangian into Equation ($1$) and then compute the variation of the action, you get Equation ($4$) below (**Please correct this phrasing as I imagine this is wrongly or poorly done on my part.**)
$
\delta{S} = \frac{\partial}{\partial \epsilon}\Bigr|_{\epsilon=0} \int _{{t_{0}}} ^{t_{f}} \frac{1}{2} {^{\epsilon} \hat{\boldsymbol \omega}} \cdot J \cdot {^{\epsilon} \hat{\boldsymbol \omega}} \, dt
\tag{4}
$
Manchester and Peck shows that this eventually becomes:
$
\delta{S} = \int _{{t_{0}}} ^{t_{f}} \hat{\boldsymbol \omega} \cdot J \cdot (2\hat{\boldsymbol \omega} \hat{\boldsymbol \eta} + 2 \dot{\hat{\boldsymbol \eta}}) \, dt
\tag{5}
$
#### Angadh's work so far
The integrand of Equation (4) is:
$
^\epsilon \mathcal{L} =
\frac{1}{2}
{^{\epsilon} \hat{\boldsymbol \omega}} \cdot J \cdot {^{\epsilon} \hat{\boldsymbol \omega}}
\tag{6}
$
Using Equation $(3)$ in ($6$) is:
$
^\epsilon \mathcal{L} =
\frac{1}{2}
\bigg(\hat{\boldsymbol \omega} + \epsilon (\hat{\boldsymbol \omega}\hat{\boldsymbol \eta} - \hat{\boldsymbol \eta} \hat{\boldsymbol \omega} + 2 \dot{\hat{\boldsymbol \eta}})\bigg) \cdot
J
\cdot
\bigg(\hat{\boldsymbol \omega} + \epsilon (\hat{\boldsymbol \omega}\hat{\boldsymbol \eta} - \hat{\boldsymbol \eta} \hat{\boldsymbol \omega} + 2 \dot{\hat{\boldsymbol \eta}})\bigg)
\tag{7}
$
And we can expand this as:
$
\begin{align}
^\epsilon \mathcal{L} =
\frac{1}{2}
\hat{\boldsymbol \omega} &\cdot J \cdot
\hat{\boldsymbol \omega}\\
+\frac{1}{2}
&\hat{\boldsymbol \omega}
\cdot J \cdot
\bigg(\epsilon (\hat{\boldsymbol \omega}\hat{\boldsymbol \eta} - \hat{\boldsymbol \eta} \hat{\boldsymbol \omega} + 2 \dot{\hat{\boldsymbol \eta}})\bigg)\\
+
\frac{1}{2}
&\bigg(\epsilon (\hat{\boldsymbol \omega}\hat{\boldsymbol \eta} - \hat{\boldsymbol \eta} \hat{\boldsymbol \omega} + 2 \dot{\hat{\boldsymbol \eta}})\bigg)
\cdot
J
\cdot
\hat{\boldsymbol \omega}
\end{align}
\tag{8}
$
So.... how does Equation ($8$) eventually become Equation ($5$) above...?
```ad-todo
title: Steven's Mission If-Possible
I was unable to show myself how Equation ($4$) becomes Equation $(5)$. Can you derive this below, please?
```