In this paper on the Burgers equation (1.1),
$$ u_t = \left(\nu u_{xx} - \sigma u^2\right)_x, $$
the authors use a simple change of variables (2.1a,b)
$$\begin{align} u(x,t) &= q(y,t) \\ x &= y + s(t) \end{align}$$
to write the equation as (2.2)
$$ q_t - \left(q_{yy} + 2\sigma qq_y\right)=\dot{s}(t)q_y. $$
However, should not the $2\sigma qq_y$ term have the opposite sign?
It may be because $\sigma$ is defined as $\pm 1$, so either sign is appropriate.