A point in each rectangle is individuated vectorially by
$$
\eqalign{
& {\bf P}_{\,1} = {\bf t}_{\,1} + a{\bf u}_{\,1} + b{\bf v}_{\,1} \quad \left| {\;0 \le a,b \le 1} \right. \cr
& {\bf P}_{\,2} = {\bf t}_{\,2} + c{\bf u}_{\,2} + d{\bf v}_{\,2} \quad \left| {\;0 \le c,d \le 1} \right. \cr}
$$
with an obvious meaning of the vectors
- ${\bf t} _k$: vector of a chosen vertex;
- ${\bf u} _k$: vector parallel to a side from the chosen vertex;
- ${\bf v} _k$: vector parallel to the other side from the chosen vertex;
Therefore, for the points to coincide we shall have
$$
{\bf P}_{\,1} = {\bf P}_{\,2} \quad \Rightarrow \quad a{\bf u}_{\,1} + b{\bf v}_{\,1} - c{\bf u}_{\,2} - d{\bf v}_{\,2} = {\bf t}_{\,2} - {\bf t}_{\,1}
$$
which is a linear system of $3$ equations in $4$ unknowns .
For the rectangles to intersecate each other, we shall verify that
- the system is solvable (in $a,b,c,d$);
- the set of solutions intersect the domain $[0,1]^4$.
There are various approaches to verify the first condition.
We can for instance go algebraically and check the the rank of the coefficients matrix, and that of the complete matrix:
- if both have rank $3$, we have a set of $\infty ^1$ solutions (a line), depending on one of the unknowns taken as a parameter,
geometrically that means that the planes of the ractangles intersect on a line;
- if the coefficients matrix has rank $2$, then the rectangles planes are parallel,
and they will be coincident or not depending if the complete matrix has rank $2$ or $3$,
if they have both rank $2$ then we have $\infty ^2$ solutions (two parameters);
- if the coefficients matrix has rank $1$, one or both the rectangles are degenerated
(and you can check this case in advance).