I think this will be hard for TI to implement.
Thinking of the parametric equations as representing the paths of two particles in time, are we looking just for a place where the paths cross, or a time when the particles will be in the same place simultaneously?
By doing graph trace, you can get good estimates of where the graphs cross and see whether the t-values are the same, or at least close.
If you are looking for values of t where the particles are in the same spot simultaneously, one approach is to define f1(x) = (x2(x)-x1(x))*(y2(x)-x2(x)) and look at its zeros. These will be candidates for the t-values you want, but they have to be checked.