Well I'm not the expert, but I don't think that a regular collision between large objects would ever have the force to cause a fusion reaction. Hydrogen and other lighter elements already are immensly difficult to fuse, and solid objects would never have the force to cause such a reaction (if the objects were going at realistic speeds and weren't lets say, traveling at the speed of light for some reason).
Actually if a planet the size of mars were to crash into Jupiter it would be enough to trigger fusion
EDIT: Solids? So frozen hydrogen? idk
Hmm, well I guess with a gaseous planet yes. And maybe with frozen hydrogen? Someone who is good at da sciences halp us D:
meta-metallic hydrogen?
If two suitably massive (ie. with lots of mass) or dense objects collide there can be fusion due to gravity compressing substances together so close fusion can actually occur, this sort of pressure can start up stars. But if you want one collision bound to start fusion wait for the collision between our galaxy and the one next door- http://en.wikipedia.org/wiki/Andromeda%E2%80%93Milky_Way_collision which if you believe the way in which we currently understand black holes will do all sorts of interesting things to the fabric of space time.
But since even now, without 4 billion years of accerating still to do, Andromeda is heading for the Milky way at 110 km/second and while it is unlikely due to the low densisty of space that there would be a collision between two planets, if there was it would certainly cause runaway fission in any fissile materials, and probably fusion too due to the massive heat/pressure.
Actually, Jupiter would have to gain another ten to fifteen times its current mass just to reach the brown dwarf stage, so a collision with a Mars sized body would be insufficient to trigger nuclear fusion, unless said Mars sized body were travelling at near relativistic velocities, and even then, the resulting fusion would be a short burst, and not a sustained reaction.