I've been thinking of doing a kinda riff on Planet of the Apes, where a space faring civilization sets out to colonize the stars only for one of the ships to wind up back on the home planet.
I like the idea of the round trip happening in geological time (like hundreds of millions of years) so that by the time they return the remnants of their civilization are not obvious, and in fact don't get discovered until the final act in an archeological dig or something like that.
I'm not sure how to contrive the return in such a way that it is not obvious to the returning craft. My first instinct is that maybe the ship gets cast a-drift somehow. Perhaps a miscalculation sends the ship ever so slightly off course leading it to drift in deep space in a low-zero power mode. I could imagine it drifting long enough that any internal clock may overflow an indefinite number of times before the ship approaches near enough to any star that it might recharge it's solar cells. In that time perhaps the navigation systems get corrupted, perhaps it uses time to measure how far it has traveled? But that seams a little contrived.
I figure that if we suppose enough time has passed that a whole new set of species are thriving on the planet, then any recognizable astronomical objects may not be in positions that the crew recognize them once they are awakened from stasis. Are hundreds of millions of years too short a span to suppose the solar system would look sufficiently different to fool trained space farers?
Edit:
So many really cool answers! I'm partial to the orbital mechanics solution, that feels subjectively nicer than any of the universe is actually curved solutions. Not because those are bad, just because you need such a small universe for that to feel like the right solution. I'm really torn about who to award this to, because the orbiting the galaxy idea is great in my opinion, fits the time scale I need perfectly!
A couple of issues seam to crop up about the details of the ship and how the navigation computer could last so long. Honestly I hadn't settled on anything there yet, and was toying with ideas along the lines of the ship is a (relatively) tiny object, just a collection of high density information stores and a couple of nano-machines. Small enough to not need astronomical amounts of energy to get reasonable fractions (like $0.01 \space c$) of light speed. Basically the ship would rebuild the crew and survival units from available resources on the planet it lands on, without having to transport them physically (of course they were scanned pre-travel with memories etc).
Either that or it is a standard issue large hulled ship full of frozen people, over engineered to the nth degree. Such that the original designers had a round trip of 100 000 years in mind but used a safety factor of 1000 for everything. Because if a thing needs to be frozen for 100 000 years, what are a few 100 million?