The Rivest methods in the CSR13 paper - ThreeBallot, VAV, and Twin - seem to be relatively simple. Not directly applicable to online voting, though, but perhaps they would be simple enough to prove to the people that regular voting has no voter fraud?