Even if the fundamental workings of the brain remained a mystery forever, would that necessarily prevent us from building a robotic simulation? I read about a recent study which looked into the inputs entering, and the outputs leaving a rat's hippocampus. With enough observations, it might be possible to predict the output statisically from the inputs, even though the intervening processes occurring within the hippocampus remain unknown.
If such a simulation could be conducted for every sub-unit of a brain, their interactions could be modelled and their outputs relayed to the rest of the body. Although you would know the architecture of the computer programme performing this feat, you could still be ignorant of that of the brain... it all sounds so simple - Ha.
People would argue that was has been accomplished is a mere duplication of the outward appearances of consciousness. This, of course, would be true, but it would be just as much evidence for an underlying consciousness entity as is available from talking with you lot, or even face to face with friends and family.