Well that's an even more complicated problem. That would depend on what type of cell you wanted to target and in what tissue. Most tissues are quite heterogeneo

Well that's an even more complicated problem. That would depend on what type of cell you wanted to target and in what tissue. Most tissues are quite heterogeneous and constantly remodeling. Even if you could get the coordinates of a particular cell before it moved, targeting the specific cell (and not the others that are likely to be present in the vicinity)is a problem unto itself. Don't diss diffusion and biorecognition, they're quite effective. In my mind, any robot designed on the scale to target individual cells would have to mostly rely on circulation, diffusion, and chemistry to target a particular cell type. The mechanics are far from deterministic on the nano scale. I highly recommend this book if you're considering a project like this: http://www.garlandscience.com/product/isbn/0815341636 http://www.garlandscience.com/product/isbn/0815341636. If this is for a visualization or a game, I would just fake the coordinate system.