Mobile ad-hoc networks are a favorite research topic, everybody and his brother has written thick papers on it or built some testbed. Problem is that none of these has led to any practically significant developments, the routing, clock synchronisation, power mangement and dozens of other issues are just too complicated. So if there will ever be a breakthough on this, it will most likely be an extremely boring (for Joe Average) discovery in the field of graph algorithms or something similar.