Using a sextant and a clock you can navigate by measuring the position of the observable stars and comparing them with the known map of the fixed stars. However, if it was cloudy couldn't you theoretically use the CMBR map instead?
The cosmic microwave background radiation (CMBR) map is a constant pattern in the sky through which the earth travels and rotates. Given an accurate clock, microwave antenna and model of the earth's orbit, shouldn't it in principle be possible to match the patch of the the CMBR pattern above you with the overall CMBR map and therefore calculate where you are located on the earth's surface?
Yes, I appreciate that measuring the CMBR sufficiently accurately at 160GHz is not trivial and needs a large antenna (preferably a phased array), probably with a cryogenically cooled microwave receiver. However, an antenna of this size should be easy to accomodate onboard a large military ship for example. There would be a tactical advantage in the USA turning off the GPS satellites and switching to CMBR-based navigation prior to launching an attack.
Presumably for this application some of the the major sources of microwave interference that scientists normally have to carefully subtract in order to get the classic CMBR map could be left in the data if they are from outside the solar system (e.g. milky-way sources of microwaves)?
What are the other practical issues and how might they be resolved to make a working system?