In 1960, NASA launched the "Echo 1A" balloon which had amongst its objectives to test the triangulation principle with satellites. I wonder why we don't have more "ballon" satellites for our GPS system : they are really less expensive to operate. The only problem I see is the wind, which makes their localisation quite unpredictable. At which altitude a balloon would not be affected by the wind?
In the range where a balloon can fly, there is always wind.
The NASA Echo 1A satellite was used to test the satellite triangulation. GPS works in some other way...
The GPS system used active satellites (non-pasive as Echo 1A) that constantly supplies their position, their time and additional coding information. So the satellite position should be known and corrected constantly. With a balloon constellation it will be very difficult.
Supposing that we can install all the equipment in a balloon, the top we could reach will be 53 km. The actual Navstar GPS constellation it is working at 20.200 km from our surface. So it means that we will need a lot of balloons in order to create a similar network, with a lot of ground stations to manage it.
Even if that point could be possible, there will be additional problems, as the areas with no islands or no continents when additional ground stations should be required to correct the signal and position.
The actual GPS satellite geometry was decided in order to reduce the number of satellites with the maximum cover.
A balloon satellite like Echo is obviously more affected by the 'wind' (aerodynamic drag) than a normal satellite in low earth orbit would be, but the GPS satellites are so far out that they are unaffected by Earth's atmosphere. Whether this would be true for balloons, with their much greater drag,I am not sure, but in any case they would still have the solar wind to contend with. In the course of time this would be enough to move them out of position.There is no altitude between Earth and the GPS system where Echo-type balloons would not be affected by a wind of some kind.