Wow, a chance to reply to pg! Yes, it could and has been done. Glider pilot Helmut Reichmann describes the technique for how to do this in a glider in his book, "Cross-Country Soaring". He has also performed the maneuver a couple of times.
You have to be flying in an area where there is strong wind shear, which means a sharp difference in wind speed with altitude. You could detect such wind shear when climbing or descending through it, either with GPS or by watching the ground and seeing how much you drift in relation to it. Cloud movements are another option. You also need a very maneuverable and aerodynamically efficient plane, so you won't lose a lot of energy from the sharp banking maneuvers required.
So assume that the wind speed increases sharply with altitude, maybe around 20 knots over 100 meters of altitude. You start out flying the same direction as the wind and then sharply dive 100-200 meters down into the space where the wind is weaker. This turns part of your altitude into kinetic energy, so your ground speed increases by about 100-150 kph. But due to the change in wind speed, your airspeed velocity has changed less than it would in dead air. So your total energy has increased. You can then make a sharp (>120 degree) turn in the direction you want to go, so you face partly into the wind again. You then sharply pull up, gaining about 100 meters of altitude and losing some of your velocity. The relative wind speed has increased with altitude, so your kinetic energy loss in relation to the air from pulling up is less than it would be in dead air. Overall, you have gained velocity, moved the aircraft and maintained your altitude "for free".
This process can be repeated, and you can keep doing it (albeit with some nausea, unless you're used to these sharp maneuvers) as long as there is sufficient wind shear. It will be easier to move perpendicular to the wind direction than directly with or against the wind, but the energy (altitude, velocity) you gain from this maneuver could be used to glide in any direction.
I don't want to make any grand claims that this technique can be used for anything practical (i.e. passenger transport). Glider pilots don't use it in competitions, because there are lots of techniques for moving around without an engine that are a lot better and easier to exploit. (You can gain altitude in thermals, ridge lift or mountain waves - and translated to horsepowers, a thermal carrying a 500kg glider upwards by 3m/s is a very powerful engine). But the technique is very cool as an intellectual curiosity.
[Edit: Actually, it's interesting that the authors mention possible applications for this to robotic aircraft. I'm sure you could make a robotic glider that used the meteorological principles that glider pilots use to move around without engines. The "Albatross" technique would only be a small part of this - glider pilots have extensively studied techniques for moving around without an engine, and there are lots of them. Glider pilots manage >100kph average velocities over >500km journeys on days with good weather, and robotic aircraft could in principle do the same].
There isn't a direct glider analogy to being "in irons" (you can always dive to gain airspeed, assuming you have any altitude to lose of course - not having any altitude would mean you are guaranteed to crash, since gliding is basically flying slightly downhill all the time).
But there are plenty of ways to get in trouble when gliding, most notably flying into an area where there aren't any sources of lift. This means you can't gain any more altitude, and you'll gradually lose your remaining altitude and probably have to land in a field or something.
You have to be flying in an area where there is strong wind shear, which means a sharp difference in wind speed with altitude. You could detect such wind shear when climbing or descending through it, either with GPS or by watching the ground and seeing how much you drift in relation to it. Cloud movements are another option. You also need a very maneuverable and aerodynamically efficient plane, so you won't lose a lot of energy from the sharp banking maneuvers required.
So assume that the wind speed increases sharply with altitude, maybe around 20 knots over 100 meters of altitude. You start out flying the same direction as the wind and then sharply dive 100-200 meters down into the space where the wind is weaker. This turns part of your altitude into kinetic energy, so your ground speed increases by about 100-150 kph. But due to the change in wind speed, your airspeed velocity has changed less than it would in dead air. So your total energy has increased. You can then make a sharp (>120 degree) turn in the direction you want to go, so you face partly into the wind again. You then sharply pull up, gaining about 100 meters of altitude and losing some of your velocity. The relative wind speed has increased with altitude, so your kinetic energy loss in relation to the air from pulling up is less than it would be in dead air. Overall, you have gained velocity, moved the aircraft and maintained your altitude "for free".
This process can be repeated, and you can keep doing it (albeit with some nausea, unless you're used to these sharp maneuvers) as long as there is sufficient wind shear. It will be easier to move perpendicular to the wind direction than directly with or against the wind, but the energy (altitude, velocity) you gain from this maneuver could be used to glide in any direction.
I don't want to make any grand claims that this technique can be used for anything practical (i.e. passenger transport). Glider pilots don't use it in competitions, because there are lots of techniques for moving around without an engine that are a lot better and easier to exploit. (You can gain altitude in thermals, ridge lift or mountain waves - and translated to horsepowers, a thermal carrying a 500kg glider upwards by 3m/s is a very powerful engine). But the technique is very cool as an intellectual curiosity.
http://www.cumulus-soaring.com/books/CrossCountrySoaring/Cro...
[Edit: Actually, it's interesting that the authors mention possible applications for this to robotic aircraft. I'm sure you could make a robotic glider that used the meteorological principles that glider pilots use to move around without engines. The "Albatross" technique would only be a small part of this - glider pilots have extensively studied techniques for moving around without an engine, and there are lots of them. Glider pilots manage >100kph average velocities over >500km journeys on days with good weather, and robotic aircraft could in principle do the same].