HAWKS and albatrosses soar for hours or even days without having to land. Soon robotic gliders could go one better, soaring on winds and thermals indefinitely. Cheap remote sensing for search and rescue would be possible with this technology, or it could be used to draw up detailed maps of a battlefield.
Glider pilots are old hands at using rising columns of heated air to gain altitude. In 2005 researchers at NASA's Dryden Flight Research Center in Edwards, California, flew a glider fitted with a custom autopilot unit 60 minutes longer than normal, just by catching and riding thermals. And in 2009 Dan Edwards, who now works at the US Naval Research Laboratory in Washington DC, kept a glider soaring autonomously for 5.3 hours this way.
Both projects relied on the glider to sense when it was in a thermal and then react to stay in the updraft. But thermals can be capricious, and tend to die out at night, making flights that last several days impossible, says Salah Sukkarieh of the Australian Centre for Field Robotics in Sydney. He is designing an autopilot system that maps and plans a glider's route so it can use a technique known as dynamic soaring when thermals are scarce. The glider first flies in a high-speed air current to gain momentum, then it turns into a region of slower winds, where the newly gained energy can be converted to lift. By cycling back and forth this way, the glider can gain either speed or altitude.
"Theoretically you can stay aloft indefinitely, just by hopping around and catching the winds," says Sukkarieh, who presented his research at a robotics conference in Shanghai, China, last month.
Inspired by albatrosses and frigate birds, the operators of radio-controlled gliders have used dynamic soaring to reach speeds of more than 600 kilometres per hour by flying between two regions of differing wind speeds.
To plan a path for dynamic soaring you need a detailed map of the different winds around the glider. So Sukkarieh is working on ways to accurately measure and predict these winds. He recently tested his autopilot on a real glider, which made detailed wind-speed estimates as it flew.
The system has on-board sensors, including an accelerometer and altimeter, which measure changes in the aircraft's velocity and altitude to work out how the winds will affect the glider. From its built-in knowledge of how wind currents move, the system was able to work out the location, speed, and direction of nearby winds to create a local wind map.
By mapping wind and thermal energy sources this way and using a path-planning program, the glider autopilot should be able to calculate the most energy-efficient routes between any two points. The system would be able to plot a path up to a few kilometres away when the wind is calm but only over a few metres when turbulent, as the winds change so quickly, says Sukkarieh.
He says that the amount of energy available to a glider is usually enough to keep it aloft for as long as it can survive the structural wear and tear. He plans to test the mapping and route-planning systems more extensively in simulations, to be followed by actual soaring experiments.
"I think we have some examples from nature that mean this should be possible," says Edwards, who is not involved in Sukkarieh's research. "We're just taking our first baby steps into doing it autonomously."