Now testing: pilotless planes to grab weather data over Pacific

When a single-engine Aero-sonde airplane crossed the Atlantic last year, it did so on only a gallon and a half of fuel. Mere efficiency, however, wasn't the point. The project sent the message that model-size, pilotless planes are comers as long-distance scientific helpers.

The real goal of the transatlantic flight, says Scott Eberhardt, an associate professor of aeronautics and astronautics at the University of Washington, is to develop planes that can gather hard-to-retrieve weather data over the Pacific Ocean. The problem, Professor Eberhardt explains, is the northern Pacific region known as the Gulf of Alaska, where much potentially helpful weather data sits in limbo.

Planes heading to Asia fly at 35,000 feet, which is above most of the weather. Boats are limited to collecting surface data, and even satellites are hard-pressed to gather accurate readings. That's where a fleet of pilotless airplanes at the University of Washington come in. Not to be confused with remotely piloted vehicles, these aircraft, can fly thousands of miles with no human in the loop. Instead, they are piloted by computer, with the assistance of the Global Positioning System. To save weight, the plane has no landing gear and must make grass-cushioned, belly landings. Takeoffs are from a cradle attached to the top of a car. One of the challenges of the project, Eberhardt says, is in finding approved landing areas.

During last year's transoceanic test flight, the University of Washington team wanted to recreate the first North Atlantic flight path, from Newfoundland to Ireland, but Irish officials balked at the idea. At the 11th hour, Scotland became the alternative landing site.

"There's concern," Eberhardt says, "that you've got this airplane with no pilot that can't see anything. It's highly unlikely it would slam into another plane." The US Federal Aviation Administration is also cautious, requiring that all autonomous planes come equipped with transponders to receive radar signals and respond to them.

This means more weight and more power is required with the net result of delaying the plane's development. Eberhardt believes Pacific weather-tracking is several years off for the Aerosondes, which need to fly higher and farther than presently possible. Even with the planes flying only one direction, they need a range of 3,000 to 5,000 miles.

Right now each Aerosonde costs roughly $25,000, which makes them the bargain-basement entry in the unmanned aerial vehicles race. The Aerosondes can be used again and again, but attrition is a factor. The university researchers estimate that if each plane is able to make 10 flights, the cost to record key weather data would be roughly $120 per data point, making the plane's $25,000 cost competitive with that of weather balloons and rocket soundings.

(c) Copyright 1999. The Christian Science Publishing Society