UW News

April 22, 1998

UW developing fleet of unmanned airplanes to gather Pacific Ocean weather data needed to improve accuracy of Northwest forecasts

The fog may be lifting for Northwest weather prognosticators who struggle to make forecasts without critical data on atmospheric conditions over the Pacific Ocean. Aeronautical engineering researchers at the University of Washington have been awarded a $456,000 grant from the Office of Naval Research to launch a fleet of unmanned airplanes to gather this missing weather data.

With the advent of global positioning satellite technology, these unmanned, or autonomous, aircraft can be piloted by computer on flights of over a thousand miles to conduct weather reconnaissance and a multitude of other missions.

“There’s no doubt that autonomous aircraft are the wave of the future,” explains Scott Eberhardt, associate professor in the UW Department of Aeronautics and Astronautics and director of the weather reconnaissance project. “The questions we have to answer are how to maintain and operate these planes over the long term, how to improve their design performance and how to integrate them into the air traffic control system.”

The UW will acquire 10 Aerosonde planes from Insitu, Inc., of White Salmon, Wash., creating what Eberhardt jokingly describes as the largest university a ir force in the country. While five of the planes will be deployed for the weather reconnaissance project, the remaining five will be used for instruction and research. This will allow UW students and faculty to test new design concepts and control system s for the first time on planes in flight and eventually may launch efforts to design and build unmanned aircraft at the university, Eberhardt says.

The Aerosondes are about six feet long with 11-foot wingspans. They are powered by five-horsepower, prope ller-driven engines. Most of the plane, including the engine, is made of fiberglass to reduce weight and improve flight endurance. Current models can fly for up to 30 consecutive hours with a maximum range of 1,000 miles. They reach cruising speeds of 90 miles per hour at altitudes of up to 18,000 feet.

Aerosondes already have been tested for remote weather-data collection in Australia and Canada with promising results. Eberhardt hopes to launch the first weather reconnaissance flights from the Washingt on coast this summer. The airplanes will be equipped with instruments to measure wind speed and direction, temperature, humidity, air pressure and other data that Puget Sound meteorologists can plug into forecasting models.

While East Coast forecasters have reams of data on upwind conditions supplied by weather stations across the country, West Coast forecasters have only limited data from ships, satellites and commercial air traffic over the Pacific. Furthermore, ships and planes tend to avoid severe w eather that would provide the most meaningful forecasting data. And even in good weather, there is a vast data void between ships on the sea surface and commercial airplanes traveling at 30,000 feet.

“We’ve got huge observational gaps, and we know they are causing major forecasting busts” explains Cliff Mass, professor of atmospheric sciences at the UW.

For example, he says, the major Seattle snow storm of November 1996 was not accurately predicted because forecasters had woefully incomplete data on t he level of moisture in the atmosphere over the Pacific Ocean. Without complete data, the models produced forecasts calling for little or no snow.

“These major mistakes in our forecasts are usually due to missing data that prevent our models from accura tely representing atmospheric conditions,” Mass says. “If we could plug those gaps, it would make revolutionary improvements in our forecasts. These Aerosondes could be the answer.”

However, existing unmanned aircraft have a limited range that enables t hem to travel only about 500 miles out into the Pacific before having to turn back to the coast. Even data from that relatively small area should significantly improve short-term forecasts in Puget Sound, Mass says, but data from 1,500 to 2,000 miles out into the Pacific will be needed to greatly improve long-term forecasts. Ultimately, he says, autonomous aircraft could be launched from Hawaii and Alaska to fly toward the Pacific Northwest coast, transmitting weather data from pre-programmed locations al ong the way. After landing, the aircraft could be transported back to Hawaii and Alaska by cargo ship or plane for subsequent flights.

Such an operation would be expensive, Mass concedes, but the cost would pale in comparison with the potential savings from more accurate forecasts of severe weather. “An improved forecast for a single storm could pay for this project 1,000 times over,” he says.

The challenge, according to Eberhardt, is to make the needed design improvements for enhancing the range and performance of autonomous aircraft and to develop strategies enabling manned and unmanned planes to safely share the skies. The Navy picked the UW to lead this effort, according to Eberhardt, because of the Department of Aeronautics and Astronautics’ Kirsten Wind Tunnel laboratory and its faculty expertise in design, control systems and aerodynamics.

Unlike radio-controlled airplanes and other remotely operated aircraft that rely on pilots being able to see the planes they are operating, autonomous airplanes are flown entirely by computer. A flight path is programmed into an onboard computer navigation system that controls the plane via global positioning satellite technology.

Until recently, most autonomous aircraft have been designed and built at high cost for specialized military and government functions. Before long, however, Eberhardt envisions unmanned airplanes replacing blimps at sporting events, monitoring rush-hour traffic and taking over tedious jobs such as scanning oil and gas pipelines or counting livestock on large ranches. The UW is one of the few research institutions in the country developing an optimal design for versatile autonomous aircraft that can be mass produced at rel atively low cost.

In addition, Eberhardt says, the UW is among the first universities training students in the unique design and control issues related to unmanned airplanes. Undergraduates will have the opportunity to take the Aerosonde planes apart to see how they work, reconfigure them for different missions and test them in the wind tunnel. Graduate students and faculty will be able to fly the aircraft to test new design configurations, control systems and instrumentation. Eventually, Eberhardt says , the UW may even build an unmanned airplane.

“While we can do all of the designs we want to for a Boeing 777, this department will never be able to build a 777 or even a small commercial plane,” Eberhardt says. “But we can design and build an autonomou s airplane. Instead of just testing a model in the wind tunnel we can actually put something in the air and look at how it flies. This takes the learning experience to new heights.”


For more information, contact Eberhardt at (206) 543-6508 and scott@aa.washington.edu or Mass at (206) 685-0910 and cliff@atmos.washington.edu