Fullscreen HTML5Med Res (??MB)HTML4

When most people think of radio controlled aircraft, they imagine small models that fly around a small field. But this Australian designed and manufactured aircraft has crossed the Atlantic and performed many other record feats. The idea of a robot aircraft flying at 40,000 feet and with a range of up to 7000km takes some getting used to, especially when you realise that commercial jet aircraft fly at the same height and have a similar range. Add in the fact that this radiocontrolled aircraft has a wing‑span of only three metres and weighs only 15kg and the feat is all the more incredible. In what must be one of the leastpublicised epics for some time, the North Atlantic was crossed by the Australian-designed Aerosonde robotic aircraft in August 1998. The Aerosonde was the first robotic aircraft to cross the North Atlantic Ocean and it was also the smallest aircraft ever to do so. 8  Silicon Chip As you might imagine, for a crossing of the Atlantic the aircraft is not under radio control for most of the flight. Instead, the Aerosonde employs an autopilot and GPS fixes to guide it most of the way. So notable has this aircraft become that it is now a joint development with the US military and its future uses could be quite widespread. The history‑making Aerosonde, nicknamed “Laima,” landed smoothly on a field at the Benbecula military range in the Outer Hebrides, Scotland, after a 27‑hour non‑stop flight from St Johns, Newfoundland, Canada. By BOB YOUNG Powered by a tiny one‑cylinder 20cc engine, the aircraft autonomously guided itself across the 3200km stretch of the North Atlantic while burning less than six litres of fuel! The Aerosonde rigorously maintained a flight path approved by aviation authorities and landed exactly as scheduled while collecting meteorological data throughout the flight. The tiny aircraft is packed with computers, a communications radio, a GPS satellite guidance system and meteorological instruments. This crossing followed extensive trials held in Australia, Canada and Asia over the previous year. It followed a path similar to that taken by the first Atlantic manned crossing by Alcock and Brown. Hard to believe, but as the tiny Aerosonde makes a low pass over an airfield it could be coming in to land after a flight of thousands of kilometres from who knows where. The flight was conducted by the University of Washington and US engineering company, The Insitu Group, using aircraft purchased by the University from co‑developer Environmental Systems and Services. “We’ve flown the same mission as a $10 million unmanned craft at a fraction of the cost,” said Professor Juris Vagners of the University of Washington Aeronautics and Astronautics department. The aircraft cost $US25,000. Aerosonde development has been underway since 1992. Phase I Aero-sondes were given their full operational trial by the Bureau of Meteorology in early 1998 and passed with flying colours. In addition, Aerosonde RA have conducted several missions in Australia, Taiwan, Canada and the United States, including flights of over 30 hours and at 16,000 feet. To date over 30 Phase I Aerosondes have been delivered. Their specifications are as shown in Table 1. Aerosonde is currently working on a Phase 2 version which will have a range up to 7000km, up to 5 days endurance and a ceiling of 40,000 feet. While Aerosonde resembles a model aircraft externally, this resemblance is purely superficial. True, some components are essentially model aircraft components, however the operationing systems are structured along traditional military lines. Take‑off and landings are arranged so that manual or automatic control can be engaged. Manual control is en- gaged when the pilot plugs his control box into the computer control console. Currently, all take‑offs and landings are done under manual control. The Aerosonde uses a gyroscopic autopilot and standard model aircraft servos but the details of these have not been released. The aircraft is a joint Australian/ American design and manufacture has commenced at Melbourne. Component manufacture is contracted to a number of Australian and interna- Table 1: Phase 1 Aerosonde Specifications Wingspan: .............. 3 metres Weight: ................... 15kg Engine:.................... 20cc petrol (Avgas) Performance:........... Cruise 20‑30m/s Range:..................... >3,000km, Endurance .............. >30 hours Height Range: ........ Surface to 16,000 feet Payload:................... 1‑2 kg Operation:................ Autonomous Navigation:.............. GPS Communication:...... UHF Radio, Satellite Observations:.......... Wind, Pressure, Height, Temperature, Moisture MAY 1999  9 tional groups. Following a series of engineering demonstrators built in 1992-94, the first Aerosonde suitable for field testing was flown in June 1995. In November, the Aerosonde Development Consortium took several examples to Melville Island north of Darwin for the Maritime Continent Thunderstorm Experiment. This was primarily for engineering trials, since at the time of deployment they had flown less than 50 hours Since then, Aerosondes have come a long way. They can be used for meteorological and environmental monitoring. For example, they are able to do some very useful work in monitoring sea breeze fronts, gust fronts and storms, working with Doppler radar. Winds and thermodynamic data measured during the more interesting missions, along with more details on the aircraft, are available on the MCTEX web page at www.aerosonde.com An interesting point is that the Aerosonde cannot determine wind by the standard wind‑triangle method whereby wind is calculated directly by differencing groundspeed and airspeed vectors. This is because while it has vector groundspeed from its GPS, it does not have a heading sensor. Hence true airspeed is known only as a scalar. It turns out that vector groundspeed and scalar airspeed provide sufficient information for wind‑finding if they are compared through the course of a turn, say through about a quarter of a circle. The algorithm is given in the Aerosonde RA publications. Aerosonde flight‑plan segments therefore include a specified interval for wind‑finding S‑turns. Wind-finding requires about 10 seconds manoeuvring (spatial resolution of about 200 metres). The following flight reports downloaded from the Aerosonde web site make interesting reading: “1996: 24 HOUR FLIGHT ‑ 21st November 1996 At about 5 in the afternoon of 21 November 1996, Aerosonde Morti‑ cia landed at Geelong, having flown around the local model‑aircraft field for 24 hours at about 300m altitude. It Fig.1: this is the flight track of the record‑breaking North Atlantic flight. This consisted of a series of way points for a route that went slightly south of a great circle (shortest distance) to the landing site at DERA Benbecula Range in the Outer Hebrides. The altitude was specified at 1680m, dropping to around 150m on approach to Benbecula. Before launch, complete flight simulations had been made using winds provided by the US NOAA/NCEP model to provide approximate times at each way point. 10  Silicon Chip some tall clover. Overall the performance was quite comparable to a good manual land‑ ing. Although the landing was done under autopilot, it was not quite au‑ tonomous; guidance onto the runway centreline was done visually from the ground station rather than being left to the onboard tracker. However the test produced good results in position measurement by differential GPS, so the next step to fully autonomous landing will be straightforward.” The Aerosonde is normally launched from a cradle atop a car roof rack. Takeoff is normally under manual control but can be be completely automatic, as can the landing on a remote field. had enough fuel on board for another 10 hours or so of flying. Meteorological data were reported throughout, in conditions ranging from fair at the start to blustery, with heavy showers as a cold front moved through early on the 21st. For us this was a milestone in not only basic performance, but also reliability and readiness for routine operations. Several more such flights will have to be successful before we feel con‑ fident but certainly the program is steadily developing towards reliable and repeatable operations.” “1997: AUTOMATIC TAKEOFF AND LANDING ‑ 22nd September 1997 On 22nd September 1997 an impor‑ tant step was taken toward automatic rather than manual control of takeoff and landing. In a one‑hour test at Trout Lake in Washington, Aerosonde Millionaire flew under autopilot con‑ tinuously from launch to touchdown. Figures show the landing as plotted on ground‑station displays. The aircraft touched down smoothly on the Trout Lake runway, made one small bounce and a large‑angle yaw, and then decelerated rapidly through All in all the Aerosonde project is a credit to the dreamers who dared to make it happen. What an audacious project: to send a single engine, miniature aircraft across one of the most hostile stretches of ocean in the world. Once again we see vividly demonstrated, that by standing on the shoulders of giants, we can see past the crowds who would otherwise limit our vision. Where will this all lead? The developers envisage a global robotic airline operating out of Australia with a distributed set of launch and recovery sites (“airports” if you like) and a global command site possibly located SC in Melbourne. Acknowledgement: Much of the material in this article courtesy of Aerosonde Robotic Aircraft Pty Ltd. For more information, visit their website, www.aerosonde.com.au MAY 1999  11