In 1896, the Locomotive Act was passed and stated that all road locomotives - cars to you and us - could travel at a maximum speed of only 4mph and had to be manned by a crew of three, one of whose job it was to walk 60 yards in front of the vehicle waving a red flag. It wasn’t until 30 years later, when automobiles had been tried and tested, that they were deemed safe enough to be driven by just one and as fast as 14mph. BAE Systems are looking to push the same kinds of boundaries in aerospace, except this time, they’re hoping to have no crew at all.
ASTRAEA is a six-year programme with the aim of enabling the use of unmanned aircraft in normal airspace: planes that fly themselves. Yes, there are drones flown about battlefields already by people on the ground with a screen and a joystick, but they're very much held on a lead. What BAE and its partners in the project have been trying to get together is a system that will do it all by itself with only a pilot on the ground watching - with hands hovering by some controls just in case.
The mechanics of what would seem like the more complicated flight operations have already been mastered. Aircraft have been quite capable of taking off and landing all by themselves for some time. It’s the more subtle, reaction-based decision-making that’s proved trickier. As a result, unmanned flight has been allowed only in segregated airspace for the time being but the goal of ASTRAEA is to make something that is able to make decisions and operate just as a pilot would so that it is “safe to fly and flown safely”, as the motto goes - so much so that the governing body, the Civil Aviation Authority, will allow the system to control flights anywhere across the skies without any special conditions attached whatsoever.
With the programme due to come to an end in March 2013, Pocket-lint was invited to hear and see just how far ASTRAEA has come, and the good news is that, according to the director of project, BAE Systems’s Lambert Dopping-Hepenstal, it’s come a long way indeed.
Rather than start with an entirely new-built and purpose-designed aircraft, the ASTRAEA engineers made the wise choice simply to retro-fit a vehicle they already knew everything about, therefore cutting out on quite a lot of the preliminary work and calculations. What they chose was a Jetstream 31 - a twin-turboprop passenger plane introduced in 1982 - and stripped out the innards replacing the rows of 19 passenger seats with an enormous computer and communications array for both flying and recording the data from the test sorties; a mobile sky laboratory, in other words.
With one half of the fuselage taken up with the equipment, there’s just about room for three engineers to sit and take care of any necessary readings and operations as well as two pilots at the front with their hands over the yoke just in case anything goes wrong.
“It’s just like a normal plane,” jokes one of the crew, Rod Buchanan, as he shows Pocket-lint around, “except without the two best parts - the stewardesses and the toilets.”
The surrogate craft has a few important additions to its outer as well. Underneath are an infra-red sensor and an ADS-B antenna. The former is to identify any heat signatures in the event of an emergency landing. It’s all very well having a database full of known safe locations to land but, if the chosen field happens to be hosting a car boot sale at the time, it’s something ASTRAEA needs to know. The ADS-B is for detecting other aircraft in the sky, most of which broadcast and receive signals from a box of their own.
On the topside of the Jetstream 31 is a second antenna, used to stay in contact with radio sources on the ground and satellite communications, but the most curious one of all is the piece of kit just inside the cockpit window facing out to the sky ahead - a video camera, the electronic eye. The unit itself is standard enough but the data it captures is wired direct to its own dedicated eight-core computer where the image processing takes place. ASTRAEA uses the information to detect, range and avoid thick cloud and storms by flying around or underneath them. It also relies on the camera for sensing and avoiding other aircraft or objects such as hot air balloons in the sky - all important when that other vehicle isn’t itself fitted with an ADS-B antenna.
The result is a system that can integrate with air traffic control, perform its own decision making, replace the pilot’s responsibility for sensing and avoiding traffic and will work even when the comms link to the ground is lost. So, does it work?
Well, the "Flying Test Bed" Jetstream 31 with ASTRAEA on board will be undergoing some 20 flights up to the winter with all of the data from each being recorded on to around 300GB of a 500GB SSD, and later uploaded on the ground to what Buchanan describes as “essentially a massive RAID array”. They used to have the unit on the aircraft but the workings of spinning, mechanical HDDs weren’t too fond of the odd rough landing. And, if you think that the computational hardware on which ASTRAEA runs is state of the art, think again.
“It was literally hardware we had on our shelves bolted into this aircraft,” explains Buchanan.
“We’re only using two of the cores at the moment. The main work on ASTRAEA is autonomy, decision making, sense and avoidance, adaptive routing, contingency management, emergency planning and managing the communication system. All the of that software is currently running on one core. We’re only beginning to use a second core of the eight a little bit now.”
Simple hardware with some complicated mathematics, and the findings have been that ASTRAEA can effectively see better than a human pilot and fly accordingly, with just as much skill. So, it works and, doubtless, soon enough the CAA will decide that it works too and that leaves one big question - just why do we want unmanned aircraft anyway?
“Well, we’re not talking about unmanned passenger aircraft,” says the project boss, Dopping-Hepenstal. “You're not going to be going on your holidays without a pilot. It's about opening up a whole new branch of aviation; about looking at things that are difficult or that you can't already do.”
Search and rescue is one of the areas that could most benefit. Not only can you send aircraft into conditions where you normally couldn’t risk a human pilot, but there’s also the fact that a UAV (Unmanned Aerial Vehicle) could work 24 hours a day without any degradation in performance. Bush fires and disaster fighting work could be carried out to greater effect with planes able to fly far lower and deliver their payloads with much greater efficiency, but the first uses, which we expect to see by the end of the decade, will most likely be out at sea monitoring fisheries and patrolling.
Ultimately, as Dopping-Hepenstral admits though, the answer is that we simply don’t know. The point of ASTRAEA is to open up the possibilities in this field, perhaps the more exciting part will be to see how it’s put into practice.