The cost of the hardware may be low, but it is certainly labor intensive, and the functions are technical. I would assume that grouping these things will reduce the personnel overhead, but this doesn’t appear to be very efficient.

DARPA just finished its latest robo-car challenge, a vehicle that can autonomously drive through a city environment in traffic, with some decent results as far as navigation goes, but how do you “teach” a computer what is militarily “interesting”? What does the “Shoot/Don’t Shoot” course for robots look like, if you are going to arm them?

We had that ability to fly missions based on stored instructions for the navigation system, from engine start to engine shut down, but that didn’t do much good when you found something “hot” in the middle and needed to loiter.

We are on the trail, but there’s a ways to go yet, and buying too much of today’s technology means there’s no money for tomorrow’s.

I remember telling people rushing to get into the tech market that there was a huge bubble caused by Y2K and that sales would drop off afterwards because business was generally on a 5-year life cycle for equipment, so hardware sales would be down from 2001 through 2004, but no one listened. Today’s hardware may be great, but save something for the wonderful things that will be created tomorrow, or the creativity will die.

On the edge of chaos

In a significant breakthrough for Australia’s defence strategy, and a world first to boot, an Unmanned Aerial Vehicle (UAV) operated by the Defence Science and Technology Organisation (DSTO), has achieved autonomous flight.

DSTO scientists are also researching and designing programs for groups, or swarms, of small, inexpensive UAVs that may become part of Australia’s future defence arsenal.

Recently, a Codarra ‘Avatar’ UAV was guided by software that directed the aircraft’s autopilot in flightpath selection during a short mission at the Australian Army’s Graytown range near Melbourne.

Though only of short duration and involving simple choices, this ‘UAV first’ demonstrated in-flight intelligent agent control of the aircraft and fully autonomous mission selection capabilities.

DSTO Flight Systems Research Leader, David Graham, said while the task achieved by the UAV, “was simple”, he pointed out that it was the first step in accomplishing far more complex autonomous flight control.

“The hard part was probably the integration of all the components,” David said. “The achievement means that eventually one person would be able to manage many UAVs, rather than have several people managing the operation of one UAV. Flight-testing will start next year on multiple UAV teams.

In other developments, DSTO mathematician Alex Ryan is part of a research team using advanced mathematics and state-of-the-art computing to design ‘collective intelligence’ for swarms of UAVs.

Like a plague of locusts or a swarm of killer bees, these small expendable unmanned aircraft could be used for surveillance of possible enemy activity, as small weapons carriers, or to investigate areas too dangerous for humans.

“We’re working at the edge of chaos,” says Alex. “There’s a fine line between systems that are too ordered and stagnate; or systems that are too chaotic and collapse into total disorder.

“Swarm behaviour as such is not what we are after,” he says. “Swarms — like the notorious killer bees — concentrate on attacking a single enemy in vast numbers. Our aim is rather to develop an intelligent and communicating network.

“Each ‘agent’ in the network has its own utility function while there is an over-arching utility function for the whole system. It is vital that the agents don’t work at cross-purposes, and they must each be able to react to unexpected circumstances.”

Alex says that many small, simple and inexpensive UAVs, costing less than $20,000 each, are a more practical answer than larger, more sophisticated vehicles costing millions of dollars.

He says that much work needs to be done to reduce the imbalance between unmanned vehicles and the people who control them. “At present, each unmanned aircraft needs a ground crew of about thirty people,” he says.

Australia too poses its own special problems, as imported technology may be quite unsuitable for our topography. “We are in a ‘littoral’ region, with miles of coastline and chains of islands. And we have a vast interior. These geographical features make quite special demands on the design of unmanned aerial vehicles,” Alex adds.

Seems they agree with your assessment of current UAV technology! They quote 30 people for one UAV currently. That’s quite ridiculous really, can hardly be said to be *cost effective*!

To continue… One of the things they were testing was a collision avoidance system they developed. It was a success apparently. They also developed a real-time, 3D terrain mapping system for a UAV. I heard the idea was to send a UAV ahead of a flight of ground-hugging FB-111’s flying with only passive systems operating (so they would be much harder to detect) to target. Another UAV project I know DSTO worked on was called ‘Nervana’. I think the premise was ‘Automating the Battlespace’. I believe DSTO developed the Aerosonde, or at least were a major participant, and are now in partnership with NASA & the US military (and NATO I believe) to develop them further. I think they recently had successful trials in Texas… not sure. I hear a lot of things! LOL.

You sound like you had a lot of fun with your dad, though I bet he was busy! 🙂 He sounds like a very interesting man. 🙂

Cheers!

During the time I spent liasing with DSTO in Salisbury here, they had successfully tested a squadron of UAV’s flown by a single operator. Remember, these are the guys who developed Jindalee (the working over-the-horizon radar system). There was a reason for that… and it wasn’t about watching what was coming so much as what was going! 😉 This is why Rockwell spent a huge fortune buying Jindalee when the retarded Howard stole Australia. DSTO are very advanced in Comuter Science, and I saw some AI’s and expert systems that made my mout drop (and drool!) LOL And was kinda scary too! Mostly because most of the Scientists at DSTO are crazy!

The first time I arrived at DSTO, and I was cleared through security (and I was escorted the entire time, and mobile phones and laptop’s were confiscated), we drove for several KM’s past pastoral land with sheep grazing. Eventually I saw a building with a huge golfball on it (Radome of course), and more sheep. I commented that that was a good idea to have organic lawnmowers! LOL The Scientist I was with said, dead pan, “Oh. I suppose they are good for that too.” And of course my curiosity was peaked! Eventually I discovered that the real reason was that when they tested Jindalee (or other Radar & microwave systems) sometimes, if an error in alignment or an atmospheric anomaly occurred, the microwaved would damage or kill the sheep and they could measure what happened! I asked if anyone had been hurt or killed this way, and was told “That almost never happens.”

Crazy! LOL

Helicopters are just too noisy.

The UAVs are fine for surveillance, but two rounds, even of AGMs is not a lot of support and the whole thing of switching among 4 different drones while only actively controlling one at a time, makes a massed attack impossible.

I would note that my Dad spent many productive years with the 3205 Drone Squadron, and was involved in all of the guided weapons development from WWII to Vietnam, so I know the value of drones [UAVs]. Also know you can turn anything into a drone, as in the massed B-17 shots in the movie Twelve O’Clock High were of his unit’s QB-17 fleet being flown from a single “mother ship” in the formation. The crashes in the movie were QB-17s slated for destruction after being used during nuclear tests for air sampling.

They had QT-33s, QF-100s, used as targets, as well as purpose-built Q-2 Firebees. I’m a fan of drones, in their place, but ground controlled systems will not have the capabilities of a human at the scene for a very long time.

But seriously, nice writeup! Thanks m8. 🙂