Template 67

Defending Australia

Jennie Gibson

Defence scientist, Defence Science and Technology Organisation, Edinburgh, South Australia

(1) C130 Hercules dispensing flares by EDWINA BURGESS KSB by H.Ward

(2) ASRAAM (Advanced Short Range Air-to-Air Missile) on an ARDU (Aircraft Research & Development Unit (RAAF)) FA18 ©RAAF, ARDU and 92 Wing Photographic Section

Jennie has a PhD in low energy electron scattering physics from the Australian National University and an honours degree from the same institution. It took about 17 years from the first year of the degree (1982) to the award of her PhD (1999).

In that time I worked off and on as a clerk with the (then) Department of Aviation (where I met my husband) while I finished my first degree. I find working full time and studying part time impossibly hard and admire those who can manage it. I worked (full time) one year on/ one year off for four years instead. I then joined the patent office and stayed for five years. When I thought that one more optical fibre cable winding method patent application would drive me insane I looked up my honours supervisor and asked him for advice. He introduced me to my PhD supervisor and I left the patent office for life as a student again, this time on a scholarship so I didn’t need to try it one year on/one year off. When I started looking for a new position toward the end of my PhD I thought for a while I’d educated myself out of the workforce. I felt I couldn’t take the post-doc route so I was looking for a permanent position right from the start. Fortunately, Defence Science and Technology Organisation (DSTO) offered me a position. It did mean a move to Adelaide which required some ‘life-style’ changes. Five years on Adelaide still has great cloud formations.

What do I do here? I’m a defence scientist at DSTO in Edinburgh, South Australia. For the past five years I’ve worked in Weapons Systems Analysis (WSA). We work to so-called “understand and model the behaviour of weapons systems to provide the defence forces with the best advice on how to use the weapon to its fullest advantage”.. Yes, I’ve been infected with the ‘management’ language virus. I’ve also read Don Watson’s book (Death Sentence¹) and am trying to fight the good fight against that language so let me re-phrase that last bit. We try to work out exactly how weapon systems work and how best to use them. We (WSA) largely do this by computer modelling and simulating and bench testing where possible.

On first look this ‘system analysis’ may seem pretty simple – aim and fire sort of thing - but it turns out that weapons systems are sometimes extremely complicated and sophisticated. For example the seeker is the subsystem which sends homing information to the missile guidance system. Seekers can be visual or infra red detectors or have radar or GPS antennas. Sometimes they have more than one sensor system. Some of the early infra red air-to-air missiles had simple infra red detectors which just followed the hottest spot they could find, whether it was an aircraft tail pipe, the sun or a counter measure like a flare. Many of the newest missiles on the market have IIR (Imaging Infra Red) seekers. If you have delved into the murky world of image processing algorithms you can imagine the myriad combinations of algorithms you could choose to detect, identify and recognise a target against background clutter. Trying to determine the best set of algorithms for air-to-air, air-to-surface or surface-to-surface operation for example is a major research programme of itself within the Weapons Systems Division. Add to that the fact that new generation missiles are ‘software the field to take advantage of the actual conditions and desired strategy – and you can see how complicated one single component of a missile, the seeker, can be. WSD has research programmes in all aspects of missile systems: motors, fuses, warheads, guidance, electro-optic and radar seekers.

I can imagine every physicist and engineer, not to mention the other experimentalists, asking why modelling and simulation – why not cut the thing open and see how it works? That’s the question I asked when I first arrived for work: ‘Where’s the data base?’ The response went something along the lines of: ‘Data base? You’ll probably never even see a missile, let alone see one fired or fire enough to produce a data base’. Missiles are incredibly expensive, getting more so and we couldn’t fire enough to test every possible scenario anyway. A 1998 report of the US General Accounting
Office to a House of Representatives sub-committee quoted the then unit price for an AIM9x (a new generation imaging infra red missile) as approximately
$US 264,000². Targets are also very expensive (think about a target sophisticated enough to mimic an aircraft to an IIR seeker and which can’t have a pilot on board), for example the easiest publicly available price tag I can find for the Kalkara unmanned aerial target system is $750 000 from a senate committee
of May 2000³. So we fall back on the tool kit of the physicist: write down the equations which describe the system and solve them for the cases you want to test, i.e. we build computer models of missiles and ‘fly’ them in simulations.

The next question might be ‘The Department of Defence buys these systems from manufacturers. Why don’t you (they) demand they (the manufacturers) provide full information on how the systems
work?’ Well, they usually do - after all we wouldn’t buy them if we didn’t know how they work, but that doesn’t mean they tell us exactly what happens within each subsystem e.g. seeker or guidance algorithms. Weapons manufacturing is a very competitive business and companies do not like handing out proprietary information to anyone, even potential customers. Many manufacturers specifically deny access to that information.
So it’s up to our military weapons specialists and DSTO’s defence scientists to analyse the system, understand it, show how best to use it and/or recommend its purchase (or not!).

In the last few years I’ve been involved in the development of a new approach to weapon modelling in WSA. Some years ago the senior members of the unit realised that the old ways of computer modelling were not going to serve us in the future. What used to happen was a modelling expert wrote a model in a linear fashion in a programming language like FORTRAN (or C or PASCAL or whatever their favourite language was. As far as I know no-one ever used COBOL for a missile model). They talked to the subsystem experts then interpreted that information into the program. This usually meant one modelling expert for each system. When that person left or retired there was no whole of system expert to provide advice. There’s also the typical problem of reading someone else’s computer code. I don’t know about you but if I don’t comment my code when I write it I can’t read it myself three months later. The only way I can read someone else’s code is on paper with a pencil to trace variables through the code to figure out exactly what it does. This is the exact problem WSA had – how to use the codes of people who’ve left or codes bought in from overseas or codes from manufacturers when available?

One way we’re trying to make code more re-useable is a system originally developed in response to this problem by the Munitions Directorate of the US Air Force Research Laboratory at Eglin, Florida. The system is called MSTARS (Munition Simulation Tools And ResourceS)
and relies on the new visual languages like VisSim© and Simulink©⁴. These languages have proven popular with engineers and scientists and lend themselves to functional break-down of missile systems (and almost any other control based system). The key to re-use of code is to require modellers to write to the rules (functional decomposition and interface specification). We have the system linked to other architectures (eg our in-house C++ architecture) and all the components can be kept in libraries to allow plug and play interchange of components for research, development or analysis. The system includes telemetry and visualisation tools. Yes, it’s very hard to get people to follow rules in programming, so we don’t try to regulate how they write their component models, we specify the architecture requirements. The lure of libraries of re-useable components and analysis and visualisation tools has given us a small user base which we hope will expand.

We get re-use by requiring interface specifications; we get expert input at the code level, rather than filtered through a programmer. A scientist can directly test a change to the code of a subsystem rather than tracking down someone who can read the missile model code and make the change for them. We train the scientists and engineers in the architecture of the system.

I have actually seen telemetry missiles attached
to an FA18 Hornet but the second part of that response I mentioned earlier is still true. I’ve never seen a missile fired (except on TV). I hope our defence forces never have to fire one of those missiles in anger.

In mid-2004 I moved across the corridor to a new division, Electronic Warfare and Radar Division to work with ways of defeating missiles.

The mention of Don Watson’s book? I really meant it. I think it should be compulsory reading for everyone. I realise it will never catch on with the people who use obfuscating language deliberately, but for those of us befuddled and/or infuriated by every new management speak trend at least we know we’re not alone.

¹Watson, Don Death Sentence: The Decay of Public Language, Random House Australia, 2003.

²United States General Accounting Office: GAO Report to the Chairman, Subcommittee on Military Research and Development Committee on National Security, House of Representatives April 1998 AIM-9X ACQUISITION.

http://www.gao.gov/archive/1998/ns98045.pdf

³Kalkara price from Official Committee Hansard SENATE FOREIGN AFFAIRS, DEFENCE AND TRADE LEGISLATION COMMITTEE Consideration of Budget Estimates TUESDAY, 30 MAY 2000

http://www.aph.gov.au/hansard/senate/commttee/s1058.pdf

| Issue 67 Contents |