American Patriot anti-aircraft missiles shot down two coalition aircraft in Iraq last year, and almost brought down a third. The problem was blamed on software that was unable to distinguish between incoming enemy missiles and friendly aircraft returning from a mission. The friendly aircraft were equipped with IFF (Identify, Friend or Foe) electronic devices that send out a signal to warn friendly anti-aircraft systems not to fire. The IFF did not work, and the Patriots were set to fire automatically (as incoming missiles would be moving too fast for a human operator to make the call.) 

This was not the first time a complex electronic system worked in unexpected ways in combat. Electrical engineers have known, for nearly a century, that electronic signals can, in unexpected combinations, produce unpredictable results. This has become a growing problem in the last two decades as more radars and electronic warfare systems are put into service. Anti-missile systems, like the American Patriot, are particularly dangerous because, to be effective against missiles, the system has to be on "automatic." This is a known danger, because when you go to war, you bring together a lot of friendly, and enemy, electronic gear that has never been used in close proximity before. For example, during the 1991 Gulf War, it was found that Patriots electronic systems interacted in unexpected ways when certain jamming systems were used nearby. In one instance, American aircraft flying north, over Patriot batteries, found that when their jamming systems were tested (a standard procedure before entering hostile air space), those signals would sometimes trigger a nearby Patriot battery to automatically identify the passing aircraft and prepare to fire on it. The Patriot operators were able to stop the firing sequence before missiles were fired, and pilots had to be warned not to test their electronic gear while flying near a Patriot battery.

But there's still another problem that is not much talked about by pundits. Basically, all these electronic devices, be they radars, IFF, electronic warfare gear or whatever, do not operate in a world of black and white, but rather shades of gray. Radars, for example, are subject to all manner of interference and lowered performance. In many cases, especially in combat, what you see on your radar screen is what might be there, not what is 100 percent-definitely there. This is why commercial aircraft carry a powerful electronic beacon ("transponder") that continuously broadcasts the aircrafts location and identity. This is a lot safer than depending on radar alone. But in combat, conditions are always less than ideal. Take counterbattery radar, for example. These radars spot incoming shells and rockets, and plot where they were fired from (so your artillery battery can counter that fire, thus "counterbattery"). These radars have a lot of false positives, despite decades of tweaking the software and hardware. But it's better to have some false positives than to allow shells to get in without being properly tracked and the source of the fire identified. False positives by counterbattery radars rarely result in fatalities, but false signals from anti-missile systems are another matter, as was discovered in Iraq last year. Such problems will persist, because as quickly as solutions are found, new electronic devices appear, usually only when you've gone to war. Getting various services and nations to bring their electronic systems together periodically just to test for these fatal glitches hasn't even gotten to the proposal stage. It's too expensive, and the number of lives actually lost doesn't justify the cost when there are so many more demanding projects that need the money (like better medical gear, or better weapons.)

So expect more of the electronic unexpected. What happened in 1991 and 2003 were not exceptions, they were the inevitable.