Wednesday, January 9, 2013

You can't hide from quantum radar


Radar is, broadly speaking, the standard way to recognize and identify incoming objects. Aircraft and ships usually broadcast a signal that identifies them anyway, but even in the absence of that signal, you still want to ensure that you accurately identify passing aircraft—and not by the wreckage they leave after you have shot them down.

This is also critical because the approaching aircraft could broadcast a signature that makes it look innocent when, in fact, it isn't. This form of sophisticated jamming would be very difficult to detect using a standard radar system. When you add the magic of quantum, however, life suddenly becomes a lot harder for the jammer.
The nice thing about an imaging radar system is that you can get the speed, direction, and the shape of the object from different aspects of the signal. The doppler shift on the radar signal gives you speed, the time between sending a pulse and receiving the scattered radiation at your detector gives you distance. The signal intensity from several detectors allows you to create an image. And repeated measurements tell you where the object is going as well as the speed again.
But, since radar is an active signal, it is possible to calculate a pattern of signals to send from the target to make it appear like something it's not. A plane equipped with the right transmitters could spoof its identity via radar. The ideal solution to this would be for the radar transmitter to tag each photon of microwave energy so that the it can verify that the radiation it detects comes from its transmitter. And this is exactly what a group from University of Rochester claims to have done (but hasn't).
What they noted is that a quantum key distribution system offers a way to analyze the statistics of detected photons and determine if they came from the intended source. The key to the security comes from the nature of a quantum measurement. In a classical measurement, you ask questions like "How long is that table?" and get an answer. In quantum mechanics, measurements don't work that way. Instead you ask the question "Is the table 1.60m long?" and the answer is either "yes" or "no."

The same holds for polarization. You place a polarization sensitive mirror in the path of the light. The mirror reflects horizontally polarized photons and transmits vertically polarized photons. This measurement can only report that the photons are vertically or horizontally polarized, even if they have an entirely different polarization, such as 45°. And, no matter what orientation the photon is, the detector will always provide a horizontal or vertical answer.
Now consider this from the perspective of someone who wants to jam a radar system. They want to detect the incoming photons and broadcast new ones so that they show incorrect radar information. But, if they're trying to match the polarization, they have to choosewhat measurement to make on the incoming photons. And, no matter what choice they make, they will always get an answer, even though it may be entirely wrong. As a consequence, their jamming signal will almost certainly contain enough photons with the wrong polarization to make it easily detectable.
Every photon transmitted by the radar unit is polarized either along the vertical/horizontal direction, or along the two diagonals (diagonal and anti-diagonal). This choice of which orientation is used is randomized for each photon. The jammer has a 50 percent chance of choosing the right polarization orientation. If they choose correctly, the photon that they send back will be indistinguishable from the one sent from the radar system.
Now, let's say the jammer chooses to measure diagonally when the sender is sending horizontally polarized photons. The photon will be detected as either diagonal and anti-diagonal with 50 percent probability. The jammer duly resends whichever polarization the detector told it to, with incorrect image information.
Meanwhile, the detector on the radar unit knows that the photon should be horizontal and is set up accordingly. The jammer has sent a photon that is actually in one of the diagonal states. Once again, even though it's diagonally polarized, the photon will be detected as either vertical or horizontal. Half the time, the photon is measured to be vertically oriented, which reveals the presence of the jammer.
The upshot is that 25 percent of the time, the radar unit receives a photon with the wrong polarization. (The jammer makes the wrong choice half the time, and the photon goes the wrong direction in the receiver's detector half the time—combined, they account for the 25 percent.) Such a high error rate is clearly observable compared to imaging without the jammer.
And, it works, of course. The researchers showed that they could tell the difference between jammed images and unjammed images very easily. And polarization is an ideal choice because it is not used for other aspects of radar, so the verification scheme does not get in the way of the radar measurements.
But we're apparently not ready for a jam-proof radar. The team did this with lasers rather than an actual radar system. Why? Well, they wanted a proof-of-principle* demonstration and they are an optics group. There is also the limitation that, although you know you are being jammed, you still end up with the image that the jammer projects. So, you still have no idea what is out there or where, but at least you know not to trust your radar image.
I am very skeptical that this will ever see the light of day outside of the lab. You would need a single photon source, and even with efficient single photon sources and detectors, there would be problems. On reflection from a 3D object, the polarization of the photon will be altered. What's more, that change will be different for every photon, since the object will be moving and changing orientation.
It is certainly possible that the changes are small enough that the error rate doesn't reach the 25 percent threshold—but if these reflections increase the error rate to 25 percent, then the real object would be indistinguishable from the jammer's projected object.
Still, the entire world of science is devoted to proving curmudgeons like me wrong and I look forward to that happening again.
*Proof-of-principle is science code for "we don't have a clue how to do this in any useful context, so we will do something useless and pretend that the rest is easy."

No comments:

Post a Comment