Quantum radar has been on the… ahem… radar for a while now. Unfortunately, the theoretical and practical results from our explorations of the concept have been underwhelming. But before we get to the disappointments, let me give all you radar enthusiasts a reason for hope. A new paper demonstrates that, under conditions of low signal-to-noise ratios (at the edge of the radar’s classical range), employing quantum technologies may offer a very significant boost in accuracy.
Quantum radar?
Radar, at its simplest, involves sending out pulses of radiation that reflect off an object. The reflected signal is detected, and the time of flight is measured. The time of flight is then translated into a range, while the direction that the radar antenna was pointed when it picked up the reflection tells us the direction.
The horrible thing about radar is that the signal drops off very rapidly—as the fourth power of the distance. This is because the power of the radiation we send out drops as the square of the distance between the transmitter and the object. And then it drops as the square of the range again after it’s reflected and has to travel back to the receiver. You get clobbered by the inverse square rule twice.
Let me make this concrete with a very rough estimate: a radar with a 1 kW transmitter and an antenna with a gain of 10 will need to be able to detect a few nW (10-9 W) of received power to see a 1 m2 object at 5 km.
A quantum radar makes use of quantum entanglement to increase the receiver’s sensitivity. To make quantum radar work, we no longer send all our photons out to look for objects. Instead, we only send one half of an entangled pair of photons out to reflect off objects; the other half is kept at the receiver. When the photon that is sent out returns, it matches its partner more perfectly than any single other photon that might be detected by the receiver. We can detect these matches, called correlations, with high sensitivity.
In terms of microwave engineering, think of it as better than the best possible narrow-band filter. In other words, a quantum radar doesn’t increase the absolute level of the signal, but it does raise your certainty in distinguishing signal from noise.
Wake me up when it gets interesting
On the face of it, this sounds exciting. Early calculations showed that entanglement should provide a factor of 2-4x improvement in certainty. Nice—but probably not worth the additional complications of working with entangled photons when it comes to practical applications. Even worse, the first experiments with quantum radar all used optical frequencies rather than microwave frequencies, and they operated over such short distances that the signal loss was tiny. Even on the brightest day, the noise at optical frequencies is orders of magnitude lower than it is for microwaves.
So practical applications, which would need to use microwave frequencies, imply huge losses. The snoring from uninterested radar engineers was deafening.
To make quantum radar interesting again, theoreticians dove deeper into radar theory and practice. It turns out that range accuracy (how good your average estimate of the range is) and range resolution (how confidently you can separate the range of two objects) are not entirely the best of bedfellows. Range accuracy happens to get really bad when the ratio of returned signal to background noise is below a certain threshold. And it is at this point that quantum entanglement can seemingly provide a big advantage.
Stretch that pulse out
To improve accuracy, you must stretch and chirp the pulse. Essentially, you sweep the radar frequency from high to low during the pulse (this type of pulse is also used in some classical radars). This stretches each photon out in time so that its frequency becomes much better defined. This also makes its entangled partner better defined so that they they can be jointly detected with greater certainty.
On the face of it, this reduces accuracy. An individual photon may be detected at any time over the entire pulse duration, which is now very long. But the microwave pulse consists of billions of photons per frequency, so there are lots and lots of individual photons to be detected. The statistical variation in their detection time narrows with the number of photons, allowing you to generate an accurate time of flight.
This really shows its power when the signal to noise levels drop below the classical threshold for accurate detection. When the signal is four times greater than the noise, the quantum radar is about 500 times more accurate than the classical radar (assuming the same transmitter power). Even when the signal-to-noise ratio is one (about when I would give up), the quantum radar remains three to four times more accurate than classical radar.
How stretched are your pulses?
The advantage of the quantum radar really depends on how much the pulse is stretched out. The researchers demonstrate this by calculating the quantum advantage of a W-band radar locating a small drone (a radar cross section of 1 cm2). At 100 m, the drone is detected by a 10 ms pulse from a quantum radar with about 60 times more accuracy than with a classical radar. But the window of utility is limited; when the drone is one km away, the same advantage is obtained only if the radar pulse is about two minutes long, by which time the drone has been and gone.
The bigger problem is, unfortunately, practicality. To make this work, high-powered sources of maximally entangled microwave photons are required. At present, the best entangled photon sources operate at optical frequencies and emit somewhere up to a million photons per second, corresponding to a power of about a fW (10-15 W). There are a few orders of magnitude between where we are now and where we need to be.
But, before you get too depressed, note that microwave sources are actually easier to build (and have a longer engineering history) than optical sources. And scientists have already demonstrated entangled microwave sources. So perhaps there is a future here…
Physical Review Letters, 2022, DOI: 10.1103/PhysRevLett.128.010501 (About DOIs)
https://arstechnica.com/?p=1824413