Two problems with conventional radar make it unsuitable for many applications: Anyone with a radar receiver can tell when you activate it, and it can't image objects closer than about 100 feet. Granted, radar automatically opens the door for you at the grocery store, and Stealth bombers are supposedly transparent to radar. But the grocery store radar uses a Doppler algorithm that can only sense movement, not make images, and an aircraft can only be made invisible to radar directed at it from the ground.
Now Ohio State University electrical engineer Eric Walton claims to solve both problems with $100 worth of parts. Walton's "noise radar" hides its signal in wideband noise, making it undetectable by the enemy, and it can image objects right through concrete walls.
"Noise radar has been known as a laboratory curiosity since the 1950s, but it's only been in the last few years that commercial chips have gotten to the gigahertz frequencies that make it practical. Early prototypes were the size of refrigerators," said Walton, who received a patent on noise radar just this year.
"Today, using fairly commonplace components, you can build a noise radar for under $100," Walton said. "All you need is a high-speed shift register for random noise, or a FIFO [first-in first-out shift register] if you want to load pseudorandom waveforms. The receiver is just a microwave mixer and a low-pass filter with a cutoff of 10 to 100 kHz."
By spreading low-level noise across gigahertz of radio spectrum, the noise radar signal becomes undetectable to normal radar receivers, which are designed to look for high-level signals and to filter out weak signals assumed to be noise. Spread-spectrum transmitters and receivers are widely used today, but Walton claims his patented approach is unique. "The typical spread-spectrum radio system in cell phones has only a few megahertz of bandwidth," he said, "but our noise radar has a few gigahertz of bandwidth--a factor of 1,000 difference."
According to Walton, spread-spectrum receivers cannot decode noise radar signals, because the signals are spread across gigahertz of bandwidth--simply too much territory for the receivers to cover when searching for correlations. In fact, even two of Walton's own noise radar transceivers, sitting side by side, cannot detect each other, because the signals they send out are random and unique.
The only receiver that can detect the signal from a noise radar, Walton said, is the very device that sent out the signal in the first place, its unique code being the random signal itself.
"We are not sending a carrier or anything like that; it's just very, very wide-bandwidth random noise, which the universe is filled with already, thus making it impossible for anybody else to receive or even detect it,"said Walton.
A copy of the random signal is kept for use in a cross-correlation algorithm between what is sent and what is transmitted, Walton added. Thus, each noise radar device inherently incorporates a random key that is unique to that device and that changes randomly as it is used.
Because random noise generators constitute the radar signal, the signal reflections, which can be used to image objects, are also unique. Any number of noise radars can operate side by side with absolutely no statistical chance that any two will interfere with each other, according to Walton.
"You could imagine General Motors deciding to put these on the backs of cars to warn of obstacles," he said. "You could have a thousand people all backing up at the same time after the Super Bowl, and none of their radars would interfere" with their neighbors' radar systems.
Whereas conventional radar sends out a signal and then switches off the transmitter to listen for the reflection, noise radar constantly transmits and receives. Because transmission and reception occur simultaneously, noise radar can "see" through walls to detect objects inside, even if the objects in the room are only inches away. That could prove a boon for search-and-rescue missions.
Conventional radar is not very useful for finding survivors after an earthquake, for example, because it takes too long to switch off the transmitter and begin analysis; reflections from the nearby objects are already past the receiver before it is switched over. But with noise radar, reflections from objects only inches away can be used for imaging.
"Even with just a single gigahertz of bandwidth--say, between dc and 1 GHz--noise radar penetrates walls and foliage and down into the ground. You can see the reinforcement bars in a concrete road bed or a land mine buried under the ground, or look at the heart valve in somebody's chest,"said Walton. "For long-range imaging, you can give your antenna some more gain by upconverting your signal to 10 GHz, so you have a gigahertz of bandwidth in the X band [8 to 12 GHz]."
By adjusting the center frequency and the statistical composition of the gigahertz-wide spread-spectrum signal, different applications can be addressed. Military applications, for example, could analyze reflections not only for the size and shape of objects imaged, but also for the objects' composition and function. To do that, a pseudorandom noise consisting of 1,000 bits or more would be loaded into a FIFO and used as the signal source. Users could then create a catalog of what reflections of certain objects look like when transmitting a certain sequence from the FIFO.
"There are all sorts of signals that can be transmitted, each with statistically desirable properties, such as maximal-length sequences--pseudoran- dom noise that produces a set of orthogonal frequencies whose reflections are easier to recognize," said Walton.
"For instance, the military could compile a catalog of reflection patterns for certain maximal-length sequences, since when a waveform returns after being reflected off an object it's predictably different," said Walton. "A catalog of those differences could be used as a matched-filter radar, which only responds to particular types of targets. For instance, it could tell the difference between a Scud missile and a school bus."
Walton claims to have several license deals in the works, including one with a major aerospace manufacturer. His work was funded by the ElectroScience Laboratory at Ohio State.