Spy in the sky that sees round corners

by David Hambling
20 March 2010
WHY jump in a cab to “follow that car” when an airborne drone could do the job for you? The US Pentagon’s Defense Advanced Research Projects Agency is developing a radar system which sees around corners and down into “urban canyons”. DARPA hopes to be able to track vehicles across an entire city using just a few uncrewed aircraft.

Traditional radar relies on direct line of sight, so it’s tricky to track a vehicle that keeps nipping behind buildings. But DARPA believes that by using buildings as mirrors, it will be possible to identify a target vehicle from radar reflections. The experimental system is called Multipath Exploitation Radar.

By using buildings as mirrors, it will be possible to identify a target vehicle from radar reflections
The agency has been exploring how MER might work by driving vehicles around a simulated urban area and collecting returns from an overhead radar. Its researchers are aiming to combine the radar data with a three-dimensional map of the test environment to calculate how the radar reflects off and between vehicles and buildings. This process should highlight which signals in the returning radar data can be used to plot the target vehicle’s path.

MER is expected to be compatible with the radar systems currently used to track vehicles, a DARPA spokesman told New Scientist. The team anticipates that using reflected radar will cover more ground than a line-of-sight system, making it possible to monitor a city of about 1000 square kilometres, such as Baghdad, with just three airborne radars. The three-dimensional model of a city needed to make sense of the reflection pattern could be created using LIDAR, the optical surveying technology which is routinely carried on aircraft.

MER makes use of Ku-band radar – frequencies of between 12 and 18 gigahertz. It is sensitive enough to produce distinct signatures for apparently similar vehicles, by detecting slight differences, such as the angle of an aerial or a wing mirror.

DARPA is also looking to develop an algorithm which would enable the system to track multiple vehicles.

Ain Sume of the Swedish Defence Research Agency says the “sound, well-known physical principles” behind MER make it feasible. His team built a radar system that detects people around a corner by using reflections from the opposite wall.

But Sume reckons it will take some time to turn DARPA’s plans into a viable system. Key challenges include maintaining a radar lock as the view shifts from line-of-sight to reflection and back, and establishing a unique radar “fingerprint” for each vehicle.