DARPA shows off a crowd-sourcing radiation detector

The gadget, on show at this week's Wait, What conference in St. Louis, costs about $400 in volume -- significantly cheaper than existing detectors used by public safety agencies -- and provides a more accurate picture of any potential threats, said Vincent Tang, an applied physicist working for DARPA.

It works by sending a radiation reading to a companion smartphone over Bluetooth once every second. In St. Louis, attendees could borrow a phone and detector and try it out, searching for radioactive sources hidden around the room.

Walking along, most of the time the phone displayed a reading of around 20 counts per second of gamma radiation -- the background level in the event hall -- but as one of the hidden sources was approached, the reading started to rise rapidly.

At one point it hit more than 2,000 counts per second -- driven that high by a radioactive source (probably a rock) in a sealed container under a table.

All the time, the phone was sending the data back to a server that was recording the readings from the sensor and others being used at the conference. The result was a crowd-sourced heat map showing the levels of radiation in the event space.

In actual use, Tang envisions the sensors on things like public safety vehicles, constantly monitoring and reporting back levels. But that's only step one. What wasn't demonstrated in St. Louis, because the radiation sources were all safe, was that the signature of the radiation can be analyzed and its likely source determined.

So, someone fresh from a nuclear medicine procedure at a hospital would be recognized but ignored by the system, while someone carrying illicit nuclear materials would set off alarm bells.

It's the first product of a DARPA program called Sigma, which is attempting to revolutionize detection and deterrent capabilities to counter nuclear terrorism.

"The idea of Sigma is to try to leverage as much as possible existing commercial technology that could be applied to this area that would enable us to go from a low-volume, high-cost model to high-volume, low-cost model," Tang said.

Here's Tang's 5-minute presentation on the Sigma project and where it's going.