The problem, hardware engineers warn, is the integrity of the data on which the system would be based. Smartphones have four sources of data to help determine their location, and all have accuracy problems.
The civilian version of GPS only gives a four-metre resolution and is easily interfered with by buildings. Smartphones also use databases of WiFi routers to work out where they are, but their accuracy is even worse. FON, which maintains a global database of hotspots, only claims accuracy of around 20 metres. Data from connecting to mobile masts is also less accurate than GPS, and even worse in the countryside.
NHSX has chosen to focus instead on the final source: the short-range wireless technology Bluetooth, which a panoply of digital devices use to talk to each other. On paper Bluetooth is promising. Fixed Bluetooth “beacons” in shops are already advanced enough to provide accurate location data used by retailers to track how customers move through the aisles.
However, such retail beacons don’t themselves move. Using a Bluetooth link between two moving targets is a big challenge.
“You can’t measure proximity accurately with Bluetooth – that’s the thing,” says Ted Wood, an experienced engineer who most recently headed up Bluetooth at Kano, which makes computers for learning coding.
Cambridge Professor Ross Anderson suggests how false positives threaten to undermine the data too. The passengers of two buses next to each other at traffic lights might appear to have been in contact as Bluetooth would not detect the separation between them. A lot of people could be given a bogus instruction to isolate.