More from the article ...

... After enough trips, software may treat some of those devices as a recurring electronic signature associated with the vehicle. Weeks later, one of the same device signals appears alongside a different car connected to an investigation. The signal itself may not contain its owner's name, but its previous association with a known vehicle gives investigators another clue they can use to work out who was carrying the device.

SignalTrace, a system marketed by the security company Leonardo, is designed to work alongside automatic license plate readers. The company says it can recognize groups of consumer devices that regularly move together, then associate them with license plate records and time-stamped locations. The pattern can then be searched even when a police investigator does not know the plate number. ...

Leonardo's new explanatory sheet states that SignalTrace "does not identify people." It says the system "only collects electronic signatures" from signals already being broadcast, such as Bluetooth or radio frequency identification tags. Those signatures, by themselves, do not disclose a person's identity. The company says the output must be corroborated through ordinary investigative methods. ...

Leonardo's SignalTrace product page says the system stores electronic fingerprints for later queries and can recognize a vehicle without seeing its license plate. A separate product sheet says the technology helps identify suspects through the mix of devices they carry. The patent behind the system describes targets that may be people or vehicles. It also describes searchable signatures that can be correlated with visual identifiers and used to track a target across locations. ...

@#3

The issue I see here is the merging of Flock data with phone data.

That merged data creates a surveillance problem that is significantly larger than either of them used individually.