Tuesday, April 09, 2013

Historical data needs protection as cell-phone location tracking moves indoors

At recent hearings regarding whether law enforcement should have to get a warrant for cell-phone location tracking, police from Dallas and Houston as well as an ostensibly (if not actually) neutral witness from the US Marshal's Service claimed that historical location data should not require a warrant because it is relatively inaccurate. At the House panel in particular that suggestion was rebuffed by the 2009 example of Malte Spitz, a German politician who sued for access to his cell-phone location data and collaborated with a newspaper to generate this graphic tracking his every move for six months. In general, though, it turns out the notion that historical location data is inaccurate was an ill-timed argument. In the past month, a flurry of new developments have shown location data is rapidly becoming more accurate than ever, particularly a new study showing cell-phone users can be identified via location data using relatively few "pings."

Now, this week's IT news features discussion of emerging "indoor location tracking" by smart phones that in the near future will provide ever-more precise location data on anyone carrying one:
Claburn's piece is the only critical item of the bunch, pointing out that, "this technology isn't really for consumers. It's for businesses, so they can track you and glean more data about what you're looking at, where you're going, how long you spend in a place and so on." He could have added that it's also for law enforcement, which makes thousands of requests for cell-phone location data every day - more every year as the technology becomes more ubiquitous and precise.

Most discussions of location tracking at the Texas Legislature have been about either GPS, which doesn't work indoors or when line of sight to a satellite is blocked, or cell-phone tower triangulation, which generally does work indoors but in some circumstances is less accurate. Increasingly, though, both those methods are being enhanced by an array of technologies that collaborate to generate extremely accurate tracking data. Find below the jump an expurgated version of a startling list compiled by Mr. Dodge of some of the means being used to make location tracking data more accurate in the next few years, even indoors.
  • Wifi Triangulation – Wifi Triangulation measures signal loss or strength from multiple wifi hotspots to triangulate position. It is not necessary to connect to these wifi hotspots, only to measure the signal strength. Your phone displays signal strength in terms of  3 or 4 bars, but inside it is actually measuring signal strength very precisely. These services have a database of known wifi hotspots, and adds new hotspots as they are discovered by users. ...
  • GPS/Cellular/Wifi Triangulation – Uses inputs from GPS/Cellular/Wifi, when available, to determine position. This is important for smooth transition from outdoor to indoor positioning. ...  
  • Wifi Fingerprinting – Smartphones turn on wifi for a few seconds to get a Wifi Fingerprint and associate it with a Check-In location. Compares the current Wifi Fingerprint to a known database of Fingerprint/Location pairs. Often used in conjunction with Check-in services like Google Places or FourSquare. This allows a more accurate location within a building....
  • Dedicated Beacons - Cheap, low power, radio beacons located at known positions within a building. The only purpose of the beacons is to transmit a unique signal that can be received by your Smartphone. Uses the same location triangulation methods as wifi, but can be more accurate due to their specific location and purpose. ...
  • Bluetooth Sensors - Many electronic devices contain Bluetooth, including every smartphone. These Bluetooth sensors can read signals from dedicated beacons, or dynamically create a mesh network of Bluetooth signals that constantly corrects and refines relative position and location.
  • Tracking Sensors from known positions - Most smartphones contain multiple sensors including a compass, gyroscope, accelerometer, altimeter, and barometer. These sensors can measure your direction, turns, speed, and height above sea level to create a three dimensional view of your location. Starting with a known position from other methods such as GPS, cellular, or wifi which work outside, the smartphone sensors can be used to track your position inside a building.
  • Magnetic sensors - Magnetic sensors can pick up the Earth’s natural magnetic forces to determine lat/long position similar to the way a compass works, but two dimensional, and much more accurate.
  • LED Lights - lights in the ceiling can be programmed to pulse in milli-seconds, so fast the human eye can’t detect the pulse. But, your smartphone camera can detect the pulses and distinguish between different lights and triangulate your position. The LED lights each have a unique pulse fingerprint. They can be used with standard light fixtures and remain in fixed positions within the building, making it easy to calculate location.
  • Cameras - A ceiling or wall mounted camera within a building can cover up to 100 square meters. The camera on your smartphone can automatically take many snapshot photos per second. Object recognition software uses pattern matching to compare those smartphone snapshots to the wall-mounted camera to determine precise location. 
One could already see from Malte Spitz's example in 2009 that historical location data from cell phones had become shockingly accurate. And the trajectory is only headed in one direction: In the near future, smart-phone location tracking will be incredibly precise, no matter where you are. In that protean environment, it makes no sense for legislators or the courts to distinguish between real-time and historical tracking data as though the latter is somehow less invasive. Thanks to advancing technology, that distinction hasn't really been true in several years and, in the near future, as smart phones evolve, it will become flat-out meaningless, if it hasn't already.

No comments: