Simply put, using phone-based bluetooth for contact tracing is way more complicated it appears it should be. This must-read article from MIT does a great job of pointing out more than a half dozen variables that influence the accuracy of using a phone to measure the distance between itself and another receiver or transmitter. https://www.technologyreview.com/2020/04/22/1000353/bluetooth-contact-tracing-needs-bigger-better-data/
However, in our opinion, the article leaves out one very critical detail, perhaps because it is a more technical detail than the average reader might care to consider, but I believe that it’s one that can be explained nonetheless.
For reasons obvious to anyone who’s ever noticed their phone battery dying more quickly than normal, phones are intentionally designed to optimize the output of power between different apps as well as different wireless connections to preserve battery life and prevent overheating. Power is assigned to each app differently each time depending on what other apps are open, how much battery life is available, ambient light factors and many other reasons. The different radios in the phone control their power output to balance the quality of a wireless connection with conserving battery life.
In order to allow for optimal performance between a phone and bluetooth devices, the output power can vary between -20 dBm (0.01 mW) and +20 dBm (100 mW). That’s a range of 40 dBm.
A 40dBm power output difference if sent to a contact tracing app could mean the difference of your phone thinking that it’s 500 ft away from another phone, receiver or transmitter vs. just 6 feet. That’s a critical problem.
Here’s how that works.
Distances between devices are calculated using a “Received Signal Strength Indicator” (RSSI). This is a measurement of how well a device can hear a signal from an transmitter. If you think of how we hear sounds, the further away a sound is, the lower the volume appears to us. Similarly the further away a received signal is, the lower a device indicates the strength to be. Received Signal Strength Indicator correlates to distance and can be calibrated so long as the power being used is consistent and known.
With that in mind, assume that you calibrate your RSSI system algorithms by placing a receiver 6 feet away using +20 dBm of power. Now imagine that same receiver, at the same 6 feet away receives a signal transmitted using -20dBm of power. With -40dBm difference in signal strength, that receiver that is physically 6 feet away would ‘think’ it’s nearly 500 feet away based on the RSSI calculation. RSSI diminishes exponentially meaning the distance would double 6.5 times (6 *(2^6.5)) or roughly 500 feet.
Clearly this is the extreme. Hopefully it clarifies why we felt it was important to explain.
Even if the difference between the assumed power output and actual power is as little as 6 dBm the calculated distance would be off by a factor of 2. The difference between 6 feet and 12 feet when it comes to contact tracing is not a small detail. Herein lies the challenge for using our phones to determine the distance between people and objects even if we ignore the other challenges mentioned in the referenced article.
This is not to say that other RSSI systems can’t do the job. It just means that purpose-built systems, calibrated at manufacture with fixed power are more predictable.
For example, RDCrease™ by Recon Dynamics is a purpose-built tracking and tracing system that is calibrated at manufacture based on a fixed power level. This system is accurate to inches.
Unlike Bluetooth on a worker’s phone, the challenges are no longer about the placement or orientation of the sensors and receivers because the system is helmet-mounted and only has one task. RDCrease tracks* the RSSI between people’s heads because that’s the most likely place for a contagion to be transmitted.
Phones are good for so many things. However, contact tracing at this point does not seem to be one of them.
* RDCrease™ by Recon Dynamics houses the tracking data for each company in their own separate private cloud server and and takes all industry standard precautions to ensure that personal or sensitive data is never shared or accessible by unauthorized parties.