Author Topic: Bluetooth Low Energy is unsuitable for COVID-19 contact tracing, say inventors  (Read 9132 times)

0 Members and 1 Guest are viewing this topic.

Offline SparkyFX

  • Frequent Contributor
  • **
  • Posts: 676
  • Country: de
We're not epidemiologists and I doubt the developers are either, so we can't form reliable opinions on how well or not it will work in real life. It's definitely worth a try. Any intervention in disease control has costs and risks, which is why extensive in-field testing is required to determine whether they're worth the benefits.
That's for sure, it could add to the set of tools for contact tracing, but imho never replace it, because the false positives might be so widespread that you immediatly go to lockdown anyway or the threshold is set so high that too many cases are under the radar and you get an outbreak.

It surely helps in cases like where you can't name the person you sat next to in public transport, trying to identify all riders is next to impossible.
Support your local planet.
 

Online tom66

  • Super Contributor
  • ***
  • Posts: 7331
  • Country: gb
  • Electronics Hobbyist & FPGA/Embedded Systems EE
Well, if you would combine BLE with GPS and the gyrometer/inertial measurement unit you could get rather accurate histograms of a persons position. If another device gets registered via BLE you wait until the signal drops out and then compare the (more or less) exact position of both persons over the time they were in proximity. Should yield way better results (you basically eliminate the RSSI/walls issue), but you would have to share your movement/exact position with everyone in your proximity-> won't happen as it would be a data security/privacy nightmare.

GPS doesn't work indoors, underground, or even sometimes if the device is in a less than favourable orientation.   Plus, there are considerable privacy implications: you are now giving the government full positional data on all citizens. Will they give that up? 

The present system using BLE assumes that infection occurs if you spend more than X minutes around a given individual.  For instance, while queueing at a shop.  While not perfect, it reduces a lot of concerns around privacy because it can be based on entirely anonymous keys.
 

Offline Syntax Error

  • Frequent Contributor
  • **
  • Posts: 584
  • Country: gb
Here in the UK we are recruiting 14,000 contact tracers. So far, a week before the BLE app launches, we have about 900 - mostly in managerial roles. From a real world job advert, this is the kind of plug and play team player that they are seeking to employ part-time...

Someone who has... Experience of working in telephone or face to face customer services. Is passionate about exceeding customers expectations and delivering an outstanding customer experience. A great communicator, both written and verbal, who wakes up every day wanting to exceed client and customer expectations. Strong desire to learn, take on board feedback, strive to improve and work to personal goals as well as part of a driven, energetic team!

And also knows f'all about Covid-19.
 

Online Marco

  • Super Contributor
  • ***
  • Posts: 7043
  • Country: nl
The present system using BLE assumes that infection occurs if you spend more than X minutes around a given individual.  For instance, while queueing at a shop.  While not perfect, it reduces a lot of concerns around privacy because it can be based on entirely anonymous keys.

Assuming it works, false positive/negative rates in complex environments with lots of steel around (ie. supermarkets) are an open question AFAICS.
 

Offline Buriedcode

  • Super Contributor
  • ***
  • Posts: 1718
  • Country: gb
The present system using BLE assumes that infection occurs if you spend more than X minutes around a given individual.  For instance, while queueing at a shop.  While not perfect, it reduces a lot of concerns around privacy because it can be based on entirely anonymous keys.

Assuming it works, false positive/negative rates in complex environments with lots of steel around (ie. supermarkets) are an open question AFAICS.

Which is why several posters here, as well as the devs themselves have commented just that.  It's less than ideal, will require some testing - or rather testing it in the wild before making any kind of major decisions based on its results.  There is the danger of the government relying too much on this, but it seems the UK hasn't done too badly in their response in terms of listening to their experts, so I can only assume/hope that the results will be taken wit a pinch of salt, at least until its relative efficacy is known. 

I get the impression that some here believe such apps are seen by others as techno salvation, possibly based on the sensationalist media, but I don't know anyone who believes such an app will be a magic bullet.
« Last Edit: May 19, 2020, 08:32:21 pm by Buriedcode »
 

Online nctnico

  • Super Contributor
  • ***
  • Posts: 28429
  • Country: nl
    • NCT Developments
The Dutch government has become awfully quiet about a tracking app. The new mantra is testing testing testing and manual contact tracing.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Online Zero999

  • Super Contributor
  • ***
  • Posts: 20355
  • Country: gb
  • 0999
Where its accurate, the risk of infection is almost zero.

Where its inaccuate, or unavailable, the risk of infection is high. That doesn't make sense to me, does it to you?
It depends on how it's inaccurate. Yes, if it doesn't generate enough positives, when it should, then it will allow the infection to spread too much and be useless, but if it errs on the side of caution and generates more positives, than necessary, then it will reduce the rate of infection. If everyone on the same train carriage goes into self-isolation for a week, because one person had it, then any of them who have caught it won't infect anyone else and it would have done its job.

And you've completely missed the point that just because GPS is unavailable in some buildings it doesn't make it useless. If some people are at a bus stop, then disappear off the radar, only to reappear somewhere else, a few miles down the road, they've obviously been on the bus together and if one of them is infected, then the others should self-isolate for a week. It doesn't take much imagination to see how this could be effective for tracking people to other indoor spaces.

We're not epidemiologists and I doubt the developers are either, so we can't form reliable opinions on how well or not it will work in real life. It's definitely worth a try. Any intervention in disease control has costs and risks, which is why extensive in-field testing is required to determine whether they're worth the benefits.
That's for sure, it could add to the set of tools for contact tracing, but imho never replace it, because the false positives might be so widespread that you immediatly go to lockdown anyway or the threshold is set so high that too many cases are under the radar and you get an outbreak.

It surely helps in cases like where you can't name the person you sat next to in public transport, trying to identify all riders is next to impossible.
I agree, there needs to be some manual intervention for it to work well. Even if GPS isn't used to track the person continuously, position data could be stored on the users phone and only uploaded when they test positive, with their consent and the condition that it's deleted within a week. For maximum transparency it could be in plain text format. Heck GPS might even be useful if it's only stored locally and never uploaded. It can help the user who's had a hit decide whether they need to self-isolate and get tested or not.

Testing will help to reduce the number of people self-isolating due to false positives, but simply asking someone questions can cut down on the number of wasted tests. For example, I cycle to work and being outdoors it's unlikely I'll infect anyone, or pick up it up off anyone else, so if my phone were to get a hit during my journey to work, there's little point in being tested or self-isolating. If the person who's infected is on a bus which overtakes me and I get the hit, then it's impossible for them to infect me, but possible other people on the bus could have caught it. All it would take is for someone to ask me if I ride the bus at that time of day to eliminate me.
 

Online tom66

  • Super Contributor
  • ***
  • Posts: 7331
  • Country: gb
  • Electronics Hobbyist & FPGA/Embedded Systems EE
Well, the good news is the contract tracers will all know each other, at least in the UK.

https://www.bbc.co.uk/news/uk-52732818

Seriously, do people still not understand Bcc vs Cc?
 

Offline Syntax Error

  • Frequent Contributor
  • **
  • Posts: 584
  • Country: gb
Here in the UK, public sector service provider Serco does not exactly enjoy the best reputation when it comes to IT security. As one of the UK goverment's go-to cheapest is best providers, it's more 'Ryan Air' than 'Emirates'. So employing a 'high flyer' who doesn't know how to send an email, is no real surprise. Knowing that Serco is in the supply chain for UK contact tracing leaves me to consider, should the app be avoided?
 

Online nctnico

  • Super Contributor
  • ***
  • Posts: 28429
  • Country: nl
    • NCT Developments
It depends on how it's inaccurate. Yes, if it doesn't generate enough positives, when it should, then it will allow the infection to spread too much and be useless, but if it errs on the side of caution and generates more positives, than necessary, then it will reduce the rate of infection. If everyone on the same train carriage goes into self-isolation for a week, because one person had it, then any of them who have caught it won't infect anyone else and it would have done its job.
GPS is useless for tracking. The inaccuracy in a city can easely be +/-80m (if not more with high buildings around). A much better option is to use triangulation between cell towers.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Online Zero999

  • Super Contributor
  • ***
  • Posts: 20355
  • Country: gb
  • 0999
It depends on how it's inaccurate. Yes, if it doesn't generate enough positives, when it should, then it will allow the infection to spread too much and be useless, but if it errs on the side of caution and generates more positives, than necessary, then it will reduce the rate of infection. If everyone on the same train carriage goes into self-isolation for a week, because one person had it, then any of them who have caught it won't infect anyone else and it would have done its job.
GPS is useless for tracking. The inaccuracy in a city can easely be +/-80m (if not more with high buildings around). A much better option is to use triangulation between cell towers.
That must be a worst case scenario. I've found satellite navigation systems to be much more accurate than that and we don't need 2m accuracy, just whether two people are in the same building or not will do. It can be combined with other data such as maps to narrow it down further. If triangulation is better, then use that too.

Irrespective, combining it with manual contract tracing will help a lot. Even 1 km accuracy would be better than nothing at all.

Well, the good news is the contract tracers will all know each other, at least in the UK.

https://www.bbc.co.uk/news/uk-52732818

Seriously, do people still not understand Bcc vs Cc?
I can't believe anyone would be that dumb. :palm:
 

Offline cdevTopic starter

  • Super Contributor
  • ***
  • !
  • Posts: 7350
  • Country: 00
City, as in big city, and country driving are two totally different situations..

They are testing the situation on the Isle of Wright. Thats a location that seems likely to me to have almost an ideal situation for GPS. Probably very few if any tall buildings or underground areas, and even the largest buildings seem likely to be as large as the smallest buildings in most big cities.

Many people who dont live in big cities never go driving in them. So their own GPS which looks and is accurate enough to spot lane changes under local use seems accurate enough to them.

But its an entirely different situation in cities with lots of tall steel buildings, (urban canyons) or inside a network of streets where traffic often descends underground to get from point A to point B.

Newer phones can and sometimes are more accurate but there are hard limits to the coverage -

Outdoors, coverage is fairly good, but accuracy declines greatly in dense cities, exactly where people are forced more closely together.  (Risk of infection is by all accounts very low outdoors)


1. whether a fix is available or not largely comes down to whether somebody is outdoors or indoors.

2. in the places where people would be most likely to pick up infections, (indoors) it is basically guaranteed

to either be unavailable or very very inaccurate and unpredictable.

Nonetheless, if this system was available, it would invariably be used more than its accuracy would dictate it should be.

Does anybody remember the report that came out several years ago from the UK government on the safety of cell phones? That was a good example, of a badly done investigation of a subject that clearly had only one decision in mind and was determined to block off any adult discussion of the unknowns.


It depends on how it's inaccurate. Yes, if it doesn't generate enough positives, when it should, then it will allow the infection to spread too much and be useless, but if it errs on the side of caution and generates more positives, than necessary, then it will reduce the rate of infection. If everyone on the same train carriage goes into self-isolation for a week, because one person had it, then any of them who have caught it won't infect anyone else and it would have done its job.
GPS is useless for tracking. The inaccuracy in a city can easely be +/-80m (if not more with high buildings around). A much better option is to use triangulation between cell towers.

Thats what TruePosition used to do but a few years ago they were pushed out by some other folks and I dont know if it emerged why or not.

« Last Edit: May 20, 2020, 05:57:43 pm by cdev »
"What the large print giveth, the small print taketh away."
 

Online nctnico

  • Super Contributor
  • ***
  • Posts: 28429
  • Country: nl
    • NCT Developments
It depends on how it's inaccurate. Yes, if it doesn't generate enough positives, when it should, then it will allow the infection to spread too much and be useless, but if it errs on the side of caution and generates more positives, than necessary, then it will reduce the rate of infection. If everyone on the same train carriage goes into self-isolation for a week, because one person had it, then any of them who have caught it won't infect anyone else and it would have done its job.
GPS is useless for tracking. The inaccuracy in a city can easely be +/-80m (if not more with high buildings around). A much better option is to use triangulation between cell towers.
That must be a worst case scenario. I've found satellite navigation systems to be much more accurate than that and we don't need 2m accuracy, just whether two people are in the same building or not will do. It can be combined with other data such as maps to narrow it down further. If triangulation is better, then use that too.
Navigation systems rely heavily on mapping your movements on a map and only use GPS as a hint to where you are. If you look at the raw GPS data you'll see it can be total garbage in urban areas. Just take a different turn then your navigation system is pointing to. You'll see it will have trouble catching up.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf