Systems for observing vehicle occupancy in realtime

[squizzler]

If the NHS app gets introduced it is hoped that over 60% of people will use it, as that is the proportion required to make it effective. If all busses and train cars had the necessary bluetooth transceivers and were enabled to talk to the app, it would provide benefits:

• There is a hard and fast record of what vehicle a person who subsequently tests positive has travelled on, providing another means of identifying contacts, beyond direct person-to-person swapping of ID through people's phones.
• The bluetooth system could, presumably regardless of any app compatibility, count the number of live devices (and thus people) onboard at any one time. In conjunction with 'open data' and a suitable choice of third party apps, prospective passengers would be able to see how many people are onboard those services they are likely to use.

I feel the second one would provide the most benefit for people and perhaps be reason to get the NHS app in the first place. A technical system would also indicate that we are taking a 'can-do' attitude to the issue of making transport networks safe, and encourage people to make journeys by routes and times that are less busy.

 

[Domh245]

The NHS app needs to be able to do what it does well, and nothing else. It just needs to tell what other devices you've been in prolonged contact with, complicating it by trying to also make it some sort of public transport loading system just means it'll suffer more delay before release (and generally perform worse) - indeed, if you make it do that for public transport then you'd have to do it for supermarkets and other places where you are far more likely to have people in contact, it just won't work

That isn't to say that the idea of showing what services are lightly/heavily loaded isn't a bad one. Where trains are fitted with live passenger count technology (which is quite a lot now IIRC) then ideally TOCs could show that live loading data in their own apps.

 

[squizzler]

complicating it by trying to also make it some sort of public transport loading system just means it'll suffer more delay before release (and generally perform worse)

Okay, I have confused the post by talking about two things that would be handled separately. I initially proposed that vehicles are made clients - almost considered as separate people - in the NHS system, since their interior surfaces are a possible vector for transmission. The follow-on idea was that the bluetooth on phones, now expected to be turned on by default, could be used for realtime vehicle loading data, and be consumed via a third party app (not the NHS one).

Am I right in thinking that passenger counting systems simply measure the loading on the suspension? If bluetooth is now expected to be turned on with most handsets, a counting system using that might be cheaper to install (where it does not already exist) and more accurate than effectively turning the train car into a giant weighing scale.

My hope would be that a system such as this would help transport operators be seen as proactive in preventing spread of virus, and able to counter government advice not to use local transport which is very lightly loaded outside the major cities.

 

[Nym]

Or why not get over this government's intense case of not invented here and big brother syndrome, swallow one's pride, realise that what they want to do isn't permitted by the Apple or Android APIs and just use Apple/Google's solution? (That was ready months ago!)

 

[Domh245]

Okay, I have confused the post by talking about two things that would be handled separately. I initially proposed that vehicles are made clients - almost considered as separate people - in the NHS system, since their interior surfaces are a possible vector for transmission. The follow-on idea was that the bluetooth on phones, now expected to be turned on by default, could be used for realtime vehicle loading data, and be consumed via a third party app (not the NHS one).

Whilst the vehicle interior surfaces are a transmission vector, giving each vehicle a bluetooth 'client' so that each passenger is marked as being in contact is a bit of a sledgehammer/nut scenario. Not only would you be getting lots of false positives from it (each vehicle being ~20m but the individual only being in contact with a small part of it) but it'd be a lot of expense to get enough transceivers to do this - you'd be better spending that on an enhanced cleaning regime (which would also be more effective as you're stopping people from catching the virus rather than making it easier to trace)

Am I right in thinking that passenger counting systems simply measure the loading on the suspension? If bluetooth is now expected to be turned on with most handsets, a counting system using that might be cheaper to install (where it does not already exist) and more accurate than effectively turning the train car into a giant weighing scale.

Most importantly, my hope would be that a system such as this would help transport operators be seen as proactive in preventing spread of virus, and able to counter government advice not to use local transport which is very lightly loaded outside the major cities.

There are different ways of doing passenger counts. Some are simple suspension based weight systems which have been around since the EP brake allowing the brake force to be altered depending on load IIRC with the added advantage of basic train loading monitoring, but you also have infrared count systems and even individual seat occupation systems. Basic IR count systems are fairly cheap now and have the advantage of being universal - they'll count everyone regardless of if they have a smartphone with bluetooth on or not.
With the prevalence of smart watches, bluetooth had started to become more "on by default" anyway, but trying to do passenger loading detection with it would present all manner of data protection issues, surely? The fact it's taken a worldwide pandemic for this sort of system to come to use should be evidence enough that bluetooth tracking is neither desirable nor practical on a regular basis.

Given that the government are currently paying the transport companies to run around without any passengers, it'd be pretty stupid of them to start countering the advice to encourage their use. There are a lot of things that they could do to encourage passenger use once we reach a phase of the outbreak where this can be done which would be both more effective and cheaper than this bluetooth idea.

 

[Bantamzen]

If the NHS app gets introduced it is hoped that over 60% of people will use it, as that is the proportion required to make it effective. If all busses and train cars had the necessary bluetooth transceivers and were enabled to talk to the app, it would provide benefits:

• There is a hard and fast record of what vehicle a person who subsequently tests positive has travelled on, providing another means of identifying contacts, beyond direct person-to-person swapping of ID through people's phones.
• The bluetooth system could, presumably regardless of any app compatibility, count the number of live devices (and thus people) onboard at any one time. In conjunction with 'open data' and a suitable choice of third party apps, prospective passengers would be able to see how many people are onboard those services they are likely to use.

I feel the second one would provide the most benefit for people and perhaps be reason to get the NHS app in the first place. A technical system would also indicate that we are taking a 'can-do' attitude to the issue of making transport networks safe, and encourage people to make journeys by routes and times that are less busy.

I'm not sure this would work very well in principle, both the NHS app and Google/Apple API work on the basis that enabled devices handshake with each other, and gather the pseudos of the devices they stay in range with for a set period of time. Then they either regularly upload the data to central database which controls who gets notified if a positive result is indicated (NHS), or directly alert users directly that may have come into contact (Apple/Google). From an individual device to individual device this might work quite well, but a public vehicle collating and informing could risk sending a lot more false positives.

As for loading, unless every single passenger complied, then it really wouldn't work. Imagine if just one of say 50 passengers had the app activated? Its not going to be very accurate.

 

[Mogster]

Big problems for the roll out in Australia it seems.

How did the Covidsafe app go from being vital to almost irrelevant?

The PM told Australians in April the contact tracing app was key to getting back to normal but just one person has been identified using its data

www.theguardian.com

Yet nearly a month since launch, the contact tracing app has barely been used – just one person has been reported to have been identified using data from it.

And the language from public officials has been toned down. No longer is it the key to freedoms, but an add-on to existing contact tracing methods, to work in concert with social distancing rules and continued testing to keep a lid on outbreaks.

So how did it go from being the key to allowing Australians to get back to the footy to being barely relevant?

 

[87 027]

Interesting idea but not sure what it adds over and above individuals downloading the app for themselves - and if you choose not to do so, how you can be traced for notification any better in this situation than before.

Depending on where the Bluetooth transceivers are situated, you might also get situations where false positives arise due to people close to the bus but not actually on board it. In theory this should be minimised because the apps measure Bluetooth signal strength as an approximation for distance, but it is a theoretical possibility.

And the National Cyber Security Centre's technical paper on how the NHS tracing app works includes a section on how they deal with the scenario where a malicious actor deliberately triggers a wave of false notifications. You'd also have to allow for this possibility.

Personally I'd spend the money on massively enhanced testing capabilities.

 

[squizzler]

Okay, so lots of good reasons here not to use the NHS app (should it appear) as a proxy for measuring traffic on public vehicles. I think this occupancy data is very important but there are apparently more ways to skin that cat than I appreciated:

There are different ways of doing passenger counts. Some are simple suspension based weight systems which have been around since the EP brake allowing the brake force to be altered depending on load IIRC with the added advantage of basic train loading monitoring, but you also have infrared count systems and even individual seat occupation systems. Basic IR count systems are fairly cheap now and have the advantage of being universal - they'll count everyone regardless of if they have a smartphone with bluetooth on or not.

This data needs to be more readily available an d the public needs to be encouraged in its use. A possible app might use what data points it can get and do a certain amount of deduction as Realtime Trains seems to do. Maybe Swlines is the best experienced developer to tackle this challenge for trains? Bus data might not be realtime but presumably they know on a day-by-day basis what their traffic is on any given diagram.

 

[squizzler]

You read it here first, folks!

Digital alerts to warn UK rail passengers of busy trains and stations
Rail industry hopes widespread sign-up will enable journeys to be staggered during pandemic.

Rail passengers can now be warned of busy trains and stations before they leave home as part of a UK industry system designed to help maintain physical distancing and safe travel during the coronavirus pandemic.

The technology will combine data on journey trends and live updates from station staff, to both inform passengers searching for journeys on the National Rail website and app, and alert those who opt in for updates on specific journeys, using their anonymised data to help predict how busy each train will be.

Article continues...

Thread title edited to reflect what transpired. Note that First Bus already provide realtime reporting, presumably other transport operators do also.