Cameras in the driving cab

[John1974]

Apparently a number of Tocs have agreed to trial cameras to monitor fatigue.

What are your thoughts on this?

Personally think this is just another device to blame the driver with....

Details.

Rail Attention and Alertness Monitoring Trial – T1193
After the tragic Croydon Trams Sandilands tram accident in 2016 in which driver fatigue was implicated, the ORR highlighted that the Rail Industry need to invest more research capacity in understanding how to manage driver attention and alertness issues more successfully.
Croydon Trams implemented attention and alertness monitoring technology after this accident and have noticed a 75% reduction in fatigue related events, a 40% reduction in the number of collisions arising from driver distraction, alongside a 316% increase in driver incident reporting – these are significantly positive safety benefits.
The T1193 project will work with three Rail companies (c2c, DB Cargo and LNER) to trial the use of a similar alertness and attention safety monitoring device located inside train cabs. We are excited to understand what sorts of safety benefits may be achieved while using this technology and anticipate that there will be systemic improvements to the way that attention and alertness issues are managed for train drivers.
We are working closely with the three Rail companies and their Trade Union representatives to ensure that we meet the needs of companies and drivers in trialling this technology. We are still developing the trial plan details and will provide regular project updates. We hope to start the monitoring trial in Summer 2023.
What does the device do?
• The device works by monitoring the driver’s facial features and detects alertness behaviours related to fatigue (e.g., yawning, long eye blinking and microsleeping) and distraction related behaviours (e.g., looking away from the line for a long period of time). [The exact device to be used is still being chosen].
• There will be an audible and physical alert for an alertness event (e.g., an audible alarm and a seat shaker) and a different audible alert for a distraction event. These will be distinct from other alerts in the cab.
• The device is not connected to other safety features of the train and will not prevent train use if it is not working.
• Drivers won’t need to make any adjustments to the device, it will work automatically when the train moves.
• It is not a CCTV device and is not constantly recording. It will only capture a snippet of video if a safety event occurs (e.g., a microsleep). You can see an example of this kind of snippet here:

• There are stringent safeguards about who can access data from the recording device – drivers will have access to their own safety data if desired.
• We are working with the three companies to determine a consistent approach if an alert is activated, and how they can be dealt with in a fair and just manner outside of any disciplinary procedures. We will collectively determine procedures to support drivers if they are not able to continue with their shift due to alertness or distraction safety events

 

[zwk500]

Personally (and not a driver) if it gets to the point where a driver is fatigued in the cab, it's gone too far. Fatigue monitoring needs to be strong enough that managers or colleagues can intervene without fear of repercussion and Drivers need to have confidence in the system that they can self-report fatigue without penalty. As I don't know what these systems currently are, I can't say what can be improved.

Active-technology cameras feel like shutting the stable door after the horse has bolted. There's a big difference between a CCTV which just records what happens and this, where an algorithm decides what snapshots to take.

I also don't like the fact that the technology won't intervene to bring the train speed down despite being used to determine what is and isn't a safety risk. I know it's for cost and performance reasons, but if it isn't worth the money to do it properly, are we really getting any value out of it at all?

 

[357]

Been waiting for a thread on this.

Company/local aslef reps had no involvement in agreeing this. The PDF was sent out by management and that was the first the reps knew of it.

How does it look to the staff, where industrial relations are already extremely bad, to see this and the reps don't know anything about it?

At at least two of the three companies, the main concern is that management will be able to watch whatever they want, and it will be downloadable after any incident or for general monitoring etc... "You were picking your nose 40 seconds before your TPWS"

We have had confirmed cases in the past of managers browsing OTDRs remotely looking for something to do with themselves.

As we have learnt the hard way before - technology that's installed to one spec can often have other features that aren't advertised.

Personally (and not a driver) if it gets to the point where a driver is fatigued in the cab, it's gone too far. Fatigue monitoring needs to be strong enough that managers or colleagues can intervene without fear of repercussion and Drivers need to have confidence in the system that they can self-report fatigue without penalty. As I don't know what these systems currently are, I can't say what can be improved.

Active-technology cameras feel like shutting the stable door after the horse has bolted. There's a big difference between a CCTV which just records what happens and this, where an algorithm decides what snapshots to take.

I also don't like the fact that the technology won't intervene to bring the train speed down despite being used to determine what is and isn't a safety risk. I know it's for cost and performance reasons, but if it isn't worth the money to do it properly, are we really getting any value out of it at all?

I'm firmly in the camp that this is more about getting cameras in cabs than it is about fatigue management.

 

[zwk500]

I'm firmly in the camp that this is more about getting cameras in cabs than it is about fatigue management.

It does feel like that.

 

[Sly Old Fox]

That is definitely the case.

If fatigue risk is so high, maybe they could try building rosters with less severe shift patterns (02.30 finishes one week, Saunday off, 02.30 starts next week anyone?), or employing more staff so drivers don’t have to work so many rest days. But it looks like they’re going for the spying option, and putting even more pressure on drivers.

Still, you could probably block it with a stop car reminder booklet.

 

[Neo9320]

Still, you could probably block it with a stop car reminder booklet.

Username checks out you Sly Old Fox

 

[357]

That is definitely the case.

If fatigue risk is so high, maybe they could try building rosters with less severe shift patterns (02.30 finishes one week, Saunday off, 02.30 starts next week anyone?), or employing more staff so drivers don’t have to work so many rest days. But it looks like they’re going for the spying option, and putting even more pressure on drivers.

Still, you could probably block it with a stop car reminder booklet.

Yep. No idea why the week of early / week of lates couldn't be changed to something easier on your body - be it starting on the earliest shifts and getting gradually later throughout the roster, or 3/4 weeks early then 3/4 weeks lates.

You're lucky, you must work for a TOC with loads of money! We haven't been issued stop car reminder books (or "bingo cards") for about ten years. Yet the company want to reduce stop short risks.... I was told off by a manager once for "making the cab messy". I'd scrawled 12 on two pieces of paper and folded them over the top of the desk so they were hanging in front of the door buttons. This way I can't release the doors without seeing a huge number 12 first. It works for me, especially when we were so rarely driving a 12 coach train.

 

[irish_rail]

My first reaction to reading this is to check the date isn't April 1st. Pretty lost for words really. What this WILL lead to is a HUGE increase in driver sickness, as the minute a driver hasn't managed 8 hours sleep for whatever reason they'll just ring in sick as fatigued rather than risk whatever the consequences are for appearing tired whilst driving. This really hasn't been thought through. A truly terrible idea, and the Dft better start training up 100s and 100s of new drivers in order to cope with the levels of sickness that will be seen (especially on early shifts!). This is going to be expensive for them!

 

[Donny_m]

I’m interested in to how this has ‘cut down’ on anything? Is it like a drug test where people may stop taking drugs knowing they may be tested at work? What do people do, all of a sudden sleep better?

Surely the money should first be spent on the cause of fatigue rather than playing a siren to tell you what you probably already know.

So what happens when the system detects your tired? Nothing? Isn’t that like a breathalyser that still lets you complete your journey.

 

[Western Lord]

I am old enough to remember when drivers of first generation DMUs were in full view of the public (or any railway official travelling incognito) all the time.

 

[Turtle]

I am old enough to remember when drivers of first generation DMUs were in full view of the public (or any railway official travelling incognito) all the time.

Yes. I recollect sitting in the front seat on Bath - Bristol journeys although I didn't pay any attention to the driver. It was what appeared to be the incredibly long pauses between gear changes which impressed me most.

 

[pdeaves]

As I see it (and my full knowledge is limited to what's written in the first post), it's no different to having AWS. "There's a signal coming up, you may already know about it but here's an alert in case" vs "You act tired, you may already have it under control but here's an alert in case". It doesn't look to me to be anything cynical, simply a last line of defence against an accidental action (the same as AWS is a last line of defence against accidentally passing a danger signal). I am not aware of anyone getting upset about having preventative alerts elsewhere (though I have no doubt when first introduced there were reservations as it was new and unfamiliar).

 

[zwk500]

As I see it (and my full knowledge is limited to what's written in the first post), it's no different to having AWS. "There's a signal coming up, you may already know about it but here's an alert in case" vs "You act tired, you may already have it under control but here's an alert in case". It doesn't look to me to be anything cynical, simply a last line of defence against an accidental action (the same as AWS is a last line of defence against accidentally passing a danger signal). I am not aware of anyone getting upset about having preventative alerts elsewhere (though I have no doubt when first introduced there were reservations as it was new and unfamiliar).

AWS isn't making judgements about you though, and (importantly) WILL intervene if you don't acknowledge. The big issue I have, as a passenger, is that these cameras are only going to be of any value after a crash. If I was a driver I'd be mildly miffed about an algorithm being used to determine when to record evidence of me potentially being tired, because it will lack context. There is already a Driver Vigilance Device (or similar name) that drivers have to consciously action so what do these cameras actually add?

 

[Bletchleyite]

There aren't many things I'd go on strike over - I'm not generally militant - but I would go on strike over this, and probably consider a new job. I suspect the RMT and ASLEF would too.

A vigilance device is one thing, effectively continuous imperfect monitoring is another entirely.

As I see it (and my full knowledge is limited to what's written in the first post), it's no different to having AWS. "There's a signal coming up, you may already know about it but here's an alert in case" vs "You act tired, you may already have it under control but here's an alert in case". It doesn't look to me to be anything cynical, simply a last line of defence against an accidental action (the same as AWS is a last line of defence against accidentally passing a danger signal). I am not aware of anyone getting upset about having preventative alerts elsewhere (though I have no doubt when first introduced there were reservations as it was new and unfamiliar).

It's much more personal than an AWS because it's recorded and will be used in performance management (it's silly to claim it won't).

 

[berwicksfinest]

Unfortunately my FOC is involved in this

The question brought up time and time again, OK a driver is tired after stupid o'clock starts all week and them a night job weekend (apparently allowed, as we alledgedly voted for this)
This device recognises your tired and "wakes you up"

The company in now complicite in recognising "and allowing" you to continue driving tired ??

Can we now have loops built so if we're tired we can pull over and have a nap ??

 

[skyhigh]

As I see it (and my full knowledge is limited to what's written in the first post), it's no different to having AWS. "There's a signal coming up, you may already know about it but here's an alert in case" vs "You act tired, you may already have it under control but here's an alert in case".

Completely different. AWS will not send a video of you to your manager if it deems you took too long to blink (which it sounds like this trial may do).

There's AWS/TPWS, vigilance and DSD fitted in cabs for a reason.

 

[zwk500]

Unfortunately my FOC is involved in this

The question brought up time and time again, OK a driver is tired after stupid o'clock starts all week and them a night job weekend (apparently allowed, as we alledgedly voted for this)
This device recognises your tired and "wakes you up"

The company in now complicite in recognising "and allowing" you to continue driving tired ??

Can we now have loops built so if we're tired we can pull over and have a nap ??

As somebody who actually has experience of these things, is there anything these cameras do that isn't covered by the Vigilance device already?

 

[Thirteen]

What do they use on the Tramlink? I would have thought rather than rolling it out to heavy duty trains, it would be better to start small and trial in other places with trams as well as the Tube, if it works there then roll it out to heavy duty trains.

 

[John1974]

Did the union not have to consult the drivers before this trial was brought in?

Unfortunately my FOC is involved in this

The question brought up time and time again, OK a driver is tired after stupid o'clock starts all week and them a night job weekend (apparently allowed, as we alledgedly voted for this)
This device recognises your tired and "wakes you up"

The company in now complicite in recognising "and allowing" you to continue driving tired ??

Can we now have loops built so if we're tired we can pull over and have a nap ??

Thats quite a good point.

The device is saying you are tired but allowing you to continue in service.

 

[Bourlea]

What do they use on the Tramlink? I would have thought rather than rolling it out to heavy duty trains, it would be better to start small and trial in other places with trams as well as the Tube, if it works there then roll it out to heavy duty trains.

This is the system that is installed on Tramlink and has been in use for at least 3 years.

 

[D6130]

I am old enough to remember when drivers of first generation DMUs were in full view of the public (or any railway official travelling incognito) all the time.

Except on services into and out of Marylebone, where the blinds were kept firmly down to avoid the prying eyes of commuting British Railways Board staff!

There aren't many things I'd go on strike over - I'm not generally militant - but I would go on strike over this, and probably consider a new job. I suspect the RMT and ASLEF would too.

So would I....glad that I'm now retired!

 

[dastocks]

Clearly the detail of these two statements needs to be gone into very carefully:

• It is not a CCTV device and is not constantly recording. It will only capture a snippet of video if a safety event occurs (e.g., a microsleep).

• There are stringent safeguards about who can access data from the recording device – drivers will have access to their own safety data if desired.

The big issue I have, as a passenger, is that these cameras are only going to be of any value after a crash. If I was a driver I'd be mildly miffed about an algorithm being used to determine when to record evidence of me potentially being tired, because it will lack context.

Don't think so, this is actually a fairly mature technology - it looks very similar to something that is already installed and/or available as an option on certain cars (I think it was available on later versions of my car) and I understand it provides alerts in a very timely fashion. It's the sort of thing that reduces insurance premiums.

• The device works by monitoring the driver’s facial features and detects alertness behaviours related to fatigue (e.g., yawning, long eye blinking and microsleeping) and distraction related behaviours (e.g., looking away from the line for a long period of time). [The exact device to be used is still being chosen].
• There will be an audible and physical alert for an alertness event (e.g., an audible alarm and a seat shaker) and a different audible alert for a distraction event. These will be distinct from other alerts in the cab.

It's much more personal than an AWS because it's recorded and will be used in performance management (it's silly to claim it won't).

I would have thought it more relevant to question the value of this over TPWS, which (AIUI) does actually intervene in the operation of a train. Regardless, if a particular driver generates a lot of alerts I would have thought that would be something that needs to be investigated and action taken.

 

[bengley]

We don't need this on the main line. We already have DSD which does this job perfectly well, plus constant AWS warnings and TPWS as a last resort in case you really do cock up.

This is a way of getting inward facing CCTV into the cab, nothing else.

 

[godfreycomplex]

Admittedly I’m not a driver, and whilst I do respect the circumstances that have brought this about (namely the Sandilands tragedy), emphatically a safer solution for a tramway (or other railway) would be some species of speed control system for speed limit changes/signals that would be affected by microsleeps/distraction etc, rather than this system that will do little more than cheerfully beep whilst major safety events occur.

As for its use on railways with speed supervision (mainline railways, in other words), if you are really concerned about fatigue, put a properly thought out and staffed roster in place.

And finally, for any managers that are enforcing these systems, you need to answer the question of why you are enforcing such horrendously immoral things as a punitive approach to fatigue based events on people who have in all good faith attempted to do a safety critical job to the best of their ability and have had, as all humans do, circumstances befall them that have prevented them from doing that.

I think we would all like to know how you consider that ethical

And “because I was told to from above” is not an answer.

 

[jon0844]

Many cars now have systems to detect a driver becoming less attentive. Some are very basic, while others have cameras that recognise particular faces (so you can customise seat and other settings) but I still don't think actually captures or saves video.

I am not sure what the car might record that could be interrogated in the event of an accident, but perhaps any additional safety system for a train can be designed so it can warn a driver in the cab but not actually do anything more. Maybe only a second or third warning will actually be recorded on the data recorder.

The driver can then decide if they really are tired or need to report that they're no longer fit for duty.

 

[Thirteen]

This is the system that is installed on Tramlink and has been in use for at least 3 years.

I assume ASLEF were involved with that given their drivers are on Tramlink which perhaps why there hasn't been any objections to it being rolled out.

 

[XIX7007177]

Nothing a bit of tipex wouldn’t sort out.

 

[Stephen42]

We don't need this on the main line. We already have DSD which does this job perfectly well, plus constant AWS warnings and TPWS as a last resort in case you really do cock up.

This is a way of getting inward facing CCTV into the cab, nothing else.

What about the incident at Kirkby? In that case the driver operated the DSD, reset the AWS warnings and did not trigger TPWS despite being distracted from the driving task. The full RAIB accident report is available here.

The first recommendation is about T1193:

The intent of this recommendation is that additional research be undertaken into systems which can detect and monitor driver alertness and awareness, and how these could be trialled in the industry.

RSSB, in consultation with relevant stakeholders and bodies representing staff, should undertake further research into how the detection and mitigation of a loss of alertness or attention in train drivers can be improved. This research should specifically consider the effectiveness of systems currently in operation and build on work already completed, such as the functional specification and proposed trials set out in the T1193 research report. It should also take into account relevant practice from other transport systems (paragraphs 114a and 114b.i).

I fully get why drivers would not want this system and the risk that it becomes another stick to beat drivers with. Expect most would agree the driver in the incident quoted actions were far below the acceptable standard and is obviously the exception rather than the rule. However, on the flip side the existing safety mechanisms did not prevent the incident and looking at possible technological advances rather than just adjustments to avoid a repeat incident seems sensible.

The industry should also look at whether there are alternative solutions that are less invasive could achieve the same goals. The fatigue element sounds the larger risk area, while roster changes and improved monitoring/self-reporting would help, a solution like this in addition when used properly and non judgmentally has the potential to identify real world risk areas/patterns that paper based assessment alone won't and alert drivers that they may be fatigued.

 

[Bletchleyite]

I assume ASLEF were involved with that given their drivers are on Tramlink which perhaps why there hasn't been any objections to it being rolled out.

Interesting.

I'd say with Tramlink it was the wrong solution. The correct solution would have been something more ATP-esque to prevent a tram overspeeding or passing a signal at danger/red traffic light, and similarly a radar based system to stop a tram hitting another from behind in an area of "drive on sight" (as fitted to almost all new cars now). Probably it was cheaper?

What about the incident at Kirkby? In that case the driver operated the DSD, reset the AWS warnings and did not trigger TPWS despite being distracted from the driving task. The full RAIB accident report is available here.

Kirkby should have been physically impossible with a proper active train control system, whether the driver was texting his mate or not.

 

[mikeb42]

A root problem with objection that nobody (so far) has addressed is this:

"Croydon Trams implemented attention and alertness monitoring technology after this accident and have noticed a 75% reduction in fatigue related events, a 40% reduction in the number of collisions arising from driver distraction, alongside a 316% increase in driver incident reporting – these are significantly positive safety benefits."

If (and it's a big if) these claims are evidentially robust and withstand peer review and all the rest of it: this is a big effect on a big problem. Even if it comes at the cost of aggrieving staff, your average passenger would read that and consider it tough luck.

The comparison immediately seems to me to fall apart though, as what safety devices are there in a tram cab? Maybe some sort of vigilance device? Forward facing CCTV to record all the pedestrians in Manchester trying their best to get run over? But no AWS, TPWS, ATP etc as trams are necessarily driven sort-of on sight at least on-street? What about "flight data recording" type monitoring? I don't know - not enough knowledge.

Were I the responsible safety engineer in a related system like the mainline railway, when confronted with claims like that it'd be both professional obligation and ethical responsibility to either forensically dismantle the original claims, prove it doesn't work in the different context I'm responsible for, or consider implementing it.

If it doesn't yield those benefits in the context of mainline railways, someone is going to have to prove it. There's only one way to do that. Personally I'd hate to be subject to monitoring like that at work. However, if with all other things being equal it does demonstrably work, persuasively arguing against it is going to be hard unless it is more expensive than very costly whole-system measures like recruiting more staff to enable less onerous shift patterns.