‘Driverless’ DLR trains?
I have been travelling around the Docklands Light Railway recently. This caused me to wonder whether those who keep suggesting that the London Underground could have ‘drivers’ wandering around the passenger compartment, rather than driving trains, have ever been on the DLR, or indeed the Underground. What I find is that when the DLR is busy the train captains neither wander around the train nor operate the doors from the doorway positions. They are actually to be found in the front passenger seat at the driver’s control position and operate the doors from there; safe train despatch is carried out using the platform-mounted mirrors provided for the purpose. It is obvious why: it is simply not possible to operate the doors effectively from within a heavily crowded cabin and it is faster to supervise train despatch from the front driving position rather than by holding one set of doors open to do so, adding several seconds at each station. These are exactly the same arguments that I have already made out for the inappropriateness of the train captain concept on the Underground and which are already known and understood by LU management, which has distanced itself from the DLR-type ‘GoA 3’ method of operation, as it is called. [GoA stands for Grade of Automation.]
Off peak, DLR captains still patrol the trains, and we must bear in mind the DLR is an ‘open’ system, so on-board revenue protection has a definite and ongoing role here which helps justify their cost. The Underground, of course, is essentially a closed and gated system where the issues are different.
Latest on the London Underground proposals
The present plan is for the Piccadilly Line to operate in GoA 2 mode (automatic train operation, ATO) in 2026 and GoA 4 (fully automated and with no technical need for an operator) in 2029, just 14 years away. The equivalent dates for the Central Line are 2029 and 2032 respectively. Both lines will be equipped with platform edge doors prior to GoA 4 operation. Bakerloo Line should achieve GoA 2 in 2035, with no mention of GoA 4. All the sub-surface lines will be GoA 2 operation when the new signalling is finally installed, perhaps complete by 2018 (but probably not), and nobody has mentioned GoA 4 on these lines even though the safety issues are slightly less challenging. You can see that even if these two line are successfully converted, more widespread loss of drivers (if there is actually a loss at all) is just not being talked about by those who run the network and one conjectures is a very long way off indeed.
Although the annoying word ‘driverless’ is still being used in the press, and by certain politicians seeking to gain attention, LU is (quite rightly) being very quiet about how GoA 4 would operate in practice, and in all probability they simply do not know at this stage. LU is forecasting that in spite of relief given by the new trains, overcrowding will still occur north of Kings Cross on the Piccadilly by 2031, and over a much wider distance east of Liverpool Street on the Central. In both cases it is also assumed that traffic levels will still be increasing, extending over time the stretches of line defined as overcrowded. For all the reasons I set out in my original paper about fully automatic operation, it seems hugely inappropriate to jam a member of staff into this swaying throng as well. I won’t repeat the argument, but you can look it up HERE in my Modern Railways article.
The questions still to be answered before GoA 4 operation is attempted in the deep tube resolve themselves into the following, which I do not pretend is complete:
1. Will staff be scheduled to accompany every train in GoA 4 mode, at least in the deep tube, on the grounds of safety?
2. If provided, where will the member of staff be located on the train and what skills will be required to deal with emergencies and provide reassurance? Naturally, not all the benefits can be realized if it remains mandatory to have a highly skilled operator on board every train.
3. If trains are permitted to run unattended, what measures will be required to mitigate the impact of a serious failure in single track tube where a train cannot be moved by remote control and evacuation is very difficult (but may be very urgent)? The little incident last year at Holland Park where passengers panicked when they perceived smoke (it was brakes) and climbed out between the carriages at a station platform is symptomatic of the issues that have to be considered. As far as I am aware, no GoA 4 system operates in an environment like the London 12ft deep tube.
4. The technology required to operate in GoA 4 mode on a very old system has yet to be fully developed, eg reliable systems to detect obstructions on the track, platform edge doors on curved platforms and in open air, automatic timely train despatch avoiding closing doors on crowds (or vulnerable individuals) still boarding, and so on. No doubt all this is capable of solutions, hopefully after extensive testing, but fitting this stuff on an old system and not a purpose designed one is very challenging.
5. Quite what happens where tube and surface stock share tracks has not yet been resolved, but GoA 2 operation along the Uxbridge branch (ie with a train operator at the front) has been mooted. Whether this requires a cab or some other arrangement is anyone’s guess.
I have no reservations myself that unattended GoA 4 operation is perfectly feasible in the open air (and about half of the Underground is in the open air). I have some reservations about GoA 4 along the sub-surface lines where, in the main, one is dealing with twin-track tunnels where it should be possible to get help quite quickly from a train going the other way.
To be able to get help quickly at any time of the day or night is a huge challenge and I have no doubt managers are already beginning to think about it. ‘Help’, to me, feels as though it might be graded. Initially it could take the form of someone to provide reassurance and gain information until a technician can get there. In a serious emergency they could carry out instructions given via the hardened radio system. What happens if there are no trains going the other way (eg first trains, last trains, or an emergency stopping traffic in both directions) is going to be more difficult to handle.
Resolving an emergency in the deep tube, where trains are effectively confined within a closely-fitting iron pipe with access only from the ends, I see as an order of magnitude harder to sort out. Just getting someone along to have a look could be a difficult task in the first place. If we had larger tunnels with a side walkway, as modern systems have, together with multiple ‘intervention’ shafts, then it would be far less of a problem, but we don’t have anything like that on the antiquated Central and Piccadilly Lines where some of the inter-station distances can be over a mile. You can see that it would be quite possible for there to be one or more passenger trains between the nearest point of access and a failed train, with all the complications that adds. It is a shame that the London 2050 Infrastructure Plan makes no mention of ever addressing the 12ft tube issue, or even looking at the feasibility of enlargement, which would hugely facilitate unstaffed trains in the deep tube. [My blog post 26 August 2014 – The Hole Just Isn’t Big Enough]
And who would we send along as the first line of response to a failed train? Historic thinking would suggest a station supervisor (an old term intended to indicate a responsible person trained in the rules and able to assess a situation). We must, however, now consider the present station staff reorganization which will result in smaller staffs, particularly at quiet stations and with less involvement with train operations than in years gone by. In an emergency the station staff will already have their hands full and losing someone to investigate an incident in the tunnel (with the likelihood they will be away some time) could easily reduce numbers to the point where the station would have to close. And are we going to train them to handle these very infrequent events? If so I can only see the training being of the most basic kind. Better perhaps to have comprehensibly-trained specialists located at strategic points along the line. Or we just keep the suitably trained person on the train and admit that the potential benefits (cost reduction and service flexibility) will be reduced. The Central Line has substantial mileage in the open air and one could envisage a hybrid staffing arrangements where an ‘operator’ only accompanies a train in the tunnel portion. This could also be done on the Piccadilly Line, but with much more tunnel operation (and the complication at Heathrow) the savings would be less.
I regard GoA 4 operation as inevitable and have no doubt it will be technically reliable and offer up some useful operational benefits. Any skepticism which is detectable in what I am saying is reserved solely to the environment of the deep-level tube and the issue of whether or not to have a member of staff on-board as part of the safety regime. If a member of staff is retained, then for the same practical reason I have identified on the DLR (amongst other practical reasons) then I cannot help thinking the best place for that person would be in a cab at the front of the train…
None of this will diminish the political sabre-rattling by ill-informed politicians intending to suggest by their inappropriate use of the word ‘driverless’ that train drivers are on the way out. I suggest the reverse is true, with GoA 4 a long way off and only introduced on two lines. In the meantime, ever increasing service levels and the construction of new London railways would appear likely to be calling for more train staff than now, not fewer.