‘Driverless’ Underground Trains – The Issues.

This blog piece was the forerunner of an article in Modern Railways in 2012 that you can read HERE

The London mayoral election campaign exposed to the public gaze Boris Johnson’s intention to ‘work towards the introduction of driverless trains within a decade’; Boris has now been re-elected as mayor and presumably believes he has a mandate to push on with this concept. The intriguing promise of ‘driverless’ trains was justified on the grounds it would make the Underground cheaper and more reliable. The intrigue is his assurance that every train will ‘always have a staff member on board to assist customers, as they do on the DLR’. The promise cannot have been made without some kind of support from within TfL and raises so many conflicting considerations that it seemed worth setting some of them down. As the word ‘driverless’ has been used, I will use the word ‘driver’ in this piece, though London Underground calls the staff ‘train operators’ and, of course, on automatic lines they do not do much driving.
The proposal is obviously only viable on lines that will have an automatic train control (ATC) system; at the moment ATC systems require the person in the cab to open and close the doors and, when it is safe to do so, to initiate automatic driving that will take the train to the next station unaided. This is a fundamental requirement of ATC and addresses the scenario where a driver becomes ill after the train starts: the train will always get to the next station. Clearly it is possible to contrive additional automation that can open and close the doors and, once the doors are proved closed, to initiate automatic operation. If we can do that, there is nothing left for a driver to do. What, then, is the problem?
In my world, for I come from the Underground and have spent happy years managing incidents on both manned and automatic lines and conducting driving tests, there are two challenges. The first challenge is to manage the station activity effectively; this process is often called ‘train despatch’. The second challenge is how to deal with failures and emergencies between stations. In these situations, the luckless driver is the only representative of the company present and is expected to handle the problem and keep everyone informed, often unaided unless the problem is so bad it warrants external help being sent in. Neither of these challenges is easy, but before we say ‘it’s all too difficult’ we need to ask two further questions. First, how well do we actually achieve perfection with the existing mode of operation, for perfect it is not. Secondly, in looking at the assertion that ‘driverless’ will be cheaper, what goes away? If the railway is to be cheaper to run, then some real source of cost must go away, and go away completely (ie not simply be renamed or redeployed). On the face of it, on driverless trains the drivers must go away. But the proposition is that every train must still have a member of staff on board. The conclusion, then, is that what goes away may be the difference in pay rates between a driver and whatever the new member of staff will be called (for want of anything else I’ll use the term ‘train captain’, as on DLR, for our roving non-driver). Whether retaining on-board staff will offer the improved service performance that has been suggested is another matter, touched on shortly.
Train Despatch
Let us look at train despatch first. Opening and closing doors sounds simple enough, but the platform-train interface is the major source of accidents on the Underground (TfL reports 164 injuries over last year). Clearly, for an organization where safety is paramount, train despatch is not as simple a matter as one might think.
Opening the doors is the least fraught part of the equation. A driver will not open the doors if the train has stopped in the wrong place, or if the station or platform is closed, or if the train is out of service, or if he can see a source of danger or he is otherwise instructed not to open the doors. In theory, some of these checks can be automated (correct stopping location already is, and partial station closure is partly catered for) and others could be done as well by someone dedicated to monitoring the platform, whether on it or in a suitable control room; the key is dedicated monitoring. I won’t elaborate on what can be done by technology and what by people as we are looking some years ahead, but it feels do-able.
Closing the doors is another matter altogether. Although drivers now get door close ‘prompts’ on some lines, the conditions vary so much between trains that it must be down to the driver to choose the moment when it is obvious all those alighting have got off and when it isn’t practical to let anyone else on. This is a judgement call because if doors close unreasonably quickly then passengers will simply hold them open, while delaying too much benefits almost nobody but seriously reduces capacity.
The next issue is managing the process between when the doors begin closing and when the train begins to pull away. Existing safety features mean trains cannot start unless doors are ‘proved’ closed. This is achieved by an interlock on every door leaf that proves it is fully closed (and a train cannot start unless every door is proved closed). So far, so good. Unfortunately something like a heavy door must put up with significant pressure from both passengers and air movement, and relative movement with a flexing carriage body. In order not to injure passengers while the doors are closing they are fitted with a flexible rubber edge. It may perhaps be seen that doors cannot close with fine precision and a small tolerance is required. It is thus possible for thin objects such as a jacket or strap to become trapped whilst still indicating to the driver that the door is closed. These conditions set the scene for unfortunate dragging incidents where someone on the platform becomes attached to the train and cannot release whatever it is that has become trapped (normally a garment with something inboard the carriage that physically prevents it from being yanked away). For this reason drivers are required to check a platform carefully after the ‘doors closed’ indication has been received and satisfy themselves everyone is standing clear, something made possible by extensive CCTV coverage available on monitors located near the driver. On recent trains, in-cab CCTV now makes it possible to check the platform even after the train has begun to pull away. These precautions all too frequently detect trapped passengers or other platform-based hazards and prevent consequent injury (including a recent one where a child fell between train and platform and was hauled out unscathed). Rarely, victims who get trapped are not so lucky.
All this is quite difficult to replicate automatically. Recent trains are fitted with sensitive door edges that will allow a degree of tolerance but will detect a dragging event. This has not proved to be trouble free and has caused some serious delays. If the sensitive edges can be made to function reliably, then it will deal with one, but only one of the platform edge hazards.
Actually, I seriously wonder if the train despatch process is best handled from the front of a train from where it is very difficult to influence events. Somebody on the platform, near the centre or near where it is most congested, might be better positioned to identify the optimal moment to close the doors, encourage people to stop boarding, see that all is clear, start the train and, if something is amiss, to stop it again. Readers may already have noticed that at busy stations part of this role is already done by a station assistant, wielding what looks like a racquet, to give a visible signal to the driver; this is an entirely pragmatic response to the safety issue, even if it not a very efficient one. Nor is the duty always carried out to optimal effect as the person is apt to face towards the front of the train whilst showing the target to the driver and not looking behind (I have myself watched the safety signal being given to the driver whilst noticing someone actually trapped in the doors one car farther back; it took the driver to notice the problem on his CCTV and reopen the doors, and the assistant still didn’t turn around…).
In short, the train despatch process could probably be managed from the station but it would be necessary to install the necessary surveillance and control equipment, and increase platform staffing. It might be possible at quieter stations to do this from a control room. With platform edge doors there is a possibility, perhaps, of further automating the process, at least at quiet times.
Before moving on, we should look at how the DLR does it. Here the doors release automatically as soon as the train stops. The train captain then selects a doorway in the passenger saloon from which train despatch will be monitored  and ‘opens up’ the adjacent control panel. From here the captain can close all the doors on the train except that one, from which a lookout can be maintained. On checking all the other doors are shut and everyone is clear, the local doors are now closed and the train automatically starts. Let us now translate this situation to (say) a Piccadilly Line train in the height of the morning peak. Under these circumstances a train captain would be hard put to get from one side of the car to the other to deal with platforms that vary from one side to the other, let alone control access within a five-foot wide doorway whilst looking out along the crowded platform to see the other doors are safe. Without a dedicated door (free of passengers, as former guards had) how would the captain have access to essential CCTV? The need to check that all doors were closed and then to have to wait while the local door closed would also add (maybe) 8 – 10 seconds a station stop, impacting on urgently needed capacity. It is hard to see how the DLR model could be applied to the circumstances of the Underground. If we are not to have a cab then it looks like full automation (technology not there yet) or control from a platform. Controlling train despatch from the platform, whilst also having a train captain, would seem perverse, though.
Dealing with Emergencies
In this context an emergency means any kind of occurrence that happens between stations where the train is rendered incapable of movement. Excluded are medical and certain other emergencies on the grounds that the best place to deal with them is at stations, though this might have implications for stations that are unmanned for parts of the day at the moment. It should be noted that operation of the passenger alarm can be configured to sound an alert in the control room but not necessarily to stop the train. Again, usually the best place to deal with passenger incidents is at a station.
A train might be unable to move either because it has developed a defect, or because of a problem on the track or in the tunnel where progress is impeded: signal failures, for example. The driver is not normally expected to deal directly with failures or other problems that are off-train, but of course is expected to pass useful and accurate information to the control room (and the passengers). In certain types of failure a driver may need to drive the train manually through the problem area. With on-train failures, the driver’s traditional role is either to fix the failure or do whatever is necessary to isolate the fault so that the train can be moved (often under manual control) to the nearest station; here passengers can be detrained before the train is driven to a depot where it can receive attention. For minor failures that can be fixed in service there are train technicians located around the network who can either perform a running repair or at least confirm a fault or isolate equipment. All this is the theory.
Let us look at these things in turn. Failures fall into several different categories.
Train can only proceed in degraded mode
Typically such a failure will include a minor track or structure failure requiring an emergency speed restriction, or a failure of signalling equipment where trains have to proceed for some distance on-sight, and therefore at caution. In the first instance the driver will have to take the train out of automatic working and drive manually, but still with the benefit of the automatic train protection system. In the second instance a mode called restricted manual is selected (limiting speed to 10 MPH) and the driver drives on sight only until ATP kicks in again; after that, the train can usually be driven at (near) normal speed to the next station. There are, of course many variations on this but I hope this gives a flavour.
On the DLR it is simple. The train captain will probably be alerted to a known problem by his personal radio, proceed to the front of the train, shift any passengers in the front seat and open up the control box, then drive the train in whatever mode is called for until the problem has been passed. The question here is to imagine translating this onto the example of a Piccadilly Line train in the peak, morning peak approaching King’s Cross, for example. Could we really expect a Piccadilly Line train captain, who could be anywhere on the train, to battle his way to the front, shift the people, maybe check the track (involving opening leading door) and safely drive the train, maybe in tunnel with glare from a fully lit car distracting him (rather than dark cab enabling driver to see out)? He will be speaking to the controller during this process, in full earshot of passengers, which is not always desirable.
To me, this feels unsatisfactory. For failures where a train is able to move, it may actually be better to engage a remote driving facility and have a dedicated shore-based control desk where a specially trained operator can drive the train through an affected section. The technology, if not there now, must surely be available shortly and is already used in military circles. This class of failure also includes the possibility of a train having to move backwards to the previous station, not something normally done without special precautions being taken; I’d be very uncomfortable doing this without someone properly trained actually on board.
A third scenario is that there is a train failure that someone on board can fix in order to provide limited movement, sufficient to get a train to the next station (if it cannot actually be fixed, it drops into the category that follows). With existing modern trains, most of the probable types of difficulty can be identified and fixed from within the driving cab. If there is to be no driving cab it raises the question of whether the train captain is trained and equipped to do a similar job from a random position on a very crowded train in full view of the passengers. Will he have a magic portable control box to operate from? Will there be static positions on a train to proceed to without having to fight through multiple carriages? I can’t begin to describe how stressful these situations can be for staff. To have to do fault analysis in full view of an anxious public and a radio blaring out things that maybe the public would be better not hearing I think is not the way to go. Could we do it remotely? I suppose it depends on what the fault actually is. This is a whole area that does not yet have an answer. I would counsel great care in drawing comparisons with the DLR, operating mainly in open air and with far lower passenger densities.
Train cannot proceed forwards under its own power
There are two reasons a train cannot proceed under its own power. One is a serious train equipment failure and the other is some kind of infrastructure failure that physically prevents progress. In either case a train full of people cannot be abandoned and external help must be provided. The help will either clear the line or fix the train, or if neither is possible extract the passengers. In all cases the job of a train captain appears to resolve itself into (1) passing accurate and timely information to control enabling optimal decisions to be made about how to deal with the fault and who to send, and (2) providing comfort to the passengers. There is often not a great deal more onboard staff can do. In the latter case one train captain effectively trapped in situ in the passenger compartment is of only limited help amongst 800 passengers, 790 of whom cannot even see the train captain, who might just as well be in the driving cab where at least they could think. I’m not sure personally I would be very comfortable seeing the person nominally in charge clearly unable to do anything (imagine that on an aircraft or a ship?).
There is a variation of the above where a train can physically move but needs assistance from a following train (not as rare as one might think). You now need two qualified drivers on site quickly, one on each of the trains. Will our low-cost train captains be trained for this kind of thing? If not, where are the drivers to come from: remember we have at least a train load of people on the first train and the possibility of another load on the second—maybe 1500 people. You really wouldn’t want to do this by remote control.
Significant in all this is that, as far as I can see, all other systems having ‘driverless’ trains, including the DLR, have a means of access to the side where staff can walk past without having to go through the train or onto which passengers can walk as part of an evacuation. There are also, often dedicated, emergency access and escape shafts. This is not the case on any part of the deep tube network except the Jubilee Line east of Green Park. This makes a huge difference to how emergencies are dealt with and places more significance on the person in charge of a train and how he is trained and what he is expected to do because it is so hard to provide help quickly. I need hardly add that a train stuck in a tunnel for five minutes is merely undesirable, but a train stuck in a tunnel for 25 minutes, under some conditions, is potentially the playing out of a real emergency.
Other events
There are lots of other things that can and do happen. Loss of electric current, for example, where a driver may need to use judgement about where best to stop if it is obvious the train cannot reach the next station. Serious smoke in a tunnel caused perhaps by arcing can happen and a driver might decide it is not safe to proceed and if necessary might have to evacuate a train on his own initiative. A train captain could in theory do this, but if there is no cab he might not perceive a problem until a train is well inside a danger area. Looking out for trespassers on the line happens frequently. Suicides (but not where there are platform edge doors). These are only examples of things to be considered. Who could have foreseen the bulging Bakerloo tunnel in April, struck by a train and where a driver at the front could take immediate action and give an accurate report — what if he’d been at the back? An automatic train would not have stopped.
What has just been described merely gives a flavour of the issues that will have to be addressed. Reference to the DLR, which has only short tunnel sections, short trains and relatively low density traffic, is really unhelpful. We must look at detailed operating methods of similar types of railway operating in a similar environment. Meteor Line 1 is of course a new railway in double track tunnel (and with platform edge doors). Lessons to learn from there and from other locations maybe? Not very many metros run deep underground in single track tubes with limited access though; even Copenhagen has large tunnels with a walkway, such as that on the Jubilee Line extension. London, with tunnels a century old, is a much tougher nut to crack. Things are much easier in the open air, or even in double track tunnel, where help can be sent via a train on another track.
Status Quo?
Sticking with ‘conventional’ drivers on otherwise automatic lines is possibly equally fraught though. With equipment more and more reliable, but equally more and more complex, we may be approaching a point where it is impossible to have confidence that staff can reasonably be expected to handle, to the standard required, all manner of foreseeable occurrences, let alone those that are hard to foresee. The range of incidents for which they have to be trained tends to increase whilst the frequency with which they occur tends to decrease, denying them real life experience. This is a combination that is a huge challenge to any organization, though it is not new and is by no means confined to rail transport. What may be new as traffic rises alarmingly is the mounting lack of acceptance of failures in the first place and any inefficiency in how long it takes to deal with them in the second. Serious Delays impact on tens of thousands of people, can create potentially dangerous situations involving hundreds of people, and (a new factor) are very conspicuous, which annoys the Mayor’s office and introduces a new form of management pain.  Clearly zero failures would be nice, but impossible to deliver and apt to reduce further experience in how best to deal with them.
I have seen this myself on even basic rolling stock where decent, conscientious staff may pass through most of their career without ever having had certain types of failure. When, inevitably, it does happen on their train it is mishandled, often as something akin to panic sets in. The outcome is a horrendous delay with thousands of people trapped. With the Underground ever busier the inconvenience and risks of mishandled delays is increasingly unacceptable.
It is obvious that we could train staff even harder, and give them continual refresher training (often using simulators). We can give them radio links to train doctors. We can give them voluminous fault-finding documentation and procedures, as pilots have. We can blue-light emergency response vehicles. Doubtless there are other options too. But what we are actually doing is giving more and more training to staff who, when lines all convert to ATC, will be doing less and less. Such an approach would create very large numbers of staff with very expensive training for very rare events. It is probably a good time to be giving some thought to whether this is the best way of proceeding. Might another approach be better all round? Might replacing on-train drivers by something altogether quite different actually be better?
Service Management and the Issues Raised by having On-Board Staff
We must now examine the thinking behind why services may be more reliable without drivers. Train drivers, through no fault of their own, are a constraint on train operations. It comes as a surprise to many that Underground trains run to a timetable. Given that many lines are wholly dependent on minimizing service intervals and (above all) providing trains at closely managed regular intervals, the apparent distraction of a timetable may seem odd. In fact even simple lines usually have numerous destinations, either for reversing some services short or where there are branches, and with anything up to 90 trains on a line some kind of detailed service plan is essential. The plan must do many things and one of these is to make sure that the correct trains stable in the correct sidings when they come out of service—these must exactly match those scheduled to start the following day and meet the all the maintenance requirements. It is not presently feasible to consider wholesale dynamic timetabling to meet every individual part of the service specification, so timetables are going to be around for a while yet (but maybe not for ever).
The timetable represents the ideal service. There are, however, any number of reasons why individual trains or groups of trains will occasionally depart from a timetable. Most obviously, there will be a failure of some kind, bringing a train to a halt. This has two effects. First it is vital to avoid sending more trains into the incident area, or one soon ends up with trains full of people stuck in tunnels, which is very undesirable, so trains have to be turned short or diverted at short notice. Secondly, it is necessary for the service controller to provide as regular service as possible over the rest of the line, improvising as well as possible; this is often quite difficult without clear knowledge about how long the original failure is going to take to fix. Clearly, in these circumstances, it doesn’t take very long before many of the trains are hopelessly out of course with the timetable. Once the incident is cleared it is the service controller’s job to get the trains back on schedule as quickly as possible, whilst at the same time avoiding lengthy gaps anywhere, depending on demand at the time.
Getting the service back on book is a bit of a black art, but generally involves minor retimings to trains not too far out of course, turning trains short (or diverting them) to get them on time for their next trip, or ‘reforming’ them, which means turning a train that is hopelessly off schedule into a different one that is supposed to be in that position at that time. Short term pain while this is going on has to be balanced against the knowledge that the scheduled service will be the optimal one. The theory is simple enough and the trains don’t complain when they are ordered about.
But it is not that simple! Each train has its driver on board. Every driver has their own ‘timetable’, called a duty schedule. Train drivers have a fixed start and finish point, a definite start and finish time, and are entitled to a meal relief during their shift, also fixed as to location and time (and variable only within limits). The schedule defines which trains they are to work, so that every train in service is allocated one and only one driver at all times. The schedules allow for one driver driving three, maybe even four, different trains during a shift to maximise their driving time. The corollary is that a train in service all day might see five or six different drivers. The relationship between the train timetable and the staff duty schedules is complex and sophisticated: all is well as long as everything runs to time.
The duty schedules allow for a small proportion of spare staff (either spare for a whole shift or part of a shift) in order to cover for the unexpected. Obviously it is any organization’s objective to keep spares to the minimum feasible since, on a good day, they are sitting around on full pay with no work to do. One cannot do without them though.
I hope it will be obvious from what has been said that when a service becomes disrupted the realities of reorganizing the trains are heavily tempered by the need to reorganize staff deployment on the fly. If a train is reformed, the driver will now be on the wrong train, so what happens next has to be carefully considered since it involves changing over at least one driver, and sometimes swapping two. This is tricky, as a mistake will leave a train stranded, blocking everything else.
Sorting out the drivers can be a huge logistical problem. If trains are delayed, this seriously eats into the productivity of driver’s shifts, and this cannot be recovered. When delays are small, the shortfall in driver productivity can be overcome by deploying spares to cover the gaps, for example where a train lacks a driver who might be stuck on a different train some way behind. When delays are large, there are not enough spares to do this and trains must be cancelled or turned short to match the effective staff available. This is very difficult to do perfectly, leaving services running with reduced service and potentially large and very conspicuous gaps. Recovering from delays is demanding of management time and provides suboptimal services whilst being done; it like this largely because trains are staffed. In fact the problem of sorting the staff out gets progressively more difficult as a delay gets longer. As a rule of thumb I would suggest it is not unusual for delays of ten minutes or more to take between six and eight times that long to recover from.
The purpose of explaining this is to set the scene for the following observation. If a fully automatic line is supposed to have (say) a 75-train service and there is a disruption for say half-an-hour then those 75 trains will still be available even though many might be in the wrong place. One could conceive that (in theory) a service controller could just press a button and the train control computer could instantly ‘reboot’ the service with all trains renumbered to their correct timetable position and the service immediately back on book. Actually, this is probably going a bit far as passengers already on a train would want to complete their expected journey, but I hope you get the point, a service on a fully automated railway could be back on book within half an hour or so if the only thing to worry about is train positioning.
I hope from this you can see why comments are made that if one could ‘get rid of’ the train drivers one could make the train service a great deal more resilient. Let us suppose that to be true, and let us suppose that in conjunction with the points made earlier, technical solutions are found for all the other issues. There is now just one problem.
Boris’s commitment to going ‘driverless’ was tempered by another promise: he stated that trains will continue to be manned. If this is taken at face value, it is obvious that (whatever the person is called) staff schedules will still be needed in order to ensure one train, one captain throughout the service day. This obviously creates a big question around where all this new flexibility is going to come from. One might almost think that the entire point of ‘driverless’ trains is then lost! I do not say that there should or should not be drivers, I simply observe that the arguments heard so far seem to be inconsistent and not thought through. It is a complex area and TfL hasn’t got all the answers let alone the political classes, happy to make promises that are not called for delivery during their tenure in office.
The art of the Possible?
What could be done, and done soon? These feel do-able (providing actual benefits can be assessed and the infrastructure can be developed):
  • Automatic door opening at stations;
  • Automatic train reversing in sidings and at terminal stations, this would improve capacity and utilization (in former case train unmanned, in latter driver/captain can walk through train in motion – incidentally this was tested in 1980s and worked);
  • Limited no-person operation in open air sections only, maybe at quiet times only, and with roving staff trained to drive: this would test concept as a first stage;
  • Train despatch from platforms.
  • Consideration about having no-driver trains in double line tunnels, possibly with alternate trains manned or similar precautions. It is possible to imagine this being extended to Jubilee Line east of Westminster where there are platform edge doors, a tunnel walkway and numerous escape and intervention shafts.
Beyond that things get a lot more difficult. Personally, whatever you called the person in charge of the train, because of the reasons set out already I’d be very reluctant to get rid of the driving cab, even if at quiet times the train captain was out and about. But really, in the peaks I just cannot see a roving brief working on a packed deep tube train and the person is probably better off in a cab.
I think I have raised enough issues without embarking on industrial relations and human factors complications and concerns, but clearly there are interesting times ahead and whatever Boris says now he probably won’t be around in a decade or so when all this might fall to be delivered. He can therefore come out with these brave proposals from a position of comparative safety (actually he’s only promised to produce a ‘timetable’ for change).
Nevertheless with technology marching forward it feels wrong to expect there to be drivers on trains for all time, so I think it is about when and not if.
The TfL view
The managing director of London Underground, Mike Brown, stated in November 2011 (Evening Standard 15 November 2011) that he thought ‘driverless’ trains were perfectly safe (he had visited some) and that TfL should plan to introduce them within a decade (so we know what Boris has latched onto). Details of how this would be done were not mentioned though there were some comments about the more general advantages of automation.
Mike Brown was speaking, perhaps before he was expecting to, following the leaking of an internal paper that had raised the subject. ‘Driverless’ trains were proposed for what is known as the deep tube upgrade; this will affect the Bakerloo and Piccadilly Lines first. The words actually used were:
‘As the requirements of the role are automated, it is inefficient to continue to maintain the competence of Train Operators in activities which are not required in the role, resulting in excessive training and competence management. The train operator role should be refocused on the customer.’
‘The deep tube programme affords an opportunity to introduce sophisticated rolling stock to the Bakerloo and Piccadilly Lines that will largely self operate. The introduction will provide opportunities to fundamentally change the role of the train operator. Rolling stock is unlikely to be manufactured with a separate cab, shifting the role of on-train staff to become fully customer facing.’
Some specific points were made, that readers might like to weigh up against what I have already said:
  • No ‘in cab’ train operators. Introduce ‘Train Attendants’, refocusing the role to customer service and revenue (possibly a cross functional role across trains and stations).
  • Remove the requirement to staff each train, allowing some unattended operation and reduce the overall cover levels, balance customer needs with operational efficiencies.
  • As technology is proven, further increase the level of unattended trains and cross functional nature of the role.
The document this comes from, dated July 2011, is prominently marked as representing ongoing work for discussion and that it does not represent formal views or policy. It was marked confidential, but ASLEF obtained a copy that was promptly leaked.
There appear to be over 20 cities around the World using driverless trains already (effectively no staff necessarily on board). These are substantially new systems (and mostly seem to have driving cabs). Few are comparable with the London deep tube and its particular challenges, though technology will obviously march on and may or may not provide cost-effective solutions.
In the run up to the election, mayoral candidate Boris Johnson has pledged to have the first driverless trains running in service on the Bakerloo Line ‘by 2021’; they will be operated by ‘train captains, like the DLR’ (Evening Standard, 26 April).
The Standard went on: ‘Mr Johnson said he would publish a feasibility study into the project – looking at the costs – immediately if re-elected. He would then put the design phase out to contract in late 2014 and the rest of the signalling contract in summer 2015.’ It is interesting to see that, apparently, no feasibility study has yet been carried out but that the decision to push on appears to have been taken by him already.
Nobody has set out what Network Rail’s opinions might be (half the Bakerloo mileage is over main line tracks) about operating ‘driverless’ underground trains over their tracks, shared with full sized main line trains and creating huge platform gaps, some on nasty curves. There are discussions about the future of this route anyway, but it won’t make an already difficult challenge any easier.


See Modern Railways – September 2012 issue!

About machorne

I have always lived in London and taken a great interest in its history and ongoing development. This extended into the history of its transport services, about which I have written a number of books - I have spent most of my working life working in the industry and observing changes from within, mostly to the good, but not always so. I continue to write, and have a website with half finished stuff in it so that it is at least available, if not complete. Several new books are in hand. I have many 'works in progress' and some of these can be found on my website; the we address is http://www.metadyne.co.uk
This entry was posted in Uncategorized. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s