Is it just me or has the bus real time information system got a whole heap worse since its supposed ‘upgrade’ over the past few months?

The bus real time system has always promised more than it has delivered. When it has worked, it is incredibly useful as you don’t have to always keep an eye out for the bus coming along, you have some sense of comfort that the bus really is actually coming and – if the bus is far enough away and there’s something nearby – you can pop into a dairy or a bookshop or wherever in the meanwhile.

Certainly the system has had problems for many years – this is what Brian Rudman wrote in 2009 as an example:

It was a bleak and stormy night when I reached the bus stop buried into the side of the TVNZ headquarters in Victoria St just after 7pm. The timetable on the pole said the next – and last bus for the night – to Herne Bay was at 7.10, so by my watch, and the electronic timetable clock it seemed I was in luck. Down at the other end of the cave-like stop were half a dozen raucous itinerants, one of whom was dicing with death, dashing out into the busy road trying to wash windscreens. A little comic relief.

By 7.15pm I should have suspected all was not well. The electronic timetable had made no mention of an 005, just listing a fleet of Link buses, the next due in 10 minutes.

Then a glimmer of hope lit up the sky: 005 Westmere DLY. DLY is short for delayed.

Perhaps I hadn’t missed it after all. It was just late. Nothing unusual about the Metrolink service there. The Link duly arrived. Did I jump on board, then walk from Ponsonby Rd in the rain, or did I wait? I blame the couple of drinks I’d had beforehand for my stupid decision to let the Link go.

The electronic helper kept reassuring me 005 was DLY until just before 7.30 when it just disappeared. The next Link was now 28 minutes away – so much for the 15-minute gap – and the rain had taken a break, so I started walking, muttering like a crazy man about lying real-time indicator boards. My head swivelled hopefully every time I thought I heard a bus approaching. But not one did in the half-hour walk up College Hill and down to Herne Bay. Nor could I persuade a taxi to stop.

The next morning I checked the bus timetable at Victoria St to ensure I hadn’t misread it in the rain and gloom. But there it was, departure time 7.10pm. It was only then I spotted at the bottom of the page, about the level where a dog would pee, the small print saying all times were “approximate”. Investigating further, I found the official bus timetable says the bus leaves the downtown terminus at 7.05, but that was only an “approximate” time as well.

What really annoys me is having let myself be suckered by the electronic timetable yet again. From 2003, Auckland City Council spent $7 million on setting up and trying to make this flawed system work and failed.

Three years ago, the city flicked the lemon on to the Auckland Regional Transport Authority for a token $1. I wrote at the time that it was a dollar too much.

ARTA claimed it could fix it and expand it out across the region over four years at an added cost of $17.4 million. At the time, ACC officials claimed they had managed to cut the initial error rate of up to 30 per cent down to 3-4 per cent. But a letter to the Herald earlier this year suggests that nothing much has changed.

The boffins blame the bus drivers. The multimillion-dollar system uses an on-board global positioning system, bouncing messages off satellites in space, to predict the timing of more than 730 buses in the Auckland network. The fatal flaw is, it depends on the driver to log on to the system at the beginning of each trip. If he/she doesn’t, then 10 minutes after the scheduled start of the service, the real-time board announces it’s been DLY. Which is a lie, particularly if you’re standing at the first stop after the starting point. It should read, SUCKER, LEFT 10 MINUTES AGO, START WALKING.

An ARTA spokeswoman says the driver on Wednesday started “on time” and they’re investigating why the bus didn’t show up on the system. As cold comfort, she added that “we are in the midst of planning a new, better, real-time system, the current one obviously has its limitations”.

Unfortunately we’ve been hearing that since 2003.

Over the past few months there has supposedly been a major upgrade to this system, with new (presumably quite expensive) equipment and even a different (although more confusing in my opinion with stars and the inclusion of both scheduled and “real” times) layout on the signs.

And there have been the stupidly impossible to read new signs which force you to stand about 3mm away from them in order to find out when the next bus is coming (completely eliminating one of the purposes of the signs in the first place, which lets you check when the bus is coming while still sitting down and reading a book/checking emails).

But it’s the unreliability of the new signs that has really driven me around the bend in recent times. A couple of examples:

  • One day waiting for the Link bus after seeing that I’d just missed one, the sign said the next bus was 15 minutes away. So I settled down and dug through my bag for a book, just noticing a second later when another bus ploughed past without stopping.
  • Another day I showed up at my regular stop to see that no bus was due for another 16 minutes, then two minutes later one randomly turned up.

A problem with the “old system” was the bus drivers tended to forget to “log in”, which meant that the bus was tracked according to its scheduled time rather than a real time, or wasn’t tracked at all. This is what led to the “DLY” showing on the sign – apparently that meant it was more than 10 minutes since the bus was scheduled to go past the stop and the system had no clue whether it had done so or not.

I’m not sure whether this has been resolved in the new system, but it seems not – pretty strange since I’m sure buses need to log in at the start of all their runs for ticketing purposes. Perhaps it’s something deliberate so the bus reliability stats can continue to put tin pot dictators to shame?

The upshot of all this is the absolutely insane situation where a supposed upgrade to the real time system has actually made things worse. I simply cannot comprehend why it’s so difficult to get this right – every bus has GPS tracking for one reason or another as far as I know (the buses I catch regularly and aren’t picked up by the system are still able to tell me the next stop all the time). So I guess the answer must be sheer incompetence somewhere in the system of making this happen, or the fact that it’s becoming increasingly clear that Auckland Transport doesn’t care much about its PT customers.

Share this

47 comments

  1. “Perhaps it’s something deliberate so the bus reliability stats can continue to put tin pot dictators to shame?” sorry every time I see that comment I have to seriously laugh (as it is a form of relief given the situation there)

    Quoting you again: “The upshot of all this is the absolutely insane situation where a supposed upgrade to the real time system has actually made things worse. I simply cannot comprehend why it’s so difficult to get this right – every bus has GPS tracking for one reason or another as far as I know (the buses I catch regularly and aren’t picked up by the system are still able to tell me the next stop all the time). So I guess the answer must be sheer incompetence somewhere in the system of making this happen, or the fact that it’s becoming increasingly clear that Auckland Transport doesn’t care much about its PT customers.”

    You know where my post is about ” Auckland Transport and its PT customers.” and you definitely know where to find me if you ever want to go for coffee somewhere and exchange errr “rants”

    But from reading the post I gather you had a least then desirable experience with the bus information system again?

    So the question and being bloody obvious – what to do next?

  2. “every bus has GPS tracking”

    Exactly. Can someone ask AT why a driver needs to log in before the most basic software can compare with known positions of bus stops and tell whether it has started its journey?

    1. Simply because it is not possible to have the system automatically determine which service a particular bus is operating, in a reliable manner. The driver needs to enter excatly which route and which service they are operating to ensure system links the correct bus with the correct service/route. I don’t see an issue with requiring the driver to log in — NZ Bus drivers generally seem to do this correctly most of the time. I’ve noticed more issues with other operators however.

    2. All buses start their day at a depot. On the journey from that depot to the start of their route they will pass many bus stops, sometimes even travelling a path that looks remarkably like a bus route even if that’s not the route the bus is taking. Trying to work out whether the bus is on a route or just travelling to a start location, given that no particular bus is dedicated to a particular route, would be much more prone to error than requiring drivers to login.
      What *should* happen is that whoever handles allocation of buses and drivers inputs the bus number and the route into a system that then updates the GPS tracking system. Take the driver entirely out of the picture.

  3. Other countries have real time systems that seam to work well. Doesn’t work in Auckland or Wellington! Wellington was behind Auckland, has a different supplier and is still no better than a paper timetable! NZ doing PT on the cheap again!

    1. I thought Wellington was using the same supplier as Auckland? Wellington’s system is good but suffers from the problem that NZ Bus drivers down there are hopeless at making sure they’re logged in correctly (which is an interesting contrast to NZ Bus up here in Auckland). I was just in Wellington for two weeks and used buses every day. Only saw the real time system provide real time info for a handful of buses — and usually my bus was not one of them! 🙁

    1. A check of a random stop on Ellerslie-Panmure Highway indicates that at least some Howick & Eastern buses do have GPS. Possibly you are looking at the older signs which showed an estimate if a specific service was not logged into the system.

    2. Yes they do and have done for most of the time that I caught them – my buses used to nearly always appear on the system. Drivers had to log in, though, which some didn’t do when they were late.

  4. The Passenger Information Displays (PID’s) on Melbourne’s Smartbus system seems to work very well. The bus information seems to be accurate, and at train stations also provide similar information about train departures. I believe the first system Melbourne employed was a lemon, but it works well now. The accuracy of PID’s should be reported monthly just like on-time running, with contract payments tied to its accuracy. If AT are looking or an overseas benchmark, Melbourne would be a good inclusion.

  5. The main annoyances with the “updated system” is that it sometimes takes ages to scroll back to when the next three buses are due, and as you point out, if a bus fails to be tracked and is also late (coincidence??) then it disappears off the display as soon as its scheduled time has passed. (I would prefer the time to remain up with a “?” in the realtime column up to ten minutes past scheduled time).

    Perhaps time to LGOIMA what bus reliabilty is, as measured by the realtime system..

  6. You are correct -the upgraded system is very broken. I have had exactly the same experience of missing buses last week because I have believed the information, so not paid attention to watching the actual street view where buses sail past contrary to the information. The street signs are difficult to read -and of course only facing one way so if you are crossing for the buses say at the Symonds St / Grafton Bridge corner, you can not see how fast you should run as you can only view the signs from the bus stop itself. Too many bus stops still have no information at all. Sometimes the app doesn’t seem to work. And last week some buses sailed past even though I tried to signal the driver.
    AT can smugly claim bus data shows 99% or whatever of buses are on time. It’s a lie that needs constant exposing. I won’t mention my Link experience on Friday night. Suffice to say I am over it. The experience of waiting ages for a bus then 2 or 3 turn up together is one of life’s great mysteries.
    None of this has been my experience in other modern cities.
    And we wonder why Aucklanders are turned off PT.

    1. I hope someone from AT is reading this. They need to come out with a statement PDQ and let people know what is happening and what is being done to rectify it. People within that organisation need to start coming under some real pressure to perform or be replaced with people who can get the job done. I know there is a lot being done and that people are doing their best but I’m sick of excuses and there is risk of some real long term damage to the use of PT in Auckland.

  7. Can I add a comment here, which I see quite often from people like Rudman over how they think GPS works, to quote him:

    “The multimillion-dollar system uses an on-board global positioning system, bouncing messages off satellites in space, to predict the timing of more than 730 buses in the Auckland network.”

    GPS receivers whether used by your smartphone, your in-car navigation system, NZ Bus or Timbuktoo Transport do not “bounce” signals off any satellites – ever. They merely pickup extremely accurate time signals and other data from multiple orbiting GPS satellites at once and using some mathematics, deduce where on the planet the GPS receiver must be in 4 dimensions. End of story.
    Direction and velocity information are determined by the GPS receiver by comparing the GPS co-ordinate difference between adjacent GPS readings, then using some more maths to work out how far and which direction you’ve travelled.

    So, please Rudman – I know you read this column – please don’t hide behind the techno-babble that shows you think GPS works like some kind of aviation radar does. The only “bouncing of signals” that happens is over the on-board bus data link back to AT “HQ” where the data is recorded and used to update to real time displays along the rest of the route.

    Now, for the matter at hand – crappy real time bus info. I agree its got way worse recently, not better.

    And I have a hard time understanding why its so difficult to design a real time bus system that requires bus drivers to “log in” to the system their current route number in the first place and thus when they don’t update this information – the system can’t handle it.

    Every bus I see has a system for handling this already – on the front of the bus are two bits of information which are usable for this.
    The first part is the service/route number – this information is intended for PT users to know which bus to hail. Yes this information can change as buses move around on to different routes but the driver already keeps this up to date as his passengers won’t be able to identify the bus to use otherwise. And drivers know this, so they make a point to change the route number over as they change routes.

    In addition to the route, each bus has a unique bus identifier already – its called a vehicle registration plate aka “number plate” – that identifies this bus from any other bus on the roads in NZ – it doesn’t change very often – if at all.

    Both very low tech I know, but there you go.

    So, using the on-bus GPS receiver its easy to let the real time system know where a particular uniquely identified bus is, along with what route the display board at the front shows – and so it should be a no-brainer for the real time system to work out via a series of semi-regular data messages from the bus over the GPS/Cellular network where the bus is and as the system also knows what route its on it can then work out what the expected time to arrive at future stops will be given what time it passed the last stop.

    Not that hard is it really? That is unless the bus are using some unreliable pigeon post method for sending the bus location and route information to HQ – and given the way it works now, pigeon post might be more reliable.

    When we get EMU’s I expect the same automated tracking system to work then too, the only slight issue is that trains sometimes go into tunnels where GPS doesn’t work. But thats not a unique problem to Auckland is it and if the MTA in New York can solve it for the NYC Metro, then I can’t see why it can’t be solved here.

    1. NYC Metro spent a lot more on the system (~US$230) than Auckland has and clearly that has resulted in a very accurate and well working system.

      1. And how much do you think ARTA, AT and Uncle Tom Cobly and all have spent collectively on getting the current shambles which is so disorganised?
        On a per-capita basis or per-bus basis, its probably more than the the figure for NYC Metro thats for sure.

        1. Per capita and per bus is irrelevant and will likely always be higher for a small system such as Auckland’s because we can’t reach the economies of scale possible in NYC. For any system such as this there will be fixed charges that don’t scale based on bus/train number.

    2. None of the technicalities are significant to the problem at hand, but since you mentioned it I will correct you that for speed, GPS generally users dopler shift being more accurate than comparing locations in succession.

      Unfortunately the info on the bus isn’t enough to identify what the realtime system is trying to do, as a route number is not the same as a particular run on a route, ie. it has to distinguish between the 08:00 and the 08:15 on the same route. That said, if it was otherwise not going to be tracked at all it would be much better for waiting passengers to show a “route 123 for unknown time” is about to show up in 5 mins than not show it at all…

  8. The Link Bus network exhibits not only a totally unreliable information system but also totally unreliable travel times confounded by 2-3 minute intervals every couple of stops to catch up with the timetable. It seems as if it’s been designed as a sort of PT sop by AT’s traffic planners just to show that they don’t think about private motor vehicles 99% of the time. Until someone has the political courage to confront this troglodytic way of thinking (by introducing dedicated bus lanes and bus priority traffic signals at least on the Link routes) this will continue to be indicative of the level of service that will always be found in PT in Auckland.

  9. Generally I’ve not had too much problems with the new system and I actuallty prefer it over the old system. At least with the new system I can see when GPS data is not being used — it makes it much easier to make better decisions on whether or not the info is reliable. It’s gotten to the point when I’m at a bus stop still using the old system I whip out my mobile and fetch real time info off the mobile website which will display the info using the new system.

    The most common failures I’ve noticed have been to do with specific operators (Airbus for example has never ever provided accurate info in my experience) and specific routes (such as the Outer Link). Generally NZ Bus routes (other than the Outer Link) has been nearly always accurate so I suspect it may not be the system itself that’s the problem but rather the operators/drivers not using the system correctly.

    1. So why are there two different systems in use – one for the PIDs at the bus stops and another for the mobile phone users?

      Seems to me there must be only 1 source of the data, and therefore unless you have (to copy a software phrase) “one version of the truth” your real time passenger info system is doomed to failure…

      1. Perhaps my choice of wording was unclear. There is one central back-end system which feeds all the public-facing interfaces such as the mobile site, physical signs at bus stops, rail stations, and so on. The difference I was referring to is how different signs present the same informaiton. Signs still using the old software shows buses like this [285 BLOCKHOUSE 5] which could either mean the 258 bus to Blockhouse Bay is scheduled to arrive in 5 mins or is estimated (based on GPS) to arrive in 5 minutes. This was confusing so a currently software is being rolled out progressively so services are presented like as follows [258 BLOCKHOUSE 1205 5] where the scheduled time (1205 in my example) is shown along with the estimated time (5 minutes in my example) when avaliable. Should a bus not be using GPS then the estimated time is omitted so only the scheduled time is shown (i.e. [258 BLOCKHOUSE 1205 ]) which alerts end-users to the fact the bus has not logged into the GPS system. Hope I’ve claraified things — things are certaintly not as bad as you think it is! 🙂

        1. The problem with the updated system is that when a bus failes to get realtime tracked for whatever reason, the your example of [258 BLOCKHOUSE BAY 1205 ] drops off the display at either 12:06 or 12:07 – so if the bus is not being tracked and is late, it’s off the display and then eventually arrives at the stop as a “mystery” bus. So a user that arrives at a bus stop after their intended bus’ scheduled time thinks it’s gone whereas it is still coming.
          Personally I think in this case the display should show [258 BLOCKHOUSE BAY ??] until 12:10 as well as the backend raising a flag about a lost bus.

  10. Easy answer is if the driver “forgets” to log in, their pay is docked to “encourage” them to remember. I say this as a proud unionist.

  11. No only does the real time system rarely work for me it also never works even at stops directly after Britomart i.e. the buses have major issues keeping to timetables when leaving from Britomart. Today I waited 30 minutes for a bus on Albert Street that was coming from Britomart, there was no traffic so clearly the bus left close to 30 minutes late. Meanwhile the board ticked the bus down to 0, sat on DLY for a few minutes and then the service was gone. Meanwhile it turned up 20 minutes later. I would love to see what the reliability stats are when determined using HOP tag on/offs as that will provide the bus number, route and the time you got on an then off – should be relatively simple to for AT to do and I hope this is something they are using the data for. However, why isn’t NZBus being fined for not logging their buses onto the RT system, this should be made part of all PTOM contracts going forward.

    1. Unfourtently there isn’t much that can be done for this particular issue. Real time systems work best when they have at least some information to work with. When the bus has not yet commenced its journey (which was probably the case in your example) obviously this makes any predicition difficult as there is no data for which to create predicitions from. Once the bus has progressed past a few stops the system can then make predicitions from the timing at those stops.

        1. My point is twofold really 1) there ate major issues with buses running on time – this bus was running on a Sunday with no traffic to slow it and it was departing from Britomart and yet was around 30 minutes late. So reliability needs major work.
          2) buses are almost never on the system, I just checked for bus arrivals at a stop on Wllesley Street and could see dozens and dozens of buses none of which had all sort of real time data. This included every single outer link bus for the next several hours. I know all link routes have the ability to provide real time info and always do the same route – so why isn’t NZBus not being dragged over the coals for not bothering to log in the vast majority of their buses?

        2. I think I know which sign you’re referring to (I’m guessing it’s the one near the Queen Street intersection) — this sign has the problem that some Outer Link services starts there (or perhaps they start over the hill at Victoria Park? I’m not sure) and therefore it will only show scheduled departures for those services. This is also the case for some other CBD departure points. For example the departure point for Tamaki Drive buses generally don’t show GPS data which is a minor annoyance for me but on the other hand buses nearly always do start on time so it’s not a big problem.

        3. Yes that may explain why that stop is particularly bad, however, it is pretty normal when checking any given bus stop using my phone that a large proportion of the inner and outer links don’t display RT data.

  12. True, some routes overlap. As others say, the system could automatically relay the on-board route number display and a unique vehicle ID (which the bus company could be responsible for uploading rather than each driver). HOP also presumably has some way of knowing where a stop is.

    Strikes me as poor software design and implementation, especially if similar systems work in other countries.

    1. I think that some overseas systems the company loads the vehicle number rather than the driver, thereby avoiding the problem we have of diver forgetting to load trips. But vehicle allocation in NZ can sometimes be a bit haphazard, possibly because of smaller operations or less resources!

      1. This is easy. Vin number (which doesn’t change), gps transmit id and service. Surely every run has an id # so you put the 3 together and then when you leave the start point (which the system will know based on gps and route #)
        it triggers the start of the service and the updates.

  13. Inaccurate PIDs are costing the bus companies customers, full stop. Last friday the PIDs at the Remuera Rd bus stop in Newmarket indicated the first bus towards Parnell/Britomart would leave in 18 minutes. This was sometime between 6 and 7PM. While at the stop there were at least two people who saw “18 minutes” on the sign, said something under their breath about the state of PT in Auckland, and scurried off. Sure enough, one minute later the ‘unscheduled’ Inner Link shows up from around the corner picking up not five but three people…

  14. Any geolocated system should be able create live predictions based on distance travelled above the sensitivity of GPS readings. Accuracy obviously improves with more distance, but is hardly restricted to only marked stops. Nor does the system have to wait before showing predictions based on previous trips on the same route. High school students could solve this one.

    1. If the service has not yet started its service, it’s hardly a good idea to just make things up by using data from previous trips as an estimate for when the bus might arrive. Once the bus has started its journey the system used in Auckland works just as you describe — it gets better as the bus progresses along the route when it has more data to compare with previous trips to make better predicitions.

  15. I think it really depends where in Auckland you live. When I used to live in Parnell and get the bus to the city, it was pretty accurate at least 95% of the time. It’s pretty worrying that things seem to have gotten worse though.

    Sydney has no bus real time displays at all which I find it incredible frustrating. So at least count yourselves lucky there.

  16. This could be one of the inefficiencies of the bus service outsourcing model. In the days when drivers worked directly for local government, it would be easy to change the driver employment rules so that their login was tied to their pay system. However with outsourcing, it could be that supplying GPS information is not an enforceable part of the contract, because it wasn’t considered important or reliable enough when the contracts were written. The new set of contracts will no doubt have it as an output, but again AT is locked into perhaps 10 years of no enforceable technological advance.

    1. I see no reason why a driver needs any separate log ins. All they should require is a roster of which vehicle they are driving and the service. The driver starts the bus, tags on using their own special HOP card and this in conjunction with the gps data etc starts the computer tracking.

      1. Veooa does something simalar to that at the moment. Currently they input a schedule showing which services each carriage will be operating during the day. But that raises the problem where carriages (or buses, or ferries) will be swapped for various reasons and the real time system ends up tracking the wrong vehicle. This happens at least a few times a day from my observations travelling every now and then on the Western Line. Having it linked to the driver doesn’t help as well if a driver is called off duty for any reason and is replaced with a different driver so you end up with the system also tracking the wrong driver. The system already in use in Auckland for buses avoids those (and other) issues with your proposed ideas.

    1. I don’t think the only problem with the info is drivers logging in, but it seems to be the main issue. If you think differently you are going to need to provide some evidence, or at least a theory as to what you think the main issue is.

    2. How is being told to do perform your job properly punishing someone? That’s like saying we shouldn’t be angry when we all get food poisoning because the person at the supermarket whose job it is to check the chickens are properly cooked isn’t paid very much.

  17. The new bus stop displays in parts of Spain are great. They even have weather forecast information and can be seen from far away, e.g. across the road for example. Given the NZD/EUR exchange rate would have thought Auckland could get a relatively cheap upgrade!

Leave a Reply

Your email address will not be published. Required fields are marked *