City GIS data is usually built in a format where streets/avenues/boulevards/etc are modeled as lines that are geographically located in the middle of the streets/avenues/boulevards/etc. Is that sufficiently precise for big transit operators to manage their buses (real-time bus management and real-time information to users) or do they create their own set of GIS data that gives a more accurate model of bus routes?

If public (or city) GIS data is not sufficient, and given that more accuracy creates larger data sets, where do you draw the line, figuratively and literally? Do you take the 'middle line' from the public GIS data and just apply an offset on each side to represent both directions of a bus line? And if so, is that offset variable between different segments (since an avenue is wider than a street)? Or do you create 'from scratch' a line that is as close to the actual bus route (and one line per direction)?

Not exactly an expert in this field, so I hope my explanation is clear!

@ Jason Baker and dassouki :

I agree that centerlines are sufficient when you are planning a bus route, my question is more from the operating side. More and more transit operators (and in all but the smallest towns, the operator is a corporation separate from the city itself, albeit with close ties to the city administration) are using real-time bus management systems. Buses are equipped with systems to continuously broadcast their position to the control center which takes that information and projects it on a city map. Controllers can then follow every bus in service and make adjustments in real-time as they are needed. The management system can also send to users real-time information on ETAs at every bus stop in the city.

From my understanding of those types of bus management, they take the coordinates of the bus and they match it to the GIS data they have. Obviously, buses don't follow centerlines. On narrow streets, the delta between the centerline and the actual bus position is probably not significant, but most buses take wider boulevards or avenues, where the difference is significant, or at least I expect it is. I'm expecting that difference could create various accuracy issues. For example, when the management system tries to determine if the bus is on-time, late or early, or what information it should send the users about ETA at a bus stop. I'm guessing that transit operators refine the city GIS data to create a new dataset that is more accurate and better answers their needs. I see two ways of refining the data (see original post), I'm wondering which one is most widely used, or if there are other ways to address the accuracy issue of city GIS data.

Do not hesitate to point out something that isn't quite clear in my explanation, English is not my native tongue. ;-)

@ Brad Nesom :

You're correct as far as displaying the information. I'm looking at it from the management system's algorithm level: any information the system calculates and/or provides logically stems from the delta between the actual bus position, given by an onboard GPS, and the expected position. Now, if the expected bus position (the planned bus route if you will) is created by using basic city GIS data (i.e. the planned bus route follows the centerlines of streets/boulevards/avenues/etc) then the delta with the actual bus location can be quite large, which would, I expect, create accuracy issues with the information the management system calculates and/or provides. If I'm correct, how are those accuracy issues addressed? Do transit operator/agencies refine the city GIS data to create a more precise dataset that then allows them to have planned bus routes that are closer to the actual path a real bus would follow in the real world?

The more I explain myself (and thanks to all for making me do so), the more I get the feeling that only a GIS expert within a transit agency could give me the answer I'm looking for.

  • 2
    Could you clarify what you mean by "manage their buses" for us? Typically centerlines are fine for routing applications, but maybe you had something else in mind? Also, it seems like in many, perhaps most cases, that the city would also be the transit operator, so the public data would contain whatever the transit department needed. Is there a particular application or analysis you're interested in? – Jason Baker Jan 13 '12 at 0:36
  • 1
    I'm confused as to what are you trying to achieve? – dassouki Jan 13 '12 at 14:44

Now, if the expected bus position (the planned bus route if you will) is created by using basic city GIS data (i.e. the planned bus route follows the centerlines of streets/boulevards/avenues/etc) then the delta with the actual bus location can be quite large, which would, I expect, create accuracy issues with the information the management system calculates and/or provides.

What kind of delta values are you expecting here? And don't forget the limited accuracy of GPS. In most cases you won't be able to get good enough GPS locations to tell which lane a vehicle is driving on.

As far as I know, you would just take the bus GPS location and match it to the nearest point on the planned bus route. From this matched point, you then calculate the ETA.

Considering all other unpredictable and difficult do predict influences on the ETA, I don't think you'll even notice the "error" introduced by this approach.

| improve this answer | |
  • For lane location usually, loup detectors installed per lane coupled with transponders are used. The transponder is a bluetooth or rf receiver that detects a "sender" signal in the bus/truck. The loop sensor identifies the bus's location. This setup usually costs $75,000 to $150,000 per site and can usually collect close to 25 variables per vehicle. The issue with pinpointing an object that is fast moving is that there still needs to be some form of a mechanical/analog system to pinpoint the bus/truck's location. Add to that the transponder is used to sort the vehicles to identify the buses – dassouki Jan 15 '12 at 16:03
  • @ undedark : I agree with you, I briefly worked in the GPS field in 2001 and understand all the inherent location errors, though there are some ways to correct those errors. But that's why I'm curious to know if transit operators go through the trouble of creating a precise GIS dataset, or at least one more precise then what is readily available from cities and such authorities. ETA is the first calculation that came to mind, but I'm sure there are other elements that could be adversely affected by a large delta. – Mat Jan 16 '12 at 15:02
  • @Mat: For Austria, I can tell you that the don't. – underdark Jan 16 '12 at 15:05

I can think of two uses for gis data in transit management.
1. route pamphlet creation.
2. route mangement.

you woud want either a very simplified or a schematic view of data for the big thick lines shown on the handouts (probably not even done with gis data). Out of scale, porportion, etc.

For route management you want standardized network data that will allow use of tools readily available for barrier, cost (not $ but speed/traffic/impediments, possible re-route.

I don't know the first (ok well maybe the second) thing about Bus systems.
But I think if I were going to utilize realtime bus location information on a map (web or desktop), I would be using something like streetmap, osm, or bing/google to display behind. Don't know if there is another use for the data during realtime.
Maybe you could fill us in.
BTW +1 for completing more info on the question.

| improve this answer | |
  • 1
    you can probably also add logistics (although this could fall under route management) and cost/benefit analysis to the list. – dchaboya Jan 13 '12 at 16:45

Given a GPS coordinate from a specific bus, and the centerline route which that bus operates along, they can simply calculate the shortest distance between the actual coordinate and the official route and essentially "place" the bus exactly on the route at that closest point in order to report "where" it is.

There's nothing that I can see to be gained from having a more detailed route with both sides of every street (and possibly even multiple lanes) digitized.

| improve this answer | |

For the types of management issues you are talking about, extremly accurate information is likely not needed. Even if the GIS system was very detailed, there is no way transit companies can affort GPS that give position to the centimeter, let alone be able to account for all the other random traffic patterns that would make a bus arrival still variable.

Accurate GIS information of routes would be more useful for non-realtime application, and more generally, less GIS type work. For example, having an accurate AutoCAD drawing of a road turn, along with average bus speeds gathered from GPS could be input for a very specialized civil engineering safety assessment software meant to inform policy about maximum driving speeds, or driver training.

| improve this answer | |

Not the answer you're looking for? Browse other questions tagged or ask your own question.