Streetcar and light rail signals

[muller877]

Commercial, non-military GPS is only accurate to about 7 meters or so, and that number is currently capped by the US administrators of the service. How do you propose that they make it more accurate than that?

Dan
Toronto, Ont.

The 7 meter number is a "worst-case" event. In practise it is generally half of this (around 3.5 meters). This gives sufficient accuracy to do GPS tracking. How often is you phone out of sync....mainly only in the financial district will there be some issues (but easy to overcome and if done correctly will also help the general population at large with Google Maps directions)

 

[innsertnamehere]

I can assure you that you can purchase GPS units that are more accurate than within 7 meters. Extremely expensive, high level GPS systems commercially available for things like surveying can bring accuracy within 3cm. Just be prepared to pay more for one than you did for your car.

 

[reaperexpress]

Thanks for your detailed posts on TSP. There's a lot of misinformation out there.

About your point above, you're dismissing any idea of transit priority serving a purpose of faster transit trips (under normal operating conditions), suggesting that it should only be used to increase reliability instead.

I certainly disagree. TSP should be designed to speed up transit. We have an obligation to do what we can to make transit a superior transportation choice. I do like the idea of the TSP having a more "aggressive" option for when vehicles are behind schedule, though.

Actually I'm far from dismissing the idea of TSP being used to speed up transit service.

It is a fundamental principle of transit operation that vehicles should not operate ahead of schedule (or below the target headway under headway-based management). Regardless of the TSP system in place, early and on-time vehicles will deliberately operate slower than is technically possible in order to stick to the pre-determined schedule and/or maintain an even headway.

The speed of transit service is dictated by the schedule. In a local transit system, the only time you'll experience travel times faster than the schedule is when a trip is late and making up time.

Conditional TSP speeds up transit service every bit as much as unconditional TSP because it has exactly the same effect on schedule determination. Schedules are not based on the typical travel time, but on a travel time that incorporates schedule padding to account for potential delays. With conditional TSP, this padding can be drastically reduced since delays can be mitigated by the TSP system itself, rather than by slowing down the normal scheduled operation.

I believe in LRT and their capacity to provide rapid transit service, however, I have absolutely no faith in the City of Toronto to make it work which is one of the main reason for the scepticism towards LRT in this city

So apparently we have an issue with the city's signal operation, but rather than consciously changing our priorities and/or strategies for signal operations, we should spend countless billions dollars to build grade-separated rapid transit regardless of context.

Yeah, it sounds like Toronto politics to me too.

 

[wopchop]

Commercial, non-military GPS is only accurate to about 7 meters or so, and that number is currently capped by the US administrators of the service. How do you propose that they make it more accurate than that?

Dan
Toronto, Ont.

Sorry, but this is bullshit.

I can assure you that you can purchase GPS units that are more accurate than within 7 meters. Extremely expensive, high level GPS systems commercially available for things like surveying can bring accuracy within 3cm. Just be prepared to pay more for one than you did for your car.

For work, our Leica total station can accurately measure distances to within approximately 1 - 1.5 mm per kilometre, and angles within about 1 arcsecond. It has a GPS receiver as well, and it is accurate in real time* - horizontally to within about 10mm and vertically to approximately 20mm. You can also program it to do longer observations and get that accuracy down to about 3mm and 6mm.
.
* accuracy of real time observations is very dependent on weather conditions - overcast skies, heat haze, etc. but certainly far, far, far more accurate than "within 7 meters"

 

[TheTigerMaster]

Sorry, but this is bullshit.

No need to be rude about it.

 

[nfitz]

For work, our Leica total station can accurately measure distances to within approximately 1 - 1.5 mm per kilometre, and angles within about 1 arcsecond. It has a GPS receiver as well, and it is accurate in real time* - horizontally to within about 10mm and vertically to approximately 20mm. You can also program it to do longer observations and get that accuracy down to about 3mm and 6mm.

It's been a while since I've worked with one, but they used to be dependant on establishing a base-station at a known location, that always took a while to calibrate.

Back in the day with a portable GPS to get sub-metre accuracy we used to hike around with a radio dish, which I think was getting a signal coming from Niagara Falls-way. Not sure if this is still done/necessary these days ... or that was just the vertical control.

 

[innsertnamehere]

I used one quite a bit over the summer, Modern systems involve finding a point with a known coordinate, taking 5-10 measurements while placed on top of it, and then going and doing your measurements.

Older GPS units still need good cloud cover and you can't really use them under trees, but the newer units are a lot better in those terms. Still have some issues, but unless you are trying to get measurements in the middle of the woods, you are usually fine. Oh, and they run on Windows so you can actually play solitaire on them if you wanted to.

This isn't a total station though, my experience with total stations is largely with much older, more primitive systems, missing a lot of the more modern total station features that are available. The company was having troubles finding parts for the SDR33, the stuff was so old.

 

[wopchop]

No need to be rude about it.

Fair enough. My apologies.

 

[Johnny Au]

For the purposes of GPS-based TSP, one-metre accuracy and precision is very much enough, as it is a good compromise between accuracy, precision, and cost. Yes, accuracy and precision are two different things.

 

[muller877]

For the purposes of GPS-based TSP, one-metre accuracy and precision is very much enough, as it is a good compromise between accuracy, precision, and cost. Yes, accuracy and precision are two different things.

Directly from the horses mouth:
http://www.gps.gov/systems/gps/performance/accuracy/

Hope it settles it...there are commercial systems that provide accuracy within 1 meter but it is generally with a secondary source. Irrespective of this, there are commercial systems with inches of accuracy...who cares if it is called GPS or GSP+....with a relatively low cost. Of course the TTC somehow would manage to pay millions for it and then take 4 years to install it...oops that's the subway thread

 

[wopchop]

Directly from the horses mouth:
http://www.gps.gov/systems/gps/performance/accuracy/

To be fair, that is American. Natural Resources Canada does maintain an accessible network that enables very high accuracy.

It's been a while since I've worked with one, but they used to be dependant on establishing a base-station at a known location, that always took a while to calibrate.

Back in the day with a portable GPS to get sub-metre accuracy we used to hike around with a radio dish, which I think was getting a signal coming from Niagara Falls-way. Not sure if this is still done/necessary these days ... or that was just the vertical control.

Same idea nowadays, but without the radio dish. Our station has this saucer shaped component on top of it, which is the GNSS receiver.

In Southern Ontario, the network is dense enough that receivers have access to enough control points to make very accurate observations in a timely manner. If you are in the far north, the network is not as dense, and very accurate observations can take upwards of an hour or more. Like insertnamehere says, you can setup your instrument, link up to the network, and take a half dozen observations to determine your first position. Then you can traverse from that known point.

 

[TheTigerMaster]

To be fair, that is American. Natural Resources Canada does maintain an accessible network that enables very high accuracy.



Same idea nowadays, but without the radio dish. Our station has this saucer shaped component on top of it, which is the GNSS receiver.

In Southern Ontario, the network is dense enough that receivers have access to enough control points to make very accurate observations in a timely manner. If you are in the far north, the network is not as dense, and very accurate observations can take upwards of an hour or more. Like insertnamehere says, you can setup your instrument, link up to the network, and take a half dozen observations to determine your first position. Then you can traverse from that known point.

Is this system practical for vehicles in motion?

 

[wopchop]

Is this system practical for vehicles in motion?

Yes. The system already exists. The major impediment to accuracy is the local availability of reference stations, which is not a problem in this region.

People often forget that modern GNSS receivers are not simply determining position based on the American GPS satellite network, but they are referencing that data against ground station reference points, and making correction. The latter is what ups the accuracy. Vehicle and aircraft based GNSS are obviously more complicated, and are also designed to interface with their own odometers, gyroscopes, etc., which enables them to perform dead reckoning even when the satellite connection is temporarily lost.

Previously, the US Military capped the accuracy of direct GPS readings (meaning, without augmentation from ground reference), but this is no longer the case, and changed around 2000 IIRC.

 

[georgevicbell]

Yes. The system already exists. The major impediment to accuracy is the local availability of reference stations, which is not a problem in this region.

People often forget that modern GNSS receivers are not simply determining position based on the American GPS satellite network, but they are referencing that data against ground station reference points, and making correction. The latter is what ups the accuracy. Vehicle and aircraft based GNSS are obviously more complicated, and are also designed to interface with their own odometers, gyroscopes, etc., which enables them to perform dead reckoning even when the satellite connection is temporarily lost.

Previously, the US Military capped the accuracy of direct GPS readings (meaning, without augmentation from ground reference), but this is no longer the case, and changed around 2000 IIRC.

Guys, I don't think the problem is with the GPS receivers, which are fairly standardized these days...but the transmission of the GPS data from the streetcar to Nextbus, which happens every 30 seconds...and hence the accuracy of where the vehicle is while it may be correct when reported, is then incorrect for the next 30 seconds....a fairly large amount of time for busses that are travelling fast, and totally useless for determining when to change traffic lights.

As for changing the order of phases....I can see this being totally foreign to most suburban drivers....and I often see pedestrians at bloor/yonge confused with the changing turning rule signs...(I've yelled at a driver or two, only to see that the sign wasn't what I expected it to be....and one time it was changing back and forth constantly)...

Perhaps some indication to pedestrians and drivers that an intersection has variable rules ("Signal Order May Change") signs...or something....

 

[reaperexpress]

As for changing the order of phases....I can see this being totally foreign to most suburban drivers....and I often see pedestrians at bloor/yonge confused with the changing turning rule signs...(I've yelled at a driver or two, only to see that the sign wasn't what I expected it to be....and one time it was changing back and forth constantly)...

Perhaps some indication to pedestrians and drivers that an intersection has variable rules ("Signal Order May Change") signs...or something....

Unlike time-of-day turn restrictions, variable signal phasing does not present any new information to drivers. It's the same red, amber and green signals, simply turning red, amber or green at a different time.

Besides, suburban drivers are perfectly familiar with varying signal phasing. For example, advanced left turn signals may or may not activate based on the time of day and the number of cars waiting, and drivers would not know until they actually do (or don't) go green.

And even if suburban drivers weren't familiar with varying signal phasing, what would happen? Do we expect people to start driving through a red light because they expected it to change sooner? No, of course not. People will wait until the light turns green like they do at every other intersection in the city.