| Versions |
  |
|
| Author |
 Topic  |
|
gman
USA
320 Posts |
 Posted - 11 juil. 2003 : 06:39:27
|
I've been having discussions via email w/ Tele Atlas - the map data provider for TomTom USA. I discussed my experiences and expressed my opinions about the routing problems that TomTom USA has.
In order to work on fixing the problems, an engineer w/ Teleatlas has asked me to ask all of you forum members to send in your poor routing experiences with as much details as you can put together. In other words, send in all of the pertinent info regarding the 'bad' route: Starting address, destination address (include counties), describe the 'poor' part of the route - i.e. improper roads used & what roads should have been used, maybe include speed limits for the various roads.
I have already done this for one of the 'bad' routes that I noticed.
The engineer I have been talking to is Frances Stanley-Jones, Lead Technical Support Engineer w/Tele Atlas.
Frances' email: frances.stanley-jones@na.teleatlas.com
If enough people make their voices heard, then maybe TeleAtlas and TomTom will work to get the routing problem fixed, and make TomTom Nav a worthwhile product.
Gpspassion - I especially hope that you will take advantage of this direct line of communication w/ TeleAtlas to discuss your 'bad' routing experiences, since you've been using the software for much longer than the rest of us.
Here is part of the email that I received from TeleAtlas:
"...I am the product support person for our customer TomTom. It is part of my job to determine whether the errors you are reporting are due to the data content or how it is implemented in their product. That said I'd like to get more information on the errors you report below..."
"...Incorrect routing functionality. Can you give me more details of the route you made that was not optimal or incorrect. Also since I see you refer to the gpspassion user forum I'd be happy to invite details from the other users as well. They only need to e-mail me with their information. It is helpful to get named locations (county, city, streets and the problem areas. For example "did not use US 101 or routed off overpass at Smith St). The more the better.
While we are not the direct support for the TomTom product, Tele Atlas certainly has an interest in making the TomTom product as good as possible. I currently work with them to make that happen. Thank you for your interest and help so far.
Sincerely,
Frances Stanley-Jones"
Ok everybody - send in those emails!
- gman |
Edited by - gman on 11 juil. 2003 06:58:37
|
|
|
Ads
|
|
|
gpspassion
93392 Posts |
Posted - 11 juil. 2003 : 22:47:35
|
Great iniative!
I think it would be useful to post the same trips here so that we can check them with other software to see how they compare.
Here's mine:
TomTom: Version 2.24
Start: Intersection of Saratoga Avenue and Moorpark in San Jose, Ca
End: Duartes's Tavern in Pescadero, Ca (intersection of Stage Road and Pescadero Creek Rd)
TT Normal: Routes through Highway 84, a mountain road instead of Highway 92 - 47.5 miles and 80'45"
TT Alt: Routes incorrectly through Highway 17 - 45.1 miles and 91'
Interestingly enough with version 2.22 (build 356), I could have sworn the alternate was correct through Highway 92...I'm going to have to dig it out...
______________________________________
It's all about sharing the info/Tout pour l'info!
Discounts and Assistance/Réductions et Assistance (Club GpsPasSion)
Best "GPS" PocketPC/Meilleur PocketPC "GPS" (affiliate) |
 |
|
|
gman
USA
320 Posts |
Posted - 12 juil. 2003 : 02:25:47
|
Originally posted by qiyang
Can anyone clearify how the speed is treated differently in TeleAtlas and NevTech database? Gman, are you sure TeleAtlas does not associate freeways and streets with link specific speeds?
If speeds are defined only for road types (classes) and the database does not provide a way to "localize" the speeds for inividual links, getting bad routes is a certain thing because roads in the same class can still be very hertergenous. One possibile solution to reduce the impact of bad speed data and choosing a good route might be (1) use generalize cost rather than travel time or distance alone; (2) compare (if rule based) or combine (if scoring based on generalized cost) relative rather than the absolute difference in both travel time and distance between alternative routes .
From a local elementary school to my home, TomTom gives me a route (10.2 miles) consists of a short section of toll highway, an interstate highway, a state freeway, and a major arterial (but only one lane each direction) with lots of hills and curves. This recommended route is 4.5 miles (78%) longer than the route I use all the time, and according to TomTom's calculation, it would "saves" me 5 seconds, compared to the alternative route. The alternative route, which is about 5.7 miles, is happen to be the fastest and shortest.
Problems similar to the one demonstrated in the example above can be easily fixed if TomTom can try the method I suggested in the 2nd paragraph.
Sometimes TomTom's alternative routes were better than its original routes, but it is not always the case. In several tests, the original route TomTom recommended is good ones; in some other cases, both routes are bad. I am curious to know in what conditions it is good and what conditions it becomes bad, but it is hard to characterize. It seems that when possible routes between an OD pair are more or less uniform in road type, the route given by TomTom is reasonable; if there are roads of different road classes, TomTom tends to favor high class roads. But when I compare the default road speeds with that used in Microsoft Street and Trips, they are very similar. Then how could the routes they give can be very different and travel time estimate is way off for the same route? It must have a bug in the code or a flaw in the shortest path algorithm.
For those of us who purchased TomTom and these who are buying it now or later, the initial curisity of a good UI will soon fade; the excitement of its 3D look will also soon become less appealing because the world outside our car windows is a much better real 3D view  . Who care about some flying 3D polygons which do not even have names? Even my 6-years boy lost interest in it after a few trips with me. As a navigator, what matters at the end is to give us good routes, if not the best. I wish TomTom developers will fix the routing problems before more customers become disappointed and decide to drop it from their PPCs. I guess many of us in this forums are willing to test their beta and even alpha versions before it releases the patch. Can someone from TomTom verify that they are listening and tell us whether there is a plan to fix the routing problem soon?
|
|
|
|
gman
USA
320 Posts |
Posted - 12 juil. 2003 : 02:26:53
|
Originally posted by wallstreet123456
Another thing, has anyone heard back from TomTom regarding the routing trouble peole have mentioned here? How the alternate routes seem to be better |
|
|
|
gman
USA
320 Posts |
Posted - 12 juil. 2003 : 02:44:55
|
qiyang:
I am 99% sure that there is no speed data that is actually attached to the individual streets in TomTom. Why else would they have implemented the 'Preferred speed' settings in the Navigator properties? I believe they have only attached a piece of 'road class' data to each street, such as 'Interstate 65 = Freeway'. Really it seems that this is a problem w/ the TeleAtlas data, and TomTom has tried their best to implement a solution given the lack of the speed information.
In order to help out the customers, TomTom needs to provide some description as to what each of the 'Preferred speed' values should be for typical US streets and highways. But, of course, this is still a 'work-around' type solution. I believe there will always be bad routing problems unless TeleAtlas provides speed information for each individual street in their map database.
Navtech has done this for their database, thus you can set Routis, for example, to enable Auto Speed Alerts when you go above the posted speed limit. |
|
|
|
gman
USA
320 Posts |
Posted - 12 juil. 2003 : 02:56:14
|
wallstreet:
I received an email back from TomTom today about the routing problems, although they really didn't answer my question. Here is their response:
"Thanks for your feedback. One thing to suggest Routing wise is that choosing Shortest route rather than Quickest route can help limit tendency to use highways. Route planning solutions of most flavors do see the world as if you are the only car on the road. Unlike us humans, they have little perception of actual traffic conditions, shortcuts around traffic spots or popular rat runs. It sees a highway that is an empty road ahead with a designated speed of 65 mph.
For map quality issues we have set up a reporting program where people can report issues directly back to TeleAtlas. For news on the map discrepancy reporting program see www.tomtom.com
Regards,
TomTom Customer Services
support@tomtom.com
http://www.tomtom.com"
I asked them directly if they are going to work on the routing problems, but they really did not answer this in their response (although one might infer from their reply that its not high on their list of priorities).
I have since emailed them again to let them know that I have tried their suggestions, and the results have basically not changed. I'm curious to see if they respond.
Based on my conversations w/ TeleAtlas, it seems they are more interested in fixing TomTom's routing problems than TomTom is. |
|
|
|
gman
USA
320 Posts |
Posted - 12 juil. 2003 : 03:10:57
|
I pulled the below bad route example straight out of my email to TeleAtlas. I simply pulled up Mapquest.com to get the proper routes (Navtech data - same routes are produced by Routis/iGuidance/Mapopolis), and cut and paste them into my email.
One example would be this route: (using TomTom Nav 2.24 w/ WI_IL_IN map)
Depart: Gradle Drive & Carmel Drive, Carmel, IN, Hamilton county
Arrive: Haverford Ave & Broad Ripple Ave, Indianapolis, IN, Marion county
TomTom tries to use US 31/Meridian St or it may use College Ave as the main roads to complete the route - both of which are not the optimal routes. I should tell you that after increasing many of the 'Preferred speeds' under the 'Navigator Properties' in TomTom Nav, the program stopped trying to route to US 31/Meridian St, however it now tries to use College Ave - which is still a poor route. TomTom travel time 17'20".
The optimal routes should use either Keystone Ave or Rangeline road (Rangeline is the straightest path, is a slower road, but has fewer stoplights). Two examples of correct routes for the given addresses are shown below:
1: Start out going South on GRADLE DR toward W CARMEL DR. 0.03 miles
2: Turn LEFT onto W CARMEL DR. 0.44 miles
3: Turn RIGHT onto S RANGE LINE RD. 0.32 miles
4: Turn LEFT onto E 116TH ST. 0.67 miles
5: Turn RIGHT onto N KEYSTONE AVE/ IN-431 S. Continue to follow N KEYSTONE AVE. 6.21 miles
6: Turn RIGHT onto BROAD RIPPLE AVE/ E 62ND ST. Continue to follow BROAD RIPPLE AVE. 0.76 miles
7: Turn LEFT onto HAVERFORD AVE. 0.06 miles
Total Estimated Time: Total Distance: 8.50 miles
15 minutes
OR
1: Start out going South on GRADLE DR toward W CARMEL DR. 0.03 miles
2: Turn LEFT onto W CARMEL DR. 0.44 miles
3: Turn RIGHT onto S RANGE LINE RD. 0.50 miles
4: S RANGE LINE RD becomes WESTFIELD BLVD. 2.12 miles
5: WESTFIELD BLVD becomes N WESTFIELD RD. 0.04 miles
6: N WESTFIELD RD becomes N WESTFIELD BLVD. 3.88 miles
7: Stay straight to go onto WINTHROP AVE. 0.07 miles
8: Turn LEFT onto BROAD RIPPLE AVE. 0.29 miles
9: Turn RIGHT onto HAVERFORD AVE. 0.06 miles
Total Estimated Time: Total Distance: 7.45 miles
15 minutes
I also sent them saved images of the map data (saved off Mapquest) that showed the correct locations of my starting point and destination, since the locations TomTom has are off by quite a bit.
|
|
|
|
gpspassion
93392 Posts |
Posted - 12 juil. 2003 : 10:12:35
|
Very nice work gman. One can only hope that TomTom will listen once they get over the "company line" stock answers. As usual it's going to take some "noise" for them to listen (see the beta program issue), but then most companies are like that. I really don't know where the problem might come from, since again PocketMap Navigator doesn't have these errors. I suspect the routing engine doesn't only look at speeds, but also has some built-in AI.
Not sure whether TeleAtlas has built-in speed. NavTech does but I haven't heard very good feedback about it as it is only a range, say 30-50 and it doesn't help a lot. They certainly provide road class info for each segment.
It could simply be that TomTom haven't mapped the TeleAtlas data properly or have tried to fit different segments in the same speed, etc...
The worst part of course is that this problem should have been identified during the beta testing, saving buyers valuable time and frustration.
______________________________________
It's all about sharing the info/Tout pour l'info!
Discounts and Assistance/Réductions et Assistance (Club GpsPasSion)
Best "GPS" PocketPC/Meilleur PocketPC "GPS" (affiliate) |
|
|
|
gpspassion
93392 Posts |
Posted - 12 juil. 2003 : 10:12:44
|
Very nice work gman. One can only hope that TomTom will listen once they get over the "company line" stock answers. As usual it's going to take some "noise" for them to listen (see the beta program issue), but then most companies are like that. I really don't know where the problem might come from, since again PocketMap Navigator doesn't have these errors. I suspect the routing engine doesn't only look at speeds, but also has some built-in AI.
Not sure whether TeleAtlas has built-in speed. NavTech does but I haven't heard very good feedback about it as it is only a range, say 30-50 and it doesn't help a lot. They certainly provide road class info for each segment.
It could simply be that TomTom haven't mapped the TeleAtlas data properly or have tried to fit different segments in the same speed, etc...
The worst part of course is that this problem should have been identified during the beta testing, saving buyers valuable time and frustration.
______________________________________
It's all about sharing the info/Tout pour l'info!
Discounts and Assistance/Réductions et Assistance (Club GpsPasSion)
Best "GPS" PocketPC/Meilleur PocketPC "GPS" (affiliate) |
|
|
|
qiyang
USA
21 Posts |
Posted - 12 juil. 2003 : 17:50:10
|
I would like ro comment on the response Gman received from TomTom.
TomTom: ... One thing to suggest Routing wise is that choosing Shortest route rather than Quickest route can help limit tendency to use highways ...
Qi: A route with shortest distance has its own problem. Take a route beween Washington DC and Buffalo as an example. You certainly do not want to cut through the local roads in the middle because it is shortest in distance. You better to use the highways on the west (I-79) or east (I-81). A few days ago, I have written to TomTom to suggest that both travel time and distance be considered in determining the order of the routes to be recommended, using the idea of generalized cost, but I have not got a response from them yet.
We all know that the speed data may not be accurate, therefore, the travel time for so-called "fastest route" is expected to be unreliable. TomTom had given me a route as the first choice which is 4.5 mile (79%) longer in distance but only an estimated 5 seconds shorter in travel time. Is this a data problem? I would say NO. And I can not believe it is not a problem with TomTom's routing engine. As GpsPassion said, some AI is needed in the routing engine (some other popular nevigation software may already have that). It is sad that TomTom did not do it and it sounds like they are not interested in doing it. Please correct me if I am wrong.
TomTom:... Route planning solutions of most flavors do see the world as if you are the only car on the road ... It sees a highway that is an empty road ahead with a designated speed of 65 mph ...
Qi: Wrong again and irrelevant! I am not saying the software is wrong (because it can use the prefered speeds assigned to road classes), but the argument above. The computation of the "fastest" route should be based on the averge (or called loaded) travel time, not the free flow travel time. Everybody knows the road is not empty, why you treat them as empty! Use the average speed to calculate the travel times. And it is better to use the link specific average speed because roads in the same class can still be very different in average speed.
If a navigation system does not have link specific speed data in the database, I doubt its usefulness in recommending you a realistic route because you are asking something they cannot not have.
TomTom: ... Unlike us humans, they have little perception of actual traffic conditions, shortcuts around traffic spots or popular rat runs ...
Qi: That is true. It would be perfect if real-time and predicted travel times (which takes into account drivers' possible enroute decisions after receiving such information) will be broadcasted to navigation device and used by the nevigator in searching for the best routes. But we are not asking for such future features here. What we want here from TomTom is a route that is as good as the one that other navigators on the market are providing. Am I asking too much? |
|
|
|
gpspassion
93392 Posts |
Posted - 12 juil. 2003 : 19:52:02
|
quote:
Qi: That is true. It would be perfect if real-time and predicted travel times (which takes into account drivers' possible enroute decisions after receiving such information) will be broadcasted to navigation device and used by the nevigator in searching for the best routes. But we are not asking for such future features here. What we want here from TomTom is a route that is as good as the one that other navigators on the market are providing. Am I asking too much?
Mapopolis ClearRoute does that, but at a cost...
______________________________________
It's all about sharing the info/Tout pour l'info!
Discounts and Assistance/Réductions et Assistance (Club GpsPasSion)
Best "GPS" PocketPC/Meilleur PocketPC "GPS" (affiliate) |
|
|
|
gman
USA
320 Posts |
Posted - 13 juil. 2003 : 01:31:17
|
I recently sent this email in to TomTom support regarding the bad routing issue, since they told me I should just use the Shortest Route option as a fix:
"I have tried using the route planner w/ shortest route selected. For one of my trips, it did result in the proper route. However, for other trips, it is not acceptable as it tries to route you through subdivisions, down little backstreets, etc. Regarding this, is there not a way to do this while using the navigator mode? Or can you only use Shortest Route while using the route planner function?
It seems the problem is knowing what to set the 'Preferred speeds' to in the Advanced Properties. There is no documentation regarding these settings. What do the road classifications correspond to in US road terminology? We don't have 'International roads', but I assume that refers to an Interstate or US Highway. What is a 'Connecting' road? Or an 'Important local road'? I noticed it has a setting for the 'Destination road'. The speed for that road could vary anywhere from 15 to 50 mph. It seems like that setting alone could throw off the route calculation.
I take it that these 'Preferred speed' settings have been implemented, because TeleAtlas does not provide speed limit data within their map database, correct?
Can you tell me what the proper values should be for the various 'Preferred speed' settings?
Thank you"
Here is TomTom's response:
Our development team is continually challenging the routing technology in Navigator and sometimes its accuracy can be limited by discrepancies in map data. Attributes other that listed speed limits for a road can play a major part in how the routing software interprets the best way to get to your destination. Turn restrictions at junctions are an active part of the data set. For example, a junction in a town nearby has a set of lights that crosses a two lane state highway. There are turn restriction at this junction that the data does not acknowledge and can add or subtract half a mile to the trip depending on which direction you are coming at the junction from. The example you give could be
down to junctions nearby being attributed incorrectly. You can test this by locating a junction and tapping on a road going into it and specifying that point as the Departure point, then tap a road lead out and setting that as the Destination. Finally, ask Navigator to plan a route between these two points by tapping on the A/B button at the bottom of the screen and tap on OK without changing the details. If Navigator thinks it can make turn when you know you can't, or the other way around, then the discrepancy could effect the way that a
route is planned. We do value your input and are working with TeleAtlas to track data quality.
Regards,
TomTom Customer Services
support@tomtom.com
http://www.tomtom.com
Their response is interesting in that they are basically instructing me how to prove that there are both problems with the map data and TomTom's routing engine!? Again, they gave me the standard 'corporate line' at the end, and will not answer my direct questions.
|
|
|
|
gman
USA
320 Posts |
Posted - 13 juil. 2003 : 01:37:08
|
I suggest everybody flood TeleAtlas/TomTom with as many bad route examples as you can, and express your opinions if you ever want them to provide a fix for this 'defective' software that you have spent your hard earned money on.
|
|
|
|
gpspassion
93392 Posts |
Posted - 13 juil. 2003 : 02:31:36
|
Yes while their tone is pleasant enough, it doesn't show a lot of commitment to the problem. Again, it will take time, but when they realize people have noticed that TomTom Navigator calculates routes that are inferior to nearly all competing products, they will probably change the rhetoric. The same thing happened with the "beta test" problems.
______________________________________
It's all about sharing the info/Tout pour l'info!
Discounts and Assistance/Réductions et Assistance (Club GpsPasSion)
Best "GPS" PocketPC/Meilleur PocketPC "GPS" (affiliate) |
|
|
|
ikesler
16 Posts |
Posted - 14 juil. 2003 : 19:20:29
|
I recieved this email re: routing from tomtom this weekend
Dear Ian,
Thanks for your mail. An update has been released that deals with some of the application stability issues you describe. The link below should take you right to the page listing the downloads for this update. Download and install both the application and GPS update. http://www.tomtom.com/support/ce/support/download_navigator2.php?Language=1
Routing...
One thing to suggest, Routing wise, is that choosing Shortest route rather than Quickest route can help limit tendency to use highways. Route planning solutions of most flavors do see the world as if you are the only car on the road. Unlike us humans, they have little perception of actual traffic conditions, road works, shortcuts around traffic spots or popular rat runs. It sees a highway that is an empty road ahead with a designated speed of 65 mph. Our development team is continually challenging the routing technology in Navigator and sometimes it accuracy can be limited by discrepancies in map data. For map quality issues we have set up a reporting program where people can report issues directly back to TeleAtlas. For news on the map discrepancy reporting program see www.tomtom.com
With regard to your feature requests, these two ideas are popular suggestions and our Product management team do acknowledge them. We'll have to see what they come up with in the future.
Regards,
TomTom Customer Services
support@tomtom.com
http://www.tomtom.com
*** Please leave the ticket identifier in the subject of your reply ***
I am going to follow their advice and see if that helps.... it didn't seem to before.
And yes, I already had downloaded the updates for software and gps..... so it must be a canned response.
Ian |
Edited by - ikesler on 14 juil. 2003 19:22:17 |
|
|
|
gpspassion
93392 Posts |
|
|
Topic |
|
|
|
| This page was generated in 1,48 seconds. |
 |
|
- 
- 
TomTom USA Routing: Tele Atlas wants your help
