Hi,

has anybody been working on a navigation system software for the empeg that is comparable to the systems that are available on the market?
I have to decide if I keep my Mk I and add an aftermarkte navigation system (becker traffic star pro, for example) or if i buy a Mk II and wait for someone to develop some navigation software. I am prepared to pay for such a software, I do not expect to get it for free.

Felix

A navigation system comparable to ones on the market?

It's not going to happen any time soon. I'm planning on an Alpine navigation unit at some point in addition to the empeg.

Calvin

Hi.

I would really be interested in developing such a beast if other developers are also interested. I do have loose connections to a company that already implemented such a thing (for psion handhelds) and to a company that provides the needed maps to many nav system producers (amoung them Blaupunkt and the above company). I also have (little to intermediate) experience in unix/linux developments and more experience in C++ (sad but true, mostly under windows).

So if any other developers would be interested in developing such a thing, which would be a commercial product, at least as much as the needed maps are concerned. I don´t have any problem with releasing the source under GPL or any other more or less free license, but the maps are (expensive) payware, so we would have to sell those or rely on a (yet to find) provider for maps of a known format.
A note on estimated volume of such a project: A reliable nav system is usually estimated to be a project of about 10-20 man-years, so I estimate that a minimum of 20 developers is needed to be able to develop that beast in a not too long time (1-2 years).
Anybody interested in developing that system is welcome to mail me. Any possible sponsor is welcome, too. I won´t accept any money though until enough developers are available. Money provided by the sponsors would be used to do the pre-financing of the maps (you get that maps in a rather "raw" format you have to work on) and (if enough funds are raised) to hire one or two developers that do the coordination and merging of the sub-projects. I envision each
developer doing their own sub-projects that range from single functions to complex modules (map conversion/reading, voice sampling). I myself would not accept/get any money until the project is finished (if the software itself gets sold) or not at all (if the software is release under some freeware/GPL license, which I would prefer as it would make life a lot easier).


cu,
sven

Umm actually the maps are very very cheap. the USGS sells a 6 cd set for $400.00USD that maps out every road in the USA. This is the same mapset that 99.9% of all navigation software companies use to make their product. (it's public domain data so therefore you can resell or even give-away the data you bought from the USGS. the only company I know of that doesnt use the USGS/CENSUS data is navtech, and their maps are the worst I have ever used/seen Navtech data is sparse at best where you need it, and is full of errors.

So the argument that the mapping data is expensive is false.. It can be free/had very cheaply (have 10 guys go together to buy the cd set, $40.00 per person!)

Umm actually the maps are very very cheap. the USGS sells a 6 cd set for $400.00USD that maps out every road in the USA. This is the same mapset that 99.9% of all navigation software companies use to make their product. (it's public domain data so therefore you can resell or even give-away the data you bought from the USGS. the only company I know of that doesnt use the USGS/CENSUS data is navtech, and their maps are the worst I have ever used/seen Navtech data is sparse at best where you need it, and is full of errors.

So the argument that the mapping data is expensive is false.. It can be free/had very cheaply (have 10 guys go together to buy the cd set, $40.00 per person!)

All this might be true for the US, but it sadly isn´t true for Germany (and probably the rest of Europe), at least not to its full extend. You can get maps from government institutions that show all roads, and these can be bundled to any software using them without further payment AFAIK. However, they are a bit more expensive than the US ones (about 2000DM/1000US$) and, what is far more annoying, they miss information on one way streets, type of road etc.

What I currently do is trying to get more detailed information on where to get usable maps, how much they are, which format they use etc. For the time being, I try to design a compressed fileformat for those maps and to create a sample map with just a few (50-100) streets in them. However, I´m not sure yet which information should be stored in those maps. Also, there are trade offs for either high speed or small map files (see next post on these two points).

I´m also in the process of creating both a website and a mailing list for this project. And I´m surely looking for some programmers that would like to get engaged in this project. Like I said in a previous post, I´m open for either a GPL project, or a commercial (resp. shareware) product.



cu,
sven

Hi.

You likely already noticed that I´m in the process of getting a navigation software project going. Now while I wait for some other programmers to join the team and for some info on available maps to arrive, I try to find out what data needs to be in the map files. For a start, this is what I planned to far:

My primary concept is hard to explain without detailed graphics, but I don´t have access to graphics software right now, so I´m goint to try and explain it by words anyway.
When going from town to town, you usually have different levels of maps involved: detailed maps for the streets in the origin and destination cities, less detailed maps for the drive from town to town and probably even less detailed ones when driving across state borders. On my good old Psion S5 organizer, I had two different navigation softwares: Street Planner for the high detail in-town part, Route Planner for the drive from town to town. Now what I really missed there was a way to interface the two, there was no link from the high detailed maps to entry points in the lower detailed maps.
Ok, what I would like to build is a system of different level maps (might be included in the same file though that would make installation tricky). The goal is to make it possible for the user to select the detail level he needs for certain areas. He could install the high detail map for Germany, including all known streets in all known towns, or he could install the low detail map for Germany, showing only streets that go from town to town (like main streets plus freeways) plus the detailed maps for Berlin, Munich and Dortmund, as those are the cities he normally visits.
Now, what should be know about streets?

1. their name


2. a unique ID (UID)


3. their sections (either from end to other street, or from end to end)
   
   NOTE: to be able to handle streets that describe a long curve, "other
   street" might also be the next section of the same street.

What do we need to know about each section?

1. a UID


2. their direction


3. wether they are one way or not


4. their road type (small city road or freeway)


5. their speed limit, if known, else guess it from the type


6. their adjunct intersection

What do we need to know about intersections?

1. a UID


2. which streets (street sections) meet at that point?


3. when coming from a specific section: Which sections may I enter?


4. geographic coordinates

Now, how could we identify a possible transition to a higher/lower detail level? Simply by matching UIDs across files would be one way, but we should not need to search through all files, so I guess each data file should have it´s limits (east/west/south/north border) in it´s header, so that we could quickly sort out every file we do not need to serach through.
My idea for searching for the ideal way from one point to another is to do a backward search from the destination to a higher level (lower detail) transition thus finding the last sections of the track, now do a forward search from the source to a higher level, than use those two higher level points for a new search.
Ok, this post got long enough now, so I´m going to stop here. Comments are highly welcome, though this posting should have gone two the upcoming mailing list, if it would exist yet.


cu,
sven

What you have described here is - near enough - close to what I know the Philips Carin nav system does (I worked on it at Philips in Eindhoven a number of years ago). There is a "parsing" mechanism to step through a detail hierarchy of the map supplied from the CD ROM.

Have you considered using the CARIN navigation disks as map sources for the project? These are freely available at CARIN dealers and you can subscribe to an update plan. The format and content is described in various techinical publications available directly from Philips on request. I am not sure what map coverage is like outside Europe; they are however a hierarchy database which allows the generation of vector based maps and are generated by the European Digital Mapping project which is a collaboration between various different national mapping concerns (such as OS in the UK).

One of the few remaining Mk1 owners... #00015

I can acknowledge that the navtech maps suck. I've had one for my Carin (benelux) and it's not even close to the maps from "Tele-atlas"

Walter
-------------------------------------
Reg:1934/Mk1:158-6Blue/Mk2:380-12Blue

All this might be true for the US, but it sadly isn´t true for Germany (and probably the rest of Europe), at least not to its full extend. You can get maps from government institutions that show all roads, and these can be bundled to any software using them without further payment AFAIK. However, they are a bit more expensive than the US ones (about 2000DM/1000US$) and, what is far more annoying, they miss information on one way streets, type of road etc.

There seems to be very clear difference in Europe/US here. It seems that you can buy good quality street-level maps from the US area cheaply but the case really isn't same in Europe.

In Finland, for example, we have detailed maps available but for very expensive price. For a fully vectorized street-level map with all possible information (street name/number, one-way traffic, turning restrictions, road quality, etc) it can cost well over 5000USD.

Though, once GPS navigation gets more used in Europe, the situation has to change. But as far as I've understood, this might still take years. Of course there are also differences inside European countries what comes to map availability and price range, but overall the difference to US is huge.

And on the above I'm mostly talking about street-level maps, meaning that you can do automatic route planning based on the data, so it has to be accurate.

Kim

Hi.

Regarding the data to be used with the software, I am going to contact AND (www.and.nl) and TeleAtlas (www.teleatlas.com) as soon as I can (time is a very limited resource for me currently, will change soon though). Meanwhile, I created a sourceforge project for it, hoping to find more developers there (especially for the database part). Check out http://sourceforge.net/projects/nav-em.

Anyway, since this thread seems to attrack quite a few people it seems (more than 400 reads so far + mailinglist readers), I am trying to get a figure about what people think about the project. So I would be glad if everyone reading this post answered the following questions and sent it to: [email protected]

----------- cut here ------------
1) What is your level of interest? 
   ranges from 1 (hardly interested) to 5 (eagerly waiting for it to come true)
2) If you are interested in getting such a software for the empeg, would
   you be willing to pay for
   a) the software
   b) the map data?
   If so, how much?
3) Are you able to support the development of the software? How? 
   (This might include technical support as well as monetary support.)
4) What are do you live in, what area do you want to be covered by the
   maps?
--------- cut here ------------

cu,
sven

Hi.

An update regarding the forementioned survey. I made it available as a sourceforge survey which is available via http://sourceforge.net/survey/survey.php?group_id=11444&survey_id=10897. The project itself is available via http://sourceforge.net/projects/nav-em/. I´m really looking for help on this project. There are various tasks that need to be done (this list is probably incomplete):

1. Map acquiration (already working on this, still need help though)


2. (Map) database conception (probably my second target)


3. Map conversion


4. route planning


5. GPS connection (this is probably the first part I´m going to work on)


6. user interface


7. speech synthesis


8. ...

Hope to hear from you.



cu,
sven

One thing - you seem to be duplicating the work of Kim Salo. Have you talked to each other?

Rob

One thing - you seem to be duplicating the work of Kim Salo. Have you talked to each other?

You are refering to the GPS part, right? If so: Might be, haven´t talked to im yet (Kim, are you reading this? If so, please contact me via private email: [email protected]). But though I would love to see his work, I´m still interested in doing this myself. It seems to be the smallest part of the project and might be a nice entry project to get familiar with the empeg and its functionality.


cu,
sven

The last I heard, Kim was moving in directions very similar to those you have discussed.

Rob

What is the actual status of this very interesting projects?
The last postings are some time ago and I would like to bring it back in dicussion so the developers get more feedback!

Fabian
S/N 080000540 12GB blue

I assume Kim is still busy with work trying to get May Payne out the door. Once thats released, hopefully he will have more time for the GPS project.

I assume Kim is still busy with work trying to get May Payne out the door. Once thats released, hopefully he will have more time for the GPS project.



Well, that's true, I've been busy with other things lately and nothing too amazing has happened lately. It's basically working with one dataset pretty okay, but as the data is not directly from any TeleAtlas/NavTech CD, it's not that applicable for any other than one specific area, at the moment. I'm planning looking into the TeleAtlas CDs and see if a proper converter to the current internal data format could be made.

I've also experienced some complications trying to analyze the vector map data format so that correct route guidance could always be given. But, what I have heard from other people who use GPS navigation in-car, everyone has said that they give wrong directions at times, so that seems to be more generic navigation problem.

Kim

I hope to use the empeg someday to navigate through the traffic jams :-))
Would be really cool!
And of course I would pay for those programs!
It is to much work to be unpayed!
Kim, are you working with other people on this project or is it one-man?
I don't think I could help you (except you have some really simple things for me to do - nothing like programming better some time-consuming but simple stuff) but propably there are other people around who can help you?!

Fabian
S/N 080000540 12GB blue

Kim, are you working with other people on this project or is it one-man?

So far I've been working on it only myself, and it's likely to stay like that for now.

Kim

btw, Kim, I also have an Empeg mk.II now (Rio Car) and I trust we both live in Helsinki, Finland so if you need any assistance testing or such - I could be of some assistance.

Amarth