[OSM-dev] Project Proposal - Waze Integration

[peter]

On Wed, 2010-03-24 at 21:30 +0100, Eric Marsden wrote:


>   Your proposal is interesting, but given the technical challenges you
>   will face, it seems overly optimistic to me.

Thats true, for sure it can't be done during GSoC. It is more or less a
vision. But I'm not planning to stop developing after GSoC or if such a
project is not accepted in GSoC.

>   I'm also not convinced
>   that many people would be sufficiently motivated to manually report
>   transient information such as "roadworks here",

I thought the same for entering ways etc. before I knew openstreetmap.
But of course I cant tell if people would enter such information. 

>  so I think your focus
>   should be on automated extraction of information from live GPS logs.

That is one important data source, but will not cover all aspects of
traffic information. E.g. when a road segment is blocked for some time
caused by some road work - it is really hard to detect that from
floating car data. 


>   Concerning the matching of GPS logs, I suggest you read the paper
>   "Hidden Markov Map Matching Through Noise and Sparseness"[1] and perhaps
>   prior art from the CarTel project[2]. 
> 
>   [1] http://research.microsoft.com/en-us/um/people/jckrumm/
>   [2] http://cartel.csail.mit.edu/doku.php

Thank you for the input, seems to be very interesting.


There were a couple of people in the last days, who have proposed
similar ideas. Wouldn't it be great to get all this students pull
together? 

regards 
peter