Don't believe everything in an article. This one is especially bad.
The phrase "along with a user's speed" is misleading and not correct if you look at the sources.

In TomTom's case you sign up for it. You get better real time traffic information in exchange for your data which helps generate this better traffic information. As a side-effect of calculating current traffic states TomTom can build a database containing aggregated historical/statistical data. This data is then sold mainly to traffic planners, traffic management operators etc.

TomTom of course pre-processes the data so that there are no individual traces left.
The information they sell is more like:
On I-95 between mile-point A and B on an average non-holiday Monday between 8am and 9am the average speed was 47mph.
This analysis is the value added to the data and this is what the police might be interested in in making decisions on where to set up speed traps...

TomTom would not have sold your GPS traces. They aggregate the data by street link.
It was a gross omission in the original post to not mention that the speed data sold is not individual speed data but speed data aggregated over millions of individual devices.

OK so there's:
1. Get data to check out where most speeders are at what times
2. Install radar traps and fine speeders
3. Profit

But also:
1. Get data to check out where most speeders are at what times
2. Install radar traps and fine speeders
3. Get safer roads.

So it's basically a win-win situation for police and society. Or win-win-loose if you look at the speeders as well...

The GPS measurements will not be used as a basis for speeding tickets but only to analyze the speeding hot spots where above mentioned speed measuring equipment should be placed an when in order to enforce the speed limit.

Time to chip in some information to this somewhat overhyped discussion: - The use of so called FCD (floating car data) is not new. It has been around in ITS (Intelligent Transportation System) circles probably for the last 10 years or more.
- FCD comes from sat-nav devices providing GPS traces which are typically send more or less in real time to a service provider. There is also a offline use case (also used by some TomTom devices), where compiled data is send in batch e.g. after you connect your device to your computer and use the software provided by the manufacturer. In the case of TomTom I know there is a note in the terms of service explaining this. - Similar to FCD, FPD (floating phone data) is derived from mobile phones checking in to different base stations when traveling in cars along a road. - In both cases the first step upon receiving data is to map match the traces to a digital road network. In the case of TomTom this will of course be its TeleAtlas data. - There obviously is a lot of filtering (using mostly statistics) done in this step. - In pretty much all use cases for FCD/FPD the interesting information is not the whole trace but the travel time per individual street link. - The most important online use case is to analyze the current traffic state one looks at the current average speed per link compared to the "normal" speed. If a lot of cars go a lot slower than they usually would, there is a traffic jam. - There are also mobile phone apps using this kind of crowd-sourced data to create traffic information (e.g. waze) - The offline use case is to provide traffic statistics for traffic planning and management applications. This is what the article above is talking about. - For this, TomTom aggregates data out of individual GPS traces. - There might be a theoretical chance that some TomTom employees could be able to track individual devices/persons. - Some devices (I don't know about TomTom) randomly change their ID at frequent intervals to make this impossible. - What was probably bought by / offered to the police was data that would tell the police what the average (or maybe highest detected) speed for any road or a subset thereof is depending probably on different day categories and e.g. per hour. - This exact same information can also be generated by traditional traffic count devices (inductive loops, side radar etc.). Of course the information is only available at the location of the count site and not on most of the road network like FCD generated data is. - Police's role/goal is to minimize speeding by enforcing speed limits using e.g. radar traps. Of course can data like this be used to optimize this effort. - Traditionally generated data from loops etc. is often provided to the police by the road operators (DOTs, municipalities, etc.).

OpenLR has a *completely* different use case than GPX and KML. So you cannot say that it either is better than GPX or that it is unnecessary because GPX already exists.

Many posts above miss the point of OpenLR. OpenLR is not in any way comparable to KML, GPX, GML and other standards for geodata. Its purpose is to allow dynamic linear referencing across multiple different base maps. It is therefore in direct competition to the (currently only viable dynamic location referencing method) AGORA-C ( http://www.ertico.com/en/themes/completed_projects/websites/agora_website.htm ), which by the way is not open and has a pretty big price tag ( http://www.vialicensing.com/licensing/AgoraC_index.cfm ). Let me give you an example of where to use these dynamic location referencing methods: Say a service provider calculates the current traffic state or traffic forecast for a certain region from detector data, traffic incident messages and maybe a comprehensive traffic model. The provider now has data on existing traffic jams and might even have information on optimal strategies to circumvent jams. This data was calculated and is available now for one specific base map, say a certain release of NavTeq street data or a Tele Atlas map or maybe even OpenStreetMap data (unlikely). Now in order to get this information out to the people on the road, the accurate location of the jams/routes have to be transferred to different devices, all of which rely on various map data formats, versions, accuracies and so forth. The process of describing (encoding) a location an a road network so that the exact same location can be decoded on the receiver's side regardless of its inherent map is called dynamic location referencing. OpenLR tries (just like AGORA-C) to accomplish this feat. So, don't worry about another standard for geodata. This is not the point of OpenLR.

One of the main european research projects behind this is CVIS: http://www.cvisproject.org/ . There is lots of documentation already...