Location Referencing

How to properly provide geospatial data

Photo by JESHOOTS.COM on Unsplash

One of the first spatial concepts we learn when we are children is the relative position of an object, like above the table, under the bed, in front of me, behind my mom, next to the chair. During our life we refine these concepts by introducing the notion of cardinal points, the idea of a reference system and coordinates (first of all to play battleship), until we understand their practical usefulness in giving road information. Imagine you are in a city you never have been and you do not have any map, neither digital or paper. You are looking for a museum and so you ask information to a person for locating it and going there. How often do you believe you will receive good, sharp and reliable indications? Despite the apparent simple question, even if the person you stopped knows perfectly the city, it is not assured he will be able to allow you to go easily to the museum because his information will be confused and puzzling.

In our digital world geo-location of any information is so important in each field, from business to politics, from science to engineering and so on, that we give it for granted. But providing properly geo-spatial data is not as immediate as can be thought. Traffic Infomation Systems (TIS) are paramount examples of how crucial is the accuracy in location referencing. In such systems the base goal is to be able to transfer traffic information from a producer actor to a consumer actor in a way that is reliable and compact. The location of the traffic information can be a point, a polyline or a two-dimensional shape on the surface of the Earth, while the Location Reference (LR) is a representation of such location. I’ll describe the most common location reference types, but before doing that a brief historical excursus is given.

RDS-TMC, TPEG, and TISA

At the end of the 80s, thanks to a founded European project called DRIVE (Dedicated Road Infrastructure for Vehicle Safety in Europe), there was the first detailed technical proposal at European level for a RDS-TMC protocol for broadcasting in real-time traffic information and weather data.

Radio Data System (RDS) is a communications protocol standard for transmitting digital information via conventional FM radio broadcast. While Traffic Message Channel (TMC) is the protocol defining the content of the traffic data to be transmitted via RDS in a compact digital form. After several proposals the final TMC protocol adopted was the ALERT-C (name inherited from a previous project called RDS-ALERT, and being the C the third specification after A and B).

A no-profit organization, TMC-Forum, was born to discuss traffic information related matters and maintained the TMC-Standard until 2007, when it merged with another organization (TPEG-Forum) giving life to the Traveller Information Services Association (TISA). The Transport Protocol Experts Group (TPEG) is a set of data protocols richer in term of content with respect to the TMC and that can be carried over different transmission media and not only over RDS. TPEG was standardized in a first form as TPEG1 (today deprecated and considered only in legacy systems), but the current ISO standard, supported by TISA too, is the TPEG2 specification.

Both TMC and TPEG technologies have worlwide coverage as you can see from the TISA web page (TMC coverage and TPEG coverage).

Location Referencing by sharing same map

The simplest way to transfer geo-spatial data between two actors is when they both have the same identical map in their hands. Despite allowing 100% accuracy, there is a huge downside: updating a map and sharing it among all actors is really cumbersome because it requires high coordination, and it implies a vendor map lock-in. For such reason this solution is almost never used.

Location Referencing by pre-coded locations

The first alternative, also from a historical point of view, to the sharing of the same map is the usage of pre-coded locations. It means that the two actors agree on a set of locations codified in some way and fixed until a common agreement of updating them. Each actors has its own map and he has to build for each pre-coded location the corresponding set of links of his map. In such a way when the actor A has to communicate an information located on his links to the actor B, thanks to the mapping to the common shared pre-coded locations, the information can be easily transferred to the links of the other map.

This location referencing schema is fairly simple even if it requires to build a mapper function between the pre-coded locations and your map, but it is not fully satisfying for the following reasons:

• the pre-coded locations usually do not fully cover the geographical area, so some information can not be transferred
• the pre-coded location descriptions can not have a high level definition and it implies that the mapper function has some degree of freedom in mapping the links of the map to them, causing lower accuracy of transferred information
• some pre-coded locations can not have a mapping to the links in one or both maps due to the map details, and it prevents the possibility to transfer any information
• when there is a map version update or a map vendor switch, there is the need to rebuild/update the mapping between the pre-coded locations and the new map. This is a costly operation that requires some time and in the meantime there will be an outdated provisioning of geo-located information
• when there is a pre-coded location reference update there is a need, as in the case of sharing the same map, of coordination among the several actors for rebuilding/updating the mapping to their own links at the same time, and it can be challenging

The master example of pre-coded locations is the ALERT-C location tables used to broadcast traffic information via RDS-TMC.

Location Referencing by dynamic locations

Another way for location referencing, designed to overcome the struggles of the previously described methods, is the Dynamic Location Refecencing (DLR). It is conceived to be a map agnostic methodology, in the sense that

• the maps of the two actors are totally indipendent (both in term of vendor and version) as for the pre-coded location referencing schema
• the maps has not to be shared among the actors and there is not any middle layer acting as bridge between the two maps as for the pre-coded location referencing schema

There are several DLR methods:

1. OpenLR
2. TPEG-Loc
3. AGORA-C
4. ULR
5. GIMME

The key common idea of all of them is their capability to encode a location on a given map in a way its information content can be decoded on the other map with a high confidence of success rate. This last one must be intended not only as the correct identified location, but also as the capability to correctly understand that it is not possible to identify a location on a map to the other one. The encoding and decoding procedures rely on algorithms which are specific for each method, and that I will describe in details in future posts.

Traffic data providers like TomTom, INRIX, HERE, use among their location referencing methods also the DLR one, becoming today a de facto standard.