.

Geoinformatics_06-2009
article figure 3: a more complicated example of the adas horizon showing the levels of complexity that may be needed for various adas applications. (courtesy of the adasis forum) (point to point) and 5-meter absolute horizontal, with the vertical error being better than 1% of the distance traveled (rise/run). it is telemapics s understanding that the data accuracy specification for navteq s navigation database (a less demanding application than adas) is 5-meters relative (horizontal) and 15-meters absolute, although, to our knowledge navteq, has never confirmed this specification publicly). the model for 3d road data, in general, defines the road geometry by defining a road centerline that is subject to a set of rules describing the most likely vehicle path on a roadway. adas systems need to know the distance and time required to reach the next curve, hill (slope), or other road feature, instantaneously and continuously. two types of accuracy are required of the map data used to calculate the vehicle s position. first, the data must provide a measure of absolute accuracy, (e.g. telling the system where an upcoming curve begins). second, the data must include a measure of the relative or point-to-point accuracy (e.g. being able to tell the system where the vehicle is along the curve as the driver goes around it while supporting the calculation of the distance/time/heading between features). it is clearly the case that collecting accurate 3d road data could be take a great amount of effort and expense. one of the interesting questions is how will these data be captured and compiled? at this stage, it appears there are three major approaches being taken by the leading providers of 3d road data. navteq claims that it has completed the roads in classes 1 and 2 in europe and that the data from classes 3 and 4 are now in production. navteq has indicated at this point that it does not intend to collect class 5 (roads that it classifies as not efficient through routes. ) unfortunately, these are the neighborhood streets where most of us live; the category also includes country roads, minor collectors, roads in areas with few outlets, low-speed neighborhood streets, and dead-end streets. it is likely that the number of passenger miles traveled on these streets and roads is enormous, raising the question of the potential effectiveness of adas applications, particularly safety applications, when a comprehensive road network is not available. intermap technologies intermap technologies (www.intermap.com), a new player in the adas marketplace, is well known in the geospatial arena for its skills in remote sensing. the company provides digital elevation models, orthorectified major adas map database suppliers navteq navteq (www.navteq.com) is capturing 3d road data using vans instrumented with gps/ins. the location of the van is determined by gps with high precision achieved through the use of differential gps. an inertial navigation system (ins) supplies the orientation of the van and adjustment of the gps readings during cycle slips (gross errors that lead to discontinuities in the tracking the path). while navteq s approach is tried and true, it is also expensive and slow. in essence, all roads need to be driven in order to create the 3droads data. figure 4: the adas horizon communicates across the can bus to various adas applications. (courtesy of the adasis forum) 30 september 2009
Powered by fluidbook tm