Precise mapping information collaboration for autonomous testing

Bosch is working with TomTom on the generation of high precision mapping data of German roads for testing autonomous vehicles.

Tesla model S automated test vehicleAutomotive technology suppler Bosch has entered into a collaboration agreement with mapping data company, TomTom to provide highly detailed maps of German and US roads based on specifications supplied by Bosch.

So far, the companies have worked together on maps for the A81 road in Germany and the I280 Interstate highway in the USA, both of which are being used by Bosch for automated vehicle testing using a Tesla Model S.

This is an important venture for the two companies, as explained by Bosch’s Dr Dirk Hoheisel: “Only with high precision maps will automated driving on freeways be possible by the end of the decade.”

“By the end of 2015, we want to have new high-precision maps for automated driving for all freeways and freeway-like roads in Germany.” Road coverage will subsequently be extended to the rest of Europe and North America,” added Jan Maarten de Vries of TomTom.

Different to navigation maps

How are automated driving maps different from navigation maps? This is a fundamental and crucial aspect of getting the testing right for driverless cars and TomTom details the key differences between the two ways of mapping the road infrastructure.

The two kinds of maps differ primarily in two respects. First, accuracy is significantly higher – down to decimeter precision. Second, the map material for highly automated driving consists of multiple layers. There are three layers of information in maps for autonomous driving:

The traditional base navigation layer – This is used to calculate routes from A to B, including the sequence of roads to be driven.

The localisation layer – This layer uses a novel positioning concept providing highly accurate map data, which the automated vehicle uses to accurately calculate its position within a lane. To do this, the vehicle compares its sensed environment with the corresponding information in the localisation layer. In this way, the vehicle can accurately define its position relative to the road and its surroundings.

The planning layer – This sits on top of the localisation layer and contains not only attributes such as lane divider types, traffic signs, speed limits, etc., but also 3D information about road geometry, including curves and slopes. With the help of this very detailed lane information, the automated vehicle can decide things such as when and how to change lane.

Keeping map data valid

In highly automated driving, safety and comfort depend crucially on map material that is up to date. For example, up-to-the-minute speed-limit information has to be available instantly. Only then can vehicles select the best proactive driving strategy. In this regard, Bosch and TomTom rely on several elements and services to keep the map data up to date: the TomTom mapping fleet will continue to be regularly on the road, accurately mapping new roads and routes. And to register recent changes on the roads, such as changed lane configurations or new traffic signs, TomTom and Bosch plan to use feedback from fleets of vehicles equipped with the necessary sensors. Information about changed road conditions captured this way will be transferred to a server, verified, and entered in the digital map database. The updated map will then be fed back to the highly automated driving vehicle, enabling it to see effectively beyond its sensors.

, , , , , , , ,

Vehicle Technology

Leave a Reply

Your email address will not be published. Required fields are marked *



WordPress Anti-Spam by WP-SpamShield