The TomTom HD Map allows automated vehicles to become location-aware, environment-aware and path-aware.
Autonomous vehicles require maps that are significantly different from the maps that are used in today’s navigation systems. Drivers today mainly use digital maps to orientate themselves, to plan a journey and to navigate to their destination. However, as the driving task gradually shifts from the driver to in-vehicle automated systems, the role and scope of digital maps shifts accordingly. This means that the user of the map is no longer the driver, but rather a machine. As a result, a new generation of maps built purposely for machines is needed. These next generation maps for machines come in the form of a highly accurate and realistic representation of the road, generally referred to as high-definition (HD) maps.
The TomTom HD Map is a highly accurate and highly attributed representation of the road, including attributes such as lane models, traffic signs, road furniture and lane geometry, with accuracy down to a few centimeters. The TomTom HD Map allows automated vehicles to precisely locate themselves on the road, to build a detailed model of the surrounding environment working together with the vehicle sensors, and to plan their path to destination. The TomTom HD Map is not limited to autonomous driving, but can also be leveraged to fulfill a broad range of ADAS applications such as Predictive Powertrain Control, Highway Pilot, and Adaptive Cruise Control.
TomTom is currently working on extracting map data from observations from a variety of sensors, amongst them also cameras. Through the acquisition of Autonomos, a Berlin-based autonomous driving start-up with heritage dating back to the DARPA Challenge, we significantly increased our expertise in computer vision and data extraction. Combining this know-how with our existing Artificial Intelligence expertise, we’re extracting map data – such as road geometry, traffic signs and landmarks – from camera images. Capturing this type of data is particularly relevant in order to update the HD Map, to ensure it matches reality quickly. This will enable scalable and efficient updates to the TomTom HD Map on a continuous basis. By exploring and testing innovative techniques to automate the update of HD Maps, TomTom is taking important steps towards setting a global standard for HD Maps.
A key challenge of autonomous driving is determining the exact location of a vehicle on the road. To operate safely and efficiently, it is key for an autonomous vehicle to be able to position itself on a specific lane with extreme accuracy. Today’s navigation systems use traditional GPS for localization; however, traditional GPS fails to deliver the needed level of accuracy and robustness to enable autonomous driving. In fact, GPS is only accurate down to a few meters, meaning that a vehicle relying on GPS data for localization would not always be able to determine what lane it is currently driving in.
To address this challenge, the TomTom HD map offers a range of content layers that provide a reference for sensor observations so that the vehicle can be located in the HD Map. For example, TomTom has developed RoadDNA, a set of continuous sensor-signal content layers in the TomTom HD Map that enable accurate and robust localization for autonomous vehicles. TomTom’s patented RoadDNA technology delivers a highly optimized, 3D lateral and longitudinal view of the roadway. A vehicle can correlate RoadDNA data with data obtained by its own sensors. By doing this correlation in real time, TomTom RoadDNA allows vehicles to precisely position themselves on the road, even while travelling at high speeds.
By converting a 3D point cloud of roadside patterns into a compressed, 2D view of the roadway, TomTom RoadDNA delivers a solution that can be used in-vehicle with limited processing requirements. Without losing roadway detail, TomTom RoadDNA follows a feature-agnostic approach which is robust and scalable. This technique eliminates the complexity of identifying each single roadway object, but rather creates a unique pattern of the roadway environment.