カテゴリー: Faq

When it comes to HD maps, there are different types of accuracy to be considered – and several different factors that can contribute to errors.

The short answer to this question is that we achieve a local/map relative accuracy of 5cm or below for objects within a 10m corridor of the drivable route.

Global/absolute accuracy is determined by GNSS/GPS, which cannot be bound. Conditions sometimes vary quite a lot, particularly in areas with poor satellite reception (e. g. urban canyons or tunnels), so this number needs to be taken with a grain of salt: For areas where the GNSS quality is good and proper base station correction data available, the global accuracy will deviate by an additional 3cm or below.

For a more detailed answer, information on how we achieve this accuracy, and how we mitigate errors in GNSS-denied areas please give our article “How Accurate Are HD Maps for Autonomous Driving and ADAS Simulation?” a read.

Yes, we do! Depending on your use case and requirements for update turnaround times, there are different options available.

We offer both “soft updates” that can be performed by editing already collected data (e. g. changing a speed limit, temporarily disabling a lane or updating right-of-way rules) as well as “hard updates” which require a re-scan of parts of your route (e. g. after roadworks/construction).

For soft updates, we connect you to our map cloud where you can easily submit precise change requests which our team processes within a week or within 24h if required. For hard updates, options differ depending on your specific use case: There are a number of factors to consider for us to make you a suitable offer. We’re definitely happy to have that conversation, though:

To see how our update services can serve your project, simply schedule a call with us!

Pricing for our HD maps depends mostly on four factors:

1. What is the desired coverage – are we looking at 40 km or 4000 km? The cost per kilometer is lower for larger routes, due to fixed base efforts for each project.
2. How complex is the route to be mapped (urban/highway/mix)? The easiest way to get a specific answer is usually for you to tell us which specific routes you’d like mapped.
3. What is the desired degree of map fidelity/annotation detail? We can include a varying number of object classes, depending on your project requirements to make sure you only pay for what you need.
4. Which data format(s) would you like to receive your map in? Since map data formats sometimes differ considerably, it may be necessary to perform certain portions of the production process multiple times, in different ways. You can see a selection of supported map formats in the article below.
   
   We’ll be happy to discuss your needs in more detail – feel free to book a meeting with us!

Since our approach to mapping is mainly computer vision-based, we do not necessarily rely on LiDAR for data acquisition.

However: If your project would profit from using it (e. g. by receiving georeferenced pointcloud data in addition to a map), we supplement our sensor pod with a 360° LiDAR scanner.

If you would like to learn more about the technology we use and how it can be leveraged for your use case, don’t hesitate to contact us!

The delivery times for your HD map depend mostly on two factors:

1. How complex is the route to be mapped (urban/highway/mix) and how big is the coverage? The easiest way to get an accurate answer is for you to tell us which specific route(s) you’d like mapped.
2. What is the desired degree of map fidelity/annotation detail? We can include a varying number of object classes, depending on your project requirements to make production as efficient as possible.

During map production, we break down your route into a number of smaller sub-projects: This allows for an incremental release of one part after another, so you can take delivery of the first map section much sooner than the entire project will be finished.

Alternatively, you can subscribe to our データベース of ready-made routes from the US, Europe and Japan.

We cater to a range of map formats, depending on your use case and toolchain in place – here is a selection of frequently requested ones:

OpenDRIVE 

Open source format available in different versions and variations, depending on your target system. Frequently used for simulation; supported tools include:

• SCANeR by AVSimulation
• Prescan by Siemens
• CARLA (open source)
• VTD/Virtual Test Drive by Hexagon (MSC/VIRES)
• ASM by dSPACE
• LGSVL by LG (alternative supported format: Lanelet2)
• VRX/VRXPERIENCE by Ansys
• Vissim by PTV
• Cognata
• RoadRunner by MathWorks (formerly VectorZero)

ROAD5

Proprietary format by IPG Automotive. Supported for:

• CarMaker/TruckMaker

CPAR

Proprietary format by MechanicalSimulation; short for consolidated parsfile. Supported for:

• CarSim/TruckSim

Lanelet2

Open source map format/framework for autonomous driving applications. Available as:

• Lanelet2 (standard)
• Lanelet2 for Autoware

Geometric formats/meshes

Often used for 3D modelling, e. g. to create assets for virtual environment models. Supported formats include:

• OBJ
• FBX
• OSGB
• SHP