HD Maps: Getting From A to B

Article By : Kayla Matthews

Sensors encompass a significant part of the technology that navigates an autonomous vehicle. Beyond them, HD maps will be substantial, too.

Autonomous cars make many people excited about what the future could bring. They eagerly anticipate being able to hop in an automobile and never have to touch the steering wheel or brake pedal during the journey. It’s easy to imagine the kind of freedom these cars could give to those who use them.

However, when forward-thinking individuals envision how this change may impact transportation, they don’t always think of the underlying technology that makes such advancements possible.

How will cars get to their destinations while staying aware of their environments and knowing which routes to take? Sensors encompass a significant part of the technology. Beyond them, HD maps will be substantial, too.

What are HD maps?

Sometimes called by their full name of high-definition maps, HD maps have many more details than traditional varieties do. They’re typically shown at the centimeter scale. Instead of merely indicating things like the locations of particular places or the distance between two destinations, HD maps feature information about lane placement, road boundaries, the severity of curves, the gradient of the road surface, and more.

Providing detail to this extent is crucial because so much inconsistency exists between roads. For example, bike lanes can be different widths depending on if they’re marked or unmarked. Plus, some city planners may choose to make those lanes substantially wider than the minimum in places with exceptionally heavy traffic.

If an autonomous car’s technology makes the vehicle maneuver as if all bike lanes are a uniform width, riders could be at risk for collisions. Bike lanes are just one example.

Traditional maps are for humans to read. People can use things like experience and depth perception when constantly assessing their environment and ensuring they don’t run off the road or veer into another lane. Since autonomous cars don’t have the same abilities as humans, they need maps that support what they can do. HD maps provide information that helps driverless vehicles move from one location to another without making errors.

How are HD maps structured?

HD maps have information presented in layers. This arrangement works well, considering the amount of data it must contain. As of now, there is no standardization between the businesses making 3D maps. The data in each layer varies depending on the company that produces the map.

For example, one business called HERE unveiled a three-layer map a few years ago. It has layers, and the content downloads as 2-kilometer by 2-kilometer tiles. The base layer is the mapping layer. It holds an HD map that shows curb-to-curb information with centimeter accuracy. The data there encompasses all the road markings and geographical boundaries, for example.

The second section is the dynamic layer. It has an HD map that gives live information. Those details could come from the sensors on surrounding cars or those embedded in the road. Local authorities may send data that shows on the dynamic layer, too.

Finally, the analytics layer of HERE’s map has information about how humans tackle particular stretches of road. The data there intends to make it so people riding in an autonomous car have experiences that feel as natural as possible, and not robotic.

Self-Healing Map

(Source: HERE)

Layers of HD maps also have information about landmarks, sometimes showing a virtual image of a car’s relationship to those places. Various industries benefit from geospatial data and use it to improve how they serve customers. For example, a bank might use it to determine the best location for a new ATM, while an office supply company might look at it to find potential buyers. HD maps need geospatial data, too.

What are some examples of HD mapping initiatives in progress?

High-definition maps are still in the early stages of development. Even so, some companies are hard at work to show the world what possibilities exist for the technology. TomTom, the brand most people associate with satellite navigation technology, is one of them. It recently started using an in-house test vehicle to improve the accuracy of its HD maps.

TomTom previously announced it was selected to provide HD mapping data to car manufacturers. That responsibility explains why the brand is especially eager to make the HD maps as reliable and complete as possible. The test car is a Volvo XC90 fitted with custom components, including eight laser scanners and six radars, plus GPS and mobile antennas. Those parts provide 360-degree situational awareness, and the collected information gets cross-checked against TomTom’s HD maps.

Similarly, an HD mapping company called CARMERA teamed up with Toyota to work on a method of creating maps autonomously with cameras. The companies believe this approach will be instrumental in creating improved maps. They say current work solely relates to highways, which comprise only 1% of the world’s roads.

This project starts in downtown Tokyo with vehicles fitted with cameras that capture footage as they move. Then, onboard processing converts the images into 3D maps. If the technology works as intended, it could help companies make HD maps more efficiently without requiring such a large labor force.

How far-fetched is a universal HD map for autonomous cars?

As mentioned earlier, the companies that deal with autonomous cars don’t have a single approach to use. When car manufacturers need HD maps, they either make them internally or rely on third-party companies. However, one collective aims to change that.

The well-named OneMap Alliance wants to make a global map for all autonomous vehicles by next year. The participating brands want to build a map that gives the same high-quality data regardless of a vehicle’s location in the world. HERE, the company mentioned in the earlier section about HD map structures, is playing a prominent role in the undertaking by delivering the HD map data.

The information should also get more current over time. That’s because it incorporates information cars collect as they move across the roadways. For example, an HD map may not initially have the details about a short-term construction project. However, if enough vehicles send information confirming its existence, the HD map would get updated.

Representatives of OneMap Alliance say their work has applications beyond autonomous cars, too. The information also supports the advanced driver assistance systems (ADAS) that are in some cars now. They include features that warn someone that they need to apply the brakes or engage in corrective steering to avoid accidents.

HD maps push autonomous cars forward

Autonomous cars can’t function safely or give the convenience that people crave without the information HD maps provide.

Although the work underway on them is fascinating and undoubtedly valuable, people can expect these maps to become even more advanced and information-rich soon.

Subscribe to Newsletter

Leave a comment