Autonomous and tech industries are moving towards self-driving cars with limited autonomy, either within geofenced locations or under certain conditions.

The autonomous vehicle aspires to be able to drive anywhere, anytime and the only human interaction being the destination. However, the main criteria for car owners to be where they want, is that the autonomous car will need to know exactly where they are. They will need to be a driving GPS.

So what does this mean? It means that new maps will need to have much more accuracy and detail. A map more complex than what a human being will be able to read. A map that companies such as ‘Here’ and ‘Civil Maps’ are gathering data to create a constantly updated 3D map of road networks and the environment environment around it.

Detailed maps are being created more for cars than humans.

Civil map and Nvidia are working on new cartography – ‘map drawing’. While civil maps have their own technology, the new car-centered cartography will include 3D point clouds combining data from 32-channel LIDAR scanners, radar, 96 mp cameras, and others.

HD mapping explained.

Here and Daimler created the first HD map for the ‘Birth of Benz’ drive through the German countryside. With automated driving, the map revolutionizes into another sensor to help the self-driving vehicle make decisions.

Sensor-encrusted mapping vehicles. The vehicle’s roof-mount sensor includes cameras, a LIDAR scanner, and a GPS measurement unit; to create a 3D road scan and its surroundings are sent to Here’s cloud, after which the data is then used in a digital recreation of the real world. Since 2016, maps of the road networks have been in process, to ensure the technology is ready to hit the roads by 2020. Similarly, Civil Map’s ‘Atlas DevKit’ advances localization, creates complex maps and crowdsource vital information with other cars. And it all happens in real time!

But road traffic is changing almost every day. It has been reported that maps used in autonomous vehicles will be updated consistently in correlation to road works, closures and other reported road obstacles such as accidents. Here has therefore built many relationships between cities and regional authorities, in order to be ready for any constructions that captures road users by surprise.

Here intends to take autonomous sensors to the next level. Here aims to ensure a forward-looking camera, forward and rear radar, and OEMs to get the sensor data transmitted on their cloud.

Here reports that algorithmic machines and aggregation algorithms will be able to drift through data and refine changes from noise in the world. In this way, Here will be able to update the map and deploy it into the vehicle. However, topnotch brand models will receive a more expensive radar sensor than a budget brand model, as well as get mounted in different view spots of the world.

Nvidia has been long renowned for its GPU’s, machine learning and neural networks. Nvidia shares the vision for self-driving cars with Here; believing that HD maps are essential for self-driving cars and that deep learning technology for the cloud-to-car mapping system will speed up car manufacturer’s ability to distribute autonomous vehicles. Here is developing a computer-system to detect changes in the cloud.

Once the maps start running autonomous vehicles on the road, Here’s goal is to have the car itself do the change detection. But, the data that the sensors will eventually send to the cloud is the difference between what the car sees and what the map say.