Mitsubishi Electric develops technologies for real-time 3D maps
Between 2019 and 2020, automated driving in Japan will evolve from advanced driver assistance systems (ADAS) towards level 3 automated driving (autonomous operation under certain conditions) – thus generating additional demand for appropriate systems. Automated driving systems require a combination of sensors within the vehicle as well as dynamic road maps. The biggest challenge: the information of these cards must always be up-to-date. The new Mitsubishi Electric technologies for automated mapping as well as to extract changes compared to previous versions in the mapping of landscape elements allow 3D maps to be created and updated faster and more efficiently, making them one of the key technologies for dynamic map generation.
In this environment, Mitsubishi Electric Corporation is developing a range of technologies for the automated creation of road maps. In addition, Mitsubishi Electric also presents applications that incorporate the modified elements in a landscape mapping. The applications are based on artificial intelligence (AI) and the mobile mapping system developed by Mitsubishi Electric (MMS), which allows the development of highly precise three-dimensional maps.
The maps provide static information about roads and surrounding objects and thus provide the basis for dynamic maps, which are indispensable for autonomous driving. They also offer continuously updated, dynamic information such as traffic lights and information about other vehicles on the road.
Mitsubishi’s automated mapping technology uses AI to quickly and accurately create three-dimensional road maps. In doing so, the solution extracts only the information from “point cloud” provided by the Lidar scanner as well as camera data collected by MMS. This includes, for example, traffic signs and road markings. Mitsubishi claims that its MMS generates 3D positioning information at an accuracy of 10 centimeters or less.
These data are recorded while driving by a system consisting of laser scanners (lidars), cameras and GPS antennas. AI enhances the extraction and recognition of the necessary data. This allows maps to be developed up to ten times faster than a traditional manual creation.
Mitsubishi Electric uses a technology for the development of dynamic road maps that only records the changes that have occurred since the last data collection. This allows faster and more efficient updating and maintenance of the cards. The technology automatically extracts the characteristic points of existing data, combines them with the latest lidar data from the MMS and thus detects differences and changes.
Mitsubishi Electric plans to distribute the software for automated map production as well as the technologies for processing the changed road data to card companies such as the Dynamic Map Planning Corporation from October. The application is used to develop extremely precise 3D maps of Japanese motorways.
Related articles:
Deep Learning makes static map data dynamic
BMW, Mobileye feed HD map data to service cloud
Intel buys into map services provider Here
Tomtom taps AI to create street map for autonomous driving
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
