Toyota Research Institute - Advanced Development, Inc. (TRI-AD) and Dynamic Map Platform Co., Ltd. (DMP) will partner to build a proof-of-concept (PoC) applying the technologies of both companies starting April 2020. This PoC aims to demonstrate the efficient update of DMP’s high definition (HD) maps by using TRI-AD’s open software platform, Automated Mapping Platform (AMP) to detect and process changes in lane-level road features based on images and other data collected from vehicle sensors.
Electronic intersection. Source: DMP.
Demand for high-accuracy and high-quality HD maps for automated driving has been growing with the increased use of advanced driver assistance systems (ADAS) on the way to automated driving. It is essential for map building to capture changes in roadway attributes accurately and quickly, and to update the maps with high efficiency.
The conventional approach detects lane-level changes through on-site surveys using expensive survey vehicles and typically based on planned road construction notices. In this PoC, TRI-AD’s image recognition technology automatically detects changes based on image data collected by commonly available cameras.
By helping focus the survey fleet only on areas of change as they occur, it is expected to demonstrate a significant reduction in the cost of operating the fleet and an acceleration of the map update to users including automotive manufacturers.
TRI-AD and DMP will demonstrate the ability to update DMP’s HD maps with an improved efficiency using the automatic change detection technology by TRI-AD’s AMP based on images and other data of roadways in Japan. Aiming to start operation from fiscal year 2021, this PoC will study methods to further improve the frequency of HD map updates, and consider the use of AMP to update map data by Ushr, Inc. (a DMP Group company) that supplies its HD maps in North America.