Institute for Frontier Science Initiative, Future society creation core, Autonomous vehicle research unit, Kanazawa University
*The organization and the title are those when awarded
Autonomous driving in an urban area with complicated road environments needs both a high accuracy digital map and localization methods to estimate a vehicle’s position precisely.
Global navigation satellite system (GNSS), e.g. GPS, is often used as a practical localization method.
However, for autonomous driving, GPS has problems in accuracy under certain conditions, such as high-rise areas or inside tunnels.
Dr. Suganuma has developed a novel, GNSS-independent localization method called “High dependable localization method”.
This method has achieved vehicle localization with an accuracy of about 14 centimeters, by comparing 2-dimentional orthoimages as a digital map and data from on-board sensors.
In addition, the method requires less memory and computing power than conventional methods because smaller image data is applied.
The productization approach and testing on public roads has already begun. A large contribution to safe and comfortable traveling and next generation mobility in the areas of venerability and depopulation is expected.