As a comprehensive effort to reduce CO2 emissions from the use of automobiles, it is expected that publicly implementing automated driving will lead to the prevention of traffic congestion and accidents that cause such congestion.
However, in-vehicle computing, which is required for automated driving, uses a significant amount of power, which can affect the driving time and distance of electric vehicles, and given the current technology, could be a bottleneck to widespread use of electric vehicles.
Therefore, in order to greatly improve the energy efficiency of in-vehicle computing (namely by reducing power consumption by 70% compared to current technology), this project will conduct R&D on improving the energy efficiency of automated driving software and sensor systems, which have a significant effect on power consumption.
At the same time, in order to strengthen the competitiveness of the entire supply chain, where development systems for electrification and automation need to be transformed, the project will develop a standard simulation model for electric vehicles overall that is compatible with automated driving.
Development of automated driving open platform software
The project will conduct R&D to reduce by 70% the software computation load related to hardware while securing the necessary level of performance (Level-4 automated driving in main travel environments), assuming architecture that mitigates the load placed on the network or cloud.
Development of automated driving sensor systems to realize high performance and low power consumption
Concerning “recognition” information processing for sensor systems used in automated driving, the project will conduct R&D geared to greatly reducing power consumption (by 70%) while satisfying the necessary level of performance (Level-4 automated driving in main travel environments) by enhancing the recognition method through more efficient processing and improving the performance of linked sensor devices by the enhancement of input values.
Development of electric vehicle simulation infrastructure
An electric vehicle simulation model, which is essential for testing and evaluating automated driving, will be developed to match the behavior of actual vehicles with a dynamic simulation precision of 90% or higher. By developing a standard model that can be widely used, the time required for performance verification will be reduced by half across the supply chain, thereby helping to shorten the time required for developing electric vehicles.