Integration of driver assistant systems as a virtual driver

Current assistant systems like ESP optimize the driver's command by realizing a transition between the command and the optimum as soon at it goes out of a safety range and does it later back when the driver's command is no more dangerous. The next generation of assistant systems will not only be reactive but also predictive. They will still work on the same principle but with a wider environment of the vehicle. Their actions will follow the same three steps: computation of the optimum of the maneuver, waiting passively until the very last moment, then generating the transitions. This article describes first the fusion of the assistant systems as a virtual driver, then a decision control that can realize this optimization depending on its confidence in order to improve the reliability of the command level