Smart driver: A research project for closed loop vehicle simulation in MSC.ADAMS

Abstract

It is almost universally accepted that predictive control and a hierarchical control architecture are at the basis of any human driver simulator. For human driving simulation, a hierarchical control structure is often subdivided in three levels: a strategic level for task planning; a tactical level for trajectory planning; an operational level for motion control. In the past 15 years, the geometric nonlinear control of mechanical systems and in particular of non-holonomic vehicles has been a very active research area. Concepts, such has differential flatness, which lead to new control paradigms have been introduced. These and other principles have all been applied in the SmartDriver Research Project, that MSC is conducting for defining a new driver model for full vehicle MSC.ADAMS models. The three basic components neeeded to perform an accurate simulation of a control-driven car on a given track are: the Vehicle Model (MSC.ADAMS), the Road Description and the Car Driver Model (SmartDriver). The paper illustrates some of the principles mentioned and shows a few application examples.