Recheneffiziente Trajektorienoptimierung für Fahrzeuge mittels quadratischem Programm

Abstract

The economical limitations on the computing capacity in series-production vehicles in combination with increased requirements on safety and comfort narrow the application of established methods for trajectory optimization for automated driving. In order to meet these requirements and to reduce the computational effort at the same time, this work makes use of the well-known constrained linear-quadratic optimal control problems that can be solved very efficiently. Therefore, choosing an elaborate problem formulation based on the specific application of slack-variables, functional improvements with respect to collision avoidance as well as to a human driving behavior can be achieved. Experimental results of the implemented lateral and longitudinal vehicle guidance algorithms demonstrate the abilities for automated driving applications of this approach.