Robust Design of a Complex, Perturbed Lateral Control System for Automated Driving

This paper presents a method related to Cooperative Design for the synthesis and analysis of a robust control system. Most control approaches make use of simplified, mostly linear models which may neglect important system properties. By contrast, our sampling based approach allows the usage of a complex and nonlinear simulation system that may also include black-box subsystems. Furthermore, disturbances are simultaneously considered. For the comprehensive view of the entire system, the dependencies between design variables and performance measures of the system are organized in a so-called dependency graph. Three distinct steps are carried out concerning disturbance rejection, trajectory transfer behavior and robustness analysis and are demonstrated by the example of a lateral position control for autonomous driving. A velocity dependent parameter set for disturbance observers, feedback controller and feedforward controller is derived. In order to achieve a robust parametrization different vehicle states and disturbances like system latencies and crosswind are taken into account.