This paper presents a cost-efficient approach for optimizing parameters of an Automated Lane Keeping System (ALKS) based on simulation and test bench measurements. Metamodels are trained and used to minimize the number of actual required measurements. The approach handles multiple criteria and the combination of objective and subjective evaluations. Different Pareto-optimal parameter sets are revealed to the calibration engineer. An exemplary calibration process using the optimization tool is illustrated. Quality and efficiency of the approach is demonstrated by objective simulation data from IPG CarMaker and subjective evaluation in the Stuttgart Driving Simulator. This research enables a higher degree of automation in the ALKS calibration process and supports the automotive industry`s efforts in front-loading during development.
Session:
Vehicle Dynamics
|
| 15:30 - 16:00