Friction estimation – optimization of sensor configuration with respect to RMSE and costs
The accuracy of friction estimation depends on the sensors used. Furthermore, the costs of sensors have to be considered during system design.
In this work, we discuss how sensor configurations (i.e. subsets of the sensors used during our measurements) can be determined minimizing RMSE and/or sensor costs. We first demonstrate that simple strategies (e.g. replacing a sensor by a better one) may fail. Then we discuss different optimization strategies (search strategies and cost functions). By applying them, we obtained several sensor configurations showing interesting trade-offs in minimizing RMSE and sensor costs. Based on these results, we argue whether general statements can be derived about the usefulness of particular sensors and sensor combinations.
KeywordsRoot Mean Square Error Side Slip Angle Sensor Channel Echo State Network Advanced Driver Assistance System
Unable to display preview. Download preview PDF.