Analysis of the Potential of Onboard Vehicle Sensors for Model-based Maximum Friction Coefficient Estimation

Advanced driver assistance systems (ADAS) have led to a steady improvement in driving comfort and safety. Knowledge of vehicle dynamics and perception of the vehicle’s environment are necessary for optimized ADAS and autonomous driving. A crucial parameter influencing vehicle dynamics is the maximum friction coefficient between the tire and the road. As this coefficient cannot be measured economically in serial production cars via existing vehicle sensors, model-based estimation algorithms are a field of interest.In this paper, maximum friction coefficient estimation is presented using an unscented Kalman filter (UKF) based on onboard vehicle sensors such as a six degrees of freedom inertial measurement unit, height level sensors, and tie rod force sensors. The goal is to analyze the potential of these sensors for maximum friction coefficient estimation. First, a variance-based sensitivity analysis is used to analyze the physical vehicle model of a Dacia Duster. Second, model-based maximum friction coefficient estimation is implemented for the test vehicle and the results using different sensor settings are compared for driving maneuvers carried out on a test track with different road surfaces. Finally, model-based maximum friction coefficient estimation using these onboard vehicle sensors shows improved results compared to the UKF from previous works.