Tire testing is an expensive affair that typically involves the use of dedicated Flat-Trac machines, on-road trailers or heavily instrumented vehicles. This presentation proposes a Gaussian Process Regression model for tire force estimation, trained using virtual handling data obtained via simulation. The model predicts tire forces and the corresponding slip and load inputs from easily measured vehicle states and the results correlate well with the ground truth. It is shown how this approach can open new possibilities in cost-effective, large-scale tire testing. This is particularly useful for augmenting on-vehicle tire particulate emissions measurements with force and slip information.

• The minimum instrumentation requirements for estimating tire forces from full-vehicle maneuvers
• How to formulate an effective GPR (Gaussian Process Regression) model for tire slip and force estimation
• The best maneuvers to train the model using full-vehicle simulation
• How to apply the GPR model in practice to obtain force and slip information
• Key application areas such as tire wear/particulate emissions, where the method may augment emissions test data with tire force and slip information