Anton Albinsson

Design of tyre force excitation for tyre–road friction estimation

ABSTRACT Knowledge of the current tyre–road friction coefficient is essential for future autonomous vehicles. The environmental conditions, and the tyre–road friction in particular, determine both the braking distance and the maximum cornering velocity and thus set the boundaries for the vehicle. Tyre–road friction is difficult to estimate during normal driving due to low levels of tyre force excitation. This problem can be solved by using active tyre force excitation. A torque is added to one or several wheels in the purpose of estimating the tyre–road friction coefficient. Active tyre force excitation provides the opportunity to design the tyre force excitation freely. This study investigates how the tyre force should be applied to minimise the error of the tyre–road friction estimate. The performance of different excitation strategies was found to be dependent on both tyre model choice and noise level. Furthermore, the advantage with using tyre models with more parameters decreased when noise was added to the force and slip ratio.

Vehicle Motion Measurements Using Front Facing Camera and Digital Image Correlation

Modern active vehicle safety systems rely on certain vehicle motion states to function. ABS requires the vehicle longitudinal speed to calculate the tire slip. The vehicle speed is typically estimated using the speed of all the wheels and is therefore dependent on the slip states of all the wheels. Electronic stability programs can also make more informed decisions if the vehicle side-slip angle is known. Currently the side-slip angle is not measured on commercial vehicles due to the cost of the sensors. The side-slip angle can however be estimated using multiple onboard vehicle measurements. However, these estimation techniques require accurate sensors and large excitations to estimate accurately. The measurement of the vehicle motion is therefore crucial for modern vehicle safety systems. This paper proposes a method whereby all 6 vehicle velocities can be measured using inexpensive forward facing mono and stereo cameras utilizing Digital Image Correlation (DIC) algorithms.

Estimation of the inertial parameters of vehicles with electric propulsion

More accurate information about the basic vehicle parameters can improve the dynamic control functions of a vehicle. Methods for online estimation of the mass, the rolling resistance, the aerodynamic drag coefficient, the yaw inertia and the longitudinal position of the centre of gravity of an electric hybrid vehicle is therefore proposed. The estimators use the standard vehicle sensor set and the estimate of the electric motor torque. No additional sensors are hence required and no assumptions are made regarding the tyre or the vehicle characteristics. Consequently, all information about the vehicle is available to the estimator. The estimators are evaluated using both simulations and experiments. Estimations of the mass, the rolling resistance and the aerodynamic drag coefficient are based on a recursive least-squares method with multiple forgetting factors. The mass estimate converged to within 3% of the measured vehicle mass for the test cases with sufficient excitation that were evaluated. Two methods to estimate the longitudinal position of the centre of gravity and the yaw inertia are also proposed. The first method is based on the equations of motion and was found to be sensitive to the measurement and parameter errors. The second method is based on the estimated mass and seat-belt indicators. This estimator is more robust and reduces the estimation error in comparison with that obtained by assuming static parameters. The results show that the proposed method improves the estimations of the inertial parameters. Hence, it enables online non-linear tyre force estimators and tyre-model-based tyre–road friction estimators to be used in production vehicles.

Evaluation of vehicle-based tyre testing methods

The demand for reduced development time and cost for passenger cars increases the strive to replace physical testing with simulations. This leads to requirements on the accuracy of the simulation models used in the development process. The tyres, the only components transferring forces from the road to the vehicle, are a challenge from a modelling and parameterization perspective. Tests are typically performed on flat belt tyre testing machines. Flat belt machines offers repeatable and reliable measurements. However, differences between the real world road surface and the flat belt can be expected. Hence, when using a tyre model based on flat belt measurements in full vehicle simulations, differences between the simulations and real prototype testing can be expected as well. Vehicle-based tyre testing can complement flat belt measurements by allowing reparameterization of tyre models to a new road surface. This paper describes an experimental vehicle-based tyre testing approach that aims to parameterize force and moment tyre models compatible with the standard tyre interface. Full-vehicle tests are performed, and the results are compared to measurements from a mobile tyre testing rig on the same surface and to measurements on a flat belt machine. The results show that it is feasible to measure the inputs and outputs to the standard tyre interface on a flat road surface with the used experimental setup. The flat belt surface and the surface on the test track show similar characteristics. The maximum lateral force is sensitive to the chosen manoeuvres, likely due to temperature differences and to vibrations at large slip angles. For tyre models that do not model these effects, it is vital to test the tyres in a manoeuvre that creates comparable conditions for the tyres as the manoeuvre in which the tyre model will be used.

Quantification of Excitation Required for Accurate Friction Estimation

Real-time information about the current tyre-road friction coefficient is important for vehicles with increasing levels of automation to adapt intervention thresholds and vehicle velocity. Online friction estimation poses large challenges in terms of availability and accuracy of the estimate. When estimating the friction based on the slip and force of a tyre, the accuracy of the estimate depends on the current tyre friction utilisation. This paper presents a method to evaluate how large friction utilisation is required to estimate the current friction level within a given error using an effect-based approach. The results indicate that large variations in the required utilisation should be expected for different tyre and road surface combinations when tyre models that only have the slip stiffness and the friction coefficient as parameters (e.g. brush tyre model) are used in the observer to estimate the friction coefficient. For the brush tyre model it varied from 54% in the best case to 94% in the worst case. To illustrate the benefit of quantification of required utilisation, an example use of an estimator based on the brush tyre model is shown at the end of this paper. A simulation is performed to show how fast a sudden drop in friction during braking can be detected using a recursive online estimator. The results show that the estimator is able to detect the change in friction and converge to the new friction level in 0.06 s, which is fast enough to potentially improve the performance of ABS systems of trucks. Tyre models for online friction estimation must hence be chosen with the intended use in mind. The results show that the friction information estimated using the brush tyre model at larger excitation levels is reliable and can therefore be shared with other road users in order to improve the overall safety and efficiency of the transportation system.