Online and Offline Identification of Tyre Model Parameters
Doktorsavhandling, 2018
The accelerating development of active safety system and autonomous vehicles put higher requirements on both environmental sensing and vehicle state estimation as well as virtual verification of these systems. The tyres are relevant in this context due to the considerable influence of the tyres on the vehicle motion and the performance boundaries set by the tyres. All forces that the driver use to control the vehicle are generated in the contact patch between the tyre and the road on a normal passenger car. Hence, the performance limits imposed by the tyres should ideally be considered in the active safety systems and in self-driving vehicles. Due to tyres influence on the vehicle motions, they are some of the key components that must be accurately modelled to correlate complete vehicle simulations models with physical testing.
This thesis investigates the possibility to estimate the tyre-road friction coefficient during normal driving using active tyre force excitation, i.e. online identification of tyre model parameters. The thesis also investigates the possibility to scale tyre Force and Moment (F&M) models for complete vehicle simulations from indoor tests to real road surfaces using vehicle-based tyre testing, i.e. offline identification of tyre model parameters.
For online identification of tyre model parameters, the focus has been on how to perform tyre force excitation to maximize the information about the tyre-road friction coefficient. Furthermore, the required excitation level, as a ratio of the maximum tyre-road friction coefficient, for different road surfaces and tyre models have been evaluated for a larger number of passenger car tyres. The thesis shows the feasibility and benefits of using active tyre force excitations and illustrates its benefits when estimating the tyre-road friction coefficient by identifying nonlinear tyre model parameters. The method shows promising results by offering tyre-road friction estimates when demanded by the driver or an on-board system. This system can also be combined with other tyre-road friction estimates to offer a continuous tyre-road friction estimate, e.g. through car-to-car communication.
For offline identification of tyre model parameters, the focus was put on rescaling tyre models from indoor testing to a real-world road surface using vehicle-based tyre testing. Sensors were fitted to the vehicle to measure all inputs and outputs of the Pacejka 2002 tyre model. Furthermore, testing was performed on both different road surfaces and using different manoeuvres for tyre model identification. The effect on the complete vehicle behaviour in simulation when using tyre models based on different manoeuvres and road surfaces was investigated. The results show the importance of using a road surface and manoeuvre that are representative for the road surface and manoeuvre in which the vehicle will be evaluated. The sensitivity to different manoeuvres are mainly related to the changes in tyre properties with tyre surface temperature and the lack of temperature effects in the tyre model. The method shows promising results as an efficient way to rescale tyre models to a new road surface.
This thesis investigates the possibility to estimate the tyre-road friction coefficient during normal driving using active tyre force excitation, i.e. online identification of tyre model parameters. The thesis also investigates the possibility to scale tyre Force and Moment (F&M) models for complete vehicle simulations from indoor tests to real road surfaces using vehicle-based tyre testing, i.e. offline identification of tyre model parameters.
For online identification of tyre model parameters, the focus has been on how to perform tyre force excitation to maximize the information about the tyre-road friction coefficient. Furthermore, the required excitation level, as a ratio of the maximum tyre-road friction coefficient, for different road surfaces and tyre models have been evaluated for a larger number of passenger car tyres. The thesis shows the feasibility and benefits of using active tyre force excitations and illustrates its benefits when estimating the tyre-road friction coefficient by identifying nonlinear tyre model parameters. The method shows promising results by offering tyre-road friction estimates when demanded by the driver or an on-board system. This system can also be combined with other tyre-road friction estimates to offer a continuous tyre-road friction estimate, e.g. through car-to-car communication.
For offline identification of tyre model parameters, the focus was put on rescaling tyre models from indoor testing to a real-world road surface using vehicle-based tyre testing. Sensors were fitted to the vehicle to measure all inputs and outputs of the Pacejka 2002 tyre model. Furthermore, testing was performed on both different road surfaces and using different manoeuvres for tyre model identification. The effect on the complete vehicle behaviour in simulation when using tyre models based on different manoeuvres and road surfaces was investigated. The results show the importance of using a road surface and manoeuvre that are representative for the road surface and manoeuvre in which the vehicle will be evaluated. The sensitivity to different manoeuvres are mainly related to the changes in tyre properties with tyre surface temperature and the lack of temperature effects in the tyre model. The method shows promising results as an efficient way to rescale tyre models to a new road surface.
state estimation
computer aided engineering
tyre testing
parameter estimation
tyre-road friction estimation
active safety
simulation
vehicle dynamics
Författare
Anton Albinsson
Chalmers, Mekanik och maritima vetenskaper, Fordonsteknik och autonoma system
Tire Force Estimation Utilizing Wheel Torque Measurements and Validation in Simulations and Experiments
12th International Symposium on Advanced Vehicle Control (AVEC '14), Tokyo Japan,;(2014)p. 294-299
Paper i proceeding
Estimation of the inertial parameters of vehicles with electric propulsion
Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering,;Vol. 230(2016)p. 1155-1172
Artikel i vetenskaplig tidskrift
Identification of tyre characteristics using active force excitation
Dynamics of vehicles on roads and tracks. 24th Symposium of the International-Association-for-Vehicle-System-Dynamics (IAVSD) 2015. Graz, Austria, 17-21 August 2015,;(2016)p. 501-510
Paper i proceeding
Friction utilization for tyre-road friction estimation on snow: an experimental study
13th International Symposium on Advanced Vehicle Control, AVEC'16,;(2017)p. 541-546
Paper i proceeding
Design of tyre force excitation for tyre–road friction estimation
Vehicle System Dynamics,;Vol. 55(2017)p. 208-230
Artikel i vetenskaplig tidskrift
Quantification of Excitation Required for Accurate Friction Estimation
19th IEEE International Conference on Intelligent Transportation Systems, ITSC 2016;Rio de Janeiro; Brazil; 1-4 November 2016,;(2016)p. Art no 7795966, Pages 2551-2558
Paper i proceeding
Required Friction Utilization for Friction Estimation on Wet Asphalt, an Experimental Study
25th International Symposium on Dynamics of Vehicles on Roads and Tracks,;Vol. 1(2018)p. 407-412
Paper i proceeding
Albinsson, A. Bruzelius, F. Schalk, Els. Jacobson, B. Bakker, E., Tire Lateral Vibration Considerations in Vehicle Based Tire Testing
Evaluation of vehicle-based tyre testing methods
Proceedings of the Institution of Mechanical Engineers. Part D, Transport engineering,;Vol. 233(2019)p. 4-17
Artikel i vetenskaplig tidskrift
Albinsson, A., Bruzelius, F. Jacobson, B. Ran, S., Validation of vehicle based tyre testing methods
Tyre may not be regarded as the most interesting component to spend money on for most car owners. However, tyres are arguably some of the most important components on a car. When the driver requests a change in velocity through steering, braking or by pressing the accelerator pedal, these requests are met by the tyres generating a horizontal force. The total area in contact between the tyres and the road, for all fours wheels, are roughly the same as an A4 sheet of paper and the forces used to control the vehicle are all generated in this small area. The tyres influence not only how the car feels in normal driving situations. They also limit the maximum acceleration that the vehicle can achieve and thus the braking distance, maximum cornering velocity and the available traction.
This thesis has investigated how tyre properties that directly affects the vehicle motion can be identified. Two situations have been investigated, the first one being in normal conditions when the vehicle is driven on public road. Focus has been given to identify the maximum force, given by the tyre-road friction coefficient, that the tyres can generate. Finding the peak achievable force require that the vehicle is close to utilizing these forces, or in other words, the vehicle must operate close to the limit. This is not the case for most normal driving situations. This issue has been addressed by introducing and investigating an invention that generate forces at the tires with the purpose of identifying the coeffect of friction. Results suggest that this could potentially be implemented in cars on the road and would enable the vehicle to probe the slipperiness of the road at any time.
The second situation of tyre property identification is related to vehicle motion simulations for improving the vehicle development process. The force that the tyres generate depend on the relative velocity between the tyre and the road. A mathematical model that connect the tyre motion and the tyre force can be identified based on test data. These tests are normally performed on special machines in a laboratory environment. The tyre models can then be used in vehicle simulations to predict the motion and response of the vehicle. This thesis has investigated a method where the tyres are tested directly on a test track with special sensors fitted to the vehicle. The test track is closer to the real conditions of which the tyres are used compared to the laboratory environment, governing for more accurate tyre models based on this data.
The results from this thesis can be used by car manufacturers and others to improve the prediction of vehicle motion in simulations and reducing development cost. By enabling a more efficient development process of for example active safety systems, these systems can be improved which could potentially reduce the number of injuries in road traffic accidents. The identification of the tyre performance limitation when the vehicle is driving on the road can be used by autonomous vehicles and active safety systems to predict the vehicle motion capabilities.
This thesis has investigated how tyre properties that directly affects the vehicle motion can be identified. Two situations have been investigated, the first one being in normal conditions when the vehicle is driven on public road. Focus has been given to identify the maximum force, given by the tyre-road friction coefficient, that the tyres can generate. Finding the peak achievable force require that the vehicle is close to utilizing these forces, or in other words, the vehicle must operate close to the limit. This is not the case for most normal driving situations. This issue has been addressed by introducing and investigating an invention that generate forces at the tires with the purpose of identifying the coeffect of friction. Results suggest that this could potentially be implemented in cars on the road and would enable the vehicle to probe the slipperiness of the road at any time.
The second situation of tyre property identification is related to vehicle motion simulations for improving the vehicle development process. The force that the tyres generate depend on the relative velocity between the tyre and the road. A mathematical model that connect the tyre motion and the tyre force can be identified based on test data. These tests are normally performed on special machines in a laboratory environment. The tyre models can then be used in vehicle simulations to predict the motion and response of the vehicle. This thesis has investigated a method where the tyres are tested directly on a test track with special sensors fitted to the vehicle. The test track is closer to the real conditions of which the tyres are used compared to the laboratory environment, governing for more accurate tyre models based on this data.
The results from this thesis can be used by car manufacturers and others to improve the prediction of vehicle motion in simulations and reducing development cost. By enabling a more efficient development process of for example active safety systems, these systems can be improved which could potentially reduce the number of injuries in road traffic accidents. The identification of the tyre performance limitation when the vehicle is driving on the road can be used by autonomous vehicles and active safety systems to predict the vehicle motion capabilities.
Styrkeområden
Transport
Ämneskategorier
Farkostteknik
ISBN
978-91-7597-720-1
Doktorsavhandlingar vid Chalmers tekniska högskola. Ny serie: 4401
Utgivare
Chalmers
KA, Kemigården 4
Opponent: Professor Manfred Plöchl, Division of Technical Dynamics and Vehicle Dynamics, TU Wien, Austria.