Tire model and road adhesion coefficient collaborative identification method and system

Abstract

The application discloses a tire model and road adhesion coefficient collaborative identification method, which comprises the following steps: firstly, a virtual simulation vehicle model is established according to real vehicle parameters, tire models with different parameters are configured in the virtual simulation vehicle model, longitudinal braking / lateral steering simulation is carried out in a road working condition with a first adhesion coefficient, a neural network identification model of the tire model is constructed, then the dynamics response of a real vehicle equipped with a tire model to be identified in the road working condition with the first adhesion coefficient is collected as the input of the neural network identification model of the tire model, the characteristic parameters of the tire model to be identified are obtained, then the virtual simulation vehicle model is combined, the whole vehicle dynamics response data on different adhesion rate road surfaces are collected as the input, a neural network identification model of the road adhesion coefficient is constructed, and the real-time online identification of the road adhesion coefficient when pure side slip or pure longitudinal slip control is adopted is realized, finally, the tire model and the road adhesion coefficient are collaboratively identified, and no special equipment is needed, and the implementation cost is low.

Description

Technical Field

[0001] This invention relates to the field of automotive technology, and in particular to a method and system for the collaborative identification of tire models and road surface adhesion coefficients. Background Technology

[0002] Tire models and road adhesion coefficients are key data for vehicle state estimation, vehicle dynamics and control, and intelligent driving path tracking control to achieve precise control effects. However, tire models exhibit strong nonlinear and saturation characteristics, while road adhesion coefficients exhibit time-varying characteristics. Both of these are often unknown at the same time.

[0003] Currently, tire model identification typically requires specialized equipment such as tire force analyzers or tire testing benches to obtain data on tire slip angle, lateral force, slip ratio, longitudinal force, and load. Then, optimization algorithms, such as genetic algorithms, are used to perform nonlinear fitting on the tire model. This method relies on specialized equipment, is expensive, and is difficult to popularize. Summary of the Invention

[0004] Therefore, one embodiment of the present invention proposes a method for the collaborative identification of tire model and road surface adhesion coefficient, so as to realize the collaborative identification of tire model and road surface adhesion coefficient, while having the advantages of not requiring special equipment and low implementation cost.

[0005] A method for collaborative identification of tire model and road adhesion coefficient includes:

[0006] A virtual simulation vehicle model is established based on the parameters of the real vehicle, and tire models with different parameters are configured in the virtual simulation vehicle model. Longitudinal braking / lateral steering simulation is performed under road conditions with a first adhesion coefficient. The dynamic response of the virtual vehicle is collected by sensors as input to the neural network training data, and the corresponding tire model feature parameters are used as output to train the neural network model and construct a neural network identification model of the tire model.

[0007] The dynamic response of a real vehicle equipped with the tire model to be identified is collected under road conditions with a first adhesion coefficient. After classification based on the classification conditions of the training data, it is used as the input of the neural network identification model of the tire model to obtain the feature parameters of the tire model to be identified, so as to determine the nonlinear tire model and realize the offline identification of the tire model.

[0008] After determining the nonlinear tire model, the tire model in the virtual simulation vehicle model is updated to the nonlinear tire model identified offline. Combined with the virtual simulation vehicle model, the vehicle dynamic response data on roads with different adhesion rates are collected as input and the corresponding road adhesion coefficient is used as output to train the neural network model and construct a neural network identification model for the road adhesion coefficient.

[0009] The instantaneous dynamic response of the actual vehicle on the road surface to be identified is collected as the input of the neural network identification model of the road surface adhesion coefficient, so as to realize the real-time online identification of the road surface adhesion coefficient when pure sideslip or pure longitudinal skid maneuvers are adopted.

[0010] The tire model and road surface adhesion coefficient collaborative identification method provided by the present invention first establishes a virtual simulation vehicle model based on real vehicle parameters, and configures tire models with different parameters in the virtual simulation vehicle model. Longitudinal braking / lateral steering simulation is performed under road conditions with a first adhesion coefficient to construct a neural network identification model for the tire model. Then, the dynamic response of a real vehicle equipped with the tire model to be identified under road conditions with a first adhesion coefficient is collected as input to the neural network identification model of the tire model to obtain the characteristic parameters of the tire model to be identified. Next, combined with the virtual simulation vehicle model, the vehicle dynamic response data on roads with different adhesion rates is collected as input to construct a neural network identification model for the road surface adhesion coefficient. Then, real-time online identification of the road surface adhesion coefficient is achieved when pure lateral slip or pure longitudinal slip maneuvering is performed. Finally, collaborative identification of the tire model and road surface adhesion coefficient is realized, without the need for specialized equipment and with low implementation cost.

[0011] Furthermore, the tire model and road adhesion coefficient collaborative identification method provided by the present invention also has the following technical features:

[0012] Furthermore, the sensors include at least a virtual IMU, a wheel speed sensor, and a steering wheel angle sensor.

[0013] Furthermore, before the step of training the neural network model and constructing a neural network identification model for the road surface adhesion coefficient, the method further includes:

[0014] Based on the different sensitivities of the influence of different tire model feature parameters on the tire longitudinal force-longitudinal slip ratio curve, the data is classified and then trained into a neural network model to construct a neural network identification model for road adhesion coefficient.

[0015] Furthermore, the virtual simulation vehicle model is the Carsim virtual simulation vehicle model.

[0016] Furthermore, the neural network model is a BP neural network model.

[0017] Another embodiment of the present invention proposes a tire model and road surface adhesion coefficient collaborative identification system to achieve collaborative identification of tire model and road surface adhesion coefficient, while having the advantages of not requiring special equipment and low implementation cost.

[0018] A tire model and road adhesion coefficient collaborative identification system includes:

[0019] The first construction module is used to build a virtual simulation vehicle model based on the parameters of the real vehicle, and configure tire models with different parameters in the virtual simulation vehicle model. It performs longitudinal braking / lateral steering simulation in road conditions with a first adhesion coefficient, uses sensors to collect the dynamic response of the virtual vehicle as input to the neural network training data, and uses the corresponding tire model feature parameters as output to train the neural network model and build a neural network identification model of the tire model.

[0020] The first identification module is used to collect the dynamic response of the actual vehicle equipped with the tire model to be identified in road conditions with a first adhesion coefficient. After classifying the data according to the classification conditions of the training data, it is used as the input of the neural network identification model of the tire model to obtain the feature parameters of the tire model to be identified, so as to determine the nonlinear tire model and realize the offline identification of the tire model.

[0021] The second construction module is used to update the tire model in the virtual simulation vehicle model to the offline identified nonlinear tire model after determining the nonlinear tire model. Combined with the virtual simulation vehicle model, the module collects vehicle dynamic response data on roads with different adhesion rates as input and the corresponding road adhesion coefficient as output to train the neural network model and construct a neural network identification model for the road adhesion coefficient.

[0022] The second identification module is used to collect the instantaneous dynamic response of the actual vehicle on the road surface to be identified as the input of the neural network identification model of the road surface adhesion coefficient, so as to realize the real-time online identification of the road surface adhesion coefficient when pure sideslip or pure longitudinal skidding is adopted.

[0023] The tire model and road surface adhesion coefficient collaborative identification system provided by the present invention first establishes a virtual simulation vehicle model based on real vehicle parameters, and configures tire models with different parameters in the virtual simulation vehicle model. Longitudinal braking / lateral steering simulation is performed under road conditions with a first adhesion coefficient to construct a neural network identification model for the tire model. Then, the dynamic response of a real vehicle equipped with the tire model to be identified under road conditions with a first adhesion coefficient is collected as input to the neural network identification model of the tire model to obtain the characteristic parameters of the tire model to be identified. Next, combined with the virtual simulation vehicle model, the system collects vehicle dynamic response data on roads with different adhesion rates as input to construct a neural network identification model for the road surface adhesion coefficient. Finally, it realizes real-time online identification of the road surface adhesion coefficient when performing pure lateral or pure longitudinal skidding maneuvers, ultimately achieving collaborative identification of the tire model and road surface adhesion coefficient without requiring specialized equipment and with low implementation cost.

[0024] Furthermore, the tire model and road adhesion coefficient collaborative identification system provided by the present invention also has the following technical features:

[0025] Furthermore, the sensors include at least a virtual IMU, a wheel speed sensor, and a steering wheel angle sensor.

[0026] Furthermore, the system also includes:

[0027] The classification module is used to classify the data based on the different sensitivities of the influence of the characteristic parameters of different tire models on the longitudinal force-longitudinal slip ratio curve, and then train the neural network model to construct a neural network identification model for the road adhesion coefficient.

[0028] Furthermore, the virtual simulation vehicle model is the Carsim virtual simulation vehicle model.

[0029] Furthermore, the neural network model is a BP neural network model. Attached Figure Description

[0030] The above and / or additional aspects and advantages of the embodiments of the present invention will become apparent and readily understood from the description of the embodiments in conjunction with the following drawings, wherein:

[0031] Figure 1 This is a flowchart of a tire model and road surface adhesion coefficient collaborative identification method according to an embodiment of the present invention;

[0032] Figure 2 This is a structural block diagram of a tire model and road surface adhesion coefficient collaborative identification system according to an embodiment of the present invention. Detailed Implementation

[0033] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0034] Please see Figure 1 The method for collaborative identification of tire model and road adhesion coefficient proposed in the first embodiment of the present invention includes steps S101 to S104:

[0035] S101. A virtual simulation vehicle model is established based on the parameters of the real vehicle, and tire models with different parameters are configured in the virtual simulation vehicle model. Longitudinal braking / lateral steering simulation is performed under road conditions with a first adhesion coefficient. The dynamic response of the virtual vehicle is collected by sensors as input to the neural network training data, and the corresponding tire model feature parameters are used as output to train the neural network model and construct a neural network identification model of the tire model.

[0036] The virtual simulation vehicle model is the Carsim virtual simulation vehicle model.

[0037] The sensors include at least a virtual IMU, a wheel speed sensor, and a steering wheel angle sensor.

[0038] The neural network model is a BP neural network model.

[0039] Furthermore, in this embodiment, it should be noted that, before the step of training a neural network model and constructing a neural network identification model for the road surface adhesion coefficient to improve the accuracy of identification, the method further includes:

[0040] Based on the different sensitivities of the influence of different tire model feature parameters on the tire longitudinal force-longitudinal slip ratio curve, the data is classified and then trained into a neural network model to construct a neural network identification model for road adhesion coefficient.

[0041] S102 collects the dynamic response of a real vehicle equipped with the tire model to be identified in road conditions with a first adhesion coefficient. After classifying the data according to the classification conditions of the training data, it is used as the input of the neural network identification model of the tire model to obtain the feature parameters of the tire model to be identified, so as to determine the nonlinear tire model and realize the offline identification of the tire model.

[0042] S103. After determining the nonlinear tire model, the tire model in the virtual simulation vehicle model is updated to the nonlinear tire model identified offline. Combined with the virtual simulation vehicle model, the vehicle dynamic response data on roads with different adhesion rates are collected as input and the corresponding road adhesion coefficient is used as output to train the neural network model and construct a neural network identification model for the road adhesion coefficient.

[0043] The neural network model mentioned is also a BP neural network model.

[0044] S104 collects the instantaneous dynamic response of the actual vehicle on the road surface to be identified as the input of the neural network identification model of the road surface adhesion coefficient, so as to realize the real-time online identification of the road surface adhesion coefficient when pure lateral slip or pure longitudinal slip maneuvering is adopted.

[0045] In summary, the tire model and road surface adhesion coefficient collaborative identification method provided by this invention first establishes a virtual simulation vehicle model based on actual vehicle parameters, and configures tire models with different parameters in the virtual simulation vehicle model. Longitudinal braking / lateral steering simulation is performed under road conditions with a first adhesion coefficient to construct a neural network identification model for the tire model. Then, the dynamic response of an actual vehicle equipped with the tire model to be identified under road conditions with a first adhesion coefficient is collected as input to the neural network identification model of the tire model, obtaining the characteristic parameters of the tire model to be identified. Next, combined with the virtual simulation vehicle model, the vehicle dynamic response data on roads with different adhesion rates is collected as input to construct a neural network identification model for the road surface adhesion coefficient. Real-time online identification of the road surface adhesion coefficient is then achieved when pure lateral or pure longitudinal skidding maneuvers are performed. Finally, collaborative identification of the tire model and road surface adhesion coefficient is realized, without the need for specialized equipment and with low implementation cost.

[0046] Please see Figure 2 The tire model and road adhesion coefficient collaborative identification system provided in the second embodiment of the present invention includes:

[0047] The first construction module is used to build a virtual simulation vehicle model based on the parameters of the real vehicle, and configure tire models with different parameters in the virtual simulation vehicle model. It performs longitudinal braking / lateral steering simulation in road conditions with a first adhesion coefficient, uses sensors to collect the dynamic response of the virtual vehicle as input to the neural network training data, and uses the corresponding tire model feature parameters as output to train the neural network model and build a neural network identification model of the tire model.

[0048] The first identification module is used to collect the dynamic response of the actual vehicle equipped with the tire model to be identified in road conditions with a first adhesion coefficient. After classifying the data according to the classification conditions of the training data, it is used as the input of the neural network identification model of the tire model to obtain the feature parameters of the tire model to be identified, so as to determine the nonlinear tire model and realize the offline identification of the tire model.

[0049] The second construction module is used to update the tire model in the virtual simulation vehicle model to the offline identified nonlinear tire model after determining the nonlinear tire model. Combined with the virtual simulation vehicle model, the module collects vehicle dynamic response data on roads with different adhesion rates as input and the corresponding road adhesion coefficient as output to train the neural network model and construct a neural network identification model for the road adhesion coefficient.

[0050] The second identification module is used to collect the instantaneous dynamic response of the actual vehicle on the road surface to be identified as the input of the neural network identification model of the road surface adhesion coefficient, so as to realize the real-time online identification of the road surface adhesion coefficient when pure sideslip or pure longitudinal skidding is adopted.

[0051] In this embodiment, the sensors include at least a virtual IMU, a wheel speed sensor, and a steering wheel angle sensor.

[0052] In this embodiment, the system further includes:

[0053] The classification module is used to classify the data based on the different sensitivities of the influence of the characteristic parameters of different tire models on the longitudinal force-longitudinal slip ratio curve, and then train the neural network model to construct a neural network identification model for the road adhesion coefficient.

[0054] In this embodiment, the virtual simulation vehicle model is the Carsim virtual simulation vehicle model.

[0055] In this embodiment, the neural network model is a BP neural network model.

[0056] The tire model and road surface adhesion coefficient collaborative identification system provided by this invention first establishes a virtual simulation vehicle model based on real vehicle parameters, and configures tire models with different parameters in the virtual simulation vehicle model. Longitudinal braking / lateral steering simulation is performed under road conditions with a first adhesion coefficient to construct a neural network identification model for the tire model. Then, the dynamic response of a real vehicle equipped with the tire model to be identified under road conditions with a first adhesion coefficient is collected as input to the neural network identification model of the tire model, obtaining the characteristic parameters of the tire model to be identified. Next, combined with the virtual simulation vehicle model, the vehicle dynamic response data on roads with different adhesion rates is collected as input to construct a neural network identification model for the road surface adhesion coefficient. Real-time online identification of the road surface adhesion coefficient is then achieved when pure lateral slip or pure longitudinal slip maneuvers are performed. Finally, collaborative identification of the tire model and road surface adhesion coefficient is realized, without the need for specialized equipment and with low implementation cost.

[0057] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0058] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for collaborative identification of tire model and road surface adhesion coefficient, characterized in that, include: A virtual simulation vehicle model is established based on the parameters of the real vehicle, and tire models with different parameters are configured in the virtual simulation vehicle model. Longitudinal braking or lateral steering simulation is performed under road conditions with a first adhesion coefficient. The dynamic response of the virtual vehicle is collected by sensors as input to the neural network training data, and the corresponding tire model feature parameters are used as output to train the neural network model and construct a neural network identification model of the tire model. The dynamic response of a real vehicle equipped with the tire model to be identified is collected under road conditions with a first adhesion coefficient. After classification based on the classification conditions of the training data, it is used as the input of the neural network identification model of the tire model to obtain the feature parameters of the tire model to be identified, so as to determine the nonlinear tire model and realize the offline identification of the tire model. After determining the nonlinear tire model, the tire model in the virtual simulation vehicle model is updated to the nonlinear tire model identified offline. Combined with the virtual simulation vehicle model, the vehicle dynamic response data on roads with different adhesion rates are collected as input and the corresponding road adhesion coefficient is used as output to train the neural network model and construct a neural network identification model for the road adhesion coefficient. The instantaneous dynamic response of the actual vehicle on the road surface to be identified is collected as the input of the neural network identification model of the road surface adhesion coefficient, so as to realize the real-time online identification of the road surface adhesion coefficient when pure sideslip or pure longitudinal skid maneuvers are adopted. The sensors include at least a virtual IMU, a wheel speed sensor, and a steering wheel angle sensor; Before the step of training the neural network model and constructing a neural network identification model for road surface adhesion coefficient, the method further includes: Based on the different sensitivities of the influence of different tire model feature parameters on the tire longitudinal force-longitudinal slip ratio curve, the data is classified and then trained into a neural network model to construct a neural network identification model for road adhesion coefficient.

2. The method for collaborative identification of tire model and road adhesion coefficient according to claim 1, characterized in that, The virtual simulation vehicle model is the Carsim virtual simulation vehicle model.

3. The method for collaborative identification of tire model and road adhesion coefficient according to claim 1, characterized in that, The neural network model is a BP neural network model.

4. A tire model and road surface adhesion coefficient collaborative identification system, characterized in that, include: The first construction module is used to build a virtual simulation vehicle model based on the parameters of the real vehicle, and configure tire models with different parameters in the virtual simulation vehicle model. It performs longitudinal braking or lateral steering simulation in road conditions with a first adhesion coefficient, uses sensors to collect the dynamic response of the virtual vehicle as input to the neural network training data, and uses the corresponding tire model feature parameters as output to train the neural network model and build a neural network identification model of the tire model. The first identification module is used to collect the dynamic response of the actual vehicle equipped with the tire model to be identified in road conditions with a first adhesion coefficient. After classifying the data according to the classification conditions of the training data, it is used as the input of the neural network identification model of the tire model to obtain the feature parameters of the tire model to be identified, so as to determine the nonlinear tire model and realize the offline identification of the tire model. The second construction module is used to update the tire model in the virtual simulation vehicle model to the offline identified nonlinear tire model after determining the nonlinear tire model. Combined with the virtual simulation vehicle model, the module collects vehicle dynamic response data on roads with different adhesion rates as input and the corresponding road adhesion coefficient as output to train the neural network model and construct a neural network identification model for the road adhesion coefficient. The second identification module is used to collect the instantaneous dynamic response of the actual vehicle on the road surface to be identified as the input of the neural network identification model of the road surface adhesion coefficient, so as to realize the real-time online identification of the road surface adhesion coefficient when pure lateral slip or pure longitudinal slip maneuvering is adopted. The sensors include at least a virtual IMU, a wheel speed sensor, and a steering wheel angle sensor; The system also includes: The classification module is used to classify the data based on the different sensitivities of the influence of the characteristic parameters of different tire models on the longitudinal force-longitudinal slip ratio curve, and then train the neural network model to construct a neural network identification model for the road adhesion coefficient.

5. The tire model and road adhesion coefficient collaborative identification system according to claim 4, characterized in that, The virtual simulation vehicle model is the Carsim virtual simulation vehicle model.

6. The tire model and road adhesion coefficient collaborative identification system according to claim 4, characterized in that, The neural network model is a BP neural network model.

Images