Road surface μ estimation device

Abstract

The present invention provides a road surface μ estimation device that can improve the accuracy of road surface μ estimation. [Solution] The road surface μ estimation device 10 has a state prediction unit 20 that predicts changes in the state of the vehicle using a state transition model, the state variables of the state prediction unit 20 include the friction coefficient, the state transition matrix of the state prediction unit 20 includes the force generated by each tire, and the state prediction unit 40 has a tire force estimation unit 40 that combines the prediction data including the friction coefficient of the state prediction unit 20 and the observation data of the vehicle 100 and feeds it back to the state prediction unit using a Kalman gain 30 to estimate the force generated by each tire using a magic formula.

Description

[Technical Field]

[0001] This disclosure relates to a device for estimating the coefficient of friction μ between a vehicle tire and a road surface. [Background technology]

[0002] Automated driving systems that allow vehicles to operate automatically without human intervention are well known. At Level 3 and above of automated driving systems, the system takes over the driving function, thus requiring an expansion of the operational range of the automated driving system. In this case, in order for the vehicle to operate safely, it is necessary to correctly estimate the friction coefficient between the vehicle's tires and the road surface (hereinafter referred to as road surface μ) and control the vehicle's operation so as not to exceed the performance limits of the tires. For example, Patent Document 1 discloses a method for estimating road surface μ using an extended Kalman filter. [Prior art documents] [Patent Documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2021-140540 [Overview of the project] [Problems that the invention aims to solve]

[0004] However, the road surface μ estimation method in Patent Document 1 can only estimate the road surface μ when the vehicle speed is above a predetermined speed and when turning. Furthermore, the accuracy of tire force estimation is low, resulting in low accuracy of road surface μ estimation.

[0005] Therefore, the present invention aims to provide a road surface μ estimation device that can improve the estimation accuracy of road surface μ. [Means for solving the problem]

[0006] A road surface μ estimation device according to the present invention is a road surface μ estimation device that estimates a friction coefficient between a tire of a four-wheeled vehicle and a road surface, the device comprising a state prediction unit that predicts a state change of the vehicle based on a state transition model, wherein a state variable of the state prediction unit includes the friction coefficient, a state transition matrix of the state prediction unit includes each tire generated force, prediction data including the friction coefficient from the state prediction unit and observation data of the vehicle are combined and fed back to the state prediction unit using Kalman gain, and the device comprises a tire generated force estimation unit that estimates each tire generated force using a Magic Formula equation.

[0007] In the road surface μ estimation device according to the present invention, it is preferable that the tire generated force estimation unit estimates each tire generated force using a vertical load of each tire and the Magic Formula equation, and comprises a load transfer calculation unit that calculates the vertical load of each tire using longitudinal acceleration and lateral acceleration of the vehicle among the prediction data from the state prediction unit.

[0008] In the road surface μ estimation device according to the present invention, it is preferable that the device comprises a suspension K&C calculation unit that calculates a corrected steering angle of a steering angle input to the state prediction unit using each of the tire generated forces, a slip angle of each of the tires, and lateral acceleration of the vehicle. [Effects of the Invention]

[0009] According to the road surface μ estimation device and the road surface μ estimation method of the present invention, the estimation accuracy of road surface μ can be improved. [Brief Description of the Drawings]

[0010] [Figure 1] It is a block diagram showing a road surface μ estimation device which is an example of an embodiment. [Figure 2] It is a schematic diagram showing an equation of motion in the longitudinal direction of a four-wheeled vehicle model. [Figure 3] It is a schematic diagram showing an equation of motion in the lateral direction of a four-wheeled vehicle model. [Figure 4] It is a schematic diagram showing an equation of motion for a slip angle of a four-wheeled vehicle model. [Figure 5]It is a flow diagram showing the calculation flow of the suspension K&C calculation unit. [Figure 6] It is a schematic diagram showing a suspension displacement model. [Figure 7] It is a schematic diagram showing a KingPin pivot model. [Figure 8] It is a schematic diagram showing a tilted vehicle model. DETAILED DESCRIPTION OF THE INVENTION

[0011] Hereinafter, an exemplary embodiment of the present invention will be described in detail. In the following description, specific shapes, materials, directions, numerical values, etc. are exemplifications for facilitating the understanding of the present invention, and can be appropriately changed according to the application, purpose, specifications, etc.

[0012] [Road surface μ estimating apparatus] A road surface μ estimating apparatus 10 according to an exemplary embodiment will be described with reference to FIG. 1.

[0013] The road surface μ estimating apparatus 10 estimates the coefficient of friction between the tires of a vehicle 100 and the road surface (hereinafter referred to as road surface μ). The road surface μ estimating apparatus 10 may be mounted on the vehicle 100, or may not be mounted on the vehicle 100 and connected to the vehicle 100 via a network. According to the road surface μ estimating apparatus 10, the estimation accuracy of road surface μ can be improved, the details of which will be described later.

[0014] To the road surface μ estimating apparatus 10, wheel speed, vehicle speed, yaw rate, vehicle braking force (accelerator, brake), and steering angle detected by sensors or the like while the vehicle 100 is traveling are input. The traveling operation of the vehicle 100 is not limited to that performed by a human, and may be an operation by a driving robot, or may be an operation by a system such as AD-ADAS.

[0015] The road surface μ estimation device 10 has a CPU (Central Processing Unit) which is an arithmetic processing unit, and memory units such as RAM (Random Access Memory) and ROM (Read Only Memory). It performs signal processing according to a program pre-stored in ROM while utilizing the temporary storage function of RAM.

[0016] The road surface μ estimation device 10 includes a state prediction unit 20, a Kalman gain 30, a tire force estimation unit 40, a load transfer calculation unit 50, and a suspension K&C calculation unit 60, each of which will be described in detail later. The state prediction unit 20, Kalman gain 30, tire force estimation unit 40, load transfer calculation unit 50, and suspension K&C calculation unit 60 are realized by the CPU executing a program stored in ROM or RAM.

[0017] According to the road surface μ estimation device 10, the state prediction unit 20 predicts the state changes of the vehicle 100 using a state transition model. The state variables of the state transition model include the friction coefficient μ, and the state transition matrix of the state transition model includes the forces generated by each tire. Then, the predicted data including the friction coefficient μ of the state transition model and the observed data of the vehicle 100 are combined and fed back to the state transition model using the Kalman gain 30. This improves the accuracy of the road surface μ estimation.

[0018] Furthermore, according to the road surface μ estimation device 10, the tire force estimation unit 40 uses a magic formula to estimate each tire force, which is an element of the state transition matrix of the state transition model.

[0019] Furthermore, according to the road surface μ estimation device 10, the load transfer calculation unit 50 calculates the vertical load of each tire required by the magic formula.

[0020] Furthermore, according to the road surface μ estimation device 10, the suspension K&C calculation unit 60 corrects the sensed steering angle by considering the influence of the forces generated by the left and right tires, the self-alliance torque moment, the moment around KingPin, etc., and uses it as input data for the state transition model.

[0021] In the following, abbreviations may be used in mathematical formulas and other explanations. Here, the abbreviations for physical quantities used in each claim are represented as shown in Formula 1 below.

number

[0022] [State prediction unit] The state prediction unit 20 will be explained using Figures 2 to 4.

[0023] The state prediction unit 20 predicts the state change of the vehicle 100 using a state transition model. The state transition model uses equation 2 to determine the vehicle speed V k Each wheel speed ω flk , ω frk , ω rlk , ω rrk The yaw rate r, vehicle slip angle β, friction coefficient μ, and lateral acceleration are represented as state variables. The friction coefficient μ is included in the state variables of this state transition model. In addition, the state transition matrix of the state transition model includes the front and rear tire generated forces F. x Force generated by left and right tires F y It includes.

number

[0024] Furthermore, the abbreviations for the physical quantities used in Equation 2 are expressed as shown in Equation 3 below.

number

[0025] Equation 2 is calculated using the following procedure. First, as illustrated in Figure 2, the following Equation 4 is obtained from the longitudinal balance of the four-wheeled vehicle model 21.

number

[0026] Furthermore, by focusing on the balance of lateral forces and moment in the vehicle yaw direction of the four-wheeled vehicle model 21 illustrated in Figure 3, the following equation 5 is obtained.

number

[0027] Furthermore, by focusing on the slip angle of the four-wheeled vehicle model 21 illustrated in Figure 4, the following equation 6 is obtained.

number

[0028] Furthermore, the abbreviations for the physical quantities used in Equation 6 are expressed as shown in Equation 7 below.

number

[0029] From the above, combining equations 4, 5, and 6 yields the aforementioned equation 2.

[0030] [Kalman filter] Let's explain the Kalman gain of 30 again using Figure 1.

[0031] The prediction data (wheel speed, vehicle speed, yaw rate, road surface μ) including the friction coefficient μ from the state prediction unit 20 and the observed vehicle data (wheel speed, vehicle speed, yaw rate) are combined and fed back to the state prediction unit 20 using the Kalman gain 30. The Kalman gain 30 is set appropriately according to the respective reliability of the measured physical quantity and the calculated physical quantity for each physical quantity.

[0032] [Tire Force Estimation Unit] Let's explain the tire force estimation unit 40 again using Figure 1.

[0033] The tire generated force estimating unit 40 estimates the tire generated force of each tire using the magic formula formula and inputs the result to the state predicting unit 20. In the tire generated force estimating unit 40, the wheel speed, vehicle speed, tire slip angle and road surface μ predicted by the state predicting unit 20, and the vertical load F of each tire calculated by the load transfer calculating unit 50 z and the magic formula formula are used to obtain the tire generated force F of each tire x , F y is estimated. Detailed description of the magic formula formula is omitted.

[0034] According to the tire generated force estimating unit 40, the tire generated force of each tire in the four-wheel vehicle model 21 can be estimated with high accuracy.

[0035] [Load Transfer Calculating Unit] The load transfer calculating unit 50 will be described with reference to FIG. 5.

[0036] The load transfer calculating unit 50 calculates the vertical load F of each tire z respectively. More specifically, the vertical load F of each tire is calculated from the longitudinal acceleration and lateral acceleration of the vehicle 100 predicted by the state predicting unit 20 z respectively, and the calculation results are input to the tire generated force estimating unit 40.

[0037] First, similarly to the suspension K&C calculating unit 60 described later, the front roll height h is calculated from the lateral acceleration f and the rear roll height h r are calculated.

[0038] Referring to the inclined vehicle model 51 in a stopped state on an arbitrary inclination θ exemplified in FIG. 5, the front load w is obtained as shown in Mathematical Formula 8 f , the rear load w r is obtained. [Mathematical Formula]

[0039] Further, as shown in Mathematical Formula 9, the lateral load transfer amount ΔW of the front tire f _lateral, lateral load transfer amount ΔW of the rear wheel r_lateral , can be obtained.

number

[0040] Furthermore, as shown in equation 10, the longitudinal load transfer amount ΔW of the front tire. f_longi ΔW, the amount of load transfer in the longitudinal direction of the rear tire. r_longi You can obtain this.

number

[0041] Based on the above, the vertical load F of each tire is as shown in Equation 11. z You can obtain this.

number

[0042] [SUS K&C Calculation Department] The SUS K&C calculation unit 60 will be explained using Figures 6 to 8.

[0043] The suspension K&C calculation unit 60 calculates the corrected steering angle according to the procedure illustrated in Figure 6. The suspension K&C calculation unit 60 calculates the roll angle φ from a map showing the correlation between the lateral acceleration of the vehicle 100 and the roll angle φ. Referring to the suspension displacement model 61 illustrated in Figure 7, the following equation 12 is obtained.

number

[0044] The suspension K&C calculation unit 60 calculates the front roll height h from the suspension displacement z obtained by equation 12 and a map showing the correlation between the suspension displacement z and the front roll center height. f Similarly, the suspension K&C calculation unit 60 calculates the rear roll height h from the suspension displacement z, a map showing the correlation between the suspension displacement z and the rear roll center height. r Calculate.

[0045] Furthermore, the suspension K&C calculation unit 60 calculates the front corrected steering angle from the suspension displacement amount z obtained in equation 12 and a map showing the correlation between the suspension displacement amount z and the front corrected steering angle. Similarly, the suspension K&C calculation unit 60 calculates the front corrected steering angle from the suspension displacement amount z and a map showing the correlation between the suspension displacement amount z and the front corrected steering angle.

[0046] The suspension K&C calculation unit 60 calculates the tire force F generated by each tire on the left and right sides. y And the tire force F generated by each tire on the left and right sides. y The corrected steering angle for each tire is calculated from a map showing the correlation between the corrected steering angle and the corrected steering angle.

[0047] The suspension K&C calculation unit 60 calculates the slip angle β of each tire and the self-alliance torque M of each tire. z From the map showing the correlation with the self-alliance torque M of each tire, z Calculate the self-alliance torque M of each tire. z And the self-alliance torque of each tire z The corrected steering angle for each tire is calculated from a map showing the correlation between the tire's position and the corrected steering angle for each tire.

[0048] From the KingPin model 62 illustrated in Figure 8, the following equation 13 is obtained, and the moment M around KingPin is obtained. kp Calculate.

number

[0049] The suspension K&C calculation unit 60 calculates the moment M around KingPin. kp And the moment M around KingPin kp The corrected steering angle for each tire is calculated from a map showing the correlation between the corrected steering angle and the tire.

[0050] The suspension K&C calculation unit 60 adds up all of the above-mentioned corrected steering angles and calculates the corrected steering angle KC for each tire. The output is as follows: The steering angle sensed from the vehicle 100 is input to the state prediction unit 20 with the correction steering angle KC added to it.

[0051] It should be noted that the present invention is not limited to the embodiments and their modifications described above, and various changes and improvements are possible within the scope of the claims of this application. [Explanation of symbols]

[0052] 10 Road surface μ estimation device, 20 State prediction unit, 21 4-wheel vehicle model, 30 Karman gain, 40 Tire force generation estimation unit, 41 Inclined vehicle model, 50 Load transfer calculation unit, 60 K&C calculation unit, 61 Suspension displacement model, 62 KingPin rotation model

Claims

1. A road surface μ estimation device for estimating the coefficient of friction between the tires of a four-wheeled vehicle and the road surface, It has a state prediction unit that predicts the state changes of the vehicle using a state transition model, The state variables of the state transition model include the friction coefficient, The state transition matrix of the aforementioned state transition model includes the force generated by each tire, The prediction data including the friction coefficient of the state transition model and the observation data of the vehicle are combined and fed back to the state transition model using the Kalman gain. It has a tire force estimation unit that estimates the force generated by each tire using a magic formula. Road surface μ estimation device.

2. The tire force estimation unit estimates the force generated by each tire using the vertical load of each tire and the magic formula equation. The system includes a load transfer calculation unit that calculates the vertical load on each tire using the longitudinal acceleration and lateral acceleration of the vehicle from the predicted data of the state prediction unit. The road surface μ estimation device according to claim 1.

3. The system includes a suspension K&C calculation unit that calculates a corrected steering angle for the steering angle input to the state prediction unit using the force generated by each tire, the slip angle of each tire, and the lateral acceleration of the vehicle, which are among the predicted data from the state prediction unit. The road surface μ estimation device according to claim 1.

Images