A road feel feedback torque estimation method based on unscented kalman filter

Abstract

The application provides a road feeling feedback torque estimation method based on an unscented Kalman filter, comprising the following steps: obtaining three-phase currents, a rotor rotation angle and automobile state parameters of a brushless direct current motor; constructing an unscented Kalman filter through an automobile two-degree-of-freedom model, a Dugoff tire model and a current torque conversion equation, and determining a rack torque according to the three-phase currents, the rotor rotation angle and the automobile state parameters of the motor; calculating an inertia torque, a damping torque, a friction torque and an EPS assisting torque of a steering system according to the automobile state parameters; and calculating a road feeling feedback torque through the rack torque calculated by the unscented Kalman filter and the inertia torque, the damping torque, the friction torque and the EPS assisting torque calculated, compared with the prior art, the application has the advantage that the road feeling feedback is more accurate.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of vehicle's drive-by-wire chassis, in particular to a road feel feedback torque estimation method based on unscented Kalman filter. BACKGROUND

[0002] As a new generation of steering system, the drive-by-wire steering system is quite different from the mechanical steering system. The drive-by-wire steering system removes the steering column and has no direct mechanical connection between the steering wheel and the wheels, so it has the advantages of higher flexibility, higher safety, variable transmission ratio and arbitrary road feel design. However, due to the lack of direct mechanical connection between the steering wheel and the wheels in the system, the driver cannot directly feel the road feedback through the steering wheel of the SBW system as in the traditional steering system. The driver needs to obtain the road conditions and vehicle body state through the steering wheel feel to ensure driving safety, so it is necessary to simulate the real road feel through devices such as motors.

[0003] The road feel feedback of the current drive-by-wire steering system mainly includes two methods, namely the dynamic model method and the parameter fitting method.

[0004] Referring to CN117068259A and CN114954640A, the principle of the dynamic model method is to establish a model, use the values obtained by the popular torque angle sensor, vehicle speed sensor and other devices as input, and directly calculate the road feel feedback torque. This method does not need to install additional sensors, has low cost, and the obtained road feel feedback torque value is close to that of traditional cars, and the driver can quickly adapt. However, this method needs to obtain real-time changing parameters such as tire drag distance and tire load, and it is difficult to obtain the dynamic changing values of these parameters. If the tire drag distance and tire load parameters are constant, the calculated road feel feedback torque will be inaccurate and cannot obtain the real road feel.

[0005] Referring to CN110509983A, the parameter fitting method is a method of fitting the road feel feedback torque by using function fitting, neural network and other methods. This method also does not need to install sensors, and only needs to use the values obtained by the existing sensors to determine the road feel feedback torque value. However, the disadvantage is that this method cannot well reflect the real road feel. SUMMARY

[0006] The present application aims to provide a road feel feedback torque estimation method based on unscented Kalman filter, to solve the problems of difficulty in obtaining tire drag distance and tire load parameters in the dynamic model method, and the problem that the parameter fitting method cannot well reflect the real road feel due to the lack of actual road feel information.

[0007] To achieve the above-mentioned purpose, the following technical solutions are adopted:

[0008] A road feel feedback torque estimation method based on a unscented Kalman filter, comprising the following steps:

[0009] Obtaining three-phase current, rotor angle and vehicle state parameters of the brushless DC motor;

[0010] The unscented Kalman filter is constructed by a two-degree-of-freedom model of the vehicle, a Dugoff tire model and a current-torque conversion equation, and the rack torque is determined according to the three-phase current, rotor angle and vehicle state parameters of the motor;

[0011] The inertia torque, damping torque, friction torque and EPS assist torque of the steering system are calculated according to the vehicle state parameters;

[0012] The rack torque calculated by the unscented Kalman filter is calculated with the calculated inertia torque, damping torque, friction torque and EPS assist torque to calculate the road feel feedback torque.

[0013] Further, the construction of the unscented Kalman filter is as follows:

[0014] Step S1: The values of the three-phase current and rotor angle of the brushless DC motor are transmitted to the observation equation of the unscented Kalman filter, and the observation equation is the current-torque conversion equation of the brushless DC motor, and the observation variable is:

[0015] In the formula, i a , i b , i c are the A, B and C phase currents of the brushless DC motor, and θ r is the current motor rotor angle;

[0016] The values of the front wheel angle and vehicle speed during vehicle motion are transmitted to the system state equation of the unscented Kalman filter, and the system state equation of the unscented Kalman filter includes the Dugoff tire model and the two-degree-of-freedom model;

[0017] The state variable of the unscented Kalman filter is:

[0018] In the formula, v y is the lateral vehicle speed, ω r is the yaw rate, F y is the tire lateral force, and T r ' is the state rack torque of the unscented Kalman filter.

[0019] Step S2: The unscented Kalman filter obtains the Sigma point set at time k according to unscented transformation:

[0020]

[0021] wherein: represents the state variable at time k; P(k|k) represents the covariance matrix at time k; ε represents a scaling parameter; i = represents the i-th Sigma point; n represents the dimension of the state variable;

[0022] Step S3: According to the state variable prediction value calculation formula of the unscented Kalman filter, a set of sampling points X (i) (k+1|k) representing the Gaussian distribution of the set of variables is obtained, and the formula is:

[0023] X (i) (k+1|k) = f[k, X (i) (k|k)]

[0024] wherein: f[k, X (i) (k|k)] represents that X (i) (k|k) is substituted into the two-degree-of-freedom model of the system state equation of the unscented Kalman filter, the Dugoff tire model, and the state variable prediction value calculation formula to calculate.

[0025] Step S4: The expectation of the further predicted state variable is obtained, which is used as the estimation value of the state variable at time k+1, and the formula is:

[0026]

[0027] wherein: ω (i) represents the i-th Sigma point weight;

[0028] According to the variable at time k, the covariance matrix P(k+1|k) at time k+1 is obtained, and the formula is:

[0029]

[0030] wherein: Q represents the covariance matrix of the calculation process error of the state equation;

[0031] Step S5: A new Gaussian distribution sampling point X (i) ’(k+1|k) is obtained using the unscented transformation, and the formula is:

[0032]

[0033] Step S6: The predicted observation Z (i) (k+1|k) is obtained, and the formula is:

[0034] Z (i) (k+1|k) = h[X (i) ’(k+1|k)]

[0035] wherein: h[X(i) (k+1|k) represents the rack torque in the state equation (i) (k+1|k) represents the rack torque in the state equation

[0036] Step S7: Obtain the mean of the predicted observation The formula is:

[0037]

[0038] Calculate the observation covariance matrix The formula is:

[0039]

[0040] In the formula, R represents the error covariance matrix of the observation variable;

[0041] Calculate the covariance matrix of the observation and the state variable The formula is:

[0042]

[0043] Step S8: Calculate the Kalman gain K(k+1), the formula is:

[0044]

[0045] Step S9: Update the output state variable of the unscented Kalman filter at time k+1 Which contains the rack torque T r required for unscented Kalman filter calculation

[0046]

[0047] In the formula, Z(k+1) represents the observation variable obtained by the sensor at time k+1.

[0048] Update the unscented Kalman filter covariance matrix P(k+1|k+1) at time k+1, the formula is:

[0049]

[0050] Further, the observation equation formula is:

[0051]

[0052] In the formula, e a , e b , e c are the back-EMF of the A, B, and C phases of the brushless DC motor, respectively is the motor rotor angle at the previous moment, t is the span time between each moment, T r is the observed rack torque of the unscented Kalman filter.

[0053] Furthermore, the two-degree-of-freedom model of the state equation of the unscented Kalman filter system is:

[0054]

[0055]

[0056] Where C αr is the rear wheel cornering stiffness; C αf is the front wheel cornering stiffness; l r is the distance from the rear axle to the center of mass of the car; l f is the distance from the front axle to the center of mass of the car; m is the mass of the car; I z is the moment of inertia around the longitudinal axis; δ f is the front wheel turning angle; v x is the longitudinal speed; a y is the lateral acceleration;

[0057] The Dugoff tire model of the unscented Kalman filter system state equation is:

[0058] F y '=-C αf tan α f f(λ)

[0059]

[0060]

[0061] Where, F y ' is the tire lateral force estimated by the tire model; F z is the vertical load on the tire; λ is the switching coefficient; μ is the road friction coefficient; f(λ) represents the switching function; α f is the tire slip angle, which is calculated as follows:

[0062]

[0063] Where: β represents the sideslip angle of the center of mass.

[0064] Furthermore, the calculation formula of the predicted value of the unscented Kalman filter state variable in step S3 is:

[0065]

[0066] Where: is the predicted value of lateral vehicle speed; is the yaw rate prediction value; is the tire lateral force prediction value; is the rack torque prediction value; τ m is the mechanical trail; τ p is the pneumatic trail; i r is the steering column to wheel ratio; c represents the tire slack length.

[0067] Further, the friction torque T f is given by the formula:

[0068]

[0069] where x represents the time from the moment when the positive or negative value of the steering wheel speed changes.

[0070] Further, the inertia torque T a is given by the formula:

[0071]

[0072] where I a is the equivalent rotational inertia of the steering wheel and steering column portion; is the steering wheel angular acceleration.

[0073] Further, the damping torque T d is given by the formula:

[0074]

[0075] where: is the steering wheel angular velocity; B d represents the damping coefficient.

[0076] Further, the EPS assist torque T e is given by the formula:

[0077]

[0078] where T s is the steering column torque.

[0079] Further, the final road feel feedback torque T fb is given by the formula:

[0080] T fb = T r " + T f + T a + T d + T e .

[0081] Compared with the prior art, the present application has the following beneficial effects: the road feel feedback torque estimation method of the steer-by-wire system disclosed by the present application adopts an unscented Kalman filter, which is more suitable for a nonlinear model such as an automobile, and comprehensively adopts a method of calculating a road feel torque by means of automobile dynamics and a method of calculating a road feel torque by means of detecting a steering motor current by using a current sensor, so that a more accurate road feel feedback torque can be estimated, thereby providing reliable road feel information for a driver. BRIEF DESCRIPTION OF DRAWINGS

[0082] Figure 1 A flowchart of the road feel feedback torque estimation method of the steer-by-wire system provided for the first embodiment of the present application is shown in the figure.

[0083] Figure 2 A flowchart of the road feel feedback torque estimation method of the steer-by-wire system provided for the second embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0084] To make the technical problems solved by the present application, the technical solutions adopted and the technical effects reached more clear, the technical solutions of the embodiments of the present application will be further described in detail below with reference to the drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0085] Embodiment one:

[0086] The present embodiment provides a road feel feedback torque estimation method based on an unscented Kalman filter, as shown in the figure, which specifically comprises the following steps: Figure 1

[0087] S11, determining three-phase currents of a brushless direct current motor, a rotor rotation angle and automobile state parameters;

[0088] S12, constructing an unscented Kalman filter by means of a two-degree-of-freedom model of an automobile, a Dugoff tire model and a current-torque conversion equation, and determining a rack torque according to the three-phase currents of the motor, the rotor rotation angle and the automobile state parameters;

[0089] S13, calculating a road feel feedback torque by means of the rack torque calculated by the unscented Kalman filter, an inertia torque of a steering system, a damping torque, a friction torque and an EPS assist torque.

[0090] In step S11, the brushless direct current motor rotates according to a rotation angle signal of a steering wheel, at which time the three-phase currents of the brushless direct current motor can be obtained by means of a current sensor; the rotor rotation angle can be obtained by means of a motor Hall sensor.

[0091] ​In step S12, the constructed unscented Kalman filter needs to take the three-phase current of the brushless direct current motor, the rotor rotation angle and the vehicle state parameters as inputs, and then calculate the required rack torque. Among them, the three-phase current of the brushless direct current motor and the rotor rotation angle are used as observation values for the observation equation of the unscented Kalman filter, and the vehicle state parameters are used for the state equation of the unscented Kalman filter. By calculating the Kalman gain, the most real rack torque can be calculated from the observed rack torque and the rack torque obtained from the state equation.

[0092] Specifically, the values of the three-phase current and the rotor rotation angle of the brushless direct current motor are transmitted to the observation equation of the unscented Kalman filter, and the observation variable is:

[0093]

[0094] The unscented Kalman filter includes a system state equation and an observation equation. The observation equation is the current torque conversion equation of the brushless direct current motor, and its formula is:

[0095]

[0096] In the formula, i a , i b , i c are the A, B and C phase currents of the brushless direct current motor respectively; e a , e b , e c are the back electromotive forces of the A, B and C phases of the brushless direct current motor respectively; θ r is the current motor rotor rotation angle; is the motor rotor rotation angle at the last time; t is the time span at each time; T r is the observed rack torque of the unscented Kalman filter.

[0097] The values of the front wheel rotation angle and the vehicle speed in the vehicle motion process are transmitted to the system state equation of the unscented Kalman filter. The front wheel rotation angle signal can be obtained through the rotation angle sensor, and the vehicle speed signal can be obtained through the vehicle speed sensor. The required vehicle speed includes the longitudinal vehicle speed.

[0098] The system state equation of the unscented Kalman filter includes Du g off tire model, two-degree-of-freedom model, and the two-degree-of-freedom model is:

[0099]

[0100]

[0101] In the formula, C αr is the rear wheel cornering stiffness; C αf is the front wheel cornering stiffness; lr is the distance from the rear axle to the center of mass of the vehicle; l f is the distance from the front axle to the center of mass of the vehicle; m is the mass of the vehicle; I z is the moment of inertia about the longitudinal axis; δ f is the front wheel steering angle; ω r is the yaw rate; v x is the longitudinal vehicle speed; v y is the lateral vehicle speed; a v is the lateral acceleration; is the yaw angular acceleration.

[0102] tire side slip angle α f The formula is:

[0103]

[0104] The Dugoff tire model of the state equation of the unscented Kalman filter system is:

[0105] F y ' = -C αf tan α f f (λ)

[0106]

[0107]

[0108] In the formula, F y ' is the tire lateral force estimated by the tire model; F z is the tire vertical load; λ is the switching coefficient; μ is the road surface friction coefficient; f (λ) represents the switching function; α f is the tire side slip angle, and the calculation formula is:

[0109]

[0110] In the formula: β represents the center of mass side slip angle.

[0111] The state variable of the unscented Kalman filter is:

[0112]

[0113] In the formula, F y is the tire lateral force, T r ' is the state rack torque;

[0114] The Sigma point set at time k obtained by the unscented Kalman filter according to the unscented transformation is:

[0115]

[0116] In the formula: X(k|k) represents the state variable at time k; P(k|k) represents the covariance matrix at time k; ε represents the scaling parameter; i represents the i-th Sigma point; n represents the dimension of the state variable.

[0117] The formula for calculating the predicted value of the state variable of the unscented Kalman filter is:

[0118]

[0119] In the formula: is the predicted value of the lateral vehicle speed; is the predicted value of the yaw rate; is the predicted value of the tire lateral force; is the predicted value of the rack torque; τ m is the mechanical trail; τ p is the pneumatic trail; i r is the steering column-to-wheel transmission ratio; c represents the tire slack length.

[0120] According to the above formula for calculating the predicted value of the state variable, a set of sampling points X (i) (k+1|k) representing the Gaussian distribution of the set of variables is obtained, and the formula is:

[0121] X (i) (k+1|k) = f[k, X (i) (k|k)]

[0122] In the formula: f[k, X (i) (k|k)] represents substituting X (i) (k|k) into the two-degree-of-freedom model of the system state equation of the unscented Kalman filter, the Dugoff tire model, and the formula for calculating the predicted value of the state variable at time k.

[0123] Further predict the expected value of the state variable, and take it as the estimated value of the state variable at time k+1, and the formula is:

[0124]

[0125] In the formula: ω (i) represents the i-th Sigma point weight.

[0126] According to the variable at time k, the covariance matrix P(k+1|k) at time k+1 is obtained, and the formula is:

[0127]

[0128] In the formula: Q represents the covariance matrix of the calculation process error of the state equation.

[0129] Use the unscented transformation to obtain new Gaussian distribution sampling points X(i) (k+1|k), whose formula is:

[0130]

[0131] The predicted observation Z (i) (k+1|k), whose formula is:

[0132] Z (i) (k+1|k) = h[X (i) (k+1|k)]

[0133] In the formula, h[X (i) (k+1|k)] represents that the rack torque in X (i) (k+1|k) is brought into the observation equation for calculation.

[0134] According to the predicted observation, its mean value is obtained whose formula is:

[0135]

[0136] The observation covariance matrix is calculated whose formula is:

[0137]

[0138] In the formula, R represents the error covariance matrix of the observation variable;

[0139] The covariance matrix of the observation and the state variable is calculated whose formula is:

[0140]

[0141] The Kalman gain K(k+1) is calculated, whose formula is:

[0142]

[0143] The output state variable of the unscented Kalman filter at time k+1 is updated which contains the rack torque T r required for unscented Kalman filter calculation

[0144]

[0145] In the formula, Z(k+1) represents the observation variable obtained by the sensor at time k+1.

[0146] The unscented Kalman filter covariance matrix P(k+1|k+1) at time k+1 is updated, whose formula is:

[0147]

[0148] In step S13, the EPS assist torque, the friction torque, the damping torque, the inertia torque and the rack torque are included in the conventional steering system, the steering wheel angular velocity, the steering wheel angular acceleration, the steering column torque and the longitudinal vehicle speed are taken as the inputs, the EPS assist torque, the friction torque, the damping torque and the inertia torque can be calculated, the rack torque has been obtained through the unscented Kalman filter, and the final road feel feedback torque can be calculated by synthesizing these torques.

[0149] Specifically, the friction torque T f is calculated, and the formula is as follows:

[0150]

[0151] In the formula, x represents the time that starts timing when the positive and negative values of the steering wheel rotational speed change.

[0152] The inertia torque T a is calculated, and the formula is as follows:

[0153]

[0154] In the formula, I a is the equivalent rotational inertia of the steering wheel and the steering column part; is the steering wheel angular acceleration.

[0155] The damping torque T d is calculated, and the formula is as follows:

[0156]

[0157] In the formula, ω is the steering wheel angular velocity; B d represents the damping coefficient.

[0158] The EPS assist torque T e is calculated, and the formula is as follows:

[0159]

[0160] In the formula, T s is the steering column torque.

[0161] The final road feel feedback torque T fb is calculated, and the formula is as follows:

[0162] T fb = T r ”+T f +T a +T d +T e

[0163] After the road feel feedback torque is obtained, it is provided to the road feel motor, so that the driver can feel the road information through the steering wheel.

[0164] Embodiment two:

[0165] The embodiment provides a road feel feedback torque estimation method based on a unscented Kalman filter, as shown in the formula (1), comprising the following steps: Figure 2

[0166] S21, three-phase current, rotor rotation angle and automobile state parameters of the brushless direct current motor are determined, and the automobile state parameters include steering wheel angular velocity, steering wheel angular acceleration, steering column torque and longitudinal vehicle speed;

[0167] S22, the damping torque, the inertia torque, the friction torque and the EPS assisting torque are obtained;

[0168] S23, the unscented Kalman filter is constructed through the automobile two-degree-of-freedom model, the Dugoff tire model and the current torque conversion equation, and the rack torque is determined according to the motor three-phase current, the rotor rotation angle and the automobile state parameters;

[0169] S24, the road feel feedback torque is synthesized.

[0170] The difference between the embodiment and the embodiment one is that the embodiment first calculates the damping torque, the inertia torque, the friction torque and the EPS assisting torque according to the steering wheel angular velocity, the steering wheel angular acceleration, the steering column torque and the longitudinal vehicle speed, and then calculates the rack torque through the unscented Kalman filter, while the embodiment one first calculates the rack torque, and then calculates the damping torque, the inertia torque, the friction torque and the EPS assisting torque. Specifically, the related formulas of the damping torque, the inertia torque, the friction torque, the EPS assisting torque and the rack torque are completely consistent with the embodiment one.

[0171] The above is the preferred embodiment of the application, and any change made according to the technical solution of the application, as long as the function generated does not exceed the scope of the technical solution of the application, belongs to the protection scope of the application.​

Claims

1. A method for estimating road feel feedback torque based on an unscented Kalman filter, characterized by, The method comprises the following steps: Obtaining three-phase current, rotor rotation angle and automobile state parameters of the brushless direct current motor; Building an unscented Kalman filter through a two-degree-of-freedom model of the automobile, a Dugoff tire model and a current-torque conversion equation, and determining the rack torque according to the three-phase current, rotor rotation angle and automobile state parameters of the motor; Calculating the inertia torque, damping torque, friction torque and EPS assisting torque of the steering system according to the automobile state parameters; Calculating the road feeling feedback torque according to the rack torque calculated by the unscented Kalman filter and the calculated inertia torque, damping torque, friction torque and EPS assisting torque; The unscented Kalman filter is built as follows: Step S1: transmit the three-phase current of the brushless direct current motor and the numerical value of the rotor rotation angle to the observation equation of the unscented Kalman filter, the observation equation is the current-torque conversion equation of the brushless direct current motor, and the observation variable is: where i a , i b , i c are the brushless DC motor A, B, C phase currents, and θ r is the current motor rotor angle; Transmitting the values of the front wheel rotation angle and the vehicle speed in the automobile motion process to the system state equation of the unscented Kalman filter, wherein the system state equation of the unscented Kalman filter comprises a Dugoff tire model and a two-degree-of-freedom model; The state variables of the unscented Kalman filter are: where v y is the lateral vehicle speed, ω r is the yaw rate, F y is the tire lateral force, T r is the state rack force moment of the unscented Kalman filter; Step S2: obtaining the Sigma point set at time k by the unscented Kalman filter according to unscented transformation: In the formula: represents the state variable at time k; P(k|k) represents the covariance matrix at time k; ε represents a scaling parameter; i = represents the i-th Sigma point; n represents the dimension of the state variable; Step S3: According to the state variable prediction formula of the unscented Kalman filter, a set of sampling points X representing the Gaussian distribution of the set of variables is obtained (i) (k+1|k), whose formula is: X (i) (k+1|k) = f[k, X (i) (k|k)] where: f[k, X (i) (k|k)] represents at time k, X (i) (k|k) is substituted into the two-degree-of-freedom model of the system state equation of the unscented Kalman filter, the Dugoff tire model, and the state variable prediction value calculation formula. Step S4: further predicting the expected state variable and taking it as the estimated value of the state variable at time k+1, and the formula is: where ωi (i) denotes the i-th Sigma point weight; Obtaining the covariance matrix P(k+1|k) at time k+1 according to the variable at time k, and the formula is: In the formula, Q represents the covariance matrix of the state equation calculation process error; Step S5: Obtain new Gaussian distribution sampling points X using unscented transformation (i) (k+1|k), whose formula is: Step S6: Calculate the predicted observation Z (i) (k+1|k), which is given by Z (i) (k+1|k) = h[X (i) ’(k+1|k)] where: h[X (i) (k + 1 | k) represents the rack torque variable in the observation equation. (i) (k + 1 | k) represents the rack torque variable in the observation equation. Step S7: Obtain the mean of the predicted observation The formula is: Computing the observation covariance matrix The formula is: In the formula, R represents the error covariance matrix of the observation variable; computing the covariance matrix of the observations and the state variables which is given by Step S8: calculating the Kalman gain K(k+1), and the formula is: Step S9: update the output state variable of the unscented Kalman filter at time k+1 where T is the rack torque required for the unscented Kalman filter calculation r , whose formula is: In the formula, Z(k+1) represents the observation variable obtained by the sensor at time k+1; Updating the covariance matrix P(k+1|k+1) of the unscented Kalman filter at time k+1, and the formula is:

2. The road feel feedback torque estimation method based on unscented Kalman filter according to claim 1, characterized in that, The observation equation formula is: In the formula, e a , e b , e c are back-EMF of brushless DC motor A, B, C phase respectively, is the rotor angle of the motor at the last time, t is the time span between each time, T r is the observed rack torque of the unscented Kalman filter.

3. The road feel feedback torque estimation method based on unscented Kalman filter according to claim 2, characterized in that, The two-degree-of-freedom model of the system state equation of the unscented Kalman filter is: where C αr is the rear wheel cornering stiffness; C αf is the front wheel cornering stiffness; l r is the distance from the rear axle to the center of mass of the vehicle; l f is the distance from the front axle to the center of mass of the vehicle; m is the mass of the vehicle; I z is the moment of inertia about the longitudinal axis; δ f is the front wheel steering angle; v x is the longitudinal vehicle speed; a y is the lateral acceleration; The Dugoff tire model of the system state equation of the unscented Kalman filter is: F y ,=-C αf tanα f f(λ) Where, F y ' is the tire lateral force estimated by the tire model; F z is the vertical load on the tire; λ is the switching coefficient; μ is the road friction coefficient; f(λ) represents the switching function; α f is the tire slip angle, which is calculated as follows: In the formula, β represents the center of mass side slip angle.

4. The road feel feedback torque estimation method based on unscented Kalman filter according to claim 3, characterized in that, The calculation formula of the state variable prediction value of the unscented Kalman filter in step S3 is: where: is the lateral vehicle speed prediction; is the yaw rate prediction; is the tire lateral force prediction; is the rack torque prediction; τ m is the mechanical trail; τ p is the air trail; i r is the steering column to wheel ratio; c represents the tire slack length.

5. The road feel feedback torque estimation method based on unscented Kalman filter according to claim 4, characterized in that, Friction torque T f The formula is: In the formula, x represents the time when the positive and negative values of the steering wheel rotation speed change.

6. The road feel feedback torque estimation method based on unscented Kalman filter according to claim 5, characterized in that, Moment of inertia T a The formula is: wherein: I a is the equivalent moment of inertia of the steering wheel column portion; is the steering wheel angular acceleration.

7. The road feel feedback torque estimation method based on unscented Kalman filter according to claim 6, characterized in that, Damping torque T d The formula is: In the formula: B is the steering wheel angular velocity d denotes the damping coefficient.

8. The road feel feedback torque estimation method based on unscented Kalman filter according to claim 7, characterized in that, EPS assist torque T e The formula is: In the formula, T s is the steering column torque.

9. The road feel feedback torque estimation method based on unscented Kalman filter according to claim 8, characterized in that, The final road feel feedback torque T fb The calculation formula is: T fb = T r ' + T f + T a + T d + T e .

Images