JP2011057036A - Road surface friction coefficient estimation device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To continuously estimate a road surface friction coefficient with high accuracy by a natural value without applying complicated processing such as weighting even in a motion state of any vehicle. <P>SOLUTION: This road surface friction coefficient estimation device calculates actually occurring an estimation rack driving force Fr_star, estimation wheel braking/driving force Fx_star, and estimation lateral force Fy_star, calculates a reference rack driving force Fr_model, a reference wheel braking/driving force Fx_model, and a reference lateral force Fy_model by a brush model of a tire including the road surface friction coefficient μ and a parameter λ representing a lengthwise strain and a lateral strain of the tire as parameters, and obtains, by optimization calculation, a value of the road surface friction coefficient μ such that a difference between the estimation rack driving force Fr_star and the reference rack driving force Fr_model, a difference between the estimation wheel braking/driving force Fx_star and the reference wheel braking/driving force Fx_model, and a difference between the estimation lateral force Fy_star and the reference lateral force Fy_model become the minimum. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a road surface friction coefficient estimating device that appropriately estimates a road surface friction coefficient even in various vehicle motion states.

  In recent years, various control techniques for traction control, braking force control, torque distribution control, and the like have been proposed and put into practical use in vehicles. Many of these techniques use a road surface friction coefficient for calculation or correction of a necessary control amount, and it is necessary to estimate an accurate road surface friction coefficient in order to reliably execute the control. For example, Japanese Patent No. 3393654 (hereinafter referred to as Patent Document 1) estimates the first road surface friction coefficient based on the lateral acceleration acting on the vehicle body. A vehicle slip control device that estimates a second road surface friction coefficient based on a longitudinal acceleration acting, compares the first road surface friction coefficient with a second road surface friction coefficient, and selects the higher road surface friction coefficient as the road surface friction coefficient. Is disclosed.

Japanese Patent No. 3393654

  However, in the road surface friction coefficient estimation method disclosed in Patent Document 1 described above, the higher road surface friction coefficient is adopted by comparing the first road surface friction coefficient and the second road surface friction coefficient. Depending on the motion state of the vehicle, there is a problem that the value of the road surface friction coefficient fluctuates intermittently, making it difficult to perform natural and smooth control. Moreover, in the road surface friction coefficient estimation method disclosed in Patent Document 1 described above, the first road surface friction coefficient calculated from the lateral motion information does not take into account the longitudinal strain of the tire, and is calculated from the longitudinal motion information. 1 Road surface friction coefficient does not take into account the lateral strain of the tire, so the estimation accuracy of the road surface friction coefficient becomes worse during cornering driving and cornering braking that require both tire longitudinal strain and lateral strain. There is. Furthermore, in the road surface friction coefficient estimation method disclosed in Patent Document 1 described above, when the road surface friction coefficient is finally determined, it is necessary to weight the vehicle according to conditions such as vehicle body speed, longitudinal acceleration, lateral acceleration, and the like. There is also a problem that becomes complicated.

  The present invention has been made in view of the above circumstances, and in any vehicle motion state, even when turning control is performed, the road surface friction coefficient is continuously calculated without adding complicated processing such as weighting. An object of the present invention is to provide a road surface friction coefficient estimating apparatus that can accurately estimate a natural value.

  The present invention relates to a tire brush model including, as parameters, estimated rack thrust calculation means for calculating an actually generated rack thrust as an estimated rack thrust, parameters representing at least a longitudinal strain and a lateral strain of the tire, and a road surface friction coefficient. A reference rack thrust calculation means for calculating a rack thrust calculated by the above as a reference rack thrust, an estimated wheel braking / driving force calculation means for calculating an actual wheel braking / driving force as an estimated wheel braking / driving force, and at least a tire A reference wheel braking / driving force calculating means for calculating a wheel braking / driving force calculated as a reference wheel braking / driving force as a reference wheel braking / driving force by using a tire brush model including parameters representing longitudinal strain and lateral strain and a road surface friction coefficient, and Estimated lateral force calculating means for calculating the generated lateral force as an estimated lateral force, and at least the tire Reference lateral force calculation means for calculating a lateral force calculated by a tire brush model including a parameter expressing longitudinal strain and lateral strain and a road surface friction coefficient as parameters, and the estimated rack thrust and the reference rack thrust. Of the road surface friction coefficient and the estimated wheel braking / driving force and the estimated wheel force / driving force and the estimated lateral force and the reference lateral force are minimized by optimization calculation. The road surface friction index estimation means to be obtained is provided.

  According to the road surface friction coefficient estimating apparatus according to the present invention, the road surface friction coefficient is continuously and naturally obtained without adding complicated processing such as weighting in any vehicle motion state, even during turning control. It is possible to estimate with high accuracy with high accuracy.

It is a functional block diagram of the road surface friction coefficient estimation apparatus by one Embodiment of this invention. It is a flowchart of the road surface friction coefficient estimation program by one Embodiment of this invention. FIG. 3 is a flowchart continued from FIG. 2 according to an embodiment of the present invention. It is explanatory drawing of (lambda) by one Embodiment of this invention.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

In FIG. 1, reference numeral 1 denotes a road surface friction coefficient estimating device for estimating a road surface friction coefficient. Four wheel speeds (left front wheel speed Vfl, right front wheel speed Vfr, left rear wheel speed Vrl, right rear wheel) A four-wheel wheel speed sensor 11 for detecting a speed Vrr), a steering angle sensor 12 for detecting a front wheel steering angle δf, a lateral acceleration sensor 13 for detecting a lateral acceleration (d ² y / dt ² ), and a yaw rate sensor for detecting a yaw rate γ. 14, the steering torque sensor 15 for detecting the steering torque Th and the brake hydraulic pressure sensor 16 for detecting the brake hydraulic pressure PB are connected, and these sensors 11, 12, 13, 14, 15, 16 are connected to the four wheels. Wheel speeds Vfl, Vfr, Vrl, Vrr, front wheel steering angle δf, lateral acceleration (d ² y / dt ² ), yaw rate γ, steering torque Th, and brake fluid pressure PB are input. Further, the road surface friction coefficient estimating device 1 is connected to an engine control device 17, a transmission control device 18, and an electric power steering control device 19, and an engine output torque Teg, an engine speed Ne, a main transmission gear ratio i, a turbine rotation. The number Nt, the front / rear driving force distribution ratio rd, and the current value iq of the electric power steering motor are input. Further, a brake switch 20 is connected to the road surface friction coefficient estimation device 1 and a brake ON-OFF signal is input.

  The road surface friction coefficient estimation device 1 calculates the actual rack force generated as the estimated rack force Fr_star according to the road surface friction coefficient estimation program described later based on the above input signals, and expresses the longitudinal and lateral strains of the tire. The rack thrust calculated by the tire brush model including the parameter λ and the road surface friction coefficient μ as parameters is calculated as the reference rack thrust Fr_model, and the braking / driving force of the wheel actually generated is calculated as the estimated wheel braking / driving force Fx_star. The wheel braking / driving force calculated by the tire brush model including the parameter λ representing the longitudinal strain and lateral strain of the tire and the road surface friction coefficient μ as parameters is calculated as the reference wheel braking / driving force Fx_model, and is actually generated. Is calculated as the estimated lateral force Fy_star, and represents the tire's longitudinal and lateral strain parameters. The lateral force calculated by the tire brush model including the data λ and the road surface friction coefficient μ as parameters is calculated as the reference lateral force Fy_model, and the deviation between the estimated rack thrust Fr_star and the reference rack thrust Fr_model and the estimated wheel braking / driving force Fx_star And the reference wheel braking / driving force Fx_model, and the difference between the estimated lateral force Fy_star and the reference lateral force Fy_model is determined by optimization calculation so that the deviation is minimized.

  That is, as shown in FIG. 1, the road surface friction coefficient estimation device 1 includes a vehicle body speed calculation unit 1a, a vehicle body acceleration / deceleration calculation unit 1b, a transmission output torque calculation unit 1c, an operation state determination unit 1d, a ground load calculation unit 1e, and a front wheel. Slip angle calculation unit 1f, front wheel slip ratio calculation unit 1g, estimated wheel braking / driving force calculation unit 1h, estimated rack thrust calculation unit 1i, estimated lateral force calculation unit 1j, reference wheel braking / driving force calculation unit 1k, reference rack thrust calculation unit 1l, a standard lateral force calculation unit 1m, and a road surface friction coefficient calculation unit 1n.

The vehicle body speed calculation unit 1a receives four wheel speeds (left front wheel speed Vfl, right front wheel speed Vfr, left rear wheel speed Vrl, right rear wheel speed Vrr) from the four wheel speed sensor 11. . Then, for example, the vehicle body speed V is calculated by the following equation (1), and is output to the vehicle body acceleration / deceleration calculation unit 1b and the front wheel slip angle calculation unit 1f.
V = ((Vfl + Vfr + Vrl + Vrr) −Vw_max) / 3 (1)
Here, Vw_max indicates the maximum wheel speed, and the vehicle body speed V is calculated from the average value of the three wheels other than the wheel speed of Vw_max in order to avoid erroneous determination due to single wheel floating.

  The vehicle body acceleration / deceleration calculation unit 1b receives the vehicle body speed V from the vehicle body speed calculation unit 1a. Then, for example, the vehicle body acceleration / deceleration, that is, (dV / dt) is calculated by differentiating the vehicle body speed V, and is output to the ground load calculation unit 1e and the estimated wheel braking / driving force calculation unit 1h.

The transmission output torque calculation unit 1c receives the engine output torque Teg and the engine speed Ne from the engine control device 17, and the main transmission gear ratio i and the turbine speed Nt from the transmission control device 18. Then, for example, the transmission output torque TM_trq is calculated by the following equation (2) and output to the front wheel slip ratio calculation unit 1g.
TM_trq = Teg · t · i (2)
Here, t is the torque ratio of the torque converter, and is obtained, for example, by referring to a map of a preset rotation speed ratio e (Nt / Ne) of the torque converter and a torque ratio t of the torque converter.

  The driving state determination unit 1 d receives a brake ON-OFF signal from the brake switch 20. Then, it is determined whether the vehicle is in a braking state (the brake switch 20 is ON) or other driving state (a driving state including inertial driving; the brake switch 20 is OFF), and the determination result is a front wheel slip ratio calculating unit 1g, Output to the estimated wheel braking / driving force calculation unit 1h.

The ground load calculation unit 1e receives lateral acceleration (d ² y / dt ² ) from the lateral acceleration sensor 13, and receives vehicle acceleration / deceleration (dV / dt) from the vehicle body acceleration / deceleration calculation unit 1b. Then, for example, the front / rear load movement amount ΔWx is calculated by the following equation (3), the left / right load movement amount ΔWy is calculated by the equation (4), and the left front wheel grounding is calculated by the equations (5) to (8). A load Fz_fl, a right front wheel ground contact load Fz_fr, a left rear wheel ground load Fz_rl, and a right rear wheel ground load Fz_rr are calculated to obtain a reference wheel braking / driving force calculation unit 1k, a reference rack thrust calculation unit 1l, a reference lateral force calculation unit 1m, It outputs to the road surface friction coefficient calculation part 1n.
ΔWx = (1/2) · (h / (lf + lf)) · m · (dV / dt) (3)
Here, h is the height of the center of gravity, lf is the distance between the front shaft and the center of gravity, and m is the vehicle mass.
ΔWy = (1/2) · (h / d) · m · (d ² y / dt ² ) (4)
Here, d is a tread.
Fz_fl = (1/2) · (lr / (lf + lr)) · m · g−ΔWx−ΔWy (5)
Fz_fr = (1/2) · (lr / (lf + lr)) · m · g−ΔWx + ΔWy (6)
Fz_rl = (1/2) · (lf / (lf + lr)) · m · g + ΔWx−ΔWy (7)
Fz_rr = (1/2) · (lf / (lf + lr)) · m · g + ΔWx + ΔWy (8)
Here, lr is the distance between the rear axis and the center of gravity, and g is the gravitational acceleration.

The front wheel slip angle calculation unit 1f receives the front wheel steering angle δf from the steering angle sensor 12 and the vehicle body speed V from the vehicle body speed calculation unit 1a. Then, for example, the front wheel slip angle βf is calculated by the following equation (9).
βf = β + (lf / V) · γ−δf (9)
Here, β is the vehicle slip angle, and γ is the yaw rate, and is calculated by the following equations (10) and (11).
β = ((1- (m / (2 · (lf + lr))) · (lf / (lr · Kr))
(V ² ) / (1 + A · V ² )) (lr / (lf + lr)) δf (10)
γ = (1 / (1 + A · V ² )) · (V / (lf + lr)) · δf (11)
Here, A is a stability factor of the vehicle, and is calculated by the following equation (12).
A = − (m / (2 · (lf + lr) ² ))
(Lf · Kf−lr · Kr) / (Kf · Kr) (12)
Here, Kf is the equivalent cornering power of the front wheels, and Kr is the equivalent cornering power of the rear wheels.

となっている。 Further, when the front wheel slip angle calculation unit 1f calculates a new front wheel slip angle, the front wheel slip angle βf, which is a vector amount, is newly set based on a plurality of front wheel slip angles having different sampling times calculated in the past. It outputs to the wheel braking / driving force calculation unit 1k, the reference rack thrust calculation unit 1l, the reference lateral force calculation unit 1m, and the road surface friction coefficient calculation unit 1n. In the present embodiment, as an example of the front wheel slip angle βf, βf [0], βf [1],..., Βf [m],..., Βf [18], βf [ 19], which is composed of a total of 20 components. That is,

It has become.

The front wheel slip ratio calculation unit 1g receives the brake hydraulic pressure PB from the brake hydraulic pressure sensor 16, the front / rear driving force distribution ratio rd from the transmission control device 18, and the transmission output torque TM_trq from the transmission output torque calculation unit 1c. Then, a determination result as to whether the vehicle is in a braking state (the brake switch 20 is ON) or other driving state (a driving state including inertial driving; the brake switch 20 is OFF) is input from the driving state determination unit 1d. When the vehicle is in a braking state, for example, the slip ratio Sf of the front wheel (one wheel) is calculated by the following equation (14). When the vehicle is in a driving state including inertia traveling, for example, by the following equation (15): The slip ratio Sf of the front wheel (one wheel) is calculated.
Sf = (rB / 2) · KB · PB (14)
Here, rB is a braking force distribution ratio, and KB is a braking force coefficient.

Sf = (rd / 2) · TM_trq · η · if / r (15)
Here, η is the drive system transmission efficiency, if is the final gear ratio, and r is the dynamic load radius.

となっている。 Further, when the front wheel slip ratio calculating unit 1g calculates the slip ratio of a new front wheel (single wheel), the front wheel slip ratio Sf, which is a vector amount, is calculated based on the slip ratios of a plurality of front wheels (single wheel) having different sampling times calculated in the past. Are newly set and output to the reference wheel braking / driving force calculation unit 1k, the reference rack thrust calculation unit 1l, the reference lateral force calculation unit 1m, and the road surface friction coefficient calculation unit 1n. In the present embodiment, as an example of the front wheel slip ratio Sf, Sf [0], Sf [1], ..., Sf [m], ..., Sf [18], Sf [ 19], which is composed of a total of 20 components. That is,

It has become.

  The estimated wheel braking / driving force calculation unit 1h receives the front / rear driving force distribution ratio rd from the transmission control device 18, receives the vehicle body acceleration / deceleration (dV / dt) from the vehicle body acceleration / deceleration calculation unit 1b, and receives the driving state determination unit 1d. A determination result of whether the vehicle is in a braking state (the brake switch 20 is ON) or other driving state (a driving state including inertial driving; the brake switch 20 is OFF) is input. When the vehicle is in a braking state, for example, the braking / driving force of the actually generated wheel (front one wheel) is calculated as the estimated wheel braking / driving force Fx_star by the following equation (17). In the case of a driving state including inertial running, for example, the braking / driving force of the actually generated wheel (front one wheel) is calculated as the estimated wheel braking / driving force Fx_star by the following equation (18).

となっている。このように、推定車輪制駆動力算出部１ｈは、推定車輪制駆動力算出手段として設けられている。 That is, when the vehicle is in a braking state, the relationship among the front and rear forces of each wheel (left front wheel front and rear force Fx_fl, right front wheel front and rear force Fx_fr, left rear wheel front and rear force Fx_rl, right rear wheel front and rear force Fx_rr)
Fx_fl = Fx_fr = (1/2) · rB · m · (dV / dt)
Fx_rl = Fx_rr = (1/2) · rB · m · (dV / dt)
As
Fx_star = (Fx_fl + Fx_fr) / 2 = (1/2) · rB · m · (dV / dt) (17)
In the case of a driving state including inertial running,
Fx_star = (1/2) · rd · m · (dV / dt) (18)
Further, when the estimated wheel braking / driving force calculating unit 1h calculates a new estimated wheel braking / driving force, the estimated wheel braking / driving force Fx_star, which is a vector amount, is obtained from a plurality of estimated wheel braking / driving forces having different sampling times calculated in the past. Is newly set and output to the road surface friction coefficient calculating unit 1n. In the present embodiment, as an example of the estimated wheel braking / driving force Fx_star, Fx_star [0], Fx_star [1],..., Fx_star [m],. A description will be given of a configuration composed of a total of 20 components of Fx_star [19]. That is,

It has become. Thus, the estimated wheel braking / driving force calculating unit 1h is provided as an estimated wheel braking / driving force calculating means.

The estimated rack thrust calculation unit 1 i receives the steering torque Th from the steering torque sensor 15 and the current value iq of the electric power steering motor from the electric power steering control device 19. Then, for example, the rack thrust actually generated is calculated as the estimated rack thrust Fr_star by the following equation (20).
Fr_star = ζ1 · (2 · π / hs) · Tp) (20)
Here, ζ1 is the rack and pinion efficiency, and hs is the specific stroke. Tp is the pinion gear torque, and is calculated by the following equation (21), for example.
Tp = Th + ζ2 · nw · Tm (21)
Here, ζ2 is the worm gear efficiency and nw is the worm gear ratio. Tm is the electric power steering motor torque, and is calculated by the following equation (22), for example.
Tm = sign (Th) · km · (1/3 ^1/2 ) · iq (22)
Here, sign (Th) is a sign of the steering torque Th, and km is a motor torque constant.

となっている。このように、推定ラック推力算出部１ｉは、推定ラック推力算出手段として設けられている。 Further, when the estimated rack thrust calculation unit 1i calculates a new estimated rack thrust, the estimated rack thrust Fr_star, which is a vector quantity, is newly set based on a plurality of estimated rack thrusts having different sampling times calculated in the past, and the road surface It outputs to the friction coefficient calculation part 1n. In this embodiment, as an example of the estimated rack thrust Fr_star, Fr_star [0], Fr_star [1], ..., Fr_star [m], ..., Fr_star [18], Fr_star [ 19], which is composed of a total of 20 components. That is,

It has become. Thus, the estimated rack thrust calculation unit 1i is provided as estimated rack thrust calculation means.

The estimated lateral force calculation unit 1j receives the lateral acceleration (d ² y / dt ² ) from the lateral acceleration sensor 13 and the yaw rate γ from the yaw rate sensor 14. For example, the lateral force actually generated (only on the front wheel side) is calculated as the estimated lateral force Fy_star by the following equation (24). Fy_star = (1/2) · (m · (d ² y / dt ² ) + I · (dγ / dt))
/ (Lf + lr) (24)
Here, I is the yawing moment of inertia, and (dγ / dt) is the yaw angular acceleration.

となっている。このように、推定横力算出部１ｊは、推定横力算出手段として設けられている。 Further, when the estimated lateral force calculation unit 1j calculates a new estimated lateral force, the estimated lateral force Fy_star, which is a vector amount, is newly set based on a plurality of estimated lateral forces with different sampling times calculated in the past, and the road surface It outputs to the friction coefficient calculation part 1n. In this embodiment, as an example of the estimated lateral force Fy_star, Fy_star [0], Fy_star [1], ..., Fy_star [m], ..., Fy_star [18], Fy_star [ 19], which is composed of a total of 20 components. That is,

It has become. Thus, the estimated lateral force calculation unit 1j is provided as estimated lateral force calculation means.

The reference wheel braking / driving force calculation unit 1k receives the ground load Fz_fl, Fz_fr, Fz_rl, Fz_rr of each wheel from the ground load calculation unit 1e, the front wheel slip angle βf from the front wheel slip angle calculation unit 1f, and the front wheel slip ratio. The front wheel slip ratio Sf is input from the calculation unit 1g. Then, based on a command from the road surface friction coefficient calculation unit 1n, for example, as shown in the following equation (26), a parameter λ representing the tire longitudinal strain and lateral strain and a road surface friction coefficient μ are included as parameters. Based on the tire brush model, the braking / driving force of the wheel (front one wheel) is calculated as the reference wheel braking / driving force Fx_model and output to the road surface friction coefficient calculating unit 1n.
Fx_model = −Ks · Sf · ξs ² −6 · μ · Fz_f · cos θ · ((1/6)
− (1/2) · ξs ² + (1/3) · ξs ³ ) (26)
Here, Ks is the driving stiffness, and Fz_f is the contact load of the front wheel, and is calculated by the following equation (27).
Fz_f = (Fz_fl + Fz_fr) / 2 (27)
Ξs is calculated by the following equation (28).

ξs = 1− (Ks / (3 · μ · Fz_f)) · λ (28)
Further, λ is calculated by the following equation (29), and is represented by the equation (30) using θ as shown in FIG. 4 when the diagram is shown using the front wheel slip angle βf and the front wheel slip ratio Sf.

λ = (Sf ² + (Kf / Ks) ² · (1 + Sf) ² · tan ² βf) ^1/2 (29)
cos θ = Sf / λ (30)

となっている。このように、基準車輪制駆動力算出部１ｋは、基準車輪制駆動力算出手段として設けられている。 In this embodiment, 20 data are set for the reference wheel braking / driving force Fx_model corresponding to the front wheel slip angle βf and the front wheel slip ratio Sf to be used, and the road surface friction coefficient used is the previous value μ [ n-1], Fx_model [0] [n-1], Fx_model [1] [n-1],..., Fx_model [m] [n-1],. .., Fx_model [18] [n-1] and Fx_model [19] [n-1] are composed of a total of 20 components. That is,

It has become. Thus, the reference wheel braking / driving force calculation unit 1k is provided as a reference wheel braking / driving force calculation unit.

The reference rack thrust calculation unit 11 receives the ground load Fz_fl, Fz_fr, Fz_rl, Fz_rr of each wheel from the ground load calculation unit 1e, and the front wheel slip angle βf from the front wheel slip angle calculation unit 1f, and the front wheel slip ratio calculation unit A front wheel slip ratio Sf is input from 1 g. Then, based on a command from the road surface friction coefficient calculation unit 1n, for example, as shown in the following equation (32), a parameter λ representing the tire longitudinal strain and lateral strain and a road surface friction coefficient μ are included as parameters. The rack thrust of the wheel (front one wheel) is calculated as the reference rack thrust Fr_model by the tire brush model, and is output to the road surface friction coefficient calculation unit 1n.
Fr_model = (1 / ln) · (SAT + εc · Fy) (32)
Here, ln is a pneumatic trail, and εc is a caster trail.

The SAT is calculated by the following equation (33), for example.

SAT = (lt / 2) · Kf · (1 + Sf) · tan βf · ξs ² · (1− (4/3) · ξs ² ) − (3/2) · lt · μ · Fz_f · sinθ · ξs ² · ( 1−ξs) ² + (2/3) · lt · Ks · (1 + Sf) · Sf · tan βf · ξs ³ + ((3 · lt · (μ · Fz_f) ² · sinθ · cosθ) / (5 · Ks) ) ・ (1-10 ・ ξs ³ +15 ・ ξs ⁴ −6 ・ ξs ⁵ )
... (33)
Here, lt is a tire ground contact length. sinθ is calculated by the following equation (34).

    sinθ = (Kf · tanβf · (1 + Sf)) / (Ks · λ) (34)

Fy is calculated by, for example, the following equation (35).
Fy = −Kf · (1 + Sf) · tan βf · ξs ² −6 · μ · Fz_f · sin θ · ((1/6)
− (1/2) · ξs ² + (1/3) · ξs ³ ) (35)

となっている。このように、基準ラック推力算出部１ｌは、基準ラック推力算出手段として設けられている。 In this embodiment, 20 data are set for the reference rack thrust Fr_model corresponding to the front wheel slip angle βf and the front wheel slip rate Sf to be used, and the road surface friction coefficient used is the previous value μ [n− 1], Fr_model [0] [n-1], Fr_model [1] [n-1],..., Fr_model [m] [n-1],. , Fr_model [18] [n-1], Fr_model [19] [n-1], which is composed of a total of 20 components. That is,

It has become. Thus, the reference rack thrust calculation unit 1l is provided as a reference rack thrust calculation means.

The reference lateral force calculation unit 1m receives the contact loads Fz_fl, Fz_fr, Fz_rl, and Fz_rr of each wheel from the contact load calculation unit 1e, and the front wheel slip angle βf from the front wheel slip angle calculation unit 1f, and the front wheel slip ratio calculation unit A front wheel slip ratio Sf is input from 1 g. Then, based on a command from the road surface friction coefficient calculation unit 1n, for example, as shown in the following equation (37), a parameter λ expressing the tire longitudinal strain and lateral strain and a road surface friction coefficient μ are included as parameters. The lateral force of the wheel (front one wheel) is calculated as a reference lateral force Fy_model by the tire brush model, and is output to the road surface friction coefficient calculating unit 1n.
Fy_model = −Kf · (1 + Sf) · tanβf · ξs ² −6 · μ · Fz_f · sinθ
((1/6) − (1/2) · ξs ² + (1/3) · ξs ³ ) (37)

となっている。このように、基準横力算出部１ｍは、基準横力算出手段として設けられている。 In the present embodiment, as the reference lateral force Fy_model, 20 data are set corresponding to the front wheel slip angle βf and the front wheel slip ratio Sf to be used, and the road surface friction coefficient used is the previous value μ [n− 1], Fy_model [0] [n-1], Fy_model [1] [n-1],..., Fy_model [m] [n-1],. , Fy_model [18] [n-1], Fy_model [19] [n-1], which is composed of a total of 20 components. That is,

It has become. Thus, the reference lateral force calculation unit 1m is provided as a reference lateral force calculation unit.

  The road surface friction coefficient calculation unit 1n receives the contact loads Fz_fl, Fz_fr, Fz_rl, and Fz_rr of each wheel from the contact load calculation unit 1e, and the front wheel slip angle βf from the front wheel slip angle calculation unit 1f. The front wheel slip ratio Sf is input from 1g, the estimated wheel braking / driving force Fx_star is input from the estimated wheel braking / driving force calculating unit 1h, the estimated rack thrust Fr_star is input from the estimated rack thrust calculating unit 1i, and the estimated lateral force calculating unit 1j To input the estimated lateral force Fy_star. Then, at the same timing as the estimated wheel braking / driving force Fx_star, estimated rack thrust Fr_star, and estimated lateral force Fy_star, the reference wheel braking / driving force Fx_model, reference rack thrust Fr_model, and reference lateral force Fy_model are converted into the above-described reference wheel braking / driving force calculation unit. 1k, the reference rack thrust calculation unit 1l, and the reference lateral force calculation unit 1m are calculated by a tire brush model including the parameter λ and the road surface friction coefficient μ, which express the longitudinal strain and lateral strain of the tire, as parameters, The value of the road surface friction coefficient μ that minimizes the deviation is obtained by optimization calculation.

  Specifically, the following equation (39) is a vector whose elements are changes in the reference wheel braking / driving force Fx_model, the reference rack thrust Fr_model, and the reference lateral force Fy_model when the road surface friction coefficient μ changes slightly. The Jacobian Jmodel [n-1] is calculated using the previous value μ [n-1] of the road surface friction coefficient. The subscript [n-1] of the Jacobian Jmodel [n-1] represents the previous value μ [n-1] of the road surface friction coefficient. When the iterative calculation n-1 = 0, the road surface friction coefficient. Since there is no previous value μ [n−1], the calculation result of the road surface friction coefficient μ at the time of the previous sampling is substituted.

ここで、（３９）式中のＪx_model[n-1]は、以下の（４０）式である。

Here, Jx_model [n-1] in the equation (39) is the following equation (40).

そして、（４０）式中の各要素は、それぞれ以下の各式で求められるものである。

And each element in (40) Formula is calculated | required by each following formula, respectively.

∂Fx_model / ∂μ [0] [n-1] = (1 / (27 · μ ³ [n-1] · Fz_f ² ))
(Ks ² · λ · Sf [0] · (−9 · μ [n-1] · Fz_f + 2 · Ks · λ))
∂Fx_model / ∂μ [1] [n-1] = (1 / (27 · μ ³ [n-1] · Fz_f ² ))
(Ks ² · λ · Sf [1] · (−9 · μ [n-1] · Fz_f + 2 · Ks · λ))
:
∂Fx_model / ∂μ [m] [n-1] = (1 / (27 · μ ³ [n-1] · Fz_f ² ))
(Ks ² · λ · Sf [m] · (−9 · μ [n-1] · Fz_f + 2 · Ks · λ))
:
∂Fx_model / ∂μ [18] [n-1] = (1 / (27 · μ ³ [n-1] · Fz_f ² ))
(Ks ² · λ · Sf [18] · (−9 · μ [n-1] · Fz_f + 2 · Ks · λ))
∂Fx_model / ∂μ [19] [n-1] = (1 / (27 · μ ³ [n-1] · Fz_f ² ))
(Ks ² · λ · Sf [19] · (−9 · μ [n-1] · Fz_f + 2 · Ks · λ))

  Further, Jr_model [n-1] in the equation (39) is the following equation (41).

そして、（４１）式中の各要素は、それぞれ以下の各式で求められるものである。

And each element in (41) Formula is calculated | required by each following formula, respectively.

∂Fr_model / ∂μ [0] [n -1] = (- 1 / (135 · Fz_f 3 · μ 4 [n-1] · ln))
(Ks · λ · (1 + Sf [0]) · tanβf [0] (Ks ² · λ ² · lt
(−5 · Kf + 8 · Ks · Sf [0]) − 15 · Fz_f ² · μ ² [n−1] · (6 · Kf · εc + 3 · Kf · lt−8 · Ks · Sf [0] · lt) + 10 · Ks · λ · Fz_f · μ [n-1] · (2 · Kf · εc + 3 · Kf · lt-6 · Ks · Sf [0] · lt))
∂Fr_model / ∂μ [1] [n -1] = (- 1 / (135 · Fz_f 3 · μ 4 [n-1] · ln))
・ (Ks ・ λ ・ (1 + Sf [1]) ・ tanβf [1] ・ (Ks ²・ λ ²・ lt
(−5 · Kf + 8 · Ks · Sf [1]) − 15 · Fz_f ² · μ ² [n−1] · (6 · Kf · εc + 3 · Kf · lt−8 · Ks · Sf [1] · lt) + 10 · Ks · λ · Fz_f · μ [n-1] · (2 · Kf · εc + 3 · Kf · lt-6 · Ks · Sf [1] · lt))
:
∂Fr_model / ∂μ [m] [n -1] = (- 1 / (135 · Fz_f 3 · μ 4 [n-1] · ln))
(Ks · λ · (1 + Sf [m]) · tanβf [m] (Ks ² · λ ² · lt
(−5 · Kf + 8 · Ks · Sf [m]) − 15 · Fz_f ² · μ ² [n−1] · (6 · Kf · εc + 3 · Kf · lt−8 · Ks · Sf [m] · lt) + 10 · Ks · λ · Fz_f · μ [n-1] · (2 · Kf · εc + 3 · Kf · lt-6 · Ks · Sf [m] · lt))
:
∂Fr_model / ∂μ [18] [n-1] = (− 1 / (135 · Fz_f ³ · μ ⁴ [n-1] · ln))
(Ks · λ · (1 + Sf [18]) · tanβf [18] (Ks ² · λ ² · lt
(−5 · Kf + 8 · Ks · Sf [18]) − 15 · Fz_f ² · μ ² [n−1] · (6 · Kf · εc + 3 · Kf · lt−8 · Ks · Sf [18] · lt) + 10 · Ks · λ · Fz_f · μ [n-1] · (2 · Kf · εc + 3 · Kf · lt-6 · Ks · Sf [18] · lt))
∂Fr_model / ∂μ [19] [n-1] = (− 1 / (135 · Fz_f ³ · μ ⁴ [n-1] · ln))
(Ks · λ · (1 + Sf [19]) · tanβf [19] (Ks ² · λ ² · lt
(−5 · Kf + 8 · Ks · Sf [19]) − 15 · Fz_f ² · μ ² [n−1] · (6 · Kf · εc + 3 · Kf · lt−8 · Ks · Sf [19] · lt) + 10 · Ks · λ · Fz_f · μ [n-1] · (2 · Kf · εc + 3 · Kf · lt-6 · Ks · Sf [19] · lt))

  Further, Jy_model [n-1] in the equation (39) is the following equation (42).

そして、（４２）式中の各要素は、それぞれ以下の各式で求められるものである。

And each element in (42) Formula is calculated | required by each following formula, respectively.

∂Fy_model / ∂μ [0] [n-1] = (1 / (27 · μ ³ [n-1] · Fz_f ² ))
(Kf.Ks..lambda. (1 + Sf [0]). Tan.beta.f [0]. (-9..mu. [N-1] .Fz_f + 2.Ks..lambda.))
∂Fy_model / ∂μ [1] [n-1] = (1 / (27 · μ ³ [n-1] · Fz_f ² ))
(Kf · Ks · λ · (1 + Sf [1]) · tanβf [1] · (−9 · μ [n-1] · Fz_f + 2 · Ks · λ))
:
∂Fy_model / ∂μ [m] [n-1] = (1 / (27 · μ ³ [n-1] · Fz_f ² ))
(Kf · Ks · λ · (1 + Sf [m]) · tanβf [m] · (−9 · μ [n-1] · Fz_f + 2 · Ks · λ))
:
∂Fy_model / ∂μ [18] [n-1] = (1 / (27 · μ ³ [n-1] · Fz_f ² ))
(Kf · Ks · λ · (1 + Sf [18]) · tanβf [18] · (−9 · μ [n-1] · Fz_f + 2 · Ks · λ))
∂Fy_model / ∂μ [19] [n-1] = (1 / (27 · μ ³ [n-1] · Fz_f ² ))
(Kf.Ks..lambda. (1 + Sf [19]). Tan.beta.f [19]. (-9..mu. [N-1] .Fz_f + 2.Ks..lambda.))
Next, the road surface friction coefficient calculation unit 1n calculates a road surface friction coefficient change amount δμ using the following equation (43).
^{δμ = (Jmodel [n-1} ] T · Jmodel [n-1] + Wμ) -1 · Jmodel [n-1] · (F_star-F_model [n-1]) ... (43)
Here, Wμ is a weight function of the road surface friction coefficient change amount δμ set in advance by experiments or the like. Increasing the weight leads to improvement in stability of estimation of the road surface friction coefficient, but the convergence time is long. Become. Jmodel [n-1] is as shown by the above equation (39), and F_star and F_model [n-1] are shown by the following equations (44) and (45), respectively.

そして、路面摩擦係数算出部１ｎは、以下の（４６）式により、路面摩擦係数の今回の値μ[n]を算出する。
μ[n]＝μ[n-1]＋δμ …（４６）
次いで、路面摩擦係数算出部１ｎは、（４６）式で算出した路面摩擦係数の今回の値μ[n]を用いて、基準車輪制駆動力Ｆx_model、基準ラック推力Ｆr_model、基準横力Ｆy_modelを、以下の（４７）式のように算出する。

Then, the road surface friction coefficient calculation unit 1n calculates the current value μ [n] of the road surface friction coefficient by the following equation (46).
μ [n] = μ [n−1] + δμ (46)
Next, the road surface friction coefficient calculating unit 1n uses the current value μ [n] of the road surface friction coefficient calculated by the equation (46) to obtain the reference wheel braking / driving force Fx_model, the reference rack thrust Fr_model, and the reference lateral force Fy_model, It is calculated as in the following equation (47).

尚、反復演算回数ｎ＝０の場合は、前サンプリング時間における結果を代入する。

If the number of iterations n = 0, the result at the previous sampling time is substituted.

Next, an evaluation function L [n] of the following equation (48) is calculated.
L [n] = (F_star−F_model [n]) ^T (F_star−F_model [n]) + Wμ · δμ ²
... (48)
Then, the previous value L [n-1] and the current value L [n] of this evaluation function are compared to determine whether or not the evaluation function has converged below a preset value ε. Therefore, the convergence calculation is stopped, and the calculated value μ [n] of the road friction coefficient is output as the road friction coefficient μ. If it has not converged below ε, the calculation from Jacobian J [n−1] is repeated again. Thus, the road surface friction coefficient calculation unit 1n has a function as a road surface friction coefficient estimation unit.

  Next, a road surface friction coefficient estimation program executed by the road surface friction coefficient estimation device 1 will be described with reference to the flowcharts of FIGS.

First, in step (hereinafter abbreviated as “S”) 101, necessary parameters, that is, four wheel speeds Vfl, Vfr, Vrl, Vrr, front wheel steering angle δf, lateral acceleration (d ² y / dt ² ), Yaw rate γ, steering torque Th, brake hydraulic pressure PB, engine output torque Teg, engine speed Ne, main transmission gear ratio i, turbine speed Nt, front / rear driving force distribution ratio rd, electric power steering motor current value iq, brake ON-OFF signal is read.

  Next, proceeding to S102, the transmission output torque calculation unit 1c calculates the transmission output torque TM_trq by the above-described equation (2).

  Next, the process proceeds to S103, where the vehicle body speed calculation unit 1a calculates the vehicle body speed V by the above-described equation (1), and the vehicle body acceleration / deceleration calculation unit 1b calculates the vehicle body acceleration / deceleration (dV / dt).

  Next, proceeding to S104, the ground load calculation unit 1e calculates the left front wheel ground load Fz_fl, the right front wheel ground load Fz_fr, the left rear wheel ground load Fz_rl, and the right rear wheel ground load according to the above formulas (5) to (8). Fz_rr is calculated.

  Next, the process proceeds to S105, where the front wheel slip angle calculation unit 1f calculates the front wheel slip angle βf by the above equation (9) and outputs the front wheel slip angle βf of the above equation (13).

  Next, the process proceeds to S106, and the driving state determination unit 1d determines whether the vehicle is in a braking state (the brake switch 20 is ON) or other driving state (a driving state including inertial driving; the brake switch 20 is OFF). .

  As a result of this determination, if the vehicle is in a braking state (the brake switch 20 is ON), the process proceeds to S107, and the front wheel slip ratio calculation unit 1g calculates the slip ratio Sf of the front wheel (one wheel) by the above-described equation (14). Calculate and output the slip ratio Sf of the front wheel (one wheel) of the above equation (16).

  Thereafter, the process proceeds to S108, where the estimated wheel braking / driving force calculation unit 1h calculates the estimated wheel braking / driving force Fx_star according to the above equation (17) and outputs the estimated wheel braking / driving force Fx_star according to the above equation (19). Then, the process proceeds to S111.

  On the other hand, if the result of determination in S106 is that the vehicle is in a traveling state other than the braking state (the brake switch 20 is ON) (driving state including inertial traveling; the brake switch 20 is OFF), the process proceeds to S109 and the front wheel slip ratio is calculated. The part 1g calculates the slip ratio Sf of the front wheel (single wheel) according to the above equation (15), and outputs the slip ratio Sf of the front wheel (single wheel) according to the above equation (16).

  Thereafter, the process proceeds to S110, where the estimated wheel braking / driving force calculation unit 1h calculates the estimated wheel braking / driving force Fx_star according to the above equation (18) and outputs the estimated wheel braking / driving force Fx_star according to the above equation (19). Then, the process proceeds to S111.

  Then, when proceeding from S108 or S110 to S111, the estimated rack thrust calculation unit 1i calculates the estimated rack thrust Fr_star from the above equation (20), and the estimated rack thrust Fr_star of the above equation (23) is calculated. Output.

  Next, proceeding to S112, the estimated lateral force calculation unit 1j calculates the estimated lateral force Fy_star by the above-described equation (24), and outputs the estimated lateral force Fy_star of the above-described equation (25).

  Next, the process proceeds to S113, and the road surface friction coefficient calculation unit 1n calculates Jacobian Jmodel [n-1] of the above equation (39).

  Next, the process proceeds to S114, and the road surface friction coefficient calculating unit 1n calculates the road surface friction coefficient change amount δμ by the above-described equation (43).

  Next, proceeding to S115, the road surface friction coefficient calculation unit 1n calculates the current value μ [n] of the road surface friction coefficient by the above-described equation (46).

  Next, the process proceeds to S116, where the road surface friction coefficient calculating unit 1n uses the current value μ [n] of the road surface friction coefficient calculated above by the reference wheel braking / driving force calculating unit 1k according to the above equation (26). A reference wheel braking / driving force Fx_model [n] composed of 20 components is calculated.

  Next, proceeding to S117, the road surface friction coefficient calculating unit 1n uses the current value μ [n] of the above calculated road surface friction coefficient by the reference rack thrust force calculating unit 11 to calculate the above equation (32). , A reference rack thrust Fr_model [n] composed of 20 components is calculated.

  Next, the process proceeds to S118, where the road surface friction coefficient calculation unit 1n uses the current value μ [n] of the road surface friction coefficient calculated above by the reference lateral force calculation unit 1m to calculate 20 A reference lateral force Fy_model [n] composed of individual components is calculated.

  Next, proceeding to S119, the road surface friction coefficient calculating unit 1n calculates the current value L [n] of the evaluation function shown in the above equation (48).

  In S120, the difference (L [n] -L [n-1]) between the current value L [n] and the previous value L [n-1] of the evaluation function converges below a preset value ε. (L [n] −L [n−1] <ε). As a result of the determination, if the value converges to less than ε (if L [n] −L [n−1] <ε), the process proceeds to S121, and the current value μ of the road surface friction coefficient obtained in S115. [n] is output as the estimation result of the road surface friction coefficient, and the program is exited.

  On the other hand, if it has not converged below ε (in the case of L [n] −L [n−1] ≧ ε), the process proceeds to S122, and the value calculated this time is set as the previous value, that is, Fx_model [n− 1] ← Fx_model [n], Fr_model [n-1] ← Fr_model [n], Fy_model [n-1] ← Fy_model [n], μ [n-1] ← μ [n], L [n-1] As ← L [n], the calculation from S113 is repeated again.

  As described above, according to the embodiment of the present invention, the actually generated rack thrust is calculated as the estimated rack thrust Fr_star, and the parameter λ expressing the longitudinal strain and the lateral strain of the tire and the road surface friction coefficient μ are set as parameters. The rack thrust calculated by the brush model of the tire included as the reference rack thrust Fr_model is calculated, and the braking / driving force of the actual wheel is calculated as the estimated wheel braking / driving force Fx_star to express the longitudinal strain and lateral strain of the tire. The wheel braking / driving force calculated by the tire brush model including the parameter λ and the road surface friction coefficient μ as parameters is calculated as the reference wheel braking / driving force Fx_model, and the actual lateral force is calculated as the estimated lateral force Fy_star. Tire parameters including a parameter λ expressing the longitudinal strain and lateral strain of the tire and a road surface friction coefficient μ as parameters. The lateral force calculated by the cache model is calculated as the reference lateral force Fy_model, the deviation between the estimated rack thrust Fr_star and the reference rack thrust Fr_model, the deviation between the estimated wheel braking / driving force Fx_star and the reference wheel braking / driving force Fx_model, and the estimated lateral force Fy_star The value of the road surface friction coefficient μ is determined by optimization calculation so that the deviation from the reference lateral force Fy_model is minimized. In this way, the road surface μ value is weighted in any vehicle motion state by estimating the road surface friction coefficient using the brush model formula considering the turning motion information, the longitudinal motion information, and the lateral motion information. Can be obtained continuously without any problem. In addition, since the brush model takes into account longitudinal strain and lateral strain, it is possible to accurately estimate even the road surface friction coefficient during turning driving and turning braking that require both strains. As described above, according to the embodiment of the present invention, the road surface friction coefficient can be continuously obtained without adding complicated processing such as weighting even when the vehicle is in any motion state, even when turning control is being driven. The natural value can be estimated with high accuracy.

  In this embodiment, the convergence determination of the evaluation function L [n] is performed until it becomes less than ε. However, the number of convergence calculations may be set in advance. Also, a limit value for the number of operations may be provided.

DESCRIPTION OF SYMBOLS 1 Road surface friction coefficient estimation apparatus 1a Car body speed calculation part 1b Car body acceleration / deceleration calculation part 1c Transmission output torque calculation part 1d Driving state determination part 1e Ground load calculation part 1f Front wheel slip angle calculation part 1g Front wheel slip ratio calculation part 1h Estimated wheel braking / driving Force calculation unit (estimated wheel braking / driving force calculation means)
1i Estimated rack thrust calculation unit (estimated rack thrust calculation means)
1j Estimated lateral force calculation unit (estimated lateral force calculation means)
1k Reference wheel braking / driving force calculating unit (reference wheel braking / driving force calculating means)
1l reference rack thrust calculation unit (reference rack thrust calculation means)
1m Standard lateral force calculation unit (Standard lateral force calculation means)
1n Road surface friction coefficient calculation unit (road surface friction coefficient estimation means)

Claims (3)

An estimated rack thrust calculating means for calculating an actual rack thrust as an estimated rack thrust;
A reference rack thrust calculation means for calculating, as a reference rack thrust, a rack thrust calculated by a tire brush model including at least a parameter expressing a longitudinal strain and a lateral strain of the tire and a road surface friction coefficient as parameters;
An estimated wheel braking / driving force calculating means for calculating an actual wheel braking / driving force as an estimated wheel braking / driving force;
A reference wheel braking / driving force calculating means for calculating, as a reference wheel braking / driving force, a wheel braking / driving force calculated by a tire brush model including at least a parameter expressing the longitudinal strain and lateral strain of the tire and a road surface friction coefficient as parameters; ,
An estimated lateral force calculating means for calculating an actual lateral force as an estimated lateral force;
A reference lateral force calculating means for calculating a lateral force calculated by a tire brush model including at least a parameter expressing a longitudinal strain and a lateral strain of the tire and a road surface friction coefficient as a parameter;
The road surface so that a deviation between the estimated rack thrust and the reference rack thrust, a deviation between the estimated wheel braking / driving force and the reference wheel braking / driving force, and a deviation between the estimated lateral force and the reference lateral force is minimized. Road surface friction coefficient estimating means for obtaining the value of the friction coefficient by optimization calculation;
A road surface friction coefficient estimating device comprising:   2. The road surface friction coefficient estimating apparatus according to claim 1, wherein the parameter expressing the longitudinal strain and lateral strain of the tire is a parameter calculated by at least a tire slip angle and a slip ratio.   The road friction coefficient estimation means adds a value obtained by multiplying a value obtained by squaring the correction amount of the current road friction coefficient with respect to the previously calculated road friction coefficient to a value obtained by squaring each of the deviations and a weight function set in advance. Forming an evaluation function, calculating the amount of correction of the road surface friction coefficient using the fact that the value obtained by partial differentiation of the evaluation function with the road surface friction coefficient is 0, and obtaining the current road surface friction coefficient The road surface friction coefficient estimating device according to claim 1 or 2.

Images