JP5407215B2 - Electric power steering device - Google Patents

Description

  The present invention relates to an electric power steering apparatus.

  In recent years, electric power steering devices (EPS) using a motor as a drive source have been widely adopted as vehicle power steering devices. In many of such EPSs, compensation control based on an absolute steering angle (steering angle) based on a steering neutral position (middle point) such as so-called steering return control (steering return compensation control) is performed. (For example, refer to Patent Document 1).

  That is, in the case of EPS, for example, the frictional force of the speed reduction mechanism may exceed the self-aligning torque, and may not return to the steering neutral position. Therefore, by executing the steering return control and calculating the compensation component (steering return compensation amount) acting in the steering neutral direction, good steering return performance is ensured.

  By the way, it goes without saying that compensation control based on such an absolute steering angle is based on the assumption that the steering neutral position (midpoint) is accurate. That is, for example, in the above-described steering return control, when the steering neutral position is out of order, the steering does not return completely or overshoots beyond the steering neutral position.

  However, usually, a vehicle is not provided with a sensor that can directly detect the steering angle as an absolute steering angle, and the steering angle is detected by a steering sensor or the like based on prestored midpoint information. Is calculated by counting the number of rotations of the relative steering angle (mechanical angle 360 °). For this reason, if the midpoint information is lost due to battery exhaustion or the like, accurate steering angle detection becomes impossible, and compensation control based on the absolute steering angle such as the steering return control can be executed in practice. Disappear.

Therefore, for example, the EPS described in Patent Document 2 learns the steering neutral position by estimation based on various vehicle state quantities such as the detected steering torque and vehicle speed in an actual running state, and further, a predetermined statistic. By gradually improving the midpoint accuracy by the operation, both the quick start of the steering return control and the securing of high midpoint accuracy are achieved. Then, by gradually narrowing the width of the control dead zone set in the vicinity of the steering neutral position with the lapse of time from the start of the learning, in addition, a malfunction caused by the low midpoint accuracy during the learning, for example, steering As a feeling, it is configured to suppress the occurrence of discomfort transmitted to the driver.
JP 2006-123827 A Japanese Patent Laid-Open No. 7-132845

  However, in the configuration in which the learning accuracy is improved by the statistical operation as in the conventional example, the finally obtained midpoint accuracy depends on the learning condition set at the beginning. Therefore, if the final midpoint accuracy is to be increased, the learning conditions must also be set strictly in advance, and the start of compensation control will be delayed due to the decrease in learning opportunities accompanying this. There's a problem. Furthermore, there is no special correlation other than the statistical meaning of improving accuracy by increasing the number of samples in the relationship between the elapsed time and learning accuracy, so the effect of narrowing the control dead zone with the elapsed time Will also be limited. In other words, since the specific relationship between the time elapsed from the start of learning and the degree of improvement in the midpoint accuracy is not grasped, the adverse effects caused by narrowing the control dead zone beyond the improvement in the midpoint accuracy are eliminated. In order to avoid this, the control dead zone cannot be significantly narrowed. For this reason, even if the steering return control can be started early, there is a problem that the range in which the steering return control can be performed is actually limited over a long period of time. The room was left.

  The present invention has been made to solve the above-described problems, and its object is to quickly learn the steering neutral position and start compensation control based on the absolute steering angle at an early stage, while achieving high midpoint accuracy. It is another object of the present invention to provide an electric power steering apparatus capable of ensuring the above and effectively preventing the occurrence of problems during learning.

  In order to solve the above problems, the invention according to claim 1 is directed to a steering force assisting device for applying an assist force for assisting a steering operation to a steering system, and the steering force assisting device based on an assist force target value. And the assist force target value includes a compensation component calculated based on an absolute steering angle with respect to a steering neutral position, and the control means detects a vehicle state quantity detected Is an electric power steering apparatus having a learning function for updating a steering neutral position newly estimated under the learning condition as a steering neutral position as a reference for the absolute steering angle when the learning condition is satisfied. The control means tightens the learning condition every time the steering neutral position is updated, and changes the learning condition in accordance with the tightening of the learning condition. Increasing the compensation component, and the gist.

  According to the above configuration, it is possible to quickly learn the steering neutral position and start compensation control based on the absolute steering angle at an early stage, and finally, it is possible to ensure higher midpoint accuracy. In addition, by suppressing the compensation component at the start of the compensation control to a small value, the influence of the low accuracy of the midpoint during learning is limited, thereby enabling more rapid start of compensation control. . Furthermore, since stricter learning conditions directly lead to an improvement in midpoint accuracy, increasing the compensation component in accordance with the stricter learning conditions makes it appropriate for the degree of improvement in the midpoint accuracy. The compensation component can be increased. As a result, it is possible to effectively prevent the occurrence of problems during the learning.

The gist of the invention described in claim 2 is that the compensation component is a steering return control component for applying the assist force in the direction of the steering neutral position.
That is, in the steering return control based on the steering neutral position, the high accuracy of the midpoint is extremely important. Therefore, a more remarkable effect can be obtained by applying to a device having such steering return control.

The gist of the invention described in claim 3 is that the learning condition relates to a straight traveling requirement indicating that the vehicle is traveling straight ahead.
In other words, the learning conditions can be easily and appropriately strict as long as they relate to the requirements for straight traveling of the vehicle. Therefore, a more remarkable effect can be acquired by applying to such a thing.

  According to the present invention, it is possible to quickly learn the steering neutral position and start compensation control based on the absolute rudder angle at an early stage, while ensuring high midpoint accuracy, and at the same time, effectively generating problems during the learning process It is possible to provide an electric power steering device that can be prevented.

DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, an embodiment of the invention will be described with reference to the drawings.
FIG. 1 is a schematic configuration diagram of an electric power steering apparatus (EPS) 1 according to the present embodiment. As shown in the figure, the steering shaft 3 to which the steering wheel 2 is fixed is connected to a rack 5 via a rack and pinion mechanism 4, and the rotation of the steering shaft 3 accompanying the steering operation is performed by the rack and pinion mechanism 4. Is converted into a reciprocating linear motion of the rack 5. The steering angle of the steered wheels 6, that is, the steered angle is varied by the reciprocating linear motion of the rack 5, whereby the traveling direction of the vehicle is changed.

  The EPS 1 includes an EPS actuator 10 as a steering force assisting device that applies an assist force for assisting a steering operation to the steering system, and an ECU 11 as a control unit that controls the operation of the EPS actuator 10. .

  The EPS actuator 10 of the present embodiment is a so-called rack assist type EPS actuator in which a motor 12 as a driving source thereof is arranged coaxially with the rack 5, and the motor torque generated by the motor 12 is a ball feed mechanism (not shown). ) Is transmitted to the rack 5. In addition, the motor 12 of this embodiment is a brushless motor, and rotates by receiving supply of three-phase (U, V, W) driving power from the ECU 11.

  On the other hand, the ECU 11 is connected with a plurality of sensors for detecting various vehicle state quantities, such as a torque sensor 14, a vehicle speed sensor 15, a steering sensor (steering angle sensor) 16, and a yaw rate sensor 17. The ECU 11 calculates an assist force target value (target assist force) based on the state quantities detected by these sensors, that is, the steering torque τ, the vehicle speed V, the steering angle θs (and the steering speed ωs), and the like. Then, the ECU 11 operates the EPS actuator 10, that is, assist force applied to the steering system through supply of drive power to the motor 12 as a drive source in order to cause the EPS actuator 10 to generate the calculated target assist force. To control.

Next, an aspect of assist control in the EPS of the present embodiment will be described.
As shown in FIG. 2, the ECU 11 includes a microcomputer 21 that outputs a motor control signal, and a drive circuit 22 that supplies drive power to the motor 12 that is a drive source of the EPS actuator 10 based on the motor control signal. ing.

  In the present embodiment, the ECU 11 is connected to a current sensor 23 for detecting an actual current value I supplied to the motor 12 and a rotation angle sensor 24 for detecting the rotation angle θ of the motor 12. The microcomputer 21 drives based on the actual current value I and the rotation angle θ of the motor 12 detected based on the output signals of these sensors, and the steering torque τ, the vehicle speed V, the steering angle θs, and the steering speed ωs. A motor control signal is output to the circuit 22.

  Each control block in the microcomputer 21 shown below is realized by a computer program executed by the microcomputer 21. Then, the microcomputer 21 detects each state quantity at a predetermined sampling period, and generates a motor control signal by executing each arithmetic processing shown in the following control blocks at every predetermined period.

  Specifically, the microcomputer 21 includes a current command value calculation unit 25 that calculates a current command value Iq * corresponding to an assist force target value (target assist force) to be generated by the EPS actuator 10, and a current command value calculation unit 25. And a motor control signal output unit 26 that outputs a motor control signal based on the calculated current command value Iq *.

  The current command value calculation unit 25 of this embodiment includes a basic assist control unit 27 that calculates a basic assist control amount Ias *, which is a basic control component of the target assist force, and the steering 2 as a neutral component (θs Steering return control unit 28 for calculating a steering return control amount Isb * for returning to (= 0).

  In this embodiment, the steering torque τ and the vehicle speed V are input to the basic assist control unit 27, and the basic assist control unit 27 performs the steering based on the steering torque τ and the vehicle speed V. The larger the torque τ and the smaller the vehicle speed V, the larger the basic assist control amount Ias * is calculated (see FIG. 3).

  Here, in the EPS 1 of the present embodiment, the steering sensor 16 uses a rotation angle sensor that outputs a rotation angle displacement of the steering wheel 2 caused by a steering operation as a relative steering angle θs_rl (mechanical angle 360 °). The microcomputer 21 then calculates a steering angle θs as an absolute steering angle based on the steering neutral position (θs = 0 °) based on the relative steering angle θs_rl output from the steering sensor 16. 29 is provided.

  More specifically, the steering angle calculation unit 29 holds (stores) midpoint information θ0 indicating the steering neutral position as a reference. The steering angle calculation unit 29 is configured to calculate the steering angle θs as the absolute steering angle by counting the number of laps of the input relative steering angle θs_rl based on the midpoint information θ0. ing.

  In the present embodiment, the steering angle θs as the absolute steering angle and the vehicle speed V are input to the steering return control unit 28 as described above. Then, the steering return control unit 28 calculates the steering return control amount Isb *, that is, a compensation component for generating the steering return force acting in the steering neutral direction based on each of these state quantities (steering return control). ).

  More specifically, the steering return control unit 28 of the present embodiment performs the steering return control amount by executing the steering speed feedback control so that the actual steering speed ωs follows the steering speed target value ωs * calculated based on the steering angle θs. Isb * is calculated. In the present embodiment, the steering speed ωs is obtained by differentiating the steering angle θs.

  That is, the steering angle θs input to the steering return control unit 28 is input to the steering speed target value calculation unit 30, and the steering speed target value calculation unit 30 calculates the steering speed target value ωs *. Specifically, the steering return control unit 28 of the present embodiment includes a map 30a in which the steering angle θs and the steering speed target value ωs * are associated (see FIG. 4, steering speed target value calculation map). In the map 30a, the steering speed target value ωs * is set to be a value in the return direction having a larger absolute value as the absolute value of the steering angle θs is larger (larger steering angle). The steering speed target value calculation unit 30 calculates the steering speed target value ωs * by referring to the input steering angle θs to the map 30a, that is, by executing map calculation using the map 30a.

  The steering speed target value ωs * calculated by the steering speed target value calculation unit 30 is input to the subtractor 31 together with the steering speed ωs, and the subtractor 31 calculates the steering speed target value ωs * and the actual steering speed ωs. A deviation between them (steering speed deviation Δωs) is calculated. Then, the F / B control calculation unit 32 calculates a basic control amount εsb that is a basic component of the steering return control amount Isb * by multiplying the steering speed deviation Δωs by a predetermined gain.

  The vehicle speed V is input to the vehicle speed gain calculation unit 33, and the vehicle speed gain Kv calculated by the vehicle speed gain calculation unit 33 is input to the multiplier 34 together with the basic control amount εsb. Then, the steering return control unit 28 of the present embodiment outputs a value obtained by multiplying the basic control amount εsb, the vehicle speed gain Kv, and a learning gain K_lr described later as a steering return control amount Isb * in the multiplier 34. It has a configuration.

  The basic assist control amount Ias * calculated by the basic assist control unit 27 and the steering return control amount Isb * calculated by the steering return control unit 28 are input to the adder 35. Then, the current command value calculation unit 25 superimposes the steering return control amount Isb * on the basic assist control amount Ias * in the adder 35, whereby the current command value Iq * as the assist force target value (target assist force). Is calculated.

  The motor control signal output unit 26 includes the current command value Iq * calculated by the current command value calculation unit 25, the actual current value I detected by the current sensor 23, and the rotation angle detected by the rotation angle sensor 24. θ is input. The motor control signal output unit 26 calculates a motor control signal by executing feedback control so that the actual current value I follows the current command value Iq * corresponding to the target assist force.

  Specifically, in the present embodiment, the motor control signal output unit 26 converts the phase current value (Iu, Iv, Iw) of the motor 12 detected as the actual current value I into the d, q axes of the d / q coordinate system. The current feedback control is performed by converting into a current value (d / q conversion).

  That is, the current command value Iq * is input to the motor control signal output unit 26 as a q-axis current command value, and the motor control signal output unit 26 determines the phase current value based on the rotation angle θ detected by the rotation angle sensor 24. (Iu, Iv, Iw) is d / q converted. The motor control signal output unit 26 calculates the d and q axis voltage command values based on the d and q axis current values and the q axis current command value. Then, the phase voltage command values (Vu *, Vv *, Vw *) are calculated by performing d / q inverse conversion on the d and q axis voltage command values, and a motor control signal is generated based on the phase voltage command values. To do.

  In this embodiment, the microcomputer 21 outputs the motor control signal generated in this way to the drive circuit 22, and the drive circuit 22 supplies the motor 12 with three-phase drive power based on the motor control signal. Thus, the operation of the EPS actuator 10 is controlled.

(Midpoint learning)
Next, the steering neutral position learning control (middle point learning) and the steering return control in cooperation with this will be described.

  As described above, in the configuration in which the steering angle θs as the absolute steering angle is calculated based on the previously stored midpoint information θ0 and the relative steering angle θs_rl, when the midpoint information θ0 is lost due to battery exhaustion or the like. In effect, compensation control based on the steering angle θs such as steering return control cannot be performed.

  Therefore, in such a case, the microcomputer 21 of the present embodiment estimates the steering neutral position based on the detected various vehicle state quantities, and further, new neutral point information indicating the estimated steering neutral position. The midpoint information θ0 that is the basis of the steering angle calculation is updated by θ0 ′. That is, it has a function of learning the steering neutral position at any time (middle point learning).

  More specifically, as shown in FIG. 2, the microcomputer 21 of the present embodiment includes a midpoint learning control unit 36. The midpoint learning control unit 36 includes a vehicle speed V, a steering torque τ, and the above-described relative values. The steering angle θs_rl and the yaw rate γ are input. Then, the midpoint learning control unit 36 executes learning control of the steering neutral position based on these vehicle state quantities.

  Specifically, the midpoint learning control unit 36 of the present embodiment has the vehicle speed V equal to or higher than a predetermined speed (V0), and the change amount (| Δθs_rl |) of the relative steering angle θs_rl is equal to or lower than a predetermined threshold (α). In addition, it is determined whether or not the steering torque τ (absolute value thereof) is equal to or less than a predetermined value (τ0). Then, a new steering neutral position is estimated with the learning condition that the average value (| γ_av |) of the yaw rate γ when the state continues for a predetermined time or more is equal to or less than a predetermined threshold value (γn).

  That is, if it is possible to continue straight running without performing a special steering operation, the rotational position of the steering wheel 2 can be estimated as the steering neutral position. Thus, new midpoint information θ 0 ′ indicating the estimated steering neutral position is generated and output to the steering angle calculation unit 29. Then, the steering angle calculation unit 29 updates the midpoint information θ0 that is the basis of the steering angle calculation with the new midpoint information θ0 ′, thereby learning a new steering neutral position. In the present embodiment, the midpoint information θ0 is updated by resetting the rotation counter of the relative steering angle θs_rl.

  Further, the midpoint learning control unit 36 of the present embodiment performs the learning every time the new midpoint information θ0 ′ is output, that is, every time the midpoint information θ0 that is the basis of the steering angle calculation is updated. Tighten the conditions. Specifically, as shown in FIG. 5, among the straight running requirements indicating that the vehicle is running straight, the number of updates is counted for the threshold value (γn) related to the average value (| γ_av |) of the yaw rate γ. The learning count value N obtained by doing so is changed to a smaller value. As a result, the steering neutral position can be quickly learned so that the steering return control can be started at an early stage, and finally, higher midpoint accuracy can be ensured.

  Further, the midpoint learning control unit 36 of the present embodiment outputs the learning count value N to the steering return control unit 28. Then, the steering return control unit 28 gradually increases the steering return control amount Isb * to be output as the learning count value N increases.

  Specifically, the steering return control unit 28 of the present embodiment is provided with a learning gain calculation unit 37, and the learning gain calculation unit 37 performs the learning based on the input learning count value N. The larger the count value N, the larger the learning gain K_lr is calculated (0 <K_lr ≦ 1, see FIG. 6).

  The steering return control unit 28 superimposes the learning gain K_lr on the basic control amount εsb and the vehicle speed gain Kv in the multiplier 34 and outputs the superimposed value as the steering return control amount Isb *. The steering return control amount Isb * is increased according to the learning level.

  That is, even if there is some deviation in the steering neutral position, if the compensation component based on the steering angle θs, that is, the steering return control amount Isb * is small, the influence is limited. Therefore, it is possible to start compensation control more quickly by keeping the compensation component at the beginning of the process small. In particular, in the present embodiment, the learning condition is tightened every time the midpoint information θ0 is updated. Therefore, an increase in the learning count value N directly leads to an improvement in the midpoint accuracy. Therefore, it is possible to grasp the midpoint accuracy from the learning level indicated by the learning count value N, and to appropriately increase the steering return control amount Isb * according to the degree of improvement of the midpoint accuracy. it can. And in this embodiment, it becomes the structure which prevents effectively generation | occurrence | production of the malfunction in the course of the learning by this.

Next, the processing procedure of the midpoint learning control configured as described above will be described.
As shown in the flowcharts of FIGS. 7 and 8, when the midpoint learning control unit 36 acquires each vehicle state quantity (step 101), it first determines whether the vehicle speed V is equal to or higher than a predetermined speed V0 ( Step 102). Subsequently, in step 102, when the vehicle speed V is equal to or higher than the predetermined speed V0 (V ≧ V0, step 102: YES), the change amount Δθs_rl (the absolute value thereof) of the relative steering angle θs_rl is the predetermined threshold value. It is determined whether it is α or less (step 103). In step 103, if the change amount Δθs_rl is equal to or smaller than the threshold value α (| Δθs_rl | ≦ α, step 103: YES), the steering torque τ (the absolute value thereof) is equal to or smaller than the predetermined value τ0. Whether or not (step 104).

Furthermore, when the steering torque τ is equal to or smaller than the predetermined value τ0 in step 104 (| τ | ≦ τ0, step 104: YES), the midpoint learning control unit 36 next measures the average value γ_av of the yaw rate γ. It is determined whether it is in the middle (step 105). The determination as to whether or not the measurement is being performed is made based on whether or not a measurement flag described later is set. If measurement is not in progress (step 105: NO), the measurement counter is reset (t = 0, step 106), and the measurement flag is set (step 107). If it is determined in step 105 that measurement has already been performed (step 105: YES), the processing of step 106 and step 107 is not executed. Then, the value of the yaw rate γ detected in the calculation cycle is recorded (step 108), and the measurement counter is incremented (t = t + 1, step 109).
Next, the midpoint learning control unit 36 determines whether or not the measurement count value t of the measurement counter exceeds the predetermined value t0 (step 110), and if it exceeds (t> t0, step 110: YES). Subsequently, the calculation of the average yaw rate (average value γ_av) within the predetermined time indicated by the predetermined value t0 (step 111) and the calculation of the threshold value γn (step 112) are executed.

  The calculation of the average yaw rate in step 111 is performed using the plurality of yaw rate γ values recorded by the execution of step 108 in each calculation cycle until the measurement count value t exceeds the predetermined value t0. Further, the calculation of the threshold value γn in step 112 is performed based on the learning count value N at that time as described above (see FIG. 6). When it is determined in step 110 that the measurement count value t of the measurement counter is equal to or less than the predetermined value t0 (t ≦ t0, step 110: NO), the processing of these steps 111 and 112, and The processing after step 113 shown below is not executed.

  Next, the midpoint learning control unit 36 determines whether or not the calculated average yaw rate (the absolute value thereof, | γ_av |) is equal to or smaller than the threshold value γn (step 113). | γ | ≦ γ0, Step 113: YES), the rotational position of the steering wheel 2 at that time is estimated as the steering neutral position. Then, new midpoint information θ0 ′ indicating the steering neutral position is generated (calculated) and output to the steering angle calculator 29 (step 114), and the learning counter is further incremented (N = N + 1, step 115). The learning count value N is output to the steering return control unit 28 (step 116).

  In step 113, when the calculated average yaw rate (| γ_av |) is larger than the threshold value γn (| γ |> γ0, step 113: NO), the processing of steps 114 to 116 is not executed. . If any of the determination conditions in steps 102 to 104 is not satisfied (V <V0, | Δθs_rl |> α, or | τ |> 0, that is, step 102: NO, step 103: NO, or step 104 : NO), the processing of step 105 to step 116 is not executed.

  The midpoint learning control unit 36 resets the measurement flag except when it is determined in step 110 that the measurement count value t is equal to or less than the predetermined value t0 (t ≦ t0, step 110: NO). (Step 117), the midpoint learning calculation in the calculation cycle is ended (RETURN).

As described above, according to the present embodiment, the following operations and effects can be obtained.
The microcomputer 21 has a steering neutral position learning function. Each time the midpoint information θ0 that is the basis of the steering angle calculation is updated, the learning condition is tightened, and the compensation component based on the steering angle θs as the absolute steering angle is made according to the tightening of the learning condition. The steering return control amount Isb * is gradually increased.

  According to the above configuration, it is possible to quickly learn the steering neutral position and start the steering return control at an early stage, and finally, it is possible to ensure higher midpoint accuracy. In addition, by suppressing the steering return control amount Isb * at the start of the control to be small, the influence of the low accuracy of the midpoint during learning is limited. It becomes possible. Furthermore, since stricter learning conditions directly lead to improvement in midpoint accuracy, increasing the steering return control amount Isb * in accordance with stricter learning conditions increases the midpoint accuracy. The appropriate steering return control amount Isb * can be increased accordingly. As a result, it is possible to effectively prevent the occurrence of problems during the learning.

In addition, you may change this embodiment as follows.
In the present embodiment, the present invention is embodied in EPS that executes steering return control as compensation control based on the steering angle θs that is an absolute steering angle. However, if the absolute steering angle is used, other compensation control is performed. It may be applied to what is executed.

  In the present embodiment, the stricter learning requirement is performed by reducing the threshold value γn related to the yaw rate γ among the straight traveling requirements indicating that the vehicle is traveling straight ahead. However, the present invention is not limited to this, and a configuration in which the straight traveling requirements other than the yaw rate?

  The new steering neutral position estimation and learning method is not limited to the conventional configuration described above, and may be executed by other methods. For example, it is not always necessary to use the yaw rate γ as the vehicle state quantity serving as the basis of the vehicle state quantity, and the other vehicle may be embodied as a main learning condition.

The schematic block diagram of an electric power steering device (EPS). The control block diagram of EPS in this embodiment. Explanatory drawing which shows the aspect of a basic assist control calculation. Explanatory drawing which shows the aspect of steering speed target value calculation. Explanatory drawing which shows the relationship between a learning count value and a yaw rate threshold value. Explanatory drawing which shows the relationship between a learning count value and a learning gain. The flowchart which shows the process sequence of the midpoint learning control in this embodiment. The flowchart which similarly shows the process sequence of the midpoint learning control in this embodiment.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Electric power steering apparatus (EPS), 2 ... Steering, 10 ... EPS actuator, 11 ... ECU, 12 ... Motor, 21 ... Microcomputer, 22 ... Drive circuit, 25 ... Current command value calculating part, 27 ... Basic assist control part 28 ... steering return control unit 29 ... steering angle calculation unit 36 ... mid point learning control unit 37 ... learning gain calculation unit Iq * ... current command value Ias * ... basic assist control amount Isb * ... steering return Compensation amount, ε_sb: Basic control amount, K_lr: Learning gain, θs: Steering angle, ωs: Steering speed, θ0, θ0 ′: Midpoint information, θs_rl: Relative steering angle, Δθs_rl: Change amount, α: Threshold value, V ... Vehicle speed, V0: predetermined speed, τ: steering torque, τ0: predetermined value, γ: yaw rate, γn: threshold value, t: measurement count value, t0: predetermined value, N: learning count value.

Claims (3)

A steering force assisting device for applying an assisting force for assisting a steering operation to the steering system; and a control means for controlling the operation of the steering force assisting device based on an assisting force target value. The control means includes a compensation component calculated based on the absolute steering angle with respect to the steering neutral position, and the control means newly adds a new learning condition when the detected vehicle state quantity satisfies a predetermined learning condition. An electric power steering apparatus having a learning function for updating an estimated steering neutral position as a steering neutral position that is a reference of the absolute steering angle,
The control means tightens the learning condition every time the steering neutral position is updated, and increases the compensation component according to the tightening of the learning condition.
An electric power steering device. The electric power steering apparatus according to claim 1, wherein
The electric power steering apparatus according to claim 1, wherein the compensation component is a steering return control component for applying the assist force in the direction of the steering neutral position. In the electric power steering device according to claim 1 or 2,
The electric power steering apparatus according to claim 1, wherein the learning condition relates to a straight traveling requirement indicating that the vehicle is traveling straight ahead.

Images