JPH10181565A - Road surface state identification device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a means to identify the μ on a road surface during the stable running of the wheels of a vehicle without slipping. SOLUTION: A round tire 100 is once deformed in a flat state on ground- contact surface. In this case, a longitudinal force ΔF acting on a tire tread surface varies according to μ, etc., on a road surface. Also, because the material of each part of the tire reduces a resonation during the rotation, the tire is manufactured with the stiffness of these materials varied. Therefore the parts of the tire are deformed differently at the ground-contact surface of the tire. In this way, using as a standard those variation signals (tire rotating pitch) generated when the same position of the tire is brought in contact with the road surface, the state of the road surface is identified by the magnitude of the signals and their continuity. In addition, a tire rotation pitch component is extracted by the cepstrum analysis from wheel speed signals transmitted from a vehicle speed sensor and used to identify the road surface state.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a road condition judging device for judging a road condition during traveling, and relates to a system for notifying a driver of danger information, an anti-skid brake device, a traction control device and the like. The present invention can also be applied to a system that improves the safety of vehicle traveling by transmitting road surface information to the control of the traveling system.

[0002]

2. Description of the Related Art Conventionally, a coefficient of friction μ on a road surface (hereinafter referred to as “μ”)
As a method for determining (μ), there is known a method of making a determination based on a slip ratio in a deceleration / acceleration state of a vehicle, and a method of making a determination based on a lateral acceleration and a turning moment during turning. Further, as a device that performs control based on the detected friction coefficient μ, there are an anti-skid brake device and a traction control device.

[0003]

However, according to these methods for determining μ of the road surface, when the vehicle is in a constant speed state where no acceleration is generated or when the steering operation is not performed, μ cannot be determined. Therefore, in the above-described state, it is not possible to transmit information such as a decrease in the μ of the road surface to the driver to warn the driver or to promptly transmit the information to a control device such as an anti-skid brake device. .

SUMMARY OF THE INVENTION In view of the foregoing, it is an object of the present invention to make it possible to determine the μ of the road surface even when the vehicle is in a constant speed state in which no acceleration is generated, or when the steering operation is not performed. And

[0005]

Means for Solving the Problem A circular tire is once deformed into a flat state on a tread, but at this time, the longitudinal force acting on the tire tread surface depends on the μ of the road surface and the strength of the material of each part on the tire circumference. Change. FIG. 1A illustrates the force that the rubber block on the ground contact surface receives in the front-rear direction. After the rubber block of the rolling tire 100 is pulled forward at the first ground contact portion, It is compressed up and down and pulled to the rear as it tries to get off the ground. This force is defined as ΔF.

[0006] The force ΔF to be pulled back and forth depends on the μ characteristics of the road surface and the rigidity of the rubber block of the tire 100.
In general, the higher the μ of the road surface or the tire stiffness, the greater the force Δ
The fluctuation of F increases. Also, as the μ of the road surface is higher and the rigidity is higher or the tire rigidity is higher, the change rate (gradient) of the force ΔF is larger. By the way, in general, the material of each part of the tire is manufactured by giving strength to the material in order to reduce resonance during rotation. For this reason, each part of the tire is deformed in a different state at the ground contacting part, and changes every time the tire rotates. Further, the tire 100 rolls while forming vibrations and strong and weak waves on the tread portion and the carcass portion of the tire 100. For this reason, when the road surface μ is the same, a similar wave is formed at the same location of the tire 100, that is, at the same frequency as the rotation of the tire 100 or every integer multiple thereof.

For this reason, the magnitude of μ on the road surface and the uniformity of the road surface vary depending on the vibration of the tire 100 and the strength of the stiffness of the tire 100 and fluctuations synchronous with the rotation of the tire 100 and the rotation of the tire 100 generated in the suspension system (hereinafter referred to as “tire”). Rotation pitch). That is, the road surface condition can be determined by evaluating the magnitude and continuity of the fluctuation occurring at the same location of the tire 100 as a ruler.

[0008] Specifically, on a road surface such as artificially formed asphalt, the road surface condition stably continues. Therefore, the information of the tire rotation pitch is not disturbed. Also,
Since the road surface μ is large, the fluctuation of the force ΔF is large.
As shown in (b), the front-back distortion of the tire 100 increases, and the tire rotation pitch also increases. However, on a naturally formed low μ road surface, μ of the road surface is small, the road surface is uneven because of the natural road surface, and unevenness is more uneven than asphalt. For this reason, information on the tire rotation pitch is disturbed. Further, since the road surface μ is small, the fluctuation of the force ΔF is small, and as shown in FIG. 1C, the front-back distortion of the tire 100 is reduced, and the tire rotation pitch component is also reduced.

For the above reasons, the state of μ on the road surface can be determined by measuring the tire rotation pitch while the vehicle is running. That is, a road surface having a constant grip such as an asphalt road and a road surface having a small μ and a variation such as snow or ice can be distinguished by measuring the tire rotation pitch. In order to achieve the above object, the following means are adopted.

According to the first aspect of the present invention, a means for outputting a signal which includes a fluctuation component based on a frictional force between the tires (1 to 4) of a vehicle and a road surface and includes a fluctuation component synchronized with the rotation of the tire. 11 to 14), wherein a fluctuation component is extracted from this signal, and the road surface state is determined based on the fluctuation component. Note that the signal including the fluctuation component includes wheel acceleration, wheel torque, suspension distortion, wheel suspension vibration, and the like.

As described above, the μ of the road surface can be determined from the tire rotation pitch component. Therefore, by performing the road surface determination at the tire rotation pitch, the μ of the road surface can be determined even when the vehicle is in a constant speed state in which no slip occurs or when the steering operation is not performed. Specifically, claim 2
As shown in (1), the extraction of the tire rotation pitch can be performed by cepstrum analysis. In this case, it is possible to determine μ of the road surface based on the tire rotation pitch component.

According to the third aspect of the present invention, it is possible to extract a change based on a frictional force between a tire and a road surface based on a signal from a wheel speed sensor (11 to 14). In this case, the wheel speed signal component and the tire rotation pitch component are separated from the wheel speed signal from the wheel speed sensor,
It is possible to determine μ of the road surface from the separated tire rotation pitch components.

[0013]

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing a first embodiment of the present invention. FIG. 2 shows a configuration diagram in the present embodiment. As shown in FIG. 2, wheel speed sensors 11 to 14 are attached to the wheels 1 to 4, respectively, and signals from the wheel speed sensors 11 to 14 are used to determine an electronic control unit (hereinafter, referred to as a road surface state determining device). ECU 5). The steering angle detected by the steering sensor 7 attached to the steering 6 is also input to the ECU 5, and the road surface state is determined based on the steering angle. The road condition determined by the ECU 5 is warned by a warning lamp 8 to the driver.

Specifically, the road surface condition is determined by detecting a tire rotation pitch component using cepstrum analysis. Here, cepstrum analysis is a technique that is widely used in speech recognition and the like (Reference: Shozo Saito, Basics of Speech Information Processing), and detects each pitch component in speech based on the cepstrum principle. I have. Based on the principle of the cepstrum, a component of a tire rotation speed pitch is detected.

That is, when the signal s (t) is regarded as a signal having a pitch period T, the signal s (t)
Can be expressed by convolution of the integrated value of the signal component s0 (t) and the pitch component Σδ.

[0016]

(Equation 1)

Equation (1) is converted to a Fourier transform (hereinafter referred to as FF).
T), the power spectrum value: S (ω) at s (t) can be represented by the product of the signal component S0 (ω) and the pitch component, and is expressed by Expression 2.

[0018]

(Equation 2)

Further, when the logarithmic transformation is performed on Expression 2, the signal component and the pitch component can be separated, and are expressed by Expression 3.

[0020]

(Equation 3)

Where S (t): input signal, t: time,
n: integer, T: pitch period, S0 (t): signal component other than pitch, *: convolution integral The first term in Equation 3 is the power spectrum value of s0 (t), and the parentheses in the second term are 2 shows a line spectrum.
Here, the power spectrum value of s0 (t) shows a relatively gradual change with respect to ω, and the line spectrum shows a sharp change with ω.

For this reason, when Equation 3 is regarded as a signal sequence of the variable ω, the pitch component shows a sharp change with respect to ω as compared with the signal component. Further, in general, s0 (ω) has little abrupt change with respect to ω. Therefore, when inverse Fourier transform is applied to Equation 3, the pitch component is separated into the high frequency (high quefrency) side and the signal component is separated into the low frequency (low quefrency). be able to.

That is, since S0 (ω) mainly has a low frequency component, it does not appear as a waveform in a high frequency region, and has a waveform mainly composed of a tire pitch component in this region. The low-frequency wheel speed signal component and the high-frequency tire pitch component can be separated, and only the tire rotation pitch component can be extracted from this waveform.

Therefore, in the vehicle, the cepstrum separates the wheel speed signal into the tire rotation pitch component and the wheel speed signal component, and the road surface state can be determined from the tire rotation pitch component. FIGS. 3A to 3C show examples of waveforms detected when the road surface is in a dry state, a snow compaction state, or an on-ice state. FIG.
As shown in (a), in the dry state, the tire rotation pitch component appears as a waveform corresponding to one rotation cycle of the tire, and the tire rotation pitch component appears largely. FIG. 3 (b)
As shown in (1), in the snow compaction state, the unevenness and the road surface μ are non-uniform, so that the vibration is random and the waveform becomes irregular. As shown in FIG. 3 (c), in the on-ice state, it appears as a waveform corresponding to one rotation cycle of the tire, but only a small tire rotation pitch component appears. Thus, characteristic waveforms are obtained in each state, and cepstrum analysis is performed based on these waveforms to detect tire rotation pitch components.

FIG. 4 shows each process in the road surface discrimination detection using the above-described cepstrum analysis. FIG. 5 shows waveforms calculated based on these processes. First, in step 201, a wheel speed pulse is detected for each gear (one pulse) of the rotor from the rotation angle sensor, and a wheel speed pulse interval is obtained and input from the wheel speed pulse. For detection of the wheel speed pulse, an electromagnetic pickup system or a photocoupler system widely used widely can be used.

In step 202, from the sampled wheel speed pulse intervals, a wheel speed Vw (t) at which the sample interval is constant is determined. More specifically, as shown in FIG. 6A, for each wheel speed pulse generated for each gear 50a of the rotor 50, a wheel speed Vw as shown in FIG.
(N) is obtained, and further interpolated from the wheel speed Vw (n) by a quadratic spline function to obtain the wheel speed Vw (t) at which the sampling interval shown in FIG.

In the following step 203, a high-pass filter (hereinafter, referred to as "FFT") is used to reduce an error in performing the FFT.
HPF) removes the low frequency component of the velocity. Steps 204 to 208 are processes in the cepstrum calculation. These steps will be described based on the cepstrum principle. Step 2
At 04, window processing is performed to determine a target window area. If this area is small, the calculation accuracy will decrease,
If the area is large, the responsiveness will be reduced. By switching the calculation range according to the tire rotation speed, it is possible to cover from low speed to high speed.

In this embodiment, this window area is defined as Ts = 2πr / Vb * Nrot (where Ts: window time, r: tire radius, Vb: vehicle speed, Nrot:
Tire rotation speed). Basically, the measurement area is determined by the number of rotations of the tire, and Nrot is performed at about 20 rotations, for example. The waveform shown in FIG. 5A is the detected wheel speed Vw.
The waveform shown in FIG. 5 (t) and the waveform shown in FIG.
It is the waveform which expanded the wheel speed sample in the window area determined by window processing after PF processing. 5B is a waveform represented by the signal s (t) in Expression 1.
Thereby, the signal s (t) is determined.

In step 205, S (ω) is calculated in the window area by performing an FFT operation as shown in the above equation (2). In step 206, a power spectrum value is calculated from the result of the FFT calculation. In the following step 207, the logarithmic value logVw (ω) of the power spectrum value is calculated as shown in Expression 3. Note that plotting logVw (ω) at this time is represented by a waveform in FIG. In step 208, the logarithmic value is subjected to an inverse FFT operation.

In steps 209 and 210,
In order to extract only the component related to the tire rotation pitch component, a process of applying a window related to the vehicle speed on the quefrency axis is performed. First, in step 209, H
The PF removes offset components generated by noise or the like. At this time, the waveform after removing the offset component is represented by the waveform in FIG.

In step 210, window processing is performed to remove the wheel speed signal component and detect only the rotation pitch of the tire (corresponding to the waveform in FIG. 5 (e)). N windows are arranged at regular intervals in accordance with the rotation cycle of the tire, and the arrangement interval ΔT is ΔT = 2πr
/ Vb. Further, n is determined by the order component of the pitch generated on the cepstrum graph, but is about the fourth order in a normal tire, and therefore, in this embodiment, the processing is performed with n = 4. Further, a hamming window: Hw (i, q) is used for the window, and can be expressed as a function as shown in Expression 4.

In general, high-order components are likely to occur on a high μ road surface. In this case, the window gain: G (i) is set high as shown in Expression 5. Here, the predetermined constant is a predetermined constant used for the Hamming window, q is quefrency, and M is the window width.

[0033]

(Equation 4)

[0034]

(Equation 5)

The cepstrum value after window processing: Cw (q) is obtained by Expression 6 based on Expressions 4 and 5 above (corresponding to the waveform in FIG. 5F). Note that C
(Q) is a cepstrum value after the offset is removed by the HPF.

[0036]

(Equation 6)

In step 211, the sum of squares in the window area is calculated, and the sum of the respective components of the tire rotation pitch (tire rotation pitch parameter): Cp
Ask for.

[0038]

(Equation 7)

FIG. 7B shows the cepstrum analysis results Cp of various road surface conditions experimentally set when the vehicle travels at a speed as shown in FIG. 7A. As described above, the result of the cepstrum analysis differs depending on the type of the road surface. For this reason, the road surface state can be determined by the cepstrum analysis. Step 212
Then, the road surface is determined based on the tire rotation pitch component, the vehicle body acceleration, and the steering angle obtained earlier. That is, since the tire rotation pitch component changes according to the steering angle and the vehicle body acceleration, the road surface state is determined more completely.

More specifically, this determination is made based on the road surface condition selection map shown in FIG. 8 set based on the road surface condition, and the coordinates A (represented by the tire rotation pitch parameter Cp, the steering angle and the vehicle body acceleration). The road surface is determined based on where the tire rotation pitch pattern, steering angle, and vehicle body acceleration) are located on the coordinate axes shown in FIG. The vehicle body acceleration is calculated based on the wheel speeds of the four wheels.

More specifically, as shown in FIG. 8, a high μ road such as asphalt or concrete and a low μ road such as a snowy road or on ice are used.
There is a threshold surface B for distinguishing the road from the road. If the coordinates A are above the threshold surface B, the road is determined to be a high μ road, and if the coordinates A are below, the road is determined to be a low μ road. At this time, when the driver does not operate the steering, the steering angle is zero, and when the vehicle speed is constant, the vehicle body acceleration is zero. In such a case, the coordinates A are (Cp, 0, 0).

In step 213, the driver is notified based on the determination result. For the notification, a warning lamp indicating a decrease in the road surface μ is used. For example, when the road changes from a high μ road to a low μ road, the warning lamp is turned on for a few seconds to give a notification. In addition, by sending a signal indicating that the road is a low μ road in advance to an antilock brake device, a traction control device, and the like, it is possible to effectively suppress the slip in the initial stage of these controls.

As described above, even when the vehicle is in a constant speed state in which no acceleration is generated, or when the steering operation is not performed, the friction coefficient μ of the road surface is not changed.
Can be detected. (Second Embodiment) In this embodiment, cepstrum analysis is performed at a sample interval different from that in the first embodiment.

Since the cepstrum analysis is usually performed on the time axis because of the use for speech processing, the above analysis was performed on this time axis in the first embodiment. Perform analysis with. In this case, since the pitch of the tire rotation pitch is generated in proportion to the rotation distance of the tire, a direct analysis can be performed by performing the analysis on the distance axis.

More specifically, as shown in FIG. 5 (b), in the first embodiment, the sampling interval of the wheel speed is set to a predetermined time period, so that the wheel speed Vw
It is necessary to interpolate (t) from the wheel speed Vw (n) extracted for each gear 50a of the rotor 50. However, as shown in FIG. 5 (c), if the sampling interval at the wheel speed is set to be a cycle of a fixed distance, the distance becomes equal to the rotor 50
Is proportional to the movement of the gear 50a, so that the sampling interval can be set for each gear 50a of the rotor 50 (every pulse), eliminating the need for the interpolation.

By performing the analysis in proportion to the distance in this manner, the cepstrum values are arranged at the position of the distance synchronized with the rotation of the tire. FIG. 9 shows each process in road surface discrimination detection using cepstrum analysis. Hereinafter, these processes will be described with reference to FIG. In step 301, a period required for transmitting one pulse is defined as one cycle, and a wheel speed pulse corresponding to the wheel speed in each cycle is detected. Next, in step 302, based on the wheel speed pulse, V
= K × Δn / ΔT to calculate the wheel speed Vw (n).

Steps 303 to 309 perform the same processing as steps 203 to 209 in the first embodiment (corresponding to the waveform in FIG. 10A). Then, in step 310, window processing for detecting the tire rotation pitch is performed (FIG. 10).
(This corresponds to the waveform of (b).) However, in this case, the unit of the calculated quefrency is the number of pulses.
Assuming that the number of rotation detection pulses (number of teeth) is Ns, the window position is determined by a constant multiple of the number of pulses Ns.

Thereafter, steps 311 to 313 are executed.
, The same processing as in steps 211 to 213 in the first embodiment is performed, and the processing ends. In addition,
As described above, since the cepstrum analysis is performed on the distance axis, an effect of removing noise that is not synchronized with rotation can be obtained. That is, when cepstrum analysis is performed on the distance axis, a signal extracted at the time of sampling has a component synchronized with rotation. Therefore, the maximum frequency of the signal is a frequency proportional to the rotation cycle. Therefore, if the sample frequency is set to a value proportional to the rotation, a constant proportional relationship can be maintained with the signal taken out at the time of sampling, and aliasing is less likely to occur regardless of the vehicle speed.

(Other Embodiments) In the first and second embodiments, the detection of the tire rotation pitch component is performed based on the signal (wheel speed signal) from the wheel speed sensor. Can be output as a signal, for example, a signal from a vertical acceleration sensor, a wheel torque sensor, a suspension distortion detection sensor, a wheel suspension vibration detection sensor, or the like.

Although the rotation pitch component of the tire is obtained by the cepstrum analysis, it can be obtained by the coefficient of the LPC cepstrum. In addition, since the tire rotation pitch changes depending on the type of tire such as a studless tire and a summer tire, the road surface condition can be determined for each type of tire. For example, FIG.
A road surface condition determination map as shown in FIG. 1 is set in advance, and a symbol indicating the type of tire is input in advance when the vehicle is running, so that the road surface condition determination map according to the tire type may be corrected.

Similarly, when the vehicle is running on a high μ road such as asphalt or the like having a uniform road surface, by pressing a state setting switch or the like, the state is inputted as an initial state and the tire rotation pitch parameter Cp at that time is input. Can be stored, and the road surface state can be determined using this value as a reference value.

[Brief description of the drawings]

FIG. 1A is a schematic diagram showing a force that a tire receives from a road surface, and FIGS. 1B and 1C are schematic diagrams showing a magnitude of a tire rotation pitch with respect to a road surface state.

FIG. 2 is a schematic diagram showing a circuit configuration of each sensor of the vehicle.

FIG. 3 is an explanatory diagram showing a waveform detected as a sample according to a road surface condition.

FIG. 4 is a flowchart of a road surface state determination process in the first embodiment.

FIG. 5 is a characteristic diagram showing a waveform corresponding to the processing in FIG.

FIG. 6 is a comparison diagram in a case where time or distance is adopted as a sample interval.

FIG. 7 is an explanatory diagram showing results of cepstrum analysis according to various road surface conditions.

FIG. 8 is an explanatory diagram showing a selection map used as a reference for performing road surface state determination.

FIG. 9 is a flowchart of a road surface state determination process according to the second embodiment.

FIG. 10 is a characteristic diagram showing a waveform corresponding to the processing in FIG. 9;

[Explanation of symbols]

1-4: wheels, 5: electronic control unit, 6: steering sensor, 7: warning lamp, 11-14: wheel speed sensor, 50: rotor.

Claims (4)

[Claims] 1. A means (11-14) for outputting a signal which includes a fluctuation component synchronized with the rotation of the tire (1-4) and the rotation of the tire based on a frictional force on a road surface, the signal comprising: A road surface state determination device that extracts the fluctuation component from the road surface and determines a road surface state based on the fluctuation component. 2. The road surface state determination device according to claim 1, wherein the signal is subjected to cepstrum analysis to extract the fluctuation component. 3. The road surface condition determination device according to claim 1, wherein the means for outputting the signal is a wheel speed sensor (11 to 14). 4. A tire (1-4) for a vehicle, and means (11-14) for outputting a signal containing a fluctuation component generated in synchronization with the rotation of the tire, and a road surface condition based on the fluctuation component. A road surface condition determination device characterized in that the determination is made.