JPH08138188A - Road condition detector - Google Patents

Abstract

(57) [Summary] [PROBLEMS] To provide a road surface state detection device capable of comprehensively determining even a non-uniform road surface state and improving the running stability of a vehicle. [Structure] When a road surface state is determined using a neural network at timing D, the value stored in the previous value area in the memory for storing the previous value is changed to the before-previous value area in the memory for storing the previous-previous value. After being moved, the value stored in the current value area that stores the current value in the memory is moved to the previous area, and then the determination result (usually 0 to
Any value within the range of 1) is stored in the current value area, the average value of the content of the value area before last, the content of the previous value area and the content of the current value area is calculated to determine the road surface condition. In the example of FIG. 5, the discrimination results by the neural network at timings B and C are both dry roads, and the discrimination results by the neural network at timing D are wet roads. Therefore, even at the timing D, the result is determined to be the dry road.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a road surface state detecting device for detecting a road surface state based on road noise when a vehicle is running.

[0002]

2. Description of the Related Art Conventionally, a so-called road surface state detecting device for detecting a state of a road surface on which a vehicle is traveling based on road noise has been proposed. As such a road surface condition detecting device, the applicant of the present application, in Japanese Patent Application No. 5-340052, detects a road noise generated from a wheel and uses a neural network from a pattern of each frequency component level of the detected road noise. Then, a road surface condition detecting device for judging the road surface condition was proposed.

[0003]

However, in the above-mentioned conventional road surface state detecting device, the road surface state is discriminated at regular intervals, so that when the road surface state is non-uniform, it is judged as a whole. A road surface state different from the expected road surface state may be determined.

FIG. 6 is a diagram showing an example of a road surface condition determined by a conventional road surface condition detecting device. The timing for detecting the state and the determination result are shown. Where 10 puddle
1, 102 are not the ones that perform vehicle body control such as an anti-lock brake system or traction control, or those that warn the driver to call their attention. If anything, it is better to judge it as a dry road surface as a whole. Good quality.

As shown in FIG. 6, there are puddles 101 and 102 while the vehicle is running on a dry road surface 103, and the vehicle has a puddles 1.
When the road surface state is detected while traveling 01, 102 (detection timing D, G), the road surface state is wet (W
ET), and the vehicle body control, alarm, etc. are performed according to the result of the determination. However, even if it is not a puddle enough to perform vehicle control or alarms and it is desirable to determine the dry road surface as a whole for the traveling stability of the vehicle, it is determined that the road surface state is different each time, As a result of the above-described vehicle body control and the like, there is a problem that the running stability of the vehicle is impaired.

The present invention has been made in view of the above problems, and it is possible to comprehensively determine even a non-uniform road surface condition, and to improve the running stability of a vehicle. An object is to provide a state detection device.

[0007]

In order to achieve the above object, the present invention provides a road noise detecting means for detecting road noise generated from a wheel, and a road noise detected by the road noise detecting means. It has a first discriminating means for discriminating the road surface state at a constant cycle, and a second discriminating means for discriminating the current road surface state based on two or more continuous outputs from the first discriminating means. It is characterized by

[0008]

According to the structure of the present invention, when the road noise is detected by the road noise detecting means, the first judging means judges the road surface condition at a constant cycle based on the road noise. The second discriminating means discriminates the current road surface state based on the two or more continuous outputs discriminated by the first discriminating means.

[0009]

Embodiments of the present invention will now be described in detail with reference to the drawings.

FIG. 1 is a block diagram showing a schematic configuration of an embodiment of a road surface state detecting device according to the present invention.

In the figure, reference numeral 1 is a road noise detecting means for detecting road noise, which is constituted by a microphone in this embodiment. On the output side of the road noise detecting means 1, an amplifier 2 for amplifying the road noise detected by the road noise detecting means 1 and a filter 3 for analyzing the frequency of the road noise amplified by the amplifier 2, Although details will be described later, the determination unit 4 that determines the road surface state by using a neural network
Is connected to one input side of the
The output side of the vehicle speed detecting means 5 for detecting the speed of the vehicle is connected. The determination unit 4 determines the road surface state according to these two input signals, and outputs the determination results to an alarm device and a motion control device (neither is shown).

FIG. 2 is a diagram for explaining the positions where the road noise detecting means 1 and the vehicle speed detecting means 5 are arranged.

In FIG. 1, the vehicle C is assumed to be a front engine vehicle, and the road noise detecting means 1 which is a microphone is provided with stones or the like on at least one of the inside of the wheel house of the left and right rear wheels where the influence of engine noise is small. The vehicle speed detecting means 5 is arranged so as not to be directly hit by water, and the vehicle speed detecting means 5 is arranged at a predetermined position in at least one of the left and right sides in the rear wheel. Here, the vehicle speed detecting means 5 generates an electric pulse signal according to the wheel speed.

Next, the determination method performed by the determination unit 4 will be described with reference to FIGS.
A description will be given based on FIG.

FIG. 3 is a diagram showing a neural network model used in this embodiment. In this embodiment, a neural network model having a three-layer structure of an input layer, an intermediate layer, and an output layer is used, and the learning algorithm is back propagation (Back-P).
ropagation; hereinafter referred to as "BP") learning algorithm (for details of this BP learning algorithm, see the specification of Japanese Patent Application No. 5-340052).

As shown in FIG. 3, in this embodiment, the information input to each cell in the input layer is the vehicle speed and the sound pressure value of the specific frequency detected for each 1/3 octave band. Information is weighted by the coupling matrix and input to each cell in the middle layer. In the intermediate layer, the output is determined for each cell by, for example, a sigmoid function, and the output weighted by the coupling matrix is input to each cell in the output layer as in the processing from the input layer to the intermediate layer. Furthermore, in the output layer, the output is determined for each cell by the sigmoid function in the same manner as the processing in the intermediate layer.

Since the value output from each cell in the output layer is the value of the sigmoid function, it takes any value in the range of 0 to 1. That is, each cell in the output layer has a certainty
Is being output. Depending on the size of this “probability”, the road surface type TYPEn, n = 1, 2,
One is decided.

The learning objective is the first layer (input layer).
Is to match the output of the third layer (output layer) with the certainty of desired TYPEn according to the road surface condition currently running, that is, the road surface condition.
Specifically, as shown in FIG. 4, the road noise from the wheel is sampled by the road noise detecting means 1, and the sampled data is subjected to frequency analysis. Input to the input layer. Then, each element (weight) of the connection matrix is determined so that the current road surface state and the road surface type based on the final output (probability) from the output layer match.

The control processing performed by the discrimination unit 4 using the neural network learned as described above will be described below.
This will be described with reference to FIG.

FIG. 5 is a diagram showing the determination result of the road surface condition determined by the road surface condition detecting device of the present embodiment. As in the case of FIG. 6, the vehicle C is cleaning the dry road surface 13 with the puddle 11, 12. When the vehicle is traveling, the timing for detecting the road surface condition and the determination result thereof are shown.

When the vehicle C starts to run and the discriminating section 4 discriminates the road surface condition using the neural network at the timing A, the discrimination result, that is, the maximum output value from the above-mentioned third layer is shown in the figure. Not stored in a current value area secured in a predetermined area of the memory, and the road surface state is determined to be a dry road (DRY).

Similarly, the road surface condition is determined at the timing B, the value stored in the present value area is moved to the previous value area secured in another area, and then the value is stored. The determination result is stored in the current value area, the average value of the content of the previous value area and the content of the current value area is calculated, and the road surface state is determined. In the example of FIG. 5, since the determination results by the neural network at the timings A and B are both dry roads, the average thereof is also the dry road, and as a result, the dry roads are determined.

Next, in the same manner, the road surface condition is determined at the timing C, the value stored in the previous value area is moved to the pre-previous value area secured in this area, and at the same time, After moving the value stored in the current value area to the previous value area, store the determination result in the current value area,
The road surface condition is determined by calculating the average value of the contents of the previous-previous value area, the previous value area, and the current value area. At this time, all of the discrimination results by the neural network at the timings A to C are dry roads, so the average is also a dry road, and as a result, it is discriminated as a dry road.

Further, similarly, the road surface condition is determined at the timing D. Since the processing after the determination is the same as the processing at the timing C, its description is omitted. At this time, the determination results by the neural network at timings B and C are both dry roads, and the determination results by the neural network at timing D are wet roads, but the determination result by the average of these is a dry road. In addition, even at the timing D, it is determined that the road is a dry road.

Thereafter, the same determination as that at the timings C and D results in the determination result at the timings E to H being the dry path.

As described above, in the present embodiment, the road surface condition can be comprehensively determined by taking the moving average of the determination results of the road surface condition three times by the neural network.

In this embodiment, the number of pieces of data for which the moving average is taken is three, but the number is not limited to this, and any value may be used as long as it is two or more.

Further, the latest discrimination result may be given more weight than the discrimination results of the previous time and the previous time. For example, in the case of taking the average of three data as in the present embodiment, (the value of the current value area × 3 + the value of the previous value area × 2 + the value of the two-previous value area) / 6, Different weights may be given to the values.
Needless to say, the weighting value and method are not limited to this.

In this embodiment, the neural network is used to determine the road surface condition, but the present invention is not limited to this, and other road surface condition determining methods may be used.

[0030]

As described above, according to the present invention,
When the road noise is detected by the road noise detecting means, the first discriminating means discriminates the road surface condition at a constant cycle based on the road noise, and the second discriminating means detects the road noise from the first discriminating means. Since the current road surface condition is determined based on the two or more continuous outputs determined by, it is possible to comprehensively determine even the uneven road condition, and the running stability of the vehicle is improved. And the like can be improved.

[Brief description of drawings]

FIG. 1 is a block diagram showing a schematic configuration of an embodiment of a road surface state detecting device according to the present invention.

FIG. 2 is a diagram illustrating the positions where the road noise detecting means and the vehicle speed detecting means in FIG. 1 are arranged.

FIG. 3 is a diagram showing a neural network model used in this embodiment.

FIG. 4 is a diagram for explaining a method of making the network of FIG. 3 learn.

FIG. 5 is a diagram showing a determination result of a road surface state determined by a road surface state detecting device according to the present embodiment.

FIG. 6 is a diagram showing an example of a road surface state determined by a conventional road surface state detecting device.

[Explanation of symbols]

 1 Road Noise Detecting Means 4 Judging Section (First Discriminating Means, Second Discriminating Means)

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Atsushi Itagaki 1-4-1 Chuo, Wako-shi, Saitama Inside the Honda R & D Co., Ltd. (72) Inventor Hideki Kubota 4-1-1 Chuo, Wako, Saitama Stock Company Honda Technical Research Institute

Claims (1)

[Claims] 1. A road noise detecting means for detecting road noise generated from a wheel, and a first discriminating means for discriminating a road surface condition at a constant cycle based on the road noise detected by the road noise detecting means. And a second discriminating means for discriminating the current road surface state based on two or more continuous outputs from the first discriminating means.