JP5758759B2 - Vehicle road surface determination device and driving support device - Google Patents

Description

  The present invention relates to a vehicle road surface determination device and a driving support device that determine a road surface state based on a running sound of a vehicle.

  There are various techniques for determining the road surface condition. For example, Patent Document 1 discloses a technique for determining a road surface state based on a road environment value such as a slip ratio of each wheel calculated based on a vehicle state amount. In the driving assistance device, vehicle control and alerting are changed according to the road surface state determined in this way. For example, in the case of ABS (Anti-lock Brake System), brake control is changed when wheel slip is detected. There is also a technique for detecting lane markings (white lines, etc.) based on an image captured by a camera. For example, in the case of a lane departure warning device, the possibility that the host vehicle departs from the lane is determined based on the detected lane, and if there is a possibility of departure, an alarm is output.

JP 7-334786 A

  When the road surface state is determined based on the road environment value such as the slip ratio calculated from the vehicle state amount as described above, the road surface state cannot be determined until the vehicle actually slips. Therefore, a driving support device such as ABS can only deal with after the vehicle behavior has already changed. In addition, when the driver performs an appropriate driving in consideration of a road surface condition such as a snowy road, no slip or the like occurs, so the road surface condition cannot be determined from the vehicle state quantity. In addition, when detecting the lane using the camera image, the camera is usually installed so as to capture the image equivalent to the scenery that the driver is viewing, such as the back of the rearview mirror (for pedestrians and other vehicles). In such a position, it is impossible to accurately determine at which timing the wheel has stepped on the lane, and the timing of the departure warning cannot be set accurately. In order to accurately determine whether the wheel has stepped on the lane, it may be possible to place the camera downward so that both the wheel and the lane are visible. Difficult and cost issues arise.

  Then, an object of this invention is to provide the vehicle road surface determination apparatus and driving assistance device which can determine a road surface state at an early stage with a simple structure.

  The vehicle road surface determination device according to the present invention includes a sound collecting unit that collects a traveling sound of a vehicle, and a road surface state determining unit that determines a road surface state based on the traveling sound collected by the sound collecting unit. Features.

  In this vehicle road surface determination device, the running sound of the vehicle is collected by the sound collecting means. When the vehicle travels on the road surface, the vehicle tire and the road surface come into contact with each other, so road noise (friction noise between the tire surface and the road surface) and pattern noise (air vortex (compression / release) in the tire groove) occur. The mode of the noise sound changes depending on the road surface condition. Therefore, the road surface state (for example, a dry road surface state, a wet road surface state, a snow road surface state, a road surface state painted with a white line, etc.) can be determined from the traveling sound of the vehicle. Therefore, in the vehicle road surface determination device, the road surface state determination unit determines the road surface state based on the collected traveling sound. When the state of the road surface in contact with the tire changes, the running sound also changes immediately, so that the road surface state can be determined before the vehicle behavior changes or at the initial stage of the change. Thus, in the vehicle road surface determination device, the road surface state can be determined early with a simple configuration by determining the road surface state using the traveling sound of the vehicle. By using the road surface state determined early (for example, before the vehicle behavior change occurs), appropriate driving assistance can be performed.

  The vehicle road surface determination device of the present invention includes a feature amount extraction unit that extracts a feature amount from the traveling sound collected by the sound collection unit, and the road surface state determination unit is a feature of the traveling sound extracted by the feature amount extraction unit. It is good also as a structure which determines a road surface state based on quantity.

  In this vehicle road surface determination device, when the running sound of the vehicle is collected by the sound collecting means, the feature quantity of the running sound is extracted by the feature quantity extracting means. As the feature amount, for example, there is an intensity for each frequency (wavelength) of running sound. In the vehicle road surface determination device, the road surface state determination unit determines the road surface state based on the feature amount of the traveling sound. Thus, in the vehicle road surface determination device, the road surface state can be determined from the feature amount of the traveling sound with high accuracy and easily by determining the road surface state based on the feature amount extracted from the traveling sound of the vehicle. By using such a highly accurate road surface state, appropriate driving assistance can be performed.

  The vehicle road surface determination device of the present invention includes a learning unit that learns a feature value of a running sound for each road surface state, and the road surface state determination unit is based on the feature amount for each road surface state learned by the learning unit. It is good also as a structure which determines.

  In this vehicle road surface determination device, the characteristic amount of the traveling sound is learned in advance for each road surface state by the learning means. In the vehicle road surface determination device, when the vehicle sound is collected by the sound collecting means and the feature amount of the traveling sound is extracted by the feature amount extraction means, the extracted feature amount is learned by the road surface state determination means. The road surface state is determined by comparing the feature amount for each road surface state. Thus, in the vehicle road surface determination device, the road surface state can be determined more accurately and easily by learning the characteristic amount of the traveling sound for each road surface state in advance. By utilizing such a more accurate road surface state, more appropriate driving assistance can be performed.

  In the vehicle road surface determination device of the present invention, the road surface state determination means determines whether or not the road surface state has a low friction coefficient. By determining such a road surface state with a low friction coefficient, it is possible to perform driving support that copes with the road surface state with a low friction coefficient before the behavior change of the vehicle occurs or at the initial stage of the change. In the vehicle road surface determination device of the present invention, the road surface state determination means determines whether or not the vehicle wheel is on a lane marking. By using accurate information as to whether or not such a wheel has stepped on a lane marking, lane departure warning and lane departure prevention control can be performed with high accuracy.

  In the vehicle road surface determination device of the present invention, the means for collecting the traveling sound has directivity, the traveling sound is acquired for each wheel of the host vehicle, and the road surface state determining unit is provided for each wheel of the host vehicle. It is good also as a structure which determines a state. Thus, the vehicle road surface determination device can perform more appropriate driving support based on the road surface state for each wheel by determining the road surface state for each wheel of the host vehicle. For example, when it is determined whether the road surface state has a low friction coefficient for each wheel, brake control is performed so as not to slip for each wheel, and it is determined whether each wheel is on a lane marking. In such a case, a lane departure determination with higher accuracy is performed to perform lane departure warning and lane departure prevention control.

  The vehicle road surface judging device according to the present invention includes own vehicle beam forming means for performing beam forming using the running sounds collected by the plurality of sound collecting means and acquiring the running sounds of the respective wheels of the own vehicle. It is good also as a structure. In this way, the vehicle road surface determination device does not need to provide sound collecting means for providing directivity to each wheel by providing directivity to the means for collecting traveling sound by beam forming, so that the cost can be reduced. Can be reduced.

  In the vehicle road surface judging device of the present invention, the means for collecting the traveling sound has directivity, the traveling sound of the other vehicles around the own vehicle is acquired, and the road surface state judging means is connected to the other vehicle. It is good also as a structure which determines the state of the road surface which exists. In this way, the vehicle road surface determination device can perform more appropriate driving support based on the road surface state of the other vehicle by determining the road surface state of the other vehicle surroundings (for example, in front of the host vehicle). it can. For example, when a road surface state with a low friction coefficient is determined as the road surface state of the preceding vehicle, appropriate driving assistance is performed before reaching the road surface state for the own vehicle traveling on the same road surface several seconds later, It can also be determined that the preceding vehicle has changed lanes when the vehicle is determined to be on the lane marking.

  The vehicle road surface determination device of the present invention may be configured to include other vehicle beam forming means for performing beam forming using traveling sounds collected by a plurality of sound collecting means and acquiring traveling sounds of other vehicles. . In this way, the vehicle road surface determination device does not need to provide sound collecting means having directivity in front of the host vehicle or the like by providing directivity to the means for collecting traveling sound by beam forming. Cost can be reduced. Also, by using beam forming, the traveling sound of other vehicles and the traveling sound for each wheel of the own vehicle can be acquired by the same plurality of sound collecting means, the traveling sound of the other vehicle and the traveling sound for each wheel of the own vehicle Can be separated.

  In the vehicle road surface determination device of the present invention, the plurality of sound collecting means are configured so that the directivity position opposite to the directivity target position by beam forming is inside the own vehicle and there is no sound source in the own vehicle. It arrange | positions so that it may become a location, and the directivity position on the opposite side to the position of the directivity object by beam forming may be the position outside the own vehicle and close to the own vehicle. In the vehicle road surface determination device, by arranging a plurality of sound collecting means in this way, the sound at the directivity position opposite to the directivity target position is suppressed as much as possible. Without being received, the traveling sound for each wheel of the host vehicle and the traveling sound of other vehicles can be obtained with high accuracy.

  In the vehicle road surface determination device of the present invention, it is preferable that the number of sound collecting means is three or more. As described above, in the vehicle road surface determination device, beam forming can be performed on one point in space by performing beam forming using three or more sound collecting means. And the running sound of other vehicles can be acquired with higher accuracy.

  In the vehicle road surface determination device of the present invention, the position detection means for detecting the position of the own vehicle, the position of the own vehicle detected by the position detection means, and the feature quantity extracted by the feature quantity extraction means are stored in association with each other. The road surface state determination means compares the feature amount extracted at this time by the feature amount extraction means at the position of the host vehicle with the feature amount at the same position stored in the storage means, and compares the comparison It is good also as a structure which determines whether the road surface state is changing based on a result.

  In this vehicle road surface determination device, a feature amount is extracted from traveling sound collected during traveling of the host vehicle, the position of the host vehicle is acquired by a position detecting unit, and the position of the host vehicle and the characteristics of the traveling sound are stored in a storage unit. The amount is stored in association with each other. When the same vehicle is driven at the same position, it can be inferred that the road surface state has changed if the characteristic amount of the running sound has changed when the vehicle has traveled in the past. Therefore, in the vehicle road surface determination device, when information on the same position is stored in the storage means, the feature amount extracted this time by the feature amount extraction means by the road surface state determination means and the same position stored in the storage means. Are compared with each other, and based on the comparison result, it is determined whether or not the road surface condition has changed from when traveling in the past. Thus, in the vehicle road surface determination device, by comparing the feature amount of the current traveling sound and the past feature amount at the same position, without performing prior learning or identification of a specific road surface state, etc. A change in road surface condition can be determined with a simpler configuration.

  A driving support apparatus according to the present invention includes a sound collecting unit that collects a traveling sound of a vehicle, a road surface state determining unit that determines a road surface state based on the traveling sound collected by the sound collecting unit, and a road surface state determining unit. And driving support means for providing driving support based on the determined road surface condition. Thus, the driving assistance device can perform appropriate driving assistance by using the road surface state determined based on the traveling sound. For example, by using the road surface condition determined before or at the initial stage of the change of the vehicle behavior, the change of the behavior of the vehicle can be prevented before the change of the vehicle behavior occurs or at the initial stage of the change. Suppressing driving assistance can be performed.

  The driving support apparatus according to the present invention includes a feature amount extraction unit that extracts a feature amount from the running sound collected by the sound collection unit, and the road surface state determination unit uses the feature amount of the running sound extracted by the feature amount extraction unit. It is good also as a structure which determines a road surface state based on it. As described above, the driving assistance device can provide high-precision driving assistance by using the high-accuracy road surface state determined based on the feature amount of the traveling sound.

  The driving support apparatus according to the present invention includes a learning unit that learns a feature amount of a traveling sound for each road surface state, and the road surface state determination unit determines the road surface state based on the feature amount for each road surface state learned by the learning unit. It is good also as composition to do. Thus, the driving assistance device can perform driving assistance with higher accuracy by using the highly accurate road surface state determined based on the result learned in advance.

  In the driving support device of the present invention, the road surface state determining means determines whether or not the road surface state has a low friction coefficient, and the driving support means determines that the road surface state determining means determines that the road surface state has a low friction coefficient. Driving assistance is provided according to the road surface condition with a low coefficient of friction. In this way, the driving assistance device uses the low friction coefficient road surface state to perform appropriate driving assistance according to the low friction coefficient road surface state before the vehicle behavior change occurs or at the initial stage of the change. Can do.

  In the driving support device of the present invention, the means for collecting the traveling sound has directivity, the traveling sound is obtained for each wheel, and the road surface state determining means is a road surface state having a low friction coefficient for each wheel of the host vehicle. The driving support means supports driving according to the position of the wheel of the host vehicle determined by the road surface state determining means as the road surface state having a low friction coefficient. In this way, the driving support device uses the information on whether or not the road surface state has a low friction coefficient for each wheel, so that the road surface with a low friction coefficient is obtained before the behavior change of the vehicle for each wheel or at the initial stage of the change. Appropriate driving assistance according to the state can be performed.

  In the driving support device of the present invention, the road surface state determining means determines whether or not the wheel of the own vehicle is on the lane line of the lane, and the driving support means is the road surface state determining means and the wheel of the own vehicle is on the lane line of the lane. If it is determined, lane departure warning or lane departure prevention control is performed. As described above, the driving support device can perform the lane departure warning and the lane departure prevention control with high accuracy timing by using accurate information on whether or not the wheel of the host vehicle has stepped on the lane marking. it can.

  In the driving support device of the present invention, the means for collecting the traveling sound has directivity, the traveling sound is acquired for each wheel, and the road surface condition determining means is on the lane marking for each wheel of the host vehicle. The driving support means performs lane departure warning or lane departure prevention control according to the wheel position of the host vehicle determined by the road surface state determination means as being on the lane marking of the lane. In this way, the driving support device makes a more accurate lane departure determination by using information on whether or not the vehicle is on the lane marking for each wheel of the host vehicle, and an appropriate lane departure warning at an appropriate timing. And lane departure prevention control.

  In the driving support device of the present invention, the means for collecting the traveling sound has directivity, the traveling sound of the other vehicle around the own vehicle is acquired, and the road surface condition determining means is the road surface on which the other vehicle is grounded. The driving support means supports driving according to the road surface condition of the other vehicle determined by the road surface condition determining means. In this way, the driving support device can perform appropriate driving support before reaching the road surface state by using the road surface state of another vehicle such as the front of the host vehicle.

  In the driving support apparatus according to the present invention, a position detecting unit that detects the position of the host vehicle, and a storage unit that stores the position of the host vehicle detected by the position detecting unit and the feature amount extracted by the feature amount extracting unit in association with each other. The road surface state determination means compares the feature amount extracted this time by the feature amount extraction means at an arbitrary position of the host vehicle with the feature amount at the same position stored in the storage means, and determines the comparison result. Based on this, it is determined whether or not the road surface state has changed, and the driving support means provides driving support when the road surface state determination means determines that the road surface state has changed at the same position traveled in the past. In this way, the driving support device can determine the change in the road surface state with a simpler configuration by comparing the feature amount of the traveling sound at the same position with the past feature amount, and against the change in the road surface state. Appropriate driving assistance can be provided.

  ADVANTAGE OF THE INVENTION According to this invention, a road surface state can be determined early by simple structure by determining a road surface state using the running sound of a vehicle.

It is a block diagram of the learning apparatus which concerns on 1st-4th embodiment. It is a block diagram of the driving assistance system which concerns on 1st-4th embodiment. It is an example of the change of a road surface state. It is explanatory drawing of the feature-value of the running sound which concerns on this Embodiment. It is an example of the running sound feature-value / road surface state table which concerns on 1st Embodiment, (a) is a table for every specific road surface state, (b) is a table for every road surface friction coefficient. It is a flowchart which shows the flow of operation | movement in the driving assistance system which concerns on 1st Embodiment. The white line (lane) is an example of a painted road surface, (a) is a case where the host vehicle is not stepping on the white line, and (b) is a case where the right front wheel of the host vehicle is stepping on the white line. It is an example of the running sound feature-value / road surface state table which concerns on 2nd Embodiment, (a) is a table of a road surface and a white line, (b) is a table of the road surface and white line for every specific road surface state. . It is a flowchart which shows the flow of operation | movement in the driving assistance system which concerns on 2nd Embodiment. It is explanatory drawing of the beam forming which concerns on 3rd Embodiment. It is an example of the wheel which stepped on the white line on the road surface where the white line (lane) was painted, (a) is the case where the right front wheel of the own vehicle is stepping on the white line, and (b) is the right rear wheel of the own vehicle. This is the case where the white line is stepped on. It is another example of the wheel which stepped on the white line on the road surface where the white line (lane) was painted, (a) is a case where the right front wheel of the own vehicle is stepping on the white line, and (b) is the right front wheel of the own vehicle. And when the right rear wheel is stepping on the white line. It is an example of the wheel which entered the part of the low friction coefficient on the road surface with the part of the low friction coefficient, (a) is a case where all the wheels of the own vehicle are not in the part of the low friction coefficient, and (b) This is a case where the front wheels of the host vehicle enter a portion with a low friction coefficient, (c) is a case where all the wheels of the host vehicle enter a portion with a low friction coefficient, and (d) is a case where the rear wheels of the host vehicle are low. This is a case where the friction coefficient is entered, and (e) is a case where the left wheel of the host vehicle enters a low friction coefficient. It is a flowchart which shows the flow of operation | movement in the driving assistance system which concerns on 3rd Embodiment. It is explanatory drawing of the directivity to both sides by beam forming. It is an example of arrangement | positioning of the sound collector for performing beam forming. It is an example of beam forming to another vehicle on a road surface with a low friction coefficient portion, (a) is a case where the traveling sound of the other vehicle is not obtained by beam forming, and (b) is a portion having a low friction coefficient. This is a case where the running sound of the other vehicle not entering is obtained by beam forming, and (c) is a case where the running sound of the other vehicle entering the portion of the low friction coefficient is obtained by beam forming. It is an example of beam forming to other vehicles on a curve road. It is a flowchart which shows the flow of operation | movement in the driving assistance system which concerns on 4th Embodiment. It is a block diagram of the driving assistance system which concerns on 5th Embodiment. It is a flowchart which shows the flow of operation | movement in the driving assistance system which concerns on 5th Embodiment.

  Hereinafter, embodiments of a vehicle road surface determination device and a driving support device according to the present invention will be described with reference to the drawings. In addition, the same code | symbol is attached | subjected about the element which is the same or it corresponds in each figure, and the overlapping description is abbreviate | omitted.

  In the present embodiment, the present invention is applied to a driving support system (provided with a learning device) including a road surface state determination device and a driving support device. The driving support system according to the present embodiment determines the road surface state based on the traveling sound of the vehicle collected by the sound collector (microphone) in the road surface state determining device, and the driving support device responds to the road surface state. Provide driving assistance. In this embodiment, there are five forms, and the first to fourth embodiments are composed of a learning device that performs prior learning at the time of vehicle development and a driving support system that is mounted on each vehicle at the time of vehicle shipment. There is a fifth embodiment in which only a driving support system is installed in each vehicle at the time of vehicle shipment.

  The traveling sound of the vehicle is mainly road noise (friction sound between the tire surface and the road surface) and pattern noise (air vortex (compression / release) in the tire groove). Such road noise and pattern noise change according to the road surface condition because the contact state between the tire and the contact changes depending on the road surface condition. The range of the frequency component of the running sound of the vehicle may be measured in advance through experiments or the like.

  A learning device 1A and a driving support system 2A according to the first embodiment will be described with reference to FIGS. FIG. 1 is a configuration diagram of a learning device according to the first to fourth embodiments. FIG. 2 is a configuration diagram of the driving support system according to the first to fourth embodiments. FIG. 3 is an example of a change in road surface condition. FIG. 4 is an explanatory diagram of the feature amount of the running sound according to the present embodiment. FIG. 5 is an example of a running sound feature / road surface state table according to the first embodiment, where (a) is a table for each specific road surface state, and (b) is a table for each road surface friction coefficient. is there.

  Prior learning by the learning device 1A is performed at the time of vehicle development, and a table based on the learning is mounted on the driving support system 2A of each vehicle. When performing prior learning, an actual vehicle experiment is performed in which an experimental vehicle is traveled on roads in various road surfaces using an experimental vehicle that collects traveling sound.

  The learning device 1A will be described. The learning device 1A collects traveling sound in various road surface conditions, and generates a table of feature values of the traveling sound for each road surface state using the traveling sound data for each road surface state. The learning device 1A includes a sound collecting device 10A and a computer 11A. The configuration of the learning device 1A may be a configuration in which both the sound collecting device 10A and the computer 11A are mounted on the experimental vehicle, or the sound collecting device 10A and the data storage device are mounted on the experimental vehicle, and the indoor computer 11A. Alternatively, the running sound data may be input from the data storage device. In the first embodiment, the computer 11A of the learning device 1A corresponds to the learning means described in the claims.

  The sound collector 10A has one or more sound collectors. For example, the one or more sound collectors are arranged side by side in the vehicle width direction (left-right direction) at the front end of the vehicle (in the case of one, they are arranged at the center of the front end of the vehicle). Moreover, in a prior experiment, in order to collect the traveling sound of each wheel with high accuracy, one sound collector may be arranged near each wheel toward the installation location of the tire and the road surface. The sound collector is an acoustoelectric converter that collects ambient sounds outside the vehicle and converts the collected sounds into electrical signals. The sound collector 10A collects sound with a sound collector while the experimental vehicle is running, and stores the collected sound data.

  In the actual vehicle experiment, the experimental vehicle is run for a specific time on a road in various road surface conditions, and the sound is collected by the sound collector of the sound collector 10A. The collected sound data is stored in a data storage device. The sound data may be sequentially input to the computer 11A without being stored in the data storage device. The road surface state may be a specific state such as a dry state, a wet state, a compressed snow state, an ice film state, or an ice plate state, or a state represented by a numerical value such as a road surface friction coefficient (road surface μ) at each stage. But you can. For example, on a road such as a dry road surface DR or a wet road surface WR as shown in FIG. Alternatively, on the road having the road surface friction coefficient at each stage, the experimental vehicle is allowed to travel for a specific time to collect the traveling sound. In an actual vehicle experiment, the same road surface condition may be traveled a plurality of times and travel sound data for the same road surface condition may be acquired a plurality of times, or the vehicle may be traveled once in the same road surface condition and travel sound in the same road surface condition. Data may be acquired once.

  The computer 11A is a computer such as a personal computer, and includes a traveling sound feature amount extraction unit 12A, a learning unit 13A, and a traveling sound feature amount / road surface state table storage unit 14A. The sound data (electrical signal) of each road surface state collected by the sound collector of the sound collector 10A is input to the computer 11A. In the computer 11A, as a preprocessing, an analog electric signal is converted into a digital electric signal for each sound data, and a predetermined frequency band (a band sufficiently including the frequency band of the running sound of the vehicle) is converted from the electric signal. High high frequency band and low low frequency band).

The running sound feature extraction unit 12A extracts the running sound feature from the sound data of the pre-collected sound collector for each road surface state. In the present embodiment, as the feature amount, as shown in FIG. 4, each frequency of FFT [Fast Fourier Transform] at a specific frequency (for example, 500 Hz to 2000 Hz) of sound data for a specific time (for example, 20 msec). It is assumed that the intensity for each (for example, 10 Hz) is vectorized. The intensity may be normalized so that the norm is 1. In addition, when the traveling sound data for a plurality of times are acquired in the same road surface state, the averaged feature amount is extracted. In the example shown in FIG. 4, feature amount _{S dry} dry road surface, the intensity _i 510 at intensity _i 500, 510HZ at 500 Hz, · · ·, a vector of intensity _{i 2000} in 2000 Hz.

In the learning unit 13A, the traveling sound feature amount / road surface state table is generated by associating the traveling sound feature amount extracted by the traveling sound feature amount extracting unit 12A for each road surface state. FIG. 5A shows, for five specific road surface conditions, a running sound feature value S _{dry in} a _dry state, a running sound feature value S _{wet in} a _wet state, a running sound feature amount S _{snow in} a compressed snow state, and an ice film. running sound feature quantity of state _{S ICE1,} table of traveling sound feature value _{S ICE2} ice plate state is shown. Further, in FIG. 5B, for a 10-step road friction coefficient, a running sound feature amount S _{10 with} a road surface friction coefficient of 1.0, a running sound feature amount S _{09 with} a road surface friction coefficient of 0.9, A table including a running sound feature amount S _{08 with} a road surface friction coefficient of 0.8,... And a running sound feature amount S _{01 with} a road surface friction coefficient of 0.1 is shown.

  The traveling sound feature / road surface state table storage unit 14A is configured in a predetermined area of the storage device of the computer 11A, and stores the traveling sound feature / road surface state table generated by the learning unit 13A.

  The driving support system 2A will be described. The driving support system 2A is mounted on the vehicle, determines which road surface state the traveling sound data collected during traveling of the vehicle is, and performs driving support according to the determined road surface state. The driving support system 2A includes a road surface state determination device 3A and a driving support device 4A.

  The road surface state determination device 3A will be described. The road surface state determination device 3A collects the traveling sound of the host vehicle while the vehicle is traveling, extracts the feature amount of the traveling sound of the host vehicle, and compares the feature amount with the traveling sound feature amount / road surface state table. The road surface state that is currently running is determined. The road surface state determination device 3A includes a sound collecting device 30A and an ECU [Electronic Control Unit] 31A.

  The sound collector 30A has one or more sound collectors. For example, the one or more sound collectors are arranged side by side in the vehicle width direction (left-right direction) at the front end of the vehicle (in the case of one, they are arranged at the center of the front end of the vehicle). By arranging in this way, it can be shared with a sound collector used for recognition of other vehicles approaching. The sound collection device 30A collects sound with a sound collector while the vehicle is traveling, and transmits the collected sound data to the ECU 31A. In the first embodiment, the sound collector of the sound collecting device 30A corresponds to the sound collecting means described in the claims.

  The ECU 31A is an electronic control unit including a CPU [Central Processing Unit], a ROM [Read Only Memory], a RAM [Random Access Memory], and the like, and comprehensively controls the road surface state determination device 3A. The ECU 31A includes a traveling sound feature amount extraction unit 32A, a road surface state determination unit 33A, and a traveling sound feature amount / road surface state table storage unit 34A. The ECU 31A receives sound data (electrical signal) collected by the sound collector from the sound collector 30A. The ECU 31A performs the same processing as the computer 11A of the learning device 1A as preprocessing for the sound data. In the first embodiment, the running sound feature amount extraction unit 32A corresponds to the feature amount extraction unit described in the claims, and the road surface state determination unit 33A includes the road surface state determination unit described in the claims. It corresponds to.

  The traveling sound feature / road surface state table storage unit 34A is configured in a predetermined area of the storage device of the ECU 31A, and is stored in the traveling sound feature / road surface state table storage unit 34A of the learning device 1A. A state table is stored.

  The traveling sound feature amount extraction unit 32A performs the same processing as the traveling sound feature amount extraction unit 12A of the learning device 1A. Each time the running sound feature amount extraction unit 32A receives sound data of the sound collector of the sound collecting device 30A, the running sound feature amount extraction unit 32A extracts the running sound feature amount using the pre-processed sound data.

  The road surface state determination unit 33A compares the travel sound feature amount for each road surface state stored in the travel sound feature amount / road surface state table storage unit 34A with the feature amount extracted by the travel sound feature amount extraction unit 32A. A road surface state having a feature amount closest to the feature amount extracted by the traveling sound feature amount extraction unit 32A (that is, a road surface state in which the host vehicle is currently traveling) is determined. As this determination method, there are, for example, a k-NN [k-nearest neighbor] method (k-nearest neighbor method) and a multi-class SVM [MultiClass Support Vector Machine] using an existing identification method.

  In ECU31A, road surface state information is produced | generated based on the determination result in the road surface state determination part 33A, and road surface state information is transmitted to the driving assistance apparatus 4A. The road surface condition information includes, for example, information on whether the current road surface state is a dry state, a wet state, a compressed snow state, an ice film state, an ice plate state, or the like, or the friction of the road surface currently being traveled Information about the coefficient.

  The driving support device 4A is a device that includes various sensors, an ECU, and the like, and supports various driving operations for the driver. In particular, when the driving support device 4A receives road surface state information from the road surface state determination device 3A, the driving support device 4A performs driving support related to the road surface state. For example, in the case of ABS that changes the control according to the road surface state, if road surface state information with a low friction coefficient that is more likely to slip than a dry road surface is acquired as road surface state information, wheels are displayed when the driver is braking. Brake control is performed before the vehicle slips (for example, the vehicle speed is reduced to a safe speed range with a gentle deceleration). In addition, in the case of PCS [Pre-Crash Safety System], if road surface information with a lower friction coefficient than that of a dry road surface is acquired as road surface state information, be careful when there is an obstacle that may cause a collision. Change alerting and vehicle control selection, alert and vehicle control content and timing. As a driving support method according to the road surface condition, for example, alerting or vehicle control is set for each road surface condition, and alerting or vehicle control according to the determined road surface condition is selected. Implement the selected driving assistance. In the first embodiment, the driving support device 4A corresponds to the driving support means described in the claims.

  The operation of the first embodiment will be described with reference to FIGS. In particular, the operation in the driving support system 2A will be described along the flowchart of FIG. FIG. 6 is a flowchart showing a flow of processing in the driving support system according to the first embodiment.

  In a prior actual vehicle experiment, the experimental vehicle is driven for a specific time on each road surface road. At that time, the sound collecting device 10A of the learning device 1A mounted on the vehicle collects sounds outside the vehicle with the sound collector, converts the collected sounds into electric signals, and stores the sound data as data. Store in the device. The sound data of the sound collector collected for each road surface condition is input to the computer 11A of the learning device 1A. The computer 11A performs preprocessing on each sound data.

  In the computer 11A, for each road surface state, the feature value of the running sound is extracted from the sound data of the pre-collected sound collector. Then, the computer 11A generates a table by associating the feature amounts for each road surface state, and stores the table in the traveling sound feature amount / road surface state table storage unit 14A.

  Each vehicle is equipped with a driving support system 2A, and the driving sound feature / road surface state table storage unit 34A of the road surface determination device 3A of the driving support system 2A stores the driving sound feature / road surface state table of the learning device 1A. The traveling sound feature amount / road surface state table of the unit 14A is stored.

  In each vehicle, the sound collector of the sound collecting device 30A of the road surface condition determining device 3A collects (measures) ambient sounds outside the vehicle, converts the collected sounds into electric signals, and transmits them to the ECU 31A. (S10). The ECU 31A receives the sound data of the sound collector and performs preprocessing on the sound data.

  Every time the ECU 31A receives the sound data of the sound collector from the sound collecting device 30A, the ECU 31A extracts the characteristic amount of the running sound from the pre-processed sound data (S11). Then, the ECU 31A determines, based on the traveling sound feature / road surface state table stored in the traveling sound feature / road surface state table storage unit 34A, which road surface state the extracted feature is closest to. (S12). The ECU 31A generates road surface state information based on the determination result, and transmits the road surface state information to the driving support device 4A.

  When the driving support device 4A receives the road surface state information from the road surface state determination device 3A, the driving support device 4A performs driving support according to the road surface state (S13).

  According to the first embodiment, a table of running sound feature values for each road surface state is generated by learning by the learning device 1A, and the running sound feature values of the own vehicle actually running are based on the table. By determining which road surface state, the road surface state can be determined early with a simple configuration. When the state of the road surface in contact with the tire changes, the running sound also changes immediately. Therefore, the road surface state can be determined early before the behavior of the vehicle changes or at the initial stage of the change. By using the road surface state determined at an early stage in this way, it is possible to predict a future change in the behavior of the vehicle and to perform appropriate driving assistance. As a result, a safer driving support system can be provided.

  Further, according to the first embodiment, by determining the road surface state based on the feature amount extracted from the traveling sound of the vehicle, the road surface state can be determined with high accuracy and easily from the feature amount of the traveling sound. Furthermore, according to the first embodiment, the road surface state can be determined with higher accuracy by learning the feature amount of the traveling sound for each road surface state in advance.

  A learning device 1B and a driving support system 2B according to a second embodiment will be described with reference to FIGS. 1, 2, 4, 7, and 8. FIG. FIG. 7 is an example of a road surface painted with a white line (lane), (a) is a case where the host vehicle is not stepping on the white line, and (b) is a case where the right front wheel of the host vehicle is stepping on the white line. It is. FIG. 8 is an example of a running sound feature / road surface state table according to the second embodiment, where (a) is a table of road surfaces and white lines, and (b) is a road surface and white lines for each specific road surface state. It is a table.

  The second embodiment is different from the first embodiment in that the road surface state to be determined is a lane marking (particularly, a white line painted on the road surface). By determining whether or not the vehicle has stepped on a lane marking (white line), the driving assistance determines the possibility of lane departure. If there is a possibility of lane departure, departure warning, vehicle control, etc. I do. Prior learning in the learning device 1B is performed at the time of vehicle development, and a table based on the learning is mounted on the driving support system 2B of each vehicle. In the case of performing prior learning, an actual vehicle experiment is performed in which an experimental vehicle travels with and without the wheels stepping on a white line, using an experimental vehicle that collects traveling sound. In addition, although it was set as the white line as the section line of a lane, other division lines, such as a yellow line, are also object.

  The learning device 1B will be described. The learning device 1B is in a state where all the wheels are not stepping on the white line (lane markings) (on the road surface where all the wheels are not painted with white lines) and a state where the wheels are stepping on the white line (at least one wheel is on the white line). The travel sound is collected, and the travel sound data table on the road surface and the white line is generated using the travel sound data on the road surface and the white line. Further, as in the first embodiment, traveling sound is collected for each road surface state, and a table of feature amounts of traveling sound on the road surface and white lines for each road surface state is generated. The learning device 1B includes a sound collection device 10B and a computer 11B. In the second embodiment, the computer 11B of the learning device 1B corresponds to the learning means described in the claims.

  Since the sound collecting device 10B is the same device as the sound collecting device 10A according to the first embodiment, the description thereof is omitted. In an actual vehicle experiment, in a predetermined road surface condition (for example, dry state), all the wheels are not stepping on the white line (see FIG. 7A) and at least one wheel is stepping on the white line (see FIG. 7B). Then, the experimental vehicle is run for a specific time, and the sound is collected by the sound collector of the sound collector 10B. The collected sound data is stored in a data storage device. Sound data may be sequentially input to the computer 11B without being stored in the data storage device. Further, as in the first embodiment, on roads in various road surface conditions, the test vehicle is run for a specific time with all wheels not stepping on the white line and at least one wheel stepping on the white line, respectively. Sound is collected by the sound collector of the apparatus 10B.

  The computer 11B is a computer such as a personal computer, and includes a traveling sound feature amount extraction unit 12B, a learning unit 13B, and a traveling sound feature amount / road surface state table storage unit 14B. The computer 11B includes a state in which the white line in a predetermined road surface state collected by the sound collector of the sound collector 10B is not stepped on and a state in which the white line is stepped on (or a state in which the white line for each road surface state is not stepped on) Sound data (electrical signal) in a state where the white line is stepped on is input. In the computer 11B, as preprocessing, an analog electrical signal is converted into a digital electrical signal for each sound data, and a predetermined frequency band is removed from the electrical signal.

  In the running sound feature amount extraction unit 12B, preprocessing for a state where the white line is not stepped on and a state where the white line is stepped (or a state where the white line for each road surface state is not stepped on and a state where the white line is stepped on). The feature amount of the running sound is extracted from the sound data of the later sound collector.

In the learning unit 13B, a state where the white line is not stepped (all wheels are on the road surface) and a state where the white line is stepped (at least one wheel is on the white line) (or the state where the white line for each road surface state is not stepped on and the white line The travel sound feature quantity / road surface state table is generated by associating the travel sound feature quantity extracted by the travel sound feature quantity extraction unit 12B. FIG. 8A shows a table composed of the traveling sound feature amount S on all road surfaces and the traveling sound feature amount P on at least one wheel on a white line in a predetermined road surface state. In FIG. 8B, for five specific road surface conditions, the running sound feature amount S _dry on all road surfaces in the dry state, the running sound feature amount P _dry on at least one wheel on the white line, and the wet state , The running sound feature amount S _wet on the all-wheel road surface and the running sound feature amount P _wet on at least one wheel on the white line, and the running sound feature amount S _snow on at least one wheel on the white line in the _snow- capped state. the amount _{P snow,} an ice film state and the running sound feature value _{S ICE1} on all-wheel road surface at least one wheel is traveling sound feature value _{P ICE1} on white line, a traveling sound feature value _{S ICE2} on all-wheel road surface in an ice plate state at least A table is shown in which one wheel is composed of a running sound feature value _Pice2 on a white line. In addition, as a road surface state, the road surface state of the road surface friction coefficient of each step may be sufficient.

  The traveling sound feature / road surface state table storage unit 14B is configured in a predetermined area of the storage device of the computer 11B, and stores the traveling sound feature / road surface state table generated by the learning unit 13B.

  The driving support system 2B will be described. The driving support system 2B is mounted on the vehicle and determines whether or not the wheel is stepping on the white line (lane line of the lane) based on the traveling sound data collected during traveling of the host vehicle, and according to the determination result. Judging the possibility of lane departure and providing driving assistance. The driving support system 2B includes a road surface state determination device 3B and a driving support device 4B.

  The road surface state determination device 3B will be described. The road surface state determination device 3B collects the traveling sound of the host vehicle while the vehicle is traveling, extracts the feature amount of the traveling sound of the host vehicle, and compares the feature amount with the traveling sound feature amount / road surface state table. It is determined whether the wheel is on the white line. The road surface state determination device 3B includes a sound collection device 30B and an ECU 31B.

  The sound collector 30B has the same configuration as the sound collector 30A according to the first embodiment. The sound collection device 30B collects sound with the sound collector while the vehicle is traveling, and transmits the collected sound data to the ECU 31B. In the second embodiment, the sound collector of the sound collecting device 30B corresponds to the sound collecting means described in the claims.

  Incidentally, when a sound collector is arranged for each wheel, the traveling sound for each wheel can be collected, so it is possible to determine whether or not the white line is stepped on for each wheel. When one sound collector is arranged, the running sound for each wheel cannot be collected, so it is possible to determine whether any of the four wheels is stepping on the white line.

  The ECU 31B is an electronic control unit including a CPU, a ROM, a RAM, and the like, and comprehensively controls the road surface state determination device 3B. The ECU 31B includes a traveling sound feature amount extraction unit 32B, a road surface state determination unit 33B, and a traveling sound feature amount / road surface state table storage unit 34B. The ECU 31B receives sound data (electrical signal) collected by the sound collector from the sound collecting device 30B. The ECU 31B performs the same process as the ECU 31A according to the first embodiment as a pre-process for the sound data. In the second embodiment, the traveling sound feature amount extraction unit 32B corresponds to the feature amount extraction unit described in the claims, and the road surface state determination unit 33B includes the road surface state determination unit described in the claims. It corresponds to.

  The traveling sound feature / road surface state table storage unit 34B is configured in a predetermined area of the storage device of the ECU 31B, and is stored in the traveling sound feature / road surface state table storage unit 34B of the learning device 1B. A state table is stored.

  The traveling sound feature amount extraction unit 32B performs the same processing as the traveling sound feature amount extraction unit 32A according to the first embodiment. Each time the running sound feature quantity extraction unit 32B receives the sound data of the sound collector of the sound collecting device 30B, the running sound feature quantity extraction unit 32B extracts the running sound feature quantity using the pre-processed sound data.

  In the road surface state determination unit 33B, the running sound feature amount on the road surface and on the white line stored in the traveling sound feature amount / road surface state table storage unit 34B, and the feature amount extracted by the traveling sound feature amount extraction unit 32B are used. By comparing each of them, a road surface state having a feature amount closer to the feature amount extracted by the traveling sound feature amount extraction unit 32B (that is, whether all wheels are on the road surface or few wheels are on the white line) is determined. Furthermore, when the travel sound feature value / road surface state table is a table of travel sound feature values on the road surface and the white line for each road surface state, a specific road surface state is also determined.

  In ECU31B, road surface state information is produced | generated based on the determination result in the road surface state determination part 33B, and road surface state information is transmitted to the driving assistance apparatus 4B. The road surface state information is, for example, information on whether all the wheels are not stepping on the white line or at least one wheel is stepping on the white line, and is information on a specific road surface state.

  The driving support device 4B includes various sensors, ECUs, and the like, and is a device that supports various driving operations for the driver. In particular, when the driving support device 4B receives road surface state information from the road surface state determination device 3B, it determines whether all the wheels are stepping on the white line or at least one wheel is stepping on the white line, Judges that there is a possibility of lane departure and provides driving assistance. For example, in the case of a lane departure warning device or lane departure prevention device, it is determined that there is no possibility of lane departure when the white line L is not stepped on as shown in FIG. Thus, when the white line L is stepped on, it is determined that there is a possibility of lane departure, and when there is a possibility of lane departure, vehicle control such as preset alarm output and steering control is performed. In the second embodiment, the driving support device 4B corresponds to the driving support means described in the claims.

  The operation of the second embodiment will be described with reference to FIGS. 1, 2, 4, 7, and 8. In particular, the operation in the driving support system 2B will be described along the flowchart of FIG. FIG. 9 is a flowchart showing a flow of processing in the driving support system according to the second embodiment. Here, it is assumed that the road surface state is a specific road surface state such as a dry state, and driving assistance regarding lane departure is provided as driving assistance.

  In an actual vehicle experiment in advance, the test vehicle is allowed to travel for a specific time in a state where all the wheels are not stepping on the white line and at least one wheel is stepping on the white line. At that time, in the sound collecting device 10B of the learning device 1B mounted on the vehicle, the sound around the vehicle is collected by the sound collector, the collected sound is converted into an electric signal, and the data is stored in the data storage device. Remember me. The sound data of the sound collector collected in a state where the white line is not stepped on and a state where the white line is stepped on is input to the computer 11B of the learning apparatus 1B. In the computer 11B, pre-processing is performed on each sound data.

  In the computer 11B, the characteristic amount of the running sound is extracted from the sound data of the sound collector after the preprocessing for the state where the white line is not stepped on and the state where the white line is stepped. Then, the computer 11B generates a table by associating the feature amount with the state where the white line is not stepped on and the state where the step is stepped on, and stores the table in the running sound feature amount / road surface state table storage unit 14B.

  Each vehicle is equipped with a driving support system 2B, and the driving sound feature / road surface state table storage unit 34B of the road surface determination device 3B of the driving support system 2B stores the driving sound feature / road surface state table of the learning device 1B. The running sound feature amount / road surface state table of the unit 14B is stored.

  In each vehicle, the sound collector of the sound collecting device 30B of the road surface condition determining device 3B collects sounds outside the vehicle, converts the collected sounds into electric signals, and transmits them to the ECU 31B (S20). ). The ECU 31B receives the sound data of the sound collector and performs preprocessing on the sound data.

  Each time the ECU 31B receives the sound data of the sound collector from the sound collecting device 30B, the ECU 31B extracts the feature amount of the running sound from the sound data of the sound collector after the preprocessing (S21). Then, in the ECU 31B, based on the running sound feature / road surface state table stored in the running sound feature / road surface state table storage unit 34B, the extracted feature is the feature when the white line is not stepped on or the white line. It is determined whether it is a feature amount when stepping on (S22). The ECU 31B generates road surface state information based on the determination result, and transmits the road surface state information to the driving support device 4B.

  When receiving the road surface state information from the road surface state determination device 3B, the driving support device 4B determines whether all the wheels are on the road surface based on the road surface state (S23). When it is determined in S23 that all the wheels are on the road surface, the driving support device 4B determines that there is no lane departure (S24). If it is determined in S23 that at least one wheel is on the white line, the driving support device 4B determines that there is a possibility of lane departure (S25), and alerts the driver (S26).

  According to the second embodiment, a table of running sound feature values on the road surface and on the white line is generated by learning by the learning device 1B, and the wheels of the own vehicle actually running are based on the table. By determining whether the vehicle is on the road surface or the white line, it is possible to accurately determine whether the wheel is stepping on the white line with a simple configuration. By using information on whether or not the white line of the accurate timing is stepped on in this way, it is possible to accurately determine at which timing the own vehicle deviates from the lane, and the lane after the own vehicle actually steps on the white line. Deviation warning and vehicle control for preventing vehicle deviation can be performed with high accuracy. As a result, a more reliable driving support system can be provided.

  A learning device 1C and a driving support system 2C according to a third embodiment will be described with reference to FIGS. 1, 2, 4, and 10 to 13. FIG. FIG. 10 is an explanatory diagram of beamforming according to the third embodiment. FIG. 11 is an example of a wheel that steps on the white line on the road surface on which the white line (lane) is painted. FIG. 11A shows a case where the right front wheel of the host vehicle steps on the white line, and FIG. This is when the right rear wheel is stepping on the white line. FIG. 12 is another example of a wheel that steps on the white line (lane) on the road surface, and (a) shows a case where the right front wheel of the host vehicle steps on the white line, and (b) shows that This is a case where the right front wheel and the right rear wheel of the vehicle are stepping on the white line. FIG. 13 is an example of a wheel that has entered a location with a low friction coefficient on a road surface with a location with a low friction coefficient, and (a) is a case where all the wheels of the host vehicle are not in a location with a low friction coefficient. (B) is the case where the front wheels of the host vehicle enter a location with a low friction coefficient, (c) is the case where all the wheels of the host vehicle enter a location with a low friction coefficient, and (d) This is a case where the rear wheel enters a portion having a low friction coefficient, and (e) is a case where the left wheel of the host vehicle enters a portion having a low friction coefficient.

  Compared with the first and second embodiments, the third embodiment obtains a running sound for each wheel, extracts a feature value, determines the road surface state for each wheel, and sets the road surface state for each wheel. Depending on the driving assistance. Since the road surface condition is obtained for each lane, more detailed and highly accurate driving assistance is possible.

  In the third embodiment, in order to obtain the traveling sound for each wheel, beam forming is used, and the directivity of the sound collector is directed to each wheel of the host vehicle by beam forming so that the traveling sound of each wheel is transmitted. get. Since beam forming is possible with at least two sound collectors, the cost can be reduced as compared with a case where a sound collector is provided for each wheel to obtain traveling sound for each wheel. In addition, since beam forming can be performed at a position where a sound collector used for recognizing another vehicle approaching or the like is used, it can be shared with the sound collectors of such other systems, and the cost can be reduced.

The beam forming will be described with reference to FIG. In this example, beam forming may be performed by two sound collectors ML and MR arranged in the left-right direction at the front end portion of the vehicle, and each wheel position of the right front wheel W _fr and the left rear wheel W _rl of the own vehicle. (A grounding point between the tire and the ground) is a beam forming directivity target position. As shown in FIG. 10A, the traveling sound generated from the right front wheel W _fr first reaches the right sound collector MR, and reaches the left sound collector ML with a delay. The delay time d _fr to the left sound collector ML is constant time, the speed of sound so known, the sound collector ML, the delay time d _fr from each position and the grounding position of the right front wheel W _fr (position of the sound source) of the MR Can be requested. Further, as shown in FIG. 10B, the traveling sound generated from the left rear wheel W _rl reaches the left sound collector ML first, and reaches the right sound collector MR with a delay. The delay time d _rl to this right sound collector MR is constant time, similarly to the above, the sound collector ML, to determine the delay time d _rl from the ground position of each position and the left rear wheel W _rl of MR Can do. Detect by experiment. Here, as shown in FIG. 10 (c), the delay time d _fr min delays the sound data collected by the right sound collector MR, collected by the sound data and left sound collector ML is the delay By adding the sound data, the traveling sound data of the right front wheel W _fr can be acquired. The sound data collected by the left sound collector ML is delayed by a delay time d _rl , and the sound data and the right sound collection are delayed. The running sound data of the left rear wheel W _rl can be acquired by adding the sound data collected by the sound generator MR.

  The learning device 1C will be described. The learning device 1C is the same learning device as the learning device 1A according to the first embodiment or the learning device 1B according to the second embodiment. Therefore, detailed description is omitted. In the case of performing prior learning, an actual vehicle experiment is performed in which an experimental vehicle is run using an experimental vehicle that collects running sounds, as in the first embodiment or the second embodiment. Change the road surface condition or the state where the white line is stepped on every two or a combination of two or more wheels, run the experimental vehicle and perform the actual vehicle experiment, and each wheel as a running sound feature / road surface condition table You may generate | occur | produce the table for every combination of the above wheels.

  The driving support system 2C will be described. The driving support system 2C is mounted on the vehicle, determines the road surface state for each wheel of the host vehicle based on the traveling sound data collected during traveling of the vehicle, and performs driving support according to the determination result. The driving support system 2C includes a road surface state determination device 3C and a driving support device 4C.

  The road surface state determination device 3C will be described. The road surface condition determination device 3C collects the sound of the host vehicle with a plurality of sound collectors while the vehicle is traveling, acquires the traveling sound for each wheel of the host vehicle by beam forming using the plurality of sound data, and The feature amount of the traveling sound for each wheel is extracted, and the feature amount for each wheel is compared with the traveling sound feature amount / road surface state table to determine the road surface state for each wheel of the vehicle currently traveling. The road surface state determination device 3C includes a sound collection device 30C and an ECU 31C.

  The sound collector 30C has two or more sound collectors. For example, the two or more sound collectors are arranged side by side in the vehicle width direction (left-right direction) at the front end of the vehicle. The sound collection device 30B collects sound with each sound collector while the vehicle is traveling, and transmits the collected sound data to the ECU 31C. In the third embodiment, the sound collector of the sound collecting device 30C corresponds to the sound collecting means described in the claims.

  The ECU 31C is an electronic control unit including a CPU, a ROM, a RAM, and the like, and comprehensively controls the road surface state determination device 3C. The ECU 31C includes a traveling sound feature amount extraction unit 32C, a road surface state determination unit 33C, and a traveling sound feature amount / road surface state table storage unit 34C. The ECU 31C receives sound data (electrical signals) collected by each sound collector from the sound collector 30C. The ECU 31C performs the same process as the ECU 31A according to the first embodiment as a pre-process for the sound data. In the third embodiment, the traveling sound feature quantity extraction unit 32C corresponds to the own vehicle beam forming means and the feature quantity extraction means described in the claims, and the road surface condition determination unit 33C is included in the claims. This corresponds to the road surface state determination means described.

  The traveling sound feature / road surface state table storage unit 34C is configured in a predetermined area of the storage device of the ECU 31C, and is stored in the traveling sound feature / road surface state table storage unit 34C of the learning device 1C. A state table is stored.

  Each time the running sound feature quantity extraction unit 32C receives sound data of each sound collector of the sound collector 30A, the sound data of each sound collector after preprocessing is used to perform wheel forming of the vehicle by beam forming. The running sound data is calculated every time. Further, the running sound feature amount extraction unit 32C extracts feature amounts for each wheel of the host vehicle using running sound data by beam forming.

  In the road surface state determination unit 33C, for each wheel of the host vehicle, the traveling sound feature amount for each road surface stored in the traveling sound feature amount / road surface state table storage unit 34C or the traveling sound feature amount on the road surface and on the white line The road surface state having the feature quantity closest to the feature quantity extracted by the running sound feature quantity extraction unit 32C (that is, the vehicle is currently running) is compared with the feature quantity extracted by the running sound feature quantity extraction unit 32C. Road surface condition or on road surface or white line).

  In ECU31C, road surface state information is produced | generated based on the determination result in the road surface state determination part 33C, and road surface state information is transmitted to the driving assistance apparatus 4C. The road surface state information includes, for example, information on a specific road surface state for each wheel and information on whether or not a white line for each wheel is not stepped on.

  The driving support device 4C includes various sensors, ECUs, and the like, and is a device that supports various driving operations for the driver. In particular, in the driving assistance device 4C, when road surface state information is received from the road surface state determination device 3C, driving assistance is performed according to the road surface state of each wheel. In the third embodiment, the driving support device 4C corresponds to the driving support means described in the claims.

For example, when information on whether or not a white line for each wheel is not stepped on is obtained as road surface state information, the present invention can be applied to a lane departure warning device or a lane departure prevention device. As shown in FIG. 11A, first, when only the right front wheel W _{fr of} the own vehicle steps on the white line L, only the traveling sound of the right front wheel W _fr of the own vehicle changes, and the right front wheel W _{fr of the} own vehicle changes. Only when it is determined that the vehicle is stepping on the white line L, and when the right rear wheel _{Wrr of the host} vehicle steps on the white line L as shown in FIG. _It is determined that the traveling sound of _rr changes and only the right rear wheel _{Wrr of the host} vehicle is stepping on the white line L. In this case, since the vehicle has completely deviated from the lane, strong alerting or vehicle control is performed. As a strong alert, for example, a reaction force is applied to a steering operation in a direction deviating from the lane. Further, as vehicle control, there are, for example, deceleration control and steering control in a direction opposite to the direction of departure from the lane. When performing vehicle control, it may be better to change the control according to the situation around the host vehicle (the presence or absence of other vehicles, etc.).

Further, as shown in FIG. 12A, when only the right front wheel W _{fr of} the own vehicle steps on the white line L, only the traveling sound of the right front wheel W _fr of the own vehicle changes, and the right front wheel of the own vehicle changes. It is determined that only W _fr is stepping on the white line L, and after a predetermined time, the right rear wheel W _rr has stepped on the white line L in addition to the right front wheel W _fr of the host vehicle as shown in FIG. In this case, it is determined that the traveling sound of the right front wheel W _fr and the right rear wheel W _rr of the _host vehicle changes and the right front wheel W _fr and the right rear wheel W _{rr of the host} vehicle are stepping on the white line L. In this case, the host vehicle has not deviated from the lane, but is about to deviate, so weak attention is given. As weak alerting, for example, alerting is performed by voice output or display display. Such alerting and vehicle control may be determined in advance according to the state of the stepping on the white line for each wheel and the combination of the stepping wheels.

Further, the present invention can be applied to driving support when information as to whether or not the road surface state information is in a portion of a low friction coefficient such as a puddle or an ice burn is obtained for each wheel. As shown in FIG. 13 (a), all the wheels are normally traveling on a dry road surface DR, and after a predetermined time, only the front wheels entered the part WR having a low friction coefficient as shown in FIG. 13 (b). when the front wheels of each wheel _{W fl, W fr} of the running sound is changed, the front wheels _W fl of the _{vehicle, W fr} is determined to have entered the position WR of the low coefficient of friction. In this case, since it is dangerous to steer large, control to give a reaction force to the steering operation, control to change the gear ratio, alerting, etc. are performed. Further, when all the wheels enter the portion WR having a low friction coefficient after a predetermined time, as shown in FIG. 13C, the running sounds of all the wheels W _fl , W _fr , W _rl , W _rr change, It is determined that all the wheels W _fl , W _fr , W _rl , W _{rr of the host} vehicle have entered the part WR having a low friction coefficient. Furthermore, when all the wheels have come out of the portion WR having the low friction coefficient after the predetermined time, as shown in FIG. 13 (d), only the running sound of the rear wheels W _rl and W _{rr remains} changed. Then, it is determined that the rear wheels W _rl and W _{rr of the host} vehicle remain at the portion WR having the low friction coefficient. Further, when only the left wheel as shown in FIG. 13 (e) enters the portion WR a low coefficient of friction, each wheel of the left wheel _{W fl, W} running sound of _rl is changed, the left wheel of the vehicle _{W fl, W rl} Is determined to have entered the portion WR having a low friction coefficient. In each of these cases, appropriate vehicle control and alerting are performed according to the situation. About such alerting and vehicle control, it is good to determine beforehand according to the road surface state for every wheel and every combination of wheels.

  The operation of the third embodiment will be described with reference to FIGS. 1, 2, 4, 11, and 12. In particular, the operation in the driving support system 2C will be described along the flowchart of FIG. FIG. 14 is a flowchart showing a flow of processing in the driving support system according to the third embodiment. Here, it is assumed that the road surface state is a specific road surface state such as a dry state, it is determined whether or not a white line is stepped on for each wheel, and driving assistance related to lane departure is provided as driving assistance.

  Since the operation of the learning device 1C is the same as that of the learning device 1B according to the second embodiment, the description thereof is omitted. Each vehicle is equipped with a driving support system 2C, and the driving sound feature / road surface state table storage unit 34C of the road surface determination device 3C of the driving support system 2C stores the driving sound feature / road surface state table of the learning device 1C. The traveling sound feature amount / road surface state table of the unit 14C is stored.

  In each vehicle, two or more sound collectors of the sound collecting device 30C of the road surface condition determining device 3C collect sounds around the outside of the vehicle, respectively, and convert the collected sounds into electric signals to the ECU 31C. It is transmitting (S30). The ECU 31C receives the sound data of the two or more sound collectors and performs preprocessing on each sound data.

  Every time the ECU 31C receives sound data of each sound collector from the sound collecting device 30C, it calculates the running sound for each wheel of the host vehicle by beam forming using the sound data of each sound collector after preprocessing. (S31). Then, the ECU 31C extracts a feature amount from the running sound acquired by beam forming for each wheel (S32). Then, in the ECU 31C, the feature when the extracted feature amount does not step on the white line based on the travel sound feature amount / road surface state table stored in the travel sound feature amount / road surface state table storage unit 34C for each wheel. It is determined whether the amount is a feature amount or a feature amount when the white line is stepped on (S33). The ECU 31C generates road surface state information based on the determination result, and transmits the road surface state information to the driving support device 4C.

  In the driving assistance device 4C, when the road surface state information is received from the road surface state determination device 3B, it is determined whether or not all the wheels are on the road surface based on the road surface state (S34). When it is determined in S34 that all the wheels are on the road surface, the driving assistance device 4B determines that there is no lane departure (S35).

  When it is determined in S34 that at least one wheel is on the white line, the ECU 31C determines whether or not a predetermined time has elapsed after the determination (S36). If it is determined in S36 that the predetermined time has elapsed, the ECU 31C determines the wheel determined to be on the white line based on the traveling sound feature / road surface state table stored in the traveling sound feature / road surface state table storage unit 34C. It is determined whether the feature amount extracted from the traveling sound of the left and right wheels is a feature amount when the white line is not stepped or a feature amount when the white line is stepped (S37). The ECU 31C generates road surface state information based on the determination result, and transmits the road surface state information to the driving support device 4C.

  In the driving assistance device 4C, it is determined whether the two wheels on the left and right sides are on the white line or the other one is on the white line (S38). If it is determined in S38 that the two wheels on the same left and right sides are on the white line, the driving support device 4C determines that the vehicle has not deviated from the lane but is about to deviate (S39), and performs a weak alert (S40). On the other hand, when only one other wheel on the same left and right side is determined to be on the white line, the driving support device 4C determines that the vehicle is departing from the lane (S41), and performs strong alerting and vehicle control (S42).

  According to the third embodiment, by using the beam forming to obtain the running sound for each wheel and determining the road surface state for each wheel, more detailed alerting according to the road surface state of each wheel Driving support such as vehicle control can be performed. As a result, a more appropriate and more accurate driving support system can be provided. Further, according to the third embodiment, by providing directivity to each wheel by beam forming, it is not necessary to provide a sound collector for each wheel and the sound collector can be shared with other systems. Cost can be reduced.

  Here, a method for arranging two or more sound collectors used in beamforming will be described with reference to FIGS. 15 and 16. FIG. 15 is an explanatory diagram of directivity to both sides by beam forming. FIG. 16 is an example of the arrangement of sound collectors for performing beamforming.

When directivity is directed to the position of an arbitrary wheel by beam forming, directivity also occurs at a position opposite to the position of the wheel. For example, as shown in FIG. 15, two sound collectors ML, MR the case of arranging the left and right of the central portion of the longitudinal direction of the vehicle, the delay time d _rr and the same time in the case of beamforming delay time d _fr and the right rear wheel W _rr in case of beamforming to the right front wheel W _fr For this reason, it is impossible to distinguish the traveling sound of the right front wheel W _{fr and} the traveling sound of the right rear wheel W _rr by beam forming (the same traveling sound is obtained). Therefore, when the position of any wheel is used as the beamforming directivity position, the opposite directivity position should not be the position of other wheels or other sound sources (engine, engine fan, etc.). It is necessary to arrange two or more sound collectors.

Therefore, as an arrangement of two or more sound collectors used in beam forming, when the position of each wheel is set as the directivity position of beam forming, the directivity position on the opposite side is inside the own vehicle and the sound source is It is arranged so that there is no position (or a position where the generation of sound is very small), or a position where the directivity position on the opposite side is outside the vehicle (outside the body) and is not separated from the vehicle ( Position it so that it is close to the body. In the position close to the body of the host vehicle, it is usually very unlikely that an object serving as another sound source such as another vehicle exists. Because there is no sound source, it is not affected by noise. For example, FIG. 16 shows three sound collectors ML, MC. MRs are arranged side by side in the left-right direction at the front end of the host vehicle, and when beam forming is performed on the right front wheel _Wfr , the opposite directivity position is outside the host vehicle and close to the host vehicle. Similarly, when beam forming is performed on the other wheels, the directivity position on the opposite side is outside the host vehicle and close to the host vehicle.

  It is possible to use two or more sound collectors for beam forming. However, if beam forming is performed using two sound collectors, the position on the surface has directivity. In addition, when beam forming is performed using three sound collectors, directivity can be provided on the line. Further, if beam forming is performed using four sound collectors, directivity can be given to one point in three dimensions. Even when beam forming is performed using two sound collectors and the directivity is given to the position on the surface, there is a high possibility that the sound source is only the running sound of the wheel on the surface ( (Because the directivity is good in the sky and underground except for the wheels), there is actually no problem.

  A learning device 1D and a driving support system 2D according to a fourth embodiment will be described with reference to FIGS. 1, 2, 4, 17, and 18. FIG. FIG. 17 is an example of beam forming to another vehicle on a road surface with a portion having a low friction coefficient. FIG. 17A shows a case where the traveling sound of the other vehicle is not obtained by beam forming, and FIG. This is a case where the running sound of the other vehicle that does not enter the location of the friction coefficient is obtained by beam forming, and (c) shows that the running sound of the other vehicle entered the location of the low friction coefficient is obtained by beam forming. Is the case. FIG. 18 is an example of beam forming to another vehicle on a curved road.

  Compared with the third embodiment, the fourth embodiment uses beamforming to acquire the running sound of other vehicles in front of the host vehicle, extract the feature amount, and determine the road surface state of the preceding vehicle. However, driving assistance is different according to the road surface condition of the preceding vehicle. Since the road surface condition is obtained in the vehicle ahead, appropriate driving assistance can be performed before the road surface condition of the host vehicle changes.

  In the fourth embodiment, in order to acquire the traveling sound of the preceding vehicle, beam forming is used, the directivity of the sound collector is directed to a predetermined position in front by beam forming, and the traveling sound of the preceding vehicle is acquired. . Although beam forming is possible with at least two sound collectors, in the fourth embodiment, three sound collectors are used in order to give directivity to a specific position in front.

  The learning device 1D will be described. The learning device 1D is a learning device similar to the learning device 1C according to the third embodiment. Therefore, detailed description is omitted.

  The driving support system 2D will be described. The driving support system 2D is mounted on the vehicle, determines the road surface state of the preceding vehicle based on the traveling sound data collected during traveling of the vehicle, and performs driving support according to the determination result. The driving support system 2D includes a road surface state determination device 3D and a driving support device 4D.

  The road surface state determination device 3D will be described. The road surface condition determination device 3D collects sound with three sound collectors while the vehicle is traveling, acquires the traveling sound of the preceding vehicle by beam forming using the three sound data, and travels the preceding vehicle. The feature amount of the sound is extracted, the feature amount of the preceding vehicle is compared with the traveling sound feature amount / road surface state table, and the road surface state of the preceding vehicle currently traveling is determined. The road surface state determination device 3D includes a sound collection device 30D and an ECU 31D.

  The sound collector 30D includes three (four or more) sound collectors. For example, the three sound collectors are arranged side by side in the vehicle width direction (left-right direction) at the front end of the vehicle. The sound collection device 30D collects sound with each sound collector while the vehicle is traveling, and transmits the collected sound data to the ECU 31D. In the fourth embodiment, the sound collector of the sound collecting device 30D corresponds to the sound collecting means described in the claims.

  The ECU 31D is an electronic control unit including a CPU, a ROM, a RAM, and the like, and comprehensively controls the road surface state determination device 3D. The ECU 31D includes a traveling sound feature amount extraction unit 32D, a road surface state determination unit 33D, and a traveling sound feature amount / road surface state table storage unit 34D. The ECU 31D receives sound data (electrical signals) collected by each sound collector from the sound collecting device 30D. The ECU 31D performs the same process as the ECU 31C according to the first embodiment as a pre-process for the sound data. In the fourth embodiment, the running sound feature quantity extraction unit 32D corresponds to the other vehicle beam forming means and the feature quantity extraction means described in the claims, and the road surface condition determination unit 33D is in the claims. This corresponds to the road surface state determination means described.

  The traveling sound feature / road surface state table storage unit 34D is configured in a predetermined area of the storage device of the ECU 31D, and is stored in the traveling sound feature / road surface state table storage unit 34C of the learning device 1D. A state table is stored.

  Each time the running sound feature extraction unit 32D receives the sound data of each sound collector of the sound collector 30D, the sound data of each sound collector after the pre-processing is used to perform a predetermined predetermined front by beam forming. The position sound (running sound of the preceding vehicle when another vehicle exists around the predetermined position) is calculated. To determine where the beam is to be formed, for example, set the distance in front of beam forming based on the standard inter-vehicle distance according to the speed (the current vehicle speed of the host vehicle, the speed limit of the road that is running, etc.). In the case of a curved road, the beam forming direction is set according to the steering operation angle or the road curvature. Further, the running sound feature amount extraction unit 32D extracts feature amounts using sound data obtained by beam forming for a predetermined position in front of the running sound feature amount extraction unit 32D.

  For example, as shown in FIG. 17, in the case of a straight road, beam forming is performed to a position BP that is a predetermined distance ahead in the straight forward direction from the host vehicle MV. Even when beam forming is performed at one point BP, the sound does not actually become one point, but a sound within a range BA that is spread to some extent can be obtained by beam forming. As shown in FIG. 17A, when there is no other vehicle in a predetermined range BA ahead, the beam-formed sound does not include the traveling sound of the other vehicle. As shown in FIG. 17 (b), when the other vehicle FV is present in the predetermined range BA ahead, the beam-formed sound includes the traveling sound of the other vehicle FV. In this case, on the dry road surface DR Can be obtained. As shown in FIG. 17C, when the vehicle advances from the situation of FIG. 17B and the other vehicle FV enters the road surface WR having a low friction coefficient such as a wet state, the vehicle travels on the road surface WR having a low friction coefficient. Sound is obtained. As shown in FIG. 18, when the front is a curved road, beam forming is performed to a position BP that is a predetermined distance ahead in a direction inclined by a predetermined angle to the front right according to the steering operation angle or the like of the host vehicle MV. .

  In the road surface state determination unit 33D, a traveling sound feature amount for each road surface state stored in the traveling sound feature amount / road surface state table storage unit 34D, a traveling sound feature amount on the road surface and on the white line, and a traveling sound feature amount extraction unit The road surface state having the feature quantity closest to the feature quantity extracted by the running sound feature quantity extraction unit 32D is compared with the feature quantity extracted by 32D (that is, the road surface condition where the preceding vehicle is currently running or on the road surface or on the white line) )). In addition, when the intensity of each frequency is not obtained more than a predetermined value as the feature amount, it is determined that there is no other vehicle ahead, and the road surface state is not determined.

  In ECU31D, road surface state information is produced | generated based on the determination result in road surface state determination part 33D, and road surface state information is transmitted to driving assistance apparatus 4D. The road surface state information includes, for example, the presence / absence of a forward vehicle, information on the road surface state at the position of the forward vehicle when there is a forward vehicle, and information on whether or not the forward vehicle is stepping on a white line.

  The driving support device 4D includes various sensors, ECUs, and the like, and is a device that supports various driving operations for the driver. In particular, in the driving assistance device 4D, when road surface state information is received from the road surface state determination device 3D, driving assistance is performed according to the road surface state of the preceding vehicle. In the fourth embodiment, the driving support device 4D corresponds to the driving support means described in the claims.

  For example, as shown in FIG. 17 (a), when information indicating that no other vehicle exists ahead is obtained as road surface state information (when the traveling sound of the other vehicle cannot be obtained in front of the host vehicle MV), the front Driving assistance according to the road surface condition of the vehicle is not performed. As shown in FIG. 17 (b), when another vehicle is present ahead as the road surface state information and information indicating a dry state (high friction coefficient) is obtained as the road surface state (in front of the host vehicle MV, the other vehicle's When the driving sound in the dry state is obtained as the driving sound), normal driving is possible, so driving assistance is not performed. As shown in FIG. 17C, when another vehicle is present ahead as the road surface state information and information such as a wet state (low friction coefficient) is obtained as the road surface state (the other vehicle ahead of the host vehicle MV). Depending on the road condition of the vehicle in front, because the vehicle may reach the road surface after a few seconds and the vehicle behavior such as slip may change. In the meantime, alerting and vehicle control are performed until the road surface condition is reached. The number of seconds to reach can be calculated from the vehicle speed of the host vehicle. Moreover, when the information that the other vehicle ahead is stepping on the white line is obtained as the road surface state information, it can be determined that the other vehicle ahead has changed the lane.

  The operation of the fourth embodiment will be described with reference to FIGS. 1, 2, 4, 17, and 18. In particular, the operation in the driving support system 2D will be described along the flowchart of FIG. FIG. 19 is a flowchart illustrating a process flow in the driving support system according to the fourth embodiment. Here, it is assumed that a road surface state such as a dry state or a wet state is determined as the road surface state of the preceding vehicle.

  Since the operation of the learning device 1D is the same as that of the learning device 1A according to the first embodiment, the description thereof is omitted. Each vehicle is equipped with a driving support system 2D, and the driving sound feature / road surface state table storage unit 34D of the road surface state determination device 3D of the driving support system 2D stores the driving sound feature / road surface state table of the learning device 1D. The running sound feature amount / road surface state table of the unit 14D is stored.

  In each vehicle, the three sound collectors of the sound collecting device 30D of the road surface condition determining device 3D collect sounds around the outside of the vehicle, convert the collected sounds into electric signals, and transmit them to the ECU 31D. (S50). The ECU 31D receives the sound data of the three sound collectors and performs preprocessing on each sound data.

  Every time the ECU 31D receives the sound data of each sound collector from the sound collector 30D, the sound of the predetermined position ahead is calculated by beam forming using the sound data of each sound collector after the preprocessing (S51). ). At this time, if a vehicle is present in the vicinity of a predetermined position ahead, the traveling sound of the preceding vehicle is obtained (S51). Then, the ECU 31D extracts a feature amount from the sound acquired by beam forming (S52). At this time, if the traveling sound of the preceding vehicle is obtained, the characteristic amount of the traveling sound of the preceding vehicle is obtained (S52). Then, in the ECU 31D, when the feature amount of the traveling sound of the vehicle ahead is obtained, the extracted feature is based on the traveling sound feature amount / road surface state table stored in the traveling sound feature amount / road surface state table storage unit 34D. It is determined which amount the road surface condition is closest to (S53). The ECU 31D generates road surface state information based on the determination result, and transmits the road surface state information to the driving support device 4D.

  In the driving assistance device 4D, when road surface state information is received from the road surface state determination device 3B, driving assistance is performed based on the road surface state of the preceding vehicle when there is a preceding vehicle (S54).

  According to the fourth embodiment, the front low friction coefficient is obtained based on the road surface condition of the forward vehicle by acquiring the running sound of the forward vehicle using beam forming and determining the road surface state of the forward vehicle. Before reaching the road surface (the road surface on which the vehicle behavior changes), driving assistance such as alerting and vehicle control can be performed. As a result, a more appropriate and more accurate driving support system can be provided. Further, according to the fourth embodiment, the traveling sound of the preceding vehicle and the traveling sound of the host vehicle can be reliably separated by using beam forming. Further, according to the fourth embodiment, by setting the beam forming position in consideration of the speed, the steering operation angle, etc., it is possible to reliably obtain the traveling sound of the front vehicle necessary for the host vehicle. .

  With reference to FIGS. 4 and 20, a driving support system 2E according to a fifth embodiment will be described. FIG. 20 is a configuration diagram of a driving support system according to the fifth embodiment.

  Unlike the first to fourth embodiments, the fifth embodiment does not require a learning device that performs prior learning. In the fifth embodiment, rather than determining a specific road surface state, whether or not the traveling sound is different from that when the host vehicle has traveled in the past at the same place where the host vehicle has traveled in the past (and thus the road surface state is different). Or not). In the case of the first to fourth embodiments, it is necessary to learn and store the feature value of the running sound for each road surface state, and to identify a specific road surface state. If there is a running sound (feature value) generated during normal driving, it will affect the behavior of the vehicle if a different running sound (feature value) is emitted at the same location. It can be determined that the road surface condition has changed.

  The driving support system 2E will be described. The driving support system 2E determines whether or not the road surface condition of the same place that has traveled in the past has changed based on the traveling sound data that is mounted on the vehicle and collected during traveling of the host vehicle. Provide driving assistance accordingly. The driving support system 2E includes a road surface state determination device 3E and a driving support device 4E.

  The road surface state determination device 3E will be described. The road surface condition determination device 3E collects the traveling sound of the own vehicle while traveling the vehicle, acquires the current position of the own vehicle, extracts the feature amount of the traveling sound of the own vehicle, and travels the current position in the past. If it is, the feature amount is compared with the past feature amount, and it is determined whether or not the road surface condition has changed from the past travel. The road surface state determination device 3E includes a sound collection device 30E, a position information acquisition device 35E, and an ECU 31E.

  The sound collector 30E has the same configuration as that of the sound collector 30A according to the first embodiment. The sound collection device 30E collects sound with the sound collector while the vehicle is traveling, and transmits the collected sound data to the ECU 31E. In the fifth embodiment, the sound collector of the sound collecting device 30E corresponds to the sound collecting means described in the claims.

  The position information acquisition device 35E is a device that detects the current position of the host vehicle. As the position information acquisition device 35E, for example, a device that calculates the current position using GPS [Global Positioning System]. When the host vehicle is equipped with a navigation system, the current position information is acquired from the navigation system. Device. The position information acquisition device 35E acquires the current position of the host vehicle at regular intervals, and transmits the current position information to the ECU 31E. In the fifth embodiment, the position information acquisition device 35E corresponds to the position detection means described in the claims.

  The ECU 31E is an electronic control unit including a CPU, a ROM, a RAM, and the like, and comprehensively controls the road surface state determination device 3E. The ECU 31E includes a traveling sound feature amount extraction unit 32E, a road surface state change determination unit 36E, and a traveling sound feature amount / position database 37E. The ECU 31E receives sound data (electrical signal) collected by the sound collector from the sound collector 30E and also receives current position information from the position information acquisition device 35E. The ECU 31E performs the same process as the ECU 31A according to the first embodiment as a pre-process for the sound data. In the fifth embodiment, the running sound feature amount extraction unit 32E corresponds to the feature amount extraction unit described in the claims, and the road surface state change determination unit 36E determines the road surface state determination described in the claims. The running sound feature / position database 37E corresponds to the storage means described in the claims.

  The running sound feature / position database 37E is a database configured in a predetermined area of the storage device of the ECU 31E. Each time the running sound feature quantity extraction unit 32E extracts a running sound feature quantity in the running sound feature quantity / position database 37E, the feature quantity and the current position information of the position information acquisition device 35E are stored in association with each other. . Furthermore, when information related to the change in the road surface condition (for example, ON / OFF of the wiper, date / time, temperature, weather) can be acquired, the information related to the change in the road surface condition may be associated and stored. Furthermore, when vehicle behavior information can be acquired, vehicle behavior information may also be stored in association with each other.

  The traveling sound feature amount extraction unit 32E performs the same processing as the traveling sound feature amount extraction unit 32A according to the first embodiment. Each time the running sound feature quantity extraction unit 32E receives the sound data of the sound collector of the sound collecting device 30E, the running sound feature quantity extraction unit 32E extracts the running sound feature quantity using the pre-processed sound data.

  The road surface state change determination unit 36E refers to the travel sound feature / position database 37E to determine whether or not the current position is a position stored in the travel sound feature / position database 37E. When the current position is a position stored in the traveling sound feature value / position database 37E, the road surface state change determination unit 36E extracts the feature value and the traveling sound feature value / current position extracted by the traveling sound feature amount extraction unit 32E. The feature quantity at the same position stored in the position database 37E is compared, and it is determined whether or not the two feature quantities are different. The road surface state change determination unit 36E determines that the road surface state has changed normally (in particular, changes to a low friction coefficient state such as wetness) at the current position when the feature amount is different, and the feature amount is If they are similar, it is determined that the road surface condition is the same as usual at the current position.

  In the ECU 31E, road surface state information is generated based on the determination result in the road surface state change determination unit 36E, and the road surface state information is transmitted to the driving support device 4E. The road surface state information is, for example, information on whether or not the road surface state has changed normally at the current position. If information related to changes in road surface conditions is also stored in the running sound feature / position database 37E, such information is also added. If the vehicle behavior information is also stored in the traveling sound feature / position database 37E, the vehicle behavior information is also added.

  The driving support device 4E includes various sensors, ECUs, and the like, and is a device that supports various driving operations for the driver. In particular, when the driving support device 4E receives road surface state information from the road surface state determination device 3E, the driving support device 4E performs driving support related to the road surface state. For example, if information indicating that the road surface condition has changed normally at the current position is obtained as road surface condition information, the specific road surface condition has not been obtained, so the road surface condition will change from when the vehicle is running normally. Because it is, it will give a non-specific alert such as driving with caution. When the road surface state information includes information related to the road surface state change, the state of the road surface state change can be judged in more detail (for example, it is raining or snowing and the road surface state is changing). Execute more specific alerts. Further, when the vehicle behavior information is included in the road surface state information, strong alerting or vehicle control is performed if the vehicle behavior is the same as the feature amount when the vehicle behavior is unstable. In the fifth embodiment, the driving support device 4E corresponds to the driving support means described in the claims.

  The operation of the fifth embodiment will be described with reference to FIG. In particular, the operation in the driving support system 2E will be described along the flowchart of FIG. FIG. 21 is a flowchart showing a process flow in the driving support system according to the fifth embodiment.

  In each vehicle, the position information acquisition device 35E acquires the current position of the host vehicle and transmits the current position information to the ECU 31E (S60). Further, the sound collector of the sound collecting device 30E of the road surface condition determining device 3E collects sounds outside the vehicle, converts the collected sounds into electric signals, and transmits them to the ECU 31E (S61). The ECU 31E receives the sound data from the sound collector and performs preprocessing on the sound data.

  Each time the ECU 31E receives the sound data of the sound collector from the sound collecting device 30E, the ECU 31E extracts the feature value of the running sound from the sound data of the sound collector after the preprocessing (S62). At this time, the current position and the feature value are stored in the running sound feature value / position database 37E in association with each other.

  In the ECU 31E, with reference to the running sound feature / position database 37E, when information on the same position as the current position is stored, the past feature quantity at the same position is compared with the feature quantity extracted this time, It is determined whether or not the feature amounts are different (and whether or not the road surface state has changed) (S63). Then, the ECU 31E generates road surface state information based on the determination result, and transmits the road surface state information to the driving support device 4E.

  In the driving assistance device 4E, when the road surface state information is received from the road surface state determination device 3E, when the road surface state has changed (especially, when it can be estimated that the road surface state has changed to a low friction coefficient), driving assistance is provided. (S64).

  According to the fifth embodiment, the feature amount of the traveling sound and the current position are stored in association with each other, and the feature amount of the current traveling sound at the same position is compared with the past feature amount. A change in road surface state can be determined with a simpler configuration without prior learning or identification of specific road surface information.

  As mentioned above, although embodiment which concerns on this invention was described, this invention is implemented in various forms, without being limited to the said embodiment.

  For example, in the present embodiment, the present invention is applied to a driving support system including a road surface state determination device and a driving support device, but other configurations may be used. For example, it may be incorporated in the driving support device as a road surface state determination function, or the road surface state determination device may have an alarm function or the like.

  In the present embodiment, the learning device that performs learning in advance and the driving support system that actually determines the road surface state by using the learning table and performs driving support are configured separately to perform learning in advance. , May be configured integrally. In particular, in the case of an integrated device, a travel sound feature / road surface state table is generated offline based on sound data collected during vehicle travel, and the road surface state is determined in real time using the table.

  In this embodiment, the sound collecting device collects the traveling sound, extracts the feature amount from the traveling sound, and determines the road surface state based on the feature amount. However, the sound collecting device collects the traveling sound. The configuration may be such that the road surface condition is directly determined from the running sound without collecting the sound and extracting the feature amount. In the present embodiment, the intensity for each frequency is applied as the feature quantity, but other feature quantities may be used.

  In this embodiment, the traveling sound collected by beam forming is given directivity, and the traveling sound for each wheel of the host vehicle is acquired or the traveling sound of the preceding vehicle is acquired. You may give it directivity.

  In the third embodiment, beam forming is used to determine the road surface state of each wheel of the host vehicle and driving assistance is provided. In the fourth embodiment, beam forming is used to determine the vehicle ahead. The road surface condition is determined and the driving assistance is performed. However, the third embodiment and the fourth embodiment are combined, and the beam surface is used for each wheel of the host vehicle and the vehicle in front of the vehicle using beam forming. It is good also as a structure which determines a road surface state and performs driving assistance.

  1A, 1B, 1C, 1D ... Learning device, 2A, 2B, 2C, 2D, 2E ... Driving support system, 3A, 3B, 3C, 3D, 3E ... Road surface condition judging device, 4A, 4B, 4C, 4D, 4E ... Driving support device, 10A, 10B, 10C, 10D, 30A, 30B, 30C, 30D, 30E ... sound collector, 11A, 11B, 11C, 11D ... computer, 12A, 12B, 12C, 12D, 32A, 32B, 32C, 32D, 32E ... running sound feature quantity extraction unit, 13A, 13B, 13C, 13D ... learning unit, 14A, 14B, 14C, 14D, 34A, 34B, 34C, 34D ... running sound feature quantity / road surface state table storage unit, 31A , 31B, 31C, 31D, 31E ... ECU, 33A, 33B, 33C, 33D ... road surface state determination unit, 35E ... position information acquisition means, 36E ... Surface state change determination unit, 37E ... traveling sound feature value / location database.

Claims (18)

Sound collecting means for collecting vehicle running sound;
Road surface state determining means for determining a road surface state based on the running sound collected by the sound collecting means;
Equipped with a,
Give directionality to the means to collect the running sound, obtain the running sound for each wheel of the vehicle,
The road surface state determining means determines the road surface state for each wheel of the host vehicle,
Performs beamforming using the running noise was collected by a plurality of said sound collecting means, road-vehicle, characterized in Rukoto includes a vehicle beamforming means for acquiring running sound of each wheel of the vehicle, respectively apparatus. Sound collecting means for collecting vehicle running sound;
Road surface state determining means for determining a road surface state based on the running sound collected by the sound collecting means;
Equipped with a,
Give directionality to the means to collect the running sound, obtain the running sound of other vehicles around the host vehicle,
The road surface condition judging means, the vehicle state road device other vehicle is characterized that you determine the state of the road surface in contact with the ground. 3. The vehicle road surface according to claim 2 , further comprising: other vehicle beam forming means for performing beam forming using the traveling sound collected by each of the plurality of sound collecting means and acquiring the traveling sound of the other vehicle. Judgment device. The plurality of sound collecting means are arranged or beam forming so that the directivity position on the opposite side to the position of the directivity target by beam forming is inside the own vehicle and there is no sound source in the own vehicle. the vehicle according to claim 1 or claim 3, characterized in that the position of directional target disposed such directional position opposite to a position close to it and the vehicle outside of the vehicle Road surface determination device. 5. The vehicle road surface determination device according to claim 4 , wherein the number of the sound collecting means is three or more. Comprising a feature quantity extraction means for extracting a feature quantity from the running sound collected by the sound collection means,
The vehicle surface condition according to any one of claims 1 to 5, wherein the road surface state determination unit determines a road surface state based on a feature amount of a running sound extracted by the feature amount extraction unit. Road surface determination device. A learning means for learning the feature amount of the running sound for each road surface state,
The vehicle road surface determination device according to claim 6 , wherein the road surface state determination unit determines a road surface state based on a feature amount for each road surface state learned by the learning unit. The vehicular road surface determination device according to any one of claims 1 to 7 , wherein the road surface state determination unit determines whether or not the road surface state has a low friction coefficient. The vehicle road surface determination device according to any one of claims 1 to 8 , wherein the road surface state determination unit determines whether or not a wheel of the host vehicle is on a lane marking of a lane. Position detecting means for detecting the position of the host vehicle;
Storage means for storing the position of the host vehicle detected by the position detection means and the feature quantity extracted by the feature quantity extraction means in association with each other;
With
The road surface condition judging means compares the feature quantity extracted this time by the feature quantity extraction means at an arbitrary position of the host vehicle with the feature quantity at the same position stored in the storage means, and based on the comparison result The vehicle road surface determination device according to claim 6 , wherein it is determined whether or not the road surface state has changed. Sound collecting means for collecting vehicle running sound;
Road surface state determining means for determining a road surface state based on the running sound collected by the sound collecting means;
Driving assistance means for providing driving assistance based on the road surface condition determined by the road surface condition determining means;
Equipped with a,
Give directionality to the means to collect the running sound, obtain the running sound of other vehicles around the host vehicle,
The road surface state determining means determines the state of the road surface on which another vehicle is grounded,
The driving support unit, the driving support apparatus which is characterized that you support operation in accordance with the road surface condition of the other vehicle is determined by the road surface condition judging means. Comprising a feature quantity extraction means for extracting a feature quantity from the running sound collected by the sound collection means,
The driving support device according to claim 11 , wherein the road surface state determination unit determines a road surface state based on a feature amount of the running sound extracted by the feature amount extraction unit. A learning means for learning the feature amount of the running sound for each road surface state,
The driving support device according to claim 12 , wherein the road surface state determination unit determines a road surface state based on a feature value for each road surface state learned by the learning unit. The road surface state determining means determines whether or not the road surface state has a low friction coefficient,
The driving support means can be of any claims 11 to claim 13, characterized in that the driving support is in accordance with the road surface condition of a low friction coefficient when it is determined that the road surface condition of a low friction coefficient at the road surface condition determination means The driving support device according to claim 1. Give the driving sound collection means directivity, acquire the driving sound for each wheel,
The road surface state determining means determines whether or not the road surface state of the low friction coefficient for each wheel of the host vehicle,
The driving support device according to claim 14 , wherein the driving support means supports driving according to a position of a wheel of the host vehicle determined by the road surface state determination unit as a road surface state having a low friction coefficient. The road surface state determination means determines whether the wheel of the host vehicle is on a lane marking,
The driving support means can be of any claims 11 to claim 15, characterized in that the control lane departure warning or lane departure prevention when the wheels of the vehicle is determined and the section line of the lane in the road surface condition determination means The driving support device according to claim 1. Give the driving sound collection means directivity, acquire the driving sound for each wheel,
The road surface state determining means determines whether or not the lane marking on each vehicle wheel,
The driving support according to claim 16 , wherein the driving support means performs a lane departure warning or a lane departure prevention control according to a position of a wheel of the host vehicle determined to be on a lane marking by the road surface state determination means. apparatus. Position detecting means for detecting the position of the host vehicle;
Storage means for storing the position of the host vehicle detected by the position detection means and the feature quantity extracted by the feature quantity extraction means in association with each other;
With
The road surface condition judging means compares the feature quantity extracted this time by the feature quantity extraction means at an arbitrary position of the host vehicle with the feature quantity at the same position stored in the storage means, and based on the comparison result To determine whether the road surface condition has changed,
The driving support device according to claim 12 , wherein the driving support means provides driving support when the road surface state determination unit determines that the road surface state has changed at the same position where the vehicle has traveled in the past.

Images