阅读说明：本技术 异常声音的产生部位确定方法、非临时性存储介质及车载装置 (Abnormal sound generation location specifying method, non-temporary storage medium, and in-vehicle device ) 是由 田端淳 奥田弘一 藤井广太 今村健 于 2021-01-28 设计创作，主要内容包括：本发明公开异常声音的产生部位确定方法、非临时性存储介质及车载装置。提供一种异常声音的产生部位确定方法,执行如下处理：将规定映射的映射数据存储于存储装置,该映射是将与在车辆中感知到的声音相应的声音变量以及与所述声音同步的所述车辆的驱动系统装置的状态变量作为输入并输出作为声音的主要原因的部位的映射；声音信号获取处理,获取感知声音的麦克风输出的声音信号；状态变量获取处理,获取所述驱动系统装置的状态变量；以及确定处理,将所述声音变量以及所述状态变量作为向所述映射的输入,确定与所述声音信号对应的声音的产生部位,并且本发明还提供非临时性存储介质和车载装置。(The invention discloses a method for determining a location where an abnormal sound is generated, a non-transitory storage medium, and an in-vehicle device. Provided is a method for specifying a generation site of an abnormal sound, which executes the following processing: storing, in a storage device, mapping data defining a mapping for outputting a part that is a factor of sound, with a sound variable corresponding to sound sensed in a vehicle and a state variable of a drive system device of the vehicle synchronized with the sound as inputs; sound signal acquisition processing for acquiring a sound signal output by a microphone for sensing sound; acquiring a state variable of the driving system device; and a determination process of determining a generation location of a sound corresponding to the sound signal using the sound variable and the state variable as inputs to the map, and a non-transitory storage medium and an in-vehicle apparatus are also provided.)

1. A method for specifying a generation site of an abnormal sound, comprising the steps of:

storing, in a storage device, map data defining a map for outputting a site that is a factor of a sound, the map having as input a sound variable that is a variable corresponding to the sound sensed in a vehicle and a state variable of a drive system device of the vehicle that is synchronized with the sound;

causing an execution device to execute sound signal acquisition processing of acquiring a sound signal output from a microphone that senses sound in the vehicle in a state where the mapping data is stored;

causing the execution device to execute a state variable acquisition process of acquiring a state variable of the drive system device in a state where the mapping data is stored; furthermore, it is possible to provide a liquid crystal display device,

causing the execution device to execute determination processing in which a generation site of a sound corresponding to a sound signal is determined using, as inputs to the map, the sound variable based on the sound signal acquired by the sound signal acquisition processing and the state variable synchronized with the sound signal.

2. The method of determining a generation site of an abnormal sound according to claim 1,

the sound signal acquisition processing is processing for acquiring the sound signal under a condition that the microphone is arranged at a predetermined position indicated in advance; furthermore, it is possible to provide a liquid crystal display device,

the predetermined portion is any portion of a head portion, an instrument panel, a rear seat center portion, and a center console of a driver's seat of the vehicle.

3. The method of determining a generation site of an abnormal sound according to claim 1 or 2,

the sound signal acquisition process includes a process of acquiring the sound signal when the vehicle is set to a predetermined arrangement state; furthermore, it is possible to provide a liquid crystal display device,

the predetermined arrangement state is a state in which the vehicle is surrounded at least with respect to a direction other than one of four directions of front and rear, left and right of the vehicle, and two directions of up and down of the vehicle.

4. The method of determining a generation site of an abnormal sound according to any one of claims 1 to 3, further comprising:

when the output of the map is such that the probability of two or more candidates among a plurality of candidates of a part that is a factor is higher than the probability of the other candidates by a predetermined value or more, the execution device is caused to execute an instruction process for instructing a predetermined driving operation of the vehicle, wherein the map is a map that outputs a variable indicating a relative probability of each of the plurality of candidates,

further, the determination processing includes the processing of: when the instruction processing is performed, the sound variable based on the sound signal acquired by the sound signal acquisition processing and the state variable acquired by the state variable acquisition processing when a driving operation in accordance with an instruction based on the instruction processing is performed are input to the map, and a generation location of a sound corresponding to the sound signal is specified.

5. The method of determining a location where an abnormal sound is generated according to any one of claims 1 to 4,

and executing sound variable generation processing for generating, as the sound variable, at least one of, using the sound signal as an input, intensities of frequency components corresponding to a predetermined frequency proportional to a rotational frequency of a rotary machine serving as a thrust force generation device of the vehicle and frequencies that are integer multiples of the predetermined frequency, which are acquired by the state variable acquisition processing, in the sound signal, a prominent frequency that is a frequency of a frequency band having an intensity greater than a frequency band adjacent to each of a low frequency side and a high frequency side in the sound signal, a prominent amount that is an amount by which the intensity of the prominent frequency is prominent with respect to the adjacent frequency band, and a duration during which the intensity of the sound signal is equal to or greater than a predetermined value.

6. The method of determining a location where an abnormal sound is generated according to any one of claims 1 to 5,

the state variable acquisition process includes a process of acquiring at least one of a torque of a thrust generation device of the vehicle and an amount of change in the torque per unit time as the state variable.

7. The method of determining a location where an abnormal sound is generated according to any one of claims 1 to 6,

the vehicle includes a transmission device that changes a ratio of a rotational speed of a rotary machine serving as a thrust force generation device to a rotational speed of a drive wheel; furthermore, it is possible to provide a liquid crystal display device,

the state variable acquisition process includes a process of acquiring a gear ratio of the transmission as the state variable.

8. The method of determining a location where an abnormal sound is generated according to any one of claims 1 to 7,

the state variable acquisition process includes a process of acquiring at least one of a vehicle speed and a rotation speed of a rotary machine as a thrust generation device of the vehicle as the state variable.

9. The method of determining a location where an abnormal sound is generated according to any one of claims 1 to 8,

the vehicle includes an internal combustion engine, a rotating electrical machine mechanically connectable to a crankshaft of the internal combustion engine, and a control device that performs an overlap process of overlapping a cancellation torque, which is a torque for reducing a torque ripple of the internal combustion engine, with a required torque for the rotating electrical machine; furthermore, it is possible to provide a liquid crystal display device,

the state variable acquisition process includes a process of acquiring a magnitude of the cancellation torque as the state variable.

10. The method of determining a location where an abnormal sound is generated according to any one of claims 1 to 8,

the vehicle includes an internal combustion engine, a first motor generator, a second motor generator, and a control device;

the second motor generator may be mechanically coupled to the internal combustion engine and the first motor generator via a power split device, and may be mechanically coupled to a drive wheel without via the power split device;

the control device executes processing of causing the first motor generator to generate a propulsion torque for reducing looseness of a gear train of the power split device at a time of a no-load operation of the internal combustion engine and the first motor generator; furthermore, it is possible to provide a liquid crystal display device,

the state variable acquiring process includes a process of acquiring a magnitude of the propulsion torque as the state variable.

11. The method of determining a location where an abnormal sound is generated according to any one of claims 1 to 10,

the execution devices include a first execution device provided in the vehicle, a second execution device not provided in the vehicle, and a third execution device;

the first execution device executes the state variable acquisition processing and vehicle-side transmission processing of transmitting the state variable acquired by the state variable acquisition processing;

the second execution device is a portable terminal of a user of the vehicle equipped with the microphone, and executes a state variable reception process of receiving the state variable transmitted by the vehicle-side transmission process, the sound signal acquisition process, and a terminal-side transmission process of transmitting the sound signal acquired by the sound signal acquisition process and the state variable received by the state variable reception process; furthermore, it is possible to provide a liquid crystal display device,

the third execution device executes analysis-side reception processing for receiving the sound signal and the state variable transmitted by the terminal-side transmission processing from each of portable terminals of users of the plurality of vehicles, and the determination processing.

12. The method of determining a generation site of an abnormal sound according to claim 11,

the second execution means executes notification acquisition processing for acquiring a notification to the effect of the abnormal sound when the person perceives the abnormal sound; furthermore, it is possible to provide a liquid crystal display device,

the audio signal to be transmitted by the terminal-side transmission process is the audio signal acquired by the audio signal acquisition process in a predetermined period determined according to the timing at which the notification is acquired by the notification acquisition process.

13. The method of determining a generation site of an abnormal sound according to claim 11 or 12,

the second execution means executes additional variable acquisition processing of acquiring an additional variable that is a variable indicating at least one of an open/close state of a window of the vehicle and a situation in which the vehicle is located when the abnormal sound is sensed;

the terminal side transmission processing includes processing of transmitting the additional variable in addition to the sound signal and the state variable;

the analysis side reception processing includes processing of receiving the additional variable;

the mapped input includes the additional variable in addition to the sound variable and the state variable; furthermore, it is possible to provide a liquid crystal display device,

the determination processing includes processing for determining a generation location of a sound corresponding to the sound signal, using the sound variable, the state variable, and the additional variable as inputs to the map.

14. A non-transitory storage medium storing commands executable by one or more processors and causing the one or more processors to perform functions, wherein,

the respective processes performed by the second execution means in the abnormal sound generation site determination method according to any one of claims 11 to 13 are executed.

15. A vehicle-mounted device is characterized in that,

the in-vehicle device includes the first execution device in the abnormal sound generation site determination method according to any one of claims 11 to 13.

Technical Field

The invention relates to a method for specifying a location where an abnormal sound is generated, a non-transitory storage medium, and an in-vehicle device.

Background

For example, japanese patent application laid-open No. 2016 & 222090, which will be described below, describes a device for suppressing abnormal sounds caused by backlash in a gear train in a hybrid vehicle in which a motor generator and an internal combustion engine are mechanically coupled to a power split device provided with the gear train. In this device, when a predetermined abnormal sound generation condition is satisfied, the torque of the motor generator is controlled to apply a propulsion torque to the gear train.

Disclosure of Invention

However, the abnormal sound is not necessarily limited to be generated under the envisaged condition. Therefore, it is not always easy to determine the main cause of the abnormal sound when the user perceives it and expresses it. Thus, the present invention provides an abnormal sound generation site determination method of determining a sound generation site, a non-transitory storage medium that executes the method, and an in-vehicle apparatus that executes the method.

A first aspect of the present invention relates to an abnormal sound generation site specifying method for specifying an abnormal sound generation site. The abnormal sound generation site specifying method executes the following processing. That is, map data of a predetermined map is stored in the storage device, the map being a map in which a sound variable that is a variable corresponding to a sound sensed in a vehicle and a state variable of a drive system device of the vehicle that is synchronized with the sound are input and output as a portion that is a factor of the sound. Causing an execution device to execute sound signal acquisition processing of acquiring a sound signal output from a microphone that senses sound in the vehicle in a state where the mapping data is stored. And causing the execution device to execute a state variable acquisition process of acquiring a state variable of the drive system device in a state where the map data is stored. The execution device is caused to execute a determination process of determining a sound generation site corresponding to the sound signal, using the sound variable based on the sound signal acquired by the sound signal acquisition process and the state variable synchronized with the sound signal as inputs to the map.

When an abnormal sound is generated in a vehicle, the abnormal sound is not always generated, and may be generated when a drive system device is in a specific state. In this case, even if a vehicle generating an abnormal sound is brought into a repair shop or the like, it may be difficult to reproduce the abnormal sound.

Therefore, according to the method of identifying a location where an abnormal sound is generated of the above-described aspect, the location where the sound is generated can be identified by using a map in which a sound variable and a state variable at the time of generating the abnormal sound are input variables and a location that is a factor of the sound is output.

In the abnormal sound generation site specifying method according to the first aspect, the sound signal acquisition process may be a process of acquiring the sound signal under a condition that the microphone is disposed at a predetermined site indicated in advance. The predetermined portion may be any of a head portion of a driver's seat, an instrument panel, a center portion of a rear seat, and a center console of the vehicle.

According to the abnormal-sound-generation-portion specifying method of the above configuration, by specifying the microphone placement portion at the time of acquiring the sound signal, it is possible to suppress a difference in the conditions for acquiring the abnormal sound between the time of learning the mapping data and the time of executing the specifying process of the abnormal sound generation portion using the mapping.

In addition, when the predetermined portion is a head portion of a driver's seat of the vehicle, the sound signal can be acquired under the same condition as the abnormal sound perceived by the driver. In addition, when the predetermined portion is an instrument panel or a center console, the sound signal can be acquired under a condition similar to a condition under which the driver perceives abnormal sound. In addition, when the predetermined portion is the rear seat center portion, the sound signal can be acquired under a condition similar to a condition under which a person seated in the rear seat perceives sound.

In the abnormal sound generation site determination method according to the first aspect, the sound signal acquisition process may include a process of acquiring the sound signal when the vehicle is set to a predetermined arrangement state. The predetermined arrangement state may be a state in which the vehicle is surrounded at least in a direction other than one direction out of four directions of the front and rear, the left and right, and two directions of the upper and lower sides of the vehicle.

In the case where the surroundings of the vehicle are surrounded, the sound generated in the vehicle is reflected at the object surrounding the vehicle, so in the case where the surroundings of the vehicle are surrounded, the sound generated in the vehicle is likely to be harsh to the user in the vehicle, as compared with the case where the surroundings of the vehicle are opened. Therefore, for example, when a user who intends to reproduce an abnormal sound in a garage or the like takes a vehicle into a repair shop or the like, the abnormal sound may not be reproduced even if the user wants to reproduce the abnormal sound in a state where the periphery of the vehicle is opened.

Therefore, according to the abnormal sound generation site specifying method configured as described above, the sound signal is acquired in a state where the sound signal is enclosed in at least three of the four directions of the front and rear, the left and right, and the up and down direction of the vehicle, and the abnormal sound that cannot be reproduced when the surroundings of the vehicle are open can be reproduced.

In the abnormal sound occurrence location determining method according to the first aspect, the execution device may be caused to execute an instruction process of instructing a predetermined driving operation of the vehicle when the output of the map is such that the probability of two or more candidates among the plurality of candidates for the location that is the cause of the abnormal sound is higher than the probability of the other candidates by a predetermined value or more. Here, the map may be a map that outputs a variable indicating a relative probability of each of the plurality of candidates. Further, the determination process may also include the following processes: when the instruction processing is performed, the sound variable based on the sound signal acquired by the sound signal acquisition processing and the state variable acquired by the state variable acquisition processing when a driving operation in accordance with an instruction based on the instruction processing is performed are input to the map, and a generation location of a sound corresponding to the sound signal is specified.

In general, even when the number of candidates of the abnormal sound generation portion is not reduced to one, there is a tendency that the plurality of candidates are different from each other in the degree of easiness of generation of the abnormal sound or the like according to the state of the drive system device. Therefore, according to the abnormal sound generation site specifying method configured as described above, when there are no fewer candidates for the abnormal sound generation site, a predetermined driving operation is instructed, and the sound signal and the state variable at that time are acquired, so that it is possible to acquire the sound signal and the state variable from which a plurality of candidates are easily recognized. Therefore, in the above method, the accuracy of specifying the abnormal sound generation site can be improved.

In the abnormal sound occurrence location determining method according to the first aspect, the sound variable generation process may be executed, in the sound variable generation process, the sound signal is input, and at least one of three of a predetermined frequency proportional to a rotation frequency of a rotary machine as a thrust generator of the vehicle and an intensity of a frequency component corresponding to each of frequencies which are integral multiples of the predetermined frequency, which are acquired by the state variable acquisition process, in the sound signal, a prominent frequency which is a frequency of a frequency band having an intensity greater than a frequency band adjacent to each of a low frequency side and a high frequency side in the sound signal, a prominent amount which is an amount by which the intensity of the prominent frequency is prominent with respect to the adjacent frequency band, and a duration during which the intensity of the sound signal is equal to or greater than a predetermined value is generated as the sound variable.

According to the abnormal sound generation site specifying method configured as described above, the effective feature amount is extracted from the sound signal, and thus, although the dimension of the input variable to the map is small, the map with high general-purpose performance and high accuracy of specifying the cause can be realized. Here, since the drive system device has a member that rotates in association with the rotation of the rotary machine, when an abnormal sound is generated in association with the rotation of the drive system device, the intensity of the sound of the frequency component proportional to the rotation frequency of the rotary machine tends to be particularly large. Therefore, when an abnormal sound accompanying the rotation of the drive system device is generated, the intensity of the frequency component corresponding to each of the predetermined frequency and the frequency that is an integral multiple thereof is used as the sound variable, and the generation position can be specified with high accuracy although the dimension of the input variable to be mapped is small.

In addition, in the case of having a prominent frequency, an abnormal sound perceived by the user is easily generated. Therefore, by using the projection frequency and the projection amount as the sound variables, it is possible to input information suitable for specifying the abnormal sound to the map, although the dimension of the input variable to the map is small.

In addition, the duration in which the intensity of the sound signal is equal to or greater than the predetermined value has a correlation with the time at which the abnormal sound is generated. The duration of the abnormal sound generation tends to vary depending on the location of the abnormal sound generation. Therefore, by using the above-described duration as the sound variable, the generation site of the abnormal sound can be accurately specified although the dimension of the input variable to be mapped is small.

In the abnormal sound generation site determination method according to the first aspect, the state variable acquisition process may include a process of acquiring at least one of a torque of a thrust force generation device of the vehicle and a change amount of the torque per unit time as the state variable.

In the case where an abnormal sound is generated from the drive system of the vehicle, the abnormal sound tends to be generated when the torque of the thrust force generation device, the amount of change thereof per unit time, has a specific value. According to the abnormal sound generation site specifying method configured as described above, by using these as input variables to the map, the abnormal sound generation site can be specified with high accuracy although the dimension of the input variables to the map is small. In particular, since abnormal sounds tend to be generated when the amount of change in torque is large, if the amount of change is used as an input variable, it is easy to improve the accuracy of specifying the location where the abnormal sound is generated.

In the above-described abnormal sound generation site determination method according to the first aspect, the vehicle may further include a transmission device that changes a ratio between a rotation speed of the rotary machine serving as the thrust force generation device and a rotation speed of the drive wheels. Further, the state variable acquisition process may include a process of acquiring a gear ratio of the transmission as the state variable.

When an abnormal sound is generated from a drive system of a vehicle, the abnormal sound may be generated when a gear ratio of a transmission device is a predetermined ratio. In addition, the gear ratio can represent the vehicle speed together with the rotation speed of the in-vehicle rotary machine, or represent the rotation frequency of the rotating member of the drive system device. Here, the abnormal sound generated in the vehicle may be generated at a specific vehicle speed. In addition, when an abnormal sound is generated, the intensity of the sound signal may increase in a predetermined frequency band proportional to the rotational frequency of the drive system component.

According to the abnormal sound occurrence location specifying method of the above configuration, the abnormal sound occurrence location can be specified with high accuracy, although the dimension of the input variable to be mapped is small, by using the gear ratio which is a variable effective in specifying the type of the abnormal sound for the above reason.

In the abnormal sound generation site determination method according to the first aspect, the state variable acquisition process may include a process of acquiring at least one of a vehicle speed and a rotation speed of a rotary machine as the thrust generator of the vehicle as the state variable.

An abnormal sound generated in a vehicle may be generated at a specific vehicle speed. In addition, the rotation speed of the rotary machine is a variable effective in determining the state of the drive system device. According to the abnormal sound occurrence location specifying method of the above configuration, for the above reason, the vehicle speed and the rotational speed of the rotary machine, which are variables effective in specifying the type of the abnormal sound, are used, and thus the abnormal sound occurrence location can be specified with high accuracy although the dimension of the input variable to be mapped is small.

In the abnormal sound generation site determination method according to the first aspect, the vehicle may include an internal combustion engine, a rotating electrical machine mechanically connectable to a crankshaft of the internal combustion engine, and a control device that performs an overlap process of overlapping a cancellation torque, which is a torque for reducing a torque ripple of the internal combustion engine, with a required torque for the rotating electrical machine. Further, the state variable acquisition process may include a process of acquiring the magnitude of the cancellation torque as the state variable.

When the control device performs the control of applying the cancellation torque and when the control device does not perform the control of applying the cancellation torque, a large difference occurs in the vibration caused by the torque variation of the crankshaft of the internal combustion engine. Therefore, when an abnormal sound is generated, the cause of the abnormal sound may be different depending on whether or not the control for applying the cancellation torque is performed. According to the method for specifying a location where an abnormal sound is generated, the accuracy of specifying the location where the abnormal sound is generated can be improved by including the magnitude of the canceling torque in the input variable of the map.

In the method of identifying a location where an abnormal sound is generated according to the first aspect, the vehicle may further include an internal combustion engine, a first motor generator, a second motor generator, and a control device. The second motor generator may be mechanically coupled to the internal combustion engine and the first motor generator via a power split device, and may be mechanically coupled to a drive wheel without via the power split device. The control device may execute processing of causing the first motor generator to generate a propulsion torque for reducing looseness of a gear train of the power split device when the internal combustion engine and the first motor generator are operating under an empty load. Further, the state variable acquiring process may include a process of acquiring a magnitude of the propulsion torque as the state variable.

When the control device performs the control of applying the propulsion torque and when the control device does not perform the control of applying the propulsion torque, a large difference occurs in noise caused by the backlash of the gear train of the power split device. Therefore, when an abnormal sound is generated, the cause of the generation of the abnormal sound may be different depending on whether or not the control for applying the propulsion torque is performed. According to the method for identifying a location where an abnormal sound is generated, the accuracy of identifying the location where the abnormal sound is generated can be improved by including the propulsion torque in the input variable of the map.

In the abnormal sound generation site specifying method according to the first aspect, the execution device may include a first execution device provided in the vehicle, and a second execution device and a third execution device that are not provided in the vehicle. The first execution device may execute the state variable acquisition process and a vehicle-side transmission process of transmitting the state variable acquired by the state variable acquisition process. The second execution device may be a portable terminal of a user of the vehicle equipped with the microphone, and execute a state variable reception process of receiving the state variable transmitted by the vehicle-side transmission process, the sound signal acquisition process, and a terminal-side transmission process of transmitting the sound signal acquired by the sound signal acquisition process and the state variable received by the state variable reception process. The third execution device may execute analysis-side reception processing for receiving the audio signals and the state variables transmitted by the terminal-side transmission processing from each of portable terminals of users of the plurality of vehicles, and the determination processing.

According to the abnormal sound generation site specifying method configured as described above, since the specifying process is executed by the third execution device outside the vehicle, the abnormal sound information of the plurality of vehicles can be collected by the third execution device. Further, since the portable terminal transmits the state variable and the audio signal to the third execution device, the required items for the first execution device can be reduced.

In the abnormal sound generation site specifying method according to the above aspect, the second execution device may execute notification acquisition processing for acquiring a notification to the effect of the abnormal sound when the abnormal sound is sensed by a person. The audio signal to be transmitted by the terminal-side transmission process may be the audio signal acquired by the audio signal acquisition process during a predetermined period determined according to the timing at which the notification is acquired by the notification acquisition process.

According to the abnormal sound generation site specifying method of the above configuration, when a person senses an abnormal sound, the person notifies that the abnormal sound, and the second execution device is to transmit a sound signal in a predetermined period specified by the timing at which the notification is acquired, so that the third execution device can perform the specifying process using the sound signal at the time when the abnormal sound is actually generated while minimizing the amount of data to be handled.

In the abnormal sound generation site specifying method according to the above aspect, the second execution device may execute additional variable acquisition processing for acquiring an additional variable indicating at least one of an open/close state of a window of the vehicle when the abnormal sound is sensed and a situation in which the vehicle is present. The terminal-side transmission processing may also include processing of transmitting the additional variable in addition to the sound signal and the state variable. The parsing-side receiving process may also include a process of receiving the additional variable. The input of the map may also include the additional variable in addition to the sound variable and the state variable. The specifying process may include a process of specifying a sound generation site corresponding to the sound signal, using the sound variable, the state variable, and the additional variable as inputs to the map.

According to the abnormal sound generation site specifying method configured as described above, by using the additional variable as an input to the map in addition to the sound variable and the state variable, more detailed information on the abnormal sound generation state can be provided to the map than in the case where the additional variable is not used, and the abnormal sound generation site can be specified with higher accuracy.

A storage medium according to a second aspect of the present invention is a non-transitory storage medium that stores commands that can be executed by one or more processors and that cause the one or more processors to execute the following functions. The storage medium executes each process executed by the second execution device in the abnormal sound generation location specifying method.

A third aspect of the present invention provides an in-vehicle device including the first execution device in the abnormal sound generation location specifying method.

Drawings

Features, advantages and technical and industrial significance of exemplary embodiments of the present invention will be described below with reference to the accompanying drawings, in which like symbols represent like elements, and wherein:

fig. 1 is a block diagram showing the configuration of a system of the first embodiment of the present invention.

Fig. 2A is a flowchart showing the sequence of processing performed by the control device of the system.

Fig. 2B is a flowchart showing the sequence of processing performed by the portable terminal of the system.

Fig. 3 is a timing chart showing the feature amount of the first embodiment.

Fig. 4 is a timing chart showing the feature amount of the first embodiment.

Fig. 5 is a diagram showing an example of instructions for a location of the mobile terminal according to the first embodiment.

Fig. 6 is a flowchart showing the order of processing performed by the data analysis center of the first embodiment.

Fig. 7 is a timing chart showing the feature amount of the first embodiment.

Fig. 8 is a diagram showing feature quantities of the first embodiment.

Fig. 9 is a diagram showing a list of data of the determination results of the first embodiment.

Fig. 10A is a flowchart showing the sequence of processing performed by the data analysis center of the system of the second embodiment of the present invention.

Fig. 10B is a flowchart showing the sequence of processing performed by the portable terminal of the system.

Fig. 11 is a block diagram showing the configuration of a system according to a third embodiment of the present invention.

Fig. 12 is a diagram showing a configuration example of the vehicle of the third embodiment.

Fig. 13A is a diagram showing an example of instructions for a location of a mobile terminal in modification 1 of the above embodiment.

Fig. 13B is a diagram showing an example of instructions for a location of the mobile terminal in modification 2 of the above embodiment.

Fig. 13C is a diagram showing an example of instructions for a location of a mobile terminal in modification 3 of the above embodiment.

Detailed Description

Hereinafter, a first embodiment of the present invention relating to a method of specifying a location where an abnormal sound is generated will be described with reference to the drawings.

As shown in fig. 1, the power split device 10 of the vehicle includes a planetary gear mechanism including a sun gear S, a carrier C, and a ring gear R. A crankshaft 12a of the internal combustion engine 12 is mechanically coupled to a carrier C of the power split device 10, a rotary shaft 14a of the first motor generator 14 is mechanically coupled to the sun gear S, and a rotary shaft 16a of the second motor generator 16 is mechanically coupled to the ring gear R. The drive wheels 30 are mechanically coupled to the ring gear R via a transmission 20 including clutches C1, C2, brakes B1, B2, and a one-way clutch F1.

The transmission 20 is supplied with the hydraulic oil discharged from the oil pump 40 mechanically coupling the driven shaft and the carrier C of the power split device 10. The control device 50 controls the control amounts such as the torque of the internal combustion engine 12, the exhaust gas component ratio, the torque of the first motor generator 14, and the torque of the second motor generator 16, with the vehicle as a control target. The control device 50 refers to the output signal Scr of the crank angle sensor 90, the output signal Sm1 of the first rotation angle sensor 92 for sensing the rotation angle of the rotating shaft 14a of the first motor generator 14, and the output signal Sm2 of the second rotation angle sensor 94 for sensing the rotation angle of the rotating shaft 16a of the second motor generator 16, in order to control the control amount. Further, control device 50 refers to vehicle speed SPD detected by vehicle speed sensor 96.

The control device 50 includes a CPU52, a ROM54, a peripheral circuit 56, and a communicator 58, and can communicate via a local network 59. Here, the peripheral circuit 56 includes a circuit that generates a clock signal that defines an internal operation, a power supply circuit, a reset circuit, and the like. The control device 50 controls the control amount by executing a program stored in the ROM54 by the CPU 52.

The control device 50 can communicate with a portable terminal 60 of a user of the vehicle via a communicator 58. The portable terminal 60 includes a CPU62, a storage device 63 as an electrically rewritable nonvolatile memory, a ROM64, a microphone 65, a peripheral circuit 66, a display unit 67 such as an LCD, a touch panel 61 disposed on the display unit 67 in a superposed manner, a speaker SP, and a communicator 68, and can communicate with each other via a local network 69.

The mobile terminal 60 can communicate with the control device 50 via the communication device 68, and can also communicate with other mobile terminals 60 and the data analysis center 80 via the global network 70. In fig. 1, one mobile terminal 60 having an internal structure such as a CPU62 not shown is illustrated as another mobile terminal 60.

The data analysis center 80 includes a CPU82, a storage device 83 as an electrically rewritable nonvolatile memory, a ROM84, a peripheral circuit 86, and a communication device 88, and can communicate with each other via a local network 89.

The system shown in fig. 1 constitutes a system that specifies a generation site of an abnormal sound generated in a vehicle. The following describes processing relating to a method of determining a location where an abnormal sound is generated. Fig. 2A and 2B show the sequence of processing executed by the control device 50 of the vehicle and the portable terminal 60 of the user of the vehicle. In detail, the process shown in fig. 2A is realized by the CPU52 of the control device 50 repeatedly executing a program stored in the ROM54, for example, at predetermined cycles. The processing shown in fig. 2B is realized by the CPU62 of the portable terminal 60 repeatedly executing the application 63a stored in the storage device 63, for example, at predetermined cycles. In the following, the step number of each process is expressed by a numeral given with "S" at the head. In addition, the processes of fig. 2A and 2B are described below in accordance with a time series of processes executed by the control device 50 and the mobile terminal 60.

In the series of processing shown in fig. 2A, the CPU52 first determines whether or not synchronization with the mobile terminal 60 is established (S10). When the CPU52 determines that synchronization has been established (yes in S10), it acquires the state variables of the drive system device of the vehicle (S12). In the first embodiment, the state variables include the vehicle speed SPD, the rotation speed NE of the crankshaft 12a of the internal combustion engine 12, the rotation speed Nm1 of the rotary shaft 14a of the first motor generator 14, and the rotation speed Nm2 of the rotary shaft 16a of the second motor generator 16. Here, the CPU52 calculates the rotation speed NE using the output signal Scr of the crank angle sensor 90 as an input. The CPU52 calculates the rotation speed Nm1 using the output signal Sm1 of the first rotation angle sensor 92 as an input. The CPU52 calculates the rotation speed Nm2 using the output signal Sm2 of the second rotation angle sensor 94 as an input.

The state variables include a speed ratio variable Vsft, which is a variable indicating the speed ratio based on the transmission device 20, a required torque Trqed, which is a torque required by the internal combustion engine 12 to generate power required by the vehicle, and a change amount Δ Trqed per unit time thereof. The state variables include a required torque Trqmg1, which is a torque required by the first motor/generator 14 to generate the motive power, and a change amount Δ Trqmg1 per unit time thereof, a required torque Trqmg2, which is a torque required by the second motor/generator 16 to generate the motive power, and a change amount Δ Trqmg2 per unit time thereof. The required torques Trqmg1 and Trqmg2 are not necessarily limited to positive values. That is, for example, the required torque Trqmg1 may be a sign corresponding to power generation in order to appropriately split the power of the internal combustion engine 12 by the power split device 10. When negative power is required for the vehicle, such as when deceleration of the vehicle is required, the required torque Trqmg2 has a sign corresponding to power generation.

The variations Δ Trqed, Δ Trqmg1, and Δ Trqmg2 are variables having a strong positive correlation with the abnormal sound. That is, as illustrated in fig. 3 with respect to the change amount Δ Trqed, when the change amount Δ Trqed becomes large, abnormal sound is likely to be generated. In particular, since an abnormal sound is likely to occur when the torque is switched from one of the pair of positive and negative signs to the other, the amount of change in torque at the time of switching can be information that is useful in identifying the abnormal sound. Therefore, by using the required torques Trqed, Trqmg1, and Trqmg2 and the changes Δ Trqed, Δ Trqmg1, and Δ Trqmg2, it is possible to obtain information particularly effective in specifying abnormal sounds, such as information on changes when the signs of the required torques Trqed, Trqmg1, and Trqmg2 are switched.

Returning to fig. 2A and 2B, the state variables include a cancellation torque (cancel torque) Trqcan and a propulsion torque (push torque) Trqpush. The canceling torque Trqcan is a torque applied to the power split device 10 by the first motor generator 14 and the second motor generator 16 in order to cancel out the periodic variation in the torque of the internal combustion engine 12 in the cycle of the occurrence interval of the combustion stroke of the internal combustion engine 12 in the power split device 10. The cancellation torque Trqcan is a torque that is appropriately allocated to overlap the required torque trqmmg 1 for the first motor generator 14 and the required torque trqmmg 2 for the second motor generator 16.

In fig. 4, the torque of the internal combustion engine 12 is indicated by a broken line, and the total torque of the internal combustion engine 12 and the cancellation torque Trqcan is indicated by a solid line. The canceling torque Trqcan periodically fluctuates as shown by a chain line in fig. 4 in the same manner as the torque of the internal combustion engine 12, but in the first embodiment, an amplitude value thereof is used as the canceling torque Trqcan which is the state variable. This is one method of quantifying the magnitude of the cancellation torque Trqcan.

On the other hand, the propulsion torque Trqpush is a torque for suppressing abnormal sounds caused by looseness of the gear train of the power split device 10, and is a torque applied to the power split device 10 by the first motor generator 14. For example, when a predetermined condition such as a condition that the driving force is applied to the driving wheels 30 only by the power of the second motor generator 16 is satisfied, the CPU52 calculates the propulsion torque Trqpush.

Returning to fig. 2A and 2B, the CPU52 operates the communicator 58 to transmit the state variables when acquiring the state variables (S14).

On the other hand, as shown in fig. 2B, the CPU62 determines whether or not synchronization with the control device 50 is established (S20). When determining that synchronization is established (yes in S20), CPU62 operates display unit 67 to display instruction information indicating at which location in the vehicle portable terminal 60 should be placed on display unit 67 (S22).

In fig. 5, the instruction information displayed on the display unit 67 is exemplified. In the first embodiment, it is assumed that the portable terminal 60 is disposed on the head of the driver's seat in order to record the abnormal sound reaching the head of the user by the portable terminal 60. Therefore, the CPU62 requests the driver to place the portable terminal on the head of the driver's seat, which is the information that the driver intends to place the portable terminal. "such instruction information is displayed as visual information on the display unit 67.

Returning to fig. 2B, the CPU62 starts receiving the above state variables transmitted from the control device 50 (S24). Next, the CPU62 starts recording of the output signal of the microphone 65 (S26). Then, the CPU62 monitors whether or not there is a signal indicating that the abnormal sound is perceived from the user based on the output signal of the microphone 65 (S28: no). Here, for example, signals such as "present", "start", "noise", etc. are predetermined, and the presence or absence of the signal is monitored. When the CPU62 determines that there is a signal (yes in S28), it stores the time-series data of the sound signal captured by the microphone 65 for a predetermined period consisting of a predetermined time period before and after the time point of the presence of the signal in the storage device 63 in association with the time-series data of the received state variable (S30). That is, the CPU62 stores the state variable received before and after a predetermined time of the signaled timing and the audio signal before and after the predetermined time in the storage device 63. Further, the CPU62 removes data that deviates from the predetermined period.

The CPU62 executes the process of S30 until it is determined that a predetermined time has elapsed since the signal, in other words, until it is determined that the predetermined period has ended (S32: no). When it is determined that the predetermined time has elapsed (yes in S32), the CPU62 operates the communicator 68 to transmit a state variable transmission stop command to the control device 50 (S34).

On the other hand, as shown in FIG. 2A, when the CPU52 receives the transmission stop command (S16: YES), it stops the transmission processing of the state variables (S18). When the process of S18 is completed and when a negative determination is made in the process of S10, the CPU52 once ends the series of processes shown in fig. 2A.

On the other hand, as shown in fig. 2B, the CPU62 stops the reception processing of the state variables and the recording processing (S36). Next, the CPU62 acquires the open/close state of the window of the vehicle, and the situation identification information of the vehicle identifying several predefined situations such as whether the vehicle is located in an urban area, an expressway, or a stop in a parking lot (S38). This processing is realized by the CPU62 operating the display unit 67 to display visual information on the display unit 67, the visual information urging the user to input the open/close state of the window of the vehicle and the status identification information of the vehicle. That is, the user inputs the opening/closing state of the window and the status recognition information of the vehicle through the touch panel 61 based on the visual information displayed on the display unit 67.

Next, the CPU62 operates the communicator 68 to transmit the sound signal, the state variables, and the additional variables to the data analysis center 80 (S40). In the first embodiment, the additional variables are a window variable Vw which is a variable indicating the open/close state of the window and an identification variable Vst which is a variable indicating the condition identification information of the vehicle. When the process of S40 is completed and when a negative determination is made in the process of S20, the CPU62 once ends the series of processes shown in fig. 2B.

Fig. 6 shows the order of processing performed by the data analysis center 80. The process shown in fig. 6 is realized by the CPU82 repeatedly executing a program stored in the ROM84, for example, at predetermined cycles.

In the series of processing shown in fig. 6, the CPU62 first receives the audio signal, the state variable, and the additional variable transmitted from the mobile terminal 60 (S50). Next, the CPU82 performs a process of generating a feature quantity from the sound signal and regarding the feature quantity as a sound variable (S52). In the first embodiment, the sound variable includes a duration T1 during which the intensity (dB) of the sound signal is equal to or greater than a predetermined value.

Fig. 7 illustrates the duration T1. In the example shown in fig. 7, the intensity of the audio signal becomes equal to or greater than the predetermined value a1 over the period of time ta to tb, and the duration T1 that becomes equal to or greater than the predetermined value a1 is used as the audio variable.

Returning to fig. 6, the sound variables include a protrusion frequency fpr, which is a frequency of a frequency band in which the intensity of the sound signal is more protruded than the frequency bands adjacent to the low frequency side and the high frequency side, and a protrusion amount Ipr.

Fig. 8 illustrates the protrusion frequency fpr and the protrusion amount Ipr. As illustrated in fig. 8, since an abnormal sound is likely to occur when there is a portion where the intensity of the sound signal is prominent, the first embodiment uses the prominent frequency fpr and the prominent amount Ipr as the sound variables. The CPU82 defines the protrusion frequency fpr and the protrusion amount Ipr when the intensity of the audio signal of the target frequency band is greater than the intensities of the audio signals of the low frequency side and the high frequency side by a predetermined amount or more.

Referring back to fig. 6, the sound variables include sound pressures Ima1, Imb1, Imc1, and Imd1 of frequency components related to the meshing of the gear trains of the power split device 10 and the transmission 20, and sound pressures Iman, Imbn, Imcn, and Imdn of the "n-1: n" order harmonic components of the respective frequency components, which are 2 to 5 ". Here, the frequency related to the engagement of the gear train refers to, for example, the reciprocal of the time required for the engagement of the gears of the sun gear S and the carrier C of the power split device 10 to change. The CPU62 can calculate a frequency relating to the engagement of the sun gear S and the carrier C, which is the reciprocal of the time required for the engagement of the gears of the sun gear S and the carrier C to change, from the number of teeth of the sun gear S and the number of teeth of the carrier C, using the rotation speeds NE, Nm1, Nm2 as inputs.

In the present first embodiment, the CPU82 takes the rotation speeds NE, Nm1, Nm2 as inputs, and calculates four frequencies corresponding to a, b, c, d, respectively, as the frequencies related to meshing. In the first embodiment, the four frequencies include, for example, a frequency related to engagement of the ring gear R and the carrier C, a frequency related to engagement of the counter gear of the transmission 20, and the like, in addition to the above-described frequencies.

Specifically, the CPU82 calculates the intensities of the frequency components of the time-series data of the sound signal by fourier transform, thereby calculating the sound pressures Iman to Imdn (n is 1 to 5). Then, the CPU82 inputs the state variables and additional variables received in the processing of S50 and the sound variables generated in the processing of S52 into the feature quantity F, which is an input of the map of the generation site of the output abnormal sound defined by the map data 83a stored in the storage device 83 shown in fig. 1 (S54). Next, the CPU82 assigns the feature amounts F to the mapped input variables x (1) to x (38), respectively (S56).

Then, the CPU82 calculates the output of the mapping (S58). The map of the first embodiment is a neural network having one layer as an intermediate layer, which receives the input variables x (0) to x (38). Specifically, when the number of nodes n1 in the intermediate layer is used, the CPU82 inputs the input variables x (0) to x (38) to a linear map defined by coefficients wFjk (k is 0 to 38, and j is 0 to n1) and calculates an output value in the dimension "n 1+ 1". Then, the CPU82 obtains output values by inputting the output values to the activation function f, obtains output values for the score prototype variables yi by inputting the obtained output values to a linear map defined by coefficients wSij (j is 0 to n1), and obtains output values for the score prototype variables yi, thereby calculating the score prototype variables yi. Here, as the activation function f, for example, ReLU (Rectified Linear Unit) or hyperbolic tangent may be used. Note that the input variable x (0) is a bias parameter, and "1" is always substituted for the input variable x (0).

Then, the CPU62 normalizes the scoring prototype variables y1, y2, y3, … by the Softmax function, and calculates scores Sc (1), Sc (2), Sc (3), …. Here, the scores Sc (2), Sc (3), and Sc (4) … indicate the probabilities of the abnormal sound generation sites, respectively, and the score Sc (1) indicates the probability of the abnormality at the abnormal sound generation site that cannot be specified as the scores Sc (2), Sc (3), and Sc (4) …, respectively, indicate the probabilities.

Next, the CPU82 extracts the largest score Sc (m) among the scores Sc (1), Sc (2), Sc (3), … (S60). Then, the CPU62 specifies the factor specifying data 83b stored in the storage device 83 shown in fig. 1, from the variable m of the extracted score sc (m), the factor causing the abnormal sound (S62).

Fig. 9 illustrates a part of the relationship between the maximum score among the scores Sc (1), Sc (2), Sc (3), and …, which is defined by the factor specifying data 83b, and the generation location of the abnormal sound at that time. Fig. 9 illustrates the meaning that, when the score Sc (1) is the maximum, the determination result indicates that which candidate among the candidates of the generation part specified by the factor specifying data 83b cannot be specified. Fig. 9 illustrates the case where the score Sc (2) is the maximum, and the determination result indicates that the oil pump 40 is the abnormal sound generation site. Fig. 9 shows that the first motor generator 14 is the abnormal sound generation site when the score Sc (3) is the maximum, and that the second motor generator 16 is the abnormal sound generation site when the score Sc (4) is the maximum. Fig. 9 shows that the determination result indicates that the engagement of the ring gear R and the carrier C of the power split device 10 is the abnormal sound generation site when the score Sc (5) is the maximum, and that the engagement of the sun gear S and the carrier C of the power split device 10 is the abnormal sound generation site when the score Sc (6) is the maximum. Fig. 9 shows that the determination result indicates that the meshing of the counter gear of the transmission 20 is the abnormality occurrence portion when the score Sc (7) is the maximum.

Returning to fig. 6, the CPU82 operates the communicator 88 to notify the portable terminal 60 of the determination result when determining the primary cause (S64). Here, when the generation site of the abnormal sound is specified, the result of the specification of the generation site and the action intended by the user are notified. When the generation site of the abnormal sound cannot be specified, the intention and the action intended by the user are notified. Here, the information on the action that the user wants to perform may be information on when to go to which repair shop. This can be achieved by including information on a coping process in association with the abnormality occurrence portion in the factor specifying data 83b, for example. Incidentally, in the processing of S64, the ID of the user and the determination result may be notified to the nearest repair shop in addition to the portable terminal 60 notified to the user.

When the process of S64 is completed, the CPU82 once ends the series of processes shown in fig. 6.

Then, when the user brings the vehicle into a repair shop, the specified generation site is investigated if the generation site is specified. If the determination result in S62 is correct, the parts of the generation site are replaced. In contrast, when no abnormality is found in the generation site specified in the processing of S62, another factor is investigated. Then, as a result of the examination, when the generation site of the abnormal sound is specified, the CPU82 updates the coefficients wFj and wSij by a known method so that the error between the input variables x (1) to x (38) set by the processing of S56 and the teaching data in which the score Sc of the newly specified generation site of the abnormal sound is "1" and the other scores Sc are "0" is reduced.

On the other hand, if the determination is not possible in the process of S62, the cause of the abnormal sound is also examined. When the abnormal sound generation site is specified as a result of the examination, the CPU82 updates the coefficients wFj and wSij so that the error between the input variables x (1) to x (38) set by the processing of S56 and the teaching data in which the score Sc of the newly specified abnormal sound generation site is "1" and the other scores Sc are "0" is reduced.

Incidentally, the map data 83a at the time of shipment of a vehicle of a certain specification is a learned model that is learned by teaching and learning using training data generated by intentionally creating a situation of deepening the aged deterioration at the stage of a test piece of the vehicle to generate abnormal sounds.

Here, the operation and effect of the first embodiment will be described.

When the user senses an abnormal sound in the vehicle, the user starts the application 63a of the portable terminal 60 and transmits a signal indicating the abnormal sound when the abnormal sound is generated. On the other hand, the portable terminal 60 extracts the state variable from the control device 50 by executing the application 63a by the CPU62, records the sound around the vehicle by the microphone 65, and stores the state variable in a predetermined time period before and after the timing of transmitting the signal in the storage device 63 in association with the sound signal in the period. Then, the CPU62 transmits the state variables and the audio signals in the predetermined period to the data analysis center 80. The data analysis center 80 generates an audio variable from the received audio signal, and calculates a score Sc related to the factor of the abnormal sound by inputting the audio variable and the state variable to a map defined by the map data 83 a. Then, by identifying the factor having the largest score Sc as the factor causing the abnormal sound, the abnormal sound generation site can be identified.

According to the first embodiment described above, the following operations and effects can be obtained.

(1) The sound signal is acquired on the condition that the microphone 65 is disposed on the head of the driver's seat of the vehicle. This makes it possible to acquire the sound signal under the same condition as the abnormal sound perceived by the driver.

(2) The sound variables include sound pressures Ima1, Imb1, Imc1, and Imd1 of frequencies related to meshing, and sound pressures Iman, Imbn, Imcn, and Imdn of higher harmonics thereof (n is 2 to 5). Thus, even if the dimension of the input variable to the map is small, when an abnormal sound is generated due to engagement, it is possible to specify with high accuracy that the abnormal sound is generated due to engagement.

(3) The sound variable includes a projecting frequency fpr and a projecting amount Ipr, in view of the fact that abnormal sound is likely to occur when there is a frequency band projecting from both the low frequency side and the high frequency side. This enables the feature quantity relating to the abnormal sound to be efficiently included in the sound variable, and thus the dimension of the input variable to the map can be reduced.

(4) The sound variable includes a duration T1 in which the intensity of the sound signal is a predetermined value or more. This makes it possible to use a variable correlated with the time of occurrence of an abnormal sound as a sound variable. Since the duration of the abnormal sound generation tends to vary depending on the abnormal sound generation site, the duration T1 is used as the sound variable, so that the abnormal sound generation site can be accurately specified, and the dimension of the mapped input variable can be suppressed from increasing.

(5) The state variables include the required torque Trqed, Trqmg1 and Trqmg 2. Thus, since the specification processing can be executed using information useful as the state of the drive system device when the abnormal sound is generated, it is possible to specify the generation location of the abnormal sound with high accuracy and suppress an increase in the dimension of the input variable of the map.

(6) The state variables include variation amounts Δ Trqed, Δ Trqmg1, and Δ Trqmg 2. Since the abnormal sound tends to be generated when the amount of change in the torque is large, by including the amounts of change Δ Trqed, Δ Trqmg1, and Δ Trqmg2 in the state variables, it is possible to use variables that are useful in specifying the situation where the abnormal sound is generated as the feature quantities, and further, it is possible to improve the accuracy of specifying the location where the abnormal sound is generated. In particular, from the required torques Trqed, Trqmg1, Trqmg2 and the changes Δ Trqed, Δ Trqmg1, Δ Trqmg2, information particularly useful in capturing abnormal sounds can be obtained, such as changes Δ Trqed, Δ Trqmg1, Δ Trqmg2 when the signs of the required torques Trqed, Trqmg1, Trqmg2 are switched.

(7) The state variable includes a speed ratio variable Vsft indicating a speed ratio effective in determining the type of the abnormal sound. Here, when an abnormal sound is generated from the drive system of the vehicle, the abnormal sound may be generated when the gear ratio of the transmission device 20 is a predetermined ratio. Therefore, since the gear ratio variable Vsft is included, the feature quantity which is useful in specifying the situation where the abnormal sound is generated can be used as the state variable, the dimension of the input variable to the map can be reduced, and the generation location of the abnormal sound can be specified with high accuracy.

(8) The state variables include a vehicle speed SPD. Here, the abnormal sound generated in the vehicle may be generated at a specific vehicle speed SPD. Therefore, since the feature quantity, which is useful in specifying the condition in which the abnormal sound is generated, can be used as the state variable by including the vehicle speed SPD in the state variable, the dimension of the input variable to the map can be reduced, and the abnormal sound generation site can be specified with high accuracy.

(9) The state variables include rotation speeds NE, Nm1, Nm 2. Since the rotation speeds NE, Nm1, and Nm2 are variables effective in determining the state of the drive system device, it is possible to reduce the dimension of the input variables to the map and accurately determine the occurrence location of the abnormal sound by using them as the state variables.

(10) The state variables include the cancellation torque Trqcan. Here, when the control for applying the canceling torque Trqcan is performed and the control for applying the canceling torque Trqcan is not performed, a large difference occurs in the vibration caused by the torque variation of the crankshaft 12a of the internal combustion engine 12. Therefore, when an abnormal sound is generated, the cause of the abnormal sound may be different depending on whether or not the control for applying the cancel torque Trqcan is performed. Therefore, by including the cancel torque Trqcan in the state variable, it is possible to use a characteristic amount appropriate for grasping the magnitude of the vibration caused by the torque fluctuation as the state variable, and therefore it is possible to improve the accuracy of specifying the generation portion of the abnormal sound.

(11) The state variables include the propulsion torque Trqpush. Here, when the control to apply the propulsion torque Trqpush is performed and the control to apply the propulsion torque Trqpush is not performed, a large difference occurs in noise due to backlash of the gear train of the power split device 10. Therefore, when an abnormal sound is generated, the cause of the abnormal sound may be different depending on whether or not the control for applying the propulsion torque Trqpush is performed. Therefore, by including the propulsion torque Trqpush in the state variable, it is possible to use a characteristic amount appropriate for grasping the magnitude of the vibration caused by the torque fluctuation as the state variable, and therefore it is possible to improve the accuracy of specifying the generation location of the abnormal sound.

(12) The feature quantities are selected as input variables to the mapping according to the knowledge of the skilled person. Therefore, compared to the case where a variable that specifies the state of the drive system or an acoustic variable generated from an acoustic signal is randomly set as the feature value, the number of layers of the intermediate layer of the neural network can be reduced, and the structure of the map can be simplified.

(13) The CPU62 receives the state variables from the control device 50, and transmits the state variables associated with the audio signals to the data analysis center 80. Thus, in the data analysis center 80, since the phenomena occurring in a plurality of vehicles can be collected, the frequency of relearning the mapping data 83a so that the accuracy of specifying the abnormality occurrence portion can be improved can be increased.

(14) The CPU62 transmits the window variable Vw and the identification variable Vst to the data analysis center 80 in addition to the state variable and the sound signal. Thus, more detailed information on the abnormal sound generation state can be provided to the map, and the abnormal sound generation site can be specified with higher accuracy, as compared with the case where the window variable Vw and the identification variable Vst are not used.

(15) The CPU62 of the mobile terminal 60 receives the state variables from the control device 50 by executing the application 63a, and transmits the state variables associated with the audio signals to the data analysis center 80. Thus, for example, when an abnormal sound occurs, the user may install the application 63a in the mobile terminal 60 of the user. Therefore, redundancy can be suppressed compared to a case where, for example, a function of transmitting a state variable associated with a sound signal to the data analysis center 80 is mounted in advance on the control device 50 in order to cope with a rare situation such as occurrence of an abnormal sound.

Hereinafter, a second embodiment of the present invention will be described with reference to the drawings, focusing on differences from the first embodiment.

In the first embodiment, the main cause of an abnormality is determined from the maximum value among the scores Sc (1), Sc (2), Sc (3), and …. In contrast, in the second embodiment, when two of the scores Sc (1), Sc (2), Sc (3), and … have a value larger than the other scores, the sound signal and the state variable are acquired in a specific driving state in order to determine which of them is.

Fig. 10A and 10B show a procedure of processing specific to the second embodiment among the processing executed by the data analysis center 80 and the mobile terminal 60. The process shown in fig. 10A is realized by the CPU82 repeatedly executing a program stored in the ROM84, for example, at predetermined cycles. Fig. 10A shows a modified part of the process shown in fig. 6. On the other hand, the process shown in fig. 10B is realized by repeatedly executing a program stored in the ROM64 by the CPU62, for example, at predetermined cycles. In the processing shown in fig. 10B, the processing corresponding to the processing shown in fig. 2B is given the same step number for convenience.

In the series of processing shown in fig. 10A, the CPU82 first determines whether or not the score Sc (m) that is the maximum value among the scores Sc (1), Sc (2), Sc (3), and … selected in the processing of S60 is equal to or greater than the threshold Sth (S70). This processing is processing for determining whether or not the reliability of the factor specified by the variable m corresponding to the score sc (m) and the factor specifying data 83b is a predetermined level or more. If the CPU82 determines that the threshold Sth is equal to or greater than the threshold Sth, the process proceeds to S62.

On the other hand, if the CPU82 determines that the score is smaller than the threshold Sth (S70: no), it determines whether or not the score sc (m) that is the maximum value and the score sc (n) that is a specific score other than the maximum value are equal to or greater than a predetermined value SthL that is smaller than the threshold Sth (S72). Here, the predetermined value SthL is set so that the difference between the scores sc (m), sc (n), and the other scores becomes a predetermined level or more. This means that the reliability of which of the main causes determined from the variables m, n is the actual main cause is a predetermined level or more.

When it is determined that the predetermined value SthL or more (S72: yes), the CPU82 searches for a driving operation command for identifying which of the factors identified by the variables m and n is the actual factor from the factor identifying data 83b, and operates the communicator 88 to transmit the searched driving operation command to the portable terminal 60 (S76).

In contrast, as shown in fig. 10B, the CPU62 determines whether a driving operation command is received (S80). When the CPU62 determines that it has received a signal (yes in S80), it restarts the synchronization process with the control device 50 (S82). Then, the CPU62 performs sound guidance for instructing the user to perform driving corresponding to the received driving operation command by operating the speaker SP (S84). Then, the CPU62 starts receiving state variables when the user performs driving operations in accordance with the driving instructions (S24). Next, the CPU62 executes the processes of S28 to S36. However, the processing of S32 here is processing of whether or not a predetermined length of time has elapsed since the start of recording.

Then, the CPU62 operates the communicator 68 to transmit the sound signal and the state variables stored through the processing of S30 to the data analysis center 80 in the case where the processing of S36 is completed (S86). When the process of S86 is completed and when a negative determination is made in the process of S80, the CPU62 once ends the series of processes shown in fig. 10B.

In contrast, as shown in fig. 10A, the CPU82 receives the state variables and the sound signals transmitted through the processing of S86 (S78), and executes the processing of S52 to the processing of S64 of fig. 6. On the other hand, when a negative determination is made in the process of S72, the CPU62 substitutes "1" for the score Sc (1) (S74), and proceeds to the process of S62.

As described above, according to the second embodiment, in a situation where it is not possible to accurately identify which of the factors corresponding to the variables m and n is the factor, the user is urged to perform a predetermined driving operation, and the scores Sc (1), Sc (2), Sc (3), and … are calculated again from the sound signal and the state variable at the time of performing the predetermined driving operation. This makes it possible to accurately identify which of the factors corresponding to the variables m and n is the factor.

Hereinafter, a third embodiment of the present invention will be described with reference to the drawings, focusing on differences from the first embodiment.

In the first embodiment, when the user senses the abnormal sound, the state variable and the sound signal are transmitted to the data analysis center 80 using the mobile terminal 60, and the data analysis center 80 identifies the cause of the abnormal sound. In contrast, in the third embodiment, the user takes the vehicle into the repair shop to take the vehicle into the repair shop.

Fig. 11 shows a configuration of an abnormal sound generation site specifying system according to the third embodiment. In fig. 11, members corresponding to those shown in fig. 1 are denoted by the same reference numerals for convenience.

As shown in fig. 11, the scanning tool 100 for the repair shop RF includes a CPU102, a ROM104, a peripheral circuit 106, and a communicator 108, which are capable of communicating via a local network 109. The communicator 108 has a function of communicating with the communicator 58. The communication with the control device 50 may be, for example, a wired connection using a dedicated terminal of the control device 50.

In addition, the repair shop RF is provided with a manual 112 for coping with the abnormality. In this manual, specifications regarding the arrangement position of the microphone 114 and the arrangement position of the vehicle are described.

In particular, the manual 112 also describes, as one of the arrangement of the vehicle when recording, a microphone 114 is arranged to record sound in a state where the vehicle VC is arranged as illustrated in fig. 12. Fig. 12 shows an example in which the right side of the vehicle VC faces the side wall 120, the left side faces the side wall 122, the upper side faces the ceiling 124, and the front side faces the front wall 126. That is, in the example shown in fig. 12, the vehicle VC is surrounded on the upper, lower, left, right, and front sides thereof.

The reason why such a configuration is specified is that, in the case where the vehicle VC is disposed in a garage, noise generated inside the vehicle VC is reflected at the wall of the garage, and therefore, is likely to become abnormal sound perceived by the user, as compared with the case where the periphery of the vehicle VC is opened. That is, even if the user takes the vehicle into the repair shop RF to reproduce the abnormal sound generated in the vehicle VC in the repair shop RF, the abnormal sound may not be reproduced when the vehicle VC is placed in a place where the surroundings are open. Therefore, in the manual 112, it is specified that recording is performed in a state in which at least three directions of at least two directions of up and down, and four directions of front and back, and left and right are surrounded. In addition, the manual 112 is also defined to mean that the recording is performed in a state in which the vehicle is surrounded in all of six directions, i.e., front-rear, left-right, and up-down.

The scanning tool 100 executes a process corresponding to the process executed by the mobile terminal 60 in the first embodiment. In this way, the sound signal, the state variable, and the additional variable can be acquired in the repair shop and transmitted to the data analysis center 80.

The correspondence between the items in the first embodiment and the items described in the above "summary of the invention" is as follows. The storage device corresponds to the storage device 83. The execution devices correspond to the CPUs 52, 62, 82 and ROMs 54, 64, 84 in fig. 1, the CPUs 52, 82, 102 and the ROMs 54, 84, 104 in fig. 11. The sound signal acquisition process corresponds to the processes of S26 to S30. The state variable acquisition process corresponds to the process of S12. The determination processing corresponds to the processing of S54 to S62. The instruction processing corresponds to the processing of S76, S84. The rotary machine corresponds to the internal combustion engine 12, the first motor generator 14, and the second motor generator 16. The intensities of the predetermined frequency components correspond to the sound pressures Ima1, Imb1, Imc1, and Imd 1. Intensities of frequency components of integral multiples of the predetermined frequency correspond to sound pressures Ima2 to Ima5, Imb2 to Imb5, Imc2 to Imc5, and Imd2 to Imd 5. The duration corresponds to a duration T1. The torques correspond to the required torques Trqed, Trqmg1, Trqmg 2. The variations correspond to variations Δ Trqed, Δ Trqmg1, Δ Trqmg 2. The processing at S12 corresponds to the processing of transmission gear ratio variable Vsft. The rotation speed of the rotary machine corresponds to the rotation speeds NE, Nm1, Nm 2. The control device corresponds to the control device 50. The magnitude of the canceling torque corresponds to the canceling torque Trqcan. The control device corresponds to the control device 50. The magnitude of the propulsion torque corresponds to the propulsion torque Trqpush. The first execution means corresponds to the CPU52 and the ROM 54. The second execution means corresponds to the CPU62 and the ROM 64. The third execution means corresponds to the CPU82 and the ROM 84. The state variable acquisition process corresponds to the process of S12. The state variable transmission processing corresponds to the processing of S14. The state variable reception process corresponds to the process of S24. The terminal-side transmission processing corresponds to the processing of S40. The analysis-side reception process corresponds to the process of S50. The notification acquisition process corresponds to the process of S28. The additional variable acquisition process corresponds to the process of S38, and the additional variables correspond to the window variable Vw and the identification variable Vst. The application program corresponds to the application program 63 a. 15 corresponds to the control device 50.

The present embodiment can be modified and implemented as follows. The present embodiment and the following modifications can be implemented in combination with each other within a range not technically contradictory.

Hereinafter, the sound variables will be described.

The intensities of the frequency components related to meshing are not limited to the intensities of the four frequency components related to meshing a to d. For example, the intensity of the frequency components related to meshing may be 5 or more different components. The intensity of the high-order frequency component among the intensities of the frequency components related to meshing is not limited to the intensity of the first to fourth high-order components. For example, the intensity of the high-order component may be one to three times, or the intensity of the high-order component may be one to five times. In addition, when the intensities of a plurality of types of frequency components related to meshing are used as the feature quantity, for example, different frequency components may be used as high-order components between the frequency components related to meshing, which are different from each other in type, such as one to five times for the frequency component related to meshing of a and one to three times for the frequency component related to b.

The intensity of the frequency component related to meshing is not limited to the intensity of the fundamental frequency component and the intensity of the higher-order component. For example, the intensity of the 1/2 th component may also be included. Further, only the component of the predetermined order may be used, for example, only the intensity of the fundamental frequency component.

The sound variables are not limited to all of the variables of the categories exemplified in the above embodiments. For example, the duration T1 may be eliminated, or the protrusion frequency fpr and the protrusion amount Ipr may be eliminated.

As described in the column of "map data", time-series data of an audio signal itself may be used as an audio variable in the case of deep learning or the like.

Next, the processing for generating the sound variable will be described.

In the above embodiment, the sound variables are generated in the data analysis center 80, but the present invention is not limited thereto. For example, the sound variable may be generated in the mobile terminal 60 and transmitted from the mobile terminal 60 to the data analysis center 80.

Next, the state variables are explained.

The state variables are not limited to all of the variables exemplified in the above embodiments. For example, the speed ratio variable Vsft and the rotational speeds NE, Nm1, Nm2 are included in the state variables, but the vehicle speed SPD may not be included.

It is not necessary that all of the state variables acquired from the control device 50 be input to the map. For example, the speed ratio variable Vsft and the rotational speeds NE, Nm1, Nm2 may be used for calculating the frequency related to engagement, but not as inputs to the map.

It is not necessary for the control device 50 to generate all of the state variables. For example, if the required torques Trqed, Trqmg1, and Trqmg2 are transmitted from the control device 50 in a short cycle, the mobile terminal 60 or the data analysis center 80 may generate the changes Δ Trqed, Δ Trqmg1, and Δ Trqmg2 from the time-series data of the required torques Trqed, Trqmg1, and Trqmg 2.

Next, the establishment of the association of the state variable with the sound signal is explained.

In the above-described embodiment, the set of the sound signal and the state variable corresponding to the predetermined period is associated with each other as data used for determining the generation location of the abnormal sound, but the present invention is not limited thereto. For example, the CPU62 may store data at an arbitrary timing in the time-series data of the sound signal in association with data at a corresponding timing in the time-series data of the state variable. This can be achieved by, for example, setting numerals in parentheses as label variables for identifying different times, setting data in which the time-series data Ds (1), Ds (2), … of the audio signal and the time-series data VS (1), VS (2), … of the state variables are equal to each other as the same timing, or giving time stamps to the data.

Next, the notification process is explained.

It is not necessary to use the set of the sound signal and the state variable in a predetermined period defined in accordance with the timing at which the user sends the signal. For example, input data to the map may be generated by sequentially using the audio signal and group data of the state variables synchronized with the audio signal at a predetermined period. In this case, when the map is formed by the learned model, the teaching data in the case where no abnormal sound is generated may be, for example, a score Sc (1) of "1" and the remaining scores of "0".

Next, the arrangement of the microphones will be described.

In the above embodiment, the sound signal is acquired with the microphone 65 disposed on the head of the driver's seat, but the present invention is not limited to this. For example, the display device may be on an instrument panel. For example, the center portion of the rear seat may be used. And may be disposed in a center console, for example. In each of modifications 1 to 3 shown in fig. 13A to 13C, examples of display on the display section 67 of the mobile terminal 60 are shown.

Next, the main cause determination data is explained.

Fig. 9 illustrates the meaning of whether the abnormal portion cannot be specified or the abnormal portion is indicated, but the present invention is not limited thereto. For example, the determination result may include a determination result indicating that the noise is assumed to be due to aging or the like. This makes it possible to explain the meaning of the perceived sound to the user when it is assumed that the sound is generated to the same extent in the near future even if the sound is repaired, because the sound is caused by natural aging.

Next, the abnormal part candidates will be described.

The candidates of the abnormal portion are not limited to those illustrated in fig. 9.

Next, the mapping data is explained.

The neural network is not limited to a feedforward network in which one layer is an intermediate layer. For example, the intermediate layer may be a network having two or more layers. For example, a convolutional neural network or a recurrent neural network may be used. For example, in the case of deep learning, the feature amount is not limited to the processed sound variable such as the projection amount Ipr, the projection frequency fpr, and the sound pressures Ima1 to Ima5 of the frequency components related to the engagement, and for example, the time-series data of the sound signal itself may be used as the sound variable as the input variable to the neural network.

When the time-series data of the sound signal is input to the neural network, the rotation speeds NE, Nm1, and Nm2 are included in the input to the neural network, and the rotation speeds of the gears of the power split device 10 can be grasped. Therefore, it is possible to determine the abnormal sound caused by the engagement of the power split device 10 by referring to the information of the frequency relating to the engagement of the power split device 10 through the deep learning. Further, since the speed ratio variable Vsft is included in the input, the rotational speed of each gear of the transmission 20 can be grasped, and therefore, the abnormal sound caused by the engagement of the transmission 20 can be specified by deep learning by referring to the information on the frequency relating to the engagement of the transmission 20.

For example, in the case of deep learning, time-series data of the required torques Trqed, Trqmg1, and Trqmg2 may be used as input variables to the neural network without using the characteristic quantities after human processing, such as the torque changes Δ Trqed, Δ Trqmg1, Δ Trqmg2, and the offset torque Trqcan. When the time-series data of the audio signal and the time-series data of the state variable are input to the neural network, it is preferable to designate which input variable the data at the same timing is substituted into by associating the data constituting the time-series data with each other as described in the column "association between the state variable and the audio signal".

The learned model based on the mechanical learning is not limited to the neural network. For example, a support vector machine may be used, and instead of using the coefficients wFjk and wSij as learned data, a support vector may be used as learned data. In other words, a representative group of the groups of feature amounts may also be used as the learned data.

The learned model is not limited to a model that outputs a probability that the candidate of the generation site of the abnormal sound is actually the generation site of the abnormal sound. For example, the recognition model may be a recognition model indicating which candidate is a generation part of the abnormal sound.

The mapping data is not limited to mapping data based on machine learning, and may be data including a generation location of an abnormal sound and time-series data of a corresponding sound signal. In this case, the abnormal portion may be specified by pattern matching with the time-series data of the actual audio signal.

Next, a system for specifying a generation location of an abnormal sound will be described.

The system for specifying the generation location of the abnormal sound is not limited to the system for specifying the generation location of the abnormal sound illustrated in fig. 1 and 11. For example, in the processing of fig. 11, the terminal of the repair site RF may be provided with the map data 83a or the like, and the process of identifying the abnormal sound generation site in the repair site RF may be performed.

Modifications other than the above are described below.

The vehicle to be determined as the abnormal sound is not limited to the series/parallel hybrid vehicle, and may be a series hybrid vehicle or a parallel hybrid vehicle. However, the present invention is not limited to this, and a vehicle equipped with only an internal combustion engine, an electric vehicle, or a fuel cell vehicle may be used as the thrust generator of the vehicle.

The canceling torque Trqcan is not limited to be distributed to the first motor generator 14 and the second motor generator 16, and may be generated by a predetermined one of them.