DEVICE FOR A VEHICLE WITH A POWER TRANSMISSION DEVICE

Abstract

Device for a vehicle with a power transmission device (20), wherein the power transmission device comprises an input pulley (33) into which a torque can be applied, an output pulley (35) capable of delivering a torque in the direction of the drive wheels (50) of the vehicle, and an endless rotating element (37) wound around both the input pulley and the output pulley, wherein the device comprises: a memory (63) designed to store characteristic map data comprising data that form a characteristic map representing a relationship between an input variable and an output variable, wherein the data are learned by machine learning, where the input variable is at least one of the following: i) Input speed-related data (RDNinp) based on chronological data about input speeds (Ninp), which are speeds of the input pulley; and ii) output speed-related data (RDNoutp) based on chronological data about output speeds (Noutp), which are speeds of the output pulley, where the output variable indicates whether an anomaly is caused in the endlessly circulating element; and a processor (61) designed to to capture the input variable, to capture the output variable that matches the input variable using the characteristic map data, and to determine, based on the output variable, whether an anomaly is caused in the endless circulating element, where: the input variable includes the data related to the input speed (RDNinp); and the processor (61) is designed to than the chronological data on the input speeds to capture a multitude of the input speeds (Ninp) that are captured in each acquisition cycle in a predefined measurement period, To derive frequency characteristics of the chronological data via the input speeds by performing a Fast Fourier Transform on the chronological data, and to generate the input speed-related data based on the frequency characteristics of the chronological data, wherein the processor (61) is designed to do the following: to calculate an average of the input speeds (Ninp) during a recording period for the chronological data on the input speeds; to detect an amplitude of a primary rotational frequency of the endless rotating element (37) based on the mean of the input rotational speeds; to normalize the frequency characteristics of the input rotational speeds with the amplitude of the primary rotational frequency, wherein the frequency characteristics are derived by performing the Fast Fourier Transform; and to generate the input speed-related data (RDNinp) based on the normalized frequency characteristics of the input speeds.

Description

BACKGROUND OF THE INVENTION 1. Field of the invention

[0001] The present invention relates to a device for a vehicle with a power transmission device or a transmission cable. 2. Explanation of the state of the art

[0002] Japanese patent application No. 9-65501 (JP 9-65501A) describes an example of a vehicle comprising a motor generator, a continuously variable transmission (CVT), and a temperature sensor that detects the temperature of the oil circulating in the motor generator and the CVT. According to JP 9-65501A, an anomaly in the motor generator or the CVT is caused when a temperature sensor reading is equal to or greater than a threshold value. Further relevant prior art can be found in US 2017 / 0299050A1, US 2017 / 0242076A1, US 2019 / 0040949A1, and US 2020 / 0273786A1. The closest applicable patent, US 2017 0 263 786 A1, teaches a device for a vehicle with a power transmission device, wherein the power transmission device comprises an input pulley into which a torque can be applied, an output pulley capable ofto deliver a torque in the direction of the vehicle's drive wheels, and comprises an endless rotating element wound around both the input pulley and the output pulley, the device comprising a memory configured to store characteristic map data comprising data that form a characteristic map representing a relationship between an input variable and an output variable, the input variable being at least one of the following: i) input speed-related data based on chronological data of input speeds that are speeds of the input pulley; and ii) output speed-related data based on chronological data of output speeds that are speeds of the output pulley, the output variable indicating whether an anomaly is caused in the endless rotating element; and a processor configured toto capture the input variable, and to capture the output variable that matches the input variable, based on the characteristic map data. BRIEF EXPLANATION OF THE INVENTION

[0003] In a continuously variable transmission (CVT) that includes an endless rotating element such as a belt, the occurrence of an anomaly in the endless rotating element is preferably detectable. The temperature of the oil circulating in the CVT can vary between the time an anomaly occurs in the endless rotating element and the time an anomaly occurs in a component other than the endless rotating element within the CVT. Therefore, it is not possible to determine whether an anomaly occurs in the endless rotating element by measuring the temperature of the oil circulating in the CVT.

[0004] One aspect of the present invention relates to a device for a vehicle having the features listed in claim 1. Another aspect of the present invention relates to a device for a vehicle having the features listed in claim 2. Advantageous embodiments are the subject of the dependent claims.

[0005] When the output torque is transmitted from the power source, through the power transmission device, to the drive wheels, the endless rotating element is turned while being held in the position where it is wound around both the input and output pulleys. Consequently, the torque introduced by the power source into the power transmission device is delivered to the drive wheels. Under conditions in which the power transmission device operates in this manner, the rotational speeds of the input and output pulleys will be periodically set into oscillation if no anomaly is caused in the endless rotating element. If an anomaly is caused in the endless rotating element, however, the rotational speeds of the input and output pulleys will be superimposed with vibrational components resulting from the anomaly caused in the endless rotating element.Therefore, it is possible to estimate whether an anomaly is caused in the endless rotating element by analyzing transitions in the rotational speed of the input and / or output pulley.

[0006] In the setup described above, the memory stores characteristic map data that forms a map representing the relationship between an input variable, comprising data related to the input speed and / or data related to the output speed, and an output variable indicating whether an anomaly is caused in the endless rotating element. Based on the output variable, which is derived from the acquired input variable, it is then determined whether an anomaly is caused in the endless rotating element when the power transmission device is operating. Thus, with the setup described above, it is possible to determine whether an anomaly is caused in the endless rotating element of the power transmission device.

[0007] In the aspect mentioned above, the input variable can contain the data related to the input speed; and the processor can be configured to acquire, as the chronological data about the input speeds, a plurality of the input speeds that are acquired in each acquisition cycle in a predetermined measurement period, and to generate the input speed-related data based on data that indicate a distribution of a size of numerical values ​​of the plurality of input speeds contained in the chronological data about the input speeds.

[0008] Consider a case where chronological data on input speeds are used as input variables for the characteristic map. In this case, it is also possible to determine, based on the output variable provided by the characteristic map, whether an anomaly is caused in the continuously rotating element. When such a determination is performed, the accuracy is higher if the number of input speeds included in the chronological data is greater, but the amount of data on the characteristic map's input variables is also greater. With the setup described above, the input speed-related data is generated based on data showing the distribution of the magnitude of the numerical values ​​of the multitude of input speeds included in the chronological data.The data related to the input speed is used as a single input variable for the characteristic map. In this case, the amount of data is not very large, even if the number of input speeds is large, as evidenced by the chronological data showing the distribution of the numerical values ​​across the multitude of input speeds. Therefore, it is possible to suppress an increase in the amount of input variable data for the characteristic map without reducing the accuracy of the determination.

[0009] In the aspect explained above, the input variable can include the data related to the input speed; and the processor can be configured to acquire, as the chronological data about the input speeds, a plurality of the input speeds acquired in each acquisition cycle in a predetermined measurement time interval, to normalize the plurality of input speeds contained in the chronological data about the input speeds, and to generate data as the input speed-related data indicating a distribution of a magnitude of numerical values ​​of a plurality of normalized input speeds obtained by normalizing the input speeds.

[0010] There is a difference between the chronological data of the input speeds in a case where an anomaly is caused in the endless rotating element and the chronological data of the input speeds in a case where no anomaly is caused. However, the degree of this difference can vary between a case where the input speed is high and a case where the input speed is low. In a case where the chronological data of the input speeds is used as an input variable for the characteristic map, the accuracy of the determination tends to be lower when the degree of difference is relatively large and lower when the degree of difference is relatively small. In other words, the accuracy of the determination can fluctuate depending on the magnitude of the input speeds at that time.

[0011] In the setup described above, the multitude of input speeds contained in the chronological data on input speeds are normalized. That is, a multitude of normalized input speeds are generated. The input speed-related data are generated based on data that specify the distribution of the magnitude of the numerical values ​​of the normalized input speeds. Therefore, the degree of difference between the input speed-related data for a case where an anomaly is caused in the endless rotating element and the input speed-related data for a case where no anomaly is caused is not so different between a case where the input speeds are high and a case where the input speeds are low.This makes it possible to suppress fluctuations in the determination accuracy due to variations in the input speeds by using the data related to the input speed as an input variable for the characteristic map.

[0012] Furthermore, data related to input speeds are used, indicating the distribution of the magnitude of the numerical values ​​of the normalized input speeds, rather than the chronological data of the normalized input speeds. Therefore, it is possible to suppress an increase in the amount of input variable data for the characteristic map without reducing the accuracy of the determination.

[0013] In the aspect explained above, the output variable can contain the data related to the output speed; and the processor can be designed to capture, as the chronological data about the output speeds, a plurality of the output speeds captured in each capture cycle in a predefined measurement period, and to generate the output speed-related data based on data showing a distribution of a size of numerical values ​​of the plurality of output speeds contained in the chronological data about the output speeds.

[0014] Consider a case where the chronological data of the output speeds are used as input variables for the characteristic map. In this case, it is also possible to determine, based on the output variable provided by the characteristic map, whether an anomaly is caused in the continuously rotating element. When such a determination is performed, the accuracy is higher because the number of output speeds included in the chronological data is greater, and the amount of data about the characteristic map's input variables is also greater. With the setup described above, the output speed-related data is generated based on data showing the distribution of the magnitude of the numerical values ​​of the multitude of output speeds contained in the chronological data.The data related to the output speed is used as an input variable for the characteristic map. In this case, even if the number of output speeds included in the chronological data is large, the amount of data representing the distribution of the numerical values ​​of the output speeds is not so large. Therefore, it is possible to suppress an increase in the amount of data via the input variables for the characteristic map without reducing the accuracy of the determination.

[0015] In the aspect mentioned above, the output variable can contain the data related to the output speed; and the processor can be configured to capture, as the chronological data about the output speeds, a plurality of the output speeds captured in each capture cycle in a predetermined measurement period, to normalize the plurality of output speeds contained in the chronological data about the output speeds, and to generate data as the output speed-related data that indicate a distribution of a magnitude of numerical values ​​of a plurality of normalized output speeds obtained by normalizing the output speeds.

[0016] There is a difference between the chronological data of the output speeds in a case where an anomaly is caused in the endless rotating element and the chronological data of the output speeds in a case where no anomaly is caused. However, the degree of difference can vary between a case where the output speed is high and a case where the output speed is low. If the chronological data of the output speeds is used as an input variable for the characteristic map, the accuracy of the determination tends to be lower when the degree of difference is relatively large and lower when the degree of difference is relatively small. In other words, the accuracy of the determination can fluctuate depending on the magnitude of the output speeds at that time.

[0017] In the setup described above, the output speeds contained in the chronological data are normalized. That is, a multitude of normalized output speeds are generated. The output speed-related data are generated based on data specifying the distribution of the magnitude of the numerical values ​​of the normalized output speeds. Therefore, the degree of difference between the output speed-related data for a case where an anomaly is caused in the endless rotating element and the output speed-related data for a case where no anomaly is caused is not so different between a case where the output speeds are high and a case where the output speeds are low.This makes it possible to suppress fluctuations in the determination accuracy due to changes in the output speeds by using the data related to the output speed as an input variable for the characteristic map.

[0018] Furthermore, data related to the output speed are used, indicating the distribution of the magnitude of the numerical values ​​of the normalized output speeds, rather than the chronological data of the normalized output speeds. Therefore, it is possible to suppress an increase in the amount of data for the input variables of the characteristic map without reducing the accuracy of the determination.

[0019] In the aspect explained above, the input variable can include the data related to the input speed; and the processor can be configured to acquire, as the chronological data about the input speeds, a plurality of the input speeds acquired in each acquisition cycle in a predefined measurement period, to derive frequency characteristics of the chronological data about the input speeds by performing a Fast Fourier Transform of the chronological data, and to generate the data related to the input speed based on the frequency characteristics of the chronological data.

[0020] With the setup described above, the frequency characteristics of the chronological data are derived from the input rotational speeds by performing a Fast Fourier Transform on such chronological data. Then, the input-speed-referred data are generated based on these frequency characteristics. There can be a difference between the input-speed-referred data in a case where an anomaly is introduced in the endless rotating element, which can vary the frequency characteristics of the chronological data, and the input-speed-referred data generated when no such anomaly is introduced.Therefore, it is possible to determine that an anomaly is caused by using the output variable provided by the map if an anomaly that can characterize the frequency characteristics is caused in the endless rotating element, provided that the input speed-related data generated based on the frequency characteristics is received as an input variable for the map.

[0021] In the aspect described above, the processor can be configured to do the following: calculate an average of the input speeds during a capture period for the chronological data on the input speeds; capture an amplitude of a primary rotational frequency of the endless rotating element based on the average of the input speeds; normalize the frequency characteristics of the input speeds with the amplitude of the primary rotational frequency, the frequency characteristics being derived by performing the Fast Fourier Transform; and generate the input speed-related data based on the normalized frequency characteristics of the input speeds.

[0022] Even if an anomaly that can characterize the frequency characteristics of the chronological data over the input speeds is caused in the endless rotating element, the magnitude of the characteristic value of the anomaly in the frequency characteristics can be varied according to the magnitude of the input speeds. There can be variations in the accuracy of determining the occurrence of an anomaly between a case where the magnitude of the characteristic value of the anomaly in the frequency characteristics is large and a case where the magnitude of the characteristic value of the anomaly in the frequency characteristics is small.

[0023] With the setup described above, the mean value of the input speeds during a recording period is calculated for the chronological data on the input speeds. Based on this mean value, the amplitude of the primary rotational frequency of the endless rotating element is determined. The frequency characteristic of the input speeds, derived by a Fast Fourier Transform, is normalized by the amplitude of the primary rotational frequency. The magnitude of the characteristic value due to anomaly appearing in the frequency characteristics of the input speeds thus normalized does not differ significantly between a case with high input speeds and a case with low input speeds.This makes it possible to suppress fluctuations in the determination accuracy due to variations in the input speeds by using the input speed-related data as the input variable for the characteristic map, which are generated on the basis of the frequency characteristics of the normalized input speeds.

[0024] In the aspect explained above, the processor can be designed to generate the input speed-related data by evenly dividing the frequency bands in which the data derived by performing the Fast Fourier Transform for the chronological data over the input speeds are distributed into a predetermined number of sub-frequency bands and averaging the data for each of the sub-frequency bands.

[0025] With the setup described above, the frequency band in which the data derived by performing a Fast Fourier Transform on the chronological data of the input speeds are distributed is divided into a predetermined number of frequency bands. The input speed-related data is generated based on data that is averaged for each frequency band as described above. It is possible to suppress an increase in the number of data points for the input variables of the characteristic map by using the input speed-related data, generated as described above by averaging data for each frequency band, as the input variables for the characteristic map.

[0026] In the aspect explained above, the input variable can include the output speed data; and the processor can be configured to acquire, as the chronological data about the output speeds, a plurality of the output speeds acquired in each acquisition cycle in a predefined measurement period, to derive frequency characteristics of the chronological data about the output speeds by performing a Fast Fourier Transform of the chronological data, and to generate the output speed data based on the frequency characteristics of the chronological data.

[0027] With the setup described above, the frequency characteristics of the chronological data are derived from the output rotational speeds by performing a Fast Fourier Transform on such chronological data. Then, the output rotational speed-related data are generated based on these frequency characteristics. There can be a difference between the output rotational speed-related data for a case where an anomaly, in which the frequency characteristics of the chronological data can vary, is caused in the endless rotating element, and the output rotational speed-related data based on the frequency characteristics of the chronological data for a case where no such anomaly is caused.Therefore, it is possible to determine that an anomaly is caused when an anomaly that can characterize the frequency characteristics is caused in the endlessly rotating element by using the output variable output by the vehicle by using as an input variable for the map the data related to the output speed, which is based on the frequency characteristics of the chronological data.

[0028] In the aspect described above, the processor can be configured to do the following: calculate an average of the output speeds during a capture period for the chronological data on the output speeds; capture an amplitude of a primary rotational frequency of the endless rotating element based on the average of the output speeds; normalize the frequency characteristics of the output speeds with the amplitude of the primary rotational frequency, the frequency characteristic being derived by performing the Fast Fourier Transform; and generate the output speed-related data based on the normalized frequency characteristics of the output speeds.

[0029] Even if an anomaly that can characterize the frequency characteristics of the chronological data over the output speeds is caused in the endless rotating element, the magnitude of the characteristic value of the anomaly in the frequency characteristics can be varied according to the magnitude of the output speeds. There can be variations in the accuracy of anomaly detection between a case where the characteristic value of the anomaly in the frequency characteristics is large and a case where the characteristic value of the anomaly in the frequency characteristics is small.

[0030] Using the setup described above, the mean output speeds during a recording period are calculated for the chronological output speed data. Based on this mean, the amplitude of the primary rotational frequency of the endless rotating element is determined. The frequency characteristics of the output speeds, derived by a Fast Fourier Transform, are normalized by the amplitude of the primary rotational frequency. The magnitude of the characteristic value due to anomaly resulting in these normalized primary frequency characteristics of the output speeds does not differ significantly between cases with high and low output speeds.This makes it possible to suppress fluctuations in the determination accuracy due to variations in the output speeds by using the output speed-related data as the input variable for the characteristic map, which are generated on the basis of the frequency characteristics of the normalized output speed.

[0031] In the aspect explained above, the processor can be designed to generate the output speed-related data by dividing the frequency bands in which the data derived by performing the Fast Fourier Transform for the chronological data about the output speeds are distributed into a predetermined number of sub-frequency bands and averaging the data for each of the sub-frequency bands.

[0032] With the setup described above, the frequency band in which the data derived by performing a Fast Fourier Transform on the chronological data of the output speeds are distributed is divided into a predefined number of frequency bands. The output speed-related data is generated based on data that is averaged for each frequency band as described above. It is possible to suppress an increase in the number of data points for the input variables of the characteristic map by using the output speed-related data, generated by averaging data for each frequency band as described above, as the input variables for the characteristic map.

[0033] In the aspect described above, the input variable can include at least either a torque applied to the input pulley or a temperature of oil circulating in the power transmission device. The magnitude and oscillation period of the speeds of the input and output pulleys can differ between a case where the torque applied to the input pulley is large and a case where such torque is small. Therefore, in the setup described above, the torque applied to the input pulley is used as an input variable for the characteristic map. That is, the output variable of the characteristic map has a value that is determined taking into account the torque applied to the input pulley. Thus, the accuracy of the determination can be increased by making the determination based on the output variable.The viscosity of the oil is lower when the oil temperature is higher. The degree of slip of the endless rotating element with respect to the input and output pulleys tends to be greater when the oil viscosity is lower. The vibration mode of the input and output pulley speeds can change when the degree of slip differs. In the setup described above, the oil temperature is therefore used as an input variable for the characteristic map. That is, the output variable provided by the characteristic map has a value determined taking the oil viscosity into account. Thus, the accuracy of the determination can be increased by making the determination based on the output variable.

[0034] In the aspect described above, the memory can store an index indicating a shape of the endless circulating element for each power transmission device, and / or an index indicating a shape of a component of the endless circulating element for each power transmission device; and the input variable can include the index stored in the memory.

[0035] Vibration components due to the shape of the endless rotating element and the shape of a component of the endless rotating element can be superimposed on the rotational speeds of the input and output pulleys. Such vibration components are not attributable to an anomaly caused by the endless rotating element itself. Therefore, in the setup described above, the memory stores an index indicating the shape of the endless rotating element and / or an index indicating the shape of a component of the endless rotating element. The index stored in the memory is used as an input variable for the characteristic map. That is, the output variable provided by the characteristic map has a value determined by considering the shape of the endless rotating element and / or the shape of the basic part of the endless rotating element.Therefore, the accuracy of the determination can be increased by making the determination based on the output variable.

[0036] In the aspect explained above, the input variable can include a force with which the input pulley holds the endless rotating element, and / or a force with which the output pulley holds the endless rotating element.

[0037] The degree of slippage induced between the input pulley and the endless rotating element can be varied depending on the force with which the input pulley holds the endless rotating element. Furthermore, the degree of slippage induced between the output pulley and the endless rotating element can also be varied depending on the force with which the output pulley holds the endless rotating element. The oscillation mode of the rotational speeds of such pulleys can be changed by varying the degree of slippage induced between these pulleys and the endless rotating element. Thus, in the setup described above, a force with which the input pulley holds the endless rotating element and / or a force with which the output pulley holds the endless rotating element is used as an input variable for the characteristic map.This means that the output variable of the characteristic map has a value determined by considering the force with which the input pulley holds the endless rotating element and / or the force with which the output pulley holds the endless rotating element. Therefore, the accuracy of the determination can be increased by basing the determination on the output variable.

[0038] In the aspect explained above, the memory can store an index indicating the magnitude of a backlash or play in one direction of rotation of the endless rotating element for each power transmission device; and the input variable can contain the index indicating the magnitude of the backlash and stored in the memory.

[0039] If the magnitude of the backlash in the direction of rotation of the endless rotating element is large, vibrations corresponding to the magnitude of the backlash tend to be superimposed on the rotation of the input and output pulleys as the endless rotating element rotates. Therefore, in the setup described above, the memory stores an index indicating the magnitude of the backlash in the direction of rotation of the endless rotating element. This index, which indicates the magnitude of the backlash and is stored in the memory, is used as an input variable for the characteristic map. That is, the output variable of the characteristic map has a value that is determined taking the backlash into account. The accuracy of this determination can be increased by performing the determination based on the output variable.

[0040] In the aspect described above, the input variable can include a reading from a vehicle-mounted accelerometer, a reading from a noise sensor installed in the vehicle's engine compartment, the vehicle's braking force, a ratio between the speed of the input pulley and the speed of the output pulley, and / or an index indicating the degree of temporal variation in the properties of the endless rotating element. If an anomaly occurs in the endless rotating element, vibrations due to the anomaly can be transmitted to the vehicle body, and a reading from the vehicle's internal accelerometer can be altered. Therefore, in the setup described above, the reading from the accelerometer is used as the input variable for the characteristic map.This means that the output variable of the characteristic map has a value determined by taking into account the vibrations of the vehicle body. Therefore, the accuracy of the determination can be increased by basing it on the output variable. If an anomaly is caused in the endless rotating element, and a vibration due to this anomaly is superimposed on the rotational speeds of the input and output pulleys, abnormal noise can be generated in the power transmission device. Such abnormal noise generated in the power transmission device can be detected by the noise sensor installed in the engine compartment. Therefore, in the setup described above, the noise sensor's detection value is used as an input variable for the characteristic map.This means that the output variable of the characteristic map has a value determined by taking into account the sound detected by the sound sensor. Therefore, the accuracy of the determination can be increased by basing it on the output variable. When the vehicle applies braking force, the drive wheels decelerate, thus changing the rotational speeds of the input and output pulleys. In this case, vibration components can be superimposed on these rotational speeds, depending on the magnitude of the braking force or the rate of its increase. In the setup described above, the vehicle's braking force is therefore used as an input variable for the characteristic map. That is, the output variable of the characteristic map has a value determined by taking into account the vehicle's braking force. Therefore, the accuracy of the determination can be increased by basing it on the output variable.If an anomaly is caused in the endless rotating element, and therefore the rotational speeds of the input and output pulleys are superimposed with vibrations, the ratio between the input and output pulley speeds can differ between cases where no anomaly is present and cases where one is. Therefore, in the setup described above, the rotational speed ratio is used as an input variable for the characteristic map. That is, the output variable of the characteristic map has a value determined by considering the rotational speed ratio. Thus, the accuracy of the determination can be increased by basing it on the output variable.Even if no anomaly is caused in the endless rotating element, the magnitude and period of the oscillation of the speeds of the input and output pulleys can change if the degree of temporal variation in the characteristics is relatively small, and if the degree of such temporal variation is relatively large. Therefore, in the setup described above, the index indicating the degree of temporal variation in the characteristics is used as an input variable for the characteristic map. That is, the output variable of the characteristic map has a value determined by considering the degree of temporal variation in the characteristics. Therefore, the accuracy of the determination can be increased by making the determination based on the output variable.

[0041] In the aspect explained above, the input variable can include data related to the output speed; and the output speeds can be calculated based on the speeds of the drive wheels. There is a relationship between the speed of the output pulley and the speed of the drive wheels. Therefore, the value of the output pulley speed can be calculated based on the speed of the drive wheels. Data based on transitions of the calculated speed value can be considered as data related to the output speed.

[0042] In the aspect explained above, the processor can be designed to determine, based on the output variable, whether a component of the endless circulating element is damaged.

[0043] In a case where a component of the endless rotating element is damaged, vibration is caused when the damaged section contacts the input pulley or when the damaged section contacts the output pulley. Such vibrations manifest as fluctuations in the rotational speeds of the input and output pulleys. Therefore, with the setup described above, it is possible to determine that an anomaly in the endless rotating element is caused when the component is damaged.

[0044] In the aspect explained above, the processor can be configured to do the following: to detect the output variable matching the input variable based on the characteristic map data; and, based on the output variable, to determine whether a vibration is generated in the endless rotating element that causes a part other than the power transmission device in the vehicle to vibrate.

[0045] With the setup described above, it is possible to determine that the other part in the vehicle resonates due to vibrations of the endless rotating element.

[0046] In the aspect explained above, the processor can be configured to do the following: to detect the output variable according to the input variable based on the characteristic map data; and to determine, based on the output variable, whether a resonance is generated in the endlessly circulating element.

[0047] With the setup described above, it is possible to determine that the endless rotating element is in resonance due to the operation of the other onboard device. BRIEF EXPLANATION OF THE FIGURES

[0048] Features, advantages and the technical and industrial significance of exemplary embodiments of the invention are described below with reference to the accompanying figures, in which the same symbols denote the same elements, and in which: Fig. 1 shows a control device and a drive system of a vehicle which is controlled by the control device according to a first embodiment; Fig. 2 a schematic view showing how a belt is held between pulleys; Fig. 3 a schematic view illustrating part of the belt; Fig. 4 illustrates the first half of a flowchart that demonstrates a sequence of processes performed by the control device; Fig. 5 shows the second half of the flowchart, which illustrates the sequence of operations performed by the control device; Fig. 6 is a time graph that illustrates transitions in input speeds or output speeds; Fig. 7 is a time graph that illustrates transitions in input speeds or output speeds; Fig. 8 is a graph that presents data related to the input speed or data related to the output speed in a histogram; Fig. 9 is a graph that presents the data related to the input speed or the data related to the output speed in a histogram; Fig. 10 is a flowchart illustrating part of a sequence of processes performed by a control device according to a second embodiment; and Fig. 11 is a flowchart illustrating part of a sequence of processes performed by a control device according to a third embodiment. DETAILED EXPLANATION OF THE EXECUTION FORMS First embodiment

[0049] An anomaly detection device for a power transmission device according to a first embodiment is described below with reference to the Fig. 1 to 9 described.

[0050] As in Fig. As shown in Figure 1, a vehicle VC comprises an internal combustion engine 10, a transmission device 20, and drive wheels 50. A torque converter 21 of the speed-changing device 20 is coupled to a crankshaft 11 of the internal combustion engine 10. An input shaft 31 of a speed-changing mechanism 30 is coupled to the torque converter 21. A plurality of drive wheels 50 are coupled via a differential (not shown) to an output shaft 32 of the transmission mechanism 30.

[0051] The transmission mechanism 30 comprises an input pulley 33, an output pulley 35, and a belt 37 wound around the input pulley 33 and the output pulley 35. The output torque of the internal combustion engine 10 is transmitted to the input pulley 33 via the torque converter 21. The output pulley 35 transmits the torque to the drive wheels 50 via the output shaft 32.

[0052] The input pulley 33 comprises a stationary pulley or half-pulley 33a, to which the input shaft 31 is coupled, and a movable pulley or half-pulley 33b (hereinafter referred to as "pulley" 33a and 33b, respectively). That is, the stationary pulley 33a and the movable pulley 33b hold the belt 37 between them. The movable pulley 33b can move away from and towards the stationary pulley 33a. This movement of the movable pulley 33b is achieved by the drive of an input actuator 34. The input actuator 34 can be electrically or hydraulically driven.

[0053] The output pulley 35 comprises a stationary pulley 35a, to which the output shaft 32 is coupled, and a movable pulley 35b. That is, the stationary pulley 35a and the movable pulley 35b hold the belt 37 between them. The movable pulley 35b can move away from and towards the stationary pulley 35a. This movement of the movable pulley 35b is achieved by driving an output actuator 36. The output actuator 36 can be electrically or hydraulically driven.

[0054] The transmission ratio of the transmission mechanism 30 is controlled by adjusting the position of the movable disc 33b relative to the fixed disc 33a of the input pulley 33 and the position of the movable disc 35b relative to the fixed disc 35a of the output pulley 35.

[0055] As in the Fig. 2 and Fig. As shown in Figure 3, the belt 37 comprises endless rings 371 and a large number of elements 372 that lie on the rings 371. The elements 372 are arranged along the direction of rotation of the belt 37.

[0056] One in Fig. The control device 60 shown in Figure 1 controls the internal combustion engine 10 and operates various operating sections of the internal combustion engine 10 in order to control the torque, the proportions of components in the exhaust gas, etc., as control variables. The control device 60 also controls the speed-changing device 20 and actuates the input actuator 34 and the output actuator 36.

[0057] To control the aforementioned control variables, the control device 60 refers to an output signal Scr from a crankshaft angle sensor 101, an output signal Sinp from a rotary angle sensor 102 for the input shaft, which detects the rotation angle of the input shaft 31, and an output signal Soutp from a rotary angle sensor 103 for the output shaft, which detects the rotation angle of the output shaft 32. Additionally, the control device 60 refers to an oil temperature value Toil, which represents the oil temperature detected by an oil temperature sensor 104, a wheel speed VW, which is the rotational speed of the drive wheels 50 and is detected by a wheel speed sensor 105, a vehicle acceleration G, which represents the acceleration of the vehicle VC and is detected by an acceleration sensor 106, and a sound level value Snd, which represents the loudness of a noise and is detected by a noise sensor 107.

[0058] The oil temperature sensor 104 detects the temperature of the oil circulating in the speed-changing mechanism 30. The noise sensor 107 is installed, for example, in the engine compartment of vehicle VC. The sound detection value Snd therefore represents the loudness of a noise in the engine compartment.

[0059] The control device 60 comprises a central processing unit (CPU) 61, a read-only memory (ROM) 62, a memory 63, which is an electrically rewritable non-volatile memory, and a peripheral circuit 64. These components can be interconnected via a local area network 65. The peripheral circuit 64 includes a circuit that generates a clock signal which synchronizes the internal operation, a power supply circuit, a reset circuit, etc. The control device 60 controls the various control variables by having the CPU 61 execute a program stored in the ROM 62.

[0060] Memory 63 stores a variety of characteristic map data DM1, DM2, and DM3. The characteristic map data DM1, DM2, and DM3 comprise data that prescribes a characteristic map which, upon receiving input from various input variables (explained later), outputs a variable that matches the input variable, with the data being learned through machine learning.

[0061] Memory 63 also stores a belt shape index MVs, which is an index that specifies the shape of the belt 37 for each speed-changing device 20, and an element shape index MVEs, which is an index that specifies the shape of the elements 372 for each speed-changing device 20. The belt shape index MVs and the element shape index MVEs are values ​​that are measured, for example, at the delivery of the vehicle VC or during inspection at the delivery stage of the speed-changing device 20. Examples of the belt shape index MVs are the length in the direction of rotation of the belt 37 and the error between an actual measured value and the target value of the length in the direction of rotation of the belt 37. Examples of the element shape index MVEs are the mean of the measured thicknesses of all elements 372 and the mean of the measured widths of all elements 372.Other examples of the element shape index MVEs include an error between the mean of the measured values ​​of the thickness of all elements 372 and the target value of the thickness of elements 372, and an error between the mean of the measured values ​​of the width of all elements 372 and an actual measured value of the width of elements 372.

[0062] Memory 63 also stores a kickback index MVRa, which is an index indicating the amount of kickback of the elements 372 in the direction of rotation for each speed-changing device 20. The kickback index MVRa is a value that is measured, for example, during the delivery of the vehicle VC or during inspection of the speed-changing device 20 in its delivery state. Examples of the kickback index MVRa include the mean of the distances between adjacent elements 372 in the direction of rotation.

[0063] During operation of the speed-changing device 20, vibrations are occasionally generated in the vehicle VC. Examples of factors that can trigger vibration in the vehicle VC due to the operation of the speed-changing device 20 are: - Damage to a component of the belt 37 - Resonance generated due to the operation of the speed-changing device 20 in other vehicle-internal devices than the speed-changing device 20, or - Resonance generated in the speed-changing device 30 due to the operation of other vehicle-internal devices besides the speed-changing device 20.

[0064] For example, if one of the elements 372 is damaged, a vibration is generated in the input pulley 33 every time the damaged element 372 contacts it. As a result, the vibration due to the vibration of the input pulley 33 is superimposed on the input speeds Ninp, which are the measured speed values ​​of the input pulley 33. Furthermore, a vibration is generated in the output pulley 35 every time the damaged element 372 contacts it. As a result, the vibration due to the vibration of the output pulley 35 is superimposed on the output speeds Noutp, which are the measured speed values ​​of the output pulley 35.

[0065] This means that the transitions between the input speeds Ninp and the output speeds Noutp in a case where such an anomaly is caused by damage to the belt 37 differ from those in a case where no anomaly is caused in the belt 37. In the present embodiment, the control device 60 therefore determines whether an anomaly is caused in the belt 37 based on the transitions or profiles of the input speeds Ninp and the output speeds Noutp. In this case, the control device 60 uses the characteristic map data DM1 stored in memory 63.

[0066] Other vehicle-internal devices are also installed in the engine compartment where the speed-changing device 20 is installed. Resonance occasionally occurs in the belt 37 of the speed-changing device 20, depending on the duration of the oscillation generated in the other vehicle-internal devices. Depending on the oscillation period of the speed-changing device 20, resonances occasionally occur in the other vehicle-internal devices. To suppress the generation of such resonances, the internal combustion engine 10 and the speed-changing device 20 are preferably controlled such that the operating orThe operating point of the translation change mechanism 30 is shifted relative to an operating point of the translation change mechanism 30 at which resonances are generated in the speed change device 20, and to an operating point of the translation change mechanism 30 at which resonances are generated in the other vehicle-internal devices.

[0067] In the present embodiment, the control device 60 therefore determines, based on the transitions between the input speeds Ninp and the output speeds Noutp, whether the operation of the speed-changing device 20 generates a resonance in the other vehicle-internal devices. In this case, the control device 60 uses the characteristic map data DM2 stored in memory 63. Additionally, the control device 60 determines, based on the transitions between the input speeds Ninp and the output speeds Noutp, whether the operation of the other vehicle-internal devices generates a resonance in the belt 37. In this case, the control device 60 uses the characteristic map data DM3 stored in memory 63.

[0068] The sequence of operations performed by the control device 60 to make various determinations as described above are illustrated by the Fig. 4 and Fig. 5 described. The process described in the Fig. 4 and Fig. The sequence of operations shown in Figure 5 is implemented by the CPU 61, which executes a program stored in ROM 62. The sequence of operations is repeated at predetermined time intervals. That is, the CPU 61 restarts the execution of the sequence of operations when the time elapsed since the sequence was previously completed reaches a time that corresponds to the predetermined intervals.

[0069] First, in step S11, the CPU 61 sets a coefficient z to "1". In the next step S13, the CPU 61 determines the current input speed Ninp as the input speed Ninp (z). The CPU 61 also determines the current output speed Noutp as the output speed Noutp (z). In the next step S15, the CPU 61 increases the coefficient z by "1".

[0070] Subsequently, in step S17, the CPU 61 determines whether the coefficient z is greater than a coefficient determination value zTh. In the present embodiment, chronological data on the input speeds Ninp and chronological data on the output speeds Noutp are used to perform the aforementioned determination. The chronological data of the input speeds Ninp comprise a plurality of temporally successive input speeds Ninp. The chronological data of the output speeds Noutp comprise a plurality of temporally successive output speeds Noutp. The coefficient determination value zTh is defined as the criterion for determining whether the number of input speeds Ninp and the number of output speeds Noutp required for the aforementioned determination have been recorded.If the coefficient z is equal to or less than the coefficient determination value zTh (S17: NO), CPU 61 proceeds to step S13. This means that the input speeds Ninp and the output speeds Noutp are continuously recorded. Conversely, if the coefficient z is greater than the coefficient determination value zTh (S17: YES), chronological data of the input speeds Ninp, consisting of "z" input speeds Ninp, and chronological data of the output speeds Noutp, consisting of "z" output speeds Noutp, have been recorded, and therefore CPU 61 proceeds to the next step S19.

[0071] In step S19, the CPU 61 normalizes or standardizes the chronological data for the input speed Ninp. For example, the CPU 61 determines a reference input speed NinpB, which is the largest of the several input speeds Ninp(1), Ninp(2), ..., and Ninp(z) contained in the chronological data for the input speeds Ninp. The CPU 61 then normalizes the input speeds Ninp(1), Ninp(2), ..., and Ninp(z) by dividing them by the reference input speed NinpB. The normalized input speeds Ninp(1), Ninp(2), ..., and Ninp(z) are referred to as normalized input speeds NinpN(1), NinpN(2), ..., and NinpN(z). For example, the normalized input speed NinpN(1) has a value that results from dividing the input speed Ninp(1) by the reference input speed NinpB. Data that include the normalized input speeds NinpN(1), NinpN(2), ...and NinpN (z) are also referred to as “chronological data on the normalized input speeds NinpN”.

[0072] In step S21, the CPU 61 normalizes the chronological data for the output speeds Noutp. For example, the CPU 61 determines a reference output speed NoutpB that is the largest among the multiple output speeds Noutp(1), Noutp(2), ..., and Noutp(z) contained in the chronological data for the output speeds Noutp. The CPU 61 then normalizes the output speeds Noutp(1), Noutp(2), ..., and Noutp(z) by dividing them by the reference output speed NoutpB. The normalized output speeds Noutp(1), Noutp(2), ..., and Noutp(z) are referred to as normalized output speeds NoutpN(1), NoutpN(2), ..., and NoutpN(z). For example, the normalized output speed NoutpN (1) has a value that results from dividing the output speed Noutp (1) by the reference output speed NoutpB.Data comprising the normalized output speeds NoutpN (1), NoutpN (2), ..., and NoutpN (z) are also referred to as "chronological data on the normalized output speeds NoutpN".

[0073] In the next step S23, the CPU 61 generates input speed-related data RDNinp based on the chronological data of the normalized input speeds NinpN. In the present embodiment, the normalized input speeds NinpN(1), NinpN(2), ..., and NinpN(z) are greater than "0" and equal to or less than "1". Thus, the range of numerical values ​​from "0" to "1" is divided into several subranges. For example, the range of numerical values ​​from "0" to "1" is divided into subranges of "0,2" each. The CPU 61 counts the number of normalized input speeds NinpN contained in each subrange. For example, if there are 1, 2, ... among the normalized input speeds NinpN(1), NinpN(2), ...Given that there are four normalized input speeds NinpN greater than 0.4 and equal to or less than 0.6, and NinpN(z), the CPU 61 determines the number of normalized input speeds NinpN contained in the subrange from 0.4 to 0.6 as 4. The CPU 61 derives the result of the count for each subrange as the input speed-related data RDNinp. That is, the input speed-related data RDNinp is data that indicates the distribution of the magnitude of the numerical values ​​of the multiple normalized input speeds NinpN(1), NinpN(2), ..., and NinpN(z) contained in the chronological data for the normalized input speeds NinpN.

[0074] In the next step S25, the CPU 61 generates output speed-related data RDNoutp based on the chronological data of the normalized output speeds NoutpN. In the present embodiment, the normalized output speeds NoutpN(1), NoutpN(2), ..., and NoutpN(z) are greater than "0" and equal to or less than "1". Thus, the range of numerical values ​​from "0" to "1" is divided into several subranges. For example, the range of numerical values ​​from "0" to "1" is divided into subranges of "0.2" each. The CPU 61 counts the number of normalized output speeds NoutpN contained in each subrange. For example, if there are 1, 2, ... among the normalized output speeds NoutpN(1), NoutpN(2), ...Given that there are two normalized output speeds NoutpN greater than 0.4 and equal to or less than 0.6, and NoutpN(z), the CPU 61 determines the number of normalized output speeds NoutpN contained in the subrange from 0.4 to 0.6 as 2. The CPU 61 derives the result of the count for each subrange as the output speed-related data RDNoutp. That is, the output speed-related data RDNoutp is data that indicates the distribution of the magnitude of the numerical values ​​of the multiple normalized output speeds NoutpN(1), NoutpN(2), ..., and NoutpN(z) contained in the chronological data on the normalized output speeds NoutpN.

[0075] The Fig. 6 and Fig. Figures 7 each show chronological data for the input speeds Ninp. The in Fig. The chronological data for the input speeds Ninp shown in Figure 6 are data for a case in which the transmission ratio of the transmission mechanism 30 is constant and the vibration described above is not generated. The data shown in Figure 6 are for a case in which the transmission ratio of the transmission mechanism 30 is constant and the vibration described above is not generated. Fig. The chronological data shown in Figure 7 for the input speeds Ninp are data for a case in which the transmission ratio of the transmission mechanism 30 is constant and the vibration described above is superimposed on the input speeds Ninp. Fig. Figure 8 shows the input speed-related data RDNinp, displayed in a histogram. The input speed-related data RDNinp are based on the data in Fig. The chronological data for the input speeds Ninp shown in 6 is generated. Fig. Figure 9 shows the input speed-related data RDNinp, displayed in a histogram. The input speed-related data RDNinp are based on the Fig. The chronological data of the input speeds Ninp shown in 7 is generated. Between the in Fig. 8 shown data related to the input speed RDNinp and the data shown in Fig. In the input speed-related data RDNinp shown in Figure 9, there is a difference in the fluctuations of the counter values ​​Cnt(1) to Cnt(5). This means that it is possible to determine whether an anomaly is being caused based on the fluctuations among the counter values ​​Cnt(1) to Cnt(5) in the input speed-related data RDNinp.

[0076] Will be in the Fig. 6 and Fig. 7. The input speeds Ninp are replaced by the output speeds Noutp. Fig. 6 and Fig. 7 can be considered as chronological data about the output speeds Noutp. In this case, Fig. Figure 8 is considered as a representation of the output speed-related data RDNoutp in a histogram. The output speed-related data RDNoutp are based on the in Fig. The chronological data for the output speeds Noutp shown in 6 is generated. Fig. Figure 9 is used to display the output speed-related data RDNoutp in a histogram. The output speed-related data RDNoutp are based on the Fig. The chronological data for the output speeds Noutp shown in 7 is generated. There is a difference in the fluctuations between the counter values ​​Cnt (1) to Cnt (5) between the in Fig. 8 shown data related to the output speed RDNoutp and the data in Fig. The output speed-related data RDNoutp shown in Figure 9 indicates that it is possible to determine whether an anomaly exists based on the fluctuations between the counter values ​​Cnt (1) to Cnt (5) in the output speed-related data RDNoutp.

[0077] If the process – again with reference to the Fig. 4 and Fig. Once step S25 is complete, CPU 61 acquires the remaining data in the next step S27. This remaining data includes a gear ratio RN, an input torque Trq, the oil temperature value Toil, an input clamping force Pinp, an output clamping force Poutp, the vehicle acceleration G, the noise value Snd, a braking force BPvc of the vehicle VC, and a distance traveled SC of the vehicle VC. CPU 61 also acquires the indices MVs, MVEs, and MVRa stored in memory 63.

[0078] The speed ratio RN is a value obtained by dividing the current output speed Noutp by the current input speed Ninp. The input torque Trq is the torque exerted on the input pulley 33. The input torque Trq can be derived based on the output torque of the internal combustion engine 10 and the torque transmission efficiency of the torque converter 21. The input clamping force Pinp is a clamping force of the input pulley 33 on the belt 37, i.e., a force with which the input pulley 33 holds the belt 37. The input clamping force Pinp is greater the greater the driving force acting from the input actuator 34 on the movable pulley 33b. The output clamping force Poutp is a clamping force of the output pulley 35 on the belt 37, i.e., a force with which the output pulley 35 holds the belt 37.The output clamping force Poutp is greater the greater the driving force acting from the output actuator 36 on the movable pulley 35b. The braking force BPvc is derived based on a required braking force value due to a braking action by the driver, etc. The travel distance SC is an index that indicates the degree of temporal variation in the properties of the belt 37.

[0079] In the next step S29, the CPU 61 inserts the input speed-related data RDNinp generated in step S23, the output speed-related data RDNoutp generated in step S25, and the various data acquired in step S27 into the input variables x(1) to x(22) for the characteristic map specified by the map data. That is, the CPU 61 inserts the counter value Cnt(1) of the input speed-related data RDNinp into the input variable x(1), the counter value Cnt(2) into the input variable x(2), the counter value Cnt(3) into the input variable x(3), the counter value Cnt(4) into the input variable x(4), and the counter value Cnt(5) into the input variable x(5).The CPU 61 also inserts the counter value Cnt (1) of the output speed-related data RDNoutp into the input variable x (6), the counter value Cnt (2) into the input variable x (7), the counter value Cnt (3) into the input variable x (8), the counter value Cnt (4) into the input variable x (9), and the counter value Cnt (5) into the input variable x (10). The CPU 61 also inserts the speed ratio RN into the input variable x (11), the input torque Trq into the input variable x (12), the oil temperature measurement value Toil into the input variable x (13), the input clamping force Pinp into the input variable x (14), and the output clamping force Poutp into the input variable x (15). The CPU 61 also inserts the vehicle acceleration G into the input variable x (16), the noise detection value Snd into the input variable x (17), the braking force BPvc into the input variable x (18) and the travel distance SC into the input variable x (19).The CPU 61 also inserts the belt shape index MVs into the input variable x (20), the element shape index MVEs into the input variable x (21) and the rebound amount index or game size index MVRa into the input variable x (22).

[0080] In the next step S31, CPU 61 sets the coefficient of determination MP to "1", as in Fig. Figure 5 is shown. In the next step, S33, the CPU 61 selects map data matching the coefficient of determination MP from the map data DM1, DM2, and DM3 stored in memory 63. For example, the CPU 61 selects the map data DM1 if the coefficient of determination MP is set to "1", selects the map data DM2 if the coefficient of determination MP is set to "2", and selects the map data DM3 if the coefficient of determination MP is set to "3".

[0081] Then, in step S35, the CPU 61 calculates an output variable Y (MP) by inputting the input variables x (1) to x (22) into the map specified by the selected map data.

[0082] In the present embodiment, the characteristic map is constructed as a fully linked, forward-propagating neural network with a single intermediate layer. The neural network comprises an input coefficient wFjk (j = 0 to n, k = 0 to 22) and an activation function h(x) as an input nonlinear characteristic map, which performs a nonlinear transformation at each output of the input linear characteristic map, which is a linear characteristic map defined by the input coefficient wFjk. In the present embodiment, a hyperbolic tangent "tanh(x)" is given as an example of the activation function h(x).The neural network also includes an output coefficient wSj (j = 0 to n) and an activation function f(x) as an output nonlinear characteristic map that performs a nonlinear transformation at each output of the output linear characteristic map, as specified by the output coefficient wSj. In the present embodiment, a hyperbolic tangent "tanh(x)" is given as an example of the activation function f(x). A value n specifies the dimension of the intermediate layer. In the present embodiment, the value n is less than 22, namely less than the dimension of the input variable x. The input coefficient wFj0 ​​is a bias parameter and is a coefficient of the input variable x(0). The input variable x(0) is defined as 1. The output coefficient wS0 is a bias parameter.

[0083] The map data DM1 is a trained model that was trained on a vehicle with the same specifications as vehicle VC before being installed in vehicle VC. To train the map data DM1, training data is acquired beforehand, comprising teacher data and input data. Specifically, the chronological data for the input speeds Ninp and the chronological data for the output speeds Noutp are acquired while the vehicle is actually being driven. The input speed-related data RDNinp are acquired as input data by performing operations similar to those in steps S19 and S23 on the chronological data for the input speeds Ninp.Similarly, the output speed-related data RDNoutp are acquired as input data by performing operations similar to those in steps S21 and S25 on the chronological output speed data Noutp. At this point, the speed ratio RN, the input torque Trq, the oil temperature sensor value Toil, the input clamping force Pinp, the output clamping force Poutp, the vehicle acceleration G, the noise sensor value Snd, the braking force BPvc, and the distance traveled SC are additionally acquired as input data. Furthermore, information on the occurrence of anomalies, indicating whether an anomaly has been caused in belt 37, is acquired as teacher data. For example, the anomaly occurrence information can be set to "0" if an anomaly has been caused, and the anomaly occurrence information can be set to "1" if no anomaly has been caused.The indices MVs, MVEs and MVRa are also recorded as input data before the vehicle is instructed to drive.

[0084] A variety of training data is generated by moving the vehicle under different conditions. For example, the gear ratio changer 30, in which one of the elements 372 of the belt 37 is intentionally damaged, is mounted on the vehicle, and the vehicle is driven. In this case, various input data can be acquired for a scenario in which an anomaly is caused in the belt 37, and information about the occurrence of anomalies indicating that an anomaly is being caused can be acquired as training data. Meanwhile, the gear ratio changer 30, in which none of the elements 372 of the belt 37 are damaged, is installed in the vehicle, and the vehicle is driven.In this case, various input data can be recorded for a case in which no anomaly is caused in belt 37, and information on the occurrence of anomalies indicating that no anomaly is caused can be recorded as the teacher data.

[0085] The DM1 characteristic map data is trained using a large amount of training data. This means that input variables and an output variable are adjusted so that the error between the output variable, which is generated from the characteristic map using the input data, and the actual information about the occurrence of an anomaly at or below a predefined value converges.

[0086] Similarly, the DM2 map data is a trained model that was trained using a vehicle with the same specifications as vehicle VC before being installed in vehicle VC. To train the DM2 map data, training data is also acquired beforehand, comprising teacher data and input data. That is, various input data are acquired, as explained above, by actually driving the vehicle while the operating point of the speed-changing device 20 is changed. In this case, initial resonance generation information, indicating whether resonance is generated in other vehicle-internal devices besides the speed-changing device 20, is also acquired as teacher data.For example, the first resonance generation information can be set to "0" if resonance is generated in the other vehicle-internal devices, and the first resonance generation information can be set to "1" if no resonance is generated in the other vehicle-internal devices.

[0087] A large amount of training data, including teacher data and input data, is generated by driving the vehicle under various conditions. The DM2 map data is then trained using this training data. This means that input variables and an output variable are adjusted so that the error between the output variable (provided by the map using the input data) and the actual initial resonance generation information converges to or below a predefined value.

[0088] Similarly, the DM3 map data is a trained model that was trained on a vehicle with the same specifications as vehicle VC before being installed in vehicle VC. To train the DM3 map data, training data is also acquired beforehand, comprising teacher data and input data. This means that various input data are acquired as described above by actually driving the vehicle while the operating point of the speed-changing device 20 is changed. In this case, additional information about the generation of a second resonance, indicating whether resonance is generated in belt 37, is acquired alongside the teacher data.For example, the information about the generation of a second resonance can be set to "0" if resonance is generated in belt 37, and the information about the generation of a second resonance can be set to "1" if no resonance is generated in belt 37.

[0089] A large amount of training data, including teacher data and input data, is generated by driving the vehicle under various conditions. The DM3 map data is trained using this training data. Input variables and an output variable are therefore set so that any error between the output variable, which is generated from the map using the input data, and the actual information about the generation of a second resonance at or below a predefined value converges.

[0090] When the output variable Y (MP) is calculated in step S35, the CPU 61 determines in the next step S37 whether the determination coefficient MP is set to "1". If the determination coefficient MP is set to "1" (S37: YES), the CPU 61 proceeds to step S39. In step S39, the CPU 61 evaluates the output variable Y (1) calculated in step S35. That is, the CPU 61 uses the output variable Y (1) to determine whether an anomaly is caused in belt 37. For example, the CPU 61 determines that an anomaly is caused if the output variable Y (1) is equal to or less than an anomaly determination value. Conversely, the CPU 61 does not determine that an anomaly is present if the output variable Y (1) is greater than the anomaly determination value. The anomaly determination value is set to a value between "0" and "1". For example, the anomaly detection value can be set to "0.5".

[0091] In the next step S41, CPU 61 determines, based on the result of the evaluation of the output variable Y(1) in step S39, whether an anomaly is caused in belt 37. If it is determined that an anomaly is caused (S41: YES), CPU 61 proceeds to the next step S43. In step S43, CPU 61 instructs memory 63 to store information indicating that an anomaly has been caused in belt 37 and then continues the process in the next step S45. If step S41 does not determine that an anomaly is caused in belt 37 (NO), CPU 61 proceeds to the next step S45.

[0092] In step S45, CPU 61 increments the determination coefficient MP by "1". CPU 61 then continues with the process in step S33. If the determination coefficient MP is not set to "1" in step S37 (NO), CPU 61 proceeds to the next step S47. In step S47, CPU 61 evaluates the output variable Y (MP). If the determination coefficient MP is set to "2", CPU 61 uses the output variable Y (2) to determine whether the vibration of belt 37 generates resonance in other vehicle components. For example, if the output variable Y (2) is equal to or less than a first resonance determination value, CPU 61 determines that resonance is generated in other vehicle components due to the vibration of belt 37.If, on the other hand, the output variable Y (2) is greater than the first resonance determination value, the CPU 61 does not detect that resonance is generated in the other vehicle-internal devices due to the vibration of the belt 37. The first resonance determination value is set to a value between "0" and "1". For example, the first resonance determination value can be set to "0.5".

[0093] When the determination coefficient MP is set to "3", the CPU 61 determines, based on the output variable Y(3), whether a resonance is generated in belt 37 due to vibrations from other devices in the vehicle. For example, if the output variable Y(3) is equal to or less than a second resonance determination value, the CPU 61 determines that a resonance is generated in belt 37 due to vibrations from other vehicle components. Conversely, if the output variable Y(3) is greater than the second resonance determination value, the CPU 61 does not determine that a resonance is generated in belt 37 due to vibrations from other vehicle components. The second resonance determination value is set to a value between "0" and "1". For example, the second resonance determination value can be set to "0.5".

[0094] Subsequently, in step S49, CPU 61 determines, based on the result of the evaluation of the output variable Y (MP) in step S47, whether a resonance is generated. That is, if the determination coefficient MP is set to "2", CPU 61 determines whether resonance is generated in the other vehicle-internal devices due to the vibration of belt 37. If it is determined that resonance is generated in the other vehicle-internal devices due to the vibration of belt 37 (S49: YES), CPU 61 proceeds to step S51. If it is not determined that resonance is generated in the other vehicle-internal devices due to the vibration of belt 37 (S49: NO), CPU 61 proceeds to step S53. If the determination coefficient MP is set to "3", CPU 61 determines whether resonance is generated in belt 37 due to the operation of the other vehicle-internal devices.If it is determined that resonance is generated in belt 37 due to the operation of other vehicle-internal devices (S49: YES), the CPU 61 proceeds to step S51. If it is not determined that resonance is generated in belt 37 due to the operation of other vehicle-internal devices (S49: NO), the CPU 61 proceeds to step S53.

[0095] In step S51, the CPU 61 stores the current operating point of the gear ratio change mechanism 30 in memory 63. That is, if the determination coefficient MP is set to "2", the CPU 61 instructs memory 63 to store the current operating point of the gear ratio change mechanism 30 as an operating point for a case in which resonance is generated in vehicle components other than the speed-changing device 20 due to the operation of the speed-changing device 20. If the determination coefficient MP is set to "3", the CPU 61 instructs memory 61 to store the current operating point of the gear ratio change mechanism 30 as an operating point for a case in which resonance is generated in the belt 37 due to the operation of the other vehicle components. The CPU 61 then proceeds to the next step S53.

[0096] In step S53, the CPU 61 determines whether the determination coefficient MP is set to "3". If the determination coefficient MP is not set to "3" (S53: NO), the CPU 61 proceeds to step S55. Then, in step S55, the CPU 61 increments the determination coefficient MP by "1" and proceeds to step S33. If, however, the determination coefficient MP is set to "3" in step S53 (YES), the CPU 61 temporarily terminates the sequence of steps. The function of the present embodiment is described.

[0097] If no anomaly is caused in belt 37, the input speeds Ninp and the output speeds Noutp fluctuate with periods that correspond to a motor speed NE. Fig. Figure 6 shows the curves of the input speeds Ninp and the output speeds Noutp for the case where the motor speed NE is constant. If an anomaly, such as damage to an element 372 of the belt 37, occurs, the input speeds Ninp and the output speeds Noutp are superimposed with vibration components due to the damage. As a result, the input speeds Ninp and the output speeds Noutp change, as shown in Fig. 7 is shown.

[0098] In the present embodiment, the input speed-related data RDNinp, based on the chronological data of the input speeds Ninp, and the output speed-related data RDNoutp, based on the chronological data of the output speeds Noutp, are input as the input variables x into the characteristic map defined by the characteristic map data DM1. The output variable Y(1) is then output from the characteristic map. The output variable Y(1) has a value that is determined taking into account changes or profiles of the input speeds Ninp and changes or profiles of the output speeds Noutp.Therefore, there is a difference in the output variable Y(1) between a case in which vibration components due to an anomaly in the belt 37 are superimposed on at least either the input speeds Ninp or the output speeds Noutp, and a case in which vibration components due to an anomaly in the belt 37 are superimposed on neither the input speeds Ninp nor the output speeds Noutp. Therefore, it is possible to determine whether an anomaly in the belt 37 is caused by evaluating the output variable Y(1).

[0099] The following effects can additionally be achieved with the present embodiment. (1-1) In the present embodiment, the chronological data on the normalized input speeds NinpN are derived by normalizing the chronological data on the input speeds Ninp. Then, the input speed-related data RDNinp are generated based on the chronological data of the normalized input speeds NinpN. Therefore, the degree of difference between the input speed-related data RDNinp for a case in which an anomaly is caused in the belt 37 and the input speed-related data RDNinp for a case in which no anomaly is caused is not so large between a case in which the input speeds Ninp are high and a case in which the input speeds Ninp are low.Thus, it is possible to suppress fluctuations in the determination accuracy due to changes in the input speeds Ninp by using the input speed-related data RDNinp as the input variable for the characteristic map, which are generated based on the chronological data for the normalized input speeds NinpN. (1-2) The accuracy of the determination can be improved if the chronological data for the normalized input speeds NinpN comprise a larger data set. In the present embodiment, the input speed-related data RDNinp, which are used as input variables for the characteristic map, are obtained by displaying the chronological data for the normalized input speeds NinpN in a histogram. This does not significantly increase the data volume of the input speed-related data RDNinp, even if the chronological data is large. Thus, the determination can be carried out precisely while suppressing an increase in the data set of input variables. (1-3) In the present embodiment, the chronological data for the normalized output speeds NoutpN are derived by normalizing the chronological data for the output speeds Noutp. Then, the output speed-related data RDNoutp are generated based on the chronological data of the normalized output speeds NoutpN. Therefore, the degree of difference between the output speed-related data RDNoutp in a case where an anomaly is caused in the belt 37 and the output speed-related data RDNoutp in a case where no anomaly is caused is not so large between a case where the output speeds Noutp are high and a case where the output speeds Noutp are low.Thus, it is possible to suppress fluctuations in the determination accuracy due to fluctuations in the output speeds Noutp by using the output speed-related data RDNoutp as the input variable for the characteristic map, which are generated on the basis of the chronological data for the normalized output speeds NoutpN. (1-4) The accuracy of the determination can be further improved if the chronological data for the normalized output speeds NoutpN comprise a larger data set. In the present embodiment, the output speed-related data RDNoutp, which are used as input variables for the characteristic map, are obtained by displaying the chronological data for the normalized output speeds NoutpN in a histogram. This prevents a significant increase in the data volume of the output speed-related data RDNoutp, even if the chronological data comprises a large number of data points. Thus, the determination can be carried out precisely while suppressing an increase in the data set as an input variable. (1-5) Even under conditions where no anomaly is caused in the belt 37, the amplitude and period of the oscillation of the input speeds Ninp and the output speeds Noutp can differ between a case where the input torque Trq is large and a case where the input torque Trq is small. Therefore, in the present embodiment, the input torque Trq is used as the input variable for the characteristic map. Thus, the output variable Y(MP) provided by the characteristic map has a value determined taking into account the input torque Trq. Therefore, the accuracy of the determination can be increased by making the determination based on the output variable Y(MP). (1-6) The viscosity of oil is lower when the oil temperature is higher. The degree of slippage of the belt 37 on the input pulley 33 and the degree of slippage of the belt 37 on the output pulley 35 tend to be greater when the viscosity of the oil circulating in the transmission change mechanism 30 is lower. The vibration mode of the input speeds Ninp and the vibration mode of the output speeds Noutp can also change under conditions where no anomaly is caused in the belt 37 when the degree of slippage differs. Therefore, in the present embodiment, the oil temperature sensing value Toil is used as an input variable for the characteristic map. That is, the output variable Y (MP) provided by the characteristic map has a value determined taking into account the viscosity of the oil.Therefore, the accuracy of the determination can be increased by making the determination based on the output variable Y (MP). (1-7) Vibration components due to the shape of the belt 37 may be superimposed on the input speeds Ninp and the output speeds Noutp. Such vibration components are not attributable to an anomaly caused in the belt 37. In the present embodiment, the belt shape index MVs is stored in memory 63 for each speed change device 20. The belt shape index MVs is fed to the characteristic map as an input variable. That is, the output variable Y (MP) provided by the characteristic map has a value that is determined taking into account the shape of the belt 37. Therefore, the accuracy of the determination can be increased by making the determination based on the output variable Y (MP). (1-8) Vibration components due to the shape of elements of the belt 37 may be superimposed on the input speeds Ninp and the output speeds Noutp. Such vibration components are not attributable to an anomaly caused in the belt 37. Therefore, in the present embodiment, the element shape index MVEs is stored in memory 63 for each speed-changing device 20. The element shape index MVEs is fed to the characteristic map as an input variable. That is, the output variable Y (MP) output by the characteristic map has a value that is determined taking into account the shape of the elements 372 that are components of the belt 37. Therefore, the accuracy of the determination can be increased by making the determination based on the output variable Y (MP). (1-9) If the amount of backlash or play in the direction of rotation of the elements 372 of the belt 37 is large, vibration components corresponding to the amount of play tend to be superimposed on the input speeds Ninp and the output speeds Noutp. Therefore, in the present embodiment, the backlash amount index MVRa, which is an index for the amount of play, is stored in memory 63. The backlash amount index MVRa is fed to the characteristic map as an input variable. That is, the output variable Y (MP) provided by the characteristic map has a value that is determined taking the play into account. Therefore, the accuracy of the determination can be increased by making the determination based on the output variable Y (MP). (1-10) The degree of slip caused between the input pulley 33 and the belt 37 can be changed depending on the input clamping force Pinp. The vibration mode of the input speeds Ninp and the output speeds Noutp can change when the degree of slip changes. In the present embodiment, the input clamping force Pinp is therefore used as an input variable for the characteristic map. That is, the output variable Y(MP) provided by the characteristic map has a value that is determined taking into account the amount of slip caused between the input pulley 33 and the belt 37. Therefore, the accuracy of the determination can be increased by making the determination based on the output variable Y(MP). (1-11) The degree of slippage caused between the output pulley 35 and the belt 37 can change depending on the output clamping force Poutp. The vibration mode of the input speeds Ninp and the output speeds Noutp can change when the degree of slippage changes. In the present embodiment, the output clamping force Poutp is therefore used as an input variable for the characteristic map. That is, the output variable Y(MP) provided by the characteristic map has a value that is determined taking into account the slippage caused between the output pulley 35 and the belt 37. Therefore, the accuracy of the determination can be increased by making the determination based on the output variable Y(MP). (1-12) If an anomaly is caused in belt 37 and the vibrations due to the anomaly are superimposed on the input speeds Ninp and the output speeds Noutp, the vibrations due to the anomaly can be transmitted to the vehicle body, thereby changing the vehicle acceleration G. That is, the output variable Y (MP) provided by the map is determined taking the vehicle acceleration G into account. Therefore, the accuracy of the determination can be improved by performing the determination based on the output variable Y (MP). (1-13) If an anomaly occurs in the belt 37 and a vibration due to the anomaly is superimposed on the input speeds Ninp and the output speeds Noutp, an abnormal noise can be generated in the transmission change mechanism 30 due to the vibration. Such an abnormal noise can be detected by the noise sensor 107 installed in the engine compartment. Therefore, in the present embodiment, the noise detection value Snd is used as an input variable for the characteristic map. That is, the output variable Y (MP) provided by the characteristic map is determined taking into account the noise detection value Snd. Therefore, the accuracy of the determination can be increased by making the determination based on the output variable Y (MP). (1-14) When the braking force BPvc of the vehicle VC is generated, the drive wheels 50 are decelerated, and therefore the input speeds Ninp and the output speeds Noutp are also changed. In this case, vibration components can be superimposed on the input speeds Ninp and the output speeds Noutp, which depend on the magnitude of the braking force BPvc or the rate of increase of the braking force BPvc. In the present embodiment, the braking force BPvc is therefore used as an input variable for the characteristic map. That is, the output variable Y (MP) provided by the characteristic map has a value that is determined taking into account the braking force BPvc. Therefore, the accuracy of the determination can be increased by making the determination based on the output variable Y (MP). (1-15) If an anomaly is caused in the belt 37, and therefore a vibration is superimposed on the input speeds Ninp and the output speeds Noutp, the speed ratio RN can differ between a case where no anomaly is caused and a case where an anomaly is caused. Therefore, in the present embodiment, the speed ratio RN is used as the input variable for the characteristic map. That is, the output variable Y (MP) provided by the characteristic map has a value that is determined taking into account the speed ratio RN. Therefore, the determination accuracy can be increased by making the determination based on the output variable Y (MP). (1-16) Even if no anomaly is caused in the belt 37, the magnitude and period of the oscillation of the input speeds Ninp and the output speeds Noutp can change if the degree of such temporal changes is relatively large and the degree of change in the properties of the belt 37 over time is relatively small. Therefore, in the present embodiment, the travel distance SC, which is an example of an index indicating the degree of change in properties over time, is used as an input variable for the characteristic map. That is, the output variable Y(MP) provided by the characteristic map has a value that is determined taking into account the degree of temporal changes in the characteristics of the belt 37. Therefore, the accuracy of the determination can be increased by making the determination based on the output variable Y(MP). (1-17) In the present embodiment, it is possible to determine, by inputting the input variables into the characteristic map specified by the characteristic map data DM2 and evaluating the output variable Y (2) output from the characteristic map, whether resonance is generated in the other vehicle-internal devices due to the vibration of the belt 37. Consequently, a factor of resonance generation in the other vehicle-internal devices can be specified.

[0100] In this case, it is possible to detect an operating point of the transmission change mechanism 30 at which resonance is generated in the other vehicle-internal devices. In the present embodiment, such an operating point is stored in memory 63. It is possible to suppress the generation of resonances in the other vehicle-internal devices while the vehicle VC is in motion by controlling the internal combustion engine 10 and the speed-changing device 20 in such a way that the operating point of the transmission mechanism 30 does not coincide with the operating point stored in memory 63.

[0101] (1-18) In the present embodiment, by inputting the input variables into the characteristic map specified by the characteristic map data DM3 and evaluating the output variable Y (3) output from the characteristic map, it can be determined whether resonance is generated in the belt 37 by the operation of the other vehicle-internal devices. Thus, a factor of resonance generation in the belt 37 can be determined.

[0102] In this case, it is possible to detect an operating point of the transmission change mechanism 30 at which resonance is generated in the belt 37. In the present embodiment, such an operating point is stored in the memory 63. It is possible to suppress the generation of resonance in the belt 37 while the vehicle VC is in motion by controlling the internal combustion engine 10 and the speed-changing device 20 in such a way that the operating point of the transmission mechanism 30 does not coincide with the operating point stored in the memory 63. Second embodiment

[0103] A second embodiment is described below with reference to the figures, mainly with regard to differences from the first embodiment.

[0104] The present embodiment differs from the first embodiment in the method for generating the data RDNinp related to the input speed and the data RDNoutp related to the output speed.

[0105] The sequence of processes performed by the control device 60 to make various settings as described above are illustrated by Fig. 10 described. Fig. Figure 10 illustrates part of the sequence of processes carried out by the control device 60 to make the various settings.

[0106] First, CPU 61 acquires chronological data on the input speeds Ninp and chronological data on the output speeds Noutp by executing processes corresponding to those in steps S11 to S17. Then, CPU 61 proceeds with the process in step S61. In step S61, CPU 61 derives a mean value NinpAv of the input speeds, which is the average of the input speeds Ninp during an acquisition period for the chronological data on the input speeds Ninp. For example, CPU 61 can derive as the mean value NinpAv of the input speeds the mean value of all input speeds Ninp(1), Ninp(2), ..., and Ninp(z) contained in the acquired chronological data for the input speeds Ninp.Subsequently, in step S63, CPU 61 derives a mean value NoutpAv of the output speeds, which is the average of the output speeds Noutp during a recording period for the chronological data of the output speeds Noutp. For example, CPU 61 can derive as the mean value NoutpAv of the output speeds the mean of all output speeds contained in the recorded chronological data of the output speeds Noutp (1), Noutp (2), ..., and Noutp (z).

[0107] In the next step S65, the CPU 61 derives input frequency characteristics FCinp based on the chronological data of the input speeds Ninp and the mean value NinpAv of the input speeds. The input frequency characteristics FCinp are the frequency characteristics of the input speeds Ninp. That is, the CPU 61 derives the frequency characteristics of the chronological data for the input speeds Ninp by performing a Fast Fourier Transform on the chronological data for the input speeds Ninp. The frequency characteristics represent the relationship between frequency and any physical quantity. In the present embodiment, the frequency characteristics represent the relationship between frequency and intensity. The CPU 61 detects the amplitude of the primary frequency of the belt 37's rotation based on the mean value NinpAv of the input speed.Specifically, the primary rotation frequency is determined as a value obtained by dividing the mean input speed (in rpm) by 60. The CPU 61 derives the input frequency characteristics FCinp by normalizing the frequency response of the chronological data over the input speeds Ninp, obtained by performing a Fast Fourier Transform, based on the amplitude of the primary rotation frequency. Examples of normalization methods are Max / Min scalar and Standard scalar. That is, the input frequency characteristics FCinp are normalized frequency responses.

[0108] In the next step, S67, the CPU 61 derives output frequency characteristics FCoutp based on the chronological data of the output speeds Noutp and the mean value NoutpAv of the output speed. The output frequency characteristics FCoutp are the frequency characteristics of the output speeds Noutp. That is, the CPU 61 derives the frequency characteristics of the chronological data of the output speeds Noutp by performing a Fast Fourier Transform on the chronological data of the output speeds Noutp. The CPU 61 determines the amplitude of the primary rotational frequency of the belt 37 from the mean value NoutpAv of the output speeds. The CPU 61 derives the output frequency characteristic FCoutp by normalizing the frequency characteristic of the chronological data of the output speeds Noutp, obtained by performing a Fast Fourier Transform, using the amplitude of the primary rotational frequency.This means that the output frequency characteristics FCoutp are normalized frequency characteristics.

[0109] Subsequently, in step S69, CPU 61 generates input speed-related data RDNinp based on the input frequency characteristics FCinp. For example, CPU 61 uses the input frequency characteristics FCinp as the input speed-related data RDNinp. In the next step S71, CPU 61 generates output speed-related data RDNoutp based on the output frequency characteristic FCoutp. For example, CPU 61 uses the output frequency characteristic FCoutp as the output speed-related data RDNoutp.

[0110] CPU 61 then proceeds to the process in step S27. The subsequent processes are similar to those in the first embodiment and are therefore not described in detail here.

[0111] In addition to the effects that are equivalent to the effects described in (1-5) to (1-18) with reference to the foregoing embodiment, the following effects can be achieved with the present embodiment. (2-1) In the present embodiment, the input frequency characteristics FCinp are derived by performing a Fast Fourier Transform on the chronological data of the input speeds Ninp. Then, the input speed-related data RDNinp are generated based on the input frequency characteristics FCinp. There may be a difference between the input speed-related data RDNinp in a case where an anomaly is caused in the belt 37 that can change the input frequency characteristics FCinp, and the input speed-related data RDNinp that are generated when no such anomaly is caused.Therefore, it is possible to determine that an anomaly is caused when an anomaly which can influence the input frequency characteristics FCinp is caused in belt 37 by using the output variable Y output by the map when the speed-related input data RDNinp, which were generated based on the input frequency characteristics FCinp, are received as an input variable for the map. (2-2) In the present embodiment, the amplitude of the primary frequency of the belt's rotation is determined based on the mean value NinpAv of the input speeds. The input frequency characteristics FCinp are then derived by normalizing the frequency characteristics of the input speeds Ninp, obtained by a Fast Fourier Transform, using the amplitude of the primary frequency of the belt's rotation. The magnitude of a characteristic value due to anomaly occurring in the input frequency characteristics FCinp thus normalized does not change significantly between a case where the input speeds Ninp are high and a case where the input speeds Ninp are low.This makes it possible to suppress fluctuations in the determination accuracy due to variations in the input speeds Ninp by using the input speed-related data RDNinp as the input variable for the characteristic map, which are generated on the basis of the normalized input frequency characteristics FCinp. (2-3) In the present embodiment, the output frequency characteristics FCoutp are obtained by performing a Fast Fourier Transform on the chronological data of the output speeds Noutp. Then, the output speed-related data RDNoutp are generated based on the output frequency characteristics FCoutp. There may be a difference between the output speed-related data RDNoutp in the case where an anomaly is caused in the belt 37 that can change the output frequency characteristics FCoutp, and the output speed-related data RDNoutp that are generated when no such anomaly is caused.Therefore, by using the output variable Y provided by the characteristic map, it is possible to determine that an anomaly is caused which can generate the output frequency characteristics FCoutp when an anomaly is caused in the belt 37 when the data RDNoutp related to the output speed is received as an input variable for the characteristic map, which are generated based on the output frequency characteristics FCoutp. (2-4) In the present embodiment, the amplitude of the primary frequency of the belt's rotation is determined based on the mean value NoutpAv of the output speed. The output frequency characteristics FCoutp are then derived by normalizing the output speed frequency characteristics Noutp, obtained by a Fast Fourier Transform, with the amplitude of the primary frequency of the belt's rotation. The magnitude of the characteristic value due to anomaly occurring in the output frequency characteristics FCoutp thus normalized does not change significantly between a case where the output speeds Noutp are high and a case where the output speeds Noutp are low.This makes it possible to suppress fluctuations in the determination accuracy due to changes in the output speeds Noutp by using the output speed-related data RDNoutp, generated on the basis of the normalized output frequency characteristics FCoutp, as the input variable for the characteristic map. Third embodiment

[0112] A third embodiment is described below with reference to the figures, particularly with regard to differences from the first and second embodiments.

[0113] In the present embodiment, the frequency characteristics of the chronological data for the input speeds Ninp are derived by performing a Fast Fourier Transform on such chronological data. The speed-related input data RDNinp are then generated based on the frequency characteristics of the chronological data. Furthermore, the frequency characteristics of the chronological data for the output speeds Noutp are obtained by performing a Fast Fourier Transform on this chronological data. The output speed-related data RDNoutp are then generated based on the frequency characteristics of the chronological data.

[0114] The sequence of processes performed by the control device 60 to make various determinations as described above are illustrated by the Fig. 11 described. Fig. Figure 11 illustrates part of the sequence of processes carried out by the control device 60 to make the various determinations.

[0115] First, the CPU 61 acquires chronological data about the input speeds Ninp and chronological data about the output speeds Noutp by executing processes corresponding to those in steps S11 to S17. Then, the CPU 61 proceeds to the process in step S81. In step S81, the CPU 61 derives the input frequency characteristics FCinp1 based on the chronological data about the input speeds Ninp. The input frequency characteristics FCinp1 are the frequency characteristics of the input speeds Ninp. That is, the CPU 61 divides the frequency band in which data about the frequency characteristics derived by performing a Fast Fourier Transform on the chronological data about the input speeds Ninp are distributed into m equal frequency bands. Then, the CPU 61 derives data as the input frequency characteristics FCinp1, which are obtained by averaging the data of the frequency characteristics for each of the frequency bands. This is done, for example, by...This is achieved by performing a 1 / 3 octave band analysis of the frequency response curves. In the present embodiment, "m" is set to an integer of "2" or more.

[0116] In the next step, S83, the CPU 61 derives output frequency characteristics FCoutp1 based on the chronological data of the output speeds Noutp. The output frequency characteristics FCoutp1 are the frequency characteristics of the output speeds Noutp. That is, the CPU 61 divides the frequency band in which the data on the frequency characteristics derived by performing a Fast Fourier Transform on the chronological data of the output speeds Noutp are distributed into m equal frequency bands. Then, the CPU 61 derives data as output frequency characteristics FCoutp1, which are obtained by averaging the data on the frequency characteristics for each of the frequency bands. In the present embodiment, "m" is set to an integer of "2" or more.

[0117] Subsequently, in step S85, CPU 61 generates input speed-related data RDNinp based on the input frequency characteristics FCinp1. For example, CPU 61 uses the input frequency characteristics FCinp1 as the input speed-related data RDNinp. In the next step S87, CPU 61 generates output speed-related data RDNoutp based on the output frequency characteristics FCoutp1. For example, CPU 61 uses the output frequency characteristics FCoutp1 as the output speed-related data RDNoutp.

[0118] The CPU 61 then continues with step S27. The subsequent processes are similar to those described in the first and second embodiments and are therefore not described in detail here.

[0119] In addition to the effects that are equivalent to those described in (1-5) to (1-18), (2-1) and (2-3) above, the following effects can be achieved with the present embodiment. (3-1) In the present embodiment, the frequency band in which the data derived by performing a Fast Fourier Transform of the chronological data about the input speeds Ninp are distributed is divided into m equal frequency bands. The input speed-related data RDNinp are generated based on data averaged for each frequency band as described above. It is possible to suppress an increase in the number of data points to be inputted into the characteristic map by using the input speed-related data RDNinp, generated by averaging data for each frequency band as described above, as the input variable for the characteristic map. (3-2) In the present embodiment, the frequency band is divided into m equal frequency bands in which the data are distributed. These data are derived from the chronological output speed data Noutp by performing a Fast Fourier Transform. The output speed-related data RDNoutp are generated based on data averaged for each frequency band as described above. It is possible to prevent an increase in the number of data points to be inputted into the characteristic map by using the output speed-related data RDNoutp, generated as described above by averaging data for each frequency band, as the input variable for the characteristic map.

[0120] The control device 60 is an example of the anomaly detection device. The vehicle VC is an example of the vehicle. The internal combustion engine 10 is an example of the vehicle's energy source. The drive wheels 50 are an example of the drive wheels. The speed-changing device 20 is an example of the power transmission device. The input pulley 33 is an example of the input pulley. The output pulley 35 is an example of the output pulley. The belt 37 is an example of an endless rotating element. The CPU 61 and the ROM 62 are an example of the execution device. The memory 63 is an example of the memory. The input speeds Ninp are an example of the input speeds. The input speed-related data RDNinp are an example of the input speed-related data.The output speeds Noutp are an example of output speeds. The output speed-related data RDNoutp are an example of output speed-related data. The output variable Y (MP) is an example of an output variable. The in . Fig. The characteristic map data DM1 to DM3 shown in section 1 are an example of the characteristic map data. The processes in steps S11 to S27, which are described in the Fig. 4 and Fig. Figure 5 shows the processes in steps S61 to S71, which are described in Fig. 10 are shown, and the processes in steps S81 to S87, which are in Fig. Figures 11 are an example of the data collection process. The processes in steps S35 to S41 are for a case in which the coefficient of determination MP is in the Fig. 4 and Fig. Setting 5 to "1" is an example of the anomaly detection process.

[0121] The period from the time when the coefficient z in step S11 is set to "1" until the time when it is determined that the coefficient z in step S17 is in Fig. 4 and Fig. 5 greater than the coefficient determination value zTh is an example of the measurement period. The cycle of the process execution in step S13 in the Fig. 4 and Fig. 5 is an example of the data capture cycle. The processes in steps S11 to S17 in Fig. 4 and Fig. 5 are an example of the speed measurement process. The process in step S19 and the process in step S23, which are in the Fig. 4 and Fig. Figure 5 is an example of the production process.

[0122] The process in step S21 and the process in step S25, which in Fig. 4 and Fig. Figure 5 illustrates an example of the production process.

[0123] The process in step S65 and the process in step S69, which in Fig. Figure 10 illustrates an example of the process for generating the frequency characteristic. The process in step S81 and the process in step S85, which are shown in Fig. Figure 11 illustrates another example of the process for generating the frequency characteristic.

[0124] The period from the time at which the coefficient z is set to "1" in step S11 until the time at which it is determined that the coefficient z is greater than the coefficient determination value zTh in step S17 in the Fig. 4 and Fig. 5 is an example of the recording period for the chronological data on the input speeds. The one in Fig. The process shown in step S61 is an example of the mean calculation process. The process in step S65 and the process in step S69, which are shown in Fig. Figures 10 are an example of the process for generating frequency characteristics.

[0125] The process in step S81 and the process in step S85, which in Fig. Figure 11 shows an example of the process for generating frequency characteristics.

[0126] The in Fig. 10. Process shown in step S67, which is in Fig. 11. Process shown in step S71, which is in Fig. 11 process shown in step S83 and the one in Fig. The process shown in step S87 is an example of the process for generating the frequency characteristic.

[0127] The period from the time at which the coefficient z is set to "1" in step S11 until the time at which it is determined that the coefficient z is greater than the coefficient determination value zTh in step S17 in the Fig. 4 and Fig. 5 is an example of the recording period for the chronological data on output speeds. The in Fig. The process shown in step S63 is an example of the mean calculation procedure. The process in step S67 and the process in step S71, which are in Fig. Figures 10 are an example of the process of frequency characteristic generation.

[0128] The in Fig. The processes illustrated in step S83 and in step S87 are an example of the process for generating the frequency characteristic.

[0129] The input torque Trq is an example of torque.

[0130] The oil temperature measurement value Toil is an example of oil temperature.

[0131] The belt shape index MVs is an example of an index that specifies the shape of the endless circulating element. The element shape index MVEs is an example of an index that indicates the shape of the components of the endless circulating element.

[0132] The input clamping force Pinp is an example of the force with which the input pulley holds the endless rotating element. The output clamping force Poutp is an example of the force with which the output pulley holds the endless rotating element.

[0133] The MVRa setback amount index is an example of an index that indicates the amount of the setback or game.

[0134] The accelerometer 106 is an example of an accelerometer. The vehicle acceleration G is an example of the value measured by the accelerometer.

[0135] The noise sensor 107 is an example of a noise sensor. The sound detection value Snd is an example of the detection value of the sound sensor.

[0136] The wheel speed in a VW is an example of the speed of the drive wheels.

[0137] The braking force BPvc is an example of the vehicle's braking force.

[0138] The speed ratio RN is an example of the ratio between the speeds of the input pulley and the output pulley.

[0139] The driving distance SC is an example of the index that indicates the degree of temporal fluctuations of the properties.

[0140] The elements 372 are an example of the components of the endless circulating element.

[0141] The processes in steps S47 and S49 for a case in which the coefficient of determination MP in the Fig. 4 and Fig. Setting 5 to “2” is an example of the process for determining the oscillations.

[0142] The processes in steps S47 and S49 for a case in which the coefficient of determination MP in the Fig. 4 and Fig. Setting 5 to “3” is an example of the resonance determination process. Modifications

[0143] The embodiments described above can be modified as follows. The embodiments described above and the modifications described below can be combined with one another, unless these embodiments and modifications are technically incompatible. Characteristic map - In the embodiments described above, the activation functions of the characteristic map are exemplary and not limited to those described above. For example, a logistic sigmoid function can be used as an activation function of the characteristic map. - In the embodiments described above, a neural network with only one intermediate layer is given as an example. However, the neural network can also contain two or more intermediate layers. - In the embodiments described above, a fully connected neural network with forward propagation is given as an example of the neural network. However, the present invention is not limited to this. For example, a recurrent neural network can be used as the neural network. A recurrent neural network is preferably used when the chronological data of the normalized input speeds NinpN, the chronological data of the input speeds Ninp, the chronological data of the normalized output speeds NoutpN, and the chronological data of the output speeds Noutp, as described below, are used as input variables for the characteristic map. Characteristic map data In the embodiments described above, the memory 63 stores the map data DM1 to determine whether an anomaly is caused in the components of the gear ratio change mechanism 30, the map data DM2 to determine whether a resonance is generated in the other vehicle-internal devices due to the operation of the gear ratio change mechanism 30, and the map data DM3 to determine whether the belt 37 experiences resonance due to the operation of the other vehicle-internal devices. However, the present invention is not limited thereto. For example, the memory 63 can store map data that prescribes a map which outputs a variable with which all three determinations described above can be carried out. - In the embodiments described above, memory 63 does not need to store the characteristic map data DM2 as long as memory 63 stores the characteristic map data DM1. Furthermore, memory 63 does not need to store the characteristic map data DM3 as long as memory 63 stores the characteristic map data DM1. Even in this case, by inputting the input variables into the characteristic map defined by the characteristic map data DM1 and using the output variables provided by the characteristic map, it can be determined whether an anomaly exists in the belt 37. Input variables - A parameter other than the travel distance SC can be used as the index indicating the degree of temporal variation in the properties of the endlessly rotating element. For example, the total time that belt 37 runs and the number of times belt 37 runs can be used as input variables for the characteristic map as the index indicating the degree of temporal variation in the properties of the endlessly rotating element. - The input variables do not need to include an index indicating the degree of temporal changes in the properties of the endlessly looping element. - The translation ratio RN does not need to be among the input variables. - The difference between the input speeds Ninp and the output speeds Noutp can be used as an input variable for the characteristic map instead of the gear ratio RN. - Chronological data for the translation ratio RN can be recorded, and multiple translation ratios RN contained in the chronological data for the translation ratio RN can be used as one input variable for the characteristic map. - The braking force BPvc must not be among the input variables. - If the input variables do not include the braking force BPvc, it can be used as an input variable to determine whether a braking force is applied to the drive wheels 50. Additionally, the magnitude of the decrease in the rotational speed of the drive wheels 50 per unit of time can be used as an input variable. - If the input variables do not include the braking force BPvc, the one in Fig. 4 and Fig. The sequence of processes shown in point 5, for example, will not be carried out if the vehicle is braked. - The noise detection value Snd does not need to be among the input variables. In this case, the control device 60 can be used in the vehicle VC in which the noise sensor 107 is not installed. - The vehicle acceleration G does not need to be among the input variables. - The setback amount index MVRa does not need to be among the input variables. - The input clamping force Pinp does not need to be among the input variables. - The output clamping force Poutp does not need to be among the input variables. - Memory 63 can store the shape of elements 372 as element shape index MVEs. In this case, the element shape index MVEs can be used as an input variable in the characteristic map. - The input variables do not need to include the element form index MVEs. - The input variables do not need to include the belt shape index MVs. - The input variables do not need to include the oil temperature sensing value Toil. - The input variables do not need to include the input torque Trq. - The input variables do not need to include the output speed-related data RDNoutp, as long as the input variables include the input speed-related data RDNinp. - The input variables do not need to include the input speed-related data RDNinp, as long as the input variables include the output speed-related data RDNoutp. Data related to input speed - In the second embodiment described above, the frequency characteristics of the chronological data about the input speeds Ninp, derived by performing a Fast Fourier Transform on such chronological data, can be used as the input speed-related data RDNinp. That is, a multitude of input speeds Ninp contained in the chronological data about the input speeds Ninp does not need to be normalized. - In the first embodiment described above, the input speed-related data RDNinp is composed of five counter values ​​Cnt(1) to Cnt(5). However, the number of counter values ​​is not limited to five. For example, if the range of numerical values ​​from 0 to 1 is divided into sections of 0.1 each, the input speed-related data RDNinp can consist of ten counter values ​​Cnt(1) to Cnt(10). - In the first embodiment described above, the data related to the input speed need not be the data RDNinp related to the input speed. That is, the data related to the input speed can be the chronological data for the normalized input speeds NinpN. - The data relating to the input speed can be the chronological data for the input speeds Ninp. Data related to the output speed - In the second embodiment described above, the frequency characteristics of the chronological data of the output speeds Noutp, obtained by performing a Fast Fourier Transform on such chronological data, can be used as the output speed-related data RDNoutp. That is, a multitude of output speeds Noutp contained in the chronological data about the output speeds Noutp do not need to be normalized. - In the first embodiment described above, the output speed-related data RDNoutp is composed of five counter values ​​Cnt(1) to Cnt(5). However, the number of counter values ​​is not limited to five. For example, if the range of numerical values ​​from 0 to 1 is divided into sections of 0.1 each, the output speed-related data RDNoutp can consist of ten counter values ​​Cnt(1) to Cnt(10). - In the first embodiment described above, the output speed-related data need not be the output speed-related data RDNoutp. In other words, the output speed-related data can be the chronological data for the normalized output speeds NoutpN. - The data relating to the output speed can be the chronological data of the output speeds Noutp. Output speeds In the embodiments described above, the output speeds Noutp are determined by a value based on the output signal Soutp of the rotation angle sensor 103 for the output shaft. However, the present invention is not limited to this. For example, the output speeds Noutp can also be determined by a value calculated based on the wheel speed VW, i.e., the speed of the drive wheels 50. In this case, the output speeds Noutp can be calculated by dividing the wheel speed VW by the reduction ratio of a torque transmission path from the output pulley 35 to the drive wheels 50. Execution device The execution device is not limited to one comprising the CPU 61 and the ROM 62 for executing the software processing. For example, the execution device may include a dedicated hardware circuit that performs hardware processing for at least some of the operations that are processed as software in the embodiments described above. Examples of the dedicated hardware circuit may include an application-specific integrated circuit (ASIC). That is to say, the execution device may have one of the following configurations (a) to (c). (a) The execution device comprises a processing device that performs all the processes described above according to a program and a program storage device, such as a ROM, that stores the program. (b) The execution device comprises a processing device that performs some of the processes described above according to a program, a program storage device and a dedicated hardware circuit that performs the remaining processes. (c) The execution device comprises a dedicated hardware circuit that performs all the processes described above. The execution device may comprise a plurality of software execution devices, each comprising a processing device and a program storage device, and a plurality of dedicated hardware circuits. Power transmission device - The power transmission device may differ from the one in Fig.The speed-changing device 20 shown in Figure 1 can be constructed as such, provided the power transmission device includes an endless rotating element. For example, the speed-changing device can include a transmission mechanism that incorporates a chain as the endless rotating element. Furthermore, the power transmission device need not have a transmission function, provided it is constructed such that it includes an input pulley, an output pulley, and an endless rotating element. vehicle - The vehicle can be a hybrid vehicle. Alternatively, the vehicle can be a vehicle that includes a motor-generator but, for example, does not include an internal combustion engine. In this case, the motor-generator serves as the vehicle's energy source.

Claims

[1] Device for a vehicle with a power transmission device (20), wherein the power transmission device comprises an input pulley (33) into which a torque can be applied, an output pulley (35) capable of delivering a torque in the direction of the drive wheels (50) of the vehicle, and an endless rotating element (37) wound around both the input pulley and the output pulley, wherein the device comprises: a memory (63) designed to store characteristic map data comprising data that form a characteristic map representing a relationship between an input variable and an output variable, wherein the data are learned by machine learning, where the input variable is at least one of the following: i) Input speed-related data (RDNinp) based on chronological data about input speeds (Ninp), which are speeds of the input pulley; and ii) output speed-related data (RDNoutp) based on chronological data about output speeds (Noutp), which are speeds of the output pulley, where the output variable indicates whether an anomaly is caused in the endlessly circulating element; and a processor (61) designed to to capture the input variable, to capture the output variable that matches the input variable using the characteristic map data, and to determine, based on the output variable, whether an anomaly is caused in the endless circulating element, where: the input variable includes the data related to the input speed (RDNinp); and the processor (61) is designed to than the chronological data on the input speeds to capture a large number of the input speeds (Ninp) that are captured in each acquisition cycle in a predefined measurement period, To derive frequency characteristics of the chronological data via the input speeds by performing a Fast Fourier Transform on the chronological data, and to generate the input speed-related data based on the frequency characteristics of the chronological data, wherein the processor (61) is designed to do the following: to calculate an average of the input speeds (Ninp) during a recording period for the chronological data on the input speeds; to detect an amplitude of a primary rotational frequency of the endless rotating element (37) based on the mean of the input rotational speeds; to normalize the frequency characteristics of the input rotational speeds with the amplitude of the primary rotational frequency, wherein the frequency characteristics are derived by performing the Fast Fourier Transform; and to generate the input speed-related data (RDNinp) based on the normalized frequency characteristics of the input speeds. [2] Device for a vehicle with a power transmission device (20), wherein the power transmission device comprises an input pulley (33) into which a torque can be applied, an output pulley (35) capable of delivering a torque in the direction of the drive wheels (50) of the vehicle, and an endless rotating element (37) wound around both the input pulley and the output pulley, wherein the device comprises: a memory (63) designed to store characteristic map data comprising data that form a characteristic map representing a relationship between an input variable and an output variable, wherein the data are learned by machine learning, where the input variable is at least one of the following: i) Input speed-related data (RDNinp) based on chronological data about input speeds (Ninp), which are speeds of the input pulley; and ii) output speed-related data (RDNoutp) based on chronological data about output speeds (Noutp), which are speeds of the output pulley, where the output variable indicates whether an anomaly is caused in the endlessly circulating element; and a processor (61) designed to to capture the input variable, to capture the output variable that matches the input variable using the characteristic map data, and to determine, based on the output variable, whether an anomaly is caused in the endless circulating element, where: the input variable includes the data related to the output speed (RDNoutp); and the processor (61) is designed to than the chronological data on the output speeds to capture a multitude of output speeds (Noutp) that are captured in each acquisition cycle in a predefined measurement period, To derive frequency characteristics of the chronological data via the output speeds by performing a Fast Fourier Transform on the chronological data, and to generate the rotational speed-related output data based on the frequency characteristics of the chronological data, wherein the processor (61) is designed to do the following: to calculate an average of the output speeds (Noutp) during a recording period for the chronological data on the output speeds; to detect an amplitude of a primary rotational frequency of the endless rotating element (37) based on the mean of the output rotational speeds; to normalize the frequency characteristic of the output speeds with the amplitude of the primary rotational frequency, wherein the frequency characteristic is derived by performing the Fast Fourier Transform; and to generate the output speed-related data (RDNoutp) based on the standardized frequency characteristics of the output speeds. [3] Device according to claim 1 or 2, wherein: the input variable includes the data related to the input speed (RDNinp); and the processor (61) is designed to than the chronological data on the input speeds to capture a multitude of the input speeds (Ninp) that are captured in each acquisition cycle in a predefined measurement period, and to generate the input speed-related data based on data that show a distribution of a size of numerical values ​​of the multitude of input speeds contained in the chronological data about the input speeds. [4] Device according to claim 1 or 2, wherein the input variable includes the data related to the input speed (RDNinp); and the processor (61) is designed to than the chronological data on the input speeds to capture a large number of the input speeds (Ninp) that are captured in each acquisition cycle in a predefined measurement period, to normalize the multitude of input speeds contained in the chronological data on input speeds, and To generate data as the input speed-related data, which show a distribution of a size of numerical values ​​of a multitude of normalized input speeds obtained by normalizing the input speeds. [5] Device according to any one of claims 1 to 3, wherein: the output variable includes the data related to the output speed (RDNoutp); and the processor (61) is designed to than the chronological data on the output speeds to capture a multitude of output speeds (Noutp) that are captured in each acquisition cycle in a predefined measurement period, and to generate the output speed-related data based on data that show a distribution of a size of numerical values ​​of the multitude of output speeds contained in the chronological data about the output speeds. [6] Device according to claim 1, 2 or 4, wherein: the output variable includes the data related to the output speed (RDNoutp); and the processor (61) is designed to than the chronological data on the output speeds to capture a multitude of output speeds (Noutp) that are captured in each acquisition cycle in a predefined measurement period, to normalize the multitude of output speeds contained in the chronological data on output speeds, and To generate data as the output speed-related data, which indicate a distribution of a size of numerical values ​​of a multitude of normalized output speeds obtained by normalizing the output speeds. [7] Device according to claim 1, wherein the processor (61) is configured to generate the input speed-related data (RDNinp) by uniformly subdividing the frequency bands in which data derived by performing the Fast Fourier Transform for the chronological data about the input speeds are distributed into a predetermined number of sub-frequency bands, and to average the data for each of the sub-frequency bands. [8] Device according to claim 2, wherein the processor (61) is configured to generate the output speed-related data (RDNoutp) by uniformly subdividing frequency bands into a predetermined number of sub-frequency bands in which data obtained by performing the Fast Fourier Transform for the chronological data about the output speeds are distributed, and by averaging the data for each of the sub-frequency bands. [9] Device according to any one of claims 1 to 8, wherein the input variable comprises at least either a torque (Trq) that can be delivered to the input pulley (33) or a temperature (Toil) of oil circulating in the power transmission device. [10] Device according to any one of claims 1 to 9, wherein the memory (63) is at least either: an index (MVs) that specifies a form of the endless circulating element for each power transmission device, or stores an index (MVEs) that indicates a form of a component of the endless circulating element for each power transmission device; and the input variable comprises the index stored in the memory. [11] Device according to any one of claims 1 to 10, wherein the input variable comprises at least one of the following quantities: a force (Pinp) with which the input pulley (33) holds the endless rotating element (37), and / or a force (Poutp) with which the output pulley (35) holds the endless rotating element. [12] Device according to any one of claims 1 to 11, wherein: the memory (63) stores an index (MVRa) indicating the magnitude of a kickback in one direction of rotation of the endless rotating element (37) for each power transmission device; and The input variable includes the index that indicates the size of the backlash and is stored in memory. [13] Device according to any one of claims 1 to 12, wherein the input variable comprises at least one of the following values: a measured value from an acceleration sensor attached to the vehicle (106) a measured value from a noise sensor installed in the engine compartment of the vehicle (107), a braking force (BPvc) of the vehicle, a ratio (RN) between the speed of the input pulley and the speed of the output pulley, and an index that indicates a degree of temporal fluctuations in the properties of the endlessly circulating element. [14] Device according to any one of claims 1 to 13, wherein: the input variable includes the data related to the output speed (RDNoutp); and The output speeds (Noutp) are calculated based on the speeds of the drive wheels (50). [15] Device according to any one of claims 1 to 14, wherein the processor (61) is configured to determine, based on the output variable, whether a component (372) of the endless rotating element (37) is damaged. [16] Device according to any one of claims 1 to 14, wherein the processor (61) is configured to do the following: to determine the output variable that matches the input variable based on the characteristic curve data; and to determine, based on the output variable, whether a vibration is generated in the endless rotating element (37) that causes a vehicle part other than the power transmission device to resonate. [17] Device according to any one of claims 1 to 14, wherein the processor (61) is configured to do the following: to capture the output variable corresponding to the input variable using the characteristic map data; and to determine, based on the output variable, whether a resonance is generated in the endlessly circulating element (37).

Images