Information processing device, information processing method, and program

Abstract

This invention provides an information processing device, an information processing method, and a program for estimating vehicle deterioration indicators. [Solution] The information processing device 10 includes a pre-estimated data acquisition unit 26 that acquires pre-estimated data of deterioration indicators, which are indicators related to the deterioration of a vehicle; an observation data acquisition unit 24 that acquires observation data of deterioration indicators; and a data assimilation processing unit 28 that derives post-estimated data of deterioration indicators by performing data assimilation processing based on the pre-estimated data and the observation data.

Description

【Technical Field】 【0001】 The present disclosure relates to an information processing apparatus, an information processing method, and a program. 【Background Art】 【0002】 Japanese Unexamined Patent Application Publication No. 2024-010623 discloses an information processing apparatus. The information processing apparatus according to Japanese Unexamined Patent Application Publication No. 2024-010623 executes data assimilation processing based on an estimated value obtained by a predetermined physical property simulation and a measured value obtained by measuring a substance. By this data assimilation processing, the information processing apparatus generates a second estimated value. 【Prior Art Documents】 【Patent Documents】 【0003】 【Patent Document 1】 Japanese Unexamined Patent Application Publication No. 2024-010623 【Summary of the Invention】 【Problems to be Solved by the Invention】 【0004】 In view of estimating a deterioration index of a vehicle, a better information processing apparatus, information processing method, and program are desired. 【0005】 The present disclosure aims to solve the above-described problems. 【Means for Solving the Problems】 【0006】 A first aspect of the present disclosure includes a pre-estimation data acquisition unit that acquires pre-estimation data of a deterioration index, which is an index related to deterioration of a vehicle; an observation data acquisition unit that acquires observation data of the deterioration index; and a data assimilation processing unit that derives post-estimation data of the deterioration index by executing data assimilation processing based on the pre-estimation data and the observation data. 【0007】 A second aspect of this disclosure is an information processing method comprising: a pre-estimate data acquisition step of acquiring pre-estimate data of a deterioration index, which is an index related to the deterioration of a vehicle; an observation data acquisition step of acquiring observation data of the deterioration index; and a data assimilation processing step of deriving post-estimate data of the deterioration index by performing data assimilation processing based on the pre-estimate data and the observation data. 【0008】 A third aspect of this disclosure is a program for causing a computer to execute the information processing method relating to the second aspect of this disclosure. [Effects of the Invention] 【0009】 This disclosure provides a better information processing device, information processing method, and program in terms of estimating vehicle degradation indicators. [Brief explanation of the drawing] 【0010】 [Figure 1] Figure 1 is a block diagram showing the configuration of an information processing device according to one embodiment. [Figure 2] Figure 2A is a graph showing an example of traceability data. Figure 2B is a graph showing another example of traceability data. [Figure 3] Figure 3 is a schematic diagram illustrating the process by which the pre-estimated data candidate generation unit generates multiple pre-estimated data candidates. [Figure 4] Figure 4 is a graph illustrating multiple estimated parameters that constitute a candidate pre-estimated data set before correction, and multiple observed parameters that constitute the observed data. [Figure 5] Figure 5 is a graph illustrating multiple estimated parameters that constitute a corrected pre-estimated data candidate, and multiple observed parameters that constitute the observed data. [Figure 6] Figure 6 is a graph illustrating the data assimilation process performed by the data assimilation processing unit. [Figure 7] Figure 7 is a flowchart showing an information processing method according to one embodiment. [Figure 8] Figure 8 is a flowchart showing the method for generating pre-estimated data candidates included in the information processing method. [Figure 9] Figure 9 is a flowchart showing the degradation index estimation method included in the information processing method. [Figure 10] Figure 10 is a block diagram showing the configuration of the information processing device according to Modification Example 1. [Modes for carrying out the invention] 【0011】 Vehicles may be equipped with various sensors. These sensors can acquire information indicating various parameters (state quantities, physical quantities, etc.) related to the vehicle's degradation index. The vehicle's degradation index is an index related to the degradation of the vehicle. This degradation index may be, for example, an index related to the degradation of the power source (power plant) installed in the vehicle. The vehicle's power source includes, for example, a fuel cell. That is, for example, the degree of degradation of the electrolyte membrane of a fuel cell installed in a fuel cell vehicle (FCV) (degradation rate of the electrolyte membrane), the degree of degradation of the catalyst layer bonded to the electrolyte membrane (degradation rate of the catalyst layer), etc., are examples of degradation indexes. 【0012】 The various parameters mentioned above can be transmitted from the vehicle to a server or other device as traceability data. It is conceivable to estimate the degradation index corresponding to the traceability data transmitted from the vehicle using a machine learning model (a trained model) that has been pre-trained on the relationship between traceability data and the vehicle degradation index. In this case, in order to achieve good estimation accuracy, the input traceability data must have a certain degree of statistical reliability or stability. In this respect, traceability data obtained from vehicle models (vehicles) immediately after release is often insufficient to ensure statistical reliability or stability. Due to these circumstances, it has been difficult to estimate the degradation index of vehicles immediately after release with good accuracy. 【0013】 Based on the above preliminary explanation, one embodiment will be described below. 【0014】 The program (computer program, computer software) in the following description is also referred to as a computer program product. A computer program product is not limited to a program recorded on a recording medium, but also includes a program transmitted, distributed, or downloaded via a network such as the Internet. 【0015】 The target vehicle in the following description is a vehicle for which a deterioration index is estimated by data assimilation processing by an information processing apparatus (data assimilation processing unit) described later. In contrast, the monitoring vehicle in the following description is a vehicle that provides data (traceability data) used for data clustering described later to an information processing apparatus (user classification setting unit). The target vehicle and the monitoring vehicle may be different. The user of the target vehicle and the user of the monitoring vehicle may be different. In the following description, the user of the monitoring vehicle is also referred to as a monitoring user. In one embodiment, a case where each of the target vehicle and the monitoring vehicle is a fuel cell vehicle (FCV) will be described, but each of the target vehicle and the monitoring vehicle is not limited to an FCV (see also the modification example described later). 【0016】 (One embodiment) FIG. 1 is a block diagram showing the configuration of an information processing apparatus 10 according to one embodiment. 【0017】 The information processing apparatus 10 is an apparatus (deterioration index estimation apparatus, estimation apparatus) that estimates the deterioration index of a target vehicle using data assimilation processing. The information processing apparatus 10 is included in an electronic apparatus such as a computer, for example. The target vehicle is an FCV. The FCV as the target vehicle is equipped with a fuel cell as a power source. The deterioration index of the FCV is, as described above, for example, the degree of deterioration of the electrolyte membrane of the fuel cell, the degree of deterioration of the catalyst layer joined to the electrolyte membrane, and the like. As shown in FIG. 1, the information processing apparatus 10 includes a communication unit 12, a storage unit 14, and an arithmetic unit 16. 【0018】 The communication unit 12 includes a communication module (not shown). The communication unit 12 performs communication via a network. This network may include the Internet. 【0019】 The storage unit 14 includes one or more recording media (not shown). For example, the storage unit 14 includes one or more non-volatile memories. The storage unit 14 may further include one or more volatile memories. At least a part of the storage unit 14 may be a portable recording medium. The portable recording medium may be, for example, a memory card, a USB flash memory (USB: Universal Serial Bus), an optical disk, etc., but is not limited thereto. 【0020】 The arithmetic unit 16 includes a processing circuit (not shown). This processing circuit may have an IC (Integrated Circuit) or may have a discrete circuit. At least a part of the arithmetic unit 16 (processing circuit) may be a processor such as a CPU (Central Processing Unit), a GPU (Graphical Processing Unit), etc. The arithmetic unit 16 may include one or more processors. 【0021】 The calculation unit 16 includes a traceability data acquisition unit 18, a user classification setting unit 20, a pre-estimated data candidate generation unit 22, an observation data acquisition unit 24, a pre-estimated data acquisition unit 26, and a data assimilation processing unit 28. The traceability data acquisition unit 18, the user classification setting unit 20, the pre-estimated data candidate generation unit 22, the observation data acquisition unit 24, the pre-estimated data acquisition unit 26, and the data assimilation processing unit 28 are realized by the calculation unit 16 (processing circuit). For example, one or more processors included in the calculation unit 16 execute a program stored in the storage unit 14. This may realize at least a part of the traceability data acquisition unit 18, the user classification setting unit 20, the pre-estimated data candidate generation unit 22, the observation data acquisition unit 24, the pre-estimated data acquisition unit 26, and the data assimilation processing unit 28. The aforementioned IC or discrete circuit may implement at least a part of the traceability data acquisition unit 18, the user classification setting unit 20, the pre-estimated data candidate generation unit 22, the observation data acquisition unit 24, the pre-estimated data acquisition unit 26, and the data assimilation processing unit 28. 【0022】 The traceability data acquisition unit 18 acquires traceability data of the vehicle (target vehicle, monitoring vehicle). Traceability data is transmitted from the vehicle. For example, a communication device (not shown) may be provided in the vehicle. Traceability data may be transmitted from this communication device. The traceability data acquisition unit 18 can acquire traceability data transmitted from the vehicle via the communication unit 12. The traceability data acquisition unit 18 may also acquire traceability data via a predetermined server that stores traceability data transmitted from the vehicle. 【0023】 Traceability data includes various types of information detected by various sensors installed in the vehicle. This information may include various parameters (state quantities, physical quantities, etc.) related to degradation indicators. These parameters related to degradation indicators are also referred to as degradation-related parameters in the following explanation. For example, the current density (output current), operating pressure, hydrogen flow rate, air flow rate, voltage (output voltage), temperature, and humidity of an FCV's fuel cell can all be affected by fuel cell degradation. Therefore, the output current, operating pressure, hydrogen flow rate, air flow rate, voltage, temperature, and humidity of an FCV's fuel cell can each be considered a degradation-related parameter. In that case, the traceability data may be data related to at least one of the above-mentioned output current, operating pressure, hydrogen flow rate, air flow rate, voltage, temperature, humidity, etc. 【0024】 Figure 2A is a graph showing an example of traceability data. 【0025】 The traceability data may also be data relating to the frequency distribution of various degradation-related parameters. In this case, the traceability data represents, for example, the frequency of occurrence of each value that various degradation-related parameters could take over a predetermined period. The predetermined period is, for example, 24 hours (1 day), but is not limited to this. The predetermined period may be, for example, 7 days (1 week), or it may be the duration of one vehicle use. The duration of one vehicle use is the period from when the vehicle is ignited on (power turned on) until the vehicle is ignited off (power turned off). The traceability data acquisition unit 18 may acquire new traceability data each time the predetermined period has elapsed. Figure 2A illustrates the frequency distribution of the output current of a fuel cell, which is one of the degradation-related parameters. 【0026】 Figure 2B is a graph showing another example of traceability data. 【0027】 Traceability data may also be data relating to the time-series changes of various degradation-related parameters. In this case, the traceability data would represent, for example, the changes in various degradation-related parameters over a predetermined period. The predetermined period could be one day, one week, the duration of one vehicle use (as described above), etc., but is not limited to these. Figure 2B illustrates the time-series change in the output current of a fuel cell. 【0028】 The user classification setting unit 20 sets up multiple user classification USRs based on traceability data obtained from each of the multiple monitoring vehicles. For example, the user classification setting unit 20 performs data clustering (clustering) based on the traceability data obtained from each of the multiple monitoring vehicles. As a result, the user classification setting unit 20 sets up multiple user classification USRs (clusters) that represent different usage patterns of the monitoring vehicles. In other words, by performing clustering based on traceability data, the user classification setting unit 20 sets up multiple user classification USRs that represent different degrees of vehicle deterioration. 【0029】 The user classification setting unit 20 may perform the clustering described above using a clustering algorithm, which is an unsupervised machine learning algorithm. The clustering algorithm used by the user classification setting unit 20 is, for example, a non-hierarchical clustering algorithm such as the k-means method, but is not limited thereto. For example, a hierarchical clustering algorithm may be used by the user classification setting unit 20. 【0030】 The number of user classification USRs (USR1 to USR6) that can be set by the user classification setting unit 20 may be predetermined. This predetermined number is any natural number greater than or equal to 2. For example, the predetermined number is 6 (see also Figure 3), but is not limited to this. Based on the predetermined number, the user classification setting unit 20 can set a predetermined number of user classification USRs (USR1 to USR6). 【0031】 The user classification setting unit 20 may perform a predetermined data processing as a preprocessing step for the clustering described above. For example, if the traceability data is data relating to the time-series changes of degradation-related parameters, the user classification setting unit 20 may process the data to generate data relating to the frequency distribution (described above) of the degradation-related parameters. In that case, the user classification setting unit 20 may perform clustering using the data generated by the predetermined data processing (processed traceability data). 【0032】 The pre-estimation data candidate generation unit 22 generates a plurality of pre-estimation data candidate PECs. Each of the plurality of pre-estimation data candidate PECs is a candidate for pre-estimation data PE used in the data assimilation process to estimate the degradation index. The pre-estimation data candidate generation unit 22 generates pre-estimation data candidate PECs corresponding to each of the plurality of user classifications USR set by the user classification setting unit 20. Therefore, if the number of user classifications USR set by the user classification setting unit 20 is a predetermined number, the number of pre-estimation data candidate PECs generated by the pre-estimation data candidate generation unit 22 is also a predetermined number. For example, if the predetermined number is 6, the pre-estimation data candidate generation unit 22 generates 6 pre-estimation data candidate PECs. 【0033】 Figure 3 is a schematic diagram illustrating the process by which the pre-estimated data candidate generation unit 22 generates multiple pre-estimated data candidates PECs. 【0034】 The pre-estimation data candidate generation unit 22 generates multiple pre-estimation data candidate PECs using the estimation model MDL. The estimation model MDL is a mathematical model that outputs estimated data of a degradation index corresponding to the input data DA. The estimation model MDL is designed, for example, based on experiments. The estimation model MDL may be stored in the storage unit 14. The output estimated data may be data that shows multiple changes in the degradation index using a probability distribution (normal distribution). The estimation model MDL may also be a learned model (trained model) obtained by machine learning such as supervised learning. 【0035】 Simulated traceability data can be used as the input data DA to the estimation model MDL. Simulated traceability data is data that mimics traceability data. For example, simulated traceability data shows the frequency distribution of at least one degradation-related parameter. In that case, the estimation model MDL outputs estimated data of a degradation index corresponding to the frequency distribution of that at least one degradation-related parameter. 【0036】 The simulated traceability data corresponding to the user classification USR may be determined based on statistics of traceability data of monitoring vehicles owned by each of the multiple monitoring users belonging to that user classification USR. The simulated traceability data may be calculated by the pre-estimated data candidate generation unit 22. The user of the information processing device 10 may teach the pre-estimated data candidate generation unit 22 the simulated traceability data via an operation unit (not shown) provided in the information processing device 10. This operation unit may include, for example, an input device such as a keyboard or a pointing device. 【0037】 As input data DA corresponding to the user classification USR, traceability data of a monitoring vehicle owned by a monitoring user belonging to that user classification USR may be used. That monitoring user may be selected randomly or based on predetermined selection criteria by the pre-estimated data candidate generation unit 22. The user of the information processing device 10 may teach the pre-estimated data candidate generation unit 22 a monitoring user via an operation unit (not shown above) provided on the information processing device 10. 【0038】 For example, Figure 3 shows multiple input data sets DA (DA1 to DA6). Also, Figure 3 shows multiple pre-estimated data candidate sets PEC (PEC1 to PEC6). 【0039】 Input data DA1 is, for example, input data DA prepared according to the first user classification USR1. When this input data DA1 is input into the estimation model MDL, a candidate pre-estimated data PEC1 corresponding to the first user classification USR1 is obtained. 【0040】 Input data DA2 is input data DA prepared according to, for example, a second user classification USR2. When this input data DA2 is input into the estimation model MDL, a candidate pre-estimated data PEC2 corresponding to the second user classification USR2 is obtained. 【0041】 Input data DA3 is input data DA prepared according to, for example, a third user classification USR3. When this input data DA3 is input into the estimation model MDL, a candidate pre-estimated data PEC3 corresponding to the third user classification USR3 is obtained. 【0042】 The input data DA4 is, for example, the input data DA prepared according to the fourth user classification USR4. When this input data DA4 is input into the estimation model MDL, a candidate pre-estimated data PEC4 corresponding to the fourth user classification USR4 is obtained. 【0043】 Input data DA5 is input data DA prepared according to, for example, the fifth user classification USR5. When this input data DA5 is input into the estimation model MDL, candidate pre-estimated data PEC5 corresponding to the fifth user classification USR5 is obtained. 【0044】 The input data DA6 is, for example, the input data DA prepared according to the sixth user classification USR6. When this input data DA6 is input into the estimation model MDL, a candidate pre-estimated data PEC6 corresponding to the sixth user classification USR6 is obtained. 【0045】 The observation data acquisition unit 24 acquires observation data RD, which is used in data assimilation processing to estimate the deterioration index. Specifically, the observation data acquisition unit 24 acquires observation data RD of the deterioration index of the target vehicle. 【0046】 Observational data RD of the deterioration index can be obtained based on the traceability data of the target vehicle. For example, the observational data acquisition unit 24 may calculate the observational data RD of the deterioration index of the target vehicle based on various deterioration-related parameters indicated by the traceability data of the target vehicle. In this way, the observational data acquisition unit 24 can obtain the observational data RD of the deterioration index of the target vehicle. 【0047】 The target vehicle may include information indicating its degradation index in its traceability data. For example, an electronic device such as an ECU (Electronic Control Unit) installed in the target vehicle may calculate the degradation index of the target vehicle based on the output current of the fuel cell of the target vehicle. In that case, since the traceability data acquisition unit 18 acquires traceability data that includes information indicating the degradation index, the observation data acquisition unit 24 can extract the degradation index from that traceability data. 【0048】 The pre-estimated data acquisition unit 26 includes a pre-estimated data determination unit 30. The pre-estimated data determination unit 30 determines the pre-estimated data PE, which is used in the data assimilation process to estimate the deterioration index of the target vehicle, based on the observation data RD obtained from the target vehicle. More specifically, the pre-estimated data determination unit 30 determines the pre-estimated data PE as described below. 【0049】 The pre-estimation data determination unit 30 compares the observation data RD acquired by the observation data acquisition unit 24 with each of the multiple pre-estimation data candidate PECs. This allows the pre-estimation data determination unit 30 to identify one pre-estimation data candidate PEC from among the multiple pre-estimation data candidate PECs that has the smallest degree of deviation from the observation data RD. If the degree of deviation (minimum deviation) between the identified pre-estimation data candidate PEC and the observation data RD is less than a predetermined degree, the pre-estimation data determination unit 30 determines that pre-estimation data candidate PEC as the pre-estimation data PE. 【0050】 As described above, the candidate pre-estimation data PEC is the estimated data for the degradation index by the estimation model MDL. The estimated data may also be data that shows multiple predicted changes in the degradation index using a probability distribution. In that case, the degree of discrepancy between the candidate pre-estimation data PEC and the observed data RD can be derived, for example, by comparing the mode of the probability distribution shown by the candidate pre-estimation data PEC with the degradation index shown by the observed data RD. 【0051】 If the minimum deviation is greater than or equal to a predetermined degree, the pre-estimation data determination unit 30 corrects the identified pre-estimation data candidate PEC based on the observed data RD. In this case, the pre-estimation data determination unit 30 determines the corrected pre-estimation data candidate PEC as the pre-estimation data PE. The pre-estimation data determination unit 30 corrects the pre-estimation data candidate PEC as described below, for example. 【0052】 Figure 4 is a graph illustrating multiple estimated parameters that constitute a pre-predicted data candidate PEC before correction, and multiple observation parameters that constitute the observed data RD. 【0053】 Figure 4 schematically shows, by dashed lines, multiple observation parameters (frequency distributions) that constitute the observation data RD, which are multiple degradation-related parameters. These multiple observation parameters consist, for example, a first observation parameter A1 indicating the output current of the fuel cell, a second observation parameter A2 indicating the temperature of the fuel cell, and a third observation parameter A3 indicating the humidity of the fuel cell. These multiple observation parameters can be shown by the traceability data used when the observation data RD is calculated by the observation data acquisition unit 24 described above. 【0054】 Furthermore, Figure 4 schematically shows, with solid lines, multiple estimation parameters (frequency distributions) that constitute one pre-estimated data candidate PEC before correction, which are multiple degradation-related parameters. These multiple estimation parameters consist, for example, a first estimation parameter B1 indicating the output current of the fuel cell, a second estimation parameter B2 indicating the temperature of the fuel cell, and a third estimation parameter B3 indicating the humidity of the fuel cell. These multiple estimation parameters are obtained by reusing at least a portion of the input data DA (simulated traceability data, traceability data) used when one pre-estimated data PE is generated by the pre-estimated data candidate generation unit 22 as multiple estimation parameters. 【0055】 The pre-estimation data determination unit 30 compares each of the multiple observed parameters with the estimated parameter corresponding to that degradation-related parameter. More specifically, the pre-estimation data determination unit 30 compares a first observed parameter A1, which is a parameter relating to the output current of the fuel cell, with a first estimated parameter B1, which is also a parameter relating to the output current of the fuel cell, similar to the first observed parameter A1. Similarly, the pre-estimation data determination unit 30 compares a second observed parameter A2 with a second estimated parameter B2. Furthermore, the pre-estimation data determination unit 30 compares a third observed parameter A3 with a third estimated parameter B3. Through this, the pre-estimation data determination unit 30 identifies the degree of discrepancy between each of the multiple observed parameters and the estimated parameter corresponding to that observed parameter. This degree of discrepancy is identified, for example, based on the deviation between the mode of the observed parameter and the mode of the estimated parameter. 【0056】 The pre-estimation data determination unit 30 determines whether the degree of deviation identified for each of the multiple observation parameters is less than a predetermined tolerance. The predetermined tolerance can be predetermined for each pair of observation parameters and estimated parameters. If an estimated parameter whose corresponding degree of deviation is greater than or equal to the predetermined tolerance is identified by this determination, the pre-estimation data determination unit 30 corrects the estimated parameter. More specifically, the pre-estimation data determination unit 30 corrects the estimated parameter so that the degree of deviation, which was greater than or equal to the predetermined tolerance, becomes less than the predetermined tolerance. As a result, the estimated parameter to be corrected is changed so that it shows a value close to that of the corresponding observation parameter. 【0057】 Figure 5 is a graph illustrating the multiple estimation parameters that constitute a corrected pre-estimated data candidate PEC, and the multiple observation parameters that constitute the observed data RD. 【0058】 For example, let's consider the case where the degree of deviation between the third observed parameter A3 and the third estimated parameter B3, as shown in Figure 4, is within a predetermined tolerance. In this case, the pre-estimation data determination unit 30 corrects the third estimated parameter B3 so that the degree of deviation between the third observed parameter A3 and the third estimated parameter B3 is less than the predetermined tolerance. For example, the pre-estimation data determination unit 30 may shift the distribution of the third estimated parameter B3 along the horizontal axis of Figure 5 until the mode of the third estimated parameter B3 matches the mode of the third observed parameter A3. As a result, as shown in Figure 5, the corrected third estimated parameter B3 (B3b) is changed to show a value (distribution) close to that of the third observed parameter A3. 【0059】 The pre-estimation data determination unit 30 re-derives candidate pre-estimation data PEC based on a plurality of estimation parameters, including the corrected estimation parameters. This derives a candidate pre-estimation data PEC corrected based on the observed data RD (corrected candidate pre-estimation data PEC). The degree of deviation between the corrected candidate pre-estimation data PEC and the observed data RD can be reduced compared to the degree of deviation between the uncorrected candidate pre-estimation data PEC and the observed data RD. 【0060】 The pre-estimation data acquisition unit 26 further includes a feedback acquisition unit 32. The feedback acquisition unit 32 acquires the post-estimation data ES of the degradation index derived by the data assimilation processing unit 28 as new pre-estimation data PE. A more detailed explanation of the data assimilation processing unit 28 will be given later. 【0061】 Figure 6 is a graph illustrating the data assimilation process performed by the data assimilation processing unit 28. 【0062】 The data assimilation processing unit 28 performs data assimilation processing based on the pre-estimated data PE of the degradation index acquired by the pre-estimated data acquisition unit 26 and the observed data RD of the degradation index acquired by the observed data acquisition unit 24. Through this process, the data assimilation processing unit 28 derives the post-estimated data ES of the degradation index. The data assimilation processing unit 28 may also perform the above-mentioned data assimilation processing using a predetermined filter (filtering process) such as a Kalman filter or an ensemble Kalman filter. 【0063】 The post-hoc estimate data ES derived by the data assimilation processing unit 28 may be data that shows multiple predicted changes in the degradation index using a probability distribution (normal distribution). In that case, the deviation between the mode of the post-hoc estimate data ES and the observed data may be smaller than the deviation between the mode of the pre-hoc estimate data PE and the observed data (see also Figure 6). 【0064】 The data assimilation processing unit 28 can recursively (sequentially) derive post-hoc estimate data ES of the degradation index. For example, the data assimilation processing unit 28 can derive first post-hoc estimate data ES of the degradation index based on the prior estimate data PE determined by the prior estimate data determination unit 30 (prior estimate data acquisition unit 26) and the observed data RD. The first post-hoc estimate data ES can be acquired as new prior estimate data PE by the feedback acquisition unit 32 (prior estimate data acquisition unit 26) described above. As a result, the data assimilation processing unit 28 can further perform data assimilation processing based on the new prior estimate data PE and the observed data RD (new observed data RD) newly acquired by the observation data acquisition unit 24. Through this data assimilation processing, the data assimilation processing unit 28 can derive a second post-hoc estimate data ES, which is new post-hoc estimate data ES. The second post-hoc estimate data ES can be acquired as yet another new prior estimate data PE by the feedback acquisition unit 32. 【0065】 Figure 7 is a flowchart showing an information processing method according to one embodiment. 【0066】 The information processing device 10 described above is capable of executing the information processing method shown in Figure 7. For example, a processor provided in the information processing device 10, which is a computer, executes a predetermined program stored in memory. As a result, the information processing device 10 executes the information processing method shown in Figure 7. As shown in Figure 7, the information processing method includes a pre-estimated data candidate generation method M1 and a degradation index estimation method M2. 【0067】 Figure 8 is a flowchart showing the pre-estimated data candidate generation method M1 included in the information processing method. 【0068】 The pre-estimated data candidate generation method M1 comprises a traceability data acquisition step S1, a user classification setting step S2, and a pre-estimated data candidate generation step S3. 【0069】 In the traceability data acquisition step S1, the traceability data acquisition unit 18 acquires traceability data from multiple monitoring vehicles. 【0070】 In the user classification setting step S2, the user classification setting unit 20 sets multiple user classifications (USRs) based on the multiple traceability data acquired in the traceability data acquisition step S1. 【0071】 In the pre-estimated data candidate generation step S3, the pre-estimated data candidate generation unit 22 generates multiple pre-estimated data candidate PECs corresponding to each of the multiple user classifications USR set in the user classification setting step S2. 【0072】 Figure 9 is a flowchart showing the degradation index estimation method M2 included in the information processing method. 【0073】 The degradation index estimation method M2 includes an observation data acquisition step S4, a prior estimation data acquisition step S5, and a data assimilation processing step S6. 【0074】 In observation data acquisition step S4, the observation data acquisition unit 24 acquires observation data RD of the deterioration index of the target vehicle. In observation data acquisition step S4, the traceability data acquisition unit 18 may acquire traceability data of the target vehicle, and the observation data acquisition unit 24 may acquire observation data RD of the target vehicle based on said traceability data. 【0075】 In the pre-estimated data acquisition step S5, the pre-estimated data acquisition unit 26 acquires the pre-estimated data PE to be used in the next data assimilation processing step S6. The pre-estimated data acquisition step S5 includes a feedback feasibility determination step S51, a pre-estimated data determination step S52, and a feedback acquisition step S53. As will be described in detail later, the post-estimated data ES of the degradation index is derived by executing the data assimilation processing step S6. In the feedback feasibility determination step S51, the pre-estimated data acquisition unit 26 determines whether or not this post-estimated data ES has been derived. 【0076】 If the post-hoc estimation data ES has not been derived (S51:NO), the pre-estimation data determination step S52 is executed. In the pre-estimation data determination step S52, the pre-estimation data determination unit 30 (pre-estimation data acquisition unit 26) acquires a pre-prepared pre-estimation data candidate PEC as pre-estimation data PE. More specifically, in the pre-estimation data determination step S52, the pre-estimation data determination unit 30 selects one pre-estimation data candidate PEC from among a plurality of pre-estimation data candidate PECs based on the observed data RD. The selected one pre-estimation data candidate PEC is determined to be the pre-estimation data PE. That is, the said one pre-estimation data candidate PEC is acquired as the pre-estimation data PE. 【0077】 In the data assimilation processing step S6, the data assimilation processing unit 28 performs data assimilation processing based on the observed data RD and the prior estimated data PE. As a result, in the data assimilation processing step S6, the post-hoc estimated data ES of the degradation index is derived. 【0078】 After the data assimilation processing step S6 is completed, the observation data acquisition step S4, the prior estimation data acquisition step S5, and the data assimilation processing step S6 may be executed repeatedly. Once the data assimilation processing step S6 has been executed, the feedback feasibility determination step S51 determines that the post-estimation data ES has been derived (S51: YES). Therefore, after the data assimilation processing step S6 is completed, the feedback acquisition step S53 may be executed. 【0079】 In the feedback acquisition step S53, the feedback acquisition unit 32 (pre-estimated data acquisition unit 26) acquires the post-estimated data ES as new pre-estimated data PE. In this case, the observation data acquisition step S4 may be performed at predetermined intervals as described above. For example, if the predetermined period is 1 day, the observation data RD acquired based on the traceability data of the nth day may be acquired by the nth observation data acquisition step S4 (n: natural number). In that case, in the (n+1)th observation data acquisition step S4, new observation data RD based on the traceability data of the (n+1)th day is acquired. 【0080】 The information processing device 10 described above provides the following effects, for example. 【0081】 The information processing device 10 includes a pre-estimated data determination unit 30. The pre-estimated data determination unit 30 selects pre-estimated data PE, which is used for data assimilation processing to estimate the degradation index, from among a plurality of pre-estimated data candidate PECs based on the observed data RD. As a result, the information processing device 10 can perform data assimilation processing using the selected pre-estimated data PE. Consequently, the need to perform simulations with a huge number of patterns is reduced, and the processing load of the information processing device 10 can be suppressed. 【0082】 The pre-estimation data determination unit 30 selects a candidate pre-estimation data PEC, which has the smallest degree of deviation from the observed data RD, as the pre-estimation data PE. This rapidly improves the estimation accuracy of the degradation index by data assimilation processing. In this case, if the degree of deviation between the candidate pre-estimation data PEC and the observed data RD is less than a predetermined degree, the pre-estimation data determination unit 30 determines the candidate pre-estimation data PEC as the pre-estimation data PE. On the other hand, if the degree of deviation between the candidate pre-estimation data PEC and the observed data RD is greater than or equal to a predetermined degree, the pre-estimation data determination unit 30 obtains the pre-estimation data PE by correcting the candidate pre-estimation data PEC based on the observed data RD. This further rapidly improves the estimation accuracy of the degradation index by data assimilation processing. Furthermore, according to this embodiment, the execution condition for the correction process for the candidate pre-estimation data PEC is that the degree of deviation between the candidate pre-estimation data PEC and the observed data RD is greater than or equal to a predetermined degree. Therefore, the frequency of performing this correction can be reduced. As a result, the processing load on the information processing device 10 can be reduced. 【0083】 The pre-estimated data candidate PEC is prepared to correspond to multiple user classifications USR, each representing a different vehicle usage pattern. This allows the information processing device 10 to select pre-estimated data PE according to the degree of vehicle deterioration and perform data assimilation processing. As a result, the information processing device 10 achieves good accuracy in estimating deterioration indicators through data assimilation processing. 【0084】 The information processing device 10 includes a pre-estimated data candidate generation unit 22. The pre-estimated data candidate generation unit 22 generates the above-mentioned multiple pre-estimated data candidate PECs using the estimation model MDL. The estimation model MDL may reflect knowledge obtained during the vehicle design stage, etc. Therefore, the multiple pre-estimated data candidate PECs generated by the pre-estimated data candidate generation unit 22 may also reflect knowledge obtained during the vehicle design stage, etc. 【0085】 The information processing device 10 includes a user classification setting unit 20. The user classification setting unit 20 sets multiple user classifications (USRs) based on traceability data obtained from multiple monitoring vehicles. This allows for the setting of multiple user classifications (USRs) that accurately reflect the differences in vehicle usage patterns. 【0086】 The vehicle may be an FCV (Fuel Cell Vehicle). An FCV is equipped with a fuel cell. In that case, the traceability data may be data relating to at least one of the following: the fuel cell's output current, the fuel cell's operating pressure, the fuel cell's hydrogen flow rate, the fuel cell's air flow rate, the fuel cell's voltage, the fuel cell's temperature, and the fuel cell's humidity. Thus, the vehicle degradation index can be estimated as the degree of degradation of the fuel cell of the vehicle (FCV). 【0087】 The information processing device 10 includes a data assimilation processing unit 28. The data assimilation processing unit 28 performs the above-mentioned data assimilation processing based on the observed data RD and the prior estimated data PE to derive the post-hoc estimated data ES of the degradation index. If the degradation index is estimated using only the observed data RD (traceability data), a certain amount of data is required to improve the estimation accuracy in order to ensure the statistical reliability or stability of the observed data RD. In this respect, since the data assimilation processing unit 28 estimates the degradation index by data assimilation processing using the prior estimated data PE, it is possible to achieve better estimation accuracy earlier than when the degradation index is estimated using only estimated data. Furthermore, since the data assimilation processing unit 28 estimates the degradation index by data assimilation processing using the observed data RD, it is easier to obtain estimation results that are more in line with the user's usage than when the estimation of the degradation index is performed by simulation alone. 【0088】 The pre-estimated data acquisition unit 26 of the information processing device 10 can acquire new pre-estimated data PE of the degradation index based on the post-estimated data ES derived by the data assimilation processing unit 28. In this case, the data assimilation processing unit 28 can derive new post-estimated data ES by performing data assimilation processing based on the new pre-estimated data PE and the new observation data RD acquired by the observation data acquisition unit 24. In this way, the information processing device 10 can perform data assimilation processing sequentially. In this case, the new pre-estimated data PE reflects the results of the data assimilation processing that has been performed at the time the new pre-estimated data PE is acquired, so the reliability of the new pre-estimated data PE is good. In addition, the statistical reliability or stability of the observation data RD can be ensured by repeatedly acquiring the observation data RD of the target vehicle. As a result, the reliability of the post-estimated data ES can be improved as the data assimilation processing is repeated. 【0089】 As described above, if the data assimilation processing unit 28 has not derived the post-estimation data ES, the pre-estimation data acquisition unit 26 (pre-estimation data determination unit 30) acquires a pre-prepared pre-estimation data candidate PEC as the pre-estimation data PE. This allows the information processing device 10 to perform data assimilation processing even if it is not possible to feed back the post-estimation data ES as the pre-estimation data PE. 【0090】 One embodiment may be modified as shown in the following examples. Descriptions that overlap with the first embodiment will be omitted as appropriate below. Furthermore, among the configurations described below, those identical to those described in the first embodiment will be denoted by the same reference numerals as in the first embodiment. 【0091】 (Variation 1) Figure 10 is a block diagram showing the configuration of the information processing device 10 (information processing device 101) according to Modification Example 1. 【0092】 Compared to the information processing device 10 of one embodiment (Figure 1), the information processing device 101 differs in that it further includes a pre-estimation and evaluation unit 34 and an estimated model update unit 36. The pre-estimation and evaluation unit 34 and the estimated model update unit 36 ​​are implemented by a processing circuit (calculation unit 16), similar to the traceability data acquisition unit 18 and the user classification setting unit 20. 【0093】 The pre-estimation evaluation unit 34 compares the first distribution with the second distribution. The first distribution is the probability distribution (normal distribution) of the degradation index obtained based on multiple observation data RD. Multiple observation data RD can be cumulatively stored by the storage unit 14 as the observation data acquisition step S4 is repeatedly executed. The first distribution is obtained by performing predetermined statistical processing based on these multiple observation data RD. The second distribution, which is compared with this first distribution, is the probability distribution (normal distribution) of the degradation index indicated by the pre-estimation data PE. The second distribution is indicated, for example, by the pre-estimation data candidate PEC selected as the pre-estimation data PE by the pre-estimation data determination unit 30. The pre-estimation evaluation unit 34 may compare the first distribution with the second distribution only if predetermined execution conditions are met. The pre-estimation evaluation unit 34 may determine that predetermined execution conditions are met, for example, when the degree of deviation between the probability distribution of each of the multiple estimation parameters and the probability distribution of the observation parameter corresponding to the estimation parameter is within an acceptable range. 【0094】 The pre-estimation evaluation unit 34 determines whether it is necessary to update the estimation model MDL based on whether the degree of deviation between the first distribution and the second distribution is within a predetermined acceptable range. This estimation model MDL is a mathematical model that generates multiple pre-estimation data candidate PECs (see also one embodiment). The degree of deviation between the first distribution and the second distribution can be determined, for example, based on a comparison between the mode of the first distribution and the mode of the second distribution. If the degree of deviation between the first distribution and the second distribution is within an acceptable range, the pre-estimation evaluation unit 34 determines that it is not necessary to update the estimation model MDL. If the degree of deviation between the first distribution and the second distribution is outside an acceptable range, the pre-estimation evaluation unit 34 determines that it is necessary to update the estimation model MDL. 【0095】 The estimation model update unit 36 ​​updates the estimation model MDL based on the first distribution when the pre-estimation evaluation unit 34 determines that it is necessary to update the estimation model MDL. The estimation model update unit 36 ​​updates the estimation model MDL with the goal of reducing the degree of divergence between the first distribution and the second distribution. In this case, the estimation model update unit 36 ​​may update the estimation model MDL, which is a mathematical model, by performing regression analysis based on, for example, the least squares method. 【0096】 The estimated model MDL may be a trained model obtained by machine learning. In that case, the estimated model update unit 36 ​​may update the estimated model MDL based on a machine learning algorithm that can be applied to supervised learning. In that case, the first distribution may be used as the training data. 【0097】 According to this modified version, if the second distribution, which is a probability distribution based on estimation (prior estimated data PE), deviates excessively from the first distribution, which is a probability distribution based on actual data (observed data RD), the estimation model MDL is corrected based on the actual data. As a result, the prior estimated data acquisition unit 26 can obtain prior estimated data PE with better estimation accuracy using the updated estimation model MDL. Furthermore, according to this modified version, the estimation model MDL can be updated only when predetermined execution conditions are met and the first and second distributions deviate excessively. The predetermined execution conditions can be determined to be met when the degree of deviation between the probability distribution of each of the multiple estimated parameters and the probability distribution of the observed parameter corresponding to the estimated parameter is within an acceptable range. Even if the degree of deviation between the probability distribution of each of the multiple estimated parameters and the probability distribution of the observed parameter corresponding to the estimated parameter is within an acceptable range, if the first and second distributions deviate excessively, there is a relatively high possibility that the estimation model MDL is not properly constructed. In other words, according to this modified version, the estimated model MDL can be updated when there is a relatively high possibility that the estimated model MDL is not properly constructed. 【0098】 If the estimation model MDL is updated, the data assimilation processing unit 28 may perform data assimilation processing using the pre-estimated data PE estimated by the updated estimation model MDL. This may improve the estimation accuracy of the data assimilation processing. 【0099】 (Modification 2) Vehicle traceability data may include data relating to at least one of the following: vehicle speed, acceleration, and usage time. Usage time is the time from when the vehicle's ignition is turned on until it is turned off. 【0100】 The information processing device 10 can estimate vehicle degradation indicators related to vehicle speed, acceleration, and usage time by using traceability data related to vehicle speed, acceleration, and usage time. Furthermore, traceability data related to vehicle speed, acceleration, and usage time can be obtained from vehicle types other than FCVs. Therefore, according to this modified version, the information processing device 10 can also be applied to estimate degradation indicators for vehicles other than FCVs. 【0101】 (Variation 3) According to one embodiment, the traceability data transmitted from the target vehicle may include information indicating the deterioration index of that vehicle. Similarly, the traceability data transmitted from the monitoring vehicle may include information indicating the deterioration index of that monitoring vehicle. In this case, the information indicating the deterioration index included in the traceability data of the monitoring vehicle may be used for clustering by the information processing device 10 (user classification setting unit 20). 【0102】 (A combination of multiple variations) The various modifications described above may be combined as appropriate, within the bounds of consistency. 【0103】 With regard to the above-described embodiment, the following additional information is disclosed. 【0104】 (Note 1) The information processing device according to this disclosure is an information processing device (10) comprising: a pre-estimated data acquisition unit (26) that acquires pre-estimated data (PE) of a deterioration index, which is an index related to the deterioration of a vehicle; an observation data acquisition unit (24) that acquires observation data (RD) of the deterioration index; and a data assimilation processing unit (28) that derives post-estimated data (ES) of the deterioration index by performing data assimilation processing based on the pre-estimated data and the observation data. This makes it possible to achieve good estimation accuracy from the stage immediately after the vehicle is released. 【0105】 (Note 2) The information processing device described in Appendix 1 may be an information processing device in which the pre-estimation data acquisition unit acquires new pre-estimation data of the degradation index based on the post-estimation data derived by the data assimilation processing unit, and the data assimilation processing unit derives new post-estimation data by performing data assimilation processing based on the new pre-estimation data and the new observation data acquired by the observation data acquisition unit. As a result, the reliability of the post-estimation data can be improved as the data assimilation processing is repeated. 【0106】 (Note 3) In the information processing device described in Appendix 1 or 2, if the data assimilation processing unit has not derived the post-estimation data, the pre-estimation data acquisition unit may be an information processing device that acquires pre-prepared pre-estimation data candidates (PEC) as the pre-estimation data. This can reduce the processing load on the information processing device. 【0107】 (Note 4) In the information processing device described in Appendix 3, if the data assimilation processing unit has not derived the post-estimation data, the pre-estimation data acquisition unit may determine the pre-estimation data by selecting one of a plurality of pre-prepared pre-estimation data candidates based on the observation data. This can reduce the processing load on the information processing device. 【0108】 (Note 5) The information processing device described in Appendix 4 may be an information processing device in which the pre-estimated data candidates are prepared to correspond to a plurality of user classifications (USRs) whose vehicle usage patterns differ from each other. This allows the information processing device to select pre-estimated data according to the degree of vehicle deterioration and perform data assimilation processing. 【0109】 (Note 6) The information processing device described in Appendix 5 may further include a pre-estimated data candidate generation unit (22) that generates a plurality of pre-estimated data candidates, wherein the pre-estimated data candidate generation unit generates the pre-estimated data candidates corresponding to the usage patterns of each of the plurality of user classifications by estimation using an estimation model (MDL). This allows knowledge obtained during the vehicle design stage, etc., to be reflected in the plurality of pre-estimated data candidates. 【0110】 (Note 7) The information processing method relating to this disclosure is an information processing method comprising: a pre-estimated data acquisition step (S5) for acquiring pre-estimated data (PE) of a deterioration index, which is an index related to the deterioration of a vehicle; an observation data acquisition step (S4) for acquiring observation data (RD) of the deterioration index; and a data assimilation processing step (S6) for deriving post-estimated data (ES) of the deterioration index by performing data assimilation processing based on the pre-estimated data and the observation data. 【0111】 (Note 8) The program relating to this disclosure is a program that causes a computer to execute the information processing method described in Appendix 7. 【0112】 Furthermore, the present invention may take various configurations without departing from the gist of this disclosure, and is not limited to the disclosure described above. [Explanation of Symbols] 【0113】 10, 101… Information Processing Devices 22...Pre-estimated data candidate generation unit 24...Observation data acquisition unit 26... Pre-estimated data acquisition unit 28…Data Assimilation Processing Unit ES…Post-event estimated data MDL… Estimated Model PE…Pre-estimated data PEC… Candidate data for prior estimation RD…Observation data USR…User Classification

Claims

[Claim 1] A pre-estimate data acquisition unit acquires pre-estimate data for deterioration indicators, which are indicators related to vehicle deterioration, An observation data acquisition unit that acquires observation data of the aforementioned degradation index, A data assimilation processing unit that derives post-hoc estimated data of the degradation index by performing data assimilation processing based on the aforementioned pre-estimated data and the aforementioned observation data, An information processing device equipped with the following features. [Claim 2] An information processing apparatus according to claim 1, The pre-estimation data acquisition unit acquires new pre-estimation data for the degradation index based on the post-estimation data derived by the data assimilation processing unit. The data assimilation processing unit is an information processing device that derives new post-estimation data by performing data assimilation processing based on the new pre-estimation data and the new observation data acquired by the observation data acquisition unit. [Claim 3] An information processing apparatus according to claim 1 or 2, If the data assimilation processing unit has not derived the post-estimation data, the pre-estimation data acquisition unit acquires pre-prepared pre-estimation data candidates as the pre-estimation data, information processing device. [Claim 4] The information processing apparatus according to claim 3, If the data assimilation processing unit has not derived the post-estimation data, the pre-estimation data acquisition unit can determine the pre-estimation data by selecting one of a plurality of pre-prepared pre-estimation data candidates based on the observation data. [Claim 5] An information processing apparatus according to claim 4, The aforementioned pre-estimated data candidates are prepared to correspond to multiple user classifications whose vehicle usage patterns differ from one another, and are provided in an information processing device. [Claim 6] An information processing device according to claim 5, The system further comprises a pre-estimated data candidate generation unit that generates a plurality of the aforementioned pre-estimated data candidates, The aforementioned pre-estimated data candidate generation unit is an information processing device that generates pre-estimated data candidates corresponding to the usage patterns of each of the plurality of user classifications by estimation using an estimation model. [Claim 7] A step to acquire pre-estimated data for deterioration indicators, which are indicators related to vehicle deterioration, and An observation data acquisition step to acquire observation data of the aforementioned degradation index, A data assimilation processing step that derives post-estimation data of the degradation index by performing data assimilation processing based on the aforementioned pre-estimation data and the aforementioned observation data, An information processing method having the following characteristics. [Claim 8] A program for causing a computer to execute the information processing method described in claim 7.

Images