Driving quality estimation system, driving quality estimation server, in-vehicle device, and computer program

Abstract

A driving quality estimation system includes a storage section that acquires and stores a time series of sensor data output by a sensor related to biological information of an occupant of a vehicle, and a driving quality estimation section that estimates a quality of driving performed by the occupant based on the sensor data for a prescribed period stored in the storage section, the driving quality estimation section including a state number estimation section that estimates a state number of the occupant based on the sensor data for the entire prescribed period, and a determination section that determines the quality of driving performed by the occupant in accordance with the state number estimated by the state number estimation section.

Description

Technical Field

[0001] This disclosure relates to a driving quality estimation system, a driving quality estimation server, an in-vehicle device, and a computer program. This application claims priority based on Japanese Application No. 2024-062792, filed on April 9, 2024, and invokes all the contents described in that Japanese application. Background Technology

[0002] Distracted driving, caused by factors such as fatigue, is one of the main causes of traffic accidents. Traffic accidents caused by distracted driving due to fatigue account for approximately one-quarter of all accidents. To avoid such accidents, it is necessary to accurately monitor the driving quality of passengers (drivers). If changes in driving quality are observed, appropriate rest or leave should be arranged for the driver, or the driver's workload should be reduced as needed.

[0003] However, there are also situations where fatigue differs from one's own perception. That is, there are cases where, even if the driver doesn't feel fatigued, their body is actually fatigued, thus reducing the quality of their driving. Therefore, whether the quality of driving performed by the driver has decreased requires an objective external benchmark, rather than relying solely on the driver's own judgment.

[0004] Patent Document 1, described later, discloses a technique for classifying a driver into any of several groups based on biometric information detected from the driver in real time. Patent Document 1 describes a method for accurately estimating the driver's state by selecting the optimal feature quantity for each group.

[0005] Existing technical documents Patent documents Patent Document 1: Japanese Patent Application Publication No. 2013-164748 Summary of the Invention

[0006] A driving quality estimation system according to one aspect of this disclosure includes: a storage unit that acquires and stores time series of sensor data output by sensors related to the biometric information of a vehicle occupant; and a driving quality estimation unit that estimates the quality of driving performed by the occupant based on the sensor data stored in the storage unit for a predetermined period, the driving quality estimation unit including: a state number estimation unit that estimates the state number of the occupant based on the sensor data throughout the predetermined period; and a determination unit that determines the quality of driving performed by the occupant based on the state number estimated by the state number estimation unit.

[0007] This disclosure can be implemented not only as a driving quality estimation system with such a characteristic processing unit, but also as a driving quality estimation method that uses the characteristic processing as a step, or as a computer program for causing a computer to execute the step. Furthermore, it can be implemented as a semiconductor integrated circuit that constitutes part or all of the apparatus of the driving quality estimation system, or as a larger system including the driving quality estimation system. Attached Figure Description

[0008] Figure 1 This is a block diagram showing the general structure of a fatigue estimation system, which is an example of a driving quality estimation system according to the first embodiment of this disclosure.

[0009] Figure 2 This is a schematic diagram used to explain the principle of driving quality determination in the first embodiment of this disclosure.

[0010] Figure 3 This is a schematic diagram used to illustrate the principle of driving quality estimation in the second embodiment of this disclosure.

[0011] Figure 4 This is a block diagram showing the configuration of the vehicle-mounted device according to the first embodiment.

[0012] Figure 5 It means Figure 1 The diagram shown illustrates the structure of a fatigue assessment server.

[0013] Figure 6 It means Figure 1 The diagram shows the functional configuration of the vehicle system.

[0014] Figure 7 This is a schematic diagram illustrating an example of the placement of seating sensors in a vehicle.

[0015] Figure 8 It means Figure 1 The diagram shows the functional structure of the fatigue assessment server.

[0016] Figure 9 This is a schematic diagram illustrating an example of a sensor configuration for a fatigue estimation system.

[0017] Figure 10 This is a schematic diagram used to illustrate the method for estimating the number of states in the first embodiment of this disclosure.

[0018] Figure 11 This is a flowchart illustrating the control structure of a computer program (hereinafter referred to as the program) executed by a fatigue determination server, which is an example of a driving quality determination server according to the first embodiment of this disclosure.

[0019] Figure 12 This is a graph representing an example of the data analysis results immediately following the start of driving.

[0020] Figure 13 This is a graph representing an example of the data analysis results from the seating sensor after a specified time has elapsed since the start of driving.

[0021] Figure 14 This is a schematic diagram illustrating the method for estimating the effective number of states according to the second embodiment of this disclosure.

[0022] Figure 15 This is a block diagram illustrating the functional configuration of the vehicle system according to the third embodiment of this disclosure. Detailed Implementation

[0023] [The problem this disclosure aims to solve] In the technology disclosed in Patent Document 1, examples of biological information used include information related to eye movements, line of sight, eyelid opening, heart rate, and body movements. To obtain this information, it is necessary for the driver to wear sensors or for a camera to be installed in the vehicle. However, it is impractical for employers to have drivers wear sensors. Furthermore, according to the prior art, there is a problem that the processing of sensor data needs to be changed when the type of sensor changes. Ideally, the quality of driving performed by the driver should be estimated based on sensor data regardless of the type of sensor, but this requirement cannot be met in the technology of Patent Document 1.

[0024] Therefore, the purpose of this disclosure is to provide a driving quality estimation system, a driving quality estimation server, an in-vehicle device, and a computer program that are independent of the type of sensor.

[0025] [Effects of this disclosure] This configuration enables the provision of a driving quality estimation system, driving quality estimation server, in-vehicle device, and computer program that are independent of the type of sensor.

[0026] [Description of embodiments of this disclosure] In the following description and accompanying drawings, the same reference numerals are used to refer to the same parts. Therefore, detailed descriptions related to them will not be repeated. It should be noted that at least some of the embodiments described below can be combined arbitrarily.

[0027] (1) The driving quality estimation system of the first aspect of this disclosure includes: a storage unit that acquires and stores a time series of sensor data output by sensors related to the biological information of a vehicle occupant; and a driving quality estimation unit that estimates the quality of driving performed by the occupant based on the sensor data stored in the storage unit for a predetermined period, the driving quality estimation unit including: a state number estimation unit that estimates the state number of the occupant based on the sensor data throughout the predetermined period; and a determination unit that determines the quality of driving performed by the occupant based on the state number estimated by the state number estimation unit. With this configuration, the quality of driving performed by the occupant can be determined based on the state number estimated from the occupant's biological information, regardless of the type of sensor. As a result, a driving quality estimation system that is independent of the type of sensor can be provided.

[0028] (2) In (1) above, the sensor may also include a seating sensor disposed in the seat of the vehicle. With this configuration, the quality of driving performed by the passenger can be determined even without the passenger wearing a special sensor.

[0029] (3) In (2) above, the seating sensor may also include multiple pressure sensors, and the time series of the sensor data may be represented by multiple vectors with the outputs of the multiple pressure sensors at each measurement time point as elements. With this configuration, even without the passenger wearing special sensors, the quality of driving performed by the passenger can be determined by mathematical processing of the sensor data.

[0030] (4) In (3) above, the state number estimation unit may also include: multiple clustering units that cluster the multiple vectors throughout the specified period into multiple cluster groups consisting of clusters with different numbers of clusters; and an estimation unit that evaluates the multiple cluster groups obtained by the multiple clustering units, thereby estimating the state number. With this configuration, the state number of the passenger can be estimated based on sensor data using existing clustering methods.

[0031] (5) In (4) above, the number of the plurality of clusters can also be an integer greater than or equal to 2 and less than or equal to a specified upper limit. With this configuration, the number of passenger states can be estimated within a certain range.

[0032] (6) In (5) above, the upper limit may also be greater than or equal to 3 and less than or equal to 16. With this configuration, the number of passenger states can be estimated within a certain limited range.

[0033] (7) In (4) above, the plurality of clustering units may also include a plurality of K-means clustering units, wherein the plurality of K-means clustering units cluster the sensor data with different numbers of clusters using the K-means method. With this configuration, the existing K-means clustering method can be used to efficiently estimate the number of passenger states.

[0034] (8) In (4) above, the estimation unit may also include: an evaluation value calculation unit that calculates a predetermined evaluation value for each of the plurality of cluster groups obtained by the plurality of clustering units; a qualified cluster group selection unit that selects qualified cluster groups from the plurality of cluster groups whose evaluation values ​​calculated by the evaluation value calculation unit meet the predetermined conditions; and a cluster number confirmation unit that confirms the number of clusters included in each of the qualified cluster groups selected by the qualified cluster group selection unit, thereby estimating the number of states. With this configuration, highly reliable clustering results can be used to determine the quality of driving performed by the passenger with high reliability.

[0035] (9) In (4) above, the estimation unit may also include: an evaluation value calculation unit that calculates a predetermined evaluation value for each of the plurality of cluster groups obtained by the plurality of clustering units; and a qualified cluster group selection unit that selects qualified cluster groups from the plurality of cluster groups whose evaluation values ​​calculated by the evaluation value calculation unit meet predetermined conditions. Alternatively, the determination unit may determine whether the passenger is fatigued based on whether the number of qualified cluster groups selected by the cluster group selection unit exceeds the number of qualified cluster groups previously selected. With this configuration, it is possible to determine, based on the clustering results, that the passenger's state has changed and the driving quality has become lower.

[0036] (10) In (8) above, the determination unit may also include: a first determination unit that determines that the passenger is fatigued in response to all cluster numbers confirmed by the cluster number confirmation unit being greater than a threshold; and a second determination unit that determines whether the passenger is fatigued based on whether any one of the cluster numbers confirmed by the cluster number confirmation unit is less than or equal to the threshold and whether the number of qualified cluster groups exceeds the number of qualified cluster groups prior to this. With this configuration, the passenger's state can be determined with high accuracy. As a result, the quality of driving performed by the passenger can be estimated with high reliability.

[0037] (11) In (4) above, the state number estimation unit may also include a reduction unit, wherein the reduction unit reduces the data of the data block consisting of the multiple vectors throughout the specified period before the clustering performed by the multiple clustering units. With this configuration, the amount of computation required to estimate driving quality is reduced, and processing can be performed quickly.

[0038] (12) The vehicle-mounted device of the second aspect of this disclosure is a vehicle-mounted device equipped with any of the aforementioned driving quality estimation systems. As a result, with this vehicle-mounted device, the quality of driving performed by the passenger can be determined based on the state number estimated from the passenger's biological information, regardless of the type of sensor. As a result, a vehicle-mounted device that can estimate driving quality regardless of the type of sensor can be provided.

[0039] (13) The driving quality estimation server of the third aspect of this disclosure includes: a storage unit that acquires and stores time series of sensor data output by sensors related to the biological information of the occupants of the vehicle via communication from an on-board device mounted in the vehicle; a driving quality estimation unit that estimates the quality of driving performed by the occupants based on the sensor data stored in the storage unit for a predetermined period; and a processing unit that performs predetermined processing to address the fatigue of the occupants in response to the driving quality estimated by the driving quality estimation unit meeting predetermined conditions. The driving quality estimation unit includes: a state number estimation unit that estimates the state number of the sensor data throughout the predetermined period; and a determination unit that determines the quality of driving performed by the occupants based on the state number estimated by the state number estimation unit. With this configuration, the quality of driving performed by each occupant can be determined based on the state number estimated from the biological information of the occupants of each vehicle, regardless of the type of sensor. The result is a driving quality estimation server that can provide an estimate of the quality of driving performed by each passenger, independent of the type of sensor.

[0040] (14) The computer program of the fourth aspect of this disclosure enables a computer to function as a storage unit and a driving quality estimation unit, wherein the storage unit acquires and stores a time series of sensor data output by sensors related to the biometric information of the occupants of a vehicle, and the driving quality estimation unit estimates the quality of driving performed by the occupants based on the sensor data stored in the storage unit for a predetermined period. The driving quality estimation unit includes: a state number estimation unit that estimates the state number of the sensor data throughout the predetermined period; and a determination unit that determines the quality of driving performed by the occupants based on the state number estimated by the state number estimation unit. With this configuration, the quality of driving performed by each occupant can be determined based on the state number estimated from the biometric information of the occupants of each vehicle, regardless of the type of sensor. As a result, a computer program can be provided that estimates the quality of driving performed by each occupant regardless of the type of sensor.

[0041] (15) The computer program of the fifth aspect of this disclosure enables a computer to function as a storage unit, a driving quality estimation unit, and a processing unit, wherein the storage unit acquires and stores a time series of sensor data output by sensors related to the biological information of the occupants of the vehicle from an on-board unit mounted on the vehicle via communication; the driving quality estimation unit estimates the quality of driving performed by the occupants based on the sensor data stored in the storage unit for a predetermined period; and the processing unit performs predetermined processing to suppress a decrease in the quality of driving performed by the occupants in response to the driving quality estimated by the driving quality estimation unit meeting a predetermined condition; the driving quality estimation unit includes: a state number estimation unit that estimates the state number of the sensor data throughout the predetermined period; and a determination unit that determines the quality of driving performed by the occupants based on the state number estimated by the state number estimation unit. This configuration allows for the estimation of each passenger's state number based on biometric information received from external vehicles, independent of the type of sensor used, and the determination of the quality of driving performed by each passenger based on that state number. As a result, a computer program can be provided that can estimate the quality of driving performed by each passenger regardless of the type of sensor used.

[0042] [Details of the embodiments disclosed herein] First Embodiment 1.1 Composition Figure 1The diagram schematically illustrates the configuration of a fatigue estimation system 50 according to a first embodiment of this disclosure, wherein the fatigue estimation system 50 estimates the driver's fatigue level as an indicator of the quality of driving performed by the vehicle's occupant (driver). (See also...) Figure 1 The fatigue estimation system 50 includes a vehicle 60, a fatigue assessment server 62, and a driver management system 64. The vehicle 60, the fatigue assessment server 62, and the driver management system 64 can communicate with each other via a network 66.

[0043] The vehicle 60 is equipped with an onboard system 68 for controlling various systems within the vehicle 60 or for communicating with the driver management system 64. The onboard system 68 includes an onboard unit 80, which has the following functions: collecting data from seating sensors located within the vehicle 60 (in... Figure 1 (Not shown in the figure.) The output is accumulated and periodically sent to the fatigue assessment server 62.

[0044] The fatigue assessment server 62 receives and stores the time series output from the seat sensor of the vehicle-mounted device 80, and estimates the driver's fatigue level at regular intervals. The fatigue assessment server 62 has the following function: when it determines that the driver is fatigued, it notifies the driver management system 64 of this information. That is, the fatigue assessment server 62 determines that the driving quality has decreased due to driver fatigue and notifies the driver management system 64 of this.

[0045] The driver management system 64, in response to a notification from the fatigue assessment server 62, notifies the manager that a driver is fatigued. To restore the driving quality of the problematic driver, the manager urges the driver to rest or arranges for driver rotation. Alternatively, a system can be built that performs such processing automatically without the manager's intervention.

[0046] The method for estimating driver fatigue in this first embodiment will now be described. In this disclosure, the law of entropy increase is used to estimate driver fatigue. According to the law of entropy increase, ordered systems have low entropy, while disordered systems have high entropy. The entropy of an isolated system only changes in the direction of increase. When the driver is considered as an isolated system, the driver's entropy increases as the driver becomes fatigued. Therefore, it is believed that by knowing the change in the driver's entropy, the driver's fatigue level can be determined.

[0047] In this disclosure, it is assumed that the entropy of the driver system increases as the driver drives continuously. That is, it is assumed that the order of the driver system gradually disappears. It is believed that when the order of the driver system disappears, the quality of the driver's driving also decreases. Specifically, driver fatigue accumulates, or the driver's attention becomes scattered, increasing the likelihood of distracted driving. Based on this assumption, the entropy of the driver system is estimated in this disclosure, and when this entropy is considered to increase beyond a certain level, the driver is determined to be in a state of fatigue.

[0048] Here, the method for estimating the entropy of the driver system becomes a problem. In this disclosure, the entropy of the driver is estimated as follows. Generally, it is believed that in a low-entropy state, the number of microstates exhibited by the driver system is small, and as entropy increases, the number of microstates exhibited by the driver system increases. Therefore, in this disclosure, some methods are used to estimate the number of microstates exhibited by the driver system, and based on the results, the magnitude of the driver's entropy is estimated.

[0049] That is, refer to Figure 2 On the left, when the number of micro-states exhibited by driver 100 is constant and small, as indicated by state group 102, driver 100 is judged to be not fatigued, as indicated by icon 104. Conversely, referring to... Figure 2 On the right side, when the number of micro-states presented by the driver 100 is as many as state group 112, it is determined as driver 100 fatigue, i.e., the quality of driving performed by the driver 100 is reduced, as indicated by icon 114.

[0050] Or, such as Figure 3 As shown, driver fatigue can also be determined when the number of micro-states exhibited by driver 100 increases from only state group 102, such as by state group 120 and state group 122, to two equal "number of micro-states exhibited by driver 100". The following first embodiment uses Figure 2 As shown in the diagram.

[0051] That is, in this disclosure, the macro state of driver 100, whether it is fatigued or not, is determined based on the number of micro states of driver 100.

[0052] The problem arising from this approach is the method for estimating the number of micro-states exhibited by the driver. In this disclosure, the number of such states is estimated by acquiring some of the driver's movements using sensors. Various sensors are considered. Various biosensors used in the prior art can be used, such as sensors that acquire information related to eye movements, gaze, eyelid opening, heart rate, and body movements. In this embodiment, considering the driver's workload, a seating sensor, as a type of biosensor, is used. The seating sensor functions to simultaneously measure the distribution of pressure exerted on various parts of the seat by the driver's weight at multiple locations at various measurement time points, and outputs the measured values ​​as a vector. Therefore, a time series of measurement vectors, consisting of measurement vectors from each measurement time point, is obtained from the seating sensor. This time series of measurement vectors is considered to represent a characteristic of the driver's body movements.

[0053] - Hardware configuration of vehicle-mounted device 80 Figure 4 The hardware configuration of the vehicle-mounted device 80 is shown in the figure. (Refer to...) Figure 4 The vehicle-mounted device 80 is essentially a computer, including: a control unit 150, which controls the entire vehicle-mounted device 80; an auxiliary storage device 158, which stores various data; an in-vehicle communication unit 154, which communicates with other devices in the vehicle via an in-vehicle network (not shown); and a communication unit 156, which communicates with external wireless devices (not shown) in the vehicle. The control unit 150, auxiliary storage device 158, in-vehicle communication unit 154, and communication unit 156 are all connected to a communication bus 152, and data exchange between them is carried out via the communication bus 152.

[0054] The control unit 150 includes: a CPU (Central Processing Unit) 180; a ROM (Read-Only Memory) 182 storing the startup program of the vehicle-mounted device 80, etc.; and RAM (Random Access Memory) 184, which can be written to and read from at any time. The CPU 180 includes, for example, a CPU or an MPU (Microprocessor) as the arithmetic element (processor). The auxiliary storage device 158 includes, for example, non-volatile memory such as flash memory. The ROM 182 or the auxiliary storage device 158 stores the software (program) executed by the CPU 180 and various information (data).

[0055] The in-vehicle communication unit 154 communicates with other devices in the vehicle via the in-vehicle network according to a prescribed communication protocol. The in-vehicle network can be any network such as CAN (Controller Area Network), LIN (Local Interconnect Network), MOST (Media Oriented Systems Transport), FlexRay, 10BASE-T1S, 100BASE-T1, 1000BASE-T1, CXPI (Clock Extension Peripheral Interface), ASRB (Automotive Safety Restraints Bus), or CAN XL (Controller Area Network eXtendedLength). In this embodiment, there is no particular limitation to using CAN as the in-vehicle network.

[0056] The vehicle-mounted device 80, under the control implemented by the control unit 150, receives sensor data output from various sensors on the vehicle via the vehicle network and accumulates it in RAM 184. The vehicle-mounted device 80 periodically sends the sensor data accumulated in RAM 184 to... Figure 1 The fatigue assessment server 62 is shown. The communication unit 156 is connected via... Figure 4 An external wireless device (not shown) provides communication with the fatigue assessment server 62.

[0057] - Hardware configuration of fatigue assessment server 62 Reference Figure 5 The fatigue assessment server 62 includes a control unit 200, an auxiliary storage device 204, and a communication unit 206. The control unit 200 includes a CPU 230, a GPU (Graphics Processing Unit) 232, a ROM 234, and a RAM 236. The control unit 200, the auxiliary storage device 204, and the communication unit 206 are all connected to a communication bus 202, and data exchange between them is performed via the communication bus 202.

[0058] Auxiliary storage device 204 includes, for example, non-volatile storage devices such as flash memory or hard disk drives. Auxiliary storage device 204 stores programs and various information executed by CPU 230. Communication unit 206 provides communication with other devices, including vehicle-mounted device 80, via network 66.

[0059] Fatigue assessment server 62 connects to vehicle-mounted device 80 via network 66 and communication unit 206 (see reference). Figure 1 The fatigue assessment server 62 communicates with the vehicle-mounted device 80. It obtains data from the vehicle-mounted device 80 to estimate the fatigue level of the driver of the vehicle 60 equipped with the vehicle-mounted device 80. Based on the obtained data, the fatigue assessment server 62 estimates the fatigue level of the driver of the vehicle 60. When the estimated fatigue level exceeds a predetermined threshold, it notifies the manager of the driver's fatigue level. It should be noted that the fatigue assessment server 62 can also notify the vehicle-mounted device 80 of the vehicle 60 driven by the driver instead of the driver management system 64, or it can notify both the driver management system 64 and the vehicle-mounted device 80 of the vehicle 60 driven by the driver.

[0060] The program used to enable the fatigue assessment server 62 to function as the various functional units in this embodiment is stored and circulated on a specified storage medium such as a DVD (Digital Versatile Disc) or USB (Universal Serial Bus) memory, and further transmitted from there to the auxiliary storage device 204. Alternatively, the program can also be sent from an external device to the fatigue assessment server 62 via network 66 and stored in the auxiliary storage device 204.

[0061] - Functional configuration of vehicle system 68 and vehicle device 80 Figure 6 The diagram illustrates the functional configuration of the vehicle system 68, which is equipped with vehicle-mounted device 80. (See reference...) Figure 6 In addition to the vehicle-mounted device 80, the vehicle system 68 also includes a seating sensor 252 (acting as a biosensor) and a vehicle navigation system 254. The seating sensor 252, the vehicle-mounted device 80, and the vehicle navigation system 254 are connected to each other via a vehicle network 250. The vehicle-mounted device 80 can also communicate with various ECUs (Electronic Control Units) via the vehicle network 250.

[0062] like Figure 7 As shown, the seating sensor 252 is a body pressure sensor installed inside the driver's seat 350 in the vehicle 60. The seating sensor 252 measures the seat surface pressure at predetermined intervals, representing the distribution of body pressure of the driver seated on the seat 350, and outputs this as sensor data. In this embodiment, the sensor data output by the seating sensor 252 at a given measurement is in vector form, with seat surface pressure data as the element.

[0063] Refer again Figure 6The in-vehicle navigation system 254 includes a route guidance unit 302 that performs route guidance processing, a display device 304, a voice output unit 306, and a control unit 300 for controlling them to assist the driver. The control unit 300 is connected to the in-vehicle network 250.

[0064] The main functional components of the vehicle-mounted device 80 are described below. Specifically, the functions implemented by the vehicle-mounted device 80 include: a communication function implemented by the communication unit 156; a sensor data receiving unit 280 that receives sensor data from vehicle-mounted sensors such as the seating sensor 252 via the vehicle network 250; and a sensor data accumulation unit 282 that accumulates the sensor data received by the sensor data receiving unit 280. The functions implemented by the vehicle-mounted device 80 also include: a sensor data transmitting unit 284 that performs the process of uniformly transmitting the sensor data accumulated in the sensor data accumulation unit 282 to the fatigue assessment server 62 via the communication unit 156 at predetermined intervals; and a warning notification unit 286 that uses the vehicle navigation system 254 to notify the driver of fatigue warnings received from the outside by the communication unit 156.

[0065] -Functional components of the fatigue assessment server 62 Figure 8 The functional configuration of the fatigue assessment server 62 is shown in the diagram. (Refer to...) Figure 8 In addition to the communication function implemented by the communication unit 206, the fatigue determination server 62 also includes a timer 380, a fatigue processing unit 382, ​​and a warning notification unit 384.

[0066] The fatigue processing unit 382 includes a sensor data acquisition unit 400, a sensor data accumulation unit 402, a state number estimation unit 404, a state number increase determination unit 406, a state number storage unit 410, and a fatigue determination unit 408.

[0067] The sensor data acquisition unit 400 acquires sensor data from the on-board unit 80 of each vehicle 60. The sensor data accumulation unit 402 accumulates the sensor data acquired by the sensor data acquisition unit 400 for each vehicle (driver). The state number estimation unit 404, according to the timing implemented by the timer 380, reads the sensor data of the preceding predetermined period from the sensor data accumulation unit 402 for each vehicle (driver) at predetermined intervals, and estimates the micro-state number of the driver during that period. The state number storage unit 410 stores the micro-state number estimated by the state number estimation unit 404. The state number increase determination unit 406 determines whether the micro-state number estimated by the state number estimation unit 404 has increased compared to the micro-state number estimated by the previous processing stored in the state number storage unit 410. The fatigue determination unit 408 determines whether the predetermined condition that the fatigue level of the target driver exceeds a threshold is met based on the micro-state number estimated by the state number estimation unit 404 and the determination result obtained by the state number increase determination unit 406. The warning notification unit 384 performs prescribed processing for drivers deemed to meet the condition, such as notifying the driver management system 64 of a warning. The warning notification unit 384 may also directly notify the on-board unit 80 of the warning. In this case, consider... Figure 6 The control unit 300 shown uses the display device 304 and the voice output unit 306 to warn the driver.

[0068] - Seating sensor 252 Figure 9 An example of the configuration of the pressure sensor in the seating sensor 252 is shown. Figure 9 This is just one example; for instance, the seating sensor 252 could also be set as a rectangle, and the pressure sensors could be configured in a matrix. Alternatively, the number of pressure sensors could be reduced or increased.

[0069] like Figure 9 As shown, each pressure sensor of the seating sensor 252 is assigned a number to distinguish each pressure sensor. The vector output by the seating sensor 252 includes elements that arrange the outputs of each pressure sensor in order according to this number. In the following description, the number of pressure sensors is set to m. Multiple vectors measured over a certain period of time (e.g., 30 minutes) form a data block. It should be noted that the order of the assigned numbers is not limited. Figure 9 The order shown.

[0070] - Estimation of the number of microstates Figure 10 The diagram schematically illustrates the cluster analysis process 440 for estimating the effective number of states in this embodiment. This process... Figure 8The state number estimation section 404 shown is executed.

[0071] Reference Figure 10 Let's assume that during the specified period, the seating sensor 252 outputs sensor data at n measurement times T1, T2, ..., Tn. As described above, a measurement vector is obtained at each measurement time, with m sensor outputs as elements. The elements of this measurement vector are defined as P1, P2, ..., Tm. That is, every specified period, a data block 450 consisting of n×m data points becomes a state number estimation unit 404 ( Figure 8 The object of the processing implemented.

[0072] In this embodiment, the cluster analysis process 440 includes a K-means clustering unit 452 for elements belonging to data block 450. In this embodiment, K-means clustering is used as the clustering method. This clustering results in L-1 possible cluster numbers, ranging from 2 to L, performed using K-means clustering. L is an integer greater than 2, for example, considered to be an integer greater than 6 and less than a predetermined upper limit (e.g., 16). Alternatively, the number of vehicles processed by the fatigue assessment server 62 can be determined based on its computing power, the number of vehicles processed by the fatigue assessment server 62, and the seating sensor 252 (…). Figure 6 The number of pressure sensors a device possesses, etc., causes the upper limit value to change.

[0073] Clustering performed by the K-means clustering unit 452 yields L-1 clustering results with the number of clusters ranging from 2 to L. In this specification, the set of multiple clusters obtained through each clustering is referred to as a cluster group obtained through that clustering. That is, L-1 cluster groups are obtained through the K-means clustering unit 452. It should be noted that clustering processing performed by the K-means clustering unit 452 in this manner can be executed in parallel. Furthermore, this is purely mathematical processing. Therefore, using GPU232 (… Figure 5 These processes are performed, thereby enabling efficient processing.

[0074] The clustering analysis process 440 also includes a score calculation unit 454, which calculates a score as a predetermined evaluation value for each of the L-1 cluster groups obtained by the K-means clustering unit 452. This score can be any score used to evaluate the clustering results; any score can be used. For example, scores obtained using contour analysis, elbow method, gap statistics, Davies-Bouldin method, Calinski-Harabasz method, etc., can be used. In this embodiment, the contour score obtained using contour analysis is used as the score. The method for calculating the contour score is well known, and its value is between -1 and 1. The closer the contour score is to 1, the higher the evaluation of the clustering. The formula for calculating the contour score in this example sets the contour score to SCORE, as described below.

[0075] SCORE=(b-a) / max(a,b) In this example, 'b' is the distance between the data belonging to cluster i and the center of cluster a, and 'a' is the distance between this data and the closest cluster to cluster i. Various definitions can be used to define distance. For example, the distance between data and cluster centers can be used. Alternatively, the distance between data x and the closest data y belonging to the target cluster can be used, or conversely, the distance between data x and the furthest data z belonging to the target cluster can be used. Furthermore, the average distance between data x and all data belonging to the other cluster can also be used. It should be noted that... Figure 10 The scores shown are for reference only.

[0076] In this embodiment, the score calculation unit 454 calculates the contour score of the data belonging to each cluster within a cluster group obtained through a certain clustering. The score calculation unit 454 also calculates the score of the cluster group by averaging the contour scores of all cluster groups obtained as a result of a certain clustering. For example, in... Figure 10 In the results, the cluster with 2 clusters has a score of SCORE1 = 0.705. The cluster with 3 clusters has a score of SCORE2 = 0.598. The same applies to the following clusters.

[0077] The cluster analysis process 440 further includes: an average score calculation unit 456, which calculates the average score (mean score) of the scores obtained by the score calculation unit 454 for all cluster groups; a comparison process 462, which compares the scores of each cluster group with the average score; and a qualified cluster group selection unit 458, which selects cluster groups that are determined by the comparison process 462 to have a profile score higher than the average score. Figure 10 In the example shown, the qualified cluster selection unit 458 selects two cluster groups: cluster number 2 and cluster number 4. In this disclosure, the cluster groups selected by the qualified cluster selection unit 458 are referred to as qualified cluster groups, meaning cluster groups that meet the prescribed criteria. It should be noted that in the calculation of the average score in the average score calculation unit 456, the average score can be calculated either by weighting the number of data belonging to each cluster, or simply by dividing the sum of the scores of each cluster by the number of clusters.

[0078] The cluster analysis process 440 also includes a cluster number confirmation unit 460, which confirms the number of qualified cluster groups selected by the qualified cluster group selection unit 458 and the number of clusters in each cluster group.

[0079] Figure 11 The flowchart illustrates the process of achieving this through collaboration with computer hardware. Figure 8 The control structure of the fatigue treatment unit 382's program. This program is controlled by... Figure 8 The timer 380 is activated periodically (e.g., every 30 minutes) for each onboard device. (See reference...) Figure 11 The procedure includes: step 480, from the sensor data accumulation unit 402 ( Figure 8 The sensor data acquired and analyzed during the period of sensor data acquisition is used to form a data block consisting of sensor data from the vehicle-mounted device of the target within the preceding 30 minutes. The procedure also includes step 482, performing... Figure 10 The diagram shows a rough cluster analysis.

[0080] The program also includes: step 484, which functions as a first determination unit, wherein the first determination unit controls the process branch based on whether the number of clusters included in each of the qualified cluster groups obtained as the result of the cluster analysis processing 440 in step 482 exceeds a threshold.

[0081] In this disclosure, each cluster group includes data that is different from but similar to the data in other clusters. This data represents the output of the seating sensor. Therefore, the data belonging to each cluster is considered to represent that the driver's state is approximately the same at the time when the data is acquired. That is, the number of clusters is considered to represent the order of the driver's state. If the number of clusters is small, there is order and low entropy; if the number of clusters is large, there is no order and high entropy. When the entropy is low, the driver is not fatigued; if the entropy increases, it can be determined that the driver is fatigued. The "threshold" described in step 484 is the number of states used as a benchmark for determining whether the driver is fatigued. Ideally, this threshold uses different values ​​depending on the type of work the driver is performing, the type of vehicle being driven, the type of equipment in the vehicle, etc.

[0082] Based on the above explanation, if the determination in step 484 is affirmative, then it meets the requirements. Figure 2 The state shown on the right indicates driver fatigue. If the determination in step 484 is negative, then at least the driver may not be fatigued.

[0083] Figure 11The procedure shown also includes step 486, where, if the determination in step 484 is negative, it functions as a second determination unit, wherein this second determination unit branches the controlled flow based on whether the number of qualified cluster groups exceeds the number of qualified cluster groups obtained through previous processing. The condition that the number of qualified cluster groups has increased compared to the previous time refers to... Figure 3 The situation shown on the left changes to the situation shown on the right. That is, it is unclear whether the driver's state is maintaining order as indicated by state group 120, or gradually losing order as indicated by state group 122. As a result, if the determination in step 486 is affirmative, it is considered that the driver is fatigued for safety reasons; on the other hand, if the determination in step 486 is negative, it can be determined that the driver is not yet fatigued.

[0084] The procedure also includes: step 488, executed when the determination in step 484 is affirmative and when the determination in step 484 is negative and the determination in step 486 is affirmative, determining that the driver has accumulated fatigue and the quality of driving performed by the driver has decreased, and issuing a warning notification related to the driver to the driver management system 64 (see reference). Figure 1 () and step 490, which is executed after step 488 and when the determination in step 486 is negative, will store the number of qualified cluster groups calculated in this process as the state number of driver 352. Figure 8 The execution of the program ends when the state number storage unit 410 is reached. The value stored in the state number storage unit 410 is referenced in step 486 when the program is executed next time.

[0085] 1.2 Actions The fatigue estimation system 50 of the structure has been described above and operates as follows. It should be noted that the on-board device 80 and the fatigue determination server 62 are pre-programmed to perform predetermined actions.

[0086] Reference Figure 7 When the driver 352 sits in seat 350, the seating sensor 252 begins outputting sensor data. For example... Figure 9 As shown, this output represents the pressure distribution generated by the driver 352, and is output in vector form at predetermined intervals. By setting it to vector form, various mathematical processing methods can be applied to the sensor data.

[0087] Reference Figure 6Sensor data output by the seating sensor 252 is received by the sensor data receiving unit 280 via the vehicle network 250. The sensor data receiving unit 280 assigns a reception time to the received sensor data and accumulates it in the sensor data accumulation unit 282. The sensor data transmitting unit 284 reads the sensor data from the sensor data accumulation unit 282 at regular intervals, assigns identification information to the vehicle device 80, and transmits it to the fatigue assessment server 62 via the communication unit 156 and the external wireless device 82.

[0088] Reference Figure 8 The sensor data acquisition unit 400 of the fatigue assessment server 62 receives the sensor data via the communication unit 206 and communicates with the vehicle-mounted wireless device 82 that sent the sensor data. Figure 6 The address of the vehicle or the identification information of the on-board device 80 are accumulated together in the sensor data accumulation unit 402. Figure 8 ).

[0089] Seating sensor 252 ( Figure 6 ), vehicle-mounted device 80 ( Figure 4 , Figure 6 ) and fatigue determination server 62 ( Figure 8 The fatigue treatment unit 382 repeatedly performs this treatment.

[0090] On the other hand, fatigue assessment server 62 starts at regular intervals (e.g., every 30 minutes). Figure 11 The program showing the control structure is illustrated. For example, the state number estimation unit 404 reads from the sensor data accumulation unit 402 and acquires a data block consisting of sensor data from a specific vehicle in the 30 minutes immediately preceding the program's initiation. Figure 11 Step 480). The state number estimation unit 404 performs the following on the data block. Figure 10 The cluster analysis process shown is 440. The result is the selection of one or more eligible cluster groups.

[0091] If the number of states included in each of the selected qualified cluster groups exceeds the threshold (determined as positive in step 484), then driver fatigue is considered to be accumulating, and driving quality is reduced, resulting in a warning (step 488). Even if the determination in step 484 is negative, if the number of qualified cluster groups this time is greater than the number of qualified cluster groups selected in the previous execution (determined as positive in step 486), then the driver may be fatigued, and a warning is also issued. If the number of selected qualified cluster groups does not increase, no warning is issued, and the process ends.

[0092] The fatigue assessment server 62 performs the above-mentioned processing repeatedly at specified intervals for each vehicle that uses the fatigue assessment server 62's service.

[0093] Figure 12 and Figure 13 An example is shown in the diagram of the relationship between the number of clusters and the score during clustering in this first embodiment. Figure 12 This example shows the clustering results for sensor data from when the driver started driving until 30 minutes later. Figure 13 This example shows the clustering results for sensor data from 1 hour after the driver started driving until 1 hour later (2 hours after the start of driving).

[0094] exist Figure 12 and Figure 13 In the calculation, the number of clusters from 2 to 15 is used as the number of clusters for clustering. The number of clusters is set as the horizontal axis, and the score of the cluster groups when clustering with that number of clusters is set as the vertical axis. The results are plotted in graph 520. Figure 12 ) and graph 530 ( Figure 13 The relationship between the number of clusters and the score is shown in graph 522. Figure 12 ) and curve 532 ( Figure 13 The average score is shown below.

[0095] Reference Figure 12 In this example, the score is highest when the number of clusters is 2, and then decreases sequentially when the number of clusters is 4 and 3. Only these three cases, with scores above the average, are selected as qualified cluster groups, resulting in low entropy. If the threshold is set to, for example, "4", then... Figure 11 The decision in step 484 is negative. Since a comparison with the immediate preceding result is not possible, the decision in step 486 is also negative, indicating that the driver is not fatigued.

[0096] exist Figure 13 In the example shown, the overall score decreased, and the clusters with 2, 3, 4, 5, 11, 13, 14, and 15 clusters were selected as qualified cluster groups. If compared with... Figure 12 Similarly, if the threshold is set to "4", then the decision in step 484 will be negative. However, in Figure 13 In this case, the probability of the determination in step 484 becoming positive is high, therefore the possibility of determining that the driver is fatigued is high.

[0097] As described above, the fatigue estimation system 50 of this embodiment uses the output of a sensor, such as a seat sensor, which obtains biological information from the driver, to estimate the number of the driver's microstates. Based on the estimated number of states, the driver's fatigue state is determined. As a result, by using appropriate sensors, fatigue states that are not visible to the naked eye and are not felt by the driver themselves can be determined. Ideally, the sensor should be a sensor like a seat sensor that can collect the required information without being worn directly by the individual. However, the fatigue estimation system 50 of this embodiment is not limited to using a seat sensor and can also use other sensors. This is because the sensor data is not considered as the value of the sensor data itself, but rather as the abstract content of the driver's state as represented by the sensor data, and a clustering approach for this purpose is adopted. Therefore, regardless of the type of sensor, driver fatigue can be flexibly determined based on sensor data.

[0098] Second Embodiment In the first embodiment described above, Figure 10 In the cluster analysis process 440, the score of each cluster group is compared with the overall score, and the cluster group with a higher score than the overall score is selected as a qualified cluster group. However, this disclosure is not limited to this implementation. This second implementation selects qualified cluster groups by considering all the scores of each cluster group within each cluster group.

[0099] Reference Figure 14 The clustering analysis process 550 in this second embodiment includes a K-means clustering unit 452, which performs AND operations on the data block 450. Figure 10 The cluster analysis process 440 is the same as the cluster analysis process 440. The difference between cluster analysis process 550 and cluster analysis process 440 is that cluster analysis process 550 replaces the entire score calculation unit 560, which includes calculating the score of each cluster in each cluster group. Figure 10 The fraction calculation section 454.

[0100] In the score calculation unit 454, only one score is calculated for each cluster group. However, in the score calculation unit 560, a score is calculated for each cluster of each cluster group. That is, two scores are calculated when there are two clusters, three scores are calculated when there are three clusters, and so on.

[0101] Cluster analysis processing 550 also includes an average score calculation unit 564 that calculates the average score of the scores calculated for each cluster group instead of... Figure 10The cluster analysis processing 440 shown has an average score calculation unit 456. In this average score calculation unit 564, either a weighted average of the scores can be calculated based on the number of data belonging to each cluster, or a number obtained by simply summing the scores of each cluster and dividing by the total number of clusters can be used.

[0102] The cluster analysis process 550 also includes a comparison process 562 that compares the scores of each cluster with the average score calculated by the average score calculation unit 564, instead of... Figure 10 The cluster analysis process 440 shown includes a comparison process 462. The cluster analysis process 550 further includes: a qualified cluster group selection unit 566, which selects cluster groups whose clusters have a score higher than the average score as a result of the comparison performed by the comparison process 562 as qualified cluster groups; and a cluster number confirmation unit 568, which confirms the number of qualified cluster groups selected by the qualified cluster group selection unit 566 and the number of clusters belonging to the qualified cluster groups.

[0103] exist Figure 14 In the example shown, the scores of all clusters belonging to the cluster group with a cluster size of 2 exceed the average score. On the other hand, two of the clusters belonging to the cluster group with a cluster size of 3 exceed the average score, but one does not reach the average score. Furthermore, for the cluster group with a cluster size of 4, three exceed the average score, but one does not reach the average score. If the same applies to the case where the number of clusters is greater than or equal to 4, then only the cluster group with a cluster size of 2 is selected as a qualified cluster group by the qualified cluster group selection unit 566. In this case, the number of qualified cluster groups is 1, and the number of clusters is 2.

[0104] The second embodiment is achieved by replacing the cluster analysis process 440 of the first embodiment with this cluster analysis process 550.

[0105] The operation of the second embodiment is the same as that of the first embodiment, except for the cluster analysis processing 550. Furthermore, the processing of the cluster analysis processing 550 is clearly understood from the above description, and therefore will not be repeated here.

[0106] The same effect as the first embodiment can be achieved through this second embodiment.

[0107] Third Implementation Method In either of the first and second embodiments described above, the output of the seating sensor is accumulated as a vector, and the multiple vectors obtained within a specified time are processed as a single data block. However, this disclosure is not limited to such embodiments. In this third embodiment, preprocessing of the sensor data from the seating sensor is performed to reduce the data for clustering purposes.

[0108] Figure 15 The diagram shows a functional block diagram of the vehicle system 600 according to the third embodiment. The vehicle system 600 and... Figure 6 The difference in the vehicle system 68 shown is that it includes a vehicle device 610 instead of... Figure 6 The vehicle-mounted device 80 has the following functions: performing a prescribed reduction process on the sensor data received by the sensor data receiving unit 280 from the seat sensor 252 and accumulating it, and periodically sending the reduced sensor data to the fatigue assessment server.

[0109] Vehicle-mounted device 610 relative to Figure 6 The vehicle-mounted device 80 shown also includes a sensor data reduction unit 620, which is located at the output of the sensor data receiving unit 280 and the input of the sensor data accumulation unit 282. The sensor data reduction unit 620 has the function of performing a prescribed reduction process on the sensor data received by the sensor data receiving unit 280 and accumulating it in the sensor data accumulation unit 282.

[0110] As for the data reduction performed by the sensor data reduction unit 620, consider the following content.

[0111] (1) The sensor surface of the seating sensor is divided into several parts, and the average pressure of each part is used as the sensor data. Based on this configuration, the number of constituent elements of the sensor data, i.e. the vector, at each measurement time point is reduced.

[0112] (2) Quantize the output of each pressure sensor of the seating sensor. For example, assume that the output of each pressure sensor is a real value, and quantize the value to any level such as 16 levels, 128 levels, 256 levels, etc., and use the quantized value as the vector element of the sensor data. It can also be combined with (1).

[0113] By reducing sensor data in this way, the following effects are achieved. First, the amount of data sent from the vehicle system 68 to the fatigue assessment server 62 can be reduced. As a result, the time required to send data from the vehicle system 68 to the fatigue assessment server 62 is shortened. Second, when clustering is performed after data reduction, the computational load for clustering can be reduced. As a result, the computational load on the fatigue assessment server 62 can be reduced. It should be noted that such data reduction processing and interpretation of clustered data content are similar to various known data processing methods, such as image processing methods, and the techniques used in such techniques can be applied to the techniques disclosed herein. Furthermore, in the above embodiment, the driver's fatigue level is estimated based on the output of the seating sensor as an indicator of the quality of driving performed by the driver. By utilizing the seating sensor, it is possible to obtain information on the estimation of driving quality without requiring the driver to wear a dedicated sensor. However, this disclosure is not limited to this embodiment. As an indicator of driving quality performed by the driver, for example, drivers can wear their own dedicated sensors, or cameras can be installed in the vehicle to obtain information related to eye movements, line of sight, eyelid opening, heart rate, and body movements.

[0114] The various processes (functions) in the above-described embodiments are implemented by a processing circuit including one or more processors. This processing circuit may also be composed of an integrated circuit consisting of one or more processors combined with one or more memories, various analog circuits, and various digital circuits. The one or more memories store programs (instructions) that cause the one or more processors to execute the aforementioned processes. The one or more processors may execute the aforementioned processes according to the programs read from the one or more memories, or they may execute the aforementioned processes according to logic circuits pre-designed to execute the aforementioned processes. The processors may be various processors suitable for computer control, such as CPUs, GPUs, DSPs (Digital Signal Processors), FPGAs (Field-Programmable Gate Arrays), ASICs (Application Specific Integrated Circuits), and DPUs (Data Processing Units). It should be noted that multiple physically separate processors may also cooperate to execute the aforementioned processes. For example, the processors of each of multiple physically separate computers can collaborate via networks such as LAN (Local Area Network), WAN (Wide Area Network), and the Internet to execute the aforementioned processes. The programs can be installed in the memory from external server devices via the aforementioned network, or they can be stored on recording media such as CD-ROM (Compact Disc Read-Only Memory), DVD-ROM (Digital Versatile Disc Read-Only Memory), or semiconductor memory, and installed from the aforementioned recording media into the memory.

[0115] The embodiments disclosed herein are merely examples, and this disclosure is not limited to the embodiments described above. The scope of this disclosure is based on the detailed description of the invention and is shown by the claims, including all modifications within the meaning and scope equivalent to the words and phrases therein.

[0116] Explanation of reference numerals in the attached figures 50: Fatigue estimation system; 60: Vehicle; 62: Fatigue assessment server; 64: Driver management system; 66: Network; 68, 600: In-vehicle system; 80, 610: In-vehicle device; 82: External wireless device; 100, 352: Driver; 102, 112, 120: Status group; 104, 114: Icon; 150, 200, 300: Control unit; 152, 202: Communication bus; 154: In-vehicle communication unit; 156, 206: Communication unit; 158, 204: Auxiliary storage device; 180, 230: CPU; 182, 234: ROM; 184, 236: RAM; 232: GPU; 250: In-vehicle network; 252: Seating sensor; 254: In-vehicle navigation; 280: Sensor data receiving unit; 282, 402: Sensor data accumulation unit 284: Sensor data transmission unit; 286, 384: Warning notification unit; 302: Route guidance unit; 304: Display device; 306: Voice output unit; 350: Seat; 380: Timer; 382: Fatigue processing unit; 400: Sensor data acquisition unit; 404: State number estimation unit; 406: State number increase determination unit; 408: Fatigue determination unit; 410: State number storage unit; 440, 550: Cluster analysis processing; 450: Data block; 452: K-means clustering unit; 454, 560: Score calculation unit; 456, 564: Average score calculation unit; 458, 566: Qualified cluster group selection unit; 460, 568: Cluster number confirmation unit; 462, 562: Comparison processing; 520, 522, 530, 532: Curve graph; 620: Sensor data reduction unit.

Claims

1. A driving quality estimation system, comprising: The storage unit acquires and stores time-series sensor data output by sensors related to the biometric information of the vehicle's occupants. as well as The driving quality estimation unit estimates the quality of driving performed by the passenger based on the sensor data stored in the storage unit for a specified period. The driving quality estimation unit includes: The state number estimation unit estimates the state number of the passenger based on the sensor data throughout the entire specified period; as well as The determination unit determines the quality of driving performed by the passenger based on the state number estimated by the state number estimation unit.

2. The driving quality estimation system according to claim 1, wherein, The sensors include seating sensors configured in the seats of the vehicle.

3. The driving quality estimation system according to claim 2, wherein, The seating sensor includes multiple pressure sensors. The time series of the sensor data is represented by multiple vectors, with the outputs of the multiple pressure sensors at each measurement time point as elements.

4. The driving quality estimation system according to claim 3, wherein, The state number estimation unit includes: Multiple clustering units cluster the multiple vectors throughout the specified period into multiple cluster groups consisting of clusters with distinct numbers of clusters; and The estimation unit evaluates the multiple cluster groups obtained by the multiple clustering units, thereby estimating the number of states.

5. The driving quality estimation system according to claim 4, wherein, The number of each cluster is an integer greater than or equal to 2 and less than or equal to a specified upper limit.

6. The driving quality estimation system according to claim 5, wherein, The upper limit is greater than or equal to 3 and less than or equal to 16.

7. The driving quality estimation system according to claim 4, wherein, The plurality of clustering units include a plurality of K-means clustering units, wherein the plurality of K-means clustering units cluster the sensor data with different numbers of clusters using the K-means method.

8. The driving quality estimation system according to claim 4, wherein, The estimation unit includes: The evaluation value calculation unit calculates a predetermined evaluation value for each of the plurality of cluster groups obtained by the plurality of clustering units respectively; The qualified cluster group selection unit selects qualified cluster groups from the plurality of cluster groups whose evaluation values ​​calculated by the evaluation value calculation unit meet the prescribed conditions; and The cluster number confirmation unit confirms the number of clusters included in each of the qualified cluster groups selected by the qualified cluster group selection unit, thereby estimating the state number.

9. The driving quality estimation system according to claim 4, wherein, The estimation unit includes: The evaluation value calculation unit calculates a predetermined evaluation value for each of the plurality of cluster groups obtained through the plurality of clustering parts; and The qualified cluster group selection unit selects qualified cluster groups from the plurality of cluster groups whose evaluation values ​​calculated by the evaluation value calculation unit meet the specified conditions. The determination unit determines whether the passenger is fatigued based on whether the number of qualified clusters selected by the cluster selection unit exceeds the number of qualified clusters previously selected.

10. The driving quality estimation system according to claim 8, wherein, The determination unit includes: The first determination unit, in response to the cluster count confirmation unit confirming that the number of clusters is greater than a threshold, determines that the passenger is fatigued; and The second determination unit determines whether the passenger is fatigued based on whether any of the cluster numbers confirmed by the cluster number confirmation unit is less than or equal to the threshold and whether the number of qualified cluster groups exceeds the number of qualified cluster groups previously identified.

11. The driving quality estimation system according to claim 4, wherein, The state number estimation unit further includes a reduction unit, wherein the reduction unit reduces the data of the data block consisting of the plurality of vectors throughout the entire specified period before the clustering is performed by the plurality of clustering units.

12. An in-vehicle device equipped with a driving quality estimation system as described in any one of claims 1 to 11.

13. A driving quality estimation server, comprising: The storage unit acquires and stores time-series sensor data output by sensors related to the biological information of the occupants of the vehicle from an on-board device mounted on the vehicle via communication. The driving quality estimation unit estimates the quality of driving performed by the passenger based on the sensor data stored in the storage unit for a specified period. as well as The processing unit, in response to the driving quality estimated by the driving quality estimation unit as meeting a predetermined condition, performs a predetermined process to address changes in the driving quality performed by the passenger. The driving quality estimation unit includes: The state number estimation unit estimates the state number of the sensor data throughout the entire specified period; as well as The determination unit determines the quality of driving performed by the passenger based on the state number estimated by the state number estimation unit.

14. A computer program that enables a computer to function as a storage unit and a driving quality estimation unit, wherein, The storage unit acquires and stores time-series sensor data output by sensors related to the biometric information of the vehicle's occupants. The driving quality estimation unit estimates the quality of driving performed by the passenger based on the sensor data stored in the storage unit for a specified period. The driving quality estimation unit includes: The state number estimation unit estimates the state number of the sensor data throughout the entire specified period; and The determination unit determines the quality of driving performed by the passenger based on the state number estimated by the state number estimation unit.

15. A computer program that enables a computer to function as a storage unit, a driving quality estimation unit, and a processing unit, wherein, The storage unit acquires and stores time-series sensor data output by sensors related to the biometric information of the vehicle's occupants from an onboard device mounted in the vehicle via communication. The driving quality estimation unit estimates the quality of driving performed by the passenger based on the sensor data stored in the storage unit for a specified period. In response to the driving quality of the passenger, as estimated by the driving quality estimation unit, meeting a predetermined condition, the processing unit performs a predetermined process to address changes in the driving quality performed by the passenger. The driving quality estimation unit includes: The state number estimation unit estimates the state number of the sensor data throughout the entire specified period; and The determination unit determines the driving quality of the passenger based on the state number estimated by the state number estimation unit.

Images