JP2023002180A - Drive preparation state estimation device and drive preparation state estimation program - Google Patents

Abstract

To properly determine whether or not there is the need to shift control of a vehicle to a driver.SOLUTION: A reaction time (estimation reaction time) acquired based on drive simulation of an estimation model assumes a specific scene. When the specific scene is applied to another scene (actual scene) as it is, it is predicted that the estimation reaction time is different from the actual reaction time (referred to as actual reaction time). The reaction time is corrected (corrected reaction time tr=tm×(ts/to)) according to the ratio (ts/to) of the deviation for an error due to the deviation between the surplus time to of the driver used in the drive simulation and the actual surplus time ts of the driver when receiving notification of driving replacement for the reaction time tm estimated in the scene of the drive simulation used in the estimation model. The possibility of the driving replacement is determined on the basis of the corrected reaction time tr.SELECTED DRAWING: Figure 2

Description

The present invention relates to a driving readiness estimation device and a driving readiness estimation program.

Conventionally, there has been proposed a driving intention estimation device provided with a driving attitude estimation unit that estimates the driving attitude related to the driving consciousness of the driver based on the reaction time, inattentiveness and wakefulness of the driver of the vehicle (Patent Reference 1). In the driving intention estimation device described in Patent Document 1, the influence of the reaction time on the driving posture estimated by the driving posture estimation unit is greater than the effect of at least one of the inattentiveness time and the degree of awakening on the driving posture. big.

Patent Document 1 does not describe how long the data is used to estimate the driving posture. If this time is long, it is not possible to obtain estimation results that appropriately reflect recent changes in the driver's condition. cause aggravation of

Further, in Patent Document 2, in order to improve the safety of the vehicle when switching the automatic driving mode to the manual driving mode, a storage unit, an external measurement unit, a driver information acquisition unit, a determination unit, and a grace time calculation unit An invention is described with

The storage unit stores the reaction time required for the driver of the vehicle under automatic operation to manually operate the vehicle, and multiple numerical values obtained by analyzing multiple types of information obtained by measuring the driver. Stores information for specifying the grace time calculation function that indicates the relationship between

The external measurement unit measures the external conditions of the vehicle during automatic driving.

The driver information acquisition unit acquires a plurality of types of information regarding the driver of the vehicle.

The determination unit determines whether or not it is necessary to shift control of the vehicle to the driver based on the situation measured by the external measurement unit.

The grace time calculation unit changes the vehicle from automatic driving to driver control based on the plurality of types of information and the grace time calculation function acquired by the driver information acquisition unit when the determination unit determines that control transition is necessary. Calculate the grace time before shifting to manual operation.

JP 2020-24551 A JP 2018-45451 A

However, the threshold value for comparing the estimated grace time (driver reaction time) is a fixed value obtained from a predetermined estimation model (4 seconds or less, 4 to 10 seconds, 10 seconds or more in Patent Document 2). Although compared, if the driving conditions set by the estimation model and the actual driving conditions diverge, it may not be possible to appropriately determine whether or not it is necessary to transfer control of the vehicle to the driver.

SUMMARY OF THE INVENTION An object of the present invention is to provide a driving readiness estimation device and a driving readiness estimation program capable of appropriately determining whether or not it is necessary to transfer control of a vehicle to a driver.

The driving preparation state estimation device according to the present invention provides an estimation that depends on the state of the driver in an estimation scene that requires driving change from automatic driving to manual driving (for example, a scene obtained by conducting an experiment using a drive simulator, etc.). a storage unit for storing an estimation model for obtaining a reaction time; and when an actual scene requiring a change of driving occurs, the state of the driver in the actual scene is calculated by arithmetic processing using the estimation model similar to the actual scene. An estimating unit that estimates the dependent estimated reaction time, and based on the result of comparison between the estimated surplus time to the drive change limit set under the estimated scene and the actual surplus time to the drive change limit in the actual scene, A correcting unit that derives an actual reaction time by correcting the estimated reaction time estimated by the estimating unit; and a determination unit.

According to the present invention, the storage unit executes a drive simulation in each of the estimated scenes requiring driving change from automatic driving to manual driving, and creates an estimated model for obtaining an estimated reaction time that depends on the state of the driver. Store.

When an actual scene requiring a driver change occurs, the estimating unit estimates an estimated reaction time dependent on the state of the driver in the actual scene by arithmetic processing using an estimation model similar to the actual scene.

Here, in the correction unit, the estimated response estimated by the estimation unit is based on the result of comparison between the estimated margin time to the driver change limit set under the estimated scene and the actual margin time to the driver change limit in the actual scene. Correct the time to derive the actual reaction time.

The judging section judges whether or not to change driving based on the actual reaction time of the actual scene corrected by the correcting section.

Accordingly, it is possible to appropriately determine whether or not it is necessary to transfer control of the vehicle to the driver.

In the present invention, the actual reaction time tr corrected by the correction unit is the estimated reaction time estimated by the estimation unit tm, the estimated margin time to, and the actual margin time ts. It is characterized in that it is obtained by calculation using the formula of ts/to.

In the present invention, the actual surplus time is the time during which the factor of driving change can be avoided under a predetermined acceleration/deceleration under automatic operation, and the factor of driving change can be physically avoided at a predetermined speed under automatic driving. It is characterized in that one is selected from the avoidable time and the arrival time to the driving avoidance factor occurrence point at a predetermined speed under automatic driving.

A driving readiness estimation program according to the present invention causes a computer to function as each part of the driving readiness estimation device.

As described above, according to the present invention, it is possible to appropriately determine whether or not it is necessary to transfer control of the vehicle to the driver.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a schematic diagram of an automatic driving system that mainly includes a change availability determination control device according to an embodiment of the present invention; FIG. 2 is a block diagram showing, by function, the alternation possibility determination control executed by the alternation possibility determination control device 10 according to the present embodiment. It is an example of a driving experiment of the driving simulator according to the present embodiment. It is an example of the margin time pattern according to the present embodiment, (A) is the margin time a that indicates the limit that can be avoided without feeling uncomfortable, and (B) is the limit that the own vehicle 5 (general passenger car) can change lanes. is the spare time b indicating 4 is a flow chart showing a main routine for a change availability determination control that is executed by the change availability determination control device according to the present embodiment. 6 is a flowchart showing details of an information acquisition processing control subroutine in step 102 of FIG. 5; FIG. 6 is a flow chart showing a subroutine for determining whether a change is possible or not in step 104 of FIG. 5; FIG. 4 is a characteristic diagram showing the relationship between estimated reaction time tm and actual reaction time tr. FIG.

FIG. 1 is a schematic diagram of an automatic driving system, mainly including a change availability determination control device 10 according to the present embodiment.

In the present embodiment, the automatic driving system is intended for a case where the automatic driving by the system needs to be switched to the driver as necessary. For example, it corresponds to level 2 to 3 of SAE (Society of Automotive Engineers). In this case, the automatic driving system is required to be able to switch the driving to the driver within a certain period of time when the automatic driving by the system becomes unstable or out of the scope of operation. Therefore, the shift availability determination control device 10 according to the present embodiment monitors whether or not the driver can return to driving within a certain period of time.

In order to maintain such a situation, it is necessary to estimate the driving readiness of the driver. The driving preparation state can be represented, for example, by the time (hereinafter referred to as "reaction time") from when the driver starts manual driving after a situation requiring driving change occurs. This embodiment relates to a technique for observing at least one of the state and behavior of the driver of a vehicle during automatic driving and estimating the driving preparation state from the observation result.

As shown in FIG. 1, the automatic driving system for estimating the driving preparation state according to the present embodiment includes a sensor group 12 mounted on the vehicle, an automatic driving control device 14, and a shift availability determination control device 10. .

The sensor group 12 includes multiple types of sensors for measuring at least one of the state and behavior of the driver of the vehicle during automatic operation.

For example, the sensor group 12 includes a camera for photographing a predetermined part such as the driver's face. This camera captures an image that allows recognition of the driver's face orientation, line of sight, posture such as crossed legs and arms (posture information). This image is an example of measured values.

The sensor group 12 also includes a pressure sensor, a torque sensor, or a motor angle sensor that measures whether or not the driver touches or operates a vehicle operating device such as a brake pedal or steering wheel (vehicle driving system information).

The sensor group 12 also includes a seat position sensor that measures the position of the seat on which the driver sits, the reclining angle, and the like. The sensor group 12 also includes biosensors that measure the driver's heartbeat, perspiration, and the like (biological information).

Furthermore, the sensor group 12 measures not only measured values that directly indicate at least one of the driver's state and behavior, but also measured values necessary for calculating state information, which will be described later. For example, the sensor group 12 includes a camera for measuring the behavior of other vehicles outside the own vehicle (environmental information).

Each sensor included in the sensor group 12 outputs measured values measured at predetermined time intervals.

The automatic driving control device 14 controls the driving operation to the destination based on information necessary for automatic driving (for example, detection information from the sensor group 12) from a vehicle control device (not shown) that controls the drive system of the vehicle. Confirm and instruct the vehicle control device.

The automatic driving control device 14 is capable of communicating with the switching availability determination control device 10 .

In the change propriety determination control device 10, state information such as travel histories due to automatic driving from the vehicle is aggregated.

The alternation permission/inhibition determination control device 10 includes a microcomputer 16 . The microcomputer 16 has a CPU 16A, a RAM 16B, a ROM 16C, an I/O 16D (input/output device), and a bus 16E such as a data bus and a control bus connecting these.

A communication I/F 18 is connected to the I/O 16D, and information is exchanged with the automatic operation control device 14 via the communication I/F 18.

Further, the I/O 16D is connected to the sensor group 12 via the I/F 20, and via this I/F 20, the measured values from each sensor belonging to the sensor group 12 (for example, the attitude information, the vehicle drive system information, biological information, environmental information, etc.).

Furthermore, the I/O 16D is connected to a large-scale storage device 22 (including all recording media such as hard disks, SSDs, IC memories, etc., with no structural limitations such as built-in, external, detachable, etc.).

The large-scale storage device 22 of this embodiment includes a state information database 24 and an estimation model database 26 (details will be described later).

FIG. 2 is a block diagram showing, by function, the alternation possibility determination control executed by the alternation possibility determination control device 10 according to the present embodiment. It should be noted that each block does not limit the hardware configuration of the switching availability determination control device, and some or all of the functions may be executed by the microcomputer 16 (CPU 16A) based on a software program.

As shown in FIG. 2, the measured value acquisition unit 28 is connected to the sensor group 12, and from the sensor group 12, one or more types of information ( attitude information, vehicle drive system information, biological information, environmental information, etc.).

The measured value acquisition unit 28 is connected to the state information calculation unit 30 and sends the acquired information to the state information calculation unit 30 .

The state information calculation unit 30 calculates state information indicating at least one of visual behavior, operation behavior, and reaction to an event as state information indicating behavior of the driver. Further, for example, the state information calculation unit 30 calculates state information indicating at least one of a driving posture, arousal level, and a physiological index as state information indicating the state of the driver. Note that the measured value obtained by the measured value obtaining unit 28 may be used as the state information as it is.

An example of the procedure for calculating (acquiring) each type of state information will be described below.

(Status information indicating visual behavior)

The state information calculation unit 30 recognizes the driver's face direction and line of sight from the images captured by the cameras included in the sensor group 12, and calculates the visual behavior represented by the distribution ratio of the driver's line of sight to the specific area. Calculate the state information shown. More specifically, the state information calculation unit 30 calculates the ratio of the time the driver's line of sight is directed toward the front of the vehicle or the mirror (hereinafter referred to as "time ratio") within a predetermined time period (for example, 1 minute), Calculate the time rate that is suitable for tasks other than driving, such as smartphones and in-vehicle displays. The higher the former rate, the shorter the reaction time, and the higher the latter rate, the longer the reaction time. State information indicating visual behavior is useful for estimating driving readiness.

(Status information indicating operation behavior)

The state information calculation unit 30 calculates whether the driver is driving based on the measurement values of a pressure sensor, a torque sensor, or a motor angle sensor that measures whether or not the driver is touching or operating a vehicle operation device such as a brake pedal or steering wheel. The state information indicating the operation behavior representing the intervention operation and the stance is calculated. When a driver feels uneasy about automatic driving by an automatic driving system, the driver may take actions such as gripping the steering wheel or putting his/her foot on the brake pedal. When such behavior appears, the reaction time tends to be shortened, so the state information indicating the operation behavior is useful for estimating the driving preparation state.

(Status information indicating driving posture)

The state information calculation unit 30 recognizes the posture of the driver from the image captured by the camera, and determines whether the driver is in the correct driving posture, or whether the driver is driving in a manner that deviates from the correct driving posture, such as crossing the legs or arms. State information indicating whether or not it is in the posture is calculated. Further, the state information calculation unit 30 determines whether the seat is lowered or reclined from the normal position based on the seat position and reclining angle indicated by the measured values of the seat position sensor. Information may be calculated. If the driver takes a posture that deviates from the correct driving posture, it takes time to return to the correct driving posture. Useful for state estimation.

(State information indicating arousal level)

The state information calculation unit 30 acquires the eyelid opening degree of the driver from the image captured by the camera, and calculates state information indicating the driver's arousal degree based on the eyelid opening degree. The degree of eyelid opening can be represented, for example, by the number of times the eye blinks within a predetermined period of time, the percentage of time the eyes are closed within the predetermined period of time, or the like. The state information calculation unit 30 may also calculate state information indicating the degree of wakefulness based on changes in the driver's heart rate and perspiration amount measured by a biosensor and comparison results with thresholds.

(Status information indicating physiological index)

The state information calculation unit 30 acquires the measured value measured by the biosensor as it is as state information indicating the physiological index.

(state information indicating reaction to event)

The state information calculation unit 30 detects changes in the behavior of other vehicles other than the own vehicle as the occurrence of an event from images captured by a camera that captures the surroundings of the own vehicle. Changes in the behavior of other vehicles include, for example, a preceding vehicle decelerating to reduce the inter-vehicle distance to the own vehicle, an overtaking vehicle changing lanes ahead of the own vehicle, and the like. Further, the state information calculation unit 30 may detect a change in the behavior of the host vehicle that accompanies a change in the behavior of the other vehicle as the occurrence of an event. Changes in the behavior of the own vehicle include acceleration and deceleration due to changes in the control of the automatic driving system in order to avoid approaching or overtaking other vehicles. In addition, the state information calculation unit 30 intentionally decelerates the event in order to obtain state information indicating the reaction to the event, regardless of whether there is a change in the behavior of the other vehicle. may be generated. Also, an intentional event may be generated by giving the driver a stimulus such as sound, vibration, smell, or the like. The state information calculation unit 30 calculates at least one of the driver's visual behavior and operation behavior when the event occurs as state information indicating the reaction to the event. If the reaction to the event is good, the reaction time tends to be shortened. Therefore, the state information indicating the reaction to the event is useful for estimating the driving preparation state.

The status information calculation unit 30 stores the acquired status information in the status information database 24 shown in FIG. 2, for example, for each type of status information and for each acquisition time.

The state information database 24 is connected to the reaction time estimation unit 32. When the reaction time estimation unit 32 is notified from the outside of an instruction to determine whether the driver can switch from automatic to manual, the reaction time estimation unit 32 accesses the state information database 24, reads necessary information (state information), and estimates the reaction time.

Here, as parameters that influence the driver's driving preparation state necessary for driving change from automatic driving to manual driving, (i) the degree of deviation from driving of the driver, and (ii) the degree of situational awareness at the time of driving change It is thought that there are two

(i) is an index indicating how far the driver is away from driving, and it is considered that the larger the divergence, the longer the time required to return to driving, that is, the longer the reaction time. This degree of divergence is estimated based on the divergence from the correct driving posture, lack of confirmation of surroundings, engagement in tasks other than driving, etc. (long-term trend).

In addition, (ii) is an index indicating how much the driver recognizes the surrounding situation when the driver is instructed to switch from automatic driving to manual driving. If the situation cannot be recognized, the time required for subsequent judgments and operations, that is, the reaction time, is considered to be longer. On the other hand, if the driver checks the front and surroundings immediately before the driver change, the degree of situational awareness increases, and the reaction time is considered to be shorter (driver behavior immediately before the driver change).

The reaction time estimator 32 uses the estimation model stored in the estimation model database 26 to estimate the reaction time of the driver.

The reaction time estimating unit 32 collects each type of state information by adding, multiplying, listing elements, or the like to form integrated state information, and inputs the state information to the estimation model.

The estimation model stored in the estimation model database 26 is a model that outputs an index indicating the driving preparation state of the driver in response to the input of the integrated state information. Here, a case will be described in which the reaction time is output as an index indicating the operational preparation state. In this case, the estimation model is a correspondence between the integrated state information calculated from the state information obtained by conducting driving tests using a driving simulator, etc., and the reaction time measured when the state information was obtained. is generated by machine learning.

Fig. 3 shows an example of a running test of the driving simulator.

As shown in FIG. 3, on a road 50 with three lanes in each direction, a subject vehicle 52 (indicated as "own vehicle" in FIG. 3) is in the center lane and a preceding vehicle 54 (indicated as "preceding vehicle" in FIG. 3). In contrast, the vehicle is automatically driving with a certain inter-vehicle distance.

Here, it is assumed that the preceding vehicle 54 suddenly stops, and the own vehicle 52 corresponding to this sudden stop is in a state that cannot be avoided only by braking. In other words, it is a scene in which the driver is forced to switch from automatic driving to manual driving.

FIG. 3 shows that when a driver change instruction is given, another vehicle 56 (denoted as “other LF” in FIG. A vehicle 58 (denoted as “other LR” in FIG. 3A) is present.

Therefore, in FIG. 3, the driver must make a decision (action) to move to the right lane (lane change). On the other hand, in the driving test, there are also scenes in which it is necessary to move to the left.

That is, in the scene of driving change, the changed driver does not simply (reflexively) operate the steering wheel. must move to a lane where no vehicle is present). In other words, the driver is required to grasp the situation, and the driver's driving readiness affects the reaction time.

Therefore, in such a scene, the state of the maintenance driver for whom the driving change notification is issued is measured, and the measured value is used as a function to output the reaction time using an estimation model.

In other words, in the estimation model, experiments are conducted using a driving simulator, the state information of the driver is measured, and regression analysis is performed using the measured values (numerical values) as explanatory variables and the reaction time as the objective variable. Find the regression equation.

(Reaction time correction)

By the way, the reaction time obtained based on the drive simulation of the estimation model (hereinafter referred to as the estimated reaction time) assumes a specific scene.

If this particular scene is applied to another scene (actual scene) as it is, it is expected that the estimated reaction time will differ from the actual reaction time (referred to as actual reaction time).

Therefore, in the present embodiment, it is found that the deviation between the estimated reaction time and the actual reaction time depends on the driver's spare time, and the estimated reaction time is corrected based on the driver's spare time. , it is determined whether or not the driver can be changed.

In a broad sense, the spare time is the time from the change instruction to the avoidance limit, and the following three patterns can be given as representative examples.

(Surplus time a) Limit at which avoidance is possible without feeling discomfort

As shown in FIG. 4(A), while the vehicle 52 is traveling on a two-lane road (main road 64), at a road branch point 66, a branch road 68 (for example, an interchange of an expressway or a grade crossing) When changing lanes to a side road, etc.), the limit (see solid line La in FIG. 4A) at which the driver and fellow passengers can change lanes without feeling uncomfortable is, for example, 3 m/sec ² (approximately 0.3 G). is. The margin time a can be calculated from the relationship between the distance from the current position of the vehicle 52 until the limit can be secured and the speed of the vehicle 52 .

(Margin time b) Limit of lane change by own vehicle 52 (general passenger car)

As shown in FIG. 4(B), while the vehicle 52 is traveling on a two-lane road (main road 64), at a road branch point 66, a branch road 68 When changing lanes to a side road or the like), the limit (see solid line Lb in FIG. 4B) for changing lanes is, for example, 6 m/sec ² (approximately 0.6 G). The margin time b can be calculated from the relationship between the distance from the current position of the vehicle 52 until the limit can be secured and the speed of the vehicle 52 .

(Margin time c) TTC (Time To Collision)

It is a value obtained by dividing the distance between the own vehicle and the obstacle at the time of driving change by the speed, and is a physical limit value.

As shown in FIG. 2, the reaction time estimator 32 is connected to the reaction time corrector 34 . The reaction time correction unit 34 receives the reaction time tm estimated by the reaction time estimation unit 32 using the estimation model.

At this time, the reaction time estimating unit 32 also inputs to the reaction time correcting unit 34 together with the margin time to when constructing the applied drive simulation.

On the other hand, the shift possibility determination control device 10 includes an actual surplus time calculation unit 36 . The actual surplus time calculation unit 36 is connected to a changeover required environment information acquisition unit 38 and a surplus time determination requirement selection unit 40 .

The environment information acquisition unit 38 when change is required receives the notification of the change instruction (at the time of change required) from the change notification reception unit 42, acquires the current environment information from the measurement value acquisition unit 28, and acquires the current environment information from the measurement value acquisition unit 28. The determination requirement (one of the time margins a to c) selected in 40 is read, and the actual time margin ts is calculated.

The actual marginal time calculation unit 36 is connected to the reaction time correction unit 34 , and the actual marginal time ts calculated by the actual marginal time calculation unit 36 is sent to the reaction time correction unit 34 .

In the reaction time correction unit 34, the reaction time tm in the drive simulation is corrected according to the ratio (ts/to) of the actual surplus time ts at the time of shift to the surplus time to in the drive simulation (reaction time tr=tm xts/to).

The reaction time correction unit 34 is connected to the replacement possibility determination unit 44 , and the corrected reaction time tr (actual reaction time tr) corrected by the reaction time correction unit 34 is sent to the replacement possibility determination unit 44 . .

The alternation propriety determination unit 44 determines the driving preparation state estimation result (alternation propriety) based on the reaction time tr, and outputs it to the automatic driving control device 14 via the determination result output unit 46 . In the automatic operation control device 14, when the determination result indicates that the replacement is possible, the automatic operation is terminated and the control for shifting to the manual operation is executed. Further, when the determination result indicates that the change is impossible, the vehicle is decelerated or stopped, and a warning or the like is issued to prompt the driver to prepare for driving.

The operation of this embodiment will be described below with reference to the flow charts of FIGS. 5 to 7. FIG.

FIG. 5 is a flow chart showing a main routine for the alternation possibility determination control executed by the alternation possibility determination control device 10 . This flowchart is repeatedly executed at fixed time intervals.

At step 100 in FIG. 5, it is determined whether or not there has been a notification of switching to manual operation from the automatic operation control device 14 .

If a negative determination is made in step 100, the routine proceeds to step 102, executes information acquisition processing, and ends this routine.

If the determination in step 100 is affirmative, the routine proceeds to step 104 to execute a change availability determination control process, and this routine ends.

FIG. 6 is a flow chart showing the details of the information acquisition process control subroutine in step 102 of FIG.

At step 110 , a measurement value is obtained from each sensor of the sensor group 12 and the process proceeds to step 112 .

At step 112 , state information is calculated based on the acquired measurement values, and the process proceeds to step 114 . Note that the state information may be the sensor measurement value itself.

At step 114, the calculated state information is stored in the state information database, and this routine ends.

FIG. 7 is a flow chart showing a subroutine of the change possibility determination process control subroutine in step 104 of FIG.

At step 120 , the state information is read from the state information database 24 , the process proceeds to step 122 to read the drive simulation from the estimated model database 26 , and the process proceeds to step 124 .

In step 124, an estimated reaction time tm is calculated based on the time to spare to using the drive simulation model.

In the next step 126, the environmental information at the time when the shift instruction is notified (that is, when the shift is necessary) is acquired, and then the process proceeds to step 128 to select the leeway time determination requirement. The leeway time determination requirement is the limit that can be avoided without feeling uncomfortable as the leeway time a. As the leeway b, the limit at which the own vehicle 52 (general passenger car) can change lanes, and as the leeway c, TTC (Time To Collision) can be considered. , may be set fixedly.

In the next step 130, the actual surplus time ts is calculated. In the next step 132, the reaction time tm in the drive simulation is corrected according to the ratio (ts/to) of the actual surplus time ts at the time of shift to the surplus time to in the drive simulation (corrected reaction time tr=tm× ts/to).

In the next step 134, it is determined whether or not the change is possible based on the corrected reaction time, then the process proceeds to step 136, the determination result is notified to the automatic operation control device 14, and this routine ends.

FIG. 8 is a characteristic diagram showing the relationship between the estimated reaction time tm and the actual reaction time tr. In the straight line indicated by the dotted line in FIG. 8 (direct proportional characteristic curve: tr=a·tm), the slope a is 1 when the error is 0 (tm=tr).

On the other hand, for example, when the actual margin time ts of the driving change scene actually encountered is longer than the margin time to when setting the drive simulation (belongs to the "set of scenes with margin" in FIG. 8), the equivalent driving Even in the ready state, the reaction time will be longer than in the simulation. In other words, an error occurs in the determination in a scene where there is a margin according to the difference (degree of divergence) between the margin time to and the margin time ts.

On the other hand, in the present embodiment, the reaction time tm estimated for the scene of the drive simulation used in the estimation model is the margin time to of the driver used in the drive simulation and The reaction time is corrected according to the deviation ratio (ts/to) (corresponding to the slope a) (corrected reaction time tr = tm xts/to). Since it is determined whether or not to change the driving based on this corrected reaction time tr, errors can be suppressed.

In addition, the margin time to in the driving simulation scene used in the estimation model is set to a strict value (for example, the shortest time in the settable range, the time shorter than the average value in the settable range, etc.). It is possible to increase the probability of unidirectional correction of correcting a scene with no scene to a scene with a margin, and increase the number of judgment results that allow safe replacement. It should be noted that this does not deny correcting a scene with a margin to a scene with no margin.

Note that in the present embodiment, a case has been described in which visual behavior, manipulation behavior, driving posture, arousal level, physiological index, and state information indicating a reaction at the time of event occurrence are all used as state information, but the present invention is limited to this. not. At least one of these state information may be used. Also, the types of state information are not limited to these examples, and may be any information indicating at least one of the driver's state and behavior during automatic driving.

Also, the program of the present invention may be stored in a storage medium and provided.

REFERENCE SIGNS LIST 10 replacement availability determination control device 12 sensor group 14 automatic operation control device 16 microcomputer 16A CPU
16B RAM
16C ROM
16D I/O
16E Bus 18 Communication I/F
20 interfaces
22 large-scale storage device 24 state information database 26 estimation model database (storage unit)
28 Measured value acquisition unit 30 State information calculation unit 32 Reaction time estimation unit (estimation unit)
34 Reaction time correction unit (correction unit)
36 Actual surplus time calculation unit 38 Replacement required environment information acquisition unit 40 Sufficient time determination requirement selection unit 42 Replacement notification reception unit 44 Replacement availability determination unit (determination unit)
46 determination result output unit 50 road 52 host vehicle 54 preceding vehicle 56 other vehicle 58 other vehicle 60 other vehicle 62 other vehicle 64 main road 66 branch point 68 branch road

Claims (4)

a storage unit that stores an estimation model for obtaining an estimated reaction time that depends on the state of the driver in an estimated scene that requires driving change from automatic driving to manual driving;
an estimation unit for estimating an estimated reaction time dependent on the state of the driver in the real scene, when the real scene requiring the driver change occurs, by arithmetic processing using an estimation model similar to the real scene;
Correction of the estimated reaction time estimated by the estimator based on the result of comparison between the estimated surplus time up to the driving change limit set under the estimated scene and the actual surplus time up to the driving change limit in the actual scene. a correction unit that derives the actual reaction time by
a determining unit that determines whether or not to allow driving change based on the actual reaction time of the actual scene corrected by the correcting unit;
A driving preparation state estimating device having a The actual reaction time tr corrected by the correction unit is tr=tm×ts/to, where tm is the estimated reaction time estimated by the estimation unit, to is the estimated margin time, and ts is the actual margin time. The driving preparation state estimation device according to claim 1, obtained by calculation using a formula. the actual spare time,
Time during which driving change factors can be avoided under a predetermined acceleration/deceleration under automated driving, time during which driving change factors can be physically avoided at a predetermined speed under automated driving, and under automated driving 3. The driving preparation state estimating device according to claim 1, wherein one is selected from the arrival time to the driving avoidance factor occurrence point at a predetermined speed. the computer,
A driving readiness estimation program that functions as each part of the driving readiness estimation device according to any one of claims 1 to 3.

Images