Navigation system, navigation method, and program

Abstract

To prevent a driver from being confused when navigation is performed.SOLUTION: The navigation system SY generates navigation information based on a current position of the vehicle, map information, and destination information indicating a destination of the vehicle, and generates transmission information, which is information to be transmitted to a driver via a user interface, based on the navigation information. When it is determined that there is confusion in any of recognition, determination, and operation of the driver with respect to the piece of information to be transmitted, the piece of information to be transmitted is generated or navigation information is generated based on the piece of information to be transmitted to the driver via the user interface and the reaction of the driver with respect to the piece of information to be transmitted so that the confusion in any of recognition, determination, and operation of the driver is suppressed.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to a navigation system, a navigation method, and a program. [Background technology]

[0002] Patent document 1 describes that parameters of navigation instructions related to maneuvers are adjusted taking into account the difficulty of the maneuvers, that a machine learning model generates a difficulty metric for a set of maneuvers, and that query data is applied to the machine learning model to generate a difficulty metric for the maneuvers. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2024-20616 Summary of the Invention [Problem to be solved by the invention]

[0004] Incidentally, although Patent Document 1 describes the causes of confusion, it does not describe any technology for suppressing driver confusion. Therefore, the technology described in Patent Document 1 (technology for generating a difficulty metric) cannot suppress the risk of driver confusion when navigation is provided to the vehicle driver.

[0005] In view of the above, an object of the present disclosure is to provide a navigation system, a navigation method, and a program that can reduce the risk of the driver becoming confused when navigation is provided to the driver of a vehicle. [Means for solving the problem]

[0006] (1) One aspect of the present disclosure is a navigation system including: a navigation information generation unit that generates navigation information used to provide navigation to a driver of a vehicle based on the current position of the vehicle, map information, and destination information indicating the destination of the vehicle; a machine learning model unit that generates transmission information, which is information to be transmitted to the driver via a user interface, based on the navigation information generated by the navigation information generation unit; and a determination unit that determines whether or not there is confusion in any of the driver's cognition, judgment, and operation regarding the transmission information based on the driver's reaction to the transmission information transmitted to the driver from the user interface. When the determination unit determines that there is confusion in any of the driver's cognition, judgment, and operation regarding the transmission information, the machine learning model unit generates the transmission information or the machine learning model unit causes the navigation information generation unit to generate the navigation information so as to suppress confusion in any of the driver's cognition, judgment, and operation based on the transmission information transmitted to the driver via the user interface and the driver's reaction to the transmission information.

[0007] (2) In the navigation system of (1), the machine learning model unit may include a first machine learning model stored in a storage device of the vehicle and a second machine learning model stored in a storage device of a server, and the second machine learning model may generate advance instructions for the first machine learning model based on the destination information, and the first machine learning model may generate the transmission information in real time based on the advance instructions generated by the second machine learning model and the navigation information.

[0008] (3) In the navigation system of (1) or (2), the machine learning model unit may generate the transmission information or cause the navigation information generation unit to generate the navigation information based on personal information related to any of the driver's cognition, judgment, and operation, the transmission information, and the driver's reaction to the transmission information, so as to suppress confusion in any of the driver's cognition, judgment, and operation.

[0009] (4) Any of the navigation systems (1) to (3) may include a simulator located on the server, the simulator simulating the behavior of the navigation information generation unit, and the second machine learning model may work in conjunction with the simulator to iteratively predict the driver's behavior and optimize the transmission information.

[0010] (5) In any of the navigation systems (1) to (4), the transmission information may be transmitted to the driver from the user interface as audio guidance output from a speaker or image guidance displayed on a display.

[0011] (6) In any of the navigation systems (1) to (5), the determination unit may include a recognizer that detects information about lane markings contained in an image captured by a forward-facing camera mounted on the vehicle.

[0012] (7) In the navigation system of any one of (1) to (6), the determining unit may include an actuator control unit that controls a braking actuator or a steering actuator based on an operation by the driver.

[0013] (8) In any of the navigation systems (1) to (7), the determination unit may include a driver monitoring system that identifies the driver's reaction to the communication information based on an image of the driver captured by a driver monitor camera or the driver's voice collected by a microphone.

[0014] (9) In any of the navigation systems (1) to (8), the transmission information generated by the machine learning model unit may include any of the following information: timing, order, granularity, and specificity of guidance; additional information to be added during guidance; type of voice, tone, volume, and speaking style; display content of image guidance; display position; display size; color; and font.

[0015] (10) In any of the navigation systems (1) to (9), the machine learning model unit may add or update the communication information based on the driver's reaction to the communication information transmitted to the driver from the user interface.

[0016] (11) One aspect of the present disclosure is a navigation method comprising: a first step in which a navigation system generates navigation information used to provide navigation to a driver of a vehicle based on the current position of the vehicle, map information, and destination information indicating the destination of the vehicle; a second step in which the navigation system generates transmission information, which is information to be transmitted to the driver via a user interface, based on the navigation information generated in the first step; and a third step in which the navigation system determines whether there is confusion in any of the driver's cognition, judgment, and operation regarding the transmission information based on the driver's reaction to the transmission information transmitted to the driver from the user interface, wherein when it is determined in the third step that there is confusion in any of the driver's cognition, judgment, and operation regarding the transmission information, the transmission information is generated in the second step or the navigation information is generated in the first step based on the transmission information transmitted to the driver via the user interface and the driver's reaction to the transmission information so as to suppress confusion in any of the driver's cognition, judgment, and operation.

[0017] (12) One aspect of the present disclosure is a program for causing a processor of a computer mounted on a vehicle to execute the following steps: a first step of generating navigation information used to provide navigation to a driver of the vehicle based on the current position of the vehicle, map information, and destination information indicating the destination of the vehicle; a second step of generating transmission information, which is information to be transmitted to the driver via a user interface, based on the navigation information generated in the first step; and a third step of determining whether or not there has been confusion in any of the driver's cognition, judgment, and operation regarding the transmission information based on the driver's reaction to the transmission information transmitted to the driver from the user interface; wherein, when it is determined in the third step that there has been confusion in any of the driver's cognition, judgment, and operation regarding the transmission information, the transmission information is generated in the second step, or the navigation information is generated in the first step, based on the transmission information transmitted to the driver via the user interface and the driver's reaction to the transmission information, so as to suppress confusion in any of the driver's cognition, judgment, and operation. [Effects of the Invention]

[0018] According to the present disclosure, it is possible to reduce the risk of the driver becoming confused when navigation is provided to the driver of a vehicle. [Brief explanation of the drawings]

[0019] [Figure 1] 1 is a diagram illustrating an example of a navigation system SY according to a first embodiment. [Figure 2] 2 is a diagram showing an example of a vehicle VH, a server SV, etc. to which the navigation system SY shown in FIG. 1 is applied. [Figure 3] 3 is a diagram showing an example of the configuration of a vehicle VH shown in FIG. 2. FIG. [Figure 4] 3 is a diagram illustrating an example of the configuration of a server SV illustrated in FIG. 2. FIG. [Figure 5]FIG. 2 is a diagram for explaining a first example of navigation information (travel route plan for a vehicle VH) generated by a navigation information generating unit SY1. [Figure 6] FIG. 10 is a diagram for explaining a second example of navigation information (travel route plan for a vehicle VH) generated by the navigation information generating unit SY1. [Figure 7] FIG. 10 is a diagram for explaining a third example of navigation information (travel route plan for a vehicle VH) generated by the navigation information generating unit SY1. [Figure 8] FIG. 10 is a diagram for explaining a fourth example of navigation information (travel route plan for a vehicle VH) generated by the navigation information generating unit SY1. [Figure 9] FIG. 10 is a diagram for explaining a fifth example of navigation information (travel route plan for a vehicle VH) generated by the navigation information generating unit SY1. [Figure 10] 10 is a flowchart illustrating an example of processing executed by a processor VH73 of a computer VH7 mounted on a vehicle VH to which a navigation system SY according to a second embodiment is applied. DETAILED DESCRIPTION OF THE INVENTION

[0020] Hereinafter, embodiments of a navigation system, a navigation method, and a program according to the present disclosure will be described with reference to the drawings.

[0021] First Embodiment Fig. 1 is a diagram showing an example of a navigation system SY according to a first embodiment. Fig. 2 is a diagram showing an example of a vehicle VH, a server SV, etc. to which the navigation system SY shown in Fig. 1 is applied. Fig. 3 is a diagram showing an example of the configuration of the vehicle VH shown in Fig. 2. Fig. 4 is a diagram showing an example of the configuration of the server SV shown in Fig. 2.

[0022] 1 to 4, the navigation system SY is configured by a computer VH7 equipped in a vehicle VH, the computer VH7 having a communication interface VH71, a storage device VH72, and a processor VH73, and a server SV having a communication interface SV1, a storage device SV2, and a processor SV3. The navigation system SY includes a navigation information generation unit SY1, a machine learning model unit SY2, a determination unit SY3, and a simulator SY4. The processor VH73 of the computer VH7 of the vehicle VH functions as a navigation information generator SY1. The navigation information generator SY1 generates navigation information used to provide navigation to the driver of the vehicle VH based on the current position of the vehicle VH, map information, and destination information indicating the destination of the vehicle VH. The current position of the vehicle VH is calculated by the navigation information generator SY1, for example, based on a GPS (Global Positioning System) signal received by a GPS receiver functioning as the sensor VH6. The map information is stored, for example, in a storage device VH72 of the computer VH7. In another example, map information may be obtained from outside the vehicle VH by a communication interface VH71 of the computer VH7.

[0023] In the example shown in FIGS. 1 to 4, destination information indicating the destination of the vehicle VH is input by the driver of the vehicle VH via, for example, a user interface (UI) VH1 of the vehicle VH.

[0024] FIG. 5 is a diagram illustrating a first example of navigation information (travel route plan for vehicle VH) generated by navigation information generation unit SY1. FIG. 6 is a diagram illustrating a second example of navigation information (travel route plan for vehicle VH) generated by navigation information generation unit SY1. FIG. 7 is a diagram illustrating a third example of navigation information (travel route plan for vehicle VH) generated by navigation information generation unit SY1. FIG. 8 is a diagram illustrating a fourth example of navigation information (travel route plan for vehicle VH) generated by navigation information generation unit SY1. FIG. 9 is a diagram illustrating a fifth example of navigation information (travel route plan for vehicle VH) generated by navigation information generation unit SY1.

[0025] In the example shown in Fig. 5, the navigation information generation unit SY1 generates a driving route plan for the vehicle VH indicated by the arrow in Fig. 5 as a driving route plan (navigation information) for the vehicle VH from the current position of the vehicle VH to the destination of the vehicle VH. In order to travel from the current position of the vehicle VH to the destination of the vehicle VH, the vehicle VH needs to go straight on the road RD1 at the junction with the road RD2 without turning left onto the road RD1 at the junction with the road RD2, and then turn left onto the road RD1 at the junction with the road RD3 immediately after going straight on the road RD1 at the junction with the road RD2. Based on the driving route plan (navigation information) of vehicle VH shown by the arrow in Figure 5, if the communication information to be transmitted to the driver of vehicle VH, such as "Turn left 400 meters ahead," is generated, depending on the shapes of roads RD1, RD2, and RD3, the cognitive ability of the driver of vehicle VH, the nervousness of the driver of vehicle VH, etc., the driver of vehicle VH may not be able to accurately grasp "400 meters ahead," and vehicle VH may not be able to properly turn left onto road RD1 at the junction with road RD3 (for example, vehicle VH may turn left onto road RD1 at the junction with road RD2, or vehicle VH may go straight on road RD1 at the junction with road RD3). Furthermore, if the transmission information to be transmitted to the driver of vehicle VH, such as "Turn left at the second fork," is generated based on the driving route plan (navigation information) of vehicle VH shown by the arrow in Figure 5, depending on the shapes of roads RD1, RD2, and RD3, the cognitive ability of the driver of vehicle VH, the nervousness of the driver of vehicle VH, the spatial recognition ability of the driver of vehicle VH, etc., the driver of vehicle VH may not be able to accurately recognize the "second fork," and vehicle VH may not be able to properly turn left onto road RD1 at the fork with road RD3 (for example, vehicle VH may turn left onto road RD1 at the fork with road RD2, or vehicle VH may go straight on road RD1 at the fork with road RD3). Therefore, in the navigation system SY of the first embodiment, the processing described below is performed in the machine learning model unit SY2, etc. so that the vehicle VH can appropriately turn left onto road RD1 at the junction with road RD3, regardless of the shapes of roads RD1, RD2, RD3, the cognitive ability of the driver of the vehicle VH, the nervousness of the driver of the vehicle VH, etc.

[0026] In the example shown in Fig. 6, the navigation information generation unit SY1 generates a driving route plan for the vehicle VH indicated by the arrow in Fig. 6 as a driving route plan (navigation information) for the vehicle VH from the current position of the vehicle VH to the destination of the vehicle VH. In order to travel from the current position of the vehicle VH to the destination of the vehicle VH, the vehicle VH needs to turn left onto road RD4 at the junction with road RD5, and then turn right onto road RD5 at the junction with road RD6 immediately after that. 6, if the following information is generated to be transmitted to the driver of vehicle VH: "Turn left in 300 meters ahead," and "Turn right in 50 meters ahead" immediately after vehicle VH turns left onto road RD4 at the junction with road RD5, the driver of vehicle VH would be required to perform cognitive and judgment operations in rapid succession, which could confuse the driver of vehicle VH and cause the driver to make an incorrect decision (for example, vehicle VH may continue straight on road RD5 without turning right onto road RD5 at the junction with road RD6). Alternatively, due to processing delays, communication delays, etc., the presentation of the "Turn right in 50 meters ahead" to the driver of vehicle VH may be delayed, causing vehicle VH to be unable to turn right onto road RD5 at the junction with road RD6. Furthermore, if the information to be transmitted to the driver of vehicle VH, based on the driving route plan (navigation information) of vehicle VH shown by the arrow in Figure 6, is generated as "Turn left 300 meters ahead, then turn right 70 meters ahead," the driver of vehicle VH will be required to perform cognitive judgment operations in rapid succession, which will place a strain on the cognitive judgment operations of the driver of vehicle VH when vehicle VH turns left onto road RD4 at the junction with road RD5, and the driver of vehicle VH may forget that vehicle VH needs to turn right onto road RD5 at the junction with road RD6. Therefore, in the navigation system SY of the first embodiment, the processing described below is performed in the machine learning model unit SY2, etc. so that the vehicle VH can appropriately turn left onto road RD4 at the junction with road RD5 and appropriately turn right onto road RD5 at the junction with road RD6, regardless of the shapes of roads RD4, RD5, RD6, the cognitive ability of the driver of the vehicle VH, the nervousness of the driver of the vehicle VH, etc.

[0027] In the example shown in Fig. 7, the navigation information generation unit SY1 generates a driving route plan for the vehicle VH indicated by the arrow in Fig. 7 as a driving route plan (navigation information) for the vehicle VH from the current position of the vehicle VH to the destination of the vehicle VH. In order to travel from the current position of the vehicle VH to the destination of the vehicle VH, the vehicle VH needs to go straight on road RD7 at the junction with a larger road RD8, and then turn left onto road RD7 at the junction with a smaller road RD9 immediately after (immediately after passing through intersection IS). If inappropriate transmission information that should be transmitted to the driver of vehicle VH is generated and the driver of vehicle VH does not have time to look at the image guidance displayed on the screen, the driver of vehicle VH may make an incorrect decision (for example, vehicle VH may turn left onto road RD7 at the junction with road RD8). Therefore, in the navigation system SY of the first embodiment, the processing described below is performed in the machine learning model unit SY2, etc. so that the vehicle VH can appropriately proceed straight on road RD7 at the junction with road RD8 and appropriately turn left on road RD7 at the junction with road RD9, regardless of the shapes of roads RD7, RD8, and RD9, the cognitive ability of the driver of the vehicle VH, the nervousness of the driver of the vehicle VH, etc.

[0028] In the example shown in Fig. 8, the navigation information generation unit SY1 generates a driving route plan for the vehicle VH indicated by the arrow in Fig. 8 as a driving route plan (navigation information) for the vehicle VH from the current position of the vehicle VH to the destination of the vehicle VH. In order to travel from the current position of the vehicle VH to the destination of the vehicle VH, the vehicle VH needs to continue traveling on the general road RDA without entering the expressway RDB from the general road RDA. Generally, when a vehicle enters an expressway from an ordinary road, the vehicle often travels in the right lane of the left and right lanes of the ordinary road before entering the expressway. In the example shown in Figure 8, in which the general road RDA and the expressway RDB are configured so that the vehicle VH travels in the left lane LN1 of the left lane LN1 and right lane LN2 of the general road RDA and enters the expressway RDB, the driver of the vehicle VH may mistakenly believe that as long as the vehicle VH is traveling in the left lane LN1 of the general road RDA, the vehicle VH can continue traveling on the general road RDA, and the vehicle VH may mistakenly enter the expressway RDB. Therefore, in the navigation system SY of the first embodiment, the processing described below is performed in the machine learning model unit SY2, etc. so that the vehicle VH can continue to travel on the general road RDA regardless of the left lane LN1 and right lane LN2 of the general road RDA, the shape of the expressway RDB, the cognitive ability of the driver of the vehicle VH, the tension of the driver of the vehicle VH, etc.

[0029] In the example shown in Fig. 9, the navigation information generation unit SY1 generates a driving route plan for the vehicle VH indicated by the arrow in Fig. 9 as a driving route plan (navigation information) for the vehicle VH from the current position of the vehicle VH to the destination of the vehicle VH. In order to travel from the current position of the vehicle VH to the destination of the vehicle VH, the vehicle VH needs to enter the expressway RDB from the general road RDA. As described above, generally, when a vehicle enters an expressway from an ordinary road, the vehicle often travels in the right lane of the left and right lanes of the ordinary road before entering the expressway. In the example shown in Figure 9, in which the general road RDA and the expressway RDB are configured so that the vehicle VH travels in the left lane LN1 of the left lane LN1 and right lane LN2 of the general road RDA and enters the expressway RDB, if transmission information such as "Entering the expressway 60 meters ahead" is generated as information to be transmitted to the driver of the vehicle VH based on the driving route plan (navigation information) of the vehicle VH shown by the arrow in Figure 9, the driver of the vehicle VH may mistakenly believe that the vehicle VH needs to be traveling in the right lane LN2 of the general road RDA in order for the vehicle VH to enter the expressway RDB from the general road RDA, and the vehicle VH may mistakenly continue traveling on the general road RDA (the vehicle VH may not be able to enter the expressway RDB). Therefore, in the navigation system SY of the first embodiment, the processing described below is performed in the machine learning model unit SY2, etc. so that the vehicle VH can enter the expressway RDB from the general road RDA regardless of the left lane LN1 and right lane LN2 of the general road RDA, the shape of the expressway RDB, the cognitive ability of the driver of the vehicle VH, the nervousness of the driver of the vehicle VH, etc.

[0030] 1 to 4, the navigation information generation unit SY1 stores the generated navigation information in the storage device VH72 as a log of instructions issued from the navigation system SY to the driver of the vehicle VH. The navigation information generation unit SY1 also transmits the generated navigation information to the server SV via the communication interface VH71 of the computer VH7 as a driving route plan for the vehicle VH. The navigation information generation unit SY1 then transmits the generated navigation information to the machine learning model unit SY2.

[0031] The machine learning model unit SY2 generates transmission information, which is information to be transmitted to the driver of the vehicle VH via the user interface VH1, based on the navigation information generated by the navigation information generation unit SY1. In the example shown in FIGS. 1 to 4, the machine learning model unit SY2 includes a first machine learning model SY21 (see FIG. 2) stored in a storage device VH72 of a computer VH7 of a vehicle VH, and a second machine learning model SY22 (see FIG. 2) stored in a storage device SV2 of a server SV. The second machine learning model SY22 includes a large-scale language model (LLM) SY22A (see FIG. 2), and the first machine learning model SY21 includes a lightweight large-scale language model SY21A (see FIG. 2) that is lighter than the large-scale language model SY22A. Note that the lightweight large-scale language model SY21A may be a so-called small-scale language model as long as it is lighter than the large-scale language model SY22A. The second machine learning model SY22 generates advance instructions for the first machine learning model SY21 based on, for example, destination information input by the driver of the vehicle VH via the user interface VH1 of the vehicle VH and personal information (characteristics of the driver of the vehicle VH) of the driver of the vehicle VH input in advance by the driver of the vehicle VH via the smartphone SP used by the driver of the vehicle VH (in the example shown in FIG. 2, the personal information of the driver of the vehicle VH is stored in the server SV). In detail, the second machine learning model SY22 generates advance instructions for the first machine learning model SY21 when the driver of the vehicle VH inputs destination information (i.e., before navigation is performed for the driver of the vehicle VH while it is traveling). In addition, the second machine learning model SY22 transmits the generated advance instructions to the first machine learning model SY21. In another example, personal information of the driver of the vehicle VH may be input in advance via the user interface VH1.

[0032] In the example shown in FIGS. 1 to 4, the first machine learning model SY21 generates, in real time (i.e., while the vehicle VH is traveling), output information to be transmitted to the driver of the vehicle VH based on advance instructions generated by the second machine learning model SY22 and navigation information generated by the navigation information generation unit SY1. In other words, the output information to be transmitted to the driver of the vehicle VH is generated by the first machine learning model SY21 of the vehicle VH, rather than by the second machine learning model SY22 of the server SV. Therefore, the output information can be generated in real time without being affected by communication delays between the vehicle VH and the server SV. The first machine learning model SY21 transmits the generated output information to the user interface VH1 as audio guidance output from a speaker (part of the user interface VH1) or as image guidance displayed on a display (another part of the user interface VH1). The user interface VH1 presents the transmission information (audio guidance, image guidance) generated by the first machine learning model SY21 to the driver of the vehicle VH. In other words, the transmission information generated by the machine learning model unit SY2 is transmitted to the driver of the vehicle VH from the user interface VH1 as audio guidance or image guidance. The communication information generated by the machine learning model unit SY2 includes information on the timing, order, granularity and specificity of the guidance, additional information to be added during the guidance, type of voice, tone, volume and speaking style, content of the image guidance to be displayed, display position, display size, color and font.

[0033] In the examples shown in Figures 1 to 4, the reaction of the driver of vehicle VH to the transmission information transmitted from the user interface VH1 to the driver of vehicle VH is used so that the driver of vehicle VH can drive vehicle VH appropriately based on the transmission information generated by the machine learning model unit SY2, regardless of the shape of the road on which vehicle VH is traveling, the cognitive ability of the driver of vehicle VH, the nervousness of the driver of vehicle VH, etc. 1 to 4, the processor VH73 of the computer VH7 of the vehicle VH functions as the determination unit SY3. The determination unit SY3 determines whether or not the driver of the vehicle VH has confusion in any of the perception, judgment, and operation of the transmission information, based on the reaction of the driver of the vehicle VH to the transmission information transmitted to the driver of the vehicle VH from the user interface VH1. When the judgment unit SY3 determines that there is confusion in any of the driver's perception, judgment, or operation regarding the transmission information, the machine learning model unit SY2 generates transmission information based on the transmission information transmitted to the driver of the vehicle VH via the user interface VH1 and the driver's response to that transmission information so as to suppress confusion in any of the driver's perception, judgment, or operation regarding the transmission information.

[0034] In detail, the machine learning model unit SY2 generates transmission information based on personal information related to any of the cognition, judgment, and operation of the driver of the vehicle VH, the transmission information generated by the machine learning model unit SY2, and the reaction of the driver of the vehicle VH to the transmission information, so as to suppress confusion in any of the cognition, judgment, and operation of the driver of the vehicle VH. In the example shown in Figure 2, personal information related to any of the driver's cognition, judgment, and operation of the vehicle VH is input into a smartphone SP used by the driver of the vehicle VH, the smartphone SP transmits the personal information to a server SV, and the personal information is used in the second machine learning model SY22 of the machine learning model unit SY2. In another example, personal information relating to any of the driver's perception, judgment, and operation of the vehicle VH may be entered into something other than the smartphone SP, such as a user interface VH1, and transmitted to the server SV from something other than the smartphone SP.

[0035] In the example shown in FIGS. 1 to 4, the determining unit SY3 includes an actuator control unit SY31 that controls the actuator VH5 (for example, a braking actuator, a steering actuator, a hazard lamp, etc.) based on the operation of the driver of the vehicle VH. Specifically, when the brake pedal is unnecessarily pressed down (i.e., the brake actuator is controlled) when the driver of the vehicle VH needs to operate to turn the vehicle VH right or left, the judgment unit SY3 (actuator control unit SY31) determines that there was confusion in either the driver's perception, judgment, or operation of the transmitted information. In addition, if the steering wheel operation fluctuates left and right when the driver of the vehicle VH needs to operate to turn the vehicle VH right or left (i.e., unnecessary control of the steering actuator is performed), the judgment unit SY3 (actuator control unit SY31) determines that there was confusion (hesitation) in either the driver's perception, judgment, or operation of the transmitted information. Furthermore, if the driver of vehicle VH turns on hazard lights that are unnecessary for turning vehicle VH right or left when the driver of vehicle VH needs to operate the vehicle VH to turn right or left, the judgment unit SY3 (actuator control unit SY31) determines that there was confusion in either the driver's perception, judgment, or operation of the transmitted information. The actuator control unit SY31 transmits a log of the operation of the driver of the vehicle VH when it is determined that there is confusion (hesitation) in any of the driver's perception, judgment, and operation of the transmission information to the second machine learning model SY22 of the server SV. When the actuator control unit SY31 determines that there is confusion (hesitation) in any of the driver's perception, judgment, and operation of the transmission information, the log of the instructions issued from the navigation system SY to the driver of the vehicle VH (navigation information generated by the navigation information generation unit SY1) is also transmitted to the second machine learning model SY22 of the server SV. When the actuator control unit SY31 determines that there is confusion (hesitation) in any of the driver's perception, judgment, or operation regarding the transmission information, the machine learning model unit SY2 generates transmission information based on the transmission information transmitted to the driver of the vehicle VH via the user interface VH1 and the driver's reaction to the transmission information (operation by the driver of the vehicle VH) so as to suppress confusion in any of the driver's perception, judgment, or operation regarding the transmission information.

[0036] 1 to 4, the determination unit SY3 includes a recognizer SY32 that detects information such as lane markings included in an image captured by a front camera VH2 mounted on the vehicle VH. Ladder marks are markings provided on roads when it is necessary to appropriately guide the flow of traffic, and include, for example, zebra strips. Specifically, when the recognizer SY32 detects that the vehicle VH has unnecessarily entered a zebra zone when the driver of the vehicle VH needs to operate to turn the vehicle VH right or left, the judgment unit SY3 determines that there was confusion in either the driver's perception, judgment, or operation of the transmitted information. In addition, if the recognizer SY32 detects that the vehicle VH has unnecessarily entered the shoulder of the road when the driver of the vehicle VH is required to operate the vehicle VH to turn right or left, the judgment unit SY3 determines that there was confusion in either the driver's perception, judgment, or operation of the transmitted information. Furthermore, if the recognizer SY32 detects that the vehicle VH has unnecessarily crossed a lane marking when the driver of the vehicle VH needs to operate to turn the vehicle VH right or left, the judgment unit SY3 determines that there was confusion in either the driver's perception, judgment, or operation of the transmitted information. In addition, when the recognizer SY32 detects that the vehicle VH has driven in violation of regulations (for example, driven in the wrong direction) when the driver of the vehicle VH needs to operate to turn the vehicle VH right or left, the judgment unit SY3 determines that there has been confusion in either the driver's perception, judgment, or operation of the transmitted information. The recognizer SY32 transmits a log of the operation of the driver of the vehicle VH when it is determined that there is confusion in any of the driver's perception, judgment, and operation of the transmission information to the second machine learning model SY22 of the server SV. When the recognizer SY32 determines that there is confusion in any of the driver's perception, judgment, and operation of the transmission information, the recognizer SY32 also transmits a log of the instructions issued from the navigation system SY to the driver of the vehicle VH (navigation information generated by the navigation information generation unit SY1) to the second machine learning model SY22 of the server SV. When the recognizer SY32 determines that there is confusion in any of the driver's perception, judgment, or operation regarding the transmission information, the machine learning model unit SY2 generates transmission information based on the transmission information transmitted to the driver of the vehicle VH via the user interface VH1 and the driver's reaction to the transmission information (operation of the vehicle VH by the driver of the vehicle VH) so as to suppress confusion in any of the driver's perception, judgment, or operation regarding the transmission information.

[0037] Furthermore, in the example shown in FIGS. 1 to 4, the determination unit SY3 includes a driver monitoring system SY33 that identifies the reaction of the driver of the vehicle VH to the transmission information generated by the machine learning model unit SY2. The driver monitoring system SY33 identifies the reaction of the driver of the vehicle VH to the transmission information generated by the machine learning model unit SY2 based on an image of the driver of the vehicle VH captured by the driver monitor camera VH3 mounted on the vehicle VH (i.e., determines whether there was confusion in any of the driver's perception, judgment, and operation of the transmission information). In addition, the driver monitoring system SY33 identifies the reaction of the driver of the vehicle VH to the transmission information generated by the machine learning model unit SY2 based on the voice of the driver of the vehicle VH collected by the microphone VH4 installed in the vehicle VH (i.e., determines whether there was confusion in any of the driver's perception, judgment, and operation of the transmission information). The driver monitoring system SY33 transmits a log of the operation of the driver of the vehicle VH when it is determined that there is confusion in any of the driver's perception, judgment, and operation of the transmission information to the second machine learning model SY22 of the server SV. When the driver monitoring system SY33 determines that there is confusion in any of the driver's perception, judgment, and operation of the transmission information, the log of the instructions issued from the navigation system SY to the driver of the vehicle VH (navigation information generated by the navigation information generation unit SY1) is also transmitted to the second machine learning model SY22 of the server SV. When the driver monitoring system SY33 determines that there is confusion in the driver of the vehicle VH's perception, judgment, or operation in response to the transmission information, the machine learning model unit SY2 generates transmission information based on the transmission information transmitted to the driver of the vehicle VH via the user interface VH1 and the driver's response to the transmission information, so as to suppress confusion in the driver of the vehicle VH's perception, judgment, or operation.

[0038] In the example shown in Figure 2, when it is determined that there was confusion (hesitation) in the driver's perception, judgment, or operation of the transmitted information, a log of the driver's operation of the vehicle VH (log of incorrect operation) is stored in both the vehicle VH and the server SV. For example, a log of operations by the driver of the vehicle VH when the vehicle VH travels along a route different from the route instructed by the navigation system SY corresponds to a log of erroneous operations. For example, a log of an operation by the driver of the vehicle VH unnecessarily depressing the brake pedal when an operation by the driver of the vehicle VH to turn the vehicle VH right or left is required corresponds to a log of an erroneous operation. For example, when the driver of vehicle VH needs to turn the vehicle VH right or left, and the steering wheel operation is unstable (there is uncertainty in the operation), the log of the driver's operation of vehicle VH corresponds to the log of an incorrect operation. For example, a log of an operation by the driver of vehicle VH who unnecessarily crosses a lane marking when an operation by the driver of vehicle VH to turn vehicle VH right or left is required would be considered a log of an incorrect operation. For example, a log of an operation by the driver of vehicle VH that unnecessarily causes vehicle VH to enter a zebra zone, road shoulder, etc. when an operation by the driver of vehicle VH to turn vehicle VH right or left is required would be considered a log of an incorrect operation. For example, a log of an operation by the driver of vehicle VH that turns on unnecessary hazard lights when the driver of vehicle VH needs to turn right or left would be considered an incorrect operation log. For example, a log of an operation by the driver of vehicle VH that causes the vehicle VH to drive in a manner that violates regulations (e.g., driving in the wrong direction) when an operation by the driver of vehicle VH to turn right or left is required would be considered a log of an incorrect operation.

[0039] In the example shown in FIGS. 1 to 4, the simulator SY4 is arranged in the server SV. Specifically, the processor SV3 of the computer constituting the server SV functions as the simulator SY4. The simulator SY4 simulates the behavior of the navigation information generation unit SY1. The second machine learning model SY22, in cooperation with the simulator SY4, iteratively predicts the behavior of the driver of the vehicle VH and optimizes the transmission information generated by the machine learning model unit SY2. The machine learning model unit SY2 adds or updates the transmission information based on the reaction of the driver of the vehicle VH to the transmission information transmitted to the driver of the vehicle VH from the user interface VH1 (feedback from the driver of the vehicle VH, such as voice, facial expressions, gestures, etc.).

[0040] As described above, the navigation system SY of the first embodiment utilizes the lightweight large-scale language model SY21A and the large-scale language model SY22A to perform navigation optimized for the communication ability, personality, preconceptions, and experience of the driver of the vehicle VH. Therefore, even in the case of a complex route, road shape, or poor communication conditions between the vehicle VH and the server SV, the driver of the vehicle VH can be appropriately navigated.

[0041] In one example of the navigation system SY of the first embodiment, when the driver of the vehicle VH inputs a destination for the current trip, the large-scale language model SY22A of the server SV uses the personal information of the driver of the vehicle VH (characteristics of the driver of the vehicle VH), a log of past erroneous decisions, and the instructions of the navigation system SY at that time to predict in advance what situations on the current trip the driver of the vehicle VH is likely to make an erroneous operation and what communication (content, timing, tone of voice, type of voice, and display content) to prevent the driver of the vehicle VH from making an erroneous operation, and provides advance instructions to the lightweight large-scale language model SY21A installed in the vehicle VH. The lightweight large-scale language model SY21A receives real-time navigation information output from the navigation information generation unit SY1 and, based on the advance instructions from the large-scale language model SY22A of the server SV, provides instructions to the driver of the vehicle VH using the communication means optimal for the driver of the vehicle VH (audio guidance and image guidance of the user interface VH1). The results of that trip are saved in a log and used to further improve navigation for future trips.

[0042] If all processing were to be done using a large-scale language model installed in the vehicle VH, the processing performance of the computer VH7 (e.g., an on-board ECU (Electronic Control Unit)) installed in the vehicle VH would become a bottleneck, causing delays and preventing navigation from being performed in a timely manner. If all processing were to be done using the large-scale language model SY22A on the server SV, communication delays between the server SV and the vehicle VH would prevent navigation from being performed in a timely manner. Therefore, in the navigation system SY of the first embodiment, LLM processing that requires real-time performance is executed by a lightweight large-scale language model SY21A installed in the vehicle VH, and LLM processing that does not require real-time performance is executed by a large-scale language model SY22A installed in the server SV.

[0043] In another example of the navigation system SY of the first embodiment, the navigation information generator SY1 may be provided not only in the vehicle VH but also in the server SV, and the simulator SY4 may execute a simulation of the navigation information generator SY1 in the server SV. In this example, the large-scale language model SY22A installed in the server SV simulates the itinerary and navigation instructions for that process in cooperation with the simulator SY4 of the navigation information generation unit SY1 in the server SV based on advance information such as the destination input in advance by the driver of the vehicle VH, simulates how the driver of the vehicle VH will react to the navigation instructions, and creates advance instructions based on the results. This process may be performed through multiple iterations of simulation and improvement.

[0044] The navigation system SY is very useful when the driver of the vehicle VH is driving to an unfamiliar destination or when the route to the destination is complicated. For example, with conventional navigation systems such as the technology described in Patent Document 1, when road shapes are complex, there is a risk that instructions issued by the navigation system may not be properly conveyed to the vehicle driver, causing the vehicle driver to take the wrong route, or that the navigation system may confuse and mislead the vehicle driver. Furthermore, with conventional navigation systems, there is a risk that delays in the voice guidance from the navigation system to the vehicle driver may occur due to the processing load on the navigation system's CPU (Central Processing Unit) and communication delays in GPS and networks, which may confuse the vehicle driver. As described above, the navigation system SY of the first embodiment can solve the problems that conventional navigation systems have.

[0045] In the navigation system SY of the first embodiment, warnings are not issued only for scenes that specifically match scenes in which the driver of vehicle VH's judgment error became apparent, but rather, cases (logs of the driver's operations of vehicle VH) in which potential risks and discomfort that have not become apparent in the past, that is, cases in which the driver of vehicle VH became confused or experienced cognitive load or stress, even though they did not ultimately impair driving, are used, and the machine learning model unit SY2 (LLM) performs processing to prevent the driver of vehicle VH from operating erroneously, taking into account the characteristics of the driver of vehicle VH.

[0046] [First Example] If the driver of vehicle VH has little driving experience or is inexperienced in simultaneous processing, the driver of vehicle VH may panic when multiple instructions are issued simultaneously from the navigation system SY, or the driver of vehicle VH may perform one operation and forget the existence of another operation. In the first example, when the navigation system SY issues the instruction "Turn left in 50 meters, then turn right in 30 meters," the driver of vehicle VH panics and turns right instead of left. In the second example, when the navigation system SY issues the instruction "Turn left in 50 meters, then turn right in 30 meters," the driver of vehicle VH becomes confused about which way to go after turning left and stops temporarily in the zebra zone. In the first embodiment, the logs of the first and second examples are stored in a log database. The large-scale language model SY22A of the server SV determines that the driver of the vehicle VH will panic when faced with complex instructions or multiple instructions issued simultaneously. Furthermore, the large-scale language model SY22A of the server SV takes the determination result into consideration and issues a pre-instruction to the lightweight large-scale language model SY21A of the vehicle VH to issue, for example, the following sequential processing instructions: Based on prior instructions from the large-scale language model SY22A of the server SV, the lightweight large-scale language model SY21A of the vehicle VH issues the instruction "Enter the left lane," then issues the instruction "Slow down to ●●km / h (a speed at which the driver of the vehicle VH can react in sufficient time even if an instruction to turn right is given after turning left)," then issues the instruction "Maintain speed and turn left 50 meters ahead," and then issues the instruction "Turn right 30 meters ahead."

[0047] As described above, in one example of the navigation system SY of the first embodiment, when the judgment unit SY3 determines that there is confusion in any of the driver of the vehicle VH's perception, judgment, and operation regarding the transmission information, which is information that should be transmitted to the driver of the vehicle VH via the user interface VH1, the machine learning model unit SY2 generates transmission information based on the transmission information transmitted to the driver of the vehicle VH via the user interface VH1 and the driver's reaction to the transmission information, so as to suppress confusion in any of the driver of the vehicle VH's perception, judgment, and operation. In a modified example of the first example of the navigation system SY of the first embodiment, when the judgment unit SY3 determines that there is confusion in any of the driver of the vehicle VH's perception, judgment, and operation in response to the transmission information, the machine learning model unit SY2 causes the navigation information generation unit SY1 to generate navigation information based on the transmission information transmitted to the driver of the vehicle VH via the user interface VH1 and the driver's response to the transmission information, so as to suppress confusion in any of the driver of the vehicle VH's perception, judgment, and operation. In detail, in a modified example of the first example of the navigation system SY of the first embodiment, the machine learning model unit SY2 causes the navigation information generation unit SY1 to generate navigation information based on personal information related to any of the cognition, judgment, and operation of the driver of the vehicle VH, the transmission information, and the reaction of the driver of the vehicle VH to the transmission information, so as to suppress confusion in any of the cognition, judgment, and operation of the driver of the vehicle VH. In other words, in a modified example of the first example of the navigation system SY of the first embodiment, the machine learning model unit SY2 causes the navigation information generation unit SY1 to generate navigation information so that the navigation system SY provides route guidance that will not cause confusion to the driver of the vehicle VH.

[0048] [Second Example] For example, when driving long distances on a highway, the driver of vehicle VH tends to become distracted, and there is a risk that the driver of vehicle VH may miss the communication information that should be transmitted to the driver of vehicle VH via user interface VH1. Therefore, in the second example of the navigation system SY of the first embodiment, when the driver of the vehicle VH is distracted while driving long distances on a highway (for example, when more than 10 minutes have passed since the last navigation), instructions are given to the driver of the vehicle VH by changing the tone, volume, and speaking style of the voice guidance.

[0049] [Third Example] The degree to which the driver of vehicle VH pays attention to the communication information transmitted to the driver of vehicle VH via user interface VH1 varies depending on whether the voice guidance from user interface VH1 to the driver of vehicle VH is delivered in a mechanical tone of speech, whether the voice guidance from user interface VH1 to the driver of vehicle VH is delivered in a casual tone of speech by someone of the opposite sex, whether the voice guidance from user interface VH1 to the driver of vehicle VH is delivered in a casual tone of speech by someone of the same sex, etc. Therefore, in the third example of the navigation system SY of the first embodiment, an analysis is performed of the tone of voice guidance that attracts the driver of the vehicle VH's attention to a great extent, and based on the results of the analysis, the tone of voice guidance from the user interface VH1 to the driver of the vehicle VH is optimized (i.e., the degree to which the driver of the vehicle VH pays attention is improved).

[0050] [Fourth Example] If the driver of the vehicle VH has a poor sense of distance, for example, when a voice guidance such as "50 m ahead" is given, the driver of the vehicle VH may not be able to accurately recognize the position 50 m ahead of the vehicle VH. Therefore, in a fourth example of the navigation system SY of the first embodiment, if it is determined that the driver of vehicle VH has a poor sense of distance from the personal information of the driver of vehicle VH or the log of the driver's operations of vehicle VH, instead of a voice guidance such as "50 meters ahead," a voice guidance such as "I think there is a store called XX on the right. After you pass it, turn on your left turn signal three seconds later and turn left at the first intersection" is given.

[0051] [Fifth Example] In a fifth example of the navigation system SY of the first embodiment, when the driver of the vehicle VH responds to the transmission information (guidance) generated by the machine learning model unit SY2 and transmitted to the driver of the vehicle VH via the user interface VH1 by saying, "Huh, wait a minute. I don't understand at all!", the driver's voice is collected by the microphone VH4. The driver monitoring system SY33 determines that the driver of the vehicle VH is confused in the perception of the transmission information generated by the machine learning model unit SY2, records the driver's response to the transmission information, and feeds the log back to the server SV. The second machine learning model SY22 (large-scale language model SY22A) of the server SV updates the transmission information generated by the machine learning model unit SY2 so that the driver of the vehicle VH no longer responds in this way. In a fifth example of the navigation system SY of the first embodiment, when the driver of the vehicle VH responds to the transmission information generated by the machine learning model unit SY2 and transmitted to the driver of the vehicle VH via the user interface VH1 by saying, "Should I turn left here? Or should I go straight? I don't know," the driver's voice is collected by the microphone VH4. The driver monitoring system SY33 determines that the driver of the vehicle VH is confused in his / her perception of the transmission information generated by the machine learning model unit SY2, records the driver's response in a log, and feeds the log back to the server SV. The second machine learning model SY22 of the server SV updates the transmission information generated by the machine learning model unit SY2 so that the driver of the vehicle VH no longer responds in this way. Furthermore, in a fifth example of the navigation system SY of the first embodiment, when the driver of the vehicle VH reacts to the transmission information (guidance) generated by the machine learning model unit SY2 and transmitted to the driver of the vehicle VH via the user interface VH1 by widening their eyes, making an irritated expression, or making gestures, the driver's reaction is captured by the driver monitor camera VH3. The driver monitoring system SY33 determines that the driver of the vehicle VH is confused in their perception in response to the transmission information generated by the machine learning model unit SY2, records the driver's reaction in a log, and feeds the log back to the server SV. The second machine learning model SY22 of the server SV updates the transmission information generated by the machine learning model unit SY2 so that the driver of the vehicle VH no longer shows such reactions.

[0052] [Sixth Example] In a fifth example of the navigation system SY of the first embodiment, when the driver of vehicle VH responds to the transmission information (instructions) generated by the machine learning model unit SY2 and transmitted to the driver of vehicle VH via the user interface VH1 by vocalizing, for example, "Huh, wait a minute. I don't understand at all!" or "Should I turn left here? Or should I go straight? I don't know," the driver monitoring system SY33 determines that there is confusion in the driver of vehicle VH's perception of the transmission information generated by the machine learning model unit SY2, and the first machine learning model SY21 (lightweight large-scale language model SY21A) installed in vehicle VH promptly issues additional instructions or generates and outputs instructions that are more specific than the transmission information (instructions) transmitted to the driver of vehicle VH.

[0053] As described above, the navigation system SY of the first embodiment can minimize confusion in any of the driver's perception, judgment, and operation of the vehicle VH.

[0054] Second Embodiment The navigation system SY of the second embodiment is configured similarly to the navigation system SY of the first embodiment described above, except for the points that will be described later.

[0055] As described above, in the example of the navigation system SY of the first embodiment (the example shown in FIGS. 1 to 4), the navigation system SY is configured by the computer VH7 mounted on the vehicle VH and the computer of the server SV. On the other hand, the navigation system SY of the second embodiment is configured by a computer VH7 mounted on the vehicle VH. Specifically, in one example of the navigation system SY of the second embodiment, the processor VH73 of the computer VH7 mounted on the vehicle VH functions as a simulator SY4. Also, in one example of the navigation system SY of the second embodiment, a machine learning model corresponding to the second machine learning model SY22 (see FIG. 2) is stored in the storage device VH72 of the computer VH7 of the vehicle VH. The machine learning model corresponding to the second machine learning model SY22 (see FIG. 2) of the navigation system SY of the second embodiment includes a large-scale language model corresponding to the large-scale language model SY22A (see FIG. 2).

[0056] FIG. 10 is a flowchart for explaining an example of processing executed by the processor VH73 of the computer VH7 mounted on the vehicle VH to which the navigation system SY of the second embodiment is applied. In the example shown in FIG. 10, in step S10, the navigation information generating unit SY1 generates navigation information based on the current position of the vehicle VH, map information, and destination information indicating the destination of the vehicle VH. Received from the interface. In step S11, the machine learning model unit SY2 generates transmission information, which is information to be transmitted to the driver of the vehicle VH via the user interface VH1, based on the navigation information generated in step S10. In step S12, the determination unit SY3 determines whether the driver of the vehicle VH was confused in any of the perception, judgment, and operation of the transmission information based on the reaction of the driver of the vehicle VH to the transmission information transmitted to the driver of the vehicle VH from the user interface VH1. If the determination is YES, the process proceeds to step S13, and if the determination is NO, the process proceeds to step S14. In step S13, when it is determined in step S12 that there is confusion in the driver of vehicle VH's perception, judgment, or operation regarding the transmission information, the machine learning model unit SY2 generates transmission information based on the transmission information transmitted to the driver of vehicle VH via the user interface VH1 and the driver's response to the transmission information, so as to suppress confusion in any of the driver of vehicle VH's perception, judgment, or operation. In step S14, for example, the decision unit SY3 decides whether or not to end the navigation, and if YES, the process shown in Fig. 10 ends, and if NO, the process returns to step S10.

[0057] <Third embodiment> The navigation system SY of the third embodiment is configured similarly to the navigation system SY of the first embodiment described above, except for the points that will be described later.

[0058] As described above, in one example of the navigation system SY of the first embodiment (the example shown in FIGS. 1 to 4), the second machine learning model SY22 of the server SV generates advance instructions for the first machine learning model SY21 based on destination information and personal information of the driver of the vehicle VH, and the first machine learning model SY21 of the vehicle VH generates transmission information, which is information to be transmitted to the driver of the vehicle VH, in real time based on the advance instructions generated by the second machine learning model SY22 and the navigation information generated by the navigation information generation unit SY1. In other words, in the navigation system SY of the first embodiment, communication delays between the server SV and the vehicle SV are taken into consideration. Meanwhile, advances in communication technology are expected to shorten the communication time between the server SV and the vehicle SV. Therefore, in one example of the navigation system SY of the third embodiment, the machine learning model unit SY2 is configured solely by the second machine learning model SY22 of the server SV. In detail, in one example of the navigation system SY of the third embodiment, the second machine learning model SY22 of the server SV generates, in real time, transmission information, which is information to be transmitted to the driver of the vehicle VH, based on the navigation information generated by the navigation information generation unit SY1.

[0059] As described above, embodiments of the navigation system, navigation method, and program of the present disclosure have been described with reference to the drawings. However, the navigation system, navigation method, and program of the present disclosure are not limited to the above-described embodiments and may be modified as appropriate without departing from the spirit and scope of the present disclosure. The configurations of the above-described embodiments may be combined as appropriate. In the above-described embodiments, the processing performed in the navigation system SY has been described as software processing performed by executing a program. However, the processing performed in the navigation system SY may also be hardware processing. Alternatively, the processing performed in the navigation system SY may be a combination of both software and hardware. Furthermore, the program stored in the storage device VH72 of the computer VH7 of the vehicle VH of the navigation system SY (the program that realizes the functions of the processor VH73 of the computer VH7 of the vehicle VH of the navigation system SY) may be recorded on a computer-readable storage medium such as a semiconductor memory, a magnetic recording medium, an optical recording medium, or the like and provided, distributed, etc. [Explanation of symbols]

[0060] SY...navigation system, SY1...navigation information generation unit, SY2...machine learning model unit, SY3...judgment unit, SY4...simulator, VH...vehicle, VH1...user interface, VH2...forward camera, VH3...driver monitor camera, VH4...microphone, VH5...actuator, VH6...sensor, VH7...computer, VH71...communication interface, VH72...storage device, SY21...first machine learning model unit, SY21A...lightweight large-scale language model, VH73...processor, SY1...navigation information generation unit, SY3...judgment unit, SY31...actuator control unit, SY32...recognizer, SY33...driver monitoring system, SP...smartphone, SV...server, SV1...communication interface, SV2...storage device, SY22...second machine learning model unit, SY22A...large-scale language model, SV3...processor, SY4...simulator

Claims

1. a navigation information generating unit that generates navigation information used to provide navigation to a driver of a vehicle based on a current position of the vehicle, map information, and destination information indicating a destination of the vehicle; a machine learning model unit that generates transmission information, which is information to be transmitted to the driver via a user interface, based on the navigation information generated by the navigation information generation unit; a determination unit that determines whether or not the driver has been confused in any of recognition, judgment, and operation of the transmission information based on the driver's reaction to the transmission information transmitted to the driver from the user interface, A navigation system in which, when the judgment unit determines that there is confusion in any of the driver's cognition, judgment, and operation in response to the communication information, the machine learning model unit generates the communication information, or the machine learning model unit causes the navigation information generation unit to generate the navigation information, based on the communication information transmitted to the driver via the user interface and the driver's reaction to the communication information, so as to suppress confusion in any of the driver's cognition, judgment, and operation.

2. the machine learning model unit includes a first machine learning model stored in a storage device of the vehicle and a second machine learning model stored in a storage device of a server; the second machine learning model generates advance instructions for the first machine learning model based on the destination information; The navigation system of claim 1 , wherein the first machine learning model generates the notification information in real time based on the advance instructions generated by the second machine learning model and the navigation information.

3. 2. The navigation system of claim 1, wherein the machine learning model unit generates the communication information or causes the navigation information generation unit to generate the navigation information based on personal information related to any of the driver's cognition, judgment, and operation, the communication information, and the driver's reaction to the communication information, so as to suppress confusion in any of the driver's cognition, judgment, and operation.

4. a simulator located on the server; the simulator simulates the behavior of the navigation information generator; The navigation system according to claim 2 , wherein the second machine learning model cooperates with the simulator to iteratively predict the driver's behavior and optimize the communication information.

5. The navigation system according to claim 1 , wherein the communication information is communicated to the driver from the user interface as audio guidance output from a speaker or as image guidance displayed on a display.

6. The navigation system according to claim 1 , wherein the determining unit includes a recognizer that detects information about a lane marking included in an image captured by a forward-facing camera mounted on the vehicle.

7. The navigation system according to claim 1 , wherein the determination unit includes an actuator control unit that controls a braking actuator or a steering actuator based on an operation by the driver.

8. 2. The navigation system according to claim 1, wherein the determination unit includes a driver monitoring system that identifies the driver's reaction to the communication information based on an image of the driver captured by a driver monitor camera or a voice of the driver collected by a microphone.

9. The navigation system of claim 5, wherein the transmission information generated by the machine learning model unit includes any of information regarding the timing, order, granularity and specificity of guidance, additional information to be added during guidance, type of voice, tone, volume and speaking style, content of image guidance to be displayed, display position, display size, color and font.

10. The navigation system according to claim 1 , wherein the machine learning model unit adds or updates the output information based on a response of the driver to the output information output from the user interface to the driver.

11. a first step in which a navigation system generates navigation information used to provide navigation to a driver of a vehicle based on a current position of the vehicle, map information, and destination information indicating a destination of the vehicle; a second step in which the navigation system generates transmission information, which is information to be transmitted to the driver via a user interface, based on the navigation information generated in the first step; a third step of the navigation system determining whether or not there is confusion in any of the driver's perception, judgment, and operation regarding the communication information based on the driver's reaction to the communication information transmitted to the driver from the user interface, A navigation method in which, when it is determined in the third step that there is confusion in any of the driver's perception, judgment, and operation regarding the communication information, the communication information is generated in the second step, or the navigation information is generated in the first step, based on the communication information transmitted to the driver via the user interface and the driver's reaction to the communication information, so as to suppress confusion in any of the driver's perception, judgment, and operation.

12. A processor of a computer installed in a vehicle a first step of generating navigation information used to provide navigation to a driver of a vehicle based on a current position of the vehicle, map information, and destination information indicating a destination of the vehicle; a second step of generating transmission information, which is information to be transmitted to the driver via a user interface, based on the navigation information generated in the first step; a third step of determining whether or not there is confusion in any of the driver's perception, judgment, and operation regarding the transmission information based on the driver's reaction to the transmission information transmitted to the driver from the user interface, A program in which, when it is determined in the third step that there is confusion in any of the driver's perception, judgment, and operation regarding the communication information, the communication information is generated in the second step, or the navigation information is generated in the first step, based on the communication information transmitted to the driver via the user interface and the driver's reaction to the communication information, so as to suppress confusion in any of the driver's perception, judgment, and operation.

Images