Driving assistance device and driving assistance program

Abstract

An exemplary driving assistance device has a control unit. The control unit acquires, in advance, driving characteristics of a driver based on data such as image information of the driver and an elapsed driving time thereof, determines a first output schedule of first voice guidance for improving the driving characteristics at the start of driving, determines a second output schedule of second voice guidance for route navigation, and determines, when the first output schedule overlaps with the second output schedule, a third output schedule corrected so that the output timing of the first voice guidance of an overlapping portion is earlier than the second output schedule.

Description

Driving Support Device and Driving Support Program 【0001】 The present disclosure relates to a driving support technology using voice. 【0002】 In recent years, in a vehicle, a plurality of voices such as navigation voice, ADAS (Advanced Driver-Assistance Systems) notification sound, audio voice, assistant voice by an agent, etc. can be uttered toward an occupant. 【0003】 Patent Document 1 discloses a process to be executed when output timings of a plurality of voices conflict in a vehicle. Specifically, when the output timings of voices from an automatic driving device and a car navigation device overlap, it is disclosed that content corresponding to the content of the voice output from the automatic driving device is displayed for a longer time than the time displayed when the timings do not overlap. And it is disclosed that by configuring in this way, the guidance by the automatic driving device can be surely conveyed to the driver. 【0004】 Japanese Patent Application Laid-Open No. 2020-87019 【0005】 By the way, vehicle occupants such as a driver tend to feel notifications bothersome when, for example, the frequency of notifications issued in the vehicle is high, or when notifications are made at inappropriate timings such as when a plurality of notifications are made at overlapping timings. Vehicle occupants who feel notifications bothersome may turn off the notification function, and in this case, a situation may occur where notifications for driving support are not appropriately made. 【0006】 In the configuration disclosed in Patent Document 1, when a plurality of voices are output at the same timing, although the voice output of the automatic driving device is stopped, a notification to the driver using a display device is made. For this reason, a plurality of notifications are still made at the same time, and there is a possibility that the driver may feel bothered. 【0007】 In view of the above points, an object of the present disclosure is to provide a technology capable of appropriately performing driving support while reducing the bother felt by the driver. 【0008】An exemplary driver assistance device of this disclosure is a driver assistance device for a driver, comprising a control unit. The control unit acquires in advance the driver's driving characteristics based on image information of the driver captured by a camera mounted on the vehicle and the elapsed time of that driving, or data on the movement of the vehicle caused by the driver's driving operations and the elapsed time of that driving. At the start of driving, it determines a first output schedule for a first voice guidance to improve the driving characteristics, a second output schedule for a second voice guidance for route navigation, and if the first output schedule overlaps with the second output schedule, it determines a third output schedule modified so that the output timing of the first voice guidance in the overlapping portion is earlier than that of the second output schedule. Based on the second and third output schedules, it executes control to output the first voice guidance and the second voice guidance. 【0009】 According to the exemplary configuration of this disclosure, by scheduling the voice output, the first voice guidance for improving the driver's driving characteristics is output at a time that does not overlap with the second voice guidance for route navigation. This reduces the possibility that the driver may find the guidance for driving assistance bothersome. Specifically, if the output schedules of the first voice guidance for improving the driver's driving characteristics and the second voice guidance for route navigation overlap, the schedule is adjusted so that the output timing of the first voice guidance is earlier than that of the second voice guidance. This allows the first voice guidance for improving driving characteristics to be output at an appropriate time to obtain the effect of its voice. In other words, according to the exemplary configuration of this disclosure, driving assistance can be provided appropriately while reducing the annoyance felt by the driver. 【0010】Diagram showing the general configuration of the cockpit system Diagram showing an overview of the operation of the cockpit system Diagram explaining sound management in the cockpit system Block diagram showing the general configuration of the driver assistance device Block diagram showing the general configuration of the data server Diagram showing an example of driver concentration level data Diagram showing an example of driver environment data Diagram showing an example of in-vehicle environment data Diagram explaining a method for predicting the timing of a decrease in concentration Level Image diagram of the voice content determination process performed by the voice content determination unit Flowchart illustrating the process of determining the content of the voice content Diagram illustrating the process of generating the voice deck by the voice deck generation unit Flowchart illustrating the process of placing the voice content onto the temporary voice deck Diagram illustrating the process of adjusting the voice deck in accordance with the schedule fluctuations of the navigation voice Adjustment between the voice content and audio Figures illustrating the process; Figures illustrating the adjustment process between content audio and audio sound; Figures illustrating the feedback information; Figures illustrating the data for feedback processing; Figures illustrating the update process of the concentration recovery effect index; Figures illustrating an example of an increase / decrease value conversion table used when updating the concentration recovery effect index; Figures illustrating an example of displaying the concentration change prediction; Figures illustrating other display examples of the concentration change prediction; Figures illustrating the first modified example of the display of the concentration change prediction; Figures illustrating the second modified example of the display of the concentration change prediction; Figures illustrating a specific example of a content display icon; Figures illustrating the third modified example of the display of the concentration change prediction; Figures illustrating further modifications of the configuration described in Figure 23; A flowchart illustrating the flow of output processing for content audio and navigation audio; Figures illustrating specific examples utilizing the technology of this disclosure; Figures illustrating other specific examples utilizing the technology of this disclosure. 【0011】 Hereinafter, exemplary embodiments of this disclosure will be described in detail with reference to the drawings. In the description of the embodiments, the same parts will be denoted by the same reference numerals, and redundant descriptions will be omitted unless particularly necessary. 【0012】<1. Cockpit System> [1-1. Outline of Configuration] Figure 1 is a diagram showing the schematic configuration of the cockpit system 100 according to the embodiment of this disclosure. More specifically, the cockpit system 100 is a system that provides information and entertainment to the occupants of the vehicle 200, thereby supporting driving and improving comfort inside the vehicle. The cockpit system 100 can also be called an IVI (In-Vehicle Infotainment) system. 【0013】 In detail, vehicle 200 is an automobile, but it may be any other type of vehicle. Furthermore, the technology disclosed herein may be applied to mobile bodies other than vehicles, such as flying vehicles or robots. 【0014】 As shown in Figure 1, the cockpit system 100 comprises a driver assistance device 1, various ECUs (Electric Control Units) 2, various sensors 3, a camera 4, a microphone 5, an operating device 6, an audio output device 7, and a display device 8. Each of these elements 1 to 8 is mounted on the vehicle 200. The cockpit system 100 also further comprises a server device 10 that is configured to communicate with the driver assistance device 1 via a communication network 9 such as the Internet. 【0015】 The driver assistance device 1 has a function to assist in the driving of the vehicle 200. The driver assistance device 1 also has a control function related to the voice output to assist in the driving of the vehicle 200. 【0016】 Each of the various ECUs 2 is provided to communicate with the driver assistance system 1. The driver assistance system 1 and the various ECUs 2 may communicate via wired or wireless connections. Examples of the various ECUs 2 include a navigation ECU, audio ECU, ADAS-ECU, body ECU, and concentration level measurement ECU. 【0017】The Navigation ECU is a control unit for the navigation equipment in the cockpit system 100 and transmits navigation information to the driver assistance system 1. The Audio ECU is a control unit for the audio equipment in the cockpit system 100 and transmits audio information (music playback information, etc.) to the driver assistance system 1. The ADAS-ECU is a control unit that comprehensively manages various safety driving support functions, such as adaptive cruise control, lane keeping assist, rear side warning, and collision mitigation braking. The ADAS-ECU transmits alert information (alert sound information, etc.) to the driver assistance system 1. The Body ECU is a control unit that centrally manages the interior and exterior lighting, doors, windows, mirrors, wipers, etc. of the vehicle 200 and transmits body-related information to the driver assistance system 1. The Concentration Level Measurement ECU is a control unit provided in a system that measures the driver's concentration level and transmits driver concentration level information to the driver assistance system 1. 【0018】 Each of the various sensors 3 is provided to communicate with the driver assistance device 1. The driver assistance device 1 and the various sensors 3 may communicate via a wired connection or via a wireless connection. The various sensors 3 include sensors that detect driving operation information of the vehicle 200 by the driver. Examples of the various sensors 3 include a vehicle speed sensor, steering angle sensor, brake sensor, accelerator sensor, etc. In addition, the various sensors 3 include sensors that detect the state of the vehicle 200 and sensors that detect the state of the interior (passenger compartment) of the vehicle 200. The various sensors 3 transmit the acquired sensor information to the driver assistance device 1. 【0019】 Camera 4 is provided to communicate with the driver assistance system 1. The driver assistance system 1 and camera 4 may communicate via a wired connection or wireless connection. Specifically, camera 4 includes an in-vehicle camera that photographs the driver and other occupants inside the vehicle, and an exterior camera that photographs the area around the vehicle 200. The exterior camera includes, for example, a front camera that photographs the area in front of the vehicle 200 and a rear camera that photographs the area behind the vehicle 200. The exterior camera may also include side cameras that photograph the left and right sides of the vehicle 200. Camera 4 transmits the captured data to the driver assistance system 1. 【0020】Microphone 5 is provided to communicate with the driver assistance device 1. The driver assistance device 1 and microphone 5 may communicate via a wired connection or wireless connection. Microphone 5 collects sound from inside the vehicle and transmits the collected sound data to the driver assistance device 1. 【0021】 The operating device 6 enables external instructions to be given to the driver assistance device 1. The operating device 6 is composed of, for example, a touch panel, buttons, levers, dials, etc. If the operating device 6 is a touch panel, it may be included in the display device 8. 【0022】 The audio output device 7 is provided to communicate with the driver assistance system 1. The driver assistance system 1 and the audio output device 7 may communicate via a wired connection or via a wireless connection. In detail, the audio output device 7 comprises an amplifier and a speaker. The audio output device 7 is placed in an appropriate location within the vehicle interior. There may be one or more audio output devices 7. The audio output device 7 outputs sound under the control of the driver assistance system 1. 【0023】 The display device 8 is provided to communicate with the driver assistance system 1. The driver assistance system 1 and the display device 8 may communicate via a wired connection or via a wireless connection. The display device 8 may be, for example, a liquid crystal display or an organic EL display. The display device 8 is placed in an appropriate location in the vehicle interior. The display device 8 may be, for example, a meter panel, a center display, a HUD (Head-Up Display), etc. The display device 8 displays information under the control of the driver assistance system 1. 【0024】 The server device 10 is located outside the vehicle 200 and exchanges information with the driver assistance system 1. Although only one server device 10 is shown in Figure 1, the driver assistance system 1 may be configured to cooperate with multiple types of server devices 10. The multiple types of server devices 10 may include servers with data storage functions, servers with generation AI (Artificial Intelligence) model execution functions, web servers, etc. 【0025】The cockpit system 100 may or may not include a server device 10. Furthermore, any of the elements shown as separate elements from the driver assistance system 1 in Figure 1 may be included in the driver assistance system 1. 【0026】 [1-2. Overview of Operation, etc.] Figure 2 is a diagram showing an overview of the operation of the cockpit system 100. Inside the vehicle (cockpit), navigation voice (hereinafter referred to as navigation voice), ADAS alert sound, audio sound, and other audio content are output from the audio output device 7. The audio of other audio content is the first audio of this disclosure. The navigation voice is the second audio of this disclosure. 【0027】 Navigation voice refers to the voice guidance emitted in conjunction with the navigation function, which provides route guidance to the destination, and is voice guidance for route navigation. ADAS alert sound refers to the warning sound emitted in conjunction with the ADAS function. Warning sounds include, for example, buzzer sounds and spoken words. Audio sound refers to the sound emitted by the audio function, specifically music. 【0028】 Other audio content refers to audio other than navigation voice, ADAS alert sound, and audio sound. Other audio content is voice guidance intended to improve the driver's driving characteristics, and specifically, it is the voice spoken by an agent (software). Examples of agent speech (voice content) include rest suggestions, topic suggestions, music change notifications, and safe driving advice. In this embodiment, the agent's voice is generated in the driver assistance device 1. Hereinafter, other audio content (agent's voice) will be referred to as content audio for convenience. Also, unless there is any particular reason to be confused, other audio content will simply be referred to as audio content. 【0029】The sounds output by the cockpit system 100 include voices and music provided to prevent a decrease in the driver's concentration. The cockpit system 100 adjusts the timing and frequency of the voices and music provided so that the driver 300 does not find them bothersome. The voices and music are output from the voice output device 7 after the timing and frequency have been adjusted (see Figure 2). 【0030】 Figure 3 is a diagram illustrating sound management in the cockpit system 100. As shown in Figure 3, navigation voice and content voice are managed on a time axis by constructing an audio deck. The audio deck is generated by pre-adjusting the output timing of the navigation voice and content voice for the entire trip (from the start to the end of the vehicle 200's journey). The audio deck constitutes a management table that manages the output schedule of the navigation voice and content voice. Note that the audio deck may be created not only for the entire trip, but also for a period from the current time to a predetermined time in the future. In this case, the audio deck is repeatedly created at predetermined time intervals. 【0031】 Audio sounds and event sounds are managed separately on a time axis. Audio sounds are, in principle, output in response to operations by the occupants of the vehicle 200, such as the driver. However, audio sounds may also be output based on commands from the driver assistance device 1. Event sounds are sounds with a high level of urgency (priority) for output in the vehicle 200, and in this embodiment, they are ADAS alert sounds. Event sounds are output when an event occurs. 【0032】Event sounds are output with priority over other sounds. If there are navigation voices, content voices, or audio sounds that are output at the same time as the event sound, they will be muted. Muted navigation voices and content voices may be output after the event sound, or they may be canceled. Due to the configuration of the audio deck, navigation voices will not be output at the same time as content voices. If there are audio sounds that are output at the same time as the navigation voices, the audio sounds will be muted. Content voices are output after their output timing has been adjusted so that they do not substantially overlap with the audio sounds. Details on this point will be described later. 【0033】 <2. Driving Assistance System> [2-1. Hardware Configuration] Figure 4 is a block diagram illustrating the schematic configuration of the driving assistance system 1. In Figure 4, the components necessary to explain the features of the driving assistance system 1 of this embodiment are shown, and components that do not require explanation are omitted. 【0034】 As shown in Figure 4, the driver assistance device 1 includes a control unit (controller) 11. The driver assistance device 1 further includes a storage unit 12 and a communication unit 13. 【0035】 The control unit 11 controls the overall operation of the driver assistance device 1. The control unit 11 is a computer device that includes an arithmetic circuit. In other words, the driver assistance device 1 is equipped with a computer. More specifically, the control unit 11 includes a processor that performs arithmetic processing, etc. The processor is composed of, for example, a CPU (Central Processing Unit). The control unit 11 may consist of one processor or multiple processors. If it consists of multiple processors, those processors should be connected to each other so that they can communicate with one another. 【0036】The storage unit 12 is configured to include volatile memory and non-volatile memory. The volatile memory is specifically RAM (Random Access Memory). The non-volatile memory is specifically ROM (Read Only Memory). The non-volatile memory may also be flash memory or a hard disk drive, etc. The non-volatile memory stores a program 121 and data that can be read by the computer. The program 121 includes a driver assistance program that causes the computer to function as a means of executing processes for driver assistance. 【0037】 The program 121 stored in the storage unit 12 may be provided, for example, on a computer-readable non-volatile recording medium. The non-volatile recording medium may be, for example, an optical recording medium (e.g., optical disc), a magneto-optical recording medium (e.g., magneto-optical disc), a USB memory, or an SD card, in addition to the non-volatile memory mentioned above. As another example, the program 121 may be provided from a program provision server via a communication line such as the Internet (a configuration provided by so-called download). 【0038】 In this embodiment, the functions of the control unit 11 are realized by the processor executing arithmetic processing according to the program 121 stored in the storage unit 12. The number of programs 121 that realize the functions of the control unit 11 may be one or more. 【0039】 Furthermore, at least a portion of the functions of the control unit 11 may be implemented by methods other than software, as in this embodiment. At least a portion of the functions of the control unit 11 may be implemented using, for example, an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array). In other words, at least a portion of the functions of the control unit 11 may be implemented by hardware using a dedicated IC or the like. Also, at least a portion of the functions of the control unit 11 may be implemented by using both software and hardware. 【0040】The communication unit 13 is configured as a communication interface having an interface circuit for connecting to the communication network 9 (see FIG. 1). 【0041】 [2-2. Functional units] As shown in FIG. 4, the control unit 11 includes, as its functions, an information management unit 111, a concentration calculation unit 112, a voice content determination unit 113, a voice deck generation unit 114, an audio sound management unit 115, an event sound management unit 116, a voice output control unit 117, a feedback processing unit 118, and a display control unit 119. The functions of these functional units 111 to 119 are realized by the processor executing arithmetic processing according to the above-described driving support program. 【0042】 (2-2-1. Information management unit) The information management unit 111 acquires and outputs information. The information management unit 111 acquires information from the above-described various ECUs 2, various sensors 3, camera 4, microphone 5, server device 10, and storage unit 12. Further, the information management unit 111 outputs information to the server device 10. 【0043】 Here, leaving the description of the driving support device 1, a description of the server device 10 will be given. FIG. 5 is a block diagram showing a schematic configuration of the data server 10a of the present embodiment. The data server 10a stores the personalization information of the driver of the vehicle 200. The data server 10a is an example of the server device 10. That is, the cockpit system 100 includes the data server 10a. Note that the data server 10a may not be provided, and in this case, the driving support device 1 may be configured to store the personalization information of the driver. 【0044】 As shown in FIG. 5, the data server 10a includes a control unit 101, a storage unit 102, and a communication unit 103. Since the hardware configuration of the data server 10a is the same as that of the above-described driving support device 1, a detailed description of the hardware configuration will be omitted. 【0045】Personalized information 1021 is stored in the storage unit 102 of the data server 10a. Personalized information 1021 is information about each individual driver operating the vehicle 200. The data server 10a may be configured to store the personalized information 1021 of only one driver, but usually it stores the personalized information 1021 of multiple drivers. That is, the personalized information 1021 shown in Figure 5 is stored in the storage unit 102 in a manner that allows individual drivers to be distinguished. 【0046】 More specifically, the personalization information 1021 includes concentration level information 1021a and feedback information 1021b. The concentration level information 1021a is information indicating the driver's concentration level tendencies when driving the driver's vehicle 200. The feedback information 1021b is information obtained by collecting the driver's response when the above content audio (agent's spoken voice) is output while driving. 【0047】 The information management unit 111 of the driver assistance device 1 can acquire personalized information stored in the storage unit 102 of the data server 10a. More specifically, the information management unit 111 can acquire individual driver concentration level information 1021a and feedback information 1021b via the communication network 9. The concentration level information 1021a will be described in detail below, and the details of the feedback information will be described elsewhere. The concentration level information 1021a is information regarding the driver's driving characteristics according to the elapsed driving time. 【0048】 The concentration level information 1021a is obtained by measuring the driver's level of concentration while they are actually driving the vehicle 200 and collecting the measurement data. The concentration level is measured by a concentration level measurement system (not shown) installed in the vehicle 200. The measurement data measured by the concentration level measurement system is output to the driver assistance device 1 from the concentration level measurement ECU (one of the various sensors 2) provided in the concentration level measurement system. The information management unit 111 of the driver assistance device 1 transmits the data on the level of concentration during driving, acquired from the concentration level measurement ECU, to the data server 10a as driving concentration level data. The concentration level measurement system may also be provided as a function of the concentration level calculation unit 112. 【0049】 Incidentally, the concentration measurement system determines the driver's concentration, for example, by counting the number of times the driver performs driving actions lacking concentration within a predetermined time. The concentration measurement system determines that the concentration is decreasing as the number of driving actions lacking concentration by the driver within the predetermined time increases. Examples of actions lacking concentration include sudden braking, sudden steering, sudden acceleration, lane departure, serpentine driving, and delays in actions with respect to signal changes and the behavior of the preceding vehicle. A specific example of a delay in action is the delay in the start of the vehicle 200 when changing from a red signal to a blue signal. Further, the concentration measurement system may determine that the driver's concentration has recovered, for example, when the number of driving actions lacking concentration within a predetermined time from the time when the driver performs a driving action lacking concentration is zero. In this case, the concentration measurement system determines that the driver's concentration has increased. The concentration may be determined using, for example, the driver's drowsiness information and posture information obtained using a camera, or may be determined by a combination of multiple measurement methods. The driving actions lacking concentration by the driver may be detected using, for example, the techniques described in JP-A-2023-178844 and JP-A-2023-110269. For example, when the time the driver's eyes are closed is longer than a certain time, or when actions such as rubbing the eyes, yawning, or a change in posture from the normal state are detected based on the image information acquired from the camera provided in the vehicle, the concentration measurement system may determine that the concentration has decreased. On the other hand, the concentration measurement system may determine that the concentration has increased when the driver's drowsiness has recovered, when the posture has returned to normal, or when no delay in the driver's actions is confirmed. Incidentally, the concentration measurement function may be a function executed by the driving support device 1. The driving support device 1 can preliminarily obtain the driving characteristics of the driver as the driving time elapses from the image information of the driver captured by the camera 4 mounted on the vehicle or the behavior information of the vehicle accompanying the driver's driving operation. 【0050】Figure 6A shows an example of the driving concentration data described above. The data shown in Figure 6A is data collected from the start to the end of the vehicle 200's journey (data for one trip). In the example shown in Figure 6A, concentration data is collected every 5 minutes. However, the interval for collecting concentration data may be changed as appropriate. In this embodiment, as an example, the concentration level is measured in five stages: "20%", "40%", "60%", "80%", and "100%". A higher number indicates a higher concentration level. The concentration level at the start of driving may be set to 100%. For example, the concentration level measurement system may decrease the concentration level by 20% each time the number of times the driver has engaged in driving behavior lacking concentration exceeds a predetermined number. Alternatively, the concentration level measurement system may increase the concentration level by 20% if the number of times the driver has engaged in driving behavior lacking concentration within a predetermined time period is zero. 【0051】 Furthermore, in the example shown in Figure 6A, at the same time that concentration level data is collected, driving environment data and in-vehicle environment data are collected and linked to the concentration level data. Driving environment data is external environmental data during the operation of vehicle 200, and is external environmental information. In-vehicle environment data is internal environmental data during the operation of vehicle 200, and is in-vehicle environment information. By collecting driving environment data and in-vehicle environment data linked to concentration level data, it becomes possible to obtain detailed trends regarding the driver's concentration level. Note that these processes may be performed by the driver assistance device 1. The driver assistance device 1 may acquire external environmental information and store it linked to driving characteristics (specifically, concentration level information). The driver assistance device 1 may also acquire in-vehicle environment information and store it linked to driving characteristics. 【0052】Figure 6B shows an example of driving environment data. In the example shown in Figure 6B, the driving environment data collected includes the presence or absence of preceding and following vehicles, road type, road conditions, road state, and weather. The presence or absence of preceding and following vehicles can be obtained, for example, using the information captured by camera 4. The road type can be obtained, for example, using the location information and map information of vehicle 200. Examples of road types include general roads, expressways, national roads, prefectural roads, municipal roads, and private roads. The road type may also include information on the number of lanes. Road conditions are information on road congestion (degree of congestion, etc.) and can be obtained, for example, using the information captured by camera 4 or road traffic information. Road state (road surface condition) can be obtained, for example, using the information captured by camera 4. Weather can be obtained, for example, from sensors and cameras placed on vehicle 200, or from internet information. Note that the weather is weather information at the location (current location) of vehicle 200. 【0053】 Figure 6C shows an example of in-vehicle environmental data. In the example shown in Figure 6C, the in-vehicle environmental data collected includes the amount of conversation, the temperature inside the vehicle, and the open / closed status of the windows. The amount of conversation is given, for example, as the percentage of time during which conversation occurred inside the vehicle over a predetermined period (e.g., the past 10 minutes). In-vehicle conversation can be obtained by analyzing the sound collected by the microphone 5. The temperature inside the vehicle can be obtained, for example, by a temperature sensor (one of the various sensors 3) placed inside the vehicle 200. The open / closed status of the windows can be obtained, for example, from information from the body ECU included in the various ECUs 2. 【0054】When the data server 10a receives driving concentration data (for example, the data shown in Figure 6A), it stores the received driving concentration data in the personalization information 1021 (more specifically, the concentration information 1021a). In this embodiment, the control unit 101 of the data server 10a, in a preferred configuration, generates averaged concentration time change information during driving using the already stored data and the new data if driving concentration data has already been stored. As the amount of stored data increases, the averaged concentration time change information reflects the driver's usual time change trend in concentration during driving. 【0055】 Furthermore, the control unit 101 of the data server 10a, in a preferred configuration, stores the timing of changes in concentration levels in the driving concentration level data, linked to environmental data (specifically, driving environment data and indoor environment data), as correction candidate information. As the accumulated data increases and time-dependent information on the usual concentration level becomes available, it becomes possible to extract environmental factors that make the driver susceptible to changes in concentration levels by comparing this time-dependent information on the usual concentration level with the accumulated correction candidate information. Environmental factors that make the driver susceptible to changes in concentration levels may be factors that decrease concentration levels, or factors that increase concentration levels. The data server 10a stores the environmental factors obtained in this way in the form of environmental correction information 1021a for correcting concentration levels. 【0056】 Furthermore, the results of the averaging process using the already accumulated data and the new data (averaging information), and the environmental correction information, may be transmitted to the driver assistance device 1 after their acquisition. The information management unit 111 of the driver assistance device 1 may be configured to store the acquired averaging information and environmental correction information in the storage unit 12 of the driver assistance device 1. In addition, the averaging process using the already accumulated data and the new data, and the generation of environmental correction information may be performed in the driver assistance device 1 rather than the data server 10a. For example, the driver assistance device 1 may identify environmental factors (external environmental factors and in-vehicle environmental factors) that cause changes in driving characteristics from among the external environmental information and in-vehicle environmental information, and generate the above-mentioned environmental correction information. 【0057】 (2-2-2. Concentration Level Calculation Unit) Returning to Figure 4, the concentration level calculation unit 112 will be explained. The concentration level calculation unit 112 predicts changes in the driver's concentration level during the upcoming driving. In addition, based on the results of the concentration level change prediction, the concentration level calculation unit 112 predicts the timing at which a decrease in concentration should occur in order to provide voice output (agent speech) to suppress the decrease in concentration. 【0058】 First, let's explain the concentration level change prediction performed by the concentration level calculation unit 112. The concentration level calculation unit 112 predicts the change in the driver's concentration level over the entire trip (from the start to the end of the vehicle 200's journey). More specifically, the concentration level change prediction is a prediction of the time change in the driver's concentration level (the transition of concentration level). The concentration level change prediction is performed on the premise that the route to be driven has been determined by the navigation function of the cockpit system 100. In other words, if the vehicle 200's journey route has not been determined by the navigation function, the concentration level change prediction will not be performed. Note that the navigation function may be a function of the driver assistance device 1. 【0059】 The concentration level calculation unit 112 (control unit 11) predicts changes in concentration level based on the driver's personalization information. This makes it possible to predict changes in concentration level that take into account the individual characteristics of each driver. As described above, in this embodiment, the driver's personalization information is stored in the storage unit 102 of the data server 10a. For this reason, the concentration level calculation unit 112 acquires the personalization information from the server device 10. By using the server device 10 in this configuration, it is possible to prevent the amount of data that needs to be held in the vehicle 200 from becoming too large. 【0060】As described above, the personalized information includes information about the driver's driving characteristics according to the elapsed driving time. For example, the personalized information is information about the time change in the driver's concentration level during past driving. With this in mind, in detail, the concentration level calculation unit 112 acquires information about the time change in the driver's concentration level during past driving from the data server 10a and makes a prediction of the change in concentration level based on the acquired past information. With this configuration, since the change in concentration level can be predicted according to the driver's actual driving data, it is possible to make a prediction of the change in concentration level that reflects the change trend of each driver's concentration level during driving. Hereinafter, the prediction of the change in concentration level obtained based on the information about the time change in the driver's concentration level during past driving may be referred to as the first concentration level change prediction. 【0061】 Furthermore, as mentioned above, the information on the time-dependent changes in a driver's concentration level during past driving sessions is an average of the accumulated data when multiple driving concentration level data points (see Figure 6A) have been stored. 【0062】 Furthermore, in this embodiment, the concentration level calculation unit 112 predicts changes in concentration level based on environmental information during driving. The external and internal environment of the vehicle 200 easily affects the driver's concentration level while driving. For this reason, by considering environmental information during driving when predicting concentration level, changes in concentration level can be predicted more accurately. 【0063】 The environmental information used for predicting changes in concentration levels during driving is the environmental information during driving that follows the route information created by the navigation function, and is the environmental information for the current drive. Preferably, the environmental information includes environmental information outside the vehicle and environmental information inside the vehicle. The environmental information outside the vehicle may be, for example, environmental information with the same items as shown in Figure 6B above. The environmental information inside the vehicle may be, for example, environmental information with the same items as shown in Figure 6C above. 【0064】The concentration level calculation unit 112 modifies the previously obtained first concentration level change prediction based on the acquired environmental information for the current operation and the environmental correction information stored in the data server 10a's storage unit 102. That is, in addition to the first concentration level change prediction, the concentration level calculation unit 112 obtains a second concentration level change prediction, which is a modified version of the first concentration level change prediction. For example, if information indicating that the weather is rainy is acquired as environmental information, the concentration level calculation unit 112 refers to the environmental correction information included in the personalization information 1021 to determine whether or not it is necessary to correct the concentration level change prediction when the weather is rainy. If it is determined that correction is necessary, the first concentration level change prediction is modified according to the environmental correction information. The same process is performed for all other environmental information, thereby obtaining the second concentration level change prediction. 【0065】 Furthermore, environmental information during operation may change over time. For this reason, the second concentration level change prediction may be configured to be recreated (updated) at regular intervals. 【0066】 Furthermore, in this embodiment, environmental information is used as a factor for correcting the level of concentration, but this is merely an example. Instead of environmental information, or in addition to environmental information, the driver's pre-driving activity history information may be used as a factor for correcting the level of concentration. The activity history information may include the driver's sleep information or overtime information from the day before driving. When this information is used, the personalized information stored in the storage unit 102 of the data server 10a will need to include information related to this information. 【0067】 The concentration level calculation unit 112, upon obtaining the first concentration level change prediction and the second concentration level change prediction, uses these to predict the timing of concentration decrease, which is the timing at which voice output (agent utterance) of the audio content should be performed. This will be explained with reference to Figure 7. Figure 7 is a diagram illustrating the method for predicting the timing of concentration decrease. 【0068】In the graph shown in Figure 7, the vertical axis represents concentration level, and the horizontal axis represents time. In the graph shown in Figure 7, the dashed line represents the first predicted change in concentration level, and the solid line represents the second predicted change in concentration level. In this embodiment, the timing at which the concentration level decreases compared to when the driver is driving normally is defined as the timing at which the audio content should be output. 【0069】 The concentration level calculation unit 112 performs a difference analysis using the first concentration level change prediction and the second concentration level change prediction. The concentration level calculation unit 112 then extracts locations in the second concentration level change prediction where the decrease in concentration is large compared to the first concentration level change prediction as concentration level decrease times. In other words, if there are points on the driving path that match external or internal environmental factors that cause changes in driving characteristics, the concentration level calculation unit 112 extracts the timing when the vehicle 100 passes over those points as concentration level decrease times. In the example shown in Figure 7, times t1, t2, and t3 are extracted as concentration level decrease times. For each of the extracted concentration level decrease times, the concentration level calculation unit 112 calculates a concentration level decrease level determined according to the difference between the first concentration level change prediction and the second concentration level change prediction, as shown in Figure 7. The concentration level calculation unit 112 stores the acquired concentration level decrease times and concentration level decrease levels in the storage unit 12 for use in subsequent processing. 【0070】 In this embodiment, the concentration reduction level increases for every 20% increase in the amount of concentration reduction. For example, if the concentration reduction is 20%, the concentration reduction level is "1", if it is 40%, it is level "2", and if it is 60%, it is level "3". However, this method of determining the concentration reduction level is merely an example and may be changed as appropriate. For example, the concentration reduction level may be classified into two stages: "low" or "high". 【0071】 Furthermore, as described above, the second concentration level change forecast may be configured to be recreated (updated) at regular intervals, taking into account fluctuations in environmental information. If such a configuration is adopted, the concentration level decrease time and concentration level decrease level will also be updated at regular intervals accordingly. 【0072】Furthermore, in this embodiment, the system is configured to obtain the concentration decrease time by performing a difference analysis between the first concentration change prediction and the second concentration change prediction, so that audio content is output only at times when the driver's concentration level is significantly lower compared to when they are driving normally. However, this is merely an example, and for example, all times when the concentration level is predicted to decrease in the second concentration change prediction may be extracted as the concentration decrease time. Alternatively, the system may be configured to extract the concentration decrease time only at times when the concentration decrease level is predicted to drop significantly below a preset level. 【0073】 Furthermore, the above configuration assumes the existence of driver personalization information and performs concentration level change predictions accordingly. However, there may be cases where personalization information is unavailable. If driver personalization information is unavailable, the system may use, for example, provisional information obtained by averaging information from an unspecified number of other drivers, or information from a large number of other drivers with similar ages, driving histories, etc., to perform concentration level change predictions. Alternatively, the system may notify the driver that the voice output function (agent speech function) for voice content to prevent a decrease in concentration cannot be used unless personalization information is accumulated, and concentration level change predictions may not be performed. 【0074】 (2-2-3. Voice Content Determination Unit) Returning to Figure 4, the voice content determination unit 113 will be explained. The voice content determination unit 113 determines the content of the voice output (agent utterance) of the voice content, which is performed to prevent a decrease in concentration. Figure 8 is an image diagram of the voice content determination process performed by the voice content determination unit 113. 【0075】 As shown in Figure 8, the voice content determination unit 113 determines the voice content of the voice content using various information and the generation AI model 105. 【0076】The various types of information include information acquired by the driver assistance system 1 from external sources and information generated within the driver assistance system 1. Examples of these types of information include driver information, passenger information, driver status information, in-vehicle status information, ADAS history information, driving environment information (external environment information), navigation information, concentration level prediction information (concentration level reduction information; see Figure 7), and feedback information. Detailed examples of this information will be described later. 【0077】 The generative AI model 105 is, in detail, a so-called Large Language Model (LLM). A Large Language Model is software that is capable of natural language processing according to a model trained using a large amount of text data. In this embodiment, the functions of the generative AI model 105 are performed by the processor of the server device 10 executing processing based on the model information of the generative AI model 105. When a prompt is input, the generative AI model 105 generates a response to the prompt and sends a reply. 【0078】 The processor in the server device 10 includes arithmetic circuits such as a CPU. The model information of the generated AI model 105 includes the structure and parameters of the generated AI model 105, as well as the code instructions for executing the generated AI model 105. 【0079】 Furthermore, the server device 10 having the execution function for the generated AI model 105 may be the same as the data server 10a described above, or it may be a different server. Also, the generated AI model 105 and its execution function may be provided in the driver assistance device 1. 【0080】The voice content determination unit 113 (control unit 11) creates a prompt to be input to the generating AI model 105 based on multiple pieces of information, including the driver's driving characteristics. The voice content determination unit 113 determines the voice content of the content voice (the content of the agent's utterance) based on the response from the generating AI model 105. In this configuration, voice content suitable for each driver to improve driving characteristics can be quickly determined using a wide variety of information. For example, concentration level prediction information may be used as the driver's driving characteristics. In this case, voice content suitable for each driver is determined to suppress a decrease in concentration level. 【0081】 In detail, the determination of voice content by the voice content determination unit 113 includes determining a solution to address the decrease in concentration and determining voice content according to the determined solution. This makes it possible to propose a method suitable for the driver to suppress the decrease in concentration and to output voice (agent speech) that matches the driver's preferences in the proposed method. 【0082】 The following describes the process by which the voice content determination unit 113 determines the content of the audio (the content of the agent's utterance), using specific examples. Figure 9 is a flowchart illustrating the process of determining the content of the audio. The process shown in Figure 9 is initiated assuming that the vehicle 200 is powered on, the driver assistance device 1 is operational, and the concentration level prediction process in the concentration level calculation unit 112 has been completed. In Figure 9, the processes indicated by thick lines are processes in which the generation AI model 105 is used. 【0083】 In step S1, the voice content determination unit 113 uses the generation AI model 105 to determine a solution to suppress the decrease in concentration. Specifically, the voice content determination unit 113 uses various information (see Figure 8) to create a prompt that requests the generation AI model 105 to propose a solution to suppress the decrease in concentration. The prompt is created according to a pre-prepared format. 【0084】The following is a specific example of a prompt requesting a solution. (Example of a prompt) Based on the following <background information>, please select the most effective solution to address the decrease in concentration from the <options> below. <Options> ・Suggest a break ・Provide conversation topics ・Suggest food and drink ・Switch music <Prerequisite Information> ・Driver Information Name: AA BB Age: 32 Gender: Male ・Passenger Information Name: AA CC (Wife) Age: 32 Gender: Female Name: AA DD (Eldest Son) Age: 4 Gender: Male ・Driver's State Awareness: Medium Emotion: Irritated Phone Calls: None Eating / Drinking: None Smoking: None Drowsiness: Low ・In-Car Conditions Conversation Volume: High Vehicle Speed: 82 km / h Room Temperature: 25℃ Windows Open / Closed: Closed ・ADAS (Number of Detections in the Past 30 Minutes) Forward Collision Warning: 3 Pedestrian Detection: 0 Lane Departure: 2 Reversing Vehicle Detection: 0 Forward Vehicle Departure Alarm: 0 ・External Environment Leading Vehicle: Yes Following Vehicle: No Road Conditions: Normal Road Condition: Comfortable Road Type: Expressway Weather: Sunny ・Navigation Information Current Location: 34.662429 N 135.159718 E Route Information: 60 minutes until arrival Destination: EE Park Concentration Decreasing Assessment Concentration Decreasing Time: 2024 / 10 / 16 16:17 Concentration Decreasing Level: 1 Concentration Recovery Effect Index for Each Driver Solution Break Suggestion: 16 Topic Provision: 25 Food & Drink Suggestion: 35 Music Switching: 68 【0085】 Once the voice content determination unit 113 has finished creating the prompt, it performs a process to input the created prompt to the generating AI model 105. In this embodiment, this process is the process of sending the prompt to the server device 10. After sending the prompt to the server device 10, the voice content determination unit 113 obtains the response from the generating AI model 105. 【0086】 Below is an example of the response when the above prompt is input to the generating AI model 105. (Example Response) The most effective solution to resolve the decrease in concentration is "changing the music". Reason: In the driver's concentration recovery effect index for each solution, changing the music shows the highest effect at "68". The driver's alertness level is moderate and their emotions are irritated, so changing the music can be expected to have a relaxing effect and change their mood. 【0087】The voice content determination unit 113 determines a solution to resolve the decrease in concentration level according to the response of the generating AI model 105. If the voice content determination unit 113 obtains the above example response, it sets the solution to "music switching". Once the solution is determined, the process proceeds to the next step S2. 【0088】 In step S2, the voice content determination unit 113 determines whether the solution is a suggestion to take a break. If the solution is a suggestion to take a break (Yes in step S2), the voice content determination unit 113 proceeds to step S3. If the solution is not a suggestion to take a break (No in step S2), the voice content determination unit 113 proceeds to step S4. 【0089】 In step S3, the voice content determination unit 113 uses the generation AI model 105 to create a text for suggesting a break. Specifically, the voice content determination unit 113 uses various information to create a prompt that requests the generation AI model 105 to create a text for suggesting a break. The prompt is created according to a pre-prepared format. 【0090】 The following is a specific example of a prompt requesting the creation of a text suggesting a break. In this example, reference data is input to the generating AI model 105 separately from the prompt. A specific example of the reference data is also shown below. The personal database (personal DB), notification DB, and music DB in the reference data may be data stored in the storage unit 12 of the driver assistance device 1, or information stored in the storage unit 102 of the data server 10a. The reference data may also be included in the prompt. 【0091】(Example prompt) You are currently driving. Refer to your personal database (restaurants and areas you visit) and current location to create a message guiding you to rest stops you can visit by car. Convert the location information to an address. Do not include a description of the location in the message. Do not include time information in the message. Also, consider business hours. If there is a parking lot nearby, you can stop there. Include information about whether the parking lot is free or paid, including the cost. Also, do not include details of the information listed in your personal database in the message. Adjust the number and ages of passengers to accommodate suitable rest stops. Please provide the official name of the rest stop. If you have created a similar rest stop suggestion within the past week, please reject that message and create a different one. If multiple message options are generated, select only one and create your message within the brackets []. 【0092】(Example of reference data) <Current date and time information> October 15, 2024, 15:00 <Current location information> 34.6933075 N 135.1929661 E <Driver information> Name: AA BB Age: 32 Gender: Male <Passenger information> Name: AA CC Age: 32 Gender: Female Name: AA DD Age: 4 Gender: Male <Personal DB> Places visited: MM Cafe, NN Shop Areas visited: Observation deck, park Hobbies: Gourmet food, movies, stretching Preferences: Coffee gummies Favorite songs / artists: ZZZZ, YYYY <Notification DB> ・2024 / 09 / 20 15:30 Food and drink suggestion: How about chewing gum to change your mood? ・2024 / 09 / 25 18:30 Topic provided: Topics related to XXXX - 2024 / 09 / 30 15:30 Food and drink suggestion: How about having some coffee to change your mood? - 2024 / 10 / 05 12:30 Break suggestion: How about stopping by XXXX near WW city? <Music DB> - 2024 / 09 / 20 12:30 Switching music: Song title [ZZZZ] Artist name [YYYY] - 2024 / 09 / 25 18:30 Switching music: Song title [GGGG] Artist name [TTTT] - 2024 / 09 / 30 10:30 Switching music: Song title [HHHH] Artist name [SSSS] - 2024 / 10 / 08 23:30 Switching music: Song title [IIII] Artist name [RRRR] - 2024 / 10 / 09 14:30 Music change: Song title [JJJJ] Artist name [QQQQ] 【0093】 Once the voice content determination unit 113 has finished creating the prompt that instructs the creation of the above-mentioned text, it performs a process to input the created prompt to the generation AI model 105. As described above, this process involves sending the prompt to the server device 10. After sending the prompt to the server device 10, the voice content determination unit 113 retrieves the text created by the generation AI model 105. Once the voice content determination unit 113 has created the text, it proceeds to step S10. 【0094】In step S4, the voice content determination unit 113 determines whether the solution is providing a topic. If the solution is providing a topic (Yes in step S4), the voice content determination unit 113 proceeds to step S5. If the solution is not providing a topic (No in step S4), the voice content determination unit 113 proceeds to step S6. 【0095】 In step S5, the voice content determination unit 113 uses the generation AI model 105 to create a text for providing a topic. Specifically, the voice content determination unit 113 uses various information to create a prompt that requests the generation AI model 105 to create a text for providing a topic. The prompt is created according to a pre-prepared format. 【0096】 The following is a specific example of a prompt requesting the creation of a topic-generating text. In this example, as well as the prompt, reference data is input to the generation AI model 105 separately. The specific example of the reference data can be the same as above, so its explanation is omitted. 【0097】(Example of a prompt) Refer to the personal database and current location information and create a message to introduce topics when the issue to be resolved is "preventing a decrease in concentration," in the following order of priority. Convert the location information to an address. Do not include a description of the location information in the message. Do not include time information in the message. Priority (1): Providing common topics, trivia, and relevant up-to-date information about the personal database's hobbies and preferences and facilities and land around the current location. Priority (2): Providing topics, trivia, and relevant up-to-date information related to the personal database's hobbies and preferences. Priority (3): Providing topics, trivia, and relevant up-to-date information about facilities and land around the current location. Clearly include the official names, real place names, and store names of the information you propose. The message should be written in a conversational style to provide information to the driver or passengers. However, do not include details of the contents listed in the personal database, the driver's name, or passenger information in the message. Also, adjust the topics suggested depending on the number and age of the passengers. Also, do not include the issue to be resolved in the message. If you have created the same topic introduction within the past week, please reject that message and create a different one. If multiple sentence suggestions are generated, select only one and create your sentence within the brackets []. 【0098】 Once the voice content determination unit 113 has finished creating the prompt instructing the creation of the above-mentioned text, it proceeds with processing in the same manner as in step S3, retrieves the text created by the generation AI model 105, and proceeds to step S10. 【0099】 In step S6, the voice content determination unit 113 determines whether the solution is a suggestion to eat or drink. If the solution is a suggestion to eat or drink (Yes in step S6), the voice content determination unit 113 proceeds to step S7. If the solution is not a suggestion to eat or drink (No in step S6), the voice content determination unit 113 proceeds to step S8. 【0100】In step S7, the voice content determination unit 113 uses the generation AI model 105 to create text for restaurant recommendations. Specifically, the voice content determination unit 113 uses various information to create a prompt that requests the generation AI model 105 to create text for restaurant recommendations. The prompt is created according to a pre-prepared format. 【0101】 The following is a specific example of a prompt requesting the creation of a text for a food and beverage suggestion. Note that, in this example as well, the system is configured to input reference data into the generation AI model 105 separately from the prompt. The specific example of the reference data can be the same as above, so its explanation is omitted. 【0102】 (Example of prompt) You are currently driving. Please create a text suggesting food and drink for when you are driving and the issue to be resolved is "preventing a decrease in concentration." Do not include the issue to be resolved in the text. Since the food and drink can be consumed while driving, the suggestion will be something like gum or coffee. However, if you have created the same food and drink suggestion within the past week or in the previous food and drink suggestion, please reject that text and create something different. If there are multiple generated text options, select only one and create your text in the brackets []. 【0103】 Once the voice content determination unit 113 has finished creating the prompt instructing the creation of the above-mentioned text, it proceeds with processing in the same manner as in step S3, retrieves the text created by the generation AI model 105, and proceeds to step S10. 【0104】 In step S8, the voice content determination unit 113 determines that the solution is music switching because no other solution method was determined from the pre-prepared set of solutions, leaving only music switching as the remaining solution method. The voice content determination unit 113 then uses the generation AI model 105 to create a message to notify the user of the music switching. When the solution is music switching, unlike in the case of other solutions, two prompts are created to create the message to notify the user of the music switching. Note that music switching includes not only switching music that is already playing, but also starting music from a state where no music is playing. 【0105】The voice content determination unit 113 first uses various information to create a prompt that requests the music (song) suggestion generation AI model 105. The prompt is created according to a pre-prepared format. 【0106】 The following is a concrete example of a prompt requesting music (song) suggestions. Note that, in this example as well, reference data is input to the generation AI model 105 separately from the prompt. The specific example of the reference data can be the same as above, so its explanation is omitted. 【0107】 (Example of prompt) If the issue to be resolved is "preventing a decrease in concentration," please refer to the following rules based on driver information and personal database (favorite songs, artists) and select a song that resolves the issue. If there are multiple song candidates generated, select only one and display only the song title and artist name in each of the brackets []. No text other than the song title and artist name is required. <Rules> - Make suggestions that take into account the current time and season. - Ensure that the song does not overlap with songs suggested in the music database when switching music within the past week. - If the only occupant is the driver, prioritize artists and songs that the driver likes to listen to and select a song that resolves the issue. - If there are multiple occupants, select music that is common to all occupants and that resolves the issue. In this case, do not rely solely on songs and artists in the personal database. 【0108】 Once the voice content determination unit 113 has finished creating the prompt described above, it processes the created prompt for input to the generation AI model 105, and then obtains the response (music suggestion) from the generation AI model 105. Once the voice content determination unit 113 has obtained the music suggestion (suggestion of song title and artist name), it proceeds to step S9. 【0109】 In step S9, the voice content determination unit 113 creates a prompt requesting the generation AI model 105 to create text to notify the driver, etc., that the music should be switched to the acquired song. The prompt is created according to a pre-prepared format. 【0110】The following is a specific example of a prompt requesting the creation of text to notify the user of a music change. Note that, in this example as well, the system is configured to input reference data into the generation AI model 105, separately from the prompt. The specific example of the reference data can be the same as above, so its explanation is omitted. 【0111】 (Example of a prompt) Let's say we're going to switch to the song [JJJJ] to address the issue of "preventing a decrease in concentration." If you were to provide an audio message to a driver who is currently driving, before the song changes, what should you say? Please create a message that matches the mood of the song. If multiple message options are generated, select only one and write your message in the brackets []. 【0112】 Once the voice content determination unit 113 has finished creating the prompt instructing the creation of the above-mentioned text, it proceeds with processing in the same manner as in step S3, retrieves the text created by the generation AI model 105, and proceeds to step S10. 【0113】 In step S10, the voice content determination unit 113 adjusts the text obtained in the previous text creation process to match the predetermined voice output time frame and user settings regarding content voice output (agent utterance). The voice content determination unit 113 also uses the generation AI model 105 for this adjustment process. The voice content determination unit 113 creates a prompt to request the generation AI model 105 to adjust the text. The prompt is created according to a pre-prepared format. A specific example of a prompt requesting text adjustment is shown below. 【0114】 (Example of prompt) Please refer to the following <Personal DB (Agent Utterance Specifications)> and modify the sentence "○○" to fit the duration of the agent utterance frame. Please create the modified sentence within []. <Personal DB (Agent Utterance Specifications)> Tone: Friendly Speed: 5 / 10 Pitch: 5 【0115】In the above prompt, "agent utterance frame" refers to a predetermined time frame for voice output. Also, the sentence "○○" is a sentence created in one of the processes in steps S3, S5, S7, and S9. 【0116】 Once the voice content determination unit 113 has finished creating a prompt instructing the above text to be modified, it proceeds with processing in the same manner as in step S3 and obtains the modified text created by the generation AI model 105. As a result, the voice content determination unit 113 completes the process of determining the content of the voice output (agent utterance) of the voice content performed to prevent a decrease in concentration. If there are multiple timings for outputting the voice content performed to prevent a decrease in concentration, it is preferable to have a configuration in which the voice content is determined for each of them. 【0117】 (2-2-4. Audio Deck Generation Unit) Returning to Figure 4, the audio deck generation unit 114 will be explained. The audio deck generation unit 114 generates an audio deck (see Figure 3) that manages the output schedule of navigation voice and content voice (voice of agent utterance). 【0118】 Figure 10 is a diagram illustrating the process of generating an audio deck by the audio deck generation unit 114. As shown in the upper part of Figure 10, when generating an audio deck, the audio deck generation unit 114 first constructs a temporary audio deck in which only the navigation voice is temporarily placed. The temporary audio deck constitutes a management table that manages the output schedule (output timing) of the navigation voice for the entire trip (from the start to the end of the vehicle 200's journey). As can be seen from this explanation, in other words, the audio deck generation unit 114 (control unit 11) creates the output schedule for the navigation voice. The audio deck generation unit 114 determines the schedule for the voice guidance for route navigation. 【0119】The temporary voice deck is constructed based on navigation information acquired by the information management unit 111. The navigation information includes route information to the destination, location information for voice guidance during route guidance, and information on the content of the voice guidance output during voice guidance. Voice guidance locations include, for example, places where turns occur or where lanes merge. 【0120】 When constructing a temporary audio deck, the estimated time when the navigation voice output will start is determined based on the above route information, the voice guidance location information, and the expected speed of the vehicle 200 (e.g., legal speed limit). Since there are usually multiple voice guidance locations, multiple estimated times for the start of navigation voice output will be determined. The end time for each navigation voice is determined based on the estimated start time of each navigation voice and the output duration (speech duration) of each navigation voice obtained from the voice content information. A temporary audio deck is obtained by arranging each navigation voice in chronological order based on its estimated start and end times. 【0121】 Once a provisional audio deck is obtained, the audio deck generation unit 114 generates an audio deck based on it, adjusting the output timing of the content audio (agent utterances) (see the lower part of Figure 10). In other words, the audio deck generation unit 114 (control unit 11) adjusts the output schedule of the content audio based on the output schedule of the navigation audio that has been created. With this configuration, it is easy to identify points where the audio output timings of the navigation audio and the content audio overlap, and the output schedules of the navigation audio and the content audio can be adjusted efficiently. 【0122】As described above, in this embodiment, the audio output of audio content is performed for the purpose of preventing a decrease in concentration. To this end, the audio deck generation unit 114 determines the timing of content audio output based on a prediction of changes in the driver's concentration level. In other words, the audio deck generation unit 114 determines the output schedule of audio guidance (content audio) to improve driving characteristics. Specifically, the audio deck generation unit 114 determines the timing of content audio output based on the above-mentioned time of decrease in concentration (see Figure 3). 【0123】 Specifically, the content audio is, in principle, positioned on the temporary audio deck so that the time of decreased concentration coincides with the end time of audio output (end of speech). If the timing of the content audio determined in this way overlaps with the output timing of the navigation audio, the content audio is positioned on the temporary audio deck as follows: The content audio is positioned on the temporary audio deck with its timing shifted forward so that it does not overlap with the timing of the navigation audio. 【0124】 In other words, the audio deck generation unit 114 adjusts the audio output schedule so that the content audio is output before the navigation audio when the content audio output timing overlaps with the navigation audio output timing. To put it another way, the audio deck generation unit 114 determines a modified schedule for the overlapping portion of the former when the audio deck generation unit 114's audio guidance output schedule overlaps with the navigation audio output schedule. With this configuration, the content audio is output at a timing that does not overlap with the navigation audio, thus reducing the possibility that the driver may find the driving assistance guidance bothersome. Furthermore, the content audio is configured to be output at a timing earlier than the navigation audio when the output timing overlaps with the navigation audio. With this configuration, it is possible to avoid overlapping between the two audios while outputting each audio at a timing appropriate to its purpose. 【0125】The method for placing content audio into the temporary audio deck will be further explained with reference to Figure 11. Figure 11 is a flowchart illustrating the flow of the process for placing content audio into the temporary audio deck. The process in Figure 11 begins after the temporary audio deck is constructed as described above. In Figure 11, the time of decreased concentration is represented by "tx". 【0126】 In step S11, the audio deck generation unit 114 determines whether the concentration decrease time tx overlaps with the navigation voice output timing. If the concentration decrease time tx overlaps with the navigation voice output timing (Yes in step S11), the process proceeds to step S12. If the concentration decrease time tx does not overlap with the navigation voice output timing (No in step S11), the process proceeds to step S13. 【0127】 In step S12, the audio deck generation unit 114 places the content audio before the navigation audio with overlapping timing in the temporary audio deck. Specifically, the content audio is placed in the temporary audio deck such that its end time is before (for example, immediately before) the start time of the navigation audio with overlapping timing. 【0128】 In step S13, the audio deck generation unit 114 determines whether the content audio can be placed in the temporary audio deck without overlapping with the navigation audio, when placing the content audio in the temporary audio deck so that the end time of the content audio coincides with the concentration decrease time tx. If it is possible to place them without overlapping (Yes in step S13), the process proceeds to step S14. If it is not possible to place them without overlapping (No in step S13), the process proceeds to step S15. 【0129】 In step S14, the audio deck generation unit 114 arranges the content audio in the temporary audio deck so that the end time of the content audio coincides with the concentration decrease time tx. 【0130】In step S15, as in step S12, the audio deck generation unit 114 places the content audio on the temporary audio deck such that the end time of the content audio is before (for example, immediately before) the start time of the navigation audio whose timing overlaps. 【0131】 If the concentration level calculation unit 112 calculates only one concentration level decrease time, the process shown in Figure 11 is performed once to complete the audio deck with the content audio placed on the temporary audio deck. If the concentration level calculation unit 112 calculates multiple concentration level decrease times, the process shown in Figure 11 is performed for each concentration level decrease time, and the audio deck with the content audio placed on the temporary audio deck is completed when the process shown in Figure 11 is completed for all concentration level decrease times. In the example shown in Figure 10, there are three concentration level decrease times (t1, t2, t3), so the process shown in Figure 11 is performed three times to complete the audio deck. 【0132】 Once the audio deck is completed in this way, in principle, navigation voice and content voice will be output according to the completed audio deck. However, if audio sound or event sound is output, the audio output control unit 117 may adjust the audio schedule, as described later. In other words, the generated audio deck may be changed when audio sound or event sound is output. The generated audio deck may also be changed if the timing of navigation voice output is shifted due to traffic congestion or changes in the speed of the vehicle 200. Furthermore, if there is a change in the concentration reduction time calculated by the concentration level calculation unit 112, the generated audio deck will be changed (reconstructed). 【0133】 Figure 12 is a diagram illustrating the adjustment process of the audio deck that is performed in response to changes in the navigation voice schedule. In Figure 12, the upper section shows the audio deck before adjustment, and the lower section shows the audio deck after adjustment. In the example shown in Figure 12, there is a delay in the movement of the vehicle 200, and the output timing of the navigation voices from the third output onwards is shifted backward in time. 【0134】In the example shown in Figure 12, the content audio output corresponding to the concentration decline time t2 is originally set to an earlier timing than it should be because its output timing would overlap with the navigation audio. The timing that should be set is when the end time of the content audio coincides with the concentration decline time t2. Due to the delay in the vehicle 200's progress, the concentration decline time t2 no longer overlaps with the navigation audio, so the content audio output corresponding to the concentration decline time t2 is shifted to the timing that should be set. 【0135】 Conversely to the example shown in Figure 12, changes in the vehicle 200's movement may cause the content audio output timing to overlap with the navigation audio, even if it was originally not. In such cases, the content audio output timing can be shifted to precede the overlapping navigation audio to prevent the two timings from overlapping. 【0136】 (2-2-5. Audio Sound Management Unit) Returning to Figure 4, the audio sound management unit 115 will be explained. The audio sound management unit 115 manages audio sounds (music). The audio sound management unit 115 monitors whether or not audio sound output is necessary. When it is necessary to output audio sound, the audio sound management unit 115 manages information on interludes and breaks between songs for the audio sounds to be output. The information on interludes and breaks between songs includes the time they start and their duration. The audio sound management unit 115 transmits the audio sound management information to the audio output control unit 117, which will be described later, as appropriate. 【0137】 (2-2-6. Event Sound Management Unit) The event sound management unit 116 manages event sounds. The event sound management unit 116 monitors whether or not event sound output is necessary. The event sound management unit 116 transmits event sound management information to the audio output control unit 117, described later, as appropriate. 【0138】(2-2-7. Audio Output Control Unit) The audio output control unit 117 adjusts the timing and other parameters for various audio outputs and outputs them from the audio output device 7. The audio output control unit 117 outputs navigation audio and content audio from the audio output device 7 according to the audio deck. In other words, the audio output control unit 117 (control unit 11) outputs navigation audio and content audio according to the adjusted schedule. 【0139】 The audio output control unit 117 prioritizes outputting event sounds when there is a request to output an event sound. When the audio output control unit 117 outputs an event sound, it mutes any navigation voice, content voice, or audio sound scheduled to be output at the same time. The audio output control unit 117 may output the muted navigation voice and content voice after the event sound is output, or it may cancel the output. 【0140】 The audio output control unit 117 prioritizes navigation sounds even when there is a request for audio output. When the audio output control unit 117 outputs navigation sounds, it mutes the audio sounds. When there is a request for audio output, the audio output control unit 117 adjusts the output timing of content audio and audio sounds. In other words, the audio output control unit 117 adjusts the output schedule of content audio based on the audio output information. This makes it possible to suppress situations where the audio sounds are muted when content audio is output, or where content audio and songs (lyrics) overlap, making it impossible to hear the songs (lyrics). As a result, the possibility of drivers and other passengers finding content audio bothersome can be reduced. 【0141】 Figure 13 is a diagram illustrating the adjustment process between content audio and audio. In Figure 13, the upper section shows the state before adjustment, and the lower section shows the state after adjustment. In the audio output information diagram (time-axis output information diagram) shown in Figure 13, the white areas indicate the intervals between songs or interludes. 【0142】As shown in Figure 13, the audio output control unit 117 adjusts the output of the content audio to coincide with the intervals between songs or interludes included in the audio output information. In other words, the content audio is output during the time when the audio is between songs or interludes. This configuration reduces the possibility that the driver or other occupants may find it difficult to hear the content audio due to the audio, or that the audio is being interfered with by the content audio. The audio may or may not be muted at the time the content audio is output. In either case, the possibility that the driver or other occupants may find the content audio bothersome is reduced. 【0143】 In Figure 13, the parts indicated by (a), (b), and (c) show the parts where the content audio output is adjusted to occur during the intervals between or between tracks of audio. As can be seen from these, there are cases where the content audio is adjusted (see (a)), where the audio is adjusted (see (b)), and where both the content audio and audio are adjusted (see (c)). Note that since the interludes in the audio cannot be adjusted, when adjusting the audio, the adjustment is made between tracks. 【0144】 The timing of the output of content audio and audio is adjusted as follows. Note that if the content audio is already being output during the interval or interlude of the audio tracks, no adjustment is performed. If the content audio is not being output during the interval or interlude, the audio deck searches for a time when the content audio will be output during the interval or interlude by shifting the content audio earlier than the scheduled time. If such a time is found in the audio deck, the position of the content audio is moved to that location. 【0145】If no such location is found, but there is a timing that overlaps with a break between songs or an interlude before the scheduled time for outputting the content audio, the following processing is performed: If the found location overlaps with a break between songs, the end time of the break between songs is extended so that the content audio finishes within the time of the break between songs. If the found location overlaps with an interlude, the output time of the content audio is adjusted, or the content of the content audio is modified, so that the output of the content audio finishes within the interlude time. The adjustment of the output time of the content audio or the modification of the content audio content may be performed using the above-mentioned generation AI model 105, but may also be performed by other methods. As an alternative method, for example, at least one short version of the content audio is prepared in advance, and the prepared content audio is output during the interlude. 【0146】 It is possible that there may be no overlapping timing between songs or interludes before the scheduled output time of the content audio. In such cases, the content audio may be output at the timing that overlaps with the audio. In this case, the audio may or may not be muted. 【0147】 Figure 14 shows a specific example of the adjustment process between content audio and audio. The upper part of Figure 14 shows the state before adjustment. In the figure shown in the upper part, the content audio is not played during the intervals or interludes between tracks of the audio. For this reason, the audio output control unit 117 shifts the content audio to an earlier time than scheduled in the audio deck, in order to find a time when the content audio output falls within the intervals or interludes between tracks. 【0148】However, in the example shown in Figure 14, since no such location exists, the system searches for a timing that overlaps with a break between songs or an interlude before the scheduled time for outputting the content audio. As a result, a break between songs is found, and the placement of the content audio is changed so that the start time of the content audio output matches the start time of the break between songs (see the middle diagram in Figure 14). Then, the end time of the break between songs is extended so that the end time of the content audio falls within the duration of the break between songs (see the bottom diagram in Figure 14). 【0149】 (2-2-8. Feedback Processing Unit) Returning to Figure 4, the feedback processing unit 118 will be explained. The feedback processing unit 118 (control unit 11) performs feedback processing on the output of content audio. In detail, the feedback processing unit 118 acquires the driver's concentration level after the output of content audio and performs feedback processing on the output of content audio based on the acquired concentration level. By performing such feedback processing, it is possible to acquire the driver's preference tendencies regarding the content of audio content emitted in order to suppress a decrease in concentration. Then, by repeating the feedback processing for each output of content audio, it is possible to bring the content of the audio content closer to the driver's preference. 【0150】 In this embodiment, the feedback information obtained by the feedback processing by the feedback processing unit 118 is stored as personalized information 1021 in the storage unit 102 of the data server 10a (see Figure 5). Alternatively, the feedback information may be stored in the storage unit 12 of the driver assistance device 1 instead of the storage unit 102 of the data server 10a. In this embodiment, the feedback processing unit 118 uses the feedback information to create prompts. That is, the feedback information is used in the process of determining the content of the audio using the generated AI model 105. This makes it possible to tailor the audio content to the driver's preferences. 【0151】Figure 15 shows a specific example of feedback information. In the example shown in Figure 15, the feedback information is a concentration recovery effect index for each solution method provided by the audio content (agent utterance) to suppress the decrease in driver concentration. The concentration recovery effect index indicates the degree to which concentration is restored when the audio content of the corresponding solution is output. In other words, the concentration recovery effect index can also be called the concentration recovery effect rate. The concentration recovery effect index is given as a number between 0 and 100, and the larger the number, the greater the concentration recovery effect that can be expected. The concentration recovery effect index is updated as needed after the audio content is output. 【0152】 It should be noted that, at the initial stage of using the driver assistance device 1 of this embodiment, no feedback information (feedback data) exists. For this reason, at the initial stage of use, the system may be configured to utilize average data obtained by averaging feedback information from an unspecified number of drivers. This average data can then be updated as needed in response to the driver's reaction after the audio content is played, thereby creating personalized data that reflects the tendencies of individual drivers. 【0153】 The following describes the feedback information update process illustrated in Figure 15. While the vehicle 200 is in motion, feedback processing data is collected for updating the feedback information. Figure 16 shows a specific example of feedback processing data. In Figure 16, the items "Date" and "Time" are time information collected while the vehicle 200 is in motion. In the example shown in Figure 16, data is collected at 5-minute intervals. 【0154】 In Figure 16, the item "Solution" represents the solution implemented to suppress a decrease in the driver's concentration while the vehicle 200 is in motion. Times for which no solution was proposed are indicated with a "-". In the example shown in Figure 16, at 15:30, music switching was implemented as the solution. 【0155】Furthermore, in Figure 16, the item "Concentration Level (%)" represents the measurement result of the driver's concentration level while the vehicle 200 is in motion. Specifically, the concentration level is measured in five stages: "20%", "40%", "60%", "80%", and "100%". 【0156】 In Figure 16, the item "Proposal Accepted" indicates whether the driver accepted the proposed solution. If the driver accepted the proposal, it is indicated as "YES". If the driver did not accept the proposal (proposal rejected), it is indicated as "NO". If there is no proposed solution, or if it is not possible to determine whether the proposal was accepted or not, it is indicated as "-". 【0157】 The decision regarding the acceptance of a proposal varies depending on the solution method. Furthermore, in cases where it is not possible to determine whether a proposal has been accepted, such as when the solution method is simply providing a topic of conversation, no acceptance determination is made. For example, the following cases will be judged as a rejection of the proposal, while all other cases will be judged as acceptance: If the solution method is a rest suggestion, the proposal will be judged as rejected if the destination is reached without taking a rest or if the suggested rest point is passed. If the solution method is music switching, the proposal will be judged as rejected if the music is changed or stopped by user operation. If the solution method is safe driving advice, the proposal will be judged as rejected if a clear rejection of the advice is shown or if actions are taken that do not follow the advice. 【0158】 If the feedback processing data includes results for the output of the content audio, the concentration recovery effect index is increased or decreased based on that portion and the data before and after it. In the example shown in Figure 16, it states that "the music change at 3:30 PM restored concentration by 20%, and the proposal was accepted." Accordingly, the concentration recovery effect index for the music change is updated. 【0159】Figure 17 is a diagram illustrating the update process for the concentration recovery effect index. The upper part of Figure 17 shows the data before the update, which is the same as the data shown in Figure 15. The lower part of Figure 17 shows the data after the update. Figure 17 also illustrates the update process when the feedback processing data shown in Figure 16 is obtained. As described above, in the example shown in Figure 16, the concentration level changed due to the music change. For this reason, the concentration recovery effect index for the solution "music change" is subject to update. In the example shown in Figure 16, it can be seen that the recovery effect index for music change, which was "50" before the update, has become "55" after the update, corresponding to the 20% recovery in concentration. 【0160】 In this embodiment, the increase / decrease value (increase / decrease points) of the concentration recovery effect index is determined using the increase / decrease value conversion table 122 shown in Figure 18. Figure 18 is a diagram showing an example of the increase / decrease value conversion table 122 used when updating the concentration recovery effect index. 【0161】 When evaluating the effect of a proposed solution on restoring concentration, the method of evaluation will vary depending on the original concentration recovery index. To illustrate with an extreme example, if the original concentration recovery effect index was "100" (the maximum value), it is not possible to increase the concentration recovery effect index simply because concentration has been restored. Taking this into consideration, the system is configured to utilize the increase / decrease value conversion table 122 shown in Figure 18. 【0162】 In the increase / decrease value conversion table 122 shown in Figure 18, the recovery effect index of the original (pre-update) concentration level is classified into 11 groups in increments of 10, such as "0 or more and less than 10", "10 or more and less than 20", and "20 or more and less than 30". For each group, the effect is divided into four stages: "large effect", "medium effect", "small effect", and "no effect", and increase / decrease points are assigned according to the magnitude of the concentration level recovery value. In addition, if a proposal is rejected, the impact on the change in concentration level becomes unclear, so it would normally be excluded from the evaluation. However, since it is thought that the proposal is not what the driver wants, increase / decrease points are assigned to reflect this point in the feedback information. 【0163】 In the example shown in Figure 18, a recovery of 60% or more in concentration is considered "highly effective," a recovery of 40% to less than 60% is considered "moderately effective," a recovery of 20% to less than 40% is considered "smallly effective," and no recovery in concentration or a decrease in concentration is considered "no effect." 【0164】 In this example, the concentration recovery effect index of the solution "switching music" is 50. Therefore, in Figure 18, we refer to the "50 or more and less than 60" category. Also, in the example shown in Figure 16, the concentration recovery is 20%, so in Figure 18, we refer to the "small effect" category. As a result, we get "+5.0" as the increase / decrease point from Figure 18. Therefore, as shown in Figure 17 above, the concentration recovery effect index of the solution "switching music" is increased by 5 points. 【0165】 (2-2-9. Display Control Unit) Returning to Figure 4, the display control unit 119 will be explained. The display control unit 119 (control unit 11) displays the predicted change in concentration level on the display device 8 (see Figure 1). The predicted change in concentration level is determined by the concentration level calculation unit 112 described above. Specific examples of the display device 8 include the meter panel, center display, HUD (Head-Up Display), ambient light, etc., all of which are installed inside the vehicle. With this configuration, the driver can be informed of their current level of concentration and implicitly encouraged to concentrate on driving. 【0166】 The concentration level change prediction may be displayed at all times, and the display mode may change according to changes in the prediction content. However, in this embodiment, the concentration level change prediction is displayed in accordance with the timing of the content audio output. With this configuration, the driver can sense from the display on the display device 8 that a decrease in their concentration level is predicted. 【0167】The display indicating a predicted decrease in concentration may start, for example, at the same time as the output of the content audio, or immediately before the output of the content audio. Furthermore, the display indicating a predicted decrease in concentration may end when the output of the content audio is completed, or after a certain period of time has elapsed since its completion. 【0168】 Figure 19A shows an example of a display of a predicted change in concentration. As shown in Figure 19A, the display of the predicted change in concentration may be a bar 81 displayed on the screen 8a of the display device 8. In the example shown in Figure 19A, the bar 81 is displayed at the bottom of the screen 8a and extends horizontally. However, the configuration of the bar 81 indicating the predicted change in concentration may be changed as appropriate. For example, the bar 81 may be displayed at the top of the screen 8a and extend horizontally. Alternatively, the bar 81 may be displayed at the left or right edge of the screen 8a and extend vertically. Furthermore, there may be multiple bars 81; in this case, for example, bars 81 extending vertically may be displayed at both the left and right ends of the screen 8a. 【0169】 Figure 19B shows another example of displaying the predicted change in concentration level. As shown in Figure 19B, the predicted change in concentration level may be displayed as an icon 82 on the screen 8a of the display device 8. In the example shown in Figure 19B, the icon 82 is displayed in the lower right portion of the screen 8a, but the placement of the icon 82 may be changed as appropriate. The appearance of the icon 82 may also be changed as appropriate. 【0170】 Figure 20 illustrates a first modified example of the display of predicted changes in concentration. As shown in Figure 20, the display of predicted changes in concentration may be configured to change according to the magnitude of the decrease in concentration level. By configuring it in this way, the driver can know their own situation regarding concentration in detail. 【0171】In this embodiment, the concentration level decrease is obtained based on the difference between the first concentration level change prediction and the second concentration level change prediction. In this embodiment, as an example, the concentration level decrease is classified into levels "1", "2", "3", and "4". Level "1" corresponds to a difference of 20% between the first concentration level change prediction and the second concentration level change prediction. Level "2" corresponds to a difference of 40% between the first concentration level change prediction and the second concentration level change prediction. Level "3" corresponds to a difference of 60% between the first concentration level change prediction and the second concentration level change prediction. Level "4" corresponds to a difference of 80% between the first concentration level change prediction and the second concentration level change prediction. 【0172】 As shown in Figure 20, in a configuration where changes in concentration are predicted using a bar 81, differences in the level of concentration decline may be displayed, for example, by changing the color of the bar 81. Also, in a configuration where changes in concentration are predicted using an icon 82, differences in the level of concentration decline may be displayed, for example, by changing the appearance of the icon 82. In the case of icon display as well, differences in the level of concentration decline may be displayed by changing the color. 【0173】 Figure 21 is a diagram illustrating a second modified example of the display of predicted changes in concentration levels. As shown in the left diagram of Figure 21, the display control unit 119 may be configured to further display content display icons 83 representing the content of the audio on the display device 8 in accordance with the output timing of the audio content. With this configuration, the driver can more easily grasp the content of the audio content from the screen display. 【0174】 In Figure 21, the content display icon 83 is positioned to the right of the icon 82 that displays the predicted change in concentration level, but this position may be changed as appropriate. Furthermore, the content display icon 83 changes depending on the content of the audio content (agent utterance). The content display icon 83 shown in Figure 21 indicates that the audio content is music switching. 【0175】Figure 22 shows specific examples of content display icons 83. Figure 22 shows specific examples of content display icons 83 when the content of the audio content (solutions to suppress the decrease in concentration) is "break suggestion," "topic introduction," "music switching," "vehicle operation," and "safe driving advice." In Figure 22, there are parts where multiple icons are shown in the icon example section, which indicates that any one of the multiple icons may be displayed. 【0176】 Returning to Figure 21, let's explain further variations. The right-hand side of Figure 21 shows an example screen when the driver's concentration level recovers after the output of the content audio. In the example in Figure 21, it is assumed that the driver is informed by the content audio that switching music is a solution to suppress the decrease in concentration. It is also assumed that the concentration level recovers within a certain period of time after the output of the content audio. In such a case, the screen 8a may be configured to display a content display icon 83 indicating music switching along with a bright smile icon 84. With this configuration, the driver can understand that their concentration level has recovered due to the music switching. 【0177】 The icon indicated by reference numeral 84 is an example of an effect confirmation icon that indicates the output effect of the content audio (effect on concentration level), and may be displayed not only when concentration level recovers, but also when there is no change in concentration level or when concentration level decreases. In addition, in the example shown in Figure 21, a content display icon 83 is displayed next to the effect confirmation icon 84, but the content display icon 83 may not be displayed, and only the effect confirmation icon 84 may be displayed alone. 【0178】Figure 23 illustrates a third modified example of the display of predicted changes in concentration levels. As shown in Figure 23, when displaying the content display icon 83, the display may also display text and keywords in conjunction with it. If there is sufficient space in the display area (screen 8a) of the display device 8, detailed information such as images related to the proposed content may be displayed in addition to text and keywords. As shown in Figure 24, the detailed information may be displayed by tapping the text and other display area next to the content display icon 83. Figure 24 is a diagram illustrating a further modified example of the configuration described in Figure 23. Furthermore, the configuration shown in Figure 24 is particularly effective when the display device 8 is a large screen such as a center display. 【0179】 For example, when displaying a content display icon 83 indicating a "break suggestion," text or keywords encouraging a break may be displayed along with the content display icon 83. In addition to text encouraging a break, detailed information such as information about the rest location may also be displayed. Furthermore, when displaying a content display icon 83 indicating a "topic suggestion," the keywords of the topic being offered, the text of the topic being offered, or a part of it may be displayed. In the configuration where a part of the text is displayed, the entire text may be displayed by tapping the screen as shown in Figure 24. 【0180】 Furthermore, when displaying content icon 83 indicating a "food and drink suggestion," the configuration may show the food and drink suggestion text or a portion thereof. Also, when displaying content icon 83 indicating a "music change," the configuration may show text informing the user that the music will be changed, or, in addition to the text, detailed information about the music to be played after the change. Also, when displaying content icon 83 indicating a "vehicle operation," the configuration may show text prompting vehicle operation (such as "Let's open the window"). When displaying content icon 83 indicating a "safe driving advice," the configuration may show the advice text, a portion thereof, or keywords. 【0181】<3. Flow of Audio Output Processing> Next, we will explain the flow of audio output processing performed by the driver assistance device 1. Here, we will mainly explain the flow of content audio and navigation audio output using the audio deck. 【0182】 Figure 25 is a flowchart illustrating the output processing flow of content audio and navigation audio. The processing shown in Figure 25 represents the technical content of the computer program to be implemented by the computer (processor in the control unit 11) of the driver assistance device 1. Furthermore, the processing shown in Figure 25 is initiated on the premise that the vehicle 200 is powered on and the driver assistance device 1 is operational, and that the route (guidance route) that the vehicle 200 will take has been determined by the navigation function of the cockpit system 100. 【0183】 Furthermore, when navigation is initiated by the navigation function, the collection of feedback processing data (see Figure 16, etc.) begins in order to obtain feedback information for the output of the content audio. The collection of feedback processing data ends when navigation ends. Then, for example, at the end of navigation, feedback processing is performed by the feedback processing unit 118 to obtain feedback information. 【0184】 In step S21, the control unit 11 (concentration level calculation unit 112) predicts changes in concentration level. This allows the concentration level decrease time and concentration level decrease level to be obtained (see Figure 7, etc.). Once the concentration level change prediction is complete, the process proceeds to the next step S22. 【0185】 In step S22, the control unit 11 (voice content determination unit 113) uses the generated AI model 105 (see Figure 8, etc.) to determine the content of the voice content (agent utterance) to be output according to the time of decreased concentration. If multiple times of decreased concentration have been acquired in step S1, voice content will be output for each of them. For this reason, the voice content is determined for all of these voice content. Once the content of the voice content is determined, the process proceeds to the next step S23. 【0186】 In step S23, the control unit 11 (audio deck generation unit 114) generates an audio deck that manages the output schedules of content audio and navigation audio (see Figure 10, etc.). When generating the audio deck, the content audio and navigation audio are scheduled so that their output timings do not overlap. In this scheduling adjustment, if the output timing of the content audio, which is determined according to the time of decreased concentration, overlaps with the output timing of the navigation audio, the content audio is scheduled to be output earlier than the navigation audio. Once the audio deck is generated, the process proceeds to the next step S24. 【0187】 In step S24, the control unit 11 (audio output control unit 117) adjusts the audio output schedule as needed. In principle, content audio and navigation audio are output according to the previously generated audio deck. However, as described above, if output of event sounds (such as ADAS alert sounds) is required, if audio sound output is required, or if the vehicle 200's speed deviates from the prediction, it is necessary to adjust the schedule in the previously generated audio deck. In step S24, this schedule adjustment is performed as appropriate. Once the schedule adjustment in step S24 is completed, the process proceeds to the next step S25. 【0188】 In step S25, the control unit 11 (audio output control unit 117) determines whether or not a change in the concentration level prediction is necessary. Situations in which a change in the concentration level prediction is necessary include, for example, when the route set by the navigation function has been changed, or when the measured concentration level deviates significantly from the concentration level prediction by a value set in advance. In these cases, it is considered meaningless to output content audio to suppress the decrease in concentration level according to the current schedule. For this reason, if a change in the concentration level prediction is necessary (Yes in step S25), the process returns to step S21, and the processing from step S21 onwards is carried out. In other words, the audio deck is regenerated. If a change in the concentration level prediction is not necessary (No in step S25), the process proceeds to the next step S26. 【0189】 In step S26, the control unit 11 (audio output control unit 117) determines whether or not it is time to output navigation audio based on the audio deck. If it is time to output navigation audio (Yes in step S26), the process proceeds to the next step S27. If it is not time to output navigation audio (No in step S26), the process proceeds to step S28. 【0190】 In step S27, the control unit 11 (audio output control unit 117) causes the audio output device 7 to output navigation voice. Once the navigation voice output process is complete, the process proceeds to the next step S28. 【0191】 In step S28, the control unit 11 (audio output control unit 117) determines whether or not it is the timing to output content audio based on the audio deck. If it is the timing to output content audio (Yes in step S28), the process proceeds to the next step S29. If it is not the timing to output content audio (No in step S28), the process returns to step S24, and the processing from step S24 onwards is carried out. 【0192】 In step S29, the control unit 11 (audio output control unit 117) causes the audio output device 7 to output content audio. In conjunction with the output of the content audio, a bar 81 and an icon 82 (see Figures 19A and 19B) indicating a predicted change in concentration level (prediction of a decrease in concentration level) are displayed on the screen 8a of the display device 8. If the device is equipped with a content display icon 83 display function, the content display icon 83 (see Figure 21) is displayed in conjunction with the display of the bar 81 and icon 82. If the device is equipped with an effect confirmation icon 84 display function, the effect confirmation icon 84 (see Figure 21) is displayed after the content audio is output. Once the content audio output process is complete, the process proceeds to the next step S30. 【0193】In step S30, the control unit 11 (voice output control unit 117) checks whether there is a reason for termination. An example of a reason for termination is when the navigation ends due to the vehicle 200 arriving at its destination. If there is a reason for termination (Yes in step S30), the process shown in Figure 25 is completed. If there is no reason for termination (No in step S30), the process returns to step S24 and the processes from step S24 onward are performed. 【0194】 In the example shown in Figure 25, the output timing is checked in the order of navigation voice followed by content voice, but this order can be reversed. 【0195】 <4. Specific Examples> Figure 26 is a diagram showing a specific example of utilizing the technology of this disclosure. Figure 26 shows an example in which the route traveled by vehicle 200 includes both ordinary roads and expressways. 【0196】 Figure 26 contains two graphs showing the relationship between time and concentration level. Of these two, the upper graph shows the predicted change in concentration level, and the lower graph adds the actual change in concentration level to the upper graph. In both graphs, the dashed line represents the first predicted change in concentration level, and the thin solid line represents the second predicted change in concentration level. The thick solid line in the lower graph represents the actual change in concentration level. The first predicted change in concentration level, shown by the dashed line, can be said to be the concentration level trend during normal driving. The second predicted change in concentration level, shown by the thin solid line, can be said to be the predicted concentration level during the current driving session. 【0197】The driver assistance system 1 (control unit 11) identifies external environmental factors that cause changes in driving characteristics from external environmental information including the type of road the vehicle traveled on, congestion levels, road surface conditions, or weather information at the vehicle's location. In the example shown in Figure 26, as can be seen from the upper graph, the driver assistance system 1 (control unit 11) predicts that the driver's concentration level will drop significantly compared to normal at time t1, some time after entering the highway from an ordinary road, based on personalized information, etc. The driver assistance system 1 (control unit 11) then designates this time t1 as the concentration level reduction time. As shown in Figure 26, if there is a point on the driving route that matches the external environmental factors, the driver assistance system 1 (control unit 11) determines the timing at which the vehicle passes that point as the content audio output timing. 【0198】 Furthermore, in the example shown in Figure 26, the solution to prevent a decrease in concentration at time t1 is determined to be "music switching" using the generative AI model 105. Specifically, the generative AI model 105 is used to switch to song ZZZZ by artist YYYY. As a result, in order to suppress a decrease in concentration at time t1, the audio content "We will deliver XXXX by YYYY to chase away your sleepiness" is output so that the audio output ends before time t1. 【0199】 In the example shown in Figure 26, since the system is configured to output audio, the content audio is output between tracks of audio. Specifically, the timing of the content audio output is earlier than the timing at which the content audio output ends at time t1. Also, after the content audio is spoken, the audio switches from the originally scheduled music to the music announced by the content audio. 【0200】 Looking at the thick solid line in the lower graph of Figure 26, the concentration level at time t1 is higher than the predicted result during this operation (see the thin solid line). In other words, the output of the content audio and the switching of music suppressed the decrease in concentration level at time t1. 【0201】 Figure 27 shows another specific example of utilizing the technology of this disclosure. Figure 27 shows an example where the route traveled by vehicle 200 includes both roads with little pedestrian traffic and roads with heavy pedestrian traffic. In addition, as with Figure 26, Figure 27 also contains two graphs showing the relationship between time and concentration. These two graphs differ in their patterns of change regarding concentration, but their structure is the same as that of the graph in Figure 26. 【0202】 The driver assistance system 1 (control unit 11) identifies external environmental factors that cause changes in driving characteristics from external environmental information including the type of road the vehicle traveled on, congestion levels, road surface conditions, or weather information at the vehicle's location. In the example shown in Figure 27, as can be seen from the upper graph, the driver assistance system 1 (control unit 11) predicts, based on personalized information, that the driver's concentration level will drop significantly compared to normal at time t1, after some time has passed since driving on a less-trafficked road. The driver assistance system 1 (control unit 11) also predicts that the driver's concentration level will increase when entering a more pedestrian-heavy road, but will drop significantly compared to normal at time t2, when entering a less-trafficked road from a more pedestrian-heavy one. The driver assistance system 1 (control unit 11) then designates these times t1 and t2 as times of decreased concentration. In the example shown in Figure 27, similar to Figure 26, the driver assistance device 1 (control unit 11) determines the timing of the vehicle passing over a point on the current driving route that matches the external environmental factors as the timing of content audio output. 【0203】 Furthermore, in the example shown in Figure 27, the solution to prevent a decrease in concentration at time t1 is determined to be "vehicle operation" using the generating AI model 105. Specifically, the vehicle operation is determined to be opening a window. As a result, in order to suppress a decrease in concentration at time t1, the content voice "Why don't you open a window?" is output so that the voice output ends at time t1. 【0204】Furthermore, in the example shown in Figure 27, the solution to prevent a decrease in concentration at time t2 is determined to be a "food and drink suggestion" using the generative AI model 105. Specifically, the suggestion is to chew gum. As a result, in order to suppress the decrease in concentration at time t2, the audio content "Why not try chewing gum to change your mood?" is output so that the audio output ends before time t2. 【0205】 In the example shown in Figure 27, audio is also output. Therefore, the content audio is appropriately adjusted to be output during the intervals between tracks or interludes of the audio. 【0206】 Looking at the thick solid line in the lower graph of Figure 27, the concentration levels at times t1 and t2 are higher than the predicted results for the current operation (see the thin solid line). In other words, outputting the above-mentioned content audio suppresses the decrease in concentration levels at times t1 and t2. 【0207】 <5. Notes and Precautions> Various technical features disclosed in the embodiments for carrying out the invention in this specification can be modified in various ways without departing from the spirit of the technical creation. Furthermore, the multiple embodiments and modifications disclosed in the embodiments for carrying out the invention in this specification may be combined to the extent possible. 【0208】 <6. Addendum> The following addendum is disclosed with respect to the multiple embodiments and modifications described above. 【0209】 This disclosure can take the following configurations (1) to (8). 【0210】(1) A driver assistance device for a driver, comprising a control unit, wherein the control unit pre-acquires the driver's driving characteristics based on image information of the driver captured by a camera mounted on the vehicle and the elapsed time of that driving, or data on the movement of the vehicle caused by the driver's driving operations and the elapsed time of that driving, determines a first output schedule for a first voice guidance to improve the driving characteristics at the start of driving, determines a second output schedule for a second voice guidance for route navigation, determines a third output schedule that modifies the first output schedule so that the output timing of the first voice guidance in the overlapping portion is earlier than that of the second output schedule if the first output schedule overlaps with the second output schedule, and executes control to output the first voice guidance and the second voice guidance based on the second and third output schedules. 【0211】 (2) The driving support device according to (1), wherein the control unit acquires external environmental information including the type of road the vehicle has traveled on, congestion status, road surface conditions, or weather information at the location of the vehicle, stores it in association with the driving characteristics, identifies external environmental factors from the external environmental information that cause changes in the driving characteristics, and determines the first output schedule so that the first voice guidance is output when the vehicle passes over a point on the driving route that matches the external environmental factors. 【0212】 (3) The driving support device according to (1) or (2), wherein the control unit transmits the image information and its elapsed driving time, or the vehicle movement information and its elapsed driving time data, to a server device, and acquires the driver's driving characteristics generated by the server device. 【0213】 (4) The driving support device according to any one of (1) to (3), wherein the control unit creates a prompt to be input to a generating AI model based on a plurality of pieces of information including the driving characteristics, and determines the content of the first voice guidance based on the response from the generating AI model. 【0214】(5) The driving support device according to any one of (1) to (4), wherein the control unit further modifies the third output schedule based on audio output information. 【0215】 (6) The driving support device according to any one of (1) to (5), wherein the control unit modifies the third output schedule so that the output timing of the first voice guidance coincides with the interval or interlude included in the audio sound output information. 【0216】 (7) The driving support device according to any one of (1) to (6), wherein the control unit causes an icon representing the content of the first voice guidance to be displayed on a display device in accordance with the output timing of the first voice guidance. 【0217】 (8) A driving assistance program for causing a computer to perform the following processes: (8) Pre-acquiring the driver's driving characteristics based on image information of the driver captured by a camera mounted on the vehicle and the elapsed time of that driver, or data on the movement of the vehicle caused by the driver's driving operations and the elapsed time of that vehicle; determining a first output schedule for a first voice guidance to improve the driving characteristics at the start of driving; determining a second output schedule for a second voice guidance for route navigation; determining a third output schedule that, if the first output schedule overlaps with the second output schedule, modifies the timing of the output of the first voice guidance in the overlapping portion to be earlier than the second output schedule; and executing control to output the first voice guidance and the second voice guidance based on the second and third output schedules. 【0218】 1... Operation support device 8... Display device 10... Server device 10a... Data server 11... Control unit 83... Content display icon 105... Generated AI model 121... Program 300... Driver 1021... Personalization information 1021b... Feedback information

Claims

1. A driver assistance device for a driver, comprising a control unit, wherein the control unit pre-acquires the driver's driving characteristics based on image information of the driver captured by a camera mounted on the vehicle and the elapsed time of that driving, or data on the movement of the vehicle caused by the driver's driving operations and the elapsed time of that driving, determines a first output schedule for a first voice guidance to improve the driving characteristics at the start of driving, determines a second output schedule for a second voice guidance for route navigation, determines a third output schedule that modifies the first output schedule so that the output timing of the first voice guidance in the overlapping portion is earlier than that of the second output schedule if the first output schedule overlaps with the second output schedule, and executes control to output the first voice guidance and the second voice guidance based on the second and third output schedules.

2. The control unit acquires external environmental information including the type of road the vehicle traveled on, congestion status, road surface conditions, or weather information at the vehicle's location, stores it in association with the driving characteristics, identifies external environmental factors from the external environmental information that cause changes in the driving characteristics, and determines the first output schedule so that the first voice guidance is output when the vehicle passes over a point on the driving route that matches the external environmental factors.

3. The control unit transmits the image information and its elapsed operating time, or the vehicle movement information and its elapsed operating time data, to a server device, and obtains the driver's driving characteristics generated by the server device, the driving support device according to claim 1.

4. The driving support device according to any one of claims 1 to 3, wherein the control unit creates a prompt to be input to a generating AI model based on a plurality of pieces of information including the driving characteristics, and determines the content of the first voice guidance based on the response from the generating AI model.

5. The driving support device according to claim 1, wherein the control unit further modifies the third output schedule based on audio output information.

6. The driving support device according to claim 5, wherein the control unit modifies the third output schedule so that the output timing of the first voice guidance coincides with the interval or interlude included in the audio sound output information.

7. The driver assistance device according to claim 1, wherein the control unit causes an icon representing the content of the first voice guidance to be displayed on a display device in accordance with the output timing of the first voice guidance.

8. A driving assistance program for causing a computer to perform the following processes: acquiring in advance the driver's driving characteristics based on image information of the driver captured by a camera mounted on the vehicle and the elapsed time of that driver, or vehicle movement information caused by the driver's driving operations and the elapsed time of that driver; determining a first output schedule for a first voice guidance to improve the driving characteristics at the start of driving; determining a second output schedule for a second voice guidance for route navigation; determining a third output schedule that, if the first output schedule overlaps with the second output schedule, modifies the timing of the output of the first voice guidance in the overlapping portion to be earlier than the second output schedule; and executing control to output the first voice guidance and the second voice guidance based on the second and third output schedules.

Images