Drive support device, drive support method, and program

Abstract

To appropriately generate a voice answer to a voice question regarding ADAS by a driver.SOLUTION: A drive support device 10 includes: an acquisition part 131 acquiring a voice question regarding ADAS received from a driver on driving a vehicle 1; a voice recognition part 132 recognizing the voice question regarding the ADAS acquired by the acquisition part 131; and an answer generation part 133 generating a voice answer corresponding to the voice question regarding the ADAS recognized by the voice recognition part 132 using a large scale language model.SELECTED DRAWING: Figure 3

Description

[Technical field]

[0001] The present disclosure relates to a driving assistance device, a driving assistance method, and a program. [Background technology]

[0002] Patent Document 1 describes an agent device that receives vehicle status information from the vehicle, receives a question from a user, estimates the intent of the question, and obtains an answer to the question based on the estimated intent of the question. Patent Document 1 also describes that the question is converted into text by voice recognition. Patent Document 1 further describes that information related to the answer to the question is received from an agent server. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2021-117941 A Summary of the Invention [Problem to be solved by the invention]

[0004] Incidentally, Patent Document 1 gives examples of questions that may be asked by users, such as how to turn off the light, what the switch is for, and what the meter light (tire pressure warning light) is, but does not give examples of questions regarding ADAS (Advanced Driver-Assistance Systems). The ADAS functions include functions such as ACC (Adaptive Cruise Control) and lane change assist, which the driver cannot operate while the vehicle is stopped, but can operate only while the vehicle is moving. It is considered that questions about such ADAS functions are asked by the driver while the vehicle is moving. For example, in order to enable a driver who is unfamiliar with ADAS to operate such an ADAS function while the vehicle is moving, it is necessary to quickly generate a voice response corresponding to the question about such an ADAS function. However, the technology described in Patent Document 1 does not take into consideration the need to quickly generate a voice response to a question. Therefore, the technology described in Patent Document 1 may not be able to appropriately generate a voice response to a driver's voice question regarding the ADAS.

[0005] In view of the above, an object of the present disclosure is to provide a driving assistance device, a driving assistance method, and a program that can appropriately generate voice responses to driver's voice questions regarding ADAS. [Means for solving the problem]

[0006] One aspect of the present disclosure is a driving assistance device that includes an acquisition unit that acquires voice questions regarding an ADAS received from a driver while driving a vehicle, a voice recognition unit that recognizes the voice questions regarding the ADAS acquired by the acquisition unit, and an answer generation unit that uses large language models (LLMs) to generate voice answers corresponding to the voice questions regarding the ADAS recognized by the voice recognition unit. Effect of the Invention

[0007] According to the present disclosure, it is possible to appropriately generate voice responses to the driver's voice questions regarding the ADAS. [Brief description of the drawings]

[0008] [Figure 1] 1 is a diagram illustrating an example of a vehicle to which a driving assistance device according to a first embodiment is applied. [Diagram 2] 2 is a diagram illustrating an example of a specific configuration of the driving support device illustrated in FIG. 1. [Diagram 3] 2 is a diagram for explaining a first example of a process executed in a vehicle to which the driving assistance device of the first embodiment is applied. FIG. [Figure 4] 4 is a flowchart illustrating an example of a process executed by the driving assistance device of the first embodiment. [Diagram 5] FIG. 1 is a diagram illustrating an example of a vehicle to which a driving assistance device according to a second embodiment is applied. [Figure 6] FIG. 11 is a diagram showing an example of a flow of data in a vehicle to which a driving assistance device of a second embodiment is applied. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] Hereinafter, embodiments of a driving assistance device, a driving assistance method, and a program according to the present disclosure will be described with reference to the drawings.

[0010] First Embodiment Fig. 1 is a diagram showing an example of a vehicle 1 to which a driving support device 10 according to a first embodiment is applied. Fig. 2 is a diagram showing an example of a specific configuration of the driving support device 10 shown in Fig. 1. 1 and 2, a vehicle 1 includes a vehicle state sensor 2, a GPS (Global Positioning System) unit 3, a map information unit 4, a driver monitor camera 5, an HMI (Human Machine Interface) 6, and a driving support device 10. The vehicle state sensor 2, the GPS unit 3, the map information unit 4, the driver monitor camera 5, and the HMI 6 are connected to the driving support device 10 via an in-vehicle network 20. 1 and 2, the vehicle 1 is equipped with a vehicle state sensor 2, a GPS unit 3, a map information unit 4, a driver monitor camera 5, and an HMI 6, but in other examples, the vehicle 1 may not be equipped with all of them. In still other examples, the vehicle 1 may be equipped with other sensors (e.g., external sensors) in addition to all or some of them.

[0011] In the example shown in Figures 1 and 2, the vehicle condition sensor 2 includes, for example, a vehicle speed sensor, a brake pedal sensor, an accelerator pedal sensor, a lane marking detection device, a seat belt sensor, a door sensor, an ABS determination device that determines whether or not the ABS (anti-lock braking system) is in an operable state, and the like. In another example, the vehicle state sensor 2 may not include all of the above-mentioned sensors, etc. In yet another example, the vehicle state sensor 2 may include other sensors, etc. (for example, an external sensor, etc.) in addition to all or some of the above-mentioned sensors, etc.

[0012] 1 and 2, the driving assistance device 10 includes a communication interface (I / F) 11, a memory 12, a processor 13, and a signal line 14. The driving assistance device 10 can be configured by a computer such as a Raspberry Pi (registered trademark). The driving assistance device 10 may also be configured by a driving assistance ECU (electronic control unit). In the example shown in FIGS. 1 and 2, the processor 13 has a function as an acquisition unit 131, a function as a voice recognition unit 132, a function as a response generation unit 133, and a function as an update unit . The acquisition unit 131 acquires a voice question regarding the ADAS received by the HMI 6 from the driver who is driving the vehicle 1. The acquisition unit 131 also acquires a signal indicating the vehicle state detected by the vehicle state sensor 2 (a vehicle state signal which is an output signal of the vehicle state sensor 2). Furthermore, the acquisition unit 131 acquires position information indicating the position of the vehicle 1 identified based on a GPS signal received by the GPS unit 3 and map information stored in the map information unit 4. The acquisition unit 131 also acquires data of a user manual (instruction manual) stored in the memory 12, for example in PDF format. The voice recognition unit 132 recognizes a voice question regarding the ADAS acquired by the acquisition unit 131. The voice recognition unit 132 recognizes a voice question regarding the ADAS acquired by the acquisition unit 131 by using a technique similar to that described in paragraphs 0031 and 0033 to 0036 of Japanese Patent No. 7062958 (the explanations of FIG. 2A and FIG. 3 (voice recognition device 10A, voice recognition unit 30A, and information generation unit 31A) of corresponding U.S. Patent No. 11011167).

[0013] The answer generation unit 133 can use a large-scale language model. The large-scale language model is a language model constructed using a very large data set and deep learning technology. In the large-scale language model, the amount of calculation, the amount of data, and the number of parameters are increased compared to conventional natural language models. The large-scale language model is capable of fluent conversation close to that of a human being, and can perform various processes using natural language with high accuracy. Representative examples of large-scale language models include "BERT" announced by Google in 2018 and "GPT-3" announced by OpenAI in 2020. As an application example of a large-scale language model, "ChatGPT", which was announced in December 2022 and is fine-tuned for chat (dialogue), is known. The answer generation unit 133 can use a publicly known (for example, open-sourced) large-scale language model as the large-scale language model. The answer generation unit 133 generates a voice answer corresponding to the voice question regarding the ADAS recognized by the voice recognition unit 132 by using a large-scale language model installed in the driving assistance device 10. The update unit 134 updates the voice answer corresponding to the voice question regarding the ADAS, which is generated by the answer generation unit 133, in response to, for example, an update of the function of the ADAS.

[0014] FIG. 3 is a diagram for explaining a first example of a process executed in the vehicle 1 to which the driving assistance device 10 of the first embodiment is applied. In the example shown in Fig. 3, the HMI 6 receives a voice question such as "Can I turn on the ACC now?" from the driver of the vehicle 1, and the acquisition unit 131 acquires the voice question. The voice recognition unit 132 recognizes the voice question acquired by the acquisition unit 131. In detail, the voice recognition unit 132 converts the voice question into text. The answer generation unit 133 acquires the question converted into text by the voice recognition unit 132. When the HMI 6 receives a voice question, the vehicle speed sensor detects the vehicle speed of the vehicle 1 (55 miles per hour (approximately 88.5 km)), the brake pedal sensor detects the off state of the brake pedal, the accelerator pedal sensor detects the on state of the accelerator pedal, the seat belt sensor detects the fastening state of the driver's seat belt, the door sensor detects the closed state of the door of the vehicle 1, and the ABS determination device determines that the ABS is in an operable state (the ABS is not in an inoperable state). The acquisition unit 131 acquires these detection results and determination results as a vehicle state signal, and the answer generation unit 133 acquires the vehicle state signal. Furthermore, when the HMI 6 receives a voice question, the GPS unit 3 and the map information unit 4 identify that the vehicle 1 is traveling on a highway where lane marks (lane dividing lines) are provided. The acquisition unit 131 acquires the identification result as position information of the vehicle 1, and the answer generation unit 133 acquires the position information. Furthermore, when the HMI 6 accepts a voice question, the acquisition unit 131 acquires data of the user manual stored in the memory 12. The user manual describes the conditions for activating the ACC of the vehicle 1, such as the vehicle speed of the vehicle 1 being 20 miles per hour (approximately 32.2 km) or more (not below a threshold), the brake pedal being in an off state, the accelerator pedal being in an on state, the driver's seat belt being fastened, the doors of the vehicle 1 being in a closed state, the ABS being in an operable state, and the vehicle 1 being traveling on a highway or the like with lane marks. The answer generation unit 133 acquires the data of the user manual. In other examples, the conditions for activating the ACC of the vehicle 1 may be different from the conditions described above.

[0015] In the example shown in Figure 3, the answer generation unit 133 uses a large-scale language model to generate a voice answer corresponding to the voice question from the driver of vehicle 1, "Can you turn on ACC now?", based on the question from the driver of vehicle 1, "Can you turn on ACC now?", the vehicle state signal, the position information of vehicle 1, and the user manual. In detail, the answer generation unit 133 determines that the state of vehicle 1 and the position of vehicle 1 indicated by the vehicle state signal satisfy the conditions for activating the ACC of vehicle 1 described in the user manual, and generates a voice answer to the effect that the ACC can be activated. Furthermore, the answer generating unit 133 generates a voice answer explaining the operation of the driver of the vehicle 1 required to operate the ACC. Specifically, the answer generating unit 133 generates text information such as "The vehicle speed is higher than 20 miles per hour (approximately 32.2 km) and the vehicle 1 is traveling on a highway with lane markings, so ACC can now be activated. Click the ACC button to activate ACC," and generates a voice answer by executing voice conversion of the text information. The HMI 6 outputs the voice answer generated by the answer generating unit 133. In the example shown in FIG. 3, the answer generation unit 133 generates a voice answer explaining the operation by the driver of the vehicle 1 required to activate the ACC, but in other examples, the answer generation unit 133 does not need to generate a voice answer explaining the operation by the driver of the vehicle 1 required to activate the ACC.

[0016] 3, a voice answer corresponding to a voice question made by the driver of vehicle 1 while traveling at a speed of 55 miles per hour (approximately 88.5 km) is generated by the answer generating unit 133 without the need to set the vehicle speed of vehicle 1 to zero (without the need to reduce the vehicle speed of vehicle 1 to below a threshold). Therefore, the driver of vehicle 1 can immediately activate ACC in accordance with the voice answer generated by the answer generating unit 133 (i.e., without the need to reduce the vehicle speed of vehicle 1 to zero and then increase it to above the threshold). That is, in the example shown in FIG. 3, a voice response to the driver's voice question regarding ACC can be generated appropriately (quickly and accurately).

[0017] The name of ACC varies by automaker. For example, Toyota calls ACC Radar Cruise Control, and Mercedes calls ACC Distronic Plus. Therefore, in a second example of the process executed in the vehicle 1 to which the driving assistance device 10 of the first embodiment is applied, a measure is taken to deal with the fact that the names of the ACC differ among automobile manufacturers. Specifically, in the second example, information indicating the names of the ACC of each automobile manufacturer is stored in advance in the memory 12, for example, when the vehicle 1 is manufactured. When the HMI 6 receives a voice question from the driver of the vehicle 1, "At what vehicle speed can ACC be operated?", the answer generating unit 133 generates a voice answer of the vehicle speed of the vehicle 1 at which ACC can be operated (20 miles per hour or faster (approximately 32.2 km) or faster), and generates a voice answer explaining the name of ACC used by the automobile manufacturer that manufactured the vehicle 1. The answer generating unit 133 also generates a voice answer explaining the operation of the driver of the vehicle 1 required to operate ACC and the associated operation of the HMI 6 (for example, the display screen of the HMI 6, etc.). Furthermore, the answer generating unit 133 generates a voice answer explaining the conditions for operating the ACC of the vehicle 1, the following distance control executed while the ACC is operating, etc.

[0018] In the third example of the process executed in the vehicle 1 to which the driving assistance device 10 of the first embodiment is applied, a measure is taken against the fact that the name of the ACC differs among automobile manufacturers, similar to the above-mentioned second example. In the third example, the vehicle 1 does not need to include the GPS unit 3 and the map information unit 4. In the third example, when the HMI 6 receives a voice question from the driver of the vehicle 1 saying, "I'm driving on the △△ Expressway at 80 km / h. Can I activate ACC now?", the answer generating unit 133 generates a voice answer indicating that ACC can be activated while driving on the △△ Expressway at 80 km / h, and generates a voice answer explaining the name of ACC given by the automobile manufacturer that manufactured the vehicle 1. The answer generating unit 133 also generates a voice answer explaining the operation of the driver of the vehicle 1 required to activate ACC and the associated operation of the HMI 6 (for example, the display screen of the HMI 6, etc.). Furthermore, the answer generating unit 133 generates a voice answer explaining the change in behavior (acceleration) of the vehicle 1 in response to the change in behavior (lane change, acceleration) of a low-speed preceding vehicle while ACC is operating. The answer generating unit 133 also generates a voice answer explaining the conditions for activating the ACC of the vehicle 1, the vehicle distance control executed while ACC is operating, the driver's obligations while ACC is operating, etc.

[0019] In a fourth example of processing executed in a vehicle 1 to which the driving assistance device 10 of the first embodiment is applied, similar to the third example described above, the vehicle 1 does not need to be equipped with a GPS unit 3 and a map information unit 4. In a fourth example, when the HMI 6 receives a voice question from the driver of vehicle 1, "I'm moving vehicle 1 in a parking lot. Can ACC be activated at 5 km / h?", the answer generation unit 133 generates a voice answer explaining that the vehicle speed of vehicle 1 at which ACC can be activated is 20 miles per hour (approximately 32.2 km / h) or higher, and generates a voice answer explaining the name of ACC used by the automobile manufacturer that manufactured vehicle 1. Furthermore, the answer generation unit 133 generates a voice answer to the effect that ACC cannot be activated at 5 km / h.

[0020] In a fifth example of processing executed in a vehicle 1 to which the driving assistance device 10 of the first embodiment is applied, similar to the second example described above, information indicating the name of the ACC of each automobile manufacturer is stored in advance in the memory 12, for example, when the vehicle 1 is manufactured. When the HMI 6 receives a voice question from the driver of vehicle 1, such as "What operation is required to activate ACC?", using the name of ACC of an automobile manufacturer other than the automobile manufacturer that manufactured vehicle 1, the answer generation unit 133 generates a voice answer explaining that the name of ACC used by the driver is not the name of ACC of the automobile manufacturer that manufactured vehicle 1, but the name of ACC of the other automobile manufacturer. Furthermore, the answer generation unit 133 generates a voice answer that serves as useful advice (advice that enables the driver to easily distinguish between the ACC function of vehicle 1 and the ACC function of a vehicle manufactured by the other automobile manufacturer) to the driver who has become a user of vehicle 1 from the user of a vehicle manufactured by the other automobile manufacturer.

[0021] In a sixth example of processing executed in a vehicle 1 to which the driving assistance device 10 of the first embodiment is applied, when the HMI 6 receives a voice question from the driver of the vehicle 1, such as "What operations are required to turn off Road Sign Assist (a function that displays road signs recognized by a camera mounted on the vehicle 1 on the screen of the HMI 6)?", the answer generation unit 133 generates a voice answer explaining the operations required by the driver of the vehicle 1 to turn off Road Sign Assist and the associated operation of the HMI 6 (e.g., the display screen of the HMI 6) based on the question from the driver of the vehicle 1 and the explanation of the operations required to turn off Road Sign Assist contained in the user manual.

[0022] In a seventh example of the process executed in the vehicle 1 to which the driving assistance device 10 of the first embodiment is applied, when the HMI 6 receives a voice question from the driver of the vehicle 1, "ACC cannot be operated when the ABS is inoperable? What's the problem?", the answer generation unit 133 generates a voice answer to the effect that ACC cannot be operated when the ABS is inoperable, based on the question from the driver of the vehicle 1 and the conditions for operating the ACC of the vehicle 1 included in the user manual (such as "ABS must be in an operable state"). The answer generation unit 133 also generates a voice answer explaining the name of the ACC by the automobile manufacturer that manufactured the vehicle 1. The answer generation unit 133 also generates a voice answer explaining the reason why ACC cannot be operated when the ABS is inoperable. Furthermore, the answer generation unit 133 generates a voice answer explaining each of a plurality of conditions for operating the ACC of the vehicle 1.

[0023] In an eighth example of the process executed in the vehicle 1 to which the driving assistance device 10 of the first embodiment is applied, not only data of the user manual of the vehicle 1 but also data of the user manual of the old model of the vehicle 1 are stored in the memory 12. When the HMI 6 receives a voice question from the driver of the vehicle 1, such as "I changed from the old model of the vehicle 1 to the vehicle 1 (current model). What's new that I need to know?", the answer generation unit 133 extracts updated functions (new functions) in the vehicle 1 (current model) based on the question from the driver of the vehicle 1, the user manual of the vehicle 1 (current model), and the user manual of the old model of the vehicle 1, and generates a voice answer explaining the extracted functions.

[0024] In a ninth example of processing executed in the vehicle 1 to which the driving assistance device 10 of the first embodiment is applied, the HMI 6 receives a voice question from the driver who is driving the vehicle 1, such as "What is the range of vehicle speeds at which lane change assist (one of the functions of ADAS) can be activated?", and the acquisition unit 131 acquires the voice question. The voice recognition unit 132 recognizes the voice question acquired by the acquisition unit 131. In detail, the voice recognition unit 132 converts the voice question into text. The answer generation unit 133 acquires the question converted into text by the voice recognition unit 132. When the HMI 6 accepts a voice question, the acquisition unit 131 acquires data from the user manual stored in the memory 12. The user manual describes that the lane change assist operation condition of the vehicle 1 is that the vehicle speed of the vehicle 1 is approximately 85 to 130 km / h. The answer generation unit 133 acquires the data from the user manual. In other examples, the operating conditions for the lane change assist of the vehicle 1 may be different from the above-mentioned conditions.

[0025] In the ninth example, the answer generating unit 133 uses a large-scale language model to generate a voice answer corresponding to a voice question from the driver of the vehicle 1, "What is the vehicle speed range in which the lane change assist can be operated?", based on the question from the driver of the vehicle 1, "What is the vehicle speed range in which the lane change assist can be operated?" and a user manual. In detail, the answer generating unit 133 generates a voice answer explaining the vehicle speed range (approximately 85 to 130 km / h) in which the lane change assist can be operated.

[0026] Fig. 4 is a flowchart for explaining an example of a process executed by the driving assistance device 10 of the first embodiment. The process shown in Fig. 4 is executed, for example, while the vehicle 1 is traveling (specifically, when the vehicle speed of the vehicle 1 is higher than zero). In the example shown in FIG. 4, in step S10, the acquisition unit 131 acquires a voice question regarding the ADAS from the driver of the vehicle 1. In step S11, the acquisition unit 131 acquires a vehicle state signal. In step S12 , the acquisition unit 131 acquires the position information of the vehicle 1 . In step S13, the acquisition unit 131 acquires data of a user manual. In step S14, the voice recognition unit 132 recognizes the voice question regarding the ADAS of the driver of the vehicle 1 acquired in step S10, and converts the voice question into text. In step S15, the answer generation unit 133 uses a large-scale language model to generate a voice answer corresponding to the voice question regarding the ADAS of the driver of the vehicle 1 acquired in step S10, based on the question regarding the ADAS of the driver of the vehicle 1 that was converted into text in step S14, the vehicle state signal acquired in step S11, the position information of the vehicle 1 acquired in step S12, and the user manual indicated by the data acquired in step S13. In detail, the answer generation unit 133 generates a text answer corresponding to the voice question regarding the ADAS of the driver of the vehicle 1 acquired in step S10, and then vocalizes the text answer to generate a voice answer corresponding to the voice question regarding the ADAS of the driver of the vehicle 1.

[0027] <Second embodiment> The vehicle 1 to which the driving assistance device 10 of the second embodiment is applied is configured similarly to the vehicle 1 to which the driving assistance device 10 of the above-described first embodiment is applied, except for the points described below.

[0028] Fig. 5 is a diagram showing an example of a vehicle 1 to which the driving support device 10 of the second embodiment is applied. Fig. 6 is a diagram showing an example of a flow of data, etc. in the vehicle 1 to which the driving support device 10 of the second embodiment is applied. 5 and 6, the vehicle 1 is equipped with a communication device 7 that communicates with the outside of the vehicle 1. The answer generation unit 133 generates a voice answer corresponding to the voice question regarding the ADAS recognized by the voice recognition unit 132 by using a large-scale language model on, for example, a server outside the vehicle 1. Specifically, the answer generation unit 133 acquires a question converted into text by the voice recognition unit 132, and causes the communication device 7 to execute a process of transmitting the question to a large-scale language model outside the vehicle 1. The answer generation unit 133 also acquires a vehicle state signal generated by the vehicle state sensor 2, and causes the communication device 7 to execute a process of transmitting the vehicle state signal to a large-scale language model outside the vehicle 1. The answer generation unit 133 also acquires location information indicating the location of the vehicle 1 identified by the GPS unit 3 and the map information unit 4, and causes the communication device 7 to execute a process of transmitting the location information to a large-scale language model outside the vehicle 1. The answer generation unit 133 also acquires data of a user manual stored in the memory 12, and causes the communication device 7 to execute a process of transmitting the data to a large-scale language model outside the vehicle 1. Furthermore, the answer generation unit 133 causes the communication device 7 to execute a process of receiving a processing result of a large-scale language model outside the vehicle 1, and uses the processing result to generate a voice answer corresponding to the voice question from the driver of the vehicle 1. In detail, the answer generation unit 133 generates text information of an answer corresponding to the voice question from the driver of the vehicle 1, and then generates the voice answer by vocalizing the text information.

[0029] In one example of processing executed in a vehicle 1 to which the driving assistance device 10 of the second embodiment is applied, when the HMI 6 receives a voice question from the driver of the vehicle 1 such as "What operations are required for an OTA (Over-The-Air) update (updating in-vehicle software, etc. by the communication device 7 sending and receiving data via wireless communication with the outside of the vehicle 1)?", the answer generation unit 133 generates an audio answer explaining the operations, etc. required for the OTA update by the driver of the vehicle 1, based on the question from the driver of the vehicle 1 and the explanation of the operations required for the OTA update contained in the user manual.

[0030] As described above, the embodiments of the driving assistance device, the driving assistance method, and the program of the present disclosure have been described with reference to the drawings, but the driving assistance device, the driving assistance method, and the program of the present disclosure are not limited to the above-described embodiments, and may be modified as appropriate without departing from the spirit of the present disclosure. The configurations of the examples of the above-described embodiments may be combined as appropriate. The program stored in the memory 12 of the driving assistance device 10 (the program that realizes the functions of the processor 13 of the driving assistance device 10) may be recorded on a computer-readable storage medium such as a semiconductor memory, a magnetic recording medium, an optical recording medium (non-temporary storage medium), etc., and provided, distributed, etc. [Explanation of symbols]

[0031] 1...vehicle, 2...vehicle condition sensor, 3...GPS unit, 4...map information unit, 5...driver monitor camera, 6...HMI, 7...communication device, 10...driving assistance device, 11...communication interface, 12...memory, 13...processor, 131...acquisition unit, 132...voice recognition unit, 133...answer generation unit, 134...update unit

Claims

1. An acquisition unit that acquires a voice question related to an Advanced Driver Assistance System (ADAS) received from a driver who is driving a vehicle; a voice recognition unit that recognizes a voice question regarding the ADAS acquired by the acquisition unit; A driving assistance device comprising: a response generation unit that uses large language models (LLMs) to generate voice responses corresponding to voice questions regarding the ADAS recognized by the voice recognition unit.

2. the acquisition unit acquires a voice question received from a driver when a vehicle speed of the vehicle is greater than zero, the voice question inquiring as to whether or not a function of the ADAS that cannot be activated when the vehicle speed is less than a threshold value is in an operable state, information indicating a state of the vehicle at the time when the voice question inquiring as to whether or not a function of the ADAS is in an operable state is received from the driver, position information of the vehicle at the time when the voice question inquiring as to whether or not a function of the ADAS is in an operable state is received from the driver, and information describing conditions for activating the function of the ADAS; The voice recognition unit recognizes a voice question asking whether or not the function of the ADAS acquired by the acquisition unit is in an operable state, 2. The driving assistance device according to claim 1, wherein the answer generation unit generates, when a vehicle speed of the vehicle is greater than zero, a voice answer corresponding to the voice question asking whether or not the ADAS function is in a state in which it is possible to operate the ADAS function, and a voice answer explaining an operation by the driver that is necessary to operate the ADAS function if the ADAS function is in a state in which it is possible to operate, based on the voice question acquired by the acquisition unit, information indicating a state of the vehicle, position information of the vehicle, and information describing conditions for activating the ADAS function.

3. The ADAS functions include lane change assist, 3. The driving assistance device according to claim 1, wherein when the acquisition unit acquires a voice question regarding a range of vehicle speeds at which the lane change assist can be operated, the answer generation unit generates a voice answer explaining the range of vehicle speeds at which the lane change assist can be operated, the voice question being received from a driver who is driving the vehicle.

4. An acquisition step in which the driving assistance device acquires a voice question regarding the ADAS received from a driver who is driving the vehicle; a voice recognition step in which the driving assistance device recognizes a voice question regarding the ADAS acquired in the acquisition step; A driving assistance method comprising: a response generation step in which the driving assistance device uses a large-scale language model to generate a voice response corresponding to the voice question regarding the ADAS recognized in the voice recognition step.

5. The processor: An acquisition step of acquiring a voice question regarding the ADAS received from a driver who is driving the vehicle; a voice recognition step for recognizing a voice question regarding the ADAS acquired in the acquisition step; and a response generation step of generating a voice response corresponding to the voice question regarding the ADAS recognized in the voice recognition step using a large-scale language model.

Images