Agent control apparatus

Intelligent body control device technology field

[0001] The present disclosure relates to an agent control device that performs processing related to an agent in response to a user's request. Background Art

[0002] In Japanese Patent Application Publication No. 2018-189984, a voice dialogue intelligent agent is disclosed, which accepts an input voice signal, performs voice recognition processing on the input voice signal, and based on the results and intelligence of the voice recognition processing The agent information determines which agent to use to implement processing based on the input speech signal.

[0003] When an agent performs a function requiring continuity such as a call, it is desirable to suppress the operation of other agents. On the other hand, when other agents have functions that are more important than functions that require continuity such as making calls, such as a function to assist rescue activities in the event of an emergency, it is desirable to use the functions of other agents. Important functions are performed.

Contents of the invention

[0004] The purpose of the present disclosure is to provide an intelligent agent control device that is capable of controlling a situation in which an intelligent agent is normally executing a function that requires continuity when other intelligent agents are activated, causing the function to be interrupted. Suppression, and the ability to prioritize the execution of important functions of other agents under specific conditions.

[0005] An intelligent agent control device according to a first aspect is an intelligent agent control device capable of operating a plurality of intelligent agents, and is provided with a storage unit that stores an interrupt availability list that defines the Whether the functions of one of the running intelligent agents or the operation of the operating status of the one intelligent agent can be interrupted; the first requesting part, which requests the operation of each of the intelligent agents at a predetermined opportunity; the second request The setting part, which requests the operation of other agents at a specific opportunity; the setting part, which accompanies the operation of the one agent and refers to the interruption possibility list to determine whether the operation of the other agents is possible or not. Whether information can be set; a management unit that manages in the following manner, that is, when there is information about the other from the first requesting unit during the operation of the one agent at the predetermined opportunity. If there is a request for an agent to run and the availability information indicates that the other agent is inoperable, the request will not be responded to and the one agent will be allowed to continue. When there is a request for the operation of the other agent from the second requesting unit during the operation of one agent, the other agent is executed based on the request.

[0006] The intelligent agent control device according to the first aspect is capable of operating a plurality of intelligent agents. In this intelligent agent control device, the storage unit stores an interrupt availability list that specifies whether or not the operation of each function of an operating intelligent agent or the operation status of an intelligent agent can be interrupted. Furthermore, in this intelligent agent control device, one intelligent agent is requested through the first requesting unit, and the other intelligent agents are requested through the first requesting unit and the second requesting unit. Here, the "predetermined opportunity" for the first requesting unit to issue the request refers to, for example, the user speaking into the microphone or pressing the PTT (Push-To-Talk: push-to-talk) switch. In addition, the "specific opportunity" for the second requesting unit to issue the request refers to, for example, the user's utterance of a specific sentence or pressing of a specific switch with respect to the microphone. When the first agent is running, the setting unit refers to the interrupt permission list to set permission information related to the execution permission of other agents.

[0007] Furthermore, the management unit performs management in such a manner that, although there is a request from the first requesting unit regarding the operation of other agents during the operation of one agent, the availability information indicates that the other agents are not operable. Next, the agent does not respond to the request and allows an agent to continue. Accordingly, in this intelligent agent control device, it is possible to suppress a situation in which one intelligent agent is normally executing a function for which continuity is required and another agent is activated, causing the function to be interrupted.

[0008] On the other hand, the management unit performs management in such a manner that when a request regarding the operation of another agent is received from the second request unit during the operation of one agent, the management unit performs management based on the request. Other agents run. Therefore, in this intelligent agent control device, important functions of other intelligent agents can be executed with priority under specific conditions.

[0009] A second aspect of the intelligent agent control device is the intelligent agent control device of the first aspect, wherein the interrupt availability list stipulates, with respect to a function of the one intelligent agent that requires continuity, In order to be uninterruptible, the other agent has a function provided in an emergency, and the specific opportunity is the utterance based on a statement associated with the emergency situation.

[0010] In the intelligent agent control device described in the second aspect, although the function of an intelligent agent that requires continuity is usually not interrupted during the operation of the function, when receiving it, it will be associated with an emergency situation. When an emergency function of another agent is requested through a statement, the emergency function will be executed. Therefore, according to this intelligent agent control device, functions of assisting rescue activities in other intelligent agents can be executed with priority in an emergency.

[0011] A third aspect of the intelligent agent control device is the intelligent agent control device according to the first aspect or the second aspect, wherein the management unit performs management in the following manner: when one of the intelligent agents During operation, if there is a request for the operation of the other agent from the first requesting unit and the availability information indicates that the other agent is not operable, the request is retained and the one The agent continues, after the execution of the continued one agent ends, causing the other agent to run in response to the reserved request.

[0012] The agent control device of the third aspect does not discard the operation requests of other agents but retains them even when other agents are unable to operate during the operation of one agent. Therefore, according to this intelligent agent control device, after the termination of one intelligent agent, other intelligent agents can be operated without requesting the execution of other intelligent agents again.

[0013] According to the present disclosure, it is possible to suppress the situation in which an intelligent agent is normally executing a function requiring continuity when other agents are started, causing the function to be interrupted, and can enable it under specific conditions. Important functions of other agents are executed first.

Description of the drawings

[0014] Exemplary embodiments of the present disclosure will be described in detail based on the following drawings, in which:

[0015] FIG. 1 is a diagram showing the schematic structure of the intelligent agent system according to the first embodiment.

[0016] FIG. 2 is a block diagram showing the hardware configuration of the vehicle according to the first embodiment.

[0017] FIG. 3 is a diagram showing the contents of the interrupt availability list according to the first embodiment.

[0018] FIG. 4 is a block diagram showing the functional structure of the host computer according to the first embodiment.

[0019] FIG. 5 is a sequence diagram showing the flow of processing in the intelligent agent system according to the first embodiment.

FIG. 6 is a sequence diagram showing the flow of processing in the agent system according to the first embodiment, and is a continuation of FIG. 5 .

7 is a sequence diagram showing the flow of processing in the agent system according to the second embodiment.

Detailed ways

[0022][First Embodiment]

[0023] As shown in FIGS. 1 and 2 , the intelligent system 10 according to the first embodiment is configured to include a vehicle 12 , a company server 14 , and another company server 16 . The vehicle 12 is equipped with a head unit 20 as an intelligent control device, a communication device 22 , a touch panel 24 , a speaker 26 and a microphone 28 . The communication device 22 of the vehicle 12, the own company's server 14, and the other company's server 16 are connected to each other via the network N.

[0024] The company server 14 is a server owned by the company as the manufacturer of the vehicle 12. The company's server 14 has a function of inferring the intention of the user U's utterance in order to realize the function of the company's agent as the company's service. Our company's intelligent agent has functions such as making calls, obtaining weather/news, obtaining vehicle information, retrieving destinations, and requesting rescue. Here, requesting rescue is a function provided to the user U in an emergency. For example, the operator can deploy an emergency vehicle when the driver U becomes unwell or when the vehicle 12 has an accident. Our company agent is an example of other agents.

[0025] The other company server 16 is a server owned by another company that is not the manufacturer of the vehicle 12 . The other company server 16 has a function of inferring the intention of the user U's utterance in order to realize the function of the other company's agent as a service of the other company. In other company's intelligent agents, it has functions such as calling, settlement, obtaining weather/news, providing music, audio books, etc. The Other Company Agent is an example of an agent. In addition, the same function can also be realized by the company's agents and other companies' agents.

[0026] (vehicle)

As shown in Figure 2, the host 20 is configured to include a CPU (Central Processing Unit, central processing unit) 20A, a ROM (Read Only Memory, a read-only memory) 20B, and a RAM (Random Access Memory, a random access memory). ) 20C, memory 20D, communication I/F (Interface, connected to D) 20E, and input/output I/F 20F. CPU20A, ROM20B, RAM20C, memory 20D, communication I/F20E, and input/output I/F20F are mutually communicably connected via internal bus 20G. CPU 20A is an example of a processor, and RAM 20C is an example of a memory.

[0028] The CPU 20A is a central processing unit that executes various programs or controls various parts. That is, CPU20A starts from

The program is read from ROM 20B or storage 20D, and RAM 20C is used as a workspace to execute the program.

[0029] The ROM 20B stores various programs and various data. In the ROM 20B of this embodiment, a control program for controlling the host computer 20 is stored.

[0030] RAM 20C serves as a work area to temporarily store programs or data.

[0031] The memory 20D as a storage unit is composed of a HDD (Hard Disk Drive: Hard Disk Drive) or an SSD (Solid State Drive: Solid State Drive), and stores various programs and various data. The memory 20D of this embodiment stores the processing program 100 and the interrupt availability list 110 .

[0032] The processing program 100 is a program for controlling the company's own agent and other company's agents.

[0033] As shown in FIG. 3, the interrupt availability list 110 is a list that specifies the interrupt availability of each function in the execution of the other company's agent. In the interrupt availability list 110, if it is a function of another company's agent that requires continuity, it is stipulated that even if the own company's agent requests it, the function being executed cannot be interrupted. For example, when the functions of another company's agent to be executed are "call" and "settlement", it is specified as "impossible" to disable the interruption of the function.

[0034] In addition, for example, the functions of other company agents being executed are "get weather/news" and "music"

In this case, "Enable" is specified to enable the interruption of this function.

[0035] As shown in FIG. 2, the communication I/F 20E is an interface for connecting to the communication device 22. This interface uses, for example, a communication standard based on the CAN protocol. The communication I/F 20E is connected to the external bus 20H.

[0036] The input/output I/F 20F is an interface for communicating with the touch panel 24, the speaker 26, and the microphone 28 mounted on the vehicle 12. In addition, the touch panel 24, the speaker 26, and the microphone 28 may be directly connected to the internal bus 20G. [0037] The communication device 22 is a wireless communication module used to communicate with the own company's server 14 and other company's servers 16. This wireless communication module uses communication standards such as 5G, LTE, and Wi-Fi (registered trademark). The communication device 22 is connected to the network N.

[0038] The touch panel 24 is provided on the instrument panel and serves as a liquid crystal display as a display unit and as a touch panel as an input unit.

[0039] The speaker 26 is provided on the instrument panel, center console, front pillar, dash panel, etc., and is a device for outputting speech related to the company's intelligent agent and other companies' agents.

[0040] The microphone 28 is installed on the front pillar, dashboard, etc. of the vehicle 12, and is a device that collects the voice uttered by the user U who is the occupant of the vehicle 12.

[0041] As shown in FIG. 4, in the host computer 20 of this embodiment, the CPU 20A functions as the first requesting unit 200, the second requesting unit 210, the first execution unit 220, and the second execution unit 230 by executing the processing program 100. , the setting unit 240 and the management unit 250 to function.

[0042] The first requesting unit 200 has a function of requesting the operation of the agent at a predetermined opportunity. Here, the "predetermined opportunity" includes the user U speaking into the microphone 28, pressing a PTT (Push-ToTalk) switch (not shown) provided on the steering wheel or the like, and the like. When the voice information based on the utterance of the user U includes a normal wake-up word for the company's agent, the first requesting unit 200 sends the voice information to the company's server 14 to request the operation of the company's agent. In addition, when the voice information based on the utterance of the user U includes a wake word for the other company's agent, the first requesting unit 200 sends the voice information to the other company's server 16 to request the operation of the other company's agent.

[0043] The second requesting unit 210 has a function of requesting the operation of other agents at a specific opportunity. Here, the "specific opportunity" includes the user U uttering a specific sentence or pressing a specific switch with respect to the microphone 28 . The specific sentence is a sentence associated with an emergency situation, such as a word including "Help", "emergency situation", "request for rescue", etc. In addition, the specific switch is a switch dedicated to requesting rescue, which is provided separately from the PTT switch. The second request unit 210 is executed when the occupation described below is marked as on.

[0044] The first execution unit 220 has a function of causing the company's intelligent agent to operate. The first execution unit 220 executes the function of the own agent based on the intention information provided from the own server 14 that transmitted the voice information based on the utterance.

[0045] The second execution unit 230 has a function of causing other company agents to operate. The second execution unit 230 executes the function of the other company's agent based on the intention information provided from the other company's server 16 that transmitted the utterance-based voice information.

[0046] The setting unit 240 has a function of referring to the interrupt availability list 110 to set the availability information related to the execution availability of each agent. The permission information in this embodiment is an occupancy flag that can be set to ON or OFF, and the setting unit 240 determines the permission information based on each of the other company agents specified in the interruption permission list 110. Whether each function can be interrupted can set the occupation flag. Specifically, when a certain function of another company's agent is uninterruptible, the setting unit 240 sets the occupancy flag when the other company's agent starts running the function.

is on and sets the possession flag to off at the end of the function.

[0047] The management unit 250 has a function of managing the operation of each intelligent agent. The management unit 250 of this embodiment manages in such a manner that when there is a request for the operation of the own company's agent from the first requesting unit 200 during the operation of another company's agent and the possession flag is turned on, In this case, it does not respond to the request for the operation of its own agent and allows other companies' agents to continue. Furthermore, when there is a request for the operation of the company's agent from the second requesting unit 210 during the operation of another company's agent, the management unit 250 causes the company's agent to operate based on the request.

[0048] (Flow of control)

[0049] The flow of processing executed in the intelligent agent system 10 of this embodiment will be described using the sequence diagrams of FIGS. 5 and 6 . The processing in the host computer 20 is realized by the CPU 20A functioning as the first requesting unit 200, the second requesting unit 210, the first execution unit 220, the second execution unit 230, the setting unit 240, and the management unit 250 described above.

[0050] In step S10 of FIG. 5, the user U speaks toward the host 20. For example, user U utters "Agent, I want to go to X", "Assistant, call Ya", etc. to make a request to the agent.

[0051] In step S11, the host 20 performs speech recognition of the speech generated by the user U's utterance.

[0052] In step S12, the host computer 20 detects which of the own company's agent and the other company's agent is to be activated based on the voice information generated by speech recognition. For example, when the wake word "Agent" is included in the uttered voice, the host 20 detects the company's agent as the agent to be activated. In addition, when the wake word "Assistant" is included in the uttered voice , the host 20 detects other company's agents as the agents to be started.

[0053] In step S13, the host 20 determines whether the occupation flag is closed. When the host 20 determines that the occupation flag is not off but is on, it proceeds to step S14. On the other hand, if the host 20 determines that the occupation flag is off, the process proceeds to step S15.

[0054] In step S14, the host 20 discards the request for the agent. Then, the execution process of the agent ends. [0055] In step S15, the host 20 sends voice information to the server of the agent detected in step S12. Figure 5 is an example of issuing a request related to agents of other companies. When user U utters "Assistant, call ύ", the voice message with content such as "Call " will be sent to other companies. Server 16.

[0056] In step S16, the other company server 16 performs voice recognition based on the received voice information and performs intention extraction. For example, the intention to call Ya is extracted from the voice message "Call Ya".

[0057] In step S17, the other company server 16 sends the extracted intention to the host 20 as intention information.

[0058] In step S18, the host 20 starts the operation of other company's intelligent agents. That is, when the host 20 receives the intention information to call y from the other company server 16, it activates the call function and refers to y's contact information obtained from the other company server 16 or stored in the storage 20D. Come and call y. As a result, the call via the host 20 is started.

[0059] In step S19, the host 20 refers to the interrupt availability list 110. For example, as shown in FIG. 3 , when another company's agent is making a call, "impossible" information is obtained that disables the interruption of the other company's agent.

[0060] In step S20 of FIG. 5, the host 20 determines whether the other company's intelligent agent is uninterruptible. When the host computer 20 determines that other company agents cannot be interrupted, step S21 is entered. On the other hand, when the host computer 20 determines that the other company's agent is not uninterruptible, the process proceeds to step S22 in FIG. 6 .

[0061] In step S21, the host 20 sets the occupation flag to on.

[0062] The processing from step S22 to step S24 in FIG. 6 is the same as the processing from step S10 to step S12 described above.

[0063] Next, if the occupation flag is on, in step S25, the host 20 determines whether it is an emergency state. Specifically, when the speech recognition result in step S23 includes a sentence that associates an emergency situation, the host computer 20 determines that it is an emergency state. When the host computer 20 determines that the emergency state is not present, the process proceeds to step S26. On the other hand, if the host computer 20 determines that it is an emergency state, the process proceeds to step S30.

[0064] The processing of steps S26 and step S27 is the same as the processing of steps S13 and S14 described above.

[0065] In step S28, the host 20 determines whether other company agents have ended. If the host computer 20 determines that the other company's agent has not ended, that is, if the other company's agent is still running, the process returns to step S22. On the other hand, if the host computer 20 determines that the other company's agent has been terminated, the process proceeds to step S29.

[0066] In step S29, the host 20 sets the occupation flag to off. Then, the execution process of the agent ends.

[0067] When it is determined in step S26 that the occupation flag is off, in step S30, the host 20 sends voice information to the server of the agent detected in step S24. Figure 6 is an example of issuing a request related to our company's agent. When user U pronounces "Agent, I want to go to X", the voice message with the content of "I want to go to X" will be sent to our company. Server 14. In addition, when the user U utters "request for rescue" as a sentence associated with an emergency situation, the voice information of the utterance will be sent to the company's server 14.

[0068] In step S31, the company's server 14 performs speech recognition based on the received speech information and performs intention extraction. For example, the intention to retrieve the destination X is extracted from the voice message "I want to go to X". In addition, for example, the intention of talking to an operator who assists in requesting rescue is extracted from the voice message of "requesting rescue".

[0069] In step S32, the company's server 14 sends the extracted intention to the host 20 as intention information.

[0070] In step S33, the host 20 starts the operation of the company's intelligent agent. In the event of an emergency, other running company agents are interrupted and terminated, allowing a call with the operator to take place.

[0071] (Summary of the first embodiment)

[0072] The host computer 20 of this embodiment is configured to be able to operate the own company's agent and other company's agents as a plurality of agents. In the host computer 20 , an interrupt availability list 110 stipulating the interrupt availability for each function of the other company's agent being run is stored in the memory 20D. In the present embodiment, the first requesting unit 200 requests the operation of the own company's agent or another company's agent at a predetermined opportunity caused by utterance, operation of the PTT switch, or the like. In addition, the second requesting unit 210 requests the operation of the company's agent at a specific opportunity based on the utterance of a specific sentence or the operation of a specific switch. In the present embodiment, when another company's agent is running, the setting unit 240 refers to the interruption permission list 110 and sets the permission information regarding the permission or permission of the own company's agent's operation as an occupation flag.

[0073] Furthermore, the management unit 250 manages the situation in which the possession flag is turned on even though there is a request for the operation of the own company's agent from the first requesting unit 200 during the operation of the other company's agent. , discard the request and let other company agents continue. That is, according to this embodiment, it is possible to deal with the situation where the own company's agent is activated while another company's agent is normally executing a function that requires continuity such as a call or settlement, causing the function that requires continuity to be interrupted. To suppress.

[0074] On the other hand, the management unit 250 manages in such a manner that when there is a request regarding the operation of the own company's agent from the second requesting unit 210 during the operation of another company's agent, based on This request causes the company's agent to run. That is, according to this embodiment, the company's intelligent agent can be operated with priority under specific conditions.

[0075] In addition, in this embodiment, the important function of assisting rescue activities is set as a specific condition.

In this case, the second requesting unit 210 receives a sentence that is associated with an emergency situation and requests the emergency function of the company's agent. Therefore, according to this embodiment, it is possible to prioritize the execution of functions and the like in the company's intelligent agent that assist in rescue activities during an emergency.

[0076] In addition, according to the present embodiment, by setting the occupancy flag as the availability information, agents with different functions can be collectively managed by turning on and off the occupancy flag.

[0077][Second Embodiment]

[0078] The first embodiment adopts the following method, that is, when the occupation flag is turned on, the execution request of the new agent is discarded. On the other hand, in the second embodiment, when the occupation flag is turned on, the execution request of the new agent is reserved. Hereinafter, differences from the first embodiment will be described. In addition, other structures are the same as those of the first embodiment, and detailed description thereof is omitted.

[0079] The management unit 250 of this embodiment is configured to, when there is a request for the operation of the own company's agent from the first requesting unit 200 during the operation of another company's agent and the occupation flag is turned on, Requests regarding the operation of our own company's agents are retained and other company's agents are allowed to continue. Then, the management unit 250 causes the own company's agent whose request is in the reserved state to run after the other company's agent that continues to be run ends.

[0080] The difference between the processing executed in the intelligent agent system 10 of this embodiment and the first embodiment is as shown in FIG. 7 . That is, instead of the processing from step S27 to step S29 in FIG. 6 , the processing from step S40 to step S42 is performed.

[0081] When it is determined in step S26 that the occupation flag is not closed, in step S40, the host 20 reserves the request for the agent.

[0082] In step S41, the host 20 determines whether other company agents have ended. If the host computer 20 determines that the other company's agent has not ended, that is, if the other company's agent is still running, the process returns to step S22. On the other hand, if the host computer 20 determines that the other company's agent has been terminated, the process proceeds to step S42.

[0083] In step S42, the host 20 sets the occupation flag to off. Then, proceed to step S30.

[0084] In addition, the request for the agent that was retained in step S40 is reflected in steps S30 to step S33. As a result, in step S33, the operation of the reserved own company agent is started.

[0085] (Summary of the second embodiment)

[0086] The host 20 of the second embodiment is configured not to discard the execution request of the own company's agent but to respond to it even if the own company's agent cannot operate during the operation of another company's agent. Make a reservation. Therefore, according to the present embodiment, after the other company's agent is terminated, the own company's agent can be run without the need for the first requesting unit 200 to execute the request of the own company's agent again.

[0087][Note]

[0088] Each of the above embodiments is a method in which, for each function executed by another company's agent, whether the function can be interrupted or not is judged. However, the present invention is not limited to this, and whether the function during operation can be interrupted can also be judged based on the operation status of other company's agents. The interruption permission list 110 in this case specifies the interruption permission for other company agents at a level corresponding to the operating status of the other company agents.

[0089] For example, if "shopping" is included in the function of another company's agent, it is specified that the interruption of processing is not allowed during the settlement process, level A, or during the purchase procedure. In order to set the operation of our company's agent to level B, it is specified that the interruption of processing during the item search process is set to level C, which is allowed. As a result, when other companies' agents are performing shopping, it is possible to evaluate other companies' agents based on their operating status.

The interruption of the agent and the operation of the company's agent are controlled.

[0090] In addition, in each of the above-described embodiments, the operation of the own company's agent is usually controlled based on the operation availability of other company's agents stored in the interruption permission list 110, but it is not limited to this. That is, the operation availability of the own company's agent may be stored in the interruption permission list 110, and the operation of other companies' agents may be controlled based on the operation availability of the own company's agent.

[0091] The host 20 of the second embodiment is configured to hold the request of the company's agent and transmit the held request as voice information to the company's server 14 when the other company's agent ends. However, it is not limited to this. When the host 20 discards the request of the own company's agent, the host 20 may ask the user U again for the request at the stage when the other company's agent has finished. In other words, the request may be made again. speak out.

[0092] In addition, various processes that the CPU 20A reads and executes software (programs) in the above-described embodiment may be executed by various processors other than the CPU. Examples of processors in this case include FPGA (Field Programmable Gate Array) and other PLDs (Programmable Logic Devices) whose circuit structure can be changed after manufacturing, and ASICs (Application Specific Integrated Circuit (Application Special Integrated Circuit) and other processors that have a specially designed circuit structure in order to perform specific processing, that is, dedicated circuits, etc. In addition, the above-mentioned acceptance processing may be executed by one of these various processors, or a combination of two or more processors of the same type or different types (for example, a plurality of FPGAs, and a combination of a CPU and an FPGA). combination, etc.) to execute. Furthermore, more specifically, the hardware structure of these various processors is a circuit in which circuit elements such as semiconductor elements are combined.

[0093] Furthermore, in the above-described embodiment, the description is made in such a manner that each program is stored (installed) in advance in a computer-readable non-transitory recording medium. For example, the processing program 100 in the host 20 is stored in the storage 20D in advance. However, it is not limited to this. Each program can also be recorded on CD-ROM (Compact Disc Read Only Memory), DVD-ROM (Digital Versatile Disc Read Only Memory), and USB (Universal Serial Bus: Universal Serial Bus) memory and other non-transitory recording media are provided. In addition, the program may be downloaded from an external device via a network.

[0094] The processing flow described in the above-mentioned embodiment is also an example, and unnecessary steps may be deleted, new steps may be added, or the processing order may be replaced within the scope that does not deviate from the gist.

CN 113682320 Β