Method for operating a digital assistant of a vehicle, computer-readable medium, system, and vehicle
Patent Information
- Authority / Receiving Office
- DE · DE
- Patent Type
- Patents
- Current Assignee / Owner
- BAYERISCHE MOTOREN WERKE AG
- Filing Date
- 2023-06-14
- Publication Date
- 2026-08-06
AI Technical Summary
Existing digital assistants for vehicles lack efficient integration of rule-based and artificial intelligence-based dialogue systems, leading to inadequate or inappropriate responses to user inputs, particularly when intent recognition is uncertain.
A method combining a rule-based dialogue system with an artificial intelligence-based system, where the rule-based system provides context information to the AI system, enabling context-based responses and proactive user guidance, and ensuring accurate responses by determining intent recognition confidence thresholds.
Enhances the efficiency and accuracy of digital assistant responses by providing precise and context-aware messages, preventing inappropriate responses, and ensuring safety-relevant information is correctly communicated.
Smart Images

Figure 00000000_0000_ABST
Abstract
Description
[0001] The invention relates to a method for operating a digital assistant in a vehicle. The invention further relates to a computer-readable medium for operating a digital assistant in a vehicle, a system for operating a digital assistant in a vehicle, and a vehicle comprising the system for operating a digital assistant in a vehicle.
[0002] Digital assistants in vehicles that enable rule-based dialogues regarding vehicle functions are known from the prior art. Digital assistants based on artificial intelligence are also known. However, a combination of a rule-based digital assistant and an AI-based digital assistant in a vehicle is not known from the prior art.
[0003] It is therefore an object of the invention to operate a digital assistant more efficiently. In particular, an object of the invention is to provide improved response messages regarding a function of a vehicle to a user of the vehicle via a digital assistant of the vehicle.
[0004] This problem is solved by the features of the independent claims. Advantageous embodiments and further developments of the invention result from the dependent claims.
[0005] According to a first aspect, the invention is characterized by a method for operating a vehicle's digital assistant, wherein the digital assistant comprises a rule-based dialogue system and an artificial intelligence-based dialogue system. The method can be a computer-implemented method and / or an electronic control unit-implemented method. The artificial intelligence-based dialogue system can, for example, be a chatbot that uses artificial intelligence to generate text- and / or voice-based replies to messages from a user. The digital assistant can be a digital voice assistant. The rule-based dialogue system preferably comprises a predefined dialogue sequence determined by one or more rules.The rule-based dialogue system can directly process status data, sensor data, and / or sensor data related to a vehicle function, for example, through a rule and / or a data processing module of the rule-based dialogue system. The AI-based dialogue system is preferably text- and / or speech-based. For example, the AI-based dialogue system cannot directly process function-specific status and / or sensor data; that is, converting the function-specific status and / or sensor data into a text format understandable to the AI-based dialogue system is necessary.
[0006] The process comprises activating a dialog-based user guidance system for a vehicle function via the rule-based dialogue system of the digital assistant. The process further comprises providing an initial message of the dialog-based user guidance to a vehicle user via the rule-based dialogue system of the digital assistant after activation of the dialog-based user guidance, and transmitting this initial message of the dialog-based user guidance from the rule-based dialogue system of the digital assistant to the AI-based dialogue system of the digital assistant as contextual information.Furthermore, the method comprises receiving a user message during the dialog-based user guidance of the rule-based dialogue system by the rule-based dialogue system of the digital assistant, and transmitting the user message from the rule-based dialogue system of the digital assistant to the AI-based dialogue system of the digital assistant. The method recognizes the intent of the dialog-based user guidance of the rule-based dialogue system regarding the vehicle's function, depending on the user message received by the rule-based dialogue system of the digital assistant.If the rule-based dialogue system of the digital assistant does not detect any intention of the user to interact with the vehicle's function, the procedure provides a context-based response message to the vehicle user via the AI-based dialogue system, using the transmitted context information and depending on the user's message. If the rule-based dialogue system of the digital assistant detects an intention of the user to interact with the vehicle's function, the procedure provides a response message to the vehicle user depending on the detected intention of the user to interact with the vehicle's function.
[0007] Advantageously, this method can operate a digital assistant, such as a digital voice assistant, more efficiently. By combining the rule-based dialogue system with the AI-based dialogue system, the vehicle user can receive a context-based response message from the AI-based system, with the rule-based system providing the context for the AI-based system's response. This allows the user to receive a more precise response from the AI-based system to a message whose intent is not recognized by the rule-based system, or not recognized with a confidence level above a predefined threshold.Furthermore, the procedure efficiently prevents an inappropriate and / or incorrect response message from the AI-based dialogue system by defining the context of the AI-based dialogue system through the rule-based dialogue system.
[0008] According to an advantageous embodiment, activating the dialog-based user guidance for the vehicle's function by the rule-based dialogue system of the digital assistant can involve: providing a recommendation message for a vehicle function to a vehicle user by the rule-based dialogue system of the digital assistant; transmitting the recommendation message to the AI-based dialogue system as context information; receiving a confirmation message from the vehicle user in response to the provided recommendation message by the rule-based dialogue system of the digital assistant to activate the dialog-based user guidance for the vehicle's function; and activating the dialog-based user guidance of the vehicle's function by the rule-based dialogue system of the digital assistant.and include transmitting the confirmation message to the AI-based dialogue system as contextual information. This allows for efficient, proactive initiation of dialogue-based user guidance.
[0009] According to a further advantageous embodiment, the provision of an initial message of the dialog-based user guidance by the rule-based dialogue system of the digital assistant to a user of the vehicle after activation of the dialog-based user guidance can include receiving a system state of the vehicle's function by the rule-based dialogue system of the digital assistant, transmitting a text message representative of the system state of the vehicle's function to the AI-based dialogue system by the rule-based dialogue system of the digital assistant as context information, and providing the initial message of the dialog-based user guidance by the rule-based dialogue system of the digital assistant to a user of the vehicle depending on the system state of the vehicle's function.This allows the system state of the vehicle's function to be efficiently provided as context information to the AI-based dialogue system.
[0010] According to a further, advantageous embodiment, the recognition of the intention of the rule-based dialog system's dialog-based user guidance, depending on the user's received message, by the rule-based dialog system of the digital assistant can include determining a confidence value for the intention of the dialog-based user guidance, depending on the user's received message. This allows the rule-based dialog system to determine the intention more precisely.
[0011] According to a further advantageous embodiment, the method can also, if no intention of the rule-based dialogue system regarding the vehicle's function is detected by the rule-based dialogue system of the digital assistant, or if the confidence value of the detected intention of the rule-based dialogue system regarding the vehicle's function is less than a predefined confidence threshold, provide a context-based response message to the vehicle's user by the AI-based dialogue system of the digital assistant using the transmitted context information and depending on the user's transmitted message.and if the confidence value of the recognized intention of the rule-based dialog system's dialog-based user guidance regarding the vehicle's function by the rule-based dialog system of the digital assistant is greater than or equal to the predefined confidence threshold, the system should provide a response message to the vehicle's user depending on the recognized intention of the rule-based dialog system's dialog-based user guidance regarding the vehicle's function. This allows for more efficient control of the provision of the context-based response message.
[0012] According to a further advantageous embodiment, the context-based response message to the vehicle user can be provided by the AI-based dialogue system of the digital assistant using all contextual information transmitted to the AI-based dialogue system by the rule-based dialogue system after activation of the dialogue-based user guidance. This allows a dialogue-based context to be efficiently established for the AI-based dialogue system.
[0013] According to a further advantageous embodiment, the context-based response message to the vehicle user can be provided by the AI-based dialogue system of the digital assistant using at least one or all of the contextual information transmitted to the AI-based dialogue system by the rule-based dialogue system after activation of the dialogue-based user guidance, and depending on the intention and / or the confidence value of the intention of the rule-based dialogue system of the digital assistant. This allows a dialogue-based context to be efficiently established for the AI-based dialogue system.
[0014] According to another aspect, the invention is characterized by a computer-readable medium for operating a digital assistant of a vehicle, wherein the computer-readable medium comprises instructions which, when executed on a computer and / or a control unit, execute the method described above.
[0015] According to another aspect, the invention is characterized by a system for operating a digital assistant of a vehicle, wherein the system is configured to perform the method described above.
[0016] According to another aspect, the invention is characterized by a vehicle comprising the system described above for operating a digital assistant of the vehicle.
[0017] Further features of the invention will become apparent from the claims, the figure, and the figure description. All features and combinations of features mentioned above in the description, as well as those subsequently mentioned in the figure description and / or shown in the figure alone, are usable not only in the combinations specified, but also in other combinations or individually.
[0018] The following describes a preferred embodiment of the invention with reference to the accompanying drawing. Further details, preferred configurations, and further developments of the invention will be derived from this. Specifically, the drawing schematically illustrates... Fig. 1 An exemplary procedure for operating a digital assistant in a vehicle.
[0019] In detail, Fig.1 An exemplary procedure 100 for operating a vehicle's digital assistant. The vehicle's digital assistant can include a rule-based dialogue system and an artificial intelligence-based dialogue system. For example, the rule-based dialogue system can include a deterministic language model. The deterministic language model can recognize pre-modeled intents and generate corresponding deterministic response messages and / or provide them to a vehicle user. For example, the rule-based dialogue system can include a predefined multi-stage dialogue about a vehicle function. The predefined multi-stage dialogue about a vehicle function can include technically correct user guidance about a vehicle function.
[0020] For example, the AI-based dialogue system could be a chatbot that uses artificial intelligence, such as machine learning, to generate and / or provide voice-based and / or text-based responses to a user. The AI-based dialogue system could, for example, be trained in several steps using a language model. The AI-based dialogue system could be trained with technically accurate data relating to a vehicle function as training data. For example, the AI-based dialogue system could be trained with a user manual and / or a technical specification for the vehicle, and in particular with a user manual and / or a technical specification relating to a specific vehicle function.
[0021] Procedure 100 comprises activating 102 a dialog-based user guidance system for a vehicle function via the rule-based dialogue system of the digital assistant. The dialog-based user guidance can include several dialogue steps. The dialog-based user guidance can explain the correct use of the vehicle function to the user step by step, using a predefined sequence of dialogue steps. For example, the vehicle function could be a parking assistance function. The dialog-based user guidance for the vehicle's parking assistance function can include several dialogue steps outlining the actions the user must perform to use the parking assistance function correctly. For example, a first dialogue step could instruct the user to drive at a predefined maximum speed with a predefined maximum distance to parked vehicles.In a second dialogue step, the vehicle can offer the user a parking space suitable for their vehicle and prompt the user for further actions based on the vehicle's parking offer. In a third dialogue step, the vehicle can request the user to remove their hands from the steering wheel. In a fourth dialogue step, the vehicle can request the user to release their foot from the brake pedal. In a fifth dialogue step, the vehicle can guide the user to park the vehicle, with the user able to control the vehicle's speed by pressing the brake pedal. In a sixth dialogue step, the vehicle can send a message to the user confirming that the parking maneuver was successful.
[0022] Method 100 comprises providing 104 an initial message of dialog-based user guidance from the rule-based dialogue system of the digital assistant to a vehicle user after activation of dialog-based user guidance. The initial message can be a proactive recommendation regarding a vehicle function. Furthermore, Method 100 can transmit the initial message of dialog-based user guidance from the rule-based dialogue system of the digital assistant to the AI-based dialogue system of the digital assistant as contextual information 106. The user can optionally send a message to the rule-based dialogue system to accept or reject the recommendation regarding the vehicle function. If the user accepts the recommendation regarding the vehicle function, the method can execute the dialog-based user guidance for that vehicle function.Each step of the dialog-based user interface can be transmitted as contextual information from the rule-based dialog system to the AI-based dialog system. In the example of the dialog-based user interface for the parking assistant, the rule-based dialog system can transmit messages to the vehicle user and / or sensor information from the respective dialog step to the AI-based dialog system as contextual information.
[0023] Method 100 can receive a user message during the dialog-based user guidance of the rule-based dialogue system from the rule-based dialogue system of the digital assistant 108 and transmit the user message from the rule-based dialogue system of the digital assistant to the AI-based dialogue system of the digital assistant 110. Furthermore, Method 100 can recognize an intention of the dialog-based user guidance of the rule-based dialogue system regarding the function of the vehicle depending on the user message received by the rule-based dialogue system of the digital assistant 112. In the example of the dialog-based user guidance of the parking assistant, in the third dialogue step, taking hands off the steering wheel can be a user message that Method 100 recognizes as an intention to continue the dialog-based user guidance of the parking assistant.Furthermore, the message could, for example, be a voice message from the user, in which the vehicle user wants to ask whether they are doing everything correctly or is unsure whether the operation is correct. If no intention is defined for this voice message in the rule-based dialogue system, procedure 100 cannot recognize any intention either.
[0024] If no intention of the rule-based dialogue system of the digital assistant to the function of the vehicle is detected by the rule-based dialogue system of the digital assistant, the procedure 100 can provide a context-based response message to the user of the vehicle by the artificial intelligence-based dialogue system of the digital assistant using the transmitted context information and depending on the transmitted message of the user 114.If an intention of the dialog-based user guidance of the rule-based dialogue system to the function of the vehicle by the rule-based dialogue system of the digital assistant has been detected, the procedure can provide a response message, in particular a predefined response message, to the user of the vehicle depending on the detected intention of the dialog-based user guidance of the rule-based dialogue system to the function of the vehicle by the rule-based dialogue system of the vehicle 116.
[0025] Advantageously, an AI-based dialogue system, for example, one that uses a large language model and one or more machine learning methods, can provide a response message to a user for a message for which no intent can be recognized by the rule-based dialogue system. The rule-based dialogue system can, for example, include a deterministic language model that enables intent recognition and communicative coupling to onboard network signals and / or messages, which the rule-based dialogue system can use to control the dialog-based user guidance.The rule-based dialog system's dialog-based user guidance can use the procedure as context for the AI-based dialog system to respond to user messages in a context-based manner, particularly when there is no intent or when the confidence level of a recognized intent falls below a predefined threshold. The rule-based dialog system leads to increased functional and / or usability reliability of the digital assistant, as it provides accurate and secure, safety-relevant information within the dialog-based user guidance. Furthermore, the digital assistant can freely interpret user messages within the context of a currently executed dialog-based user guidance session and its learned factual knowledge, thus efficiently providing the user with further information in context-based response messages. List of reference symbols 100 procedures 102 Enabling a dialog-based user interface 104 Providing an initial message 106 Transmitting the first message 108 Receiving a message 110 Transmitting the message 112 Recognizing an intention 114 Providing a context-based response message 116 Providing a reply message
Claims
[1] A method for operating a digital assistant of a vehicle, the digital assistant comprising a rule-based dialogue system and an artificial intelligence-based dialogue system, the method comprising: Activating a dialogue-based user guidance of the rule-based dialogue system to a function of the vehicle through the rule-based dialogue system of the digital assistant; Providing a first message of the dialog-based user guidance by the rule-based dialog system of the digital assistant to a user of the vehicle after activating the dialog-based user guidance; Transmitting the first message of the dialog-based user guidance from the rule-based dialog system of the digital assistant to the artificial intelligence-based dialog system of the digital assistant as context information; Receiving a message from the user during the dialog-based user guidance of the rule-based dialog system by the rule-based dialog system of the digital assistant; Transmitting the user’s message from the digital assistant’s rule-based dialogue system to the digital assistant’s artificial intelligence-based dialogue system; Recognition of an intention of the dialogue-based user guidance of the rule-based dialogue system regarding the function of the vehicle depending on the received message from the user by the rule-based dialogue system of the digital assistant; If no intention of the dialogue-based user guidance of the rule-based dialogue system regarding the function of the vehicle was recognized by the rule-based dialogue system of the digital assistant: Providing a context-based response message to the user of the vehicle through the artificial intelligence-based dialogue system of the digital assistant using the transmitted context information and depending on the transmitted message of the user; and If an intention of the dialogue-based user guidance of the rule-based dialogue system regarding the function of the vehicle was recognized by the rule-based dialogue system of the digital assistant: Providing a response message to the user of the vehicle depending on the recognized intention of the dialog-based user guidance of the rule-based dialog system to the function of the vehicle by the rule-based dialog system of the vehicle. [2] The method according to claim 1, wherein activating the dialog-based user guidance to the function of the vehicle by the rule-based dialog system of the digital assistant comprises: Providing a recommendation message regarding a function of a vehicle to a user of the vehicle through the rule-based dialogue system of the digital assistant; Transmitting the recommendation message to the artificial intelligence-based dialogue system as context information; Receiving a confirmation message from the user of the vehicle in response to the provided recommendation message by the rule-based dialogue system of the digital assistant to activate the dialogue-based user guidance to the function of the vehicle; Activating the dialogue-based user guidance of the vehicle’s function through the rule-based dialogue system of the digital assistant; and Transmitting the confirmation message to the artificial intelligence-based dialogue system as context information. [3] Method according to one of the preceding claims, wherein providing a first message of the dialog-based user guidance by the rule-based dialog system of the digital assistant to a user of the vehicle after activating the dialog-based user guidance comprises: Receiving a system status of the vehicle’s function through the digital assistant’s rule-based dialogue system; Transmitting a text message representative of the system status of the vehicle's function to the artificial intelligence-based dialogue system by the rule-based dialogue system of the digital assistant as context information; and Providing the first message of the dialog-based user guidance through the rule-based dialog system of the digital assistant to a user of the vehicle depending on the system status of the vehicle's function. [4] Method according to one of the preceding claims, wherein the recognition of an intention of the dialogue-based user guidance of the rule-based dialogue system as a function of the received message of the user by the rule-based dialogue system of the digital assistant comprises: Determining a confidence value of the intention of the dialog-based user guidance of the rule-based dialog system depending on the user message received by the rule-based dialog system of the digital assistant. [5] A method according to any one of the preceding claims, the method further comprising: If no intention of the dialogue-based user guidance of the rule-based dialogue system regarding the function of the vehicle was recognized by the rule-based dialogue system of the digital assistant or the confidence value of the recognized intention of the dialogue-based user guidance of the rule-based dialogue system regarding the function of the vehicle by the rule-based dialogue system of the digital assistant is less than a specified confidence threshold: Providing a context-based response message to the user of the vehicle through the artificial intelligence-based dialogue system of the digital assistant using the transmitted context information and depending on the transmitted message of the user; and If the confidence value of the detected intention of the dialogue-based user guidance of the rule-based dialogue system regarding the function of the vehicle by the rule-based dialogue system of the digital assistant is greater than or equal to the specified confidence threshold: Providing a response message to the user of the vehicle depending on the recognized intention of the dialog-based user guidance of the rule-based dialog system to the function of the vehicle by the rule-based dialog system of the vehicle. [6] Method according to one of the preceding claims, wherein the context-based response message is provided to the user of the vehicle by the artificial intelligence-based dialogue system of the digital assistant using all context information transmitted to the artificial intelligence-based dialogue system after activation of the dialogue-based user guidance from the rule-based dialogue system. [7] Method according to one of the preceding claims, wherein the context-based response message is provided to the user of the vehicle by the artificial intelligence-based dialogue system of the digital assistant using at least one or all of the context information transmitted to the artificial intelligence-based dialogue system after activation of the dialogue-based user guidance from the rule-based dialogue system and depending on the intention and / or the confidence value of the intention of the dialogue-based user guidance of the rule-based dialogue system of the digital assistant. [8] A computer-readable medium for operating a digital assistant of a vehicle, the computer-readable medium comprising instructions which, when executed on a computer and / or a control unit, carry out the method according to any one of claims 1 to 7. [9] System for operating a digital assistant of a vehicle, wherein the system is designed to carry out the method according to one of claims 1 to 7. [10] A vehicle comprising the system for operating a digital assistant of a vehicle according to claim 9.
Citation Information
Patent Citations
Monitoring of driving functions based on neural networks
DE102018222720A1