WO2026002181A1 - 交互方法、装置和车辆 - Google Patents

交互方法、装置和车辆

Info

Publication number
WO2026002181A1
WO2026002181A1 PCT/CN2025/104214 CN2025104214W WO2026002181A1 WO 2026002181 A1 WO2026002181 A1 WO 2026002181A1 CN 2025104214 W CN2025104214 W CN 2025104214W WO 2026002181 A1 WO2026002181 A1 WO 2026002181A1
Authority
WO
WIPO (PCT)
Prior art keywords
vehicle
output
information
input
environmental information
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
PCT/CN2025/104214
Other languages
English (en)
French (fr)
Inventor
金舟
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Yinwang Intelligent Technology Co Ltd
Original Assignee
Shenzhen Yinwang Intelligent Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shenzhen Yinwang Intelligent Technology Co Ltd filed Critical Shenzhen Yinwang Intelligent Technology Co Ltd
Publication of WO2026002181A1 publication Critical patent/WO2026002181A1/zh
Pending legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/16—Devices for psychotechnics; Testing reaction times ; Devices for evaluating the psychological state
    • A61B5/18—Devices for psychotechnics; Testing reaction times ; Devices for evaluating the psychological state for vehicle drivers or machine operators
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W40/00—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/08—Speech classification or search
    • G10L15/18—Speech classification or search using natural language modelling
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/26—Speech to text systems
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
    • G10L25/48—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use
    • G10L25/51—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination
    • G10L25/63—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination for estimating an emotional state
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/12—Protocols specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks

Definitions

  • This application relates to the field of smart cockpits, and more specifically, to an interaction method, device, and vehicle.
  • This application provides an interaction method, device, and vehicle that can prevent drivers from experiencing anger and retaliatory behavior during interactions between vehicles, thereby avoiding traffic accidents and improving the user's driving experience.
  • an interaction method comprising: obtaining a first intention of a user in a vehicle cabin, the first intention being associated with another vehicle and the first intention being associated with the user's negative emotions; performing de-emotionalization processing on the first intention to obtain a second intention; and sending the second intention to the other vehicle.
  • the first intention is de-emotionally processed to obtain the second intention, including: de-emotionally processing the first intention to obtain the de-emotionally processed intention; and adding polite language to the de-emotionally processed intention to obtain the second intention.
  • the method before obtaining the first intention of the user in the vehicle cabin, the method further includes: obtaining the user's first voice command; determining the first intention based on the first voice command; wherein, before sending the second intention to the other vehicle, the method further includes: when the first voice command does not include slot information corresponding to the first intention, determining the information of the other vehicle based on the first intention and data collected by sensors outside the vehicle cabin; or, when the first voice command includes slot information corresponding to the first intention, determining the information of the other vehicle based on the first slot information and data collected by the sensors.
  • the method before obtaining the first intention of the user in the vehicle cabin, the method further includes: obtaining a second voice command of the user, the second voice command including the first intention; wherein, the first intention is de-emotionally processed to obtain the second intention, including: inputting the second voice command, the vehicle information, and the vehicle's surrounding environment information into an inference model to obtain the second intention and the information of the other vehicle.
  • obtaining the first intention of a user inside the vehicle cabin includes: obtaining the user's driving behavior and the driving record of the other vehicle; determining the first intention based on the user's driving behavior and the driving record of the other vehicle; wherein, before sending the second intention to the other vehicle, the method further includes: determining information about the other vehicle based on the first intention and data collected by sensors outside the vehicle cabin.
  • sending the second intention to the other vehicle includes: sending the second intention to the other vehicle based on the signal strength and/or signal quality of the environment in which the vehicle is located.
  • sending the second intention to the other vehicle based on the signal strength and/or signal quality of the environment in which the vehicle is located includes: when the signal strength is greater than or equal to a preset signal strength, and/or the signal quality is greater than or equal to a preset signal quality, sending the information of the other vehicle and the second intention to a cloud server, so that the cloud server sends the second intention to the other vehicle based on the information of the other vehicle.
  • sending the second intention to the other vehicle includes: sending the second intention to the other vehicle via near-field communication when the signal strength is less than a preset signal strength and/or the signal quality is less than a preset signal quality.
  • an interaction method comprising: receiving a second intention from a vehicle, the second intention being an intention obtained by deemotionally processing a first intention of a user in the vehicle's cabin; and a control prompting device prompting the user with the second intention.
  • control prompting device prompts the user with the second intention, including: determining a first driving opinion based on the second intention; and controlling the prompting device to prompt the user with the second intention and the first driving opinion.
  • control prompting device prompts the user with the second intention, including: prompting the user with the second intention based on the user's state in the driver's area.
  • the second intent indicates abnormal driving behavior
  • the method further includes: controlling the vehicle lights and/or external projection information display to show first information, the first information being used to apology to and/or thank the user inside the vehicle.
  • an interaction method includes: acquiring a first input from a user in a vehicle cabin, information about the vehicle, and environmental information about the surrounding environment; determining a first output based on the first input, the vehicle information, and the environmental information; and sending the first output to another vehicle.
  • determining the first output based on the first input, the vehicle information, and the environmental information includes: inputting the first input, the vehicle information, and the environmental information into the inference model to obtain the first output.
  • the vehicle information includes at least one of the vehicle's speed, the type of road on which the vehicle is located, and the type of lane on which the vehicle is located.
  • the environmental information includes data collected by the vehicle's sensors.
  • an interactive device which includes a unit or module for performing a method for any of the possible implementations of the first to third aspects described above.
  • an interactive device including a processing unit and a storage unit, wherein the storage unit is used to store instructions, and the processing unit executes the instructions stored in the storage unit to cause the interactive device to perform any one of the possible methods of the first to third aspects.
  • an interactive system which includes a sensing system and a computing platform, wherein the computing platform includes any one of the possible devices in the fourth or fifth aspect.
  • a vehicle in a seventh aspect, includes the device described in the fourth or fifth aspect, or the system described in the sixth aspect.
  • a computer program product comprising: computer program code, which, when run on a computer, causes the computer to perform any one of the possible methods described in the first to third aspects.
  • a computer-readable medium that stores program code, which, when run on a computer, causes the computer to perform any one of the possible methods described in the first to third aspects.
  • embodiments of this application provide a chip system including a processor for calling a computer program or computer instructions stored in a memory, such that the processor performs any one of the possible methods described in the first to third aspects above.
  • the processor is coupled to the memory via an interface.
  • the chip system further includes a memory in which computer programs or computer instructions are stored.
  • this application provides an interaction method, which includes: acquiring a first input from a user in a vehicle cabin and environmental information around the vehicle, the environmental information including information about another vehicle; determining a first output based on the first input and the environmental information; and sending the first output to another vehicle or a terminal device, the terminal device being associated with the other vehicle.
  • the first output can be determined by the user's input and environmental information within the vehicle's cabin, and can then be sent to another vehicle or terminal device. This breaks down the information isolation between drivers in two vehicles, enabling interaction between vehicles.
  • the first input may include the user's voice input.
  • the first input may include user input to one or more components in the vehicle.
  • the one or more components may include, but are not limited to, a steering wheel, lights, a horn, an accelerator pedal, or a brake pedal.
  • the first input may include the user's physiological characteristics.
  • the physiological characteristics may include the user's blood pressure, heart rate, facial expressions, etc.
  • the first input is associated with the user's negative emotions
  • the first output includes the de-emotionalized output
  • the information that needs to be interacted with another vehicle can be further processed based on the user's input and environmental information, thereby achieving de-emotional processing.
  • This enables effective and amicable communication between vehicles, achieving the goal of transmitting effective information without transmitting negative emotions, and helping to avoid conflicts between vehicles caused by the driver's negative emotions.
  • the first input includes impolite language
  • the first output includes polite language converted from the impolite language
  • the vehicle while performing de-emotional processing, the vehicle can also convert impolite language used by users and send the converted polite language to another vehicle. This helps to further achieve effective and friendly communication between vehicles, conveying effective information without transmitting negative emotions, and helps reduce the probability of conflicts between drivers.
  • the method further includes: controlling the prompting device to output the third output, which includes output results for reassuring the user in the cabin.
  • the system can also use this first input and environmental information to provide a soothing output to the user in the cabin.
  • this first input and environmental information can also soothe the user in the cabin, helping to alleviate negative emotions and improve the user's driving experience.
  • the third output may include reassuring phrases.
  • the third output may include execution instructions for one or more components within the cockpit.
  • the first inference model can be a multimodal model.
  • the method further includes: acquiring vehicle information; wherein, determining the first output based on the first input and environmental information includes: determining the first output based on the first input, vehicle information, and environmental information.
  • a first output can be determined, which can then be sent to another vehicle or terminal device. This breaks down information barriers between drivers in two vehicles, enabling vehicle-to-vehicle interaction. Furthermore, by incorporating vehicle information, the accuracy of the first output result can be further improved.
  • the vehicle information includes the vehicle's historical driving records.
  • the vehicle information includes one or more of the vehicle's speed, acceleration, and position.
  • the first output is determined based on the first input, vehicle information, and environmental information, including: inputting the first input, vehicle information, and environmental information into the second inference model to obtain the first output.
  • the first input, vehicle information, and environmental information can be input into the first inference model to obtain the first output.
  • end-to-end input and output can be achieved through the second inference model.
  • the first inference model and the second inference model can be the same model.
  • the method further includes: determining information about another vehicle based on the first input and environmental information.
  • the information of the other vehicle includes the license plate information of the other vehicle; wherein, sending the first output to the other vehicle or terminal device includes: sending the first output and the license plate information of the other vehicle to the cloud server, so that the cloud server sends the first output to the other vehicle or terminal device based on the license plate information of the other vehicle.
  • a vehicle can send the first output and the license plate information of another vehicle to a cloud server.
  • This allows the cloud server to send the first output to the other vehicle using the license plate information.
  • the cloud server By forwarding information through the cloud server, the information isolation between the drivers of the two vehicles can be broken down, enabling interaction between vehicles.
  • the cloud server stores the association between the vehicle's license plate and the identification information of the terminal device (e.g., a mobile phone or vehicle).
  • the terminal device e.g., a mobile phone or vehicle.
  • control commands can be obtained through the first input and environmental information.
  • control commands can be obtained through the first input and environmental information.
  • control commands on one or more actuators it is possible to help the user perform corresponding operations, such as editing light signals, expressing gratitude, or expressing apologies.
  • this application provides an interaction method, the method comprising: acquiring a first output of another vehicle and environmental information surrounding the vehicle; determining a second output based on the first output and the environmental information; and controlling a prompting device to output the second output.
  • a vehicle can obtain a second output based on the output of another vehicle and the environmental information around the vehicle, and control the prompting device to output the second output. In this way, the information isolation between the drivers of the two vehicles can be broken down, realizing interaction between vehicles.
  • driver information can be incorporated when determining the second output. This allows the second output to better match the driver's profile, resulting in different outputs for different driver profiles. This helps improve the vehicle's intelligence and enhances the user's driving experience.
  • a second output is determined based on the first output and environmental information, including: determining the second output based on the historical driving records of other vehicles around the vehicle, the first output, and environmental information.
  • the historical driving records of other vehicles around the vehicle can be considered when determining the second output. This makes the second output more accurate and more acceptable to users in the cabin.
  • the second output is determined based on the first output and environmental information, including: inputting the first output and environmental information into the inference model to obtain the second output.
  • this application provides an interactive device, comprising: an acquisition unit for acquiring a first input from a user in a vehicle cabin and environmental information surrounding the vehicle, the environmental information including information about another vehicle; a determination unit for determining a first output based on the first input and the environmental information; and a sending unit for sending the first output to another vehicle or a terminal device, the terminal device being associated with the other vehicle.
  • the first input is associated with the user's negative emotions
  • the first output includes de-emotionalized output.
  • a unit is defined for: inputting a first input and environmental information into a first inference model to obtain a first output.
  • a unit is defined for: inputting the first input, vehicle information, and environmental information into the second inference model to obtain the first output.
  • the determining unit is also used to: determine information about another vehicle based on the first input and environmental information.
  • the information of the other vehicle includes the license plate information of the other vehicle; wherein, the sending unit is configured to: send a first output and the license plate information of the other vehicle to a cloud server, so that the cloud server sends the first output to the other vehicle or terminal device based on the license plate information of the other vehicle.
  • the apparatus further includes a control unit and a determining unit, which are further configured to determine a first control command based on the first input and environmental information, the first control command being associated with one or more actuators; the control unit is configured to control one or more actuators to execute the first control command.
  • this application provides an interactive device, the device comprising: an acquisition unit for acquiring a first output of another vehicle and environmental information surrounding the vehicle; a determination unit for determining a second output based on the first output and the environmental information; and a control unit for controlling a prompting device to output the second output.
  • a determining unit is used to: determine a second output based on data collected by sensors in the vehicle cabin, a first output, and environmental information.
  • a determining unit is used to: determine a second output based on first information of the driver in the vehicle, a first output, and environmental information, wherein the first information includes one or more of driving proficiency, driving habits, or physiological characteristics.
  • a determining unit is used to: determine a second output based on historical driving records of other vehicles around the vehicle, a first output, and environmental information.
  • a unit is defined for: inputting a first output and environmental information into the inference model to obtain a second output.
  • an interactive device comprising a processing unit and a storage unit, wherein the storage unit is used to store instructions, and the processing unit executes the instructions stored in the storage unit to cause the interactive device to perform any one of the possible methods in the eleventh or twelfth aspects.
  • an interactive system comprising a sensing system and a computing platform, wherein the computing platform comprises any of the possible devices of aspects thirteen through fifteen.
  • a vehicle in a seventeenth aspect, includes any of the possible devices from the thirteenth to the fifteenth aspects, or includes the system described in the sixteenth aspect.
  • a computer program product comprising: computer program code, which, when run on a computer, causes the computer to perform any one of the possible methods of the eleventh or twelfth aspect.
  • a computer-readable medium in a nineteenth aspect, storing program code that, when run on a computer, causes the computer to perform any of the possible methods of the eleventh or twelfth aspect described above.
  • embodiments of this application provide a chip system including a processor for calling a computer program or computer instructions stored in a memory, such that the processor executes any one of the possible methods in the eleventh or twelfth aspect described above.
  • the processor is coupled to the memory via an interface.
  • the chip system further includes a memory in which computer programs or computer instructions are stored.
  • FIG. 1 is a functional block diagram of the vehicle provided in an embodiment of this application.
  • FIGS 2A-2C are schematic diagrams of the interactive scenarios provided in the embodiments of this application.
  • Figure 3 is another schematic diagram of the interactive scenario provided in the embodiment of this application.
  • Figure 4 is another schematic diagram of the interactive scenario provided in the embodiments of this application.
  • Figure 5 is another schematic diagram of the interactive scenario provided in the embodiments of this application.
  • Figures 6A-6B are another schematic diagram of the interactive scenario provided in the embodiments of this application.
  • Figure 7 is another schematic diagram of the interactive scenario provided in the embodiments of this application.
  • Figures 8A-8C are another schematic diagram of the interactive scenario provided in the embodiments of this application.
  • Figure 9 is a schematic flowchart of the interaction method provided in an embodiment of this application.
  • FIG 10 is another schematic flowchart of the interaction method provided in the embodiments of this application.
  • FIG 11 is another schematic flowchart of the interaction method provided in the embodiments of this application.
  • Figure 12 is a system architecture diagram provided in an embodiment of this application.
  • Figure 13 is another system architecture diagram provided in an embodiment of this application.
  • Figure 14 is a schematic flowchart of the interaction method provided in the embodiments of this application.
  • Figure 15 is a system architecture diagram provided in an embodiment of this application.
  • Figure 16 is a schematic block diagram of the interactive device provided in an embodiment of this application.
  • Figure 17 is a schematic block diagram of the interactive device provided in an embodiment of this application.
  • prefixes such as “first” and “second” used in this application embodiment are merely for distinguishing different descriptive objects and do not limit the position, order, priority, quantity, or content of the described objects.
  • the use of ordinal numbers and other prefixes used to distinguish descriptive objects in this application embodiment does not constitute a limitation on the described objects.
  • the description of the described objects is given in the claims or the context of the embodiments, and should not constitute unnecessary restrictions due to the use of such prefixes.
  • “multiple" means two or more.
  • FIG. 1 is a functional block diagram of a vehicle 100 provided in an embodiment of this application.
  • the vehicle 100 may include a sensing system 110, a computing platform 120, a display device 130, and a sound-emitting device 140.
  • the sensing system 110 may include one or more sensors for sensing information about the environment surrounding the vehicle 100.
  • the sensing system 110 may include a positioning system, which may be a Global Positioning System (GPS), a BeiDou system, or another positioning system.
  • GPS Global Positioning System
  • BeiDou system BeiDou system
  • the sensing system 110 may also include one or more of the following: an inertial measurement unit (IMU), lidar, millimeter-wave radar, ultrasonic radar, and a camera device.
  • IMU inertial measurement unit
  • Computing platform 120 may include one or more processors, such as processors 121 to 12n (n is a positive integer).
  • a processor is a circuit with signal processing capabilities.
  • the processor may be a circuit with instruction read and execute capabilities, such as a central processing unit (CPU), a microprocessor, a graphics processing unit (GPU) (which can be understood as a type of microprocessor), or a digital signal processor (DSP).
  • the processor can implement certain functions through the logical relationship of hardware circuits. The logical relationship of the hardware circuits is fixed or reconfigurable.
  • the processor is a hardware circuit implemented by an application-specific integrated circuit (ASIC) or a programmable logic device (PLD), such as a field-programmable gate array (FPGA).
  • ASIC application-specific integrated circuit
  • PLD programmable logic device
  • FPGA field-programmable gate array
  • the process of a processor loading a configuration document and configuring the hardware circuit can be understood as the process of a processor loading instructions to implement the functions of some or all of the aforementioned units.
  • the processor can also be a hardware circuit designed for artificial intelligence, which can be understood as an ASIC, such as a neural network processing unit (NPU), tensor processing unit (TPU), or deep learning processing unit (DPU).
  • the computing platform 120 may also include a memory for storing instructions. Some or all of the processors 121 to 12n can call and execute instructions from the memory to implement the corresponding functions.
  • the in-cabin display devices 130 are mainly divided into two categories: the first is the in-vehicle display screen; the second is the projection display screen, such as the head-up display (HUD).
  • An in-vehicle display screen is a physical display screen and an important component of the in-vehicle infotainment system. Multiple displays can be installed in the cabin, such as the digital instrument cluster display, the central control screen, the display screen in front of the front passenger (also known as the front-facing passenger), the display screen in front of the left rear passenger, the display screen in front of the right rear passenger, and even the windows can be used as displays.
  • a head-up display also known as a head-up display system, is mainly used to display driving information such as speed and navigation on a display device in front of the driver (such as the windshield). This reduces the driver's eye-shift time, avoids pupil changes caused by eye-shifting, and improves driving safety and comfort.
  • HUDs include, for example, combiner-HUD (C-HUD) systems, windshield-HUD (W-HUD) systems, and augmented reality HUD (AR-HUD) systems. It should be understood that other types of HUD systems may emerge as technology evolves, and this application does not limit them.
  • the sound-generating device 140 can be a loudspeaker, audio system, or horn, etc.
  • This application provides an interaction method, device, and vehicle that locates vehicles based on information described in natural language, achieving the effect of establishing information connections. Simultaneously, it allows for secondary processing of user interaction information to achieve effective and friendly communication, conveying useful information without transmitting negative emotions. This helps prevent drivers from experiencing anger or retaliatory behavior during vehicle interactions, thereby avoiding traffic accidents and improving the user's driving experience.
  • Figures 2A-2C show schematic diagrams of the interactive scenarios provided in the embodiments of this application.
  • driver A when driver A, driving vehicle 100, notices vehicle 200 changing lanes from its current lane to the lane where vehicle 100 is located, driver A will slow down vehicle 100. This may cause driver A to feel dissatisfied with driver B inside vehicle 200. At this time, driver A may honk the horn of vehicle 100 to express their dissatisfaction.
  • driver B in vehicle 200 hears vehicle 100 honking, it's clear that driver A in vehicle 100 is unhappy. At this point, driver B can send a voice message: "Please tell the driver behind me, I'm sorry, there's an emergency at home.”
  • vehicle 200 After receiving voice information from the user, vehicle 200 can first convert the voice information into text content using an automatic speech recognition (ASR) module. Vehicle 200 can then understand the user's true intentions and current emotions based on the text content.
  • ASR automatic speech recognition
  • a natural language understanding (NLU) engine can be used to obtain the user's intention and slot.
  • the user's intention might be "I apology because of a family emergency," the slot might be "the driver behind,” and the user's current emotion might be apology.
  • Vehicle 200 can determine the license plate information of vehicle 100 using images captured by an external camera. Vehicle 200 can then send the user's intention and vehicle 100's license plate information to a cloud server. The cloud server can then locate vehicle 100's address based on its license plate information and, according to that address, send the user's intention to vehicle 100. Simultaneously, in response to receiving the user's intention, vehicle 200 can issue a voice message to the user, "Okay, I'll tell him now.”
  • vehicle 200 can determine its abnormal driving records and vehicle 100's driving records by data collected from sensors outside the cabin. For example, if vehicle 200 detects that vehicle 100 brakes suddenly (e.g., vehicle 100's deceleration is greater than or equal to a preset deceleration) while changing from the right lane to the left lane, and detects that vehicle 100 honks its horn or switches between low and high beams, it can determine that vehicle 100 gave way to vehicle 200 without its consent. After obtaining the user's intention, vehicle 200 can determine that the user's intention includes an expression of apology.
  • vehicle 100 brakes suddenly e.g., vehicle 100's deceleration is greater than or equal to a preset deceleration
  • vehicle 100 honks its horn or switches between low and high beams it can determine that vehicle 100 gave way to vehicle 200 without its consent.
  • vehicle 200 can determine that the user's intention includes an expression of apology.
  • vehicle 200 can send the user's intention to the cloud server, and also send a message of gratitude (e.g., sending a heart to thank them for giving way). At this time, vehicle 200 can send a voice message to the user, "Okay, I'll tell him now, and I'll also send him a heart for you.”
  • a message of gratitude e.g., sending a heart to thank them for giving way.
  • vehicle 200 can send a voice message to the user, "Okay, I'll tell him now, and I'll also send him a heart for you.”
  • Vehicle 100 After receiving the intent and gratitude from the user of Vehicle 200 from the cloud server, Vehicle 100 generates an apology text, "Dear driver, the driver of the car in front apologizes to you, there is an emergency at home," and a gratitude text, "He sent you a heart, thank you for your courtesy.”
  • vehicle 100 can also determine abnormal driving records of vehicle 200 (e.g., vehicle 200 failing to activate its left turn signal when changing lanes to the left) and driving records of vehicle 100 (e.g., vehicle 100 performing sudden braking and honking its horn after sudden braking) based on data collected by sensors outside the cabin. Based on the abnormal driving records of vehicle 200 and the driving records of vehicle 100, vehicle 100 can determine that driver A inside vehicle 100 is currently quite angry, and vehicle 100 can further process this text content. For example, in addition to expressing apologies and gratitude to driver A in vehicle 100, it can add emotionally reassuring text content (e.g., "Don't be angry").
  • abnormal driving records of vehicle 200 e.g., vehicle 200 failing to activate its left turn signal when changing lanes to the left
  • driving records of vehicle 100 e.g., vehicle 100 performing sudden braking and honking its horn after sudden braking
  • vehicle 100 can use the text-to-speech (TTS) module to convert the secondary processed text content into voice information, and send a voice message to driver A through the speaker in the vehicle 100 cabin: "Dear driver, the driver of the car in front apologizes to you. He has a family emergency. Please don't be angry. He sent you a message of kindness and thanks you for your courtesy.”
  • TTS text-to-speech
  • vehicle 200 detects that during a change from the right lane to the left lane, vehicle 100 brakes suddenly (e.g., the deceleration of vehicle 100 is greater than or equal to a preset deceleration) and detects that vehicle 100 honks its horn or switches between low beam and high beam, thus determining that vehicle 100 has unintentionally yielded to vehicle 200.
  • vehicle 200 can generate and send to the cloud server vehicle 100's license plate information, an apology text "Apologies to the driver of the following vehicle," and a thank-you text "Thank you for yielding" based on vehicle 200's abnormal driving record and vehicle 100's driving record.
  • the cloud server can find vehicle 100's address based on its license plate information and send the apology and thank-you text to vehicle 100 based on that address. Based on the abnormal driving records of vehicle 200 and the driving records of vehicle 100, vehicle 100 can determine that driver A in vehicle 100 is currently quite angry. Vehicle 100 can also further process this text content. For example, in addition to expressing apologies and gratitude to driver A in vehicle 100, it can add emotionally reassuring text content (e.g., "Don't be angry"). Vehicle 100 can use the speaker in the cabin to issue a voice message to driver A: "Dear driver, the driver of the car in front apologizes to you. There was an emergency at home. Please don't be angry. He showed you kindness and thanked you for yielding.”
  • driver A when driver A is driving vehicle 100 on the road and notices vehicle 200 changing lanes from its current lane to the lane where vehicle 100 is located, driver A will slow down vehicle 100 to give way to vehicle 200 out of courtesy.
  • vehicle 200 After driver B in vehicle 200 notices vehicle 100 yielding to them, they can issue a voice message saying, "Please thank the driver behind me.” Upon receiving this voice message, vehicle 200 can first convert it into text using its ASR (Automatic Speech Recognition) module. Vehicle 200 can then understand the user's true intentions and current emotions based on this text.
  • ASR Automatic Speech Recognition
  • vehicle 200 can determine that the user's intent is "to express gratitude,” the desired location is “the driver behind,” and the user's current emotion is gratitude.
  • vehicle 200 can send a voice message to the user, "Okay, I'll tell him now.”
  • vehicle 200 can use data collected by an external camera to determine the license plate number of vehicle 100.
  • Vehicle 200 can then send the user's intent and vehicle 100's license plate information to a cloud server.
  • the cloud server can then determine vehicle 100's address based on its license plate information and, according to that address, send the user's intent to vehicle 100.
  • vehicle 200 can determine abnormal driving records using data collected by sensors outside the cabin. For example, if vehicle 200 detects that vehicle 100 braked slowly while changing from the right lane to the left lane and no honking or switching between low and high beams is detected, it can be determined that vehicle 100 yielded to vehicle 200. After obtaining the user's intent, vehicle 200 can determine that the user's intent includes expressing gratitude. Based on the abnormal driving records of vehicle 200 and vehicle 100, vehicle 200 can also add other expressions of gratitude (e.g., sending a heart) to the user's intent. At this time, vehicle 200 can send a voice message to the user, "Okay, I'll tell him right away, and I'll also send him a heart for you.”
  • other expressions of gratitude e.g., sending a heart
  • vehicle 100 When vehicle 100 receives the user's intention, it can generate a text message expressing gratitude: "Dear driver, the driver of the car in front thanks you for yielding and sends you a heartfelt message.” Vehicle 100 can then use its TTS module to convert this text message into a voice message and transmit it to driver A through the speakers in the vehicle's cabin: "Dear driver, the driver of the car in front thanks you for yielding and sends you a heartfelt message.”
  • driver A when driver A, driving vehicle 100, notices vehicle 200 changing lanes from its current lane to the lane where vehicle 100 is located, driver A will slow down vehicle 100. This may cause driver A to feel dissatisfied with driver B inside vehicle 200's cabin. At this time, driver A will honk the horn of vehicle 100 to express their dissatisfaction.
  • vehicle 200 Upon noticing vehicle 100 honking its horn, driver B in vehicle 200 sends a voice message: "It's just cutting in line, I'll apologize and that's it.” After receiving the voice message, vehicle 200 can first convert it into text using its ASR (Automatic Speech Recognition) module. Vehicle 200 can then understand the user's true intentions and current emotions based on this text.
  • ASR Automatic Speech Recognition
  • Vehicle 200 can determine abnormal driving records using data collected by sensors outside the cabin. For instance, if vehicle 200 detects that vehicle 100 braked suddenly (e.g., vehicle 100's deceleration is greater than or equal to a preset deceleration) while changing from the right lane to the left lane, and detects that vehicle 100 honked its horn or switched between low and high beams, it can determine that vehicle 100 did not actively yield to vehicle 200. After obtaining the user's intent, vehicle 200 can determine the slot information as vehicle 100 based on its abnormal driving records and vehicle 100's driving records. Vehicle 200 can also determine vehicle 100's license plate information using images captured by an external camera.
  • vehicle 100's license plate information using images captured by an external camera.
  • the vehicle 200 can edit the user's intent, for example, by de-emotionalizing the user's intent and transforming dissatisfaction or annoyance into a sincere apology.
  • the user's intent might be "Apologize to the driver of the car behind me for cutting in line.”
  • vehicle 200 can also add other expressions of gratitude (e.g., sending a heart) to the user's intention.
  • vehicle 200 can send a voice message to the user: "Dear car owner, we have apologized to the driver of the car behind and sent him a heart.”
  • vehicle 100 After receiving the user's intention and other expressions of gratitude from vehicle 200, vehicle 100 generates a text message expressing thanks: "Dear driver, the driver of the car in front thanks you for yielding and sent you a message of appreciation.” Vehicle 100 can then convert this text message into voice information via a TTS module and transmit the corresponding voice message to driver A through the speakers in vehicle 100's cabin.
  • Figure 3 shows a schematic diagram of the interaction scenario provided in the embodiment of this application.
  • driver A when driver A is driving vehicle 100 at night and notices an oncoming vehicle 200 with its high beams on, it affects driver A's driving and may cause driver A to feel annoyed. At this point, driver A sends a voice message: "Could you please turn off your high beams?" After receiving the voice message, vehicle 100 can first convert it into text using the ASR (Automatic Speech Recognition) module. Vehicle 200 can then understand the user's true intentions and current emotions based on this text.
  • ASR Automatic Speech Recognition
  • the NLU engine might determine the user's intent as “Turn off your high beams,” the context as “oncoming vehicles,” and the user's current emotion as dissatisfaction or frustration.
  • Vehicle 100 can edit it. For instance, it can de-emotionally process the intent, resulting in “Turn off high beams,” or it can add polite language, such as changing the intent to "Please turn off your high beams.”
  • vehicle 100 can also translate the user's intent.
  • the user's intent after de-emotional processing is "turn off the high beams”.
  • the user's true intent can be obtained as "switch the high beams to low beams”.
  • vehicle 100 can obtain information about vehicle 200 traveling in the opposite direction based on the slot information. For example, vehicle 100 can determine the license plate information of vehicle 200 using images captured by an external camera. Vehicle 100 can then send the user's true intent and vehicle 200's license plate information to a cloud server. The cloud server can then locate vehicle 200's address based on its license plate information and, according to that address, send the de-emotionalized and translated version of the user's true intent to vehicle 200.
  • vehicle 200 can transform that intention. For example, vehicle 200 can translate the user's intention into an expression that is easy for driver B to understand and simultaneously provide driver B with effective suggestions.
  • vehicle 200 when vehicle 200 detects that driver B is a novice driver, it can explain to driver B the reason for turning off the high beams (e.g., affecting oncoming drivers) and provide driving advice (suggesting switching to low beams). For example, vehicle 200 can issue a voice message to driver B: "Dear driver, you are currently using high beams, which may be affecting oncoming drivers. The oncoming drivers would appreciate it if you could switch to low beams.”
  • driver B when vehicle 200 detects that driver B is an experienced driver, it can give driver B a concise message, such as "Oncoming vehicle, please switch to low beam headlights".
  • Figure 4 shows a schematic diagram of the interaction scenario provided in the embodiment of this application.
  • driver A while driving vehicle 100 in the overtaking lane, notices that vehicle 200 is traveling at a slow speed (e.g., 65 kph). This might cause driver A to feel dissatisfied with driver B inside vehicle 200.
  • driver A can issue a voice message: "Please tell the car in front of me that if you're driving so slowly, don't occupy the fast lane; you're obstructing others.”
  • vehicle 100 can first convert it into text using the ASR module. Vehicle 100 can then understand the user's true intentions and current emotions based on this text content.
  • the NLU engine might determine the user's intent as "Relay the message that you're driving so slowly, don't occupy the fast lane," the slot as "the vehicle in front,” and the user's current emotion as dissatisfaction or frustration.
  • Vehicle 100 can edit it, for example, by de-emotionalizing the intent to obtain a de-emotionalized intent, such as "Relay the message that you are driving too slowly in the fast lane.”
  • Vehicle 100 can send a voice message to the user, "Okay, I've conveyed that to the other party.”
  • vehicle 100 can also translate the user's intent.
  • the user's intent after de-emotional processing is "to convey to the vehicle that the current speed in the fast lane is too slow.”
  • the user's true intent can be obtained as "to convey to the vehicle that the speed in the fast lane should be increased or that the user should switch to the slow lane.”
  • Vehicle 100 can obtain information about vehicle 200 located in front of it based on the slot information. Vehicle 100 can determine the license plate information of vehicle 200 through data collected by a camera outside the cockpit. Vehicle 100 can send the de-emotionalized user intent and vehicle 200's license plate information to a cloud server. The cloud server can find the address of vehicle 200 based on the license plate information and send the de-emotionalized and translated true user intent to vehicle 200 based on that address.
  • vehicle 200 After receiving the user's true intention from the cloud server, vehicle 200 can transform that intention. For example, vehicle 200 can transform the user's intention into an expression that is easy for driver B to understand.
  • vehicle 200 can inform driver B about the speed limit information of the current fast lane and translate the user's intention into a way that is easy for driver B to understand. For example, vehicle 200 can issue a voice message to driver B: "Dear driver, Xiao A reminds you that you are currently in the fast lane, the speed limit is 80-100 km/h, and your current speed is 65 km/h. The driver behind you is a bit impatient, please speed up.”
  • vehicle 200 can translate the user's intention into a way that driver B can easily understand. For example, it can issue a more concise expression to driver B, such as the voice message "The driver behind you would like you to speed up or change lanes.”
  • Figure 5 shows a schematic diagram of the interaction scenario provided in the embodiment of this application.
  • driver A while driving vehicle 100 in their designated lane, notices vehicle 200 driving over the lane lines. This might cause driver A to feel dissatisfied with driver B inside vehicle 200. At this point, driver A can issue a voice message: "What's wrong with this person? How can they drive like that?" After receiving the voice message, vehicle 100 can first convert it into text using its ASR (Automatic Responsiveness) module. Vehicle 100 can then understand the user's true intentions and current emotions based on this text.
  • ASR Automatic Responsiveness
  • the NLU engine might determine the user's intent as "complaining about abnormal driving behavior" and their current emotion as either complaining or angry.
  • Vehicle 100 can issue calming voice messages (e.g., "Don't be angry, it's not worth getting upset") and recommend actions to the user to soothe their emotions (e.g., asking if they want to play soft music or turn on the air freshener).
  • calming voice messages e.g., "Don't be angry, it's not worth getting upset
  • recommend actions to the user to soothe their emotions e.g., asking if they want to play soft music or turn on the air freshener.
  • Vehicle 100 could use the cabin speakers to issue the voice message, "Dear owner, don't be angry, it's not worth getting upset. I can play some soft music for you, how about that?"
  • vehicle 100 can determine information about vehicle 200 with abnormal driving records based on the user's intent and data collected by sensors outside the cabin. For example, vehicle 100 may determine that vehicle 200 is driving across lane lines and obtain vehicle 200's license plate information based on sensor data. Vehicle 100 can then translate the user's intent based on this abnormal driving record. For example, the translated user's true intent might be "Please ask the driver of the vehicle in front to control their vehicle and stay within the lane.” Vehicle 100 can then send the user's true intent and vehicle 200's license plate information to a cloud server. The cloud server can then determine vehicle 200's address based on its license plate information and send the user's true intent to vehicle 200 based on that address.
  • Vehicle 200 can generate an expression that is easy for driver B to accept after receiving the true intent of the user of vehicle 100 from the cloud server.
  • vehicle 200 when vehicle 200 detects that driver B is a novice driver, it can inform driver B of the current abnormal driving behavior and provide driving suggestions. For example, vehicle 200 can issue a voice message to driver B: "Dear driver, you are currently driving over the lane lines, which may affect oncoming drivers. Please drive your vehicle in the lane as soon as possible.”
  • driver B when vehicle 200 detects that driver B is an experienced driver, it can issue a concise expression to driver B, such as issuing a voice message "Caution, driving over the line".
  • Figures 6A and 6B show schematic diagrams of the interactive scenarios provided in the embodiments of this application.
  • vehicle 200 is in intelligent driving mode and is making a U-turn ahead.
  • Vehicle 200 can determine that vehicle 100 is behind it and obtain vehicle 100's license plate information by using data collected from external sensors (e.g., a camera located at the rear of vehicle 200).
  • vehicle 200 can send vehicle 100's license plate information and indication information 1 to the cloud server.
  • Indication information 1 indicates that vehicle 200 is in intelligent driving mode and is about to make a U-turn ahead.
  • the cloud server can determine vehicle 100's address based on vehicle 100's license plate information and send indication information 1 to vehicle 100 based on that address.
  • vehicle 100 After receiving instruction information 1 from vehicle 200 sent by the cloud server, vehicle 100 can generate a prompt message.
  • This prompt message can include the status of vehicle 200 and the content to remind driver A.
  • vehicle 100 can use a TTS module to convert the prompt message into voice information and play the voice message "Dear driver, the vehicle ahead is in intelligent driving mode and wants to make a U-turn. Please be careful.”
  • vehicle 100 is in intelligent driving mode.
  • the driver in vehicle 200 notices an abnormal driving trajectory of vehicle 100, they issue a voice message, "How is this car being driven?" and express their dissatisfaction by honking the horn.
  • Vehicle 200 can determine the abnormal driving trajectory of vehicle 100 (e.g., excessive lateral movement) based on its own information and the driving information of surrounding vehicles. Vehicle 200 can then obtain the user's intent, "Please, vehicle ahead, stop moving laterally.”
  • Vehicle 200 can send the user's intent and vehicle 100's license plate information to a cloud server.
  • the cloud server can determine the address of vehicle 100 based on its license plate information and send the user's intent to vehicle 100 based on that address.
  • vehicle 100 can determine that vehicle 200 is behind vehicle 100 and obtain vehicle 200's license plate information based on data collected by external sensors (e.g., a camera located at the rear of vehicle 200) and the user's intent. At this time, vehicle 100 can send vehicle 200's license plate information and instruction information 2 to the cloud server. Instruction information 2 indicates that vehicle 100 is in intelligent driving mode. The cloud server can determine vehicle 200's address based on its license plate information and send instruction information 2 to vehicle 200 based on that address. Instruction information 2 includes an indication that vehicle 100 is in intelligent driving mode and an apology message.
  • external sensors e.g., a camera located at the rear of vehicle 200
  • vehicle 100 can send vehicle 200's license plate information and instruction information 2 to the cloud server.
  • Instruction information 2 indicates that vehicle 100 is in intelligent driving mode.
  • the cloud server can determine vehicle 200's address based on its license plate information and send instruction information 2 to vehicle 200 based on that address. Instruction information 2 includes an indication that vehicle 100 is in intelligent driving mode and an apology message.
  • vehicle 200 After receiving instruction information 2 from vehicle 100 sent by the cloud server, vehicle 200 can generate a prompt message.
  • This prompt message can include the status of vehicle 100 and content to remind the driver.
  • vehicle 200 can use a TTS module to convert the prompt message into voice information and play the voice message "Dear driver, the vehicle ahead is in intelligent driving mode, please excuse us" through the speakers in the cabin.
  • user input e.g., voice input
  • vehicle information e.g., license plate information
  • environmental information surrounding the vehicle e.g., license plate information
  • the vehicle can then send this information, along with the information to be conveyed to that other vehicle, to a cloud server.
  • the cloud server can then use this information to send the conveyed information back to that other vehicle.
  • Figure 7 shows a schematic diagram of the interaction scenario provided in an embodiment of this application.
  • vehicle 200 can determine the surrounding environmental information through data collected by sensors. For example, vehicle 200 can determine from the surrounding environmental information that the driver has decided to change lanes to the left due to road construction ahead. After determining that vehicles in the left and right lanes need to take turns passing in the road construction scenario, vehicle 200 can issue a voice message through the speaker in the cabin: "The driver who honked may not know about the road construction, I will inform him.”
  • Vehicle 200 can determine the license plate information of vehicle 100 through images captured by an external camera. Vehicle 200 can then send the license plate information and instruction information 3 to a cloud server. This instruction information 3 indicates that the lane currently occupied by vehicle 200 is under construction.
  • vehicle 100 After receiving instruction information 3 from vehicle 200 sent by the cloud server, vehicle 100 can determine, based on instruction information 3, abnormal driving records of vehicle 200, and driving records of vehicle 100, whether driver A's honking might be due to unawareness of road construction in the right lane, or driver A might be aware of the right lane construction but unaware that alternating lanes are required in this scenario. Vehicle 100 can then convey driving advice to the user in this situation, for example, by issuing a voice message through the cabin speaker: "Dear driver, the right lane is currently under construction. Alternating lanes are required. Please don't rush!"
  • vehicle 100 can generate an expression that is easy for driver A to accept based on the information of driver A.
  • vehicle 100 when vehicle 100 detects that driver A is a novice driver, it can inform driver A that vehicle 200 is not driving abnormally and provide driving suggestions. For example, vehicle 100 can issue a voice message to driver A: "Dear driver, the right lane is currently under construction and requires alternating traffic. Please don't rush!
  • driver B when vehicle 200 detects that driver B is an experienced driver, it can issue a simpler expression to driver A, such as issuing a voice message "Attention, road construction is underway on the right.”
  • the inference result can include the voice message "Okay, I've already communicated with the other party" relayed to the driver of vehicle 100, vehicle 200's license plate information, and the first message conveyed to vehicle 200 (e.g., text content "Your car seems to be blocking the car behind, and the driver behind is in a hurry to pass”).
  • the input to the inference model may include one or more of the following: voice, images, text, and video streams (images and video streams may be images or video streams collected by sensors inside and outside the cockpit).
  • the inference model can be supplemented with external knowledge (e.g., road regulations, road conditions, best practices).
  • external knowledge e.g., road regulations, road conditions, best practices.
  • the inference model can search for external knowledge (e.g., playing a racing game in a car, or a real-world driving scenario) to search for different types of knowledge. For example, if a user is playing a game on the passenger-side screen and gives the voice command "That car in front is so slow,” the inference model can input this voice message and the status of the passenger-side screen to output the text or voice message "Pass it!” As another example, if a user is driving in an urban area and the car in front is slow, honking their horn and giving the voice command "That car in front is so slow," the inference model can input this voice message, the vehicle's current speed, lane information, and surrounding environmental information to output the voice message "The speed limit on this road is 50-70 km/h. The car in front is driving normally, please don't rush!"
  • external knowledge e.g., playing a racing game in a car, or a real-world driving scenario
  • Vehicle 100 can send the license plate information of vehicle 200 and the initial message to vehicle 200 to the cloud server.
  • the cloud server can determine, based on the license plate information of vehicle 200 and the stored association between license plate information and accounts, that the account corresponding to the license plate information of vehicle 200 includes multiple devices, such as vehicle 200, mobile phones, smartwatches, and tablets. The cloud server can then send the initial message to multiple devices under that account.
  • the mobile phone under the account logged into vehicle 200 when the mobile phone under the account logged into vehicle 200 receives the first message, it can display a prompt box on the phone's screen.
  • This prompt box includes the message "Dear driver, your vehicle seems to be blocking oncoming traffic. The driver behind is in a hurry to pass. Do you want to remotely control the vehicle?", a confirmation control, and a cancellation control.
  • the phone can display environmental information and a prompt box around vehicle 200.
  • This environmental information includes a dashed box, which indicates the suggested area to move to.
  • the prompt box includes the message "We suggest you move your vehicle to the dashed box to facilitate the passage of oncoming traffic", a confirmation control, and a cancellation control.
  • the phone can send a control command to the vehicle, instructing vehicle 200 to move from its current position to the area within the dashed box.
  • the phone can also display the process of vehicle 200 moving towards the dashed box.
  • the phone detects that the user sends the voice message "Please apologize to the driver behind me,” the phone can input this voice message into the inference model, thus obtaining the text message to the driver behind: "The car has been moved and I apology to you.”
  • the cloud server can forward the text content to vehicle 100.
  • Vehicle 100 can input the text content, vehicle information, and environmental information around the vehicle into the inference model, thereby obtaining the voice message to be broadcast to the driver of vehicle 100: "Dear car owner, the driver of the car in front has moved the car and apologyd to you for the delay.”
  • vehicle 100 can control the speaker to emit a voice message: "Dear car owner, the driver of the car in front has moved the car and apologyd to you for the delay.”
  • Figure 9 shows a schematic flowchart of the interaction method 900 provided in an embodiment of this application.
  • the method 900 includes:
  • S910 obtain the first intention of a user in the vehicle cabin, the first intention being associated with another vehicle and the first intention being associated with the user's negative emotions.
  • the user's initial intent could be obtained through voice command analysis, such as the voice message in Figure 3: "Could the car on the other side turn off its high beams?" Or, as shown in Figure 4: "Could you tell the car in front of you to stop driving so slowly and not occupy the fast lane, obstructing other drivers?”
  • the user's initial intent can be obtained through or based on the vehicle's driving record (e.g., the driver of the vehicle frequently switching between low beam and high beam, sudden braking, sudden steering wheel turning) and the driving record of the other vehicle (the other vehicle is currently using high beam).
  • the vehicle's driving record e.g., the driver of the vehicle frequently switching between low beam and high beam, sudden braking, sudden steering wheel turning
  • the driving record of the other vehicle the other vehicle is currently using high beam.
  • the user's initial intent could be determined by the vehicle's driving record (honking in the fast lane) and the driving record of the other vehicle (e.g., driving too slowly in the fast lane).
  • a user's initial intent can also be determined through their actions (e.g., body language) and facial expressions (e.g., annoyance).
  • actions e.g., body language
  • facial expressions e.g., annoyance
  • the user's initial intent can also be determined by using infrared sensors and thermal imaging sensors to determine facial information (e.g., the face turning red).
  • the initial intention can still be obtained through the emotions of other passengers in the cabin. For instance, another passenger might send a voice message like, "Could that car over there turn off its high beams?"
  • this primary intent can also be obtained through information such as heart rate or blood pressure detected by the user's watch.
  • this initial intention can also be determined by the grip strength of the steering wheel.
  • the first intention is de-emotionally processed to obtain the second intention.
  • the primary intent can be de-emotionalized to obtain a de-emotionalized secondary intent.
  • the step of de-emotionalizing the first intention to obtain the second intention includes: de-emotionalizing the first intention to obtain a de-emotionalized intention; and adding polite language to the de-emotionalized intention to obtain the second intention.
  • polite language can be added to the intent, thereby further achieving the goal of effective and friendly communication, avoiding friction between drivers, and preventing traffic accidents. Furthermore, adjusting the language style is more conducive to communication.
  • user intent can be identified using an ASR module and an NLU module.
  • NLU identifies the user's intent
  • de-emotionalization processing can be performed.
  • the method before obtaining the first intent of the user in the vehicle cabin, the method further includes: obtaining the user's first voice command; determining the first intent based on the first voice command; wherein, before sending the second intent to the other vehicle, the method further includes: when the first voice command does not include slot information corresponding to the first intent, determining the information of the other vehicle based on the first intent and data collected by sensors outside the vehicle cabin; or, when the first voice command includes slot information corresponding to the first intent, determining the information of the other vehicle based on the first slot information and data collected by the sensors.
  • the information of another vehicle (which may be at least one of license plate information, body color, and brand) can be determined based on the slot information and the driving record of another vehicle; if slot information is not included, the information of another vehicle (which may be at least one of license plate information, body color, and brand) can be determined based on the user's intent and the data collected by the sensor.
  • the identification of the user's intent can also be achieved through an inference model (e.g., a multimodal large model).
  • an inference model e.g., a multimodal large model
  • the method before obtaining the first intent of the user in the vehicle cabin, the method further includes: obtaining a second voice command from the user, the second voice command including the first intent; wherein, performing de-emotional processing on the first intent to obtain the second intent includes: inputting the second voice command, the vehicle information, and the vehicle's surrounding environment information into an inference model to obtain the second intent and the information of the other vehicle.
  • obtaining the first intention of a user inside the vehicle cabin includes: obtaining the user's driving behavior and the driving record of the other vehicle; determining the first intention based on the user's driving behavior and the driving record of the other vehicle; wherein, before sending the second intention to the other vehicle, the method further includes: determining the information of the other vehicle based on the first intention and data collected by sensors outside the vehicle cabin.
  • the user's intent can be determined based on the user's driving behavior inside the vehicle and the driving record of another vehicle. For example, if the vehicle detects that the user is pressing the horn for a long time or frequently switching between low beam and high beam, it can determine that the other vehicle is exhibiting abnormal driving behavior, thereby determining the user's intent (which is generally to instruct the other vehicle to remove the abnormal driving behavior).
  • the method before sending the second intent to the other vehicle, the method further includes: determining that the second intent complies with road regulations.
  • the above can also be used to determine whether a user's voice command is valid. If valid, the inference model can determine how to communicate with the other party and provide an output. Determining validity can be considered part of the internal implementation process of the inference model.
  • sending the second intention to the other vehicle includes: sending the second intention to the other vehicle based on the signal strength and/or signal quality of the environment in which the vehicle is located.
  • sending the second intention to the other vehicle based on the signal strength and/or signal quality of the environment in which the vehicle is located includes: when the signal strength is greater than or equal to a preset signal strength, and/or the signal quality is greater than or equal to a preset signal quality, sending the information of the other vehicle and the second intention to the cloud server, so that the cloud server sends the second intention to the other vehicle based on the information of the other vehicle.
  • sending the second intention to the other vehicle includes: sending the second intention to the other vehicle via near-field communication when the signal strength is less than a preset signal strength and/or the signal quality is less than a preset signal quality.
  • FIG. 10 shows a schematic flowchart of the interaction method 1000 provided in an embodiment of this application.
  • the method 1000 includes:
  • S1010 Receive a second intent from the vehicle, the second intent being an intent obtained after de-emotionalizing the first intent of the user in the vehicle cabin.
  • control prompting device prompts the user with the second intention.
  • control prompting device prompts the user with the second intention, including: determining a first driving opinion based on the second intention; and controlling the prompting device to prompt the user with the second intention and the first driving opinion.
  • the receiving device can adjust at least one of the voice, tone, or manner of speaking when sending the voice message based on the user's state. For instance, if the driver is currently irritable, the second intention can be conveyed humorously to ease the tension. Or, if the driver is currently in a good mood, the second intention can be stated directly.
  • control prompting device prompts the user for the second intention, including: prompting the user for the second intention based on the user's status in the driver's area.
  • the second intent indicates abnormal driving behavior
  • the method further includes: controlling the vehicle lights and/or external projection information to display first information, the first information being used to apology to and/or thank the user inside the vehicle.
  • FIG 11 shows a schematic flowchart of the interaction method 1100 provided in an embodiment of this application.
  • the method 1100 includes:
  • S1110 acquire the first input from the user in the vehicle cabin, the vehicle information, and the environmental information around the vehicle.
  • the first input can be voice messages from the user, user actions, or user facial expressions.
  • the user here can be a user in the driver's area, or a user in another area.
  • vehicle information may include one or more of the following: vehicle speed, vehicle actuator operation (e.g., frequent switching between low beam and high beam, emergency braking, excessive rate of change of steering wheel angle, honking, grip strength on the steering wheel, etc.), location, type of lane, type of road, and data collected by sensors in the cabin (e.g., data collected by infrared sensors and thermal imaging sensors).
  • vehicle actuator operation e.g., frequent switching between low beam and high beam, emergency braking, excessive rate of change of steering wheel angle, honking, grip strength on the steering wheel, etc.
  • location e.g., type of lane, type of road
  • data collected by sensors in the cabin e.g., data collected by infrared sensors and thermal imaging sensors.
  • environmental information around the vehicle includes data collected by sensors outside the vehicle cabin, such as images and video streams captured by cameras outside the cabin.
  • the first input may also include input from the user's electronic device (e.g., a mobile phone, smartwatch, or smart bracelet). For example, heart rate or blood pressure detected by the user's watch.
  • the user's electronic device e.g., a mobile phone, smartwatch, or smart bracelet.
  • heart rate or blood pressure detected by the user's watch For example, heart rate or blood pressure detected by the user's watch.
  • the first output may include information (e.g., text content or voice information) to be conveyed to another vehicle.
  • information e.g., text content or voice information
  • the method further includes: outputting control commands for actuators in the vehicle cabin (e.g., air conditioning, ambient lighting, fragrance, or in-vehicle speakers) based on the first input, the vehicle information, and the environmental information.
  • actuators in the vehicle cabin e.g., air conditioning, ambient lighting, fragrance, or in-vehicle speakers
  • determining the first output based on the first input, the vehicle information, and the environmental information includes: inputting the first input, the vehicle information, and the environmental information into the inference model to obtain the first output.
  • the vehicle information includes at least one of the vehicle's speed, the type of road the vehicle is on, and the type of lane the vehicle is in.
  • the environmental information includes data collected by the vehicle's sensors.
  • the interactive system includes a transmitting device and a receiving device.
  • the transmitting device is configured to acquire a first intention of a user in a vehicle cabin, the first intention being associated with another vehicle and related to the user's negative emotions; the transmitting device is further configured to perform de-emotional processing on the first intention to obtain a second intention; the transmitting device is further configured to send the second intention to the other vehicle; and the receiving device is configured to control a prompting device to prompt the user with the second intention.
  • FIGS 12 and 13 respectively show schematic diagrams of the system architecture provided in the embodiments of this application.
  • external sensors e.g., cameras
  • can identify the appearance information of nearby vehicles e.g., color, brand, model
  • license plate numbers e.g., license plate numbers
  • relative positional relationships between vehicles e.g., license plate numbers
  • An external microphone can identify the horn characteristics of other vehicles (whether they honk, the number of horns, and the urgency), and identify surrounding environmental information (whether they honk).
  • ADS Autopilot System
  • the system can: understand the ambiguous semantics of user descriptions and map them to a specific nearby vehicle; effectively process the user's expression by combining it with current dynamic and static environmental information, the user's current state, and general handling procedures (knowledge) for the current situation, resulting in accurate, rational, polite, and appropriate communication; provide suitable feedback and suggestions to the user based on current state information, and assist the user in performing actions such as editing light signals, expressing gratitude, or apologizing.
  • Information is exchanged via cloud servers through information matching or mobile networks, or via near-field communication. This type of information exchange can protect the driver's privacy by not exposing their personal information; it also reduces ambiguity and improves communication efficiency and safety by eliminating the need for non-verbal indirect expressions (flashing lights, honking horns).
  • the system maps information to nearby vehicles and exchanges information, breaking down information barriers between drivers, enabling communication, and effectively managing emotions.
  • the system can process and filter the information expressed by the driver based on the driver's current state, environmental conditions, driving behavior, and other information. After extracting key information, it can perform secondary processing and summarization for interactive transmission. It can also fulfill a series of explicit expressions of intent on behalf of the driver, such as gratitude, apology, and opinions.
  • vehicle information is located based on information described in natural language to achieve the effect of establishing a connection (license plate, VIN number, etc. can be used as ID).
  • the interactive information is processed again based on the status information (person, vehicle, environment) to achieve effective and friendly communication, and to convey effective information without conveying negative emotions.
  • vehicles can perform point-to-point interaction or broadcast interaction.
  • vehicle hardware such as headlights and projectors, can be used to transmit secondary language information.
  • the projection includes DLP projection onto the ground and pedestrian interaction, or it may include projection onto a vehicle window and surrounding interaction.
  • vehicle-to-vehicle interaction may include mobile network interaction as well as near-field interaction (e.g., via star flash).
  • This application also provides an interactive device, which includes a module or unit for performing the above-described interactive method.
  • FIG. 14 shows a schematic flowchart of an interaction method 1400 provided in an embodiment of this application.
  • the method 1400 includes:
  • the first vehicle acquires the first input from the user in the first vehicle's cabin and the environmental information around the first vehicle, including information about the second vehicle.
  • the first input may include the user's voice input.
  • the first input can be the voice input of driver A: "Could the car on the other side turn off its high beams?”
  • the first input can be the voice input of driver A: "Please tell the car in front of me that if you're driving so slowly, don't occupy the fast lane and you're obstructing others.”
  • the first input can be the voice input of driver A, "What's wrong with this person? How is he driving?"
  • the first input may include user input to one or more components in the vehicle.
  • the one or more components may include, but are not limited to, a steering wheel, lights, horn, accelerator pedal, or brake pedal.
  • the first input can be the horn input of driver A.
  • the first input can be the driver's horn input in vehicle 200.
  • the first input may include the user's physiological characteristics information.
  • the physiological characteristics information may include the user's blood pressure, heart rate, facial expressions, etc.
  • the vehicle can acquire image data from cameras inside the cabin and determine the user's facial expression based on that image data.
  • the user's facial expression might be a frowning expression.
  • the first input may include the user's limb input.
  • the limb input may be a gesture input.
  • the first input can be data collected by a user's wearable device (e.g., a smartwatch, smart bracelet, etc.) inside the cabin.
  • a user's wearable device e.g., a smartwatch, smart bracelet, etc.
  • the first input can be input from one or more users in the cockpit.
  • the environmental information includes data collected by sensors outside the cockpit.
  • external sensors include, but are not limited to, one or more of cameras, lidar, millimeter-wave radar, and microphones.
  • the first vehicle determines the first output based on the first input and environmental information.
  • determining the negative emotions associated with the first input includes: determining that the first input meets preset conditions.
  • the preset condition includes that the text content corresponding to the voice input contains target semantics, which includes semantics related to negative emotions such as complaining, anger, impolite language, and modal particles.
  • target semantics related to complaining include “Why drive like that?" and "Obstructing others.”
  • semantics related to modal particles include "ah”.
  • the preset condition includes the duration for which the user presses the horn being greater than or equal to a preset duration.
  • the preset condition includes the user's grip force on the steering wheel being greater than or equal to a preset grip force; or, the preset condition includes the rate of change of the steering angle of the steering wheel being greater than or equal to a preset rate of change within a preset time period.
  • the preset conditions include the user's facial expression being a frown, an angry expression, or an angry expression.
  • the preset condition includes that the user's heart rate is greater than or equal to a preset heart rate.
  • the first input is associated with the user's negative emotions
  • the first output includes the de-emotionalized output.
  • the information that needs to be interacted with by the second vehicle can be further processed based on the user's input and environmental information, thereby achieving de-emotional processing.
  • This enables effective and friendly communication between vehicles, conveying effective information without transmitting negative emotions, and helps avoid conflicts between vehicles caused by the driver's negative emotions.
  • the first output includes impolite language, and the first output includes polite language converted from the impolite language.
  • the vehicle while performing de-emotional processing, the vehicle can also convert impolite language used by the user and send the converted polite language to the second vehicle. This helps to further achieve effective and friendly communication between vehicles, conveying effective information without transmitting negative emotions, and helps reduce the probability of conflicts between drivers.
  • the first input is associated with the user's negative emotions
  • the method 1400 further includes: controlling the prompting device to output a third output, the third output including output results for reassuring the user in the cabin.
  • the system can also use this first input and environmental information to provide a soothing output to the user in the cabin.
  • this first input and environmental information can also soothe the user in the cabin, helping to alleviate negative emotions and improve the user's driving experience.
  • the third output may include reassuring phrases.
  • the first input can be the voice input of driver A as shown in Figure 4, and the third output can be voice output 1 "Okay, I have expressed it to the other party" and voice output 2 "The driver in front may be a novice driver”.
  • the third output may include execution instructions for one or more components within the cockpit.
  • determining the first output based on the first input and environmental information includes: inputting the first input and environmental information into the first inference model to obtain the first output.
  • the first input and environmental information can be input into the first inference model to obtain the first output.
  • end-to-end input and output can be achieved through the first inference model.
  • the first inference model can be a multimodal model.
  • the voice input of driver A "Could the car on the other side turn off its high beams?”
  • the image data e.g., pictures or video streams
  • the first inference model can output the user's true intention, "Please switch the high beams to low beams.”
  • the voice input of driver A "Please tell the car in front of me not to drive so slowly and occupy the fast lane, it's obstructing others," and image data (e.g., pictures or video streams) collected by the camera outside the cabin of vehicle 100 can be input into the first inference model.
  • the first inference model can then output the user's true intention, "Please speed up in the fast lane.”
  • the voice input of driver A "What's wrong with this person? How is he driving?”
  • image data e.g., pictures or video streams
  • the first inference model can output the user's true intention, "Please do not drive over the lane lines.”
  • method 1400 further includes: obtaining information about a first vehicle; wherein the first vehicle determines a first output based on a first input and environmental information, including: the first vehicle determines the first output based on the first input, information about the first vehicle, and environmental information.
  • a first output can be determined, which can then be sent to a second vehicle or terminal device. This breaks down information isolation between drivers in the two vehicles, enabling vehicle-to-vehicle interaction. Furthermore, by incorporating vehicle information, the accuracy of the first output result can be further improved.
  • the vehicle information includes the vehicle's historical driving records.
  • the vehicle information may include one or more of the vehicle's speed, acceleration, and location.
  • the second inference model can be input into the second inference model, which can include the voice input of driver A, "Could the car on the other side turn off its high beams?”, the lighting status of vehicle 100 (vehicle 100 currently has its low beams on, or the driver is frequently switching between low beams and high beams), and the image data collected by the camera outside the cabin of vehicle 100.
  • the second inference model can then output the user's true intention, "Please switch the high beams to low beams".
  • the second inference model can be input into the second inference model, which can be the voice input of driver A, "Please tell the car in front of me that you are not driving so slowly and occupying the fast lane, which is blocking other people", the speed of vehicle 100 (e.g., 90 km/h) and the image data collected by the camera outside the cabin of vehicle 100.
  • the second inference model can output the user's true intention, "Please speed up in the fast lane”.
  • the first vehicle determines the first output based on the first input, the information of the first vehicle, and the environmental information, including: inputting the first input, the information of the first vehicle, and the environmental information into the second inference model to obtain the first output.
  • the first input, the information of the first vehicle, and the environmental information can be input into the first inference model to obtain the first output.
  • end-to-end input and output can be achieved through the second inference model.
  • the first inference model and the second inference model can be the same model.
  • the input to this multimodal model includes, but is not limited to, one or more of the following: speech, images, text content, or video streams.
  • This multimodal large model can be trained from a general multimodal large model.
  • this multimodal large model includes a knowledge base.
  • the knowledge base includes, but is not limited to, road regulations of one or more countries (or regions), speed limit information for road sections, and reasonable handling methods during vehicle interactions.
  • the output of the multimodal large model can be one or more of the following: the user's true intention after de-emotionalization, the converted polite language, and the translation result of the user's true intention.
  • method 1400 further includes: the first vehicle determining information about the second vehicle based on the first input and environmental information.
  • the information of the second vehicle can be determined by the user's input and environmental information in the first vehicle's cabin, thereby enabling the first output to be sent to the second vehicle or a terminal device.
  • the accuracy of the identified vehicle for interaction can be ensured by combining user input and environmental information when determining the information of the second vehicle.
  • the first vehicle determines the information of the second vehicle based on the first input and environmental information, including: the first vehicle inputs the first input and environmental information into a third inference model to obtain the information of the second vehicle.
  • the second inference model can be input into the following data: driver A's voice input "Could the car on the other side turn off its high beams?”, the headlight status of vehicle 100 (vehicle 100 currently has its low beams on, or the driver is frequently switching between low and high beams), and image data collected by a camera outside the cabin of vehicle 100.
  • This second inference model can then output the user's true intent, "Please switch the high beams to low beams," and the license plate information of vehicle 200.
  • the second inference model can be input into the second inference model, which can include the voice input of driver A, "Please tell the car in front of me not to drive so slowly and occupy the fast lane, it's obstructing others," the speed of vehicle 100 (e.g., 90 km/h), and the image data captured by the camera outside the cabin of vehicle 100.
  • the second inference model can then output the user's true intention, "Please speed up in the fast lane,” and the license plate information of vehicle 200.
  • the first inference model, the second inference model, and the third inference model can be the same model.
  • the output of the large multimodal model can also include information about the second vehicle.
  • the information about the second vehicle might include its license plate information.
  • the first inference model and the third inference model may not be the same model.
  • this third inference model can be implemented using a semantic recognition algorithm and an image segmentation algorithm.
  • the first vehicle receives the user's voice input, "Could the red car in front go faster?", it can acquire image 1 captured by a camera outside the cockpit.
  • the text content related to the target's attributes in the voice input e.g., "red car”
  • the first vehicle can extract ROI1 and ROI2 from image 1 based on an image segmentation algorithm.
  • ROI1 and ROI2 each include the red car.
  • image 1 can be divided into multiple regions; ROI1 can be region a within these regions, and ROI2 can be region b within these regions. If the user's gaze points to region a when the voice input is triggered, ROI1 can be selected as the target ROI.
  • the first vehicle can send this target ROI to a cloud server.
  • the cloud server can then analyze this target ROI to obtain the license plate information of the second vehicle.
  • method 1400 further includes: the first vehicle determining a first control command based on the first input and environmental information, the first control command being associated with one or more actuators; the first vehicle controlling one or more actuators to execute the first control command.
  • the first input can be a horn input from driver A.
  • the first vehicle can generate control commands for the sound-emitting device and ambient lighting based on the horn input and image data collected by sensors outside the cabin.
  • the control command for the sound-emitting device instructs it to play music that alleviates dissatisfaction (e.g., soothing music).
  • the control command for the ambient lighting includes controlling the ambient lighting color to be warm.
  • control commands can be obtained through the first input and environmental information.
  • control commands can be obtained through the first input and environmental information.
  • it is possible to alleviate the user's anger, resentment, or impatience.
  • the first vehicle sends a first output to the second vehicle or a terminal device, and the terminal device is associated with the second vehicle.
  • the information of the second vehicle includes the license plate information of the second vehicle; wherein, the first vehicle sending the first output to the second vehicle or the terminal device includes: the first vehicle sending the first output and the license plate information of the second vehicle to the cloud server, so that the cloud server sends the first output to the second vehicle or the terminal device based on the license plate information of the second vehicle.
  • the vehicle can send the first output and the license plate information of the second vehicle to the cloud server, thereby enabling the cloud server to send the first output to the second vehicle using the license plate information of the second vehicle.
  • the cloud server By forwarding information through the cloud server, the information isolation between the drivers of the two vehicles can be broken down, enabling interaction between vehicles.
  • the cloud server stores the association between the vehicle's license plate and the identification information of the terminal device (e.g., a mobile phone or vehicle).
  • the terminal device e.g., a mobile phone or vehicle.
  • the cloud server can store the correspondence between the identity document (ID) or address information of each account and the license plate information of one or more devices corresponding to each account.
  • Table 1 shows the correspondence between the address information and license plate information of each account, one or more devices corresponding to each account.
  • a cloud server can receive vehicle license plate information sent by multiple terminal devices. For instance, user 1 can enter the license plate number xxx associated with vehicle B while logged into account 1 on vehicle B, thereby enabling vehicle B to send the binding relationship between account 1 and license plate number xxx to the cloud server.
  • user 2 can enter the license plate number yyy associated with vehicle C on mobile phone C, which is logged into account 2, so that mobile phone C can send the binding relationship between account 2 and license plate number yyy to the cloud server.
  • the output of the aforementioned multimodal large model may include information about the license plate number xxx.
  • the cloud server After receiving the first output sent by the first vehicle and the license plate number xxx, the cloud server can determine to send the first output to vehicle B according to the association relationship shown in Table 1 above.
  • the output of the aforementioned multimodal large model may include information about the license plate number yyy.
  • the cloud server can determine to send the first output to mobile phone C based on the association shown in Table 1 above.
  • mobile phone C can determine whether it is located inside the cabin of vehicle C. If mobile phone C is located inside the cabin of vehicle C, then mobile phone C can provide a prompt to the user based on the first output.
  • the first vehicle sends a first output to the second vehicle or a terminal device, including: the first vehicle sending the first output to the second vehicle via near-field communication technology.
  • the vehicle 100 can send the first output to the vehicle 200 through the star flash technology.
  • the second vehicle determines the second output based on the first output and environmental information.
  • the first inference model can output the user's true intention: "Please switch the high beams to low beams.”
  • Vehicle 200 can input the user's true intention in vehicle 100 and the image data collected by the camera outside the cabin of vehicle 200 into the fourth inference model, thereby obtaining the text content: "Dear driver, you are currently using high beams, which may affect the driving of oncoming drivers. The oncoming drivers would like you to switch to low beams.”
  • the first inference model can output the user's true intention, "Please speed up in the fast lane.”
  • Vehicle 200 can input the user's true intention in vehicle 100 and the image data collected by the camera outside the cabin into the fourth inference model, thereby obtaining the text content, "Dear driver, Xiao A reminds you that you are currently in the fast lane, where the speed limit is 80-100 km/h. The vehicle is currently traveling at 65 km/h. The drivers behind you are in a hurry, please speed up.”
  • the first vehicle can be vehicle 100 and the second vehicle can be vehicle 200.
  • the first input can be the horn input from the driver inside vehicle 100.
  • Vehicle 100 can input the horn input and the data collected by sensors outside the vehicle 100's cabin into the first inference model, thereby obtaining the user's intention "Please do not cut in line". Since the sensors outside vehicle 100's cabin do not detect the cone in front of vehicle 200, vehicle 100 will mistakenly believe that the user's intention complies with traffic regulations, and can thus send the user's intention to vehicle 200.
  • vehicle 200 After receiving the user's intention, vehicle 200 can input the user's intention and the data collected by sensors outside its cabin (including data related to road construction) into the fourth inference model, thereby obtaining the text content "The driver who honked may not know about the road construction, I will relay it to him.”
  • the method 1400 further includes: the second vehicle control prompting device prompting the user that it has received the first output from the first vehicle and prompting the user whether to prompt the user; wherein, the second vehicle determining the second output based on the first output and environmental information includes: in response to detecting an input from the user indicating that they want to prompt the user, the second vehicle determines the second output based on the first output and environmental information.
  • the second vehicle determines the second output based on the first output and environmental information, including: the second vehicle determines the second output based on data collected by sensors in the vehicle cabin, the first output, and environmental information.
  • the second vehicle can output the first output and environmental information into the fourth inference model, thereby obtaining the second output.
  • the fourth inference model can be a multimodal large model.
  • This multimodal large model can be trained from a general multimodal large model.
  • this multimodal large model includes a knowledge base.
  • the knowledge base includes, but is not limited to, road regulations of one or more countries (or regions), speed limit information for road sections, and reasonable handling methods during vehicle interactions. This allows a general multimodal large model to become a dedicated multimodal large model for the vehicle interaction domain.
  • the first vehicle can be vehicle 100 and the second vehicle can be vehicle 200.
  • the first inference model can output the user's true intention: "Please switch the high beams to low beams.”
  • Vehicle 200 can input the user's true intention in vehicle 100, the data collected by sensors in the cabin (indicating that the driver in vehicle 200 is in a tense state), and the image data collected by the camera outside the cabin into the fourth inference model, thereby obtaining the text content: "Dear driver, you are currently using high beams, which may affect oncoming drivers. The oncoming driver would like you to switch to low beams. You can move the lever behind the right side of the steering wheel upwards to switch to low beams.”
  • the first vehicle can be vehicle 100 and the second vehicle can be vehicle 200.
  • the first inference model can output the user's true intention, "Please accelerate in the fast lane.”
  • Vehicle 200 can input the user's true intention in vehicle 100, the data collected by sensors in the cabin (indicating that the driver in vehicle 200 is in a pleasant state), and the image data collected by the camera outside the cabin into the fourth inference model, thereby obtaining the text content, "Dear driver, Xiao A reminds you that the driver behind you is a bit impatient, please accelerate!
  • the second vehicle determines the second output based on the first output and environmental information, including: the second vehicle determines the second output based on the first information of the driver in the second vehicle's cabin, the first output and environmental information, wherein the first information includes one or more of driving proficiency, driving habits or physiological characteristics.
  • the first vehicle can be vehicle 100 and the second vehicle can be vehicle 200.
  • the first inference model can output the user's true intention, "Please speed up in the fast lane.”
  • Vehicle 200 can input the user's true intention in vehicle 100 and the driver's driving skill level in vehicle 200 (indicating that the driver in the second vehicle is a novice driver) into the fourth inference model, thereby obtaining the text content, "Dear driver, Xiao A reminds you that you are currently in the fast lane, where the speed limit is 80-100 km/h. The vehicle is currently traveling at 65 km/h. The drivers behind you are a bit anxious, please speed up.”
  • Vehicle 200 can input the user's true intention in vehicle 100 and the driver's driving skill level in vehicle 200 (indicating that the driver in the second vehicle is a highly skilled driver) into the fourth inference model, thereby obtaining the text content "Dear driver, the driver behind you is in a hurry, please speed up”.
  • Vehicle 200 can input the user's true intention in vehicle 100 and the driver's driving habits in vehicle 200 (the frequency of driving in the slow lane in the past period of time is greater than or equal to a preset frequency) into the fourth inference model, so as to obtain the text content "Dear car owner, the driver behind is a bit anxious, you can switch to the slow lane".
  • Vehicle 200 can input the user's true intention in vehicle 100 and the physiological characteristics of the driver in vehicle 200 (indicating the driver is an elderly driver) into the fourth inference model, thereby obtaining the text content: "Dear driver, Xiao A reminds you that you are currently in the fast lane, where the speed limit is 80-100 km/h. The vehicle is currently traveling at 65 km/h. The drivers behind you are in a bit of a hurry. If you are not in a hurry, you can turn on your right turn signal and switch to the slow lane.”
  • driver information can be incorporated when determining the second output. This allows the second output to better match the driver's profile, resulting in different outputs for different driver profiles. This helps improve the vehicle's intelligence and enhances the user's driving experience.
  • the second vehicle determines a second output based on the first output and environmental information, including: the second vehicle determines the second output based on the historical driving records of other vehicles around the second vehicle, the first output, and environmental information.
  • the first output could be the user's intention, "Please do not cut in line.” Since the sensors outside vehicle 100's cabin do not detect the traffic cone in front of vehicle 200, they might mistakenly interpret the user's intention as complying with traffic regulations, and thus send the user's intention to vehicle 200.
  • vehicle 200 can input the user's intention, the historical driving records of other vehicles around vehicle 200 (indicating that other vehicles have alternately passed through the road segment over a past period), and data collected by the sensors outside the cabin (including data related to road construction) into the fourth inference model. This will result in the text content, "The driver who honked may not know about the road construction; I will relay this message to him.”
  • the historical driving records of other vehicles around the vehicle can be considered when determining the second output. This makes the second output more accurate and more acceptable to users in the cabin.
  • the second vehicle determines the second output based on the first output and environmental information, including: the second vehicle inputs the first output and environmental information into the inference model to obtain the second output.
  • Second vehicle control prompt device outputs second output.
  • the second vehicle can control the sound-emitting device in the second vehicle's cabin to play the corresponding voice information.
  • FIG. 15 shows a schematic diagram of the architecture of the system 1500 provided in an embodiment of this application.
  • the system 1500 includes a vehicle 1510 and a vehicle 1520.
  • the vehicle 1510 may include a first inference model
  • the vehicle 1520 may include a fourth inference model.
  • Vehicle 1510 can be vehicle 100, and vehicle 1520 can be vehicle 200. Alternatively, vehicle 1510 can be the first vehicle, and vehicle 1520 can be the second vehicle.
  • FIG. 16 shows a schematic block diagram of an interactive device 1600 provided in an embodiment of this application.
  • the device 1600 includes: an acquisition unit 1610, configured to acquire a first input from a user in a vehicle cabin and environmental information around the vehicle, the environmental information including information about another vehicle; a determination unit 1620, configured to determine a first output based on the first input and the environmental information; and a sending unit 1630, configured to send the first output to another vehicle or a terminal device, the terminal device being associated with the other vehicle.
  • the first input is associated with the user's negative emotions
  • the first output includes the de-emotionalized output.
  • the determining unit 1620 is used to: input the first input and environmental information into the first inference model to obtain the first output.
  • the acquisition unit 1610 is also used to acquire vehicle information; the determination unit is used to: determine a first output based on the first input, the vehicle information and the environmental information.
  • the determining unit 1620 is used to: input the first input, vehicle information and environmental information into the second inference model to obtain the first output.
  • the determining unit 1620 is also configured to: determine information about another vehicle based on the first input and environmental information.
  • the determining unit 1620 is used to: input the first input and environmental information into the third inference model to obtain information about another vehicle.
  • the information of the other vehicle includes the license plate information of the other vehicle; wherein, the sending unit 1630 is used to: send the first output and the license plate information of the other vehicle to the cloud server, so that the cloud server sends the first output to the other vehicle or terminal device based on the license plate information of the other vehicle.
  • the device further includes a control unit and a determining unit, which are further configured to determine a first control command based on the first input and environmental information, wherein the first control command is associated with one or more actuators; and the control unit is configured to control one or more actuators to execute the first control command.
  • Figure 17 shows a schematic block diagram of an interactive device 1700 provided in an embodiment of this application.
  • the device 1700 includes: an acquisition unit 1710 for acquiring a first output of another vehicle and environmental information around the vehicle; a determination unit 1720 for determining a second output based on the first output and the environmental information; and a control unit 1730 for controlling a prompting device to output the second output.
  • the determining unit 1720 is used to: determine a second output based on data collected by sensors in the vehicle cabin, a first output, and environmental information.
  • the determining unit 1720 is used to: determine a second output based on the first information of the driver in the vehicle, the first output, and environmental information, wherein the first information includes one or more of driving proficiency, driving habits, or physiological characteristics.
  • the determining unit 1720 is used to: determine a second output based on the historical driving records of other vehicles around the vehicle, the first output, and environmental information.
  • the determining unit 1720 is used to: input the first output and environmental information into the inference model to obtain the second output.
  • the division of units in the above device is only a logical functional division. In actual implementation, they can be fully or partially integrated into a single physical entity, or they can be physically separated.
  • the units in the device can be implemented by a processor calling software; for example, the device includes a processor connected to memory, which stores instructions. The processor calls the instructions stored in memory to implement any of the above methods or to implement the functions of each unit in the device.
  • the processor can be, for example, a general-purpose processor, such as a CPU or microprocessor, and the memory can be internal or external to the device.
  • the units in the device can be implemented as hardware circuits. The functions of some or all units can be implemented through the design of the hardware circuits, which can be understood as one or more processors.
  • the hardware circuit is an ASIC, and the functions of some or all units are implemented through the design of the logical relationships between the components within the circuit.
  • the hardware circuit can be implemented using a PLD, such as an FPGA, which can include a large number of logic gates. The connection relationships between the logic gates are configured through configuration files, thereby implementing the functions of some or all units. All units of the above devices can be implemented entirely through processor calling software, or entirely through hardware circuits, or partially through processor calling software with the remaining parts implemented through hardware circuits.
  • a processor is a circuit with signal processing capabilities.
  • the processor can be a circuit with instruction reading and execution capabilities, such as a CPU, microprocessor, GPU, or DSP.
  • the processor can implement certain functions through the logical relationships of hardware circuits. These logical relationships are fixed or reconfigurable.
  • the processor may be a hardware circuit implemented as an ASIC or PLD, such as an FPGA.
  • the process of the processor loading a configuration document and configuring the hardware circuit can be understood as the processor loading instructions to implement the functions of some or all of the above units.
  • it can also be a hardware circuit designed for artificial intelligence, which can be understood as an ASIC, such as an NPU, TPU, or DPU.
  • each unit in the above device can be one or more processors (or processing circuits) configured to implement the above methods, such as: CPU, GPU, NPU, TPU, DPU, microprocessor, DSP, ASIC, FPGA, or a combination of at least two of these processor forms.
  • processors or processing circuits configured to implement the above methods, such as: CPU, GPU, NPU, TPU, DPU, microprocessor, DSP, ASIC, FPGA, or a combination of at least two of these processor forms.
  • the units in the above devices can be integrated in whole or in part, or they can be implemented independently. In one implementation, these units are integrated together as a System-on-a-Chip (SoC).
  • SoC may include at least one processor for implementing any of the above methods or implementing the functions of the units in the device.
  • the at least one processor may be of different types, such as CPU and FPGA, CPU and AI processor, CPU and GPU, etc.
  • This application also provides an interactive device, which includes a processing unit and a storage unit.
  • the storage unit is used to store instructions, and the processing unit executes the instructions stored in the storage unit to cause the device to perform the methods or steps described in the above embodiments.
  • the aforementioned processing unit may be one or more of the processors 121-12n shown in FIG1.
  • This application also provides an interactive system, which includes the above-described interactive device and sensing system.
  • This application also provides a vehicle that may include the aforementioned interactive device or interactive system.
  • This application also provides a computer program product, which includes computer program code that, when run on a computer, causes the computer to perform the methods described in the above embodiments.
  • This application also provides a computer-readable medium storing program code that, when run on a computer, causes the computer to perform the methods described in the above embodiments.
  • This application also provides a chip, which includes a circuit for performing the methods described in the above embodiments.
  • each step of the above method can be completed by integrated logic circuits in the processor's hardware or by instructions in software.
  • the method disclosed in the embodiments of this application can be directly implemented by a hardware processor, or by a combination of hardware and software modules within the processor.
  • the software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, power-on erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method. To avoid repetition, detailed descriptions are omitted here.
  • the memory may include read-only memory and random access memory, and provides instructions and data to the processor.
  • the disclosed systems, apparatuses, and methods can be implemented in other ways.
  • the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods.
  • multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed.
  • the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.
  • the units described as separate components may or may not be physically separate.
  • the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
  • the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.
  • the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium.
  • This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application.
  • the aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Physics & Mathematics (AREA)
  • Computational Linguistics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Acoustics & Sound (AREA)
  • Multimedia (AREA)
  • Human Computer Interaction (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Signal Processing (AREA)
  • Psychiatry (AREA)
  • Hospice & Palliative Care (AREA)
  • Child & Adolescent Psychology (AREA)
  • Medical Informatics (AREA)
  • Mathematical Physics (AREA)
  • Biophysics (AREA)
  • Automation & Control Theory (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Transportation (AREA)
  • Mechanical Engineering (AREA)
  • Developmental Disabilities (AREA)
  • Educational Technology (AREA)
  • Psychology (AREA)
  • Social Psychology (AREA)
  • Computing Systems (AREA)
  • Artificial Intelligence (AREA)
  • Pathology (AREA)
  • Biomedical Technology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Molecular Biology (AREA)
  • Surgery (AREA)
  • Animal Behavior & Ethology (AREA)
  • Public Health (AREA)
  • Veterinary Medicine (AREA)
  • Traffic Control Systems (AREA)

Abstract

本申请提供了一种交互方法、装置和车辆，该交互方法可以应用于智能座舱领域。该方法包括：获取车辆座舱内用户的第一意图，该第一意图与另一车辆相关联且该第一意图关联该用户的负面情绪；对该第一意图进行去情绪化处理，得到第二意图；向该另一车辆发送该第二意图。本申请可以应用于电动汽车或者智能汽车中，可以避免在车辆之间发生交互时，驾驶员产生愤怒的情绪和报复行为，进而避免发生交通事故，有助于提升用户的驾乘体验。

Description

交互方法、装置和车辆
本申请要求在2024年6月28日提交中国国家知识产权局、申请号为202410869842.4、发明名称为“交互方法、装置和车辆”的中国专利申请的优先权，在2024年12月30日提交中国国家知识产权局、申请号为202411993129.7、发明名称为“交互方法、装置和车辆”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及智能座舱领域，并且更具体地，涉及一种交互方法、装置和车辆。
背景技术
车辆在道路上行驶时，会与附近行驶的车辆产生交互关系。例如，发生一系列交互，如超车、跟车、变道、加塞、急刹车或者排队等。车辆之间发生这些交互时，驾驶员之间无法感知对方的情绪，容易产生误会，进而导致驾驶员产生愤怒的情绪。更严重的情况下，可能会导致驾驶员的报复行为(例如，别车、刹停等)，导致发生交通事故。
发明内容
本申请提供一种交互方法、装置和车辆，可以避免在车辆之间发生交互时，驾驶员产生愤怒的情绪和报复行为，进而避免发生交通事故，有助于提升用户的驾乘体验。
第一方面，提供了一种交互方法，该方法包括：获取车辆座舱内用户的第一意图，该第一意图与另一车辆相关联且该第一意图关联该用户的负面情绪；对该第一意图进行去情绪化处理，得到第二意图；向该另一车辆发送该第二意图。
结合第一方面，在第一方面的某些实现方式中，该对该第一意图进行去情绪化处理，得到第二意图，包括：对该第一意图进行去情绪化处理，得到去情绪化处理后的意图；在该去情绪化后的意图中增加礼貌用语，得到该第二意图。
结合第一方面，在第一方面的某些实现方式中，该获取车辆座舱内用户的第一意图之前，该方法还包括：获取用户的第一语音指令；根据该第一语音指令，确定该第一意图；其中，该向该另一车辆发送该第二意图之前，该方法还包括：在该第一语音指令中不包括该第一意图对应的槽位信息时，根据该第一意图以及该车辆座舱外的传感器采集的数据，确定该另一车辆的信息；或者，在该第一语音指令中包括该第一意图对应的槽位信息时，根据该第一槽位信息以及该传感器采集的数据，确定该另一车辆的信息。
结合第一方面，在第一方面的某些实现方式中，该获取车辆座舱内用户的第一意图之前，该方法还包括：获取用户的第二语音指令，该第二语音指令包括该第一意图；其中，该对该第一意图进行去情绪化处理，得到第二意图，包括：将该第二语音指令、该车辆的信息、该车辆的周围环境信息输入推理模型中，得到该第二意图以及该另一车辆的信息。
结合第一方面，在第一方面的某些实现方式中，该获取车辆座舱内用户的第一意图，包括：获取该用户的驾驶行为和该另一车辆的行驶记录；根据该用户的驾驶行为和该另一车辆的行驶记录，确定该第一意图；其中，该向该另一车辆发送该第二意图之前，该方法还包括：根据该第一意图以及该车辆座舱外的传感器采集的数据，确定该另一车辆的信息。
结合第一方面，在第一方面的某些实现方式中，该向该另一车辆发送该第二意图之前，该方法还包括：确定该第二意图满足道路法规。
结合第一方面，在第一方面的某些实现方式中，该向该另一车辆发送该第二意图，包括：根据该车辆所处环境的信号强度和/或信号质量，向该另一车辆发送该第二意图。
结合第一方面，在第一方面的某些实现方式中，该根据该车辆所处环境的信号强度和/或信号质量，向该另一车辆发送该第二意图，包括：在该信号强度大于或者等于预设信号强度，和/或，信号质量大于或者等于预设信号质量时，向云端服务器发送该另一车辆的信息以及该第二意图，以使得该云端服务器根据该另一车辆的信息向该另一车辆发送该第二意图。
结合第一方面，在第一方面的某些实现方式中，该向该另一车辆发送该第二意图，包括：在该信号强度小于预设信号强度，和/或，信号质量小于预设信号质量时，通过近场通信向该另一车辆发送该第二意图。
第二方面，提供了一种交互方法，该方法包括：接收来自于车辆的第二意图，该第二意图为对该车辆座舱内的用户的第一意图进行去情绪化后得到的意图；控制提示装置向用户提示该第二意图。
结合第二方面，在第二方面的某些实现方式中，该控制提示装置向用户提示该第二意图，包括：根据该第二意图，确定第一驾驶意见；控制该提示装置向用户提示该第二意图和该第一驾驶意见。
结合第二方面，在第二方面的某些实现方式中，该控制提示装置向用户提示该第二意图，包括：根据主驾区域中用户的状态，控制提示装置向用户提示该第二意图。
结合第二方面，在第二方面的某些实现方式中，该第二意图指示异常驾驶行为，该方法还包括：控制车灯和/或车外投影信息显示第一信息，该第一信息用于向该车辆内的用户进行致歉和/或感谢。
第三方面，提供了一种交互方法，该方法包括：获取车辆座舱内用户的第一输入、该车辆的信息和该车辆周围的环境信息；根据该第一输入、该车辆的信息和该环境信息，确定第一输出；向该另一车辆发送第一输出。
结合第三方面，在第三方面的某些实现方式中，该根据该第一输入、该车辆的信息和该环境信息，确定第一输出，包括：将该第一输入、该车辆的信息和该环境信息输入推理模型中，得到该第一输出。
结合第三方面，在第三方面的某些实现方式中，该车辆的信息包括该车辆的速度、该车辆所处的道路的类型、该车辆所处车道的类型中的至少一项。
结合第三方面，在第三方面的某些实现方式中，该环境信息包括该车辆的传感器采集的数据。
第四方面，提供了一种交互装置，该交互装置包括用于执行上述第一方面至第三方面中任一项可能的实现方式的方法的单元或者模块。
第五方面，提供了一种交互装置，该交互装置包括处理单元和存储单元，其中存储单元用于存储指令，处理单元执行存储单元所存储的指令，以使该交互装置执行第一方面至第三方面中任一种可能的方法。
第六方面，提供了一种交互系统，该系统包括感知系统和计算平台，其中，该计算平台包括第四方面或者第五方面中任一种可能的装置。
第七方面，提供了一种车辆，该车辆包括第四方面或者第五方面所述的装置，或者，包括第六方面所述的系统。
第八方面，提供了一种计算机程序产品，所述计算机程序产品包括：计算机程序代码，当所述计算机程序代码在计算机上运行时，使得计算机执行上述第一方面至第三方面中任一种可能的方法。
需要说明的是，上述计算机程序代码可以全部或者部分存储在第一存储介质上，其中第一存储介质可以与处理器封装在一起的，也可以与处理器单独封装，本申请实施例对此不作具体限定。
第九方面，提供了一种计算机可读介质，所述计算机可读介质存储有程序代码，当所述计算机程序代码在计算机上运行时，使得计算机执行上述第一方面至第三方面中任一种可能的方法。
第十方面，本申请实施例提供了一种芯片系统，该芯片系统包括处理器，用于调用存储器中存储的计算机程序或计算机指令，以使得该处理器执行上述第一方面至第三方面中任一种可能的方法。
结合第十方面，在一种可能的实现方式中，该处理器通过接口与存储器耦合。
结合第十方面，在一种可能的实现方式中，该芯片系统还包括存储器，该存储器中存储有计算机程序或计算机指令。
第十一方面，本申请提供了一种交互方法，该方法包括：获取车辆座舱内用户的第一输入和车辆周围的环境信息，环境信息中包括另一车辆的信息；根据第一输入和环境信息，确定第一输出；向另一车辆或者终端设备发送第一输出，终端设备与另一车辆关联。
基于上述技术方案，通过车辆座舱内用户的输入以及环境信息，可以确定第一输出，从而可以向另一车辆或者终端设备发送第一输出。这样，可以打破两个车辆中驾驶员之间的信息隔离，实现车辆与车辆之间的交互。
在一些可能的实现方式中，该第一输入可以包括用户的语音输入。
在一些可能的实现方式中，该第一输入可以包括用户对车辆中一个或者多个部件的输入。示例性的，该一个或者多个部件包括但不限于方向盘、灯光、喇叭、加速踏板或者制动踏板。
在一些可能的实现方式中，该第一输入可以包括用户的生理特征信息。示例性的，该生理特征信息可以包括用户的血压、心率、面部表情等。
结合第十一方面，在第十一方面的某些实现方式中，第一输入关联用户的负面情绪，第一输出包括去情绪化后的输出。
基于上述技术方案，在用户的输入关联用户的负面情绪时，可以基于用户的输入和环境信息，对需要与另一车辆交互的信息进行二次加工，从而实现去情绪化处理。这样，可以使得车辆和车辆之间实现有效、友善沟通的目的，达到传递有效信息的同时不传递负面情绪的目的，有助于避免由于驾驶员的负面情绪而导致车辆和车辆之间发生冲突。
在一些可能的方式中，该第一输入包括非礼貌用语，该第一输出还包括对该非礼貌用语转换后的礼貌用语。
基于上述技术方案，在进行去情绪化处理的同时，车辆还可以对用户的非礼貌用语进行转换，从而向另一车辆发送转换后的礼貌用语。这样，有助于进一步实现车辆和车辆之间实现有效、友善沟通的目的，达到传递有效信息的同时不传递负面情绪的目的，有助于减小驾驶员之间发生冲突的概率。
在一些可能的实现方式中，该方法还包括：控制提示装置输出所述第三输出，该第三输出包括用于对座舱内的用户进行安抚的输出结果。
基于上述技术方案，在第一输入关联用户的负面情绪时，还可以基于该第一输入和环境信息，对座舱内的用户输出进行安抚的输出结果。这样，在实现车辆和车辆之间友善交互的同时，还可以对座舱内的用户进行安抚，有助于缓解用户的负面情绪，有助于提升用户的驾乘体验。
示例性的，该第三输出可以包括安抚用语。
示例性的，该第三输出可以包括对座舱内一个或者多个部件的执行指令。
结合第十一方面，在第十一方面的某些实现方式中，根据第一输入和环境信息，确定第一输出，包括：将第一输入和环境信息输入第一推理模型中，得到第一输出。
基于上述技术方案，可以将第一输入和环境信息输入第一推理模型中，从而可以得到该第一输出。这样，通过第一推理模型可以实现端到端的输入和输出。
示例性的，该第一推理模型可以为多模态模型。
结合第十一方面，在第十一方面的某些实现方式中，方法还包括：获取车辆的信息；其中，根据第一输入和环境信息，确定第一输出，包括：根据第一输入、车辆的信息和环境信息，确定第一输出。
基于上述技术方案，通过车辆座舱内用户的输入、车辆的信息以及环境信息，可以确定第一输出，从而可以向另一车辆或者终端设备发送第一输出。这样，可以打破两个车辆中驾驶员之间的信息隔离，实现车辆与车辆之间的交互。同时，通过结合车辆的信息，有助于进一步提升第一输出结果的准确性。
在一些可能的实现方式中，该车辆的信息包括车辆的历史行驶记录。
在一些可能的实现方式中，该车辆的信息包括车辆的速度、加速度、位置中的一项或者多项。
结合第十一方面，在第十一方面的某些实现方式中，根据第一输入、车辆的信息和环境信息，确定第一输出，包括：将第一输入、车辆的信息和环境信息输入第二推理模型中，得到第一输出。
基于上述技术方案，可以将第一输入、车辆的信息和环境信息输入第一推理模型中，从而可以得到该第一输出。这样，通过第二推理模型可以实现端到端的输入和输出。
在一些可能的实现方式中，该第一推理模型和该第二推理模型可以为同一模型。
结合第十一方面，在第十一方面的某些实现方式中，方法还包括：根据第一输入和环境信息，确定另一车辆的信息。
基于上述技术方案，通过车辆座舱内用户的输入以及环境信息，可以确定另一车辆的信息，从而可以向另一车辆或者终端设备发送第一输出。这样，在确定另一车辆的信息时可以结合用户的输入和环境信息，保证确定出的待交互的车辆的准确性。
结合第十一方面，在第十一方面的某些实现方式中，根据第一输入和环境信息，确定另一车辆的信息，包括：将第一输入和环境信息输入第三推理模型中，得到另一车辆的信息。
在一些可能的实现方式中，该第一推理模型、该第二推理模型和该第三推理模型可以为同一模型。
结合第十一方面，在第十一方面的某些实现方式中，另一车辆的信息包括另一车辆的车牌的信息；其中，向另一车辆或者终端设备发送第一输出，包括：向云端服务器发送第一输出和另一车辆的车牌的信息，以使得云端服务器基于另一车辆的车牌的信息，向另一车辆或者终端设备发送第一输出。
基于上述技术方案，车辆可以将第一输出和另一车辆的车牌的信息发送给云端服务器，从而使得云端服务器可以通过该另一车辆的车牌的信息，向另一车辆发送该第一输出。通过云端服务器对信息进行转发，可以打破两个车辆中驾驶员之间的信息隔离，实现车辆与车辆之间的交互。
在一些可能的实现方式中，该云端服务器中保存车辆的车牌与终端设备(例如，手机或者车辆)的标识信息之间的关联关系。
结合第十一方面，在第十一方面的某些实现方式中，向另一车辆或者终端设备发送第一输出，包括：通过近距离通信技术向另一车辆发送该第一输出。
结合第十一方面，在第十一方面的某些实现方式中，方法还包括：根据第一输入和环境信息，确定第一控制指令，第一控制指令关联一个或者多个执行器；控制一个或者多个执行器执行第一控制指令。
基于上述技术方案，通过第一输入和环境信息，还可以得到控制指令。这样，通过执行对一个或者多个执行器的控制指令，可以实现帮助用户执行相应操作，例如编辑灯语、表达谢意、表达歉意等。
第十二方面，本申请提供了一种交互方法，该方法包括：获取另一车辆的第一输出和车辆周围的环境信息；根据第一输出和环境信息，确定第二输出；控制提示装置输出第二输出。
基于上述技术方案，车辆可以基于另一车辆的输出和车辆周围的环境信息，得到第二输出并控制提示装置输出该第二输出。这样，可以打破两个车辆中驾驶员之间的信息隔离，实现车辆与车辆之间的交互。
结合第十二方面，在第十二方面的某些实现方式中，根据第一输出和环境信息，确定第二输出，包括：根据车辆座舱内的传感器采集的数据、第一输出和环境信息，确定第二输出。
基于上述技术方案，在确定第二输出时可以结合车辆座舱内的传感器采集的数据。这样，可以使得该第二输出更容易被当前座舱内用户所接受，有助于提升用户的驾乘体验。
结合第十二方面，在第十二方面的某些实现方式中，根据第一输出和环境信息，确定第二输出，包括：根据车辆中驾驶员的第一信息、第一输出和环境信息，确定第二输出，第一信息包括驾驶熟练度、驾驶习惯或者生理特征信息中的一项或者多项。
基于上述技术方案，在确定第二输出时可以结合驾驶员的信息。这样，可以使得该第二输出更符合驾驶员的画像，从而实现不同的驾驶员的画像得到不同的输出结果，有助于提升车辆的智能化程度，也有助于提升用户的驾乘体验。
结合第十二方面，在第十二方面的某些实现方式中，根据第一输出和环境信息，确定第二输出，包括：根据车辆周围其他车辆的历史驾驶记录、第一输出和环境信息，确定第二输出。
基于上述技术方案，在确定第二输出时可以结合车辆周围其他车辆的历史驾驶记录。这样，可以使得该第二输出更为准确，也可以使得第二输出更容易被座舱内的用户接受。
结合第十二方面，在第十二方面的某些实现方式中，根据第一输出和环境信息，确定第二输出，包括：将第一输出和环境信息输入推理模型中，得到第二输出。
第十三方面，本申请提供了一种交互装置，该装置包括：获取单元，用于获取车辆座舱内用户的第一输入和车辆周围的环境信息，环境信息中包括另一车辆的信息；确定单元，用于根据第一输入和环境信息，确定第一输出；发送单元，用于向另一车辆或者终端设备发送第一输出，终端设备与另一车辆关联。
结合第十三方面，在第十三方面的某些实现方式中，第一输入关联用户的负面情绪，第一输出包括去情绪化后的输出。
结合第十三方面，在第十三方面的某些实现方式中，确定单元，用于：将第一输入和环境信息输入第一推理模型中，得到第一输出。
结合第十三方面，在第十三方面的某些实现方式中，获取单元，还用于获取车辆的信息；确定单元，用于：根据第一输入、车辆的信息和环境信息，确定第一输出。
结合第十三方面，在第十三方面的某些实现方式中，确定单元，用于：将第一输入、车辆的信息和环境信息输入第二推理模型中，得到第一输出。
结合第十三方面，在第十三方面的某些实现方式中，确定单元，还用于：根据第一输入和环境信息，确定另一车辆的信息。
结合第十三方面，在第十三方面的某些实现方式中，确定单元，用于：将第一输入和环境信息输入第三推理模型中，得到另一车辆的信息。
结合第十三方面，在第十三方面的某些实现方式中，另一车辆的信息包括另一车辆的车牌的信息；其中，发送单元，用于：向云端服务器发送第一输出和另一车辆的车牌的信息，以使得云端服务器基于另一车辆的车牌的信息，向另一车辆或者终端设备发送第一输出。
结合第十三方面，在第十三方面的某些实现方式中，装置还包括控制单元，确定单元，还用于根据第一输入和环境信息，确定第一控制指令，第一控制指令关联一个或者多个执行器；控制单元，用于控制一个或者多个执行器执行第一控制指令。
第十四方面，本申请提供了一种交互装置，该装置包括：获取单元，用于获取另一车辆的第一输出和车辆周围的环境信息；确定单元，用于根据第一输出和环境信息，确定第二输出；控制单元，用于控制提示装置输出第二输出。
结合第十四方面，在第十四方面的某些实现方式中，确定单元，用于：根据车辆座舱内的传感器采集的数据、第一输出和环境信息，确定第二输出。
结合第十四方面，在第十四方面的某些实现方式中，确定单元，用于：根据车辆中驾驶员的第一信息、第一输出和环境信息，确定第二输出，第一信息包括驾驶熟练度、驾驶习惯或者生理特征信息中的一项或者多项。
结合第十四方面，在第十四方面的某些实现方式中，确定单元，用于：根据车辆周围其他车辆的历史驾驶记录、第一输出和环境信息，确定第二输出。
结合第十四方面，在第十四方面的某些实现方式中，确定单元，用于：将第一输出和环境信息输入推理模型中，得到第二输出。
第十五方面，提供了一种交互装置，该交互装置包括处理单元和存储单元，其中存储单元用于存储指令，处理单元执行存储单元所存储的指令，以使该交互装置执行第十一方面或者第十二方面中任一种可能的方法。
第十六方面，提供了一种交互系统，该系统包括感知系统和计算平台，其中，该计算平台包括第十三方面至第十五方面中任一种可能的装置。
第十七方面，提供了一种车辆，该车辆包括第十三方面至第十五方面中任一种可能的装置，或者，包括第十六方面所述的系统。
第十八方面，提供了一种计算机程序产品，所述计算机程序产品包括：计算机程序代码，当所述计算机程序代码在计算机上运行时，使得计算机执行第十一方面或者第十二方面中任一种可能的方法。
需要说明的是，上述计算机程序代码可以全部或者部分存储在第一存储介质上，其中第一存储介质可以与处理器封装在一起的，也可以与处理器单独封装，本申请实施例对此不作具体限定。
第十九方面，提供了一种计算机可读介质，所述计算机可读介质存储有程序代码，当所述计算机程序代码在计算机上运行时，使得计算机执行上述第十一方面或者第十二方面中任一种可能的方法。
第二十方面，本申请实施例提供了一种芯片系统，该芯片系统包括处理器，用于调用存储器中存储的计算机程序或计算机指令，以使得该处理器执行上述第十一方面或者第十二方面中任一种可能的方法。
结合第二十方面，在一种可能的实现方式中，该处理器通过接口与存储器耦合。
结合第二十方面，在一种可能的实现方式中，该芯片系统还包括存储器，该存储器中存储有计算机程序或计算机指令。
附图说明
图1是本申请实施例提供的车辆的功能框图示意。
图2A-2C是本申请实施例提供的交互场景的示意图。
图3是本申请实施例提供的交互场景的另一示意图。
图4是本申请实施例提供的交互场景的另一示意图。
图5是本申请实施例提供的交互场景的另一示意图。
图6A-6B是本申请实施例提供的交互场景的另一示意图。
图7是本申请实施例提供的交互场景的另一示意图。
图8A-8C是本申请实施例提供的交互场景的另一示意图。
图9是本申请实施例提供的交互方法的示意性流程图。
图10是本申请实施例提供的交互方法的另一示意性流程图。
图11是本申请实施例提供的交互方法的另一示意性流程图。
图12是本申请实施例提供的系统架构图。
图13是本申请实施例提供的另一系统架构图。
图14是本申请实施例提供的交互方法的示意性流程图。
图15是本申请实施例提供的系统架构图。
图16是本申请实施例提供的交互装置的示意性框图。
图17是本申请实施例提供的交互装置的示意性框图。
具体实施方式
下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行描述。其中，在本申请实施例的描述中，除非另有说明，“/”表示或的意思，例如，A/B可以表示A或B；本文中的“和/或”仅仅是一种描述关联对象的关联关系，表示可以存在三种关系，例如，A和/或B，可以表示：单独存在A，同时存在A和B，单独存在B这三种情况。
本申请实施例中采用诸如“第一”、“第二”的前缀词，仅仅为了区分不同的描述对象，对被描述对象的位置、顺序、优先级、数量或内容等没有限定作用。本申请实施例中对序数词等用于区分描述对象的前缀词的使用不对所描述对象构成限制，对所描述对象的陈述参见权利要求或实施例中上下文的描述，不应因为使用这种前缀词而构成多余的限制。此外，在本实施例的描述中，除非另有说明，“多个”的含义是两个或两个以上。
图1是本申请实施例提供的车辆100的一个功能框图示意。车辆100可以包括感知系统110、计算平台120、显示装置130和发声装置140，其中，感知系统110可以包括感测关于车辆100周边的环境的信息的一种或多种传感器。例如，感知系统110可以包括定位系统，定位系统可以是全球定位系统(global positioning system，GPS)，也可以是北斗系统或者其他定位系统。感知系统110还可以包括惯性测量单元(inertial measurement unit，IMU)、激光雷达、毫米波雷达、超声雷达以及摄像装置中的一种或者多种。
车辆100的部分或所有功能可以由计算平台120控制。计算平台120可以包括一个或多个处理器，例如处理器121至12n(n为正整数)，处理器是一种具有信号的处理能力的电路，在一种实现中，处理器可以是具有指令读取与运行能力的电路，例如中央处理单元(central processing unit，CPU)、微处理器、图形处理器(graphics processing unit，GPU)(可以理解为一种微处理器)、或数字信号处理器(digital signal processor，DSP)等；在另一种实现中，处理器可以通过硬件电路的逻辑关系实现一定功能，该硬件电路的逻辑关系是固定的或可以重构的，例如处理器为专用集成电路(application-specific integrated circuit，ASIC)或可编程逻辑器件(programmable logic device，PLD)实现的硬件电路，例如现场可编程门阵列(field programmable gate array，FPGA)。在可重构的硬件电路中，处理器加载配置文档，实现硬件电路配置的过程，可以理解为处理器加载指令，以实现以上部分或全部单元的功能的过程。此外，处理器还可以是针对人工智能设计的硬件电路，其可以理解为一种ASIC，例如神经网络处理单元(neural network processing unit，NPU)、张量处理单元(tensor processing unit，TPU)、深度学习处理单元(deep learning processing unit，DPU)等。此外，计算平台120还可以包括存储器，存储器用于存储指令，处理器121至12n中的部分或全部处理器可以调用存储器中的指令且执行指令，以实现相应的功能。
座舱内的显示装置130主要分为两类，第一类是车载显示屏；第二类是投影显示屏，例如抬头显示装置(head up display，HUD)。车载显示屏是一种物理显示屏，是车载信息娱乐系统的重要组成部分，座舱内可以设置有多块显示屏，如数字仪表显示屏，中控屏，副驾驶位上的乘客(也称为靠前部区域的乘客)面前的显示屏，左侧后排乘客面前的显示屏以及右侧后排乘客面前的显示屏，甚至是车窗也可以作为显示屏进行显示。抬头显示，也称平视显示系统。主要用于在驾驶员前方的显示设备(例如挡风玻璃)上显示例如时速、导航等驾驶信息。以降低驾驶员视线转移时间，避免因驾驶员视线转移而导致的瞳孔变化，提升行驶安全性和舒适性。HUD例如包括组合型抬头显示(combiner-HUD，C-HUD)系统、风挡型抬头显示(windshield-HUD，W-HUD)系统、增强现实型抬头显示系统(augmented reality HUD，AR-HUD)。应理解，HUD也可以随着技术演进出现其他类型的系统，本申请对此不作限定。
发声装置140可以为扬声器、音响或者喇叭等。
如前所述，车辆在道路上行驶时，会与附近行驶的车辆产生交互关系。例如，发生一系列交互，如超车、跟车、变道、加塞、急刹车或者排队等。车辆之间发生这些交互时，驾驶员之间无法感知对方的情绪，容易产生误会，进而导致驾驶员产生愤怒的情绪。更严重的情况下，可能会导致驾驶员的报复行为(例如，别车、刹停等)，导致发生交通事故。
本申请实施例提供了一种交互方法、装置和车辆，基于自然语言描述的信息进行车辆定位，达到信息建立连接的效果。同时，可以对用户的交互信息进行二次加工，达到有效、友善沟通的目的，可以实现传递有效信息且不传递负面情绪的目的。这样，可以避免在车辆之间发生交互时，驾驶员产生愤怒的情绪和报复行为，进而避免发生交通事故，有助于提升用户的驾乘体验。
示例性的，图2A-图2C示出了本申请实施例提供的交互场景的示意图。
如图2A所示，驾驶员A驾驶车辆100在道路上行驶的过程中发现车辆200从当前所在车道向车辆100所在的车道变道时，会控制车辆100减速。这样，可能会引起驾驶员A对车辆200座舱内的驾驶员B产生不满情绪。此时，驾驶员A可能会控制车辆100进行鸣笛，以表达不满的情绪。
车辆200座舱内的驾驶员B在发现车辆100鸣笛后，可以确认车辆100座舱内的驾驶员A产生了不满的情绪。此时，驾驶员B可以发出语音信息“帮我和后面的车主说下，抱歉啊，家里有点急事”。
车辆200在获取到用户发出的语音信息后，可以先通过自动语音识别(automatic speech recognition，ASR)模块将该语音信息转化为文本内容。车辆200可以根据该文本内容对用户想要表达的真实意图以及用户当前的情绪进行理解。
例如，通过自然语言理解(natural language understanding，NLU)引擎得到用户的意图(intention)和槽位(slot)，例如，用户的意图为“因为家里有点急事，表达歉意”、槽位为“后面的车主”以及用户当前的情绪为致歉。车辆200可以通过座舱外的摄像头采集的图像，确定车辆100的车牌信息。车辆200可以将该用户的意图以及车辆100的车牌信息发送给云端服务器。云端服务器可以基于该车辆100的车牌信息查找车辆100的地址，且根据该车辆100的地址，向车辆100发送该用户的意图。同时，响应于得到用户的意图，车辆200可以向用户发出语音信息“好的，我这就和他说”。
可选地，车辆200可以通过座舱外的传感器采集的数据确定车辆200的异常行驶记录以及车辆100的行驶记录。例如，车辆200检测到从右侧车道变更至左侧车道的过程中，车辆100进行了急刹车(例如，车辆100的减速度大于或者等于预设减速度)并且检测到车辆100发出鸣笛或者切换近光灯和远光灯等，从而可以确定车辆100对车辆200进行了非主动礼让。车辆200在获取到用户的意图后，可以确定用户的意图中包括表达歉意的内容。根据车辆200的异常行驶记录和车辆100的行驶记录，车辆200除了向云端服务器发送该用户的意图以外，还可以向云端服务器发送感谢的内容(例如，发送爱心，感谢他的礼让)。此时，车辆200可以向用户发出语音信息“好的，我这就和他说，顺便会代你给他发送爱心哈”。
车辆100从云端服务器接收到来自于车辆200的用户的意图以及感谢的内容后，生成表达歉意的文本内容“亲爱的车主，前车司机向您表示歉意，家里有点急事”和表达感谢的文本内容“他向您发来了爱心，感谢您的礼让”。
可选地，车辆100也可以通过座舱外的传感器采集的数据确定车辆200的异常行驶记录(例如，车辆200在向左侧车道变更时未开启左转向灯)和车辆100的行驶记录(例如，车辆100进行了急刹车以及在急刹车后进行鸣笛)。车辆100可以基于车辆200的异常行驶记录和车辆100的行驶记录，确定车辆100内的驾驶员A当前比较气愤，车辆100还可以对该文本内容进行二次加工。例如，除了向车辆100中的驾驶员A表达歉意以及感谢以外，还可以增加情绪安抚的文本内容(例如，莫生气)。
经过上述二次加工的过程，车辆100可以通过文本到语音(text to speech，TTS)模块，将二次加工后的文本内容转化为语音信息，并通过车辆100座舱内的扬声器向驾驶员A发出语音信息“亲爱的车主，前车司机向您表示歉意，家里有点急事，莫生气莫生气。他向您发来了爱心，感谢您的礼让”。
一个实施例中，车辆200检测到从右侧车道变更至左侧车道的过程中，车辆100进行了急刹车(例如，车辆100的减速度大于或者等于预设减速度)并且检测到车辆100发出鸣笛或者切换近光灯和远光灯等，从而可以确定车辆100对车辆200进行了非主动礼让。此时，车辆200在未接收到驾驶员B的语音信息的情况下，也可以基于车辆200的异常行驶记录以及车辆100的行驶记录，生成并向云端服务器发送车辆100的车牌信息、致歉的文本内容“向后车的车主致歉”以及表达感谢的文本内容“感谢您的礼让”。云端服务器可以根据车辆100的车牌信息，查找到车辆100的地址并基于该地址，向车辆100发送该致歉的文本内容和表达感谢的文本内容。车辆100可以基于车辆200的异常行驶记录和车辆100的行驶记录，确定车辆100内的驾驶员A当前比较气愤，车辆100还可以对该文本内容进行二次加工。例如，除了向车辆100中的驾驶员A表达歉意以及感谢以外，还可以增加情绪安抚的文本内容(例如，莫生气)。车辆100可以座舱内的扬声器向驾驶员A发出语音信息“亲爱的车主，前车司机向您表示歉意，家里有点急事，莫生气莫生气。他向您发来了爱心，感谢您的礼让”。
如图2B所示，驾驶员A驾驶车辆100在道路上行驶的过程中发现车辆200从当前所在车道向车辆100所在的车道变道时，出于礼让的考虑，驾驶员A会控制车辆100减速，从而对车辆200进行礼让。
车辆200座舱内的驾驶员B在发现车辆100主动礼让后，可以发出语音信息“帮我和后面的车主说谢谢”。车辆200在获取到用户发出的语音信息后，可以先通过ASR模块将该语音信息转化为文本内容。车辆200可以根据该文本内容对用户想要表达的真实意图以及用户当前的情绪进行理解。
例如，车辆200可以确定用户的意图为“表达感谢”、槽位为“后面的车主”以及用户当前的情绪为感谢。响应于得到用户的意图和槽位，车辆200可以向用户发出语音信息“好的，我这就和他说”。同时，车辆200可以通过座舱外的摄像头采集的数据，确定车辆100的车牌号的信息。车辆200可以将该用户的意图以及车辆100的车牌信息发送给云端服务器。云端服务器可以基于该车辆100的车牌信息确定车辆100的地址且根据该车辆100的地址，向车辆100发送该用户的意图。
可选地，车辆200可以通过座舱外的传感器采集的数据确定车辆200的异常行驶记录，例如，车辆200检测到从右侧车道变更至左侧车道的过程中，车辆100进行了缓慢刹车并且未检测到车辆100发出鸣笛以及切换近光灯和远光灯等，从而可以确定车辆100对车辆200进行了主动礼让。车辆200在获取到用户的意图后，可以确定用户的意图中包括表达感谢的内容。根据车辆200的异常行驶记录和车辆100的行驶记录，车辆200还可以在该用户的意图中增加其他感谢的内容(例如，发送爱心)。此时，车辆200可以向用户发出语音信息“好的，我这就和他说，顺便会代你给他发送爱心哈”。
车辆100在接收到该用户的意图时，可以根据该用户的意图，生成表达感谢的文本内容“亲爱的车主，前车车主感谢您的主动礼让并向您发送爱心”。车辆100可以通过TTS模块，将该文本内容转化为语音信息且通过车辆100座舱内的扬声器向驾驶员A发出语音信息“亲爱的车主，前车车主感谢您的主动礼让并向您发送爱心”。
如图2C所示，驾驶员A驾驶车辆100在道路上行驶的过程中发现车辆200从当前所在到的车道向车辆100所在的车道变道时，会控制车辆100减速。这样，可能会引起驾驶员A对车辆200座舱内的驾驶员B产生不满情绪。此时，驾驶员A会控制车辆100进行鸣笛，以表达不满的情绪。
车辆200座舱内的驾驶员B在发现车辆100鸣笛后，发出语音信息“不就加个塞吗，给你道歉不就完了吗”。车辆200在获取到用户发出的语音信息后，可以先通过ASR模块将该语音信息转化为文本内容。车辆200可以根据该文本内容对用户想要表达的真实意图以及用户当前的情绪进行理解。
例如，通过NLU引擎得到用户的意图为“由于加塞行为，道歉不就完了吗”以及表达该用户的意图的情绪为不满或者烦躁情绪。车辆200可以通过座舱外的传感器采集的数据确定车辆200的异常行驶记录，例如，车辆200检测到从右侧车道变更至左侧车道的过程中，车辆100进行了急刹车(例如，车辆100的减速度大于或者等于预设减速度)并且检测到车辆100发出鸣笛或者切换近光灯和远光灯等，从而可以确定车辆100对车辆200进行了非主动礼让。车辆200在获取到用户的意图后，可以基于车辆200的异常行驶记录和车辆100的行驶记录，确定该槽位信息为车辆100。车辆200可以通过座舱外摄像头采集的图像确定车辆100的车牌信息。
可选地，车辆200可以对用户的意图进行编辑，例如，将用户的意图进行去情绪化表达，将不满或者烦躁情绪转化为诚恳致歉情绪。例如，进行去情绪化表达后得到用户的意图为“由于加塞行为，向后车车主致歉”。
可选地，根据车辆200的异常行驶记录和车辆100的行驶记录，车辆200还可以在该用户的意图中增加其他感谢的内容(例如，发送爱心)。此时，车辆200可以向用户发出语音信息“亲爱的车主，已经向后车车主表达歉意并向他发送了爱心”。
车辆100接收到来自于车辆200的用户的意图以及其他感谢的内容后，生成表达感谢的文本内容“亲爱的车主，前车车主感谢您的礼让并向您发来了爱心”。车辆100可以通过TTS模块将该文本内容转化为语音信息且通过车辆100座舱内的扬声器向驾驶员A发出对应的语音信息。
示例性的，图3示出了本申请实施例提供的交互场景的示意图。
如图3所示，驾驶员A驾驶车辆100在夜间行驶的过程中发现对向的车辆200开启了远光灯时，会影响驾驶员A的驾驶。这样，可能会引起驾驶员A产生不满情绪。此时，驾驶员A发出语音信息“对面这个车能不能把远光灯关了啊”。车辆100在获取到用户发出的语音信息后，可以先通过ASR模块将该语音信息转化为文本内容。车辆200可以根据该文本内容对用户想要表达的真实意图以及用户当前的情绪进行理解。
例如，通过NLU引擎得到用户的意图为“把远光灯关了啊”、槽位为“对面的车辆”以及用户当前的情绪为不满或者烦躁。车辆100在获取到用户的意图后，可以对用户的意图进行编辑，例如，将用户的意图进行去情绪化处理，得到去情绪化后用户的意图为“关闭远光灯”，或者，车辆100还可以在该用户的意图中增加礼貌用语，如增加礼貌用语后的用户的意图为“请关闭远光灯”。
可选地，车辆100还可以对用户的意图进行翻译。例如，去情绪化处理后得到的用户的意图为“关闭远光灯”。经过翻译后可以得到用户的真实意图为“将远光灯切换为近光灯”。
同时，车辆100可以根据该槽位信息，获取与车辆100相向行驶的车辆200的信息。示例性的，车辆100可以通过座舱外的摄像头采集的图像，确定车辆200的车牌信息。车辆100可以将该用户的真实意图以及车辆200的车牌信息发送给云端服务器。云端服务器可以基于该车辆200的车牌信息查找车辆200的地址，且根据该车辆200的地址，向车辆200发送去情绪化且翻译后得到的用户的真实意图。
车辆200接收到用户的真实意图后，可以对该用户的真实意图进行转化。例如，车辆200可以将该用户的意图转化为驾驶员B易于接受的表达方式同时给驾驶员B进行有效的建议。
例如，车辆200在检测到驾驶员B为新手驾驶员时，可以向驾驶员B介绍关闭远光灯的原因(例如，影响对面司机行驶)以及驾驶建议(建议切换为近光灯)。例如，车辆200可以向驾驶员B发出语音信息“亲爱的车主，您现在开的是远光灯，可能影响对面司机行驶了，对面的车主希望您切换成近光灯哦”。
又例如，车辆200在检测到驾驶员B为经验比较丰富的驾驶员时，可以向驾驶员B发出较为简洁的表达方式，如发出语音信息“对向来车，请切换至近光灯”。
示例性的，图4示出了本申请实施例提供的交互场景的示意图。
如图4所示，驾驶员A驾驶车辆100在超车道上行驶的过程中发现车辆200的行驶速度较慢(例如，65kph)。这样，可能会引起驾驶员A对车辆200座舱内的驾驶员B产生不满情绪。此时，驾驶员A可以发出语音信息“帮我和前面这辆车说下，开这么慢就不要占快车道了，妨碍别人了啊”。车辆100在获取到用户发出的语音信息后，可以先通过ASR模块将该语音信息转化为文本内容。车辆100可以根据该文本内容对用户想要表达的真实意图以及用户当前的情绪进行理解。
例如，通过NLU引擎得到用户的意图为“转达开这么慢就不要占快车道了”、槽位为“前面的车辆”以及用户当前的情绪为不满或者烦躁。车辆100在获取到用户的意图后，可以对用户的意图进行编辑，例如，将用户的意图进行去情绪化，得到去情绪化后的用户的意图，如“转达当前在快车道内的速度过慢”。同时，车辆100可以向用户发出语音信息“好的呢，已经和对方表达了呢”。
可选地，车辆100还可以对用户的意图进行翻译。例如，去情绪化处理后得到的用户的意图为“向车辆转达当前在快车道内的速度过慢”。经过翻译后可以得到用户的真实意图为“向车辆转达在快车道内提速或者切换到慢车道”。
车辆100可以根据该槽位信息，获取位于车辆100前方的车辆200的信息。车辆100可以通过座舱外的摄像头采集的数据，确定车辆200的车牌信息。车辆100可以将去情绪化后的用户的意图以及车辆200的车牌号的信息发送给云端服务器。云端服务器可以基于该车辆200的车牌号的信息查找车辆200的地址，且根据该车辆200的地址，向车辆200发送去情绪化且翻译后得到的用户的真实意图。
车辆200从云端服务器接收到来自于车辆100的用户的真实意图后，可以对该用户的真实意图进行转化。例如，车辆200可以将该用户的意图转化为驾驶员B易于接收的表达方式。
例如，车辆200在检测到车辆200行驶在当前道路的次数小于或者等于预设次数，或者，车辆200在检测到驾驶员B为新手驾驶员时，车辆200可以向驾驶员B介绍关于当前所处的快车道的限速信息以及将该用户的意图转化为驾驶员B易于接受的表述方式。例如，车辆200可以向驾驶员B发出语音信息“亲爱的车主，小A提醒您，您现在在快车道上，限速80-100km/h，车辆现在行驶速度是65km/h，后面的司机有点着急，请您提速”。
又例如，车辆200在检测到车辆200行驶在当前道路的次数大于预设次数，或者，车辆200检测到驾驶员B为经验丰富的驾驶员时，车辆200可以将该用户的意图转化为驾驶员B易于接受的表述方式。例如，可以向驾驶员B发出较为简洁的表达方式，如语音信息“后方车主希望您提速或者更换车道”。
示例性的，图5示出了本申请实施例提供的交互场景的示意图。
如图5所示，驾驶员A驾驶车辆100在所在车道上行驶的过程中发现车辆200正在压着车道线行驶。这样，可能会引起驾驶员A对车辆200座舱内的驾驶员B产生不满情绪。此时，驾驶员A可以发出语音信息“这人怎么回事，怎么开的车”。车辆100在获取到用户发出的语音信息后，可以先通过ASR模块将该语音信息转化为文本内容。车辆100可以根据该文本内容对用户想要表达的真实意图以及用户当前的情绪进行理解。
例如，通过NLU引擎得到用户的意图为“抱怨异常驾驶行为”以及用户当前的情绪为抱怨或者生气。车辆100在获取到用户的意图以及当前的情绪后，可以向用户发出情绪安抚的语音信息(例如，“别生气啦，不值得动怒”)以及向用户推荐执行某些操作以进行情绪安抚(例如，询问用户是否播放轻音乐或者开启香氛等)。例如，车辆100可以通过座舱内的扬声器发出语音信息“亲爱的车主，别生气啦，不值得动怒，我可以给您放一点轻音乐，怎么样？”。
可选地，车辆100可以根据用户的意图以及座舱外传感器采集的数据，确定具有异常行驶记录的车辆200的信息，例如，车辆100通过传感器采集的数据确定车辆200正在压车道线行驶且获取车辆200的车牌信息。车辆100可以根据该异常驾驶记录，对用户的意图进行翻译。例如，翻译后的用户的真实意图为“请前车车主控制车辆行驶在车道内”。车辆100可以向云端服务器发送该用户的真实意图以及车辆200的车牌信息。云端服务器可以基于该车辆200的车牌信息确定车辆200的地址且根据该车辆200的地址，向车辆200发送该用户的真实意图。
车辆200可以从云端服务器来自于车辆100的用户的真实意图后，可以生成易于驾驶员B接受的表达方式。
例如，车辆200在检测到驾驶员B为新手驾驶员时，可以向驾驶员B介绍当前的异常驾驶行为以及驾驶建议。例如，车辆200可以向驾驶员B发出语音信息“亲爱的车主，您现在正在压着车道线行驶，可能影响对面司机行驶了，请尽快驾驶车辆行驶在车道中”。
又例如，车辆200在检测到驾驶员B为经验比较丰富的驾驶员时，可以向驾驶员B发出较为简洁的表达方式，如发出语音信息“注意，压线行驶中”。
示例性的，图6A-图6B示出了本申请实施例提供的交互场景的示意图。
如图6A所示，车辆200处于智能驾驶状态且在前方掉头行驶。车辆200可以向通过座舱外传感器(例如，位于车辆200尾部的摄像头)采集的数据，确定车辆100位于车辆200的后方且获取车辆100的车牌信息。此时，车辆200可以向云端服务器发送车辆100的车牌信息以及指示信息1，该指示信息1用于指示车辆200处于智能驾驶状态且即将在前方掉头行驶。云端服务器可以基于车辆100的车牌信息确定车辆100的地址并基于该地址，向车辆100发送该指示信息1。
车辆100在接收到云端服务器发送的来自于车辆200的指示信息1后，可以生成提示信息，该提示信息可以包括车辆200的状态、以及提示驾驶员A注意的内容。例如，车辆100可以通过TTS模块，将该提示信息转化为语音信息并通过座舱内的扬声器播放该语音信息“亲爱的车主，前方车辆处于智能驾驶状态，想要掉头，我们要注意哈”。
如图6B所示，车辆100处于智能驾驶状态。车辆200内的驾驶员发现车辆100的行驶轨迹异常时发出语音信息“这车怎么开的啊”并通过鸣笛表达不满。车辆200可以基于车辆200的信息以及周围车辆的行驶信息，确定车辆100的行驶轨迹异常(例如，横向移动的频次过高)。车辆200可以基于车辆200的信息和车辆100的行驶信息，获取用户的意图“请前方车辆不要总是横向移动”。车辆200可以将用户的意图以及车辆100的车牌信息发送给云端服务器。云端服务器可以基于车辆100的车牌信息确定车辆100的地址并基于该地址，向车辆100发送用户的意图。
响应于接收到该用户的意图，车辆100可以基于座舱外传感器(例如，位于车辆200尾部的摄像头)采集的数据以及来自于车辆200的用户的意图，确定车辆200位于车辆100的后方且获取车辆200的车牌信息。此时，车辆100可以向云端服务器发送车辆200的车牌信息以及指示信息2，该指示信息2用于指示车辆100处于智能驾驶状态。云端服务器可以基于车辆200的车牌信息确定车辆200的地址并基于该地址，向车辆200发送该指示信息2，该指示信息2中包括车辆100处于智能驾驶状态的指示信息以及致歉信息。
车辆200在接收到云端服务器发送的来自于车辆100的指示信息2后，可以生成提示信息，该提示信息可以包括车辆100的状态、以及提示驾驶员注意的内容。例如，车辆200可以通过TTS模块，将该提示信息转化为语音信息并通过座舱内的扬声器播放该语音信息“亲爱的车主，前方车辆处于智能驾驶状态，还请您见谅哈”。
以上实施例中是通过ASR模块和NLU模块来分析用户的意图和槽位信息为例进行说明的，本申请实施例对此并不作具体限定。例如，还可以通过将用户的输入(例如，语音输入)、车辆的信息以及车辆周围的环境信息输入推理模型(或者，称之为大模型或者多模态模型)中，输出另一车辆的信息(例如，车牌信息)以及向另一车辆传达的信息。车辆可以将该另一车辆的信息以及向另一车辆传达的信息发送给云端服务器。从而云端服务器可以通过另一车辆的信息，向该另一车辆发送该传达的信息。
示例性的，图7示出了本申请实施例提供的交互场景的示意图。
如图7所示，驾驶员A驾驶车辆100在道路上行驶的过程中发现车辆200从当前所在到的车道向车辆100所在的车道变道时，会控制车辆100减速。这样，可能会引起驾驶员A对车辆200座舱内的驾驶员B产生不满情绪。此时，驾驶员A会控制车辆100进行鸣笛，以表达不满的情绪。
车辆200在检测到车辆100鸣笛后，车辆200可以通过传感器采集的数据确定周围的环境信息。例如，车辆200通过周围的环境信息确定前方由于施工改道，导致驾驶员决策向左变更车道。在确定施工改道场景下左侧车道和右侧车道内的车辆需要交替通行后，车辆200可以通过座舱内的扬声器发出语音信息“鸣笛的车主可能不知道占道施工的情况，我这就向他转达”。
车辆200可以通过座舱外摄像头采集的图像确定车辆100的车牌信息。从而车辆200可以向云端服务器发送该车辆100的车牌信息以及指示信息3，该指示信息3用于指示车辆200当前所在的车道正在占道施工。
车辆100接收到云端服务器发送的来自于车辆200的指示信息3后，可以根据指示信息3、车辆200的异常行驶记录和车辆100的行驶记录，确定驾驶员A鸣笛可能是不清楚右侧车道内的占道施工情况，或者，驾驶员A可能知晓右侧占道施工的情况且不知晓在该场景下需要交替通行。车辆100可以向用户转达当前场景下的驾驶意见，例如，通过座舱内的扬声器发出语音信息“亲爱的车主，右侧车道正在占道施工，目前需要交替通行，不要太着急哦！”。
可选地，车辆100可以根据驾驶员A的信息，生成易于驾驶员A接受的表达方式。
例如，车辆100在检测到驾驶员A为新手驾驶员时，可以向驾驶员A介绍车辆200未异常驾驶以及驾驶建议。例如，车辆100可以向驾驶员A发出语音信息“亲爱的车主，右侧车道正在占道施工，目前需要交替通行，不要太着急哦！”。
又例如，车辆200在检测到驾驶员B为经验比较丰富的驾驶员时，可以向驾驶员A发出较为简洁的表达方式，如发出语音信息“注意，右侧占道施工中”。
示例性的，图8A-图8C示出了本申请实施例提供的交互场景的示意图。
如图8A所示，用户在驾驶车辆100行驶到窄道时发现前方车辆停在该窄道上，导致车辆100无法通过。此时，车辆100中的驾驶员可以发出语音信息“能不能和前面这个人说下，把车移开，他停在这里让别人怎么过啊”。车辆100在采集到该语音信息时，可以将该语音信息、车辆100的信息(例如，车辆的速度，所处道路的信息)和车辆100座舱外的环境信息输入推理模型中，从而可以得到推理结果，该推理结果中可以包括向车辆100的驾驶员转达的语音信息“好的呢，已经和对方表达了呢”、车辆200的车牌信息以及向车辆200传达的第一信息(例如，文本内容“您的车好像挡住后面来车了，后面的司机有点着急通过”)。
本申请实施例中，该推理模型的输入可以包括语音、图片、文字、视频流(图片和视频流可以为座舱内的传感器和座舱外的传感器采集的图像或者视频流)中的一项或者多项。
可选地，该推理模型可以外挂知识(例如，道路法规、路况，黄金处理方法(最好的方式))。这样可以让通用的多模态模型变成用户驾驶车辆时的专属多模态模型；该推理模型的输出可以为用户的意图以及另一车辆的信息。
根据当前的场景，搜索推理模型中的外挂知识(例如，车内打游戏(赛车类)，实际驾驶场景)，可以搜索不同的知识。例如，用户在车内通过副驾屏打游戏，发出语音指令“前面那辆车好慢啊”，可以将该语音信息、副驾屏的状态输入推理模型，从而可以输出文本或者语音“超过它！”。又例如，用户在城区道路中行驶前车较慢自车鸣笛，发出语音指令“前面那辆车好慢呀”，可以将该语音信息、车辆当前的速度以及所处的车道、车辆周围的环境信息输入推理模型中，从而可以输出语音信息“这段路限速50-70km/h，前车正在正常行驶，请您不要着急哦！”。
推理模型还可以输出车辆内部中其他设备的执行信息(例如，语气、语调、动效、颜色等)，例如，在输出语音信息“这段路限速50-70km/h，前车正在正常行驶，请您不要着急哦！”时，还可以输出控制发声装置播放轻音乐的指令。
车辆100可以将车辆200的车牌信息和向车辆200传达的第一信息发送给云端服务器。云端服务器可以根据该车辆200的车牌信息以及保存的车牌信息与账号的关联关系，确定车辆200的车牌信息对应的账号下包括多个设备，例如，车辆200、手机、智能手表和平板。云端服务器可以向该账号下的多个设备发送该第一信息。
如图8B所示，车辆200登录的账号下的手机接收到该第一信息时，可以通过手机的显示界面显示提示框，该提示框中包括提示信息“亲爱的车主，您的车辆好像挡住后面来车了，后面的司机有点着急通过呢，是否远程控制车辆？”、确定控件和取消控件。在检测到用户点击确定控件时，可以显示车辆200周围的环境信息和提示框，该环境信息中包括虚线框，该虚线框中为建议移动到的区域，该提示框中包括提示信息“建议您将车辆移动至虚线框中，以方便后方来车通过”、确定控件和取消控件。在检测到用户点击确定控件时，手机可以向车辆发送控制指令，该控制指令用于指示车辆200从当前位置移动至虚线框中的区域，同时手机也可以显示车辆200向该虚线框中移动的过程。在手机检测到用户发出语音信息“请帮我向后面的司机表达歉意”时，手机可以将该语音信息输入推理模型中，从而可以得到向后车驾驶员表达的文本内容“已经将车移开且向您表达歉意”。云端服务器可以将该文本内容转发给车辆100，车辆100可以将该文本内容、车辆的信息和车辆周围的环境信息输入推理模型中，从而得到向车辆100的驾驶员播报的语音信息“亲爱的车主，前车司机已经将车移开并向您表达了歉意，耽误了您宝贵的时间”。
如图8C所示，车辆100可以控制扬声器发出语音信息“亲爱的车主，前车司机已经将车移开并向您表达了歉意，耽误了您宝贵的时间”。
图9示出了本申请实施例提供的交互方法900的示意性流程图。该方法900包括：
S910，获取车辆座舱内用户的第一意图，所述第一意图与另一车辆相关联且所述第一意图关联所述用户的负面情绪。
例如，该用户的第一意图可以是通过语音指令分析得到的，例如，图3中语音信息“对面这个车能不能把远光灯关了啊”。又例如，如图4中语音信息“能不能和前面这辆车说下，开这么慢不要占快车道了，妨碍别人了啊”。
又例如，该用户的第一意图可以是通过也可以是基于车辆的驾驶记录(例如，自车驾驶员频繁切换近光灯和远光灯，急刹车，突然打方向盘)和对端车辆的驾驶记录(对端车辆此时为远光灯)得到的。
又例如，该用户的第一意图可以是通过车辆的驾驶记录(在快车道鸣笛)和对端车辆的驾驶记录(例如，在快车道的速度过慢)确定的。
又例如，该用户的第一意图还可以通过用户的动作(例如，用户的肢体动作)和表情(例如，烦躁的表情)。
又例如，该用户的第一意图还可以通过红外传感器和热成像传感器确定脸部的信息(例如，脸部变红)确定。
又例如，如果驾驶员的情绪不外露，该第一意图还可以通过座舱内其他成员的情绪得到。例如，其他成员发出语音信息“对面那个车能不能把远光灯关掉啊”。
又例如，该第一意图还可以通过用户的手表检测到的心率或者血压等信息得到。
又例如，该第一意图还可以通过方向盘的握紧力度确定。
S920，对所述第一意图进行去情绪化处理，得到第二意图。
例如，在确定用户的第一意图后，可以对该第一意图进行去情绪化处理，从而得到去情绪化后的第二意图。
S930，向所述另一车辆发送所述第二意图。
本申请实施例中，在用户发出了带有负面情绪的意图，对该负面情绪进行去情绪化处理，得到第二意图，向另一车辆发送去情绪化处理后的意图。通过对用户的意图进行二次加工，不传递负面情绪。这样，实现在车车之间存在异常交互时，可以达到有效、友善沟通的目的，达到传递有效信息。
可选地，所述对所述第一意图进行去情绪化处理，得到第二意图，包括：对所述第一意图进行去情绪化处理，得到去情绪化处理后的意图；在所述去情绪化后的意图中增加礼貌用语，得到所述第二意图。
本申请实施例中，可以在意图中增加礼貌用语，这样进一步实现有效、友善沟通的目的，避免驾驶员之间的摩擦，避免交通事故的产生。此外，语言风格的调整，更有利于沟通。
本申请实施例中识别用户的意图可以通过ASR模块和NLU模块实现。在NLU识别到用户的意图时可以进行去情绪化处理。
可选地，所述获取车辆座舱内用户的第一意图之前，所述方法还包括：获取用户的第一语音指令；根据所述第一语音指令，确定所述第一意图；其中，所述向所述另一车辆发送所述第二意图之前，所述方法还包括：在所述第一语音指令中不包括所述第一意图对应的槽位信息时，根据所述第一意图以及所述车辆座舱外的传感器采集的数据，确定所述另一车辆的信息；或者，在所述第一语音指令中包括所述第一意图对应的槽位信息时，根据所述第一槽位信息以及所述传感器采集的数据，确定所述另一车辆的信息。
本申请实施例中，通过语音指令获取到用户的意图后，如果包括槽位信息，那么可以基于槽位信息以及另一车辆的行驶记录，确定另一车辆的信息(可以是车牌信息，车身颜色，品牌中的至少一项)；如果不包括槽位信息，那么可以基于用户的意图以及传感器采集的数据，确定另一车辆的信息(可以是车牌信息，车身颜色，品牌中的至少一项)。
本申请实施例中识别用户的意图还可以通过推理模型(例如，多模态大模型)实现。
可选地，所述获取车辆座舱内用户的第一意图之前，所述方法还包括：获取用户的第二语音指令，所述第二语音指令包括所述第一意图；其中，所述对所述第一意图进行去情绪化处理，得到第二意图，包括：将所述第二语音指令、所述车辆的信息、所述车辆的周围环境信息输入推理模型中，得到所述第二意图以及所述另一车辆的信息。
可选地，所述获取车辆座舱内用户的第一意图，包括：获取所述用户的驾驶行为和所述另一车辆的行驶记录；根据所述用户的驾驶行为和所述另一车辆的行驶记录，确定所述第一意图；其中，所述向所述另一车辆发送所述第二意图之前，所述方法还包括：根据所述第一意图以及所述车辆座舱外的传感器采集的数据，确定所述另一车辆的信息。
本申请实施例中，可以基于用自车内用户的驾驶行为以及另一车辆的行驶记录，确定用户的意图。例如，自车检测到用户长按喇叭，或者频繁切换近光灯和远光灯等，可以确定他车有异常驾驶行为，从而可以确定的用户的意图(意图一般为指示他车去除异常驾驶行为)。
可选地，所述向所述另一车辆发送所述第二意图之前，所述方法还包括：确定所述第二意图满足道路法规。
本申请实施例中，在确定第一意图满足道路法规，例如图7占道施工的场景，用户希望他车不要加塞其实是不符合道路法规的(占道施工场景下应该交替通行)。
以上还可以通过推理模型判断用户的语音指令是否合法，如果合法，推理模型可以判断该如何跟对方沟通并给出输出结果。判断是否合法可以作为推理模型的内部实现过程。
可选地，所述向所述另一车辆发送所述第二意图，包括：根据所述车辆所处环境的信号强度和/或信号质量，向所述另一车辆发送所述第二意图。
可选地，所述根据所述车辆所处环境的信号强度和/或信号质量，向所述另一车辆发送所述第二意图，包括：在所述信号强度大于或者等于预设信号强度，和/或，信号质量大于或者等于预设信号质量时，向云端服务器发送所述另一车辆的信息以及所述第二意图，以使得所述云端服务器根据所述另一车辆的信息向所述另一车辆发送所述第二意图。
可选地，所述向所述另一车辆发送所述第二意图，包括：在所述信号强度小于预设信号强度，和/或，信号质量小于预设信号质量时，通过近场通信向所述另一车辆发送所述第二意图。
图10示出了本申请实施例提供的交互方法1000的示意性流程图。该方法1000包括：
S1010，接收来自于车辆的第二意图，所述第二意图为对所述车辆座舱内的用户的第一意图进行去情绪化后得到的意图。
S1020，控制提示装置向用户提示所述第二意图。
可选地，所述控制提示装置向用户提示所述第二意图，包括：根据所述第二意图，确定第一驾驶意见；控制所述提示装置向用户提示所述第二意图和所述第一驾驶意见。
示例性的，接收端设备可以基于用户的状态，调节发出语音信息的语音、语调或者语气的至少一项。例如，驾驶员当前脾气烦躁的，可以通过幽默的方式转达该第二意图，缓解紧张气氛。又例如，驾驶员当前心情比较好，可以直截了当得告知该第二意图。
可选地，所述控制提示装置向用户提示所述第二意图，包括：根据主驾区域中用户的状态，控制提示装置向用户提示所述第二意图。
可选地，所述第二意图指示异常驾驶行为，所述方法还包括：控制车灯和/或车外投影信息显示第一信息，所述第一信息用于向所述车辆内的用户进行致歉和/或感谢。
图11示出了本申请实施例提供的交互方法1100的示意性流程图。该方法1100包括：
S1110，获取车辆座舱内用户的第一输入、所述车辆的信息和所述车辆周围的环境信息。
示例性的，该第一输入可以为用户发出的语音信息、用户的动作、用户的表情。这里的用户可以是主驾区域的用户，或者，也可以是其他区域中的用户。
示例性的，车辆的信息可以包括车辆的速度、车辆的执行器的操作(例如，频繁切换近光灯和远光灯，急刹车，方向盘转角变化率过大，鸣笛、方向盘上的握紧力度等)、位置、所处车道的类型、所处道路的类型、座舱内的传感器采集的数据(例如，通过红外传感器和热成像传感器采集的数据)中的一项或者多项。
示例性的，车辆周围的环境信息包括车辆座舱外的传感器采集的数据，例如座舱外的摄像头采集的图像、视频流等。
可选地，该第一输入还包括用户的电子设备(例如，手机、智能手表或者智能手环)的输入。例如，通过用户的手表检测到的心率或者血压等。
S1120，根据所述第一输入、所述车辆的信息和所述环境信息，确定第一输出。
可选地，该第一输出包括向另一车辆传达的信息(例如，文本内容或者语音信息)。
可选地，该方法还包括：根据所述第一输入、所述车辆的信息和所述环境信息，输出车辆座舱内执行器(例如，空调、氛围灯、香氛或者车载音箱)的控制指令等。
可选地，所述根据所述第一输入、所述车辆的信息和所述环境信息，确定第一输出，包括：将所述第一输入、所述车辆的信息和所述环境信息输入推理模型中，得到所述第一输出。
可选地，所述车辆的信息包括所述车辆的速度、所述车辆所处的道路的类型、所述车辆所处车道的类型中的至少一项。
S1130，向所述另一车辆发送第一输出。
可选地，所述环境信息包括所述车辆的传感器采集的数据。
本申请实施例还提供了一种交互系统。该交互系统包括发送端设备和接收端设备，其中，所述发送端设备，用于获取车辆座舱内用户的第一意图，所述第一意图与另一车辆相关联且所述第一意图关联所述用户的负面情绪；所述发送端设备，还用于对所述第一意图进行去情绪化处理，得到第二意图；所述发送端设备，还用于向所述另一车辆发送所述第二意图；所述接收端设备，用于控制提示装置向用户提示所述第二意图。
图12和图13分别示出了本申请实施例提供的系统架构的示意图。
本申请实施例中，通过舱外传感器(例如摄像头)，可以识别邻近车辆的外观信息(例如，颜色、品牌、型号)，识别车牌号，识别车辆之间的相对位置关系，识别当前驾驶环境信息(拥堵、路口，小路)。通过舱外麦克风，识别对方车辆的鸣笛特征(是否鸣笛、鸣笛次数、紧迫程度)，识别周围环境信息(是否)通过摄像头、车辆传感器及智能驾驶(ADS)系统，识别自车或对方车辆的行驶状态以及异常驾驶行为，给驾驶员提供参考信息。
通过智能引擎系统，在大模型的辅助下，能够：理解用户描述的模糊语义，映射到邻近某辆具体车上。将当前用户表达的内容，结合当前的动态与静态环境信息、结合当前用户状态、结合对当前情况的一般处理流程(知识)，进行有效的二次加工，形成准确的、有理、有礼、有节的沟通内容。结合当前的状态信息，给用户以合适的反馈建议，同时可以帮用户执行操作，如编辑灯语、表达谢意、表示歉意等。
在云端服务器通过信息匹配或者通过移动网络，传输信息；或者，通过近场传输，完成信息交换。这种信息交换，可以不暴露驾驶员个人的信息，保护隐私；无需借助非语言的间接表达(闪灯、鸣笛)，减少理解歧义，提升沟通效率和安全性。
基于自然语言的描述，映射到邻近车辆信息，并进行信息交换，打破驾驶员之间的信息隔离，让沟通可以进行，情绪可以有效疏导。
本申请实施例中，能够根据当时的驾驶员状态、环境状态、驾驶行为等信息，处理并过滤驾驶员表达的信息，将关键信息抽取后，进行二次加工总结，进行交互传递；能够替驾驶员完成一系列意图的外显表达，如感谢、歉意、意见等。
本申请实施例中，根据自然语言描述的信息去定位车辆信息，达到信息建立连接的效果(可以使用车牌、VIN号等作为ID)。
本申请实施例中，根据状态信息(人、车、环境)对交互信息进行二次加工，达到有效、友善沟通的目的，达到传递有效信息，同时不传递负面情绪的目的。
本申请实施例中，车辆之间可以实现点对点的交互，也可以做广播式交互。
本申请实施例中，可以使用车辆其他硬件，如车灯、投影等，传递副语言信息。
本申请实施例中，投影包括DLP投到地面和行人交互，也可以包括投影到车窗上和周边交互。
本申请实施例中，车和车之间交互，可以包括移动网络交互，也包括近场交互(例如，通过星闪)。
本申请实施例还提供了一种交互装置，该交互装置包括用于执行上述交互方法的模块或者单元。
图14示出了本申请实施例提供的种交互方法1400的示意性流程图。该方法1400包括：
S1410，第一车辆获取第一车辆座舱内用户的第一输入和第一车辆周围的环境信息，环境信息中包括第二车辆的信息。
示例性的，该第一输入可以包括用户的语音输入。
示例性的，以第一车辆为图3中的车辆100且第二车辆为图3中的车辆200为例，该第一输入可以为驾驶员A的语音输入“对面这个车能不能把远光灯关了啊”。
示例性的，以第一车辆为图4中的车辆100且第二车辆为图4中的车辆200为例，该第一输入可以为驾驶员A的语音输入“帮我和前面这辆车说下，开这么慢就不要占快车道了，妨碍别人了啊”。
示例性的，以第一车辆为图5中的车辆100且第二车辆为图5中的车辆200为例，该第一输入可以为驾驶员A的语音输入“这人怎么回事，怎么开的车”。
示例性的，该第一输入可以包括用户对车辆中一个或者多个部件的输入。示例性的，该一个或者多个部件包括但不限于方向盘、灯光、喇叭、加速踏板或者制动踏板。
示例性的，以第一车辆为图2A中的车辆100且第二车辆为图2A中的车辆200为例，该第一输入可以为驾驶员A的鸣笛输入。
示例性的，以第一车辆为图6B中的车辆200且第二车辆为图6B中的车辆100为例，该第一输入可以为车辆200中驾驶员的鸣笛输入。
示例性的，该第一输入可以包括用户的生理特征信息。示例性的，该生理特征信息可以包括用户的血压、心率、面部表情等。
示例性的，以用户的面部表情为例。车辆可以获取座舱内的摄像头采集的图像数据并根据该图像数据确定用户的面部表情。示例性的，该用户的面部表情为皱眉的表情。
示例性的，该第一输入可以包括用户的肢体输入。例如，该肢体输入可以为手势输入。
示例性的，该第一输入可以为座舱内用户的可穿戴设备(例如，智能手表、智能手环等)采集的数据。
示例性的，该第一输入可以座舱内一个或者多个用户的输入。
可选地，该环境信息包括座舱外的传感器采集的数据。
示例性的，座舱外的传感器包括但不限于摄像头、激光雷达、毫米波雷达、麦克风中的一项或者多项。
S1420，第一车辆根据第一输入和环境信息，确定第一输出。
可选地，第一车辆根据第一输入和环境信息，确定第一输出之前，该方法1400还包括：确定该第一输入关联用户的负面情绪。
可选地，确定该第一输入关联用户的负面情绪，包括：确定该第一输入满足预设条件。
示例性的，以该第一输入为语音输入为例，该预设条件包括语音输入对应的文本内容中包括目标语义，该目标语义包括与抱怨、生气、非礼貌用语、语气助词等负面情绪相关的语义。示例性的，与抱怨相关的目标语义包括“怎么这样开车”、“妨碍别人”等。示例性的，与语气次助词相关的语义包括“啊”。
示例性的，以该第一输入为用户针对喇叭的输入为例，该预设条件包括用户按压喇叭的时长大于或者等于预设时长。
示例性的，以该第一输入为用户针对方向盘的输入为例，该预设条件包括用户对方向盘的握紧力大于或者等于预设握紧力；或者，该预设条件包括在预设时长内方向盘的转向角的变化率大于或者等于预设变化率。
示例性的，以该第一输入为用户的面部表情为例，该预设条件包括该用户的面部表情为皱眉、生气、愤怒的表情。
示例性的，以该第一输入为用户的心率为例，该预设条件包括用户的心率大于或者等于预设心率。
可选地，第一输入关联用户的负面情绪，第一输出包括去情绪化后的输出。
本申请实施例中，在用户的输入关联用户的负面情绪时，可以基于用户的输入和环境信息，对需要与第二车辆交互的信息进行二次加工，从而实现去情绪化处理。这样，可以使得车辆和车辆之间实现有效、友善沟通的目的，达到传递有效信息的同时不传递负面情绪的目的，有助于避免由于驾驶员的负面情绪而导致车辆和车辆之间发生冲突。
可选地，该第一输出包括非礼貌用语，该第一输出包括对该非礼貌用语转换后的礼貌用语。
基于上述技术方案，在进行去情绪化处理的同时，车辆还可以对用户的非礼貌用语进行转换，从而向第二车辆发送转换后的礼貌用语。这样，有助于进一步实现车辆和车辆之间实现有效、友善沟通的目的，达到传递有效信息的同时不传递负面情绪的目的，有助于减小驾驶员之间发生冲突的概率。
可选地，该第一输入关联用户的负面情绪，该方法1400还包括：控制提示装置输出第三输出，该第三输出包括用于对座舱内的用户进行安抚的输出结果。
基于上述技术方案，在第一输入关联用户的负面情绪时，还可以基于该第一输入和环境信息，对座舱内的用户输出进行安抚的输出结果。这样，在实现车辆和车辆之间友善交互的同时，还可以对座舱内的用户进行安抚，有助于缓解用户的负面情绪，有助于提升用户的驾乘体验。
示例性的，该第三输出可以包括安抚用语。
示例性的，该第一输入可以为如图4所示的驾驶员A的语音输入，该第三输出可以为语音输出1“好的呢，已经和对方表达了呢”以及语音输出2“前方的车主可能是个新手司机哦”。
示例性的，该第三输出可以包括对座舱内一个或者多个部件的执行指令。
可选地，根据第一输入和环境信息，确定第一输出，包括：将第一输入和环境信息输入第一推理模型中，得到第一输出。
基于上述技术方案，可以将第一输入和环境信息输入第一推理模型中，从而可以得到该第一输出。这样，通过第一推理模型可以实现端到端的输入和输出。
示例性的，该第一推理模型可以为多模态模型。
示例性的，以图3所示的场景为例，可以将驾驶员A的语音输入“对面这个车能不能把远光灯关了啊”以及车辆100的座舱外的摄像头采集的图像数据(例如，图片或者视频流)输入第一推理模型中。该第一推理模型可以输出用户的真实意图“请将远光灯切换为近光灯”。
示例性的，以图4所示的场景为例，可以将驾驶员A的语音输入“帮我和前面这辆车说下，开这么慢就不要占快车道了，妨碍别人了啊”以及车辆100的座舱外的摄像头采集的图像数据(例如，图片或者视频流)输入第一推理模型中。该第一推理模型可以输出用户的真实意图“请在快车道提速”。
示例性的，以图5所示的场景为例，可以将驾驶员A的语音输入“这人怎么回事，怎么开的车”以及车辆100的座舱外的摄像头采集的图像数据(例如，图片或者视频流)输入第一推理模型中。该第一推理模型可以输出用户的真实意图“请不要压着车道线行驶”。
可选地，方法1400还包括：获取第一车辆的信息；其中，第一车辆根据第一输入和环境信息，确定第一输出，包括：第一车辆根据第一输入、第一车辆的信息和环境信息，确定第一输出。
基于上述技术方案，通过车辆座舱内用户的输入、车辆的信息以及车辆周围的环境信息，可以确定第一输出，从而可以向第二车辆或者终端设备发送第一输出。这样，可以打破两个车辆中驾驶员之间的信息隔离，实现车辆与车辆之间的交互。同时，通过结合车辆的信息，有助于进一步提升第一输出结果的准确性。
示例性的，该车辆的信息包括车辆的历史行驶记录。
示例性的，该车辆的信息包括车辆的速度、加速度、位置中的一项或者多项。
示例性的，以图3所示的场景为例，可以将驾驶员A的语音输入“对面这个车能不能把远光灯关了啊”、车辆100的灯光状态(当前车辆100开启了近光灯，或者，检测到驾驶员频繁切换近光灯和远光灯)以及车辆100的座舱外的摄像头采集的图像数据输入第二推理模型中。该第二推理模型可以输出用户的真实意图“请将远光灯切换为近光灯”。
示例性的，以图4所示的场景为例，可以将驾驶员A的语音输入“帮我和前面这辆车说下，开这么慢就不要占快车道了，妨碍别人了啊”、车辆100的速度(例如，90km/h)以及车辆100的座舱外的摄像头采集的图像数据输入第二推理模型中。该第二推理模型可以输出用户的真实意图“请在快车道提速”。
可选地，第一车辆根据第一输入、第一车辆的信息和环境信息，确定第一输出，包括：将第一输入、第一车辆的信息和环境信息输入第二推理模型中，得到第一输出。
基于上述技术方案，可以将第一输入、第一车辆的信息和环境信息输入第一推理模型中，从而可以得到该第一输出。这样，通过第二推理模型可以实现端到端的输入和输出。
可选地，该第一推理模型和该第二推理模型可以为同一模型。
以第一推理模型和第二推理模型为同一多模态模型为例。该多模态模型的输入包括但不限于语音、图片、文本内容或者视频流中的一项或者多项。
该多模态大模型可以由通用的多模态大模型训练得到。例如，该多模态大模型包括知识库。知识库包括但不限于一个或者多个国家(或者地区)的道路法规、路段的限速信息、车辆交互时的合理处理方式。这样可以让通用的多模态大模型变成车辆交互领域的专属多模态大模型。多模态大模型的输出可以为去情绪化后用户的真实意图、转换后的礼貌用语、用户的真实意图的翻译结果中的一项或者多项。
可选地，方法1400还包括：第一车辆根据第一输入和环境信息，确定第二车辆的信息。
基于上述技术方案，通过第一车辆座舱内用户的输入以及环境信息，可以确定第二车辆的信息，从而可以向第二车辆或者终端设备发送第一输出。这样，在确定第二车辆的信息时可以结合用户的输入和环境信息，保证确定出的待交互的车辆的准确性。
可选地，第一车辆根据第一输入和环境信息，确定第二车辆的信息，包括：第一车辆将第一输入和环境信息输入第三推理模型中，得到第二车辆的信息。
示例性的，以图3所示的场景为例，可以将驾驶员A的语音输入“对面这个车能不能把远光灯关了啊”、车辆100的灯光状态(当前车辆100开启了近光灯，或者，检测到驾驶员频繁切换近光灯和远光灯)以及车辆100的座舱外的摄像头采集的图像数据输入第二推理模型中。该第二推理模型可以输出用户的真实意图“请将远光灯切换为近光灯”以及车辆200的车牌信息。
示例性的，以图4所示的场景为例，可以将驾驶员A的语音输入“帮我和前面这辆车说下，开这么慢就不要占快车道了，妨碍别人了啊”、车辆100的速度(例如，90km/h)以及车辆100的座舱外的摄像头采集的图像数据输入第二推理模型中。该第二推理模型可以输出用户的真实意图“请在快车道提速”以及车辆200的车牌信息。
可选地，该第一推理模型、该第二推理模型和该第三推理模型可以为同一模型。
以第一推理模型、第二推理模型和第三推理模型为同一多模态模型为例。多模态大模型的输出还可以包括第二车辆的信息。例如，第二车辆的信息包括第二车辆的车牌信息。
可选地，该第一推理模型和该第三推理模型可以不是同一个模型。
示例性的，该第三推理模型可以由语义识别算法和图像分割算法实现。示例性的，第一车辆在获取到用户的语音输入“前面的红色车能不能开快一点”时，可以获取座舱外的摄像头采集的图像1。通过语义识别算法，可以确定该语音输入中与目标的属性相关的文本内容(例如，“红色车”)。第一车辆可以基于图像分割算法，从图像1中提取到ROI1和ROI2，ROI1和ROI2分别包括红色的车辆。例如，图像1可以被划分为多个区域，ROI1可以为多个区域中的区域a，ROI2可以为多个区域中的区域b。若用户触发语音输入时用户的视线方向指向该区域a，可以选择ROI1作为目标ROI。第一车辆可以向云端服务器发送该目标ROI。从而可以由云端服务器基于该目标ROI分析得到第二车辆的车牌信息。
可选地，方法1400还包括：第一车辆根据第一输入和环境信息，确定第一控制指令，第一控制指令关联一个或者多个执行器；第一车辆控制一个或者多个执行器执行第一控制指令。
示例性的，以第一车辆为图2A中的车辆100且第二车辆为图2A中的车辆200为例，该第一输入可以为驾驶员A的鸣笛输入。第一车辆可以根据该鸣笛输入以及座舱外的传感器采集的图像数据，生成对发声装置和氛围灯的控制指令。示例性的，对发声装置的控制指令用于指示发声装置播放缓解不满的音乐(例如，舒缓的音乐)。示例性的，对氛围灯的控制指令包括控制氛围灯的颜色为暖色。
基于上述技术方案，通过第一输入和环境信息，还可以得到控制指令。这样，通过执行对一个或者多个执行器的控制指令，可以实现缓解用户愤怒、生气或者急躁的情绪。
S1430，第一车辆向第二车辆或者终端设备发送第一输出，终端设备与第二车辆关联。
可选地，第二车辆的信息包括第二车辆的车牌的信息；其中，第一车辆向第二车辆或者终端设备发送第一输出，包括：第一车辆向云端服务器发送第一输出和第二车辆的车牌的信息，以使得云端服务器基于第二车辆的车牌的信息，向第二车辆或者终端设备发送第一输出。
基于上述技术方案，车辆可以将第一输出和第二车辆的车牌的信息发送给云端服务器，从而使得云端服务器可以通过该第二车辆的车牌的信息，向第二车辆发送该第一输出。通过云端服务器对信息进行转发，可以打破两个车辆中驾驶员之间的信息隔离，实现车辆与车辆之间的交互。
可选地，该云端服务器中保存车辆的车牌与终端设备(例如，手机或者车辆)的标识信息之间的关联关系。
本申请实施例中，云端服务器可以保存多个账号中每个账号、每个账号对应的一个或者多个设备的身份标识号码(identity document，ID)或者地址信息与车牌信息的对应关系。
示例性的，表1示出了每个账号、每个账号对应的一个或者多个设备的地址信息与车牌信息之间的对应关系。
表1
示例性的，云端服务器可以接收多个终端设备发送的车辆的车牌的信息。例如，用户1可以在登录了账号1的车辆B输入车辆B关联的车牌号xxx的信息，从而可以使得车辆B可以向云端服务器发送账号1与车牌号为xxx的绑定关系。
又例如，用户2可以在登录了账号2的手机C输入车辆C关联的车牌号yyy的信息，从而可以使得手机C可以向云端服务器发送账号2与车牌号为yyy的绑定关系。
示例性的，上述多模态大模型的输出结果中可以包括车牌号xxx的信息。云端服务器在接收到第一车辆发送的第一输出以及车牌号xxx的信息后，可以根据上述表1所示的关联关系，确定向车辆B发送该第一输出。
示例性的，上述多模态大模型的输出结果中可以包括车牌号yyy的信息。云端服务器在接收到第一车辆发送的第一输出以及车牌号yyy的信息后，可以根据上述表1所示的关联关系，确定向手机C发送该第一输出。手机C在接收到该第一输出后可以判断手机C是否位于车辆C的座舱内。如果手机C位于车辆C的座舱内，那么手机C可以基于该第一输出，对用户进行提示。
可选地，第一车辆向第二车辆或者终端设备发送第一输出，包括：第一车辆通过近距离通信技术向第二车辆发送该第一输出。
示例性的，以图3所示的场景为例，在该环境信息指示车辆100周围存在车辆200且车辆200当前的灯光状态指示车辆200开启了远光灯时，车辆100可以通过星闪技术向车辆200发送该第一输出。
S1440，第二车辆根据第一输出和环境信息，确定第二输出。
示例性的，以图3所示的场景为例，该第一推理模型可以输出用户的真实意图“请将远光灯切换为近光灯”。车辆200可以将车辆100中的用户的真实意图以及车辆200座舱外的摄像头采集的图像数据输入第四推理模型中，从而可以得到文本内容“亲爱的车主，您现在开的是远光灯，可能影响对面司机行驶了，对面的车主希望您切换成近光灯哦”。
示例性的，以图4所示的场景为例，该第一推理模型可以输出用户的真实意图“请在快车道提速”。车辆200可以将车辆100中的用户的真实意图以及座舱外的摄像头采集的图像数据输入第四推理模型中，从而可以得到文本内容“亲爱的车主，小A提醒您，您现在在快车道上，限速80-100km/h，车辆现在行驶速度是65km/h，后面的司机有点着急，请您提速”。
示例性的，以图7所示的场景为例，第一车辆可以为车辆100且第二车辆可以为车辆200。该第一输入可以为车辆100内的驾驶员的鸣笛输入。车辆100可以将该鸣笛输入以及车辆100座舱外的传感器采集的数据输入第一推理模型中，从而可以得到用户的意图“请不要加塞”。由于车辆100的座舱外的传感器未检测到车辆200前方的锥桶，车辆100会误认为用户的意图符合道路法规，从而可以向车辆200发送该用户的意图。车辆200在接收到该用户的意图后，可以将该用户的意图以及座舱外的传感器采集的数据(包括与占道施工相关的数据)输入第四推理模型中，从而可以得到文本内容“鸣笛的车主可能不知道占道施工的情况，我这就向他转达”。
可选地，第二车辆根据第一输出和环境信息，确定第二输出之前，该方法1400还包括：第二车辆控制提示装置提示用户接收到来自于第一车辆的第一输出且提示用户是否对用户进行提示；其中，第二车辆根据第一输出和环境信息，确定第二输出，包括：响应于检测到用户确定进行提示的输入，第二车辆根据第一输出和环境信息，确定第二输出。
可选地，第二车辆根据第一输出和环境信息，确定第二输出，包括：第二车辆根据车辆座舱内的传感器采集的数据、第一输出和环境信息，确定第二输出。
示例性的，第二车辆可以将该第一输出和环境信息输出第四推理模型中，从而可以得到该第二输出。
示例性的，该第四推理模型可以为多模态大模型。该多模态大模型可以由通用的多模态大模型训练得到。例如，该多模态大模型包括知识库。知识库包括但不限于一个或者多个国家(或者地区)的道路法规、路段的限速信息、车辆交互时的合理处理方式。这样可以让通用的多模态大模型变成车辆交互领域的专属多模态大模型。
示例性的，以图3所示的场景为例，第一车辆可以为车辆100且第二车辆可以为车辆200。该第一推理模型可以输出用户的真实意图“请将远光灯切换为近光灯”。车辆200可以将车辆100中的用户的真实意图、座舱内的传感器采集的数据(指示车辆200内驾驶员处于紧张状态)以及座舱外的摄像头采集的图像数据输入第四推理模型中，从而可以得到文本内容“亲爱的车主，您现在开的是远光灯，可能影响对面司机行驶了，对面的车主希望您切换成近光灯哦。您可以将方向盘右后方的拨杆向上拨动，以切换至近光灯”。
示例性的，以图4所示的场景为例，第一车辆可以为车辆100且第二车辆可以为车辆200。该第一推理模型可以输出用户的真实意图“请在快车道提速”。车辆200可以将车辆100中的用户的真实意图、座舱内的传感器采集的数据(指示车辆200内驾驶员处于愉悦状态)以及座舱外的摄像头采集的图像数据输入第四推理模型中，从而可以得到文本内容“亲爱的车主，小A提醒您，后面的司机有点着急，抓紧提速哦”。
基于上述技术方案，在确定第二输出时可以结合车辆座舱内的传感器采集的数据。这样，可以使得该第二输出更容易被当前座舱内用户所接受，有助于提升用户的驾乘体验。
可选地，第二车辆根据第一输出和环境信息，确定第二输出，包括：第二车辆根据第二车辆座舱内驾驶员的第一信息、第一输出和环境信息，确定第二输出，第一信息包括驾驶熟练度、驾驶习惯或者生理特征信息中的一项或者多项。
示例性的，以图4所示的场景为例，第一车辆可以为车辆100且第二车辆可以为车辆200。该第一推理模型可以输出用户的真实意图“请在快车道提速”。车辆200可以将车辆100中的用户的真实意图以及车辆200中驾驶员的驾驶熟练度(指示第二车辆中驾驶员为新手司机)输入第四推理模型中，从而可以得到文本内容“亲爱的车主，小A提醒您，您现在在快车道上，限速80-100km/h，车辆现在行驶速度是65km/h，后面的司机有点着急，请您提速”。
示例性的，以图4所示的场景为例。车辆200可以将车辆100中的用户的真实意图以及车辆200中驾驶员的驾驶熟练度(指示第二车辆中驾驶员为熟练程度较高的驾驶员)输入第四推理模型中，从而可以得到文本内容“亲爱的车主，后面的司机有点着急，请您提速”。
示例性的，以图4所示的场景为例。车辆200可以将车辆100中的用户的真实意图以及车辆200中驾驶员的驾驶习惯(在过去一段时间内驾驶车辆在慢车道上行驶的频率大于或者等于预设频率)输入第四推理模型中，从而可以得到文本内容“亲爱的车主，后面的司机有点着急，您可以切换至慢车道上”。
示例性的，以图4所示的场景为例。车辆200可以将车辆100中的用户的真实意图以及车辆200中驾驶员的生理特征信息(指示驾驶员为老年驾驶员)输入第四推理模型中，从而可以得到文本内容“亲爱的车主，小A提醒您，您现在在快车道上，限速80-100km/h，车辆现在行驶速度是65km/h，后面的司机有点着急，如果您不着急的话可以向右打转向灯，切换到慢车道上”。
基于上述技术方案，在确定第二输出时可以结合驾驶员的信息。这样，可以使得该第二输出更符合驾驶员的画像，从而实现不同的驾驶员的画像得到不同的输出结果，有助于提升车辆的智能化程度，也有助于提升用户的驾乘体验。
可选地，第二车辆根据第一输出和环境信息，确定第二输出，包括：第二车辆根据第二车辆周围其他车辆的历史驾驶记录、第一输出和环境信息，确定第二输出。
示例性的，以图7所示的场景为例，该第一输出可以为用户的意图“请不要加塞”。由于车辆100的座舱外的传感器未检测到车辆200前方的锥桶，会误认为用户的意图符合道路法规，从而可以向车辆200发送该用户的意图。车辆200在接收到该用户的意图后，可以将该用户的意图、车辆200周围其他车辆的历史驾驶记录(该历史驾驶记录指示周围其他车辆在过去一段时间内交替通过该路段)以及座舱外的传感器采集的数据(包括与占道施工相关的数据)输入第四推理模型中，从而可以得到文本内容“鸣笛的车主可能不知道占道施工的情况，我这就向他转达”。
基于上述技术方案，在确定第二输出时可以结合车辆周围其他车辆的历史驾驶记录。这样，可以使得该第二输出更为准确，也可以使得第二输出更容易被座舱内的用户接受。
可选地，第二车辆根据第一输出和环境信息，确定第二输出，包括：第二车辆将第一输出和环境信息输入推理模型中，得到第二输出。
S1450，第二车辆控制提示装置输出第二输出。
示例性的，在获得该第四推理模型的文本内容后，第二车辆可以控制第二车辆座舱内的发声装置播放对应的语音信息。
图15示出了本申请实施例提供的系统1500的架构的示意图。该系统1500包括车辆1510和车辆1520。车辆1510中可以包括第一推理模型，车辆1520中可以包括第四推理模型。
以上车辆1510可以为上述车辆100，车辆1520可以为上述车辆200。或者，以上车辆1510可以为上述第一车辆，车辆1520可以为上述第二车辆。
图16示出了本申请实施例提供的交互装置1600的示意性框图。该装置1600包括：获取单元1610，用于获取车辆座舱内用户的第一输入和车辆周围的环境信息，环境信息中包括另一车辆的信息；确定单元1620，用于根据第一输入和环境信息，确定第一输出；发送单元1630，用于向另一车辆或者终端设备发送第一输出，终端设备与另一车辆关联。
可选地，第一输入关联用户的负面情绪，第一输出包括去情绪化后的输出。
可选地，确定单元1620，用于：将第一输入和环境信息输入第一推理模型中，得到第一输出。
可选地，获取单元1610，还用于获取车辆的信息；确定单元，用于：根据第一输入、车辆的信息和环境信息，确定第一输出。
可选地，确定单元1620，用于：将第一输入、车辆的信息和环境信息输入第二推理模型中，得到第一输出。
可选地，确定单元1620，还用于：根据第一输入和环境信息，确定另一车辆的信息。
可选地，确定单元1620，用于：将第一输入和环境信息输入第三推理模型中，得到另一车辆的信息。
可选地，另一车辆的信息包括另一车辆的车牌的信息；其中，发送单元1630，用于：向云端服务器发送第一输出和另一车辆的车牌的信息，以使得云端服务器基于另一车辆的车牌的信息，向另一车辆或者终端设备发送第一输出。
可选地，装置还包括控制单元，确定单元，还用于根据第一输入和环境信息，确定第一控制指令，第一控制指令关联一个或者多个执行器；控制单元，用于控制一个或者多个执行器执行第一控制指令。
图17示出了本申请实施例提供的交互装置1700的示意性框图。该装置1700包括：获取单元1710，用于获取另一车辆的第一输出和车辆周围的环境信息；确定单元1720，用于根据第一输出和环境信息，确定第二输出；控制单元1730，用于控制提示装置输出第二输出。
可选地，确定单元1720，用于：根据车辆座舱内的传感器采集的数据、第一输出和环境信息，确定第二输出。
可选地，确定单元1720，用于：根据车辆中驾驶员的第一信息、第一输出和环境信息，确定第二输出，第一信息包括驾驶熟练度、驾驶习惯或者生理特征信息中的一项或者多项。
可选地，确定单元1720，用于：根据车辆周围其他车辆的历史驾驶记录、第一输出和环境信息，确定第二输出。
可选地，确定单元1720，用于：将第一输出和环境信息输入推理模型中，得到第二输出。
应理解以上装置中各单元的划分仅是一种逻辑功能的划分，实际实现时可以全部或部分集成到一个物理实体上，也可以物理上分开。此外，装置中的单元可以以处理器调用软件的形式实现；例如装置包括处理器，处理器与存储器连接，存储器中存储有指令，处理器调用存储器中存储的指令，以实现以上任一种方法或实现该装置各单元的功能，其中处理器例如为通用处理器，例如CPU或微处理器，存储器为装置内的存储器或装置外的存储器。或者，装置中的单元可以以硬件电路的形式实现，可以通过对硬件电路的设计实现部分或全部单元的功能，该硬件电路可以理解为一个或多个处理器；例如，在一种实现中，该硬件电路为ASIC，通过对电路内元件逻辑关系的设计，实现以上部分或全部单元的功能；再如，在另一种实现中，该硬件电路为可以通过PLD实现，以FPGA为例，其可以包括大量逻辑门电路，通过配置文件来配置逻辑门电路之间的连接关系，从而实现以上部分或全部单元的功能。以上装置的所有单元可以全部通过处理器调用软件的形式实现，或全部通过硬件电路的形式实现，或部分通过处理器调用软件的形式实现，剩余部分通过硬件电路的形式实现。
在本申请实施例中，处理器是一种具有信号的处理能力的电路，在一种实现中，处理器可以是具有指令读取与运行能力的电路，例如CPU、微处理器、GPU、或DSP等；在另一种实现中，处理器可以通过硬件电路的逻辑关系实现一定功能，该硬件电路的逻辑关系是固定的或可以重构的，例如处理器为ASIC或PLD实现的硬件电路，例如FPGA。在可重构的硬件电路中，处理器加载配置文档，实现硬件电路配置的过程，可以理解为处理器加载指令，以实现以上部分或全部单元的功能的过程。此外，还可以是针对人工智能设计的硬件电路，其可以理解为一种ASIC，例如NPU、TPU、DPU等。
可见，以上装置中的各单元可以是被配置成实施以上方法的一个或多个处理器(或处理电路)，例如：CPU、GPU、NPU、TPU、DPU、微处理器、DSP、ASIC、FPGA，或这些处理器形式中至少两种的组合。
此外，以上装置中的各单元可以全部或部分可以集成在一起，或者可以独立实现。在一种实现中，这些单元集成在一起，以SoC的形式实现。该SoC中可以包括至少一个处理器，用于实现以上任一种方法或实现该装置各单元的功能，该至少一个处理器的种类可以不同，例如包括CPU和FPGA，CPU和人工智能处理器，CPU和GPU等。
本申请实施例还提供了一种交互装置，该装置包括处理单元和存储单元，其中存储单元用于存储指令，处理单元执行存储单元所存储的指令，以使该装置执行上述实施例执行的方法或者步骤。
可选地，若该交互装置位于车辆中，上述处理单元可以是图1所示的处理器121-12n中的一个或者多个。
本申请实施例还提供了一种交互系统，该交互系统包括上述交互装置和感知系统。
本申请实施例还提供了一种车辆，该车辆可以包括上述交互装置或者上述交互系统。
本申请实施例还提供了一种计算机程序产品，所述计算机程序产品包括：计算机程序代码，当所述计算机程序代码在计算机上运行时，使得计算机执行上述实施例中的方法。
本申请实施例还提供了一种计算机可读介质，所述计算机可读介质存储有程序代码，当所述计算机程序代码在计算机上运行时，使得计算机执行上述实施例中的方法。
本申请实施例还提供了一种芯片，所述芯片包括电路，所述电路用于执行上述实施例中的方法。
在实现过程中，上述方法的各步骤可以通过处理器中的硬件的集成逻辑电路或者软件形式的指令完成。结合本申请实施例所公开的方法可以直接体现为硬件处理器执行完成，或者用处理器中的硬件及软件模块组合执行完成。软件模块可以位于随机存储器，闪存、只读存储器，可编程只读存储器或者上电可擦写可编程存储器、寄存器等本领域成熟的存储介质中。该存储介质位于存储器，处理器读取存储器中的信息，结合其硬件完成上述方法的步骤。为避免重复，这里不再详细描述。
应理解，本申请实施例中，该存储器可以包括只读存储器和随机存取存储器，并向处理器提供指令和数据。
还应理解，在本申请的各种实施例中，上述各过程的序号的大小并不意味着执行顺序的先后，各过程的执行顺序应以其功能和内在逻辑确定，而不应对本申请实施例的实施过程构成任何限定。
本领域普通技术人员可以意识到，结合本文中所公开的实施例描述的各示例的单元及算法步骤，能够以电子硬件、或者计算机软件和电子硬件的结合来实现。这些功能究竟以硬件还是软件方式来执行，取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能，但是这种实现不应认为超出本申请的范围。
所属领域的技术人员可以清楚地了解到，为描述的方便和简洁，上述描述的系统、装置和单元的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
在本申请所提供的几个实施例中，应该理解到，所揭露的系统、装置和方法，可以通过其它的方式实现。例如，以上所描述的装置实施例仅仅是示意性的，例如，所述单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。另一点，所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些接口，装置或单元的间接耦合或通信连接，可以是电性，机械或其它的形式。
所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部单元来实现本实施例方案的目的。
另外，在本申请各个实施例中的各功能单元可以集成在一个处理单元中，也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个单元中。
所述功能如果以软件功能单元的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本申请各个实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(read-only memory，ROM)、随机存取存储器(random access memory，RAM)、磁碟或者光盘等各种可以存储程序代码的介质。
以上所述，仅为本申请的具体实施方式，但本申请的保护范围并不局限于此，任何熟悉本技术领域的技术人员在本申请揭露的技术范围内，可轻易想到变化或替换，都应涵盖。在本申请的保护范围之内。因此，本申请的保护范围应以所述权利要求的保护范围为准。

Claims (34)

  1. 一种交互方法，其特征在于，包括：
    获取车辆座舱内用户的第一输入和所述车辆周围的环境信息，所述环境信息中包括另一车辆的信息；
    根据所述第一输入和所述环境信息，确定第一输出；
    向所述另一车辆或者终端设备发送所述第一输出，所述终端设备与所述另一车辆关联。
  2. 根据权利要求1所述的方法，其特征在于，所述第一输入关联用户的负面情绪，所述第一输出包括去情绪化后的输出。
  3. 根据权利要求1或2所述的方法，其特征在于，所述根据所述第一输入和所述环境信息，确定第一输出，包括：
    将所述第一输入和所述环境信息输入第一推理模型中，得到所述第一输出。
  4. 根据权利要求1至3中任一项所述的方法，其特征在于，所述方法还包括：
    获取所述车辆的信息；
    其中，所述根据所述第一输入和所述环境信息，确定第一输出，包括：
    根据所述第一输入、所述车辆的信息和所述环境信息，确定所述第一输出。
  5. 根据权利要求4所述的方法，其特征在于，所述根据所述第一输入、所述车辆的信息和所述环境信息，确定所述第一输出，包括：
    将所述第一输入、所述车辆的信息和所述环境信息输入第二推理模型中，得到所述第一输出。
  6. 根据权利要求1至5中任一项所述的方法，其特征在于，所述方法还包括：
    根据所述第一输入和所述环境信息，确定所述另一车辆的信息。
  7. 根据权利要求6所述的方法，其特征在于，所述根据所述第一输入和所述环境信息，确定所述另一车辆的信息，包括：
    将所述第一输入和所述环境信息输入第三推理模型中，得到所述另一车辆的信息。
  8. 根据权利要求6或7所述的方法，其特征在于，所述另一车辆的信息包括所述另一车辆的车牌的信息；
    其中，所述向所述另一车辆或者终端设备发送第一输出，包括：
    向云端服务器发送所述第一输出和所述另一车辆的车牌的信息，以使得所述云端服务器基于所述另一车辆的车牌的信息，向所述另一车辆或者所述终端设备发送所述第一输出。
  9. 根据权利要求1至8中任一项所述的方法，其特征在于，所述方法还包括：
    根据所述第一输入和所述环境信息，确定第一控制指令，所述第一控制指令关联一个或者多个执行器；
    控制所述一个或者多个执行器执行所述第一控制指令。
  10. 一种交互方法，其特征在于，包括：
    获取另一车辆的第一输出和车辆周围的环境信息；
    根据所述第一输出和所述环境信息，确定第二输出；
    控制提示装置输出所述第二输出。
  11. 根据权利要求10所述的方法，其特征在于，所述根据所述第一输出和所述环境信息，确定第二输出，包括：
    根据车辆座舱内的传感器采集的数据、所述第一输出和所述环境信息，确定所述第二输出。
  12. 根据权利要求10或11所述的方法，其特征在于，所述根据所述第一输出和所述环境信息，确定第二输出，包括：
    根据所述车辆中驾驶员的第一信息、所述第一输出和所述环境信息，确定所述第二输出，所述第一信息包括驾驶熟练度、驾驶习惯或者生理特征信息中的一项或者多项。
  13. 根据权利要求10至12中任一项所述的方法，其特征在于，所述根据所述第一输出和所述环境信息，确定第二输出，包括：
    所根据所述车辆周围其他车辆的历史驾驶记录、所述第一输出和所述环境信息，确定所述第二输出。
  14. 根据权利要求10至13中任一项所述的方法，其特征在于，所述根据所述第一输出和所述环境信息，确定第二输出，包括：
    将所述第一输出和所述环境信息输入推理模型中，得到所述第二输出。
  15. 一种交互装置，其特征在于，包括：
    获取单元，用于获取车辆座舱内用户的第一输入和所述车辆周围的环境信息，所述环境信息中包括另一车辆的信息；
    确定单元，用于根据所述第一输入和所述环境信息，确定第一输出；
    发送单元，用于向所述另一车辆或者终端设备发送所述第一输出，所述终端设备与所述另一车辆关联。
  16. 根据权利要求15所述的装置，其特征在于，所述第一输入关联用户的负面情绪，所述第一输出包括去情绪化后的输出。
  17. 根据权利要求15或16所述的装置，其特征在于，所述确定单元，用于：
    将所述第一输入和所述环境信息输入第一推理模型中，得到所述第一输出。
  18. 根据权利要求15至17中任一项所述的装置，其特征在于，
    所述获取单元，还用于获取所述车辆的信息；
    所述确定单元，用于：根据所述第一输入、所述车辆的信息和所述环境信息，确定所述第一输出。
  19. 根据权利要求18所述的装置，其特征在于，所述确定单元，用于：
    将所述第一输入、所述车辆的信息和所述环境信息输入第二推理模型中，得到所述第一输出。
  20. 根据权利要求15至19中任一项所述的装置，其特征在于，
    所述确定单元，还用于：根据所述第一输入和所述环境信息，确定所述另一车辆的信息。
  21. 根据权利要求20所述的装置，其特征在于，所述确定单元，用于：
    将所述第一输入和所述环境信息输入第三推理模型中，得到所述另一车辆的信息。
  22. 根据权利要求20或21所述的装置，其特征在于，所述另一车辆的信息包括所述另一车辆的车牌的信息；
    其中，所述发送单元，用于：向云端服务器发送所述第一输出和所述另一车辆的车牌的信息，以使得所述云端服务器基于所述另一车辆的车牌的信息，向所述另一车辆或者所述终端设备发送所述第一输出。
  23. 根据权利要求15至22中任一项所述的装置，其特征在于，所述装置还包括控制单元，
    所述确定单元，还用于根据所述第一输入和所述环境信息，确定第一控制指令，所述第一控制指令关联一个或者多个执行器；
    所述控制单元，用于控制所述一个或者多个执行器执行所述第一控制指令。
  24. 一种交互装置，其特征在于，包括：
    获取单元，用于获取另一车辆的第一输出和车辆周围的环境信息；
    确定单元，用于根据所述第一输出和所述环境信息，确定第二输出；
    控制单元，用于控制提示装置输出所述第二输出。
  25. 根据权利要求24所述的装置，其特征在于，所述确定单元，用于：
    根据车辆座舱内的传感器采集的数据、所述第一输出和所述环境信息，确定所述第二输出。
  26. 根据权利要求24或25所述的装置，其特征在于，所述确定单元，用于：
    根据所述车辆中驾驶员的第一信息、所述第一输出和所述环境信息，确定所述第二输出，所述第一信息包括驾驶熟练度、驾驶习惯或者生理特征信息中的一项或者多项。
  27. 根据权利要求24至26中任一项所述的装置，其特征在于，所述确定单元，用于：
    所根据所述车辆周围其他车辆的历史驾驶记录、所述第一输出和所述环境信息，确定所述第二输出。
  28. 根据权利要求24至27中任一项所述的装置，其特征在于，所述确定单元，用于：
    将所述第一输出和所述环境信息输入推理模型中，得到所述第二输出。
  29. 一种交互装置，其特征在于，包括：
    存储器，用于存储计算机程序；
    处理器，用于执行所述存储器中存储的计算机程序，以使得所述装置执行如权利要求1至14中任一项所述的方法。
  30. 一种交互系统，其特征在于，所述交互系统包括计算平台和感知系统，所述计算平台包括如权利要求29所述的交互装置。
  31. 一种车辆，其特征在于，包括如权利要求15-29中任一项所述的交互装置，或者，包括如权利要求30所述的交互系统。
  32. 一种计算机可读存储介质，其特征在于，其上存储有指令，所述指令被处理器执行时，以使得处理器实现如权利要求1至14中任一项所述的方法。
  33. 一种计算机程序产品，其特征在于，所述计算机程序产品包括计算机程序代码，当所述计算机程序代码在计算机上运行时，以使得计算机实现如权利要求1至14中任一项所述的方法。
  34. 一种芯片，其特征在于，所述芯片包括电路，所述电路用于执行如权利要求1至14中任一项所述的方法。
PCT/CN2025/104214 2024-06-28 2025-06-27 交互方法、装置和车辆 Pending WO2026002181A1 (zh)

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
CN202410869842.4 2024-06-28
CN202410869842 2024-06-28
CN202411993129.7A CN121237081A (zh) 2024-06-28 2024-12-30 交互方法、装置和车辆
CN202411993129.7 2024-12-30

Publications (1)

Publication Number Publication Date
WO2026002181A1 true WO2026002181A1 (zh) 2026-01-02

Family

ID=98155351

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2025/104214 Pending WO2026002181A1 (zh) 2024-06-28 2025-06-27 交互方法、装置和车辆

Country Status (2)

Country Link
CN (1) CN121237081A (zh)
WO (1) WO2026002181A1 (zh)

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111391867A (zh) * 2020-02-28 2020-07-10 华域视觉科技(上海)有限公司 一种驾驶情绪表达系统、方法及车辆
CN111413961A (zh) * 2018-12-19 2020-07-14 本田技研工业株式会社 控制装置以及计算机可读存储介质
CN113226886A (zh) * 2021-03-31 2021-08-06 华为技术有限公司 控制车辆行驶的方法、装置及车辆
JP2022038423A (ja) * 2020-08-26 2022-03-10 トヨタ自動車株式会社 車両用エージェント装置
WO2022266844A1 (zh) * 2021-06-22 2022-12-29 华为技术有限公司 一种信息交互方法和装置
CN116394967A (zh) * 2023-06-09 2023-07-07 苏州观瑞汽车技术有限公司 一种基于传感器识别的智能前后车交互方法及系统
CN117141512A (zh) * 2023-09-28 2023-12-01 重庆赛力斯新能源汽车设计院有限公司 车载情绪调节系统的控制方法及装置

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111413961A (zh) * 2018-12-19 2020-07-14 本田技研工业株式会社 控制装置以及计算机可读存储介质
CN111391867A (zh) * 2020-02-28 2020-07-10 华域视觉科技(上海)有限公司 一种驾驶情绪表达系统、方法及车辆
JP2022038423A (ja) * 2020-08-26 2022-03-10 トヨタ自動車株式会社 車両用エージェント装置
CN113226886A (zh) * 2021-03-31 2021-08-06 华为技术有限公司 控制车辆行驶的方法、装置及车辆
WO2022266844A1 (zh) * 2021-06-22 2022-12-29 华为技术有限公司 一种信息交互方法和装置
CN116394967A (zh) * 2023-06-09 2023-07-07 苏州观瑞汽车技术有限公司 一种基于传感器识别的智能前后车交互方法及系统
CN117141512A (zh) * 2023-09-28 2023-12-01 重庆赛力斯新能源汽车设计院有限公司 车载情绪调节系统的控制方法及装置

Also Published As

Publication number Publication date
CN121237081A (zh) 2025-12-30

Similar Documents

Publication Publication Date Title
US20230110773A1 (en) Control system and method using in-vehicle gesture input
US12240502B2 (en) Presentation control device, presentation control program, and driving control device
KR102909636B1 (ko) 양방향 차내 가상 개인 비서
US10137902B2 (en) Adaptive interactive voice system
CN111016820B (zh) 智能体系统、智能体控制方法及存储介质
US12056939B2 (en) Control system using in-vehicle gesture input
US20160288708A1 (en) Intelligent caring user interface
US12296675B2 (en) Control device and control program product
JP3209387U (ja) 音声認識による車両運転者の意思表示装置
JP6150077B2 (ja) 車両用音声対話装置
CN114312797B (zh) 智能体装置、智能体方法以及记录介质
CN115859219A (zh) 多模态交互方法、装置、设备及存储介质
JP2020020987A (ja) 車内システム
JP7148296B2 (ja) 車載用ロボット
CN114291008A (zh) 车辆用智能体装置、车辆用智能体系统以及计算机可读存储介质
WO2026002181A1 (zh) 交互方法、装置和车辆
CN116547729A (zh) 关注对象共享装置、关注对象共享方法
Nakrani Smart car technologies: a comprehensive study of the state of the art with analysis and trends
CN116959481A (zh) 听障人士与周围环境的信息互动方法及装置
WO2026000340A1 (zh) 控制方法、装置和车辆
JP2020059401A (ja) 車両制御装置、車両制御方法、プログラム
JP2020086905A (ja) 情報提供装置及び車載装置
JP7570949B2 (ja) 情報処理装置、情報処理方法及び学習済みモデル
US20240227868A9 (en) Presentation control device, and automated drive control device
WO2026004565A1 (ja) 情報処理装置、情報処理方法、及び、プログラム

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 25826252

Country of ref document: EP

Kind code of ref document: A1