Semantic arbitration method

Abstract

This invention provides a semantic arbitration method, belonging to the field of vehicle-to-everything (V2X) voice interaction technology. The semantic arbitration method includes: offline recognition of user speech to obtain a first voice result; online recognition of user speech to obtain a second voice result; the first voice result and the second voice result both include at least one user intent; calculating a first weight for the first voice result and a second weight for the second voice result based on contextual relevance and the number of successful historical arbitrations, and adopting the result with the higher weight value; the number of successful historical arbitrations refers to the number of times the voice results output by offline and online recognition have been adopted within a preset number of times. This invention improves the arbitration accuracy when online and offline voice results conflict by introducing contextual relevance and historical performance as dynamic weight calculation factors.

Description

Technical Field

[0001] This invention relates to the field of vehicle-to-everything (V2X) voice interaction technology, and in particular to a semantic arbitration method. Background Technology

[0002] In existing vehicle-to-everything (V2X) voice interaction systems, the edge (offline) and cloud (online) semantic processing modules work in parallel, each outputting semantic results. However, due to factors such as network fluctuations and scenario complexity, conflicts may occur between the two sets of results. Traditional arbitration logic relies solely on a first-come, first-served basis for decision-making, leading to insufficient arbitration accuracy in complex scenarios. Summary of the Invention

[0003] This invention aims to solve at least one of the technical problems existing in the prior art, and proposes a semantic arbitration method based on dynamic weights.

[0004] In a first aspect, embodiments of the present invention provide a semantic arbitration method, comprising: offline recognition of user speech to obtain a first speech result; online recognition of user speech to obtain a second speech result; the first speech result and the second speech result include at least one user intent; calculating a first weight of the first speech result and a second weight of the second speech result based on contextual relevance and the number of successful historical arbitrations, and adopting the one with the higher weight value as the semantic result; the number of successful historical arbitrations is the number of times the speech results output by offline and online recognition have been adopted within a preset number of times.

[0005] According to an embodiment of the present invention, calculating a first weight of a first speech result and a second weight of a second speech result based on contextual relevance and historical arbitration records includes: in response to the first speech result and the second speech result including more than one user intent, calculating a first weight of the first speech result and a second weight of the second speech result for each user intent; wherein the first weight and the second weight are obtained by weighted summation of speech result confidence, contextual relevance and the number of successful historical arbitrations.

[0006] According to an embodiment of the present invention, adopting the higher weight value as the semantic result includes: in response to the first speech result and the second speech result including more than one user intent, sorting the importance of the user intent according to a preset priority; and adopting the higher weight value in order of importance to output the semantic result.

[0007] According to an embodiment of the present invention, the user intent includes at least a control intent, a navigation intent, and an entertainment intent; the preset priority includes a traffic jam scenario priority and a startup scenario priority; the importance of the traffic jam scenario priority, from high to low, is the control intent, the entertainment intent, and the navigation intent; the importance of the startup scenario priority, from high to low, is the control intent, the navigation intent, and the entertainment intent.

[0008] According to an embodiment of the present invention, adopting the higher weight value as the semantic result, the method further includes: adjusting the response timeout threshold of online identification according to the network status; the response timeout threshold is dynamically adjusted by a preset baseline threshold, the current network latency, and the packet loss rate.

[0009] According to an embodiment of the present invention, the step of adopting the higher weight value as the semantic result further includes: determining whether the difference between the first weight and the second weight is less than a preset adjustment threshold; if so, directly selecting the first speech result as the semantic result; if not, performing the step of adopting the higher weight value as the semantic result.

[0010] According to an embodiment of the present invention, the higher weight value is adopted as the semantic result, and the method further includes: executing the user intent through the vehicle control bus based on the semantic result.

[0011] Secondly, the present invention provides a semantic arbitration system capable of implementing the above-mentioned semantic arbitration method, comprising: an end-side recognition module for offline recognition of user speech to obtain a first speech result; a cloud-based recognition module for online recognition of user speech to obtain a second speech result; the first speech result and the second speech result include at least one user intent; and an integrated arbitration module for calculating a first weight of the first speech result and a second weight of the second speech result based on contextual relevance and the number of successful historical arbitrations, and adopting the one with the higher weight value as the semantic result; the number of successful historical arbitrations is the number of times the speech results output by offline and online recognition have been adopted within a preset number of times.

[0012] A third aspect of the present invention provides an electronic device comprising: one or more processors; and a memory for storing one or more programs, wherein, when the one or more programs are executed by the one or more processors, the one or more processors perform the semantic arbitration method described above.

[0013] A fourth aspect of the present invention also provides a computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, cause the processor to perform the semantic arbitration method described above.

[0014] The semantic arbitration method provided by this invention obtains a first speech result and a second speech result by recognizing user speech offline and online. It dynamically calculates weights based on contextual relevance and the number of successful arbitrations in the past to adopt the higher-weighted result. By introducing contextual relevance and historical performance as factors for dynamic weight calculation, it breaks through the limitations of traditional semantic arbitration and thus at least partially solves the technical problem of low accuracy of arbitration decisions when there is a conflict between online and offline semantic results. Attached Figure Description

[0015] Figure 1A flowchart illustrating a semantic arbitration method provided in an embodiment of the present invention;

[0016] Figure 2 This is an application block diagram of the method provided in the embodiments of the present invention;

[0017] Figure 3 This is a timing diagram of the arbitration process of the method provided in the embodiments of the present invention;

[0018] Figure 4 A structural block diagram of a semantic arbitration system provided in an embodiment of the present invention;

[0019] Figure 5 This is a structural block diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0020] To enable those skilled in the art to better understand the technical solutions of the present invention, exemplary embodiments of the present invention are described below in conjunction with the accompanying drawings, including various details of the embodiments of the present invention to aid understanding. These should be considered merely exemplary. Therefore, those skilled in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present invention. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0021] Where there is no conflict, the various embodiments of the present invention and the features thereof may be combined with each other.

[0022] As used herein, the term “and / or” includes any and all combinations of one or more related enumerated entries.

[0023] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used herein, the singular forms “a” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will also be understood that when the terms “comprising” and / or “made of” are used in this specification, the presence of the stated feature, integral, step, operation, element, and / or component is specified, but the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or groups thereof is not excluded. Terms such as “connected” or “linked” are not limited to physical or mechanical connections but can include electrical connections, whether direct or indirect.

[0024] Unless otherwise specified, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art. It will also be understood that terms such as those defined in commonly used dictionaries should be interpreted as having the meaning consistent with their meaning in the context of the relevant art and the invention, and will not be interpreted as having an idealized or overly formal meaning unless expressly so defined herein.

[0025] In the technical solution of this invention, the collection, storage, use, processing, transmission, provision, and disclosure of user personal information all comply with relevant laws and regulations and do not violate public order and good morals. The use of user data in this technical solution follows relevant national laws and regulations (e.g., the "Information Security Technology - Personal Information Security Specification"). For example: appropriate measures are taken for personal information access control; restrictions are imposed on the display of personal information; the purpose of using personal information does not exceed the scope of direct or reasonable association; and explicit identity targeting is eliminated when using personal information to avoid precisely locating a specific individual.

[0026] Traditional arbitration methods have the following drawbacks: First, they do not consider contextual relevance; second, cloud response timeout thresholds are fixed and do not dynamically adjust according to network conditions, making them prone to misjudgment in weak network environments; third, for multi-intent commands (e.g., navigating to the company and playing music), they only arbitrate sequentially without considering intent priority (e.g., navigation intent should take precedence over entertainment intent). In other words, traditional arbitration logic, which relies solely on first-come, first-served or fixed-priority decisions, leads to insufficient accuracy and a poor user experience in complex scenarios.

[0027] To address the issues of conflicting offline and online semantic results and low accuracy of arbitration decisions, this invention provides a semantic arbitration method. Figure 1 This is a flowchart illustrating a semantic arbitration method provided in an embodiment of the present invention, as shown below. Figure 1 As shown, this embodiment of the invention provides a semantic arbitration method, which includes: offline recognition of user speech to obtain a first speech result; online recognition of user speech to obtain a second speech result; the first speech result and the second speech result include at least one user intent; calculating a first weight of the first speech result and a second weight of the second speech result based on contextual relevance and the number of successful historical arbitrations, and adopting the one with the higher weight value as the semantic result; the number of successful historical arbitrations is the number of times the speech results output by offline and online recognition have been adopted within a preset number of times.

[0028] In this embodiment, offline recognition is performed by a local offline processing module in the vehicle, operating without a network. Online recognition is achieved through an online processing service on a remote server, relying on network transmission. Both offline and online recognition are implemented using Automatic Speech Recognition (ASR), converting speech signals into text. This is achieved on the device side and in the cloud via an offline ASR engine and an online ASR service, respectively. Natural language understanding is then used to analyze the text semantics of the speech results.

[0029] In this embodiment, after obtaining the semantic results, the text information is converted into voice output for broadcasting the voice interaction results.

[0030] Through the embodiments of the present invention, contextual relevance and historical records are incorporated into the weight calculation to resolve result conflicts and improve the accuracy and real-time performance of arbitration.

[0031] Based on the above embodiments, the first weight of the first voice result and the second weight of the second voice result are calculated according to the contextual relevance and historical arbitration records. This includes: in response to the fact that the number of user intents included in the first voice result and the second voice result is greater than one, the first weight of the first voice result and the second weight of the second voice result are calculated for each user intent. The first weight and the second weight are obtained by weighted summation of the voice result confidence, contextual relevance and the number of successful historical arbitrations.

[0032] In this embodiment, the dynamic weight calculation is performed based on the following formula:

[0033] W 端 =a×C 端 +b×H 端 +c×S 端 ,

[0034] W 云 =a×C 云 +b×H 云 +c×S 云 ,

[0035] Where C represents the semantic result confidence score, ranging from 0 to 100, output by offline or online recognition; H represents the number of successful arbitrations in history (e.g., the number of times the result on this side was adopted in the last 100 interactions); S represents the contextual relevance score (0 to 1, calculated based on the matching degree between the user's previous command and the current intent); and a, b, and c are preset dynamic adjustment coefficients. It should be noted that a, b, and c can be adjusted according to the scenario.

[0036] Through the embodiments of the present invention, the weights of the terminal and cloud results are calculated separately for each user intent in a multi-intent instruction. The weights are obtained by weighted summation of confidence, contextual relevance, and historical success count. This refines the processing of multi-intents and ensures that the arbitration of each intent is based on multi-dimensional dynamic evaluation, thereby improving the rationality and accuracy of arbitration in multi-intent scenarios.

[0037] Based on the above embodiments, adopting the higher weight value as the semantic result includes: in response to the first voice result and the second voice result including more than one user intent, sorting the importance of user intents according to preset priority; and adopting the higher weight value in order of importance to output the semantic result.

[0038] Through the embodiments of the present invention, instructions containing multiple user intentions are prioritized and high-priority intentions are processed first. An intention priority rule is introduced on the basis of weight arbitration to ensure that high-priority intentions are processed first, making multi-intention processing more in line with driving habits.

[0039] Based on the above embodiments, user intent includes at least control intent, navigation intent, and entertainment intent; preset priorities include traffic jam scenario priority and startup scenario priority; in the traffic jam scenario priority, the order of importance from high to low is control intent, entertainment intent, and navigation intent; in the startup scenario priority, the order of importance from high to low is control intent, navigation intent, and entertainment intent.

[0040] In this embodiment, the multi-intent priority arbitrator prioritizes vehicle control intents (such as emergency braking), navigation intents, and entertainment intents (such as playing music) based on the following rules during vehicle operation. Priorities are dynamically adjusted based on the scenario, using a preset priority template for the driving scenario. For example, navigation intents have higher priority in highway scenarios, while entertainment intents have higher priority in parking scenarios.

[0041] Optionally, the priority of specific actions in each category can be adjusted based on historical frequency. For example, if a user frequently says "turn on the air conditioner," then that action has a higher priority than a low-frequency intention.

[0042] Through the embodiments of the present invention, by combining safety priority, scenario relevance and user habits, the intent priority can be adapted to specific driving scenarios, thereby improving the rationality of multi-intent arbitration and the smoothness of interaction.

[0043] Based on the above embodiments, the semantic result with higher weight values ​​is adopted. The method also includes: adjusting the response timeout threshold of online identification according to the network status; the response timeout threshold is dynamically adjusted by a preset baseline threshold, the current network latency, and the packet loss rate.

[0044] In this embodiment, T = T0 × (1 + k1 × D + k2 × L), where T is the response timeout threshold, T0 is the preset baseline threshold, D is the current network latency, L is the packet loss rate, and k1 and k2 are constant adjustment coefficients.

[0045] Through the embodiments of the present invention, the response timeout threshold of online identification is dynamically adjusted according to the network status, so that the timeout threshold can be adaptively adjusted with network fluctuations, and the online identification results are lost under weak network conditions.

[0046] Based on the above embodiments, the method of adopting the higher weight value as the semantic result further includes: determining whether the difference between the first weight and the second weight is less than a preset adjustment threshold; if so, directly selecting the first speech result as the semantic result; if not, executing the step of adopting the higher weight value as the semantic result.

[0047] Through the embodiments of the present invention, the arbitration strategy is determined by judging whether the weight difference is less than a preset threshold, and the decision is made quickly when the weights are similar, thereby improving the processing efficiency.

[0048] Based on the above embodiments, the higher weight value is adopted as the semantic result, and then the method further includes: executing the user intent through the vehicle control bus according to the semantic result.

[0049] Through embodiments of the present invention, the user experience is improved by invoking the vehicle control bus to execute user intents based on semantic results.

[0050] Figure 2 This is an application block diagram of the method provided in the embodiments of the present invention, such as... Figure 2 As shown, the system includes a terminal-side semantic processing module, a cloud-based semantic processing module, a dynamic weight calculation engine, a dynamic timeout management module, and a multi-intent priority arbitrator. The terminal-side semantic processing module parses user commands based on an offline engine, outputting terminal-side semantic results (including intent tags, confidence levels, and parameters), and pushes the results to the dynamic weight calculation engine via the WebSocket protocol. The cloud-based semantic processing module parses commands based on a cloud protocol adaptation service, outputting cloud-based semantic results, and pushes the results to the dynamic weight calculation engine via the HTTP / 2 protocol. The dynamic weight calculation engine dynamically calculates the weights of the terminal-side and cloud-based results, considering contextual relevance, historical arbitration records, and network status. It outputs the weight difference to the multi-intent priority arbitrator. The dynamic timeout management module adjusts the cloud response timeout threshold based on real-time network status (latency, packet loss rate) and updates the network stability index to the dynamic weight calculation engine. The multi-intent priority arbitrator sorts multi-intent commands by priority and outputs the final arbitration result based on the weight difference. The result is then output to the vehicle control bus. It should be noted that if no cloud result is received within the timeout period, the terminal-side result is used.

[0051] Figure 3This is a timing diagram of the arbitration process of the method provided in the embodiments of the present invention, as shown below. Figure 3 As shown, user input triggers parallel processing on both the edge and cloud sides; the semantic processing module on the edge and the semantic processing module on the cloud side parse the user command in parallel, and the edge and cloud sides output semantic results and confidence scores respectively; the dynamic weight calculation engine integrates the results from both ends, and calculates the dynamic weight of the results from both ends based on contextual relevance, historical arbitration records and network status; the dynamic timeout module feeds back network parameters to the cloud and weight engine in real time to confirm that the cloud response has not timed out; the multi-intent arbitrator determines it as a single intent and directly adopts the cloud result with higher weight.

[0052] Figure 4 A structural block diagram of a semantic arbitration system provided in an embodiment of the present invention is shown below. Figure 4 As shown, the present invention provides a semantic arbitration system that can be used to implement the above-mentioned semantic arbitration method, including: an end-side recognition module for offline recognition of user speech to obtain a first speech result; a cloud-based recognition module for online recognition of user speech to obtain a second speech result; the first speech result and the second speech result include at least one user intent; and an integrated arbitration module for calculating a first weight of the first speech result and a second weight of the second speech result based on contextual relevance and the number of successful historical arbitrations, and adopting the one with the higher weight value as the semantic result; the number of successful historical arbitrations is the number of times the speech results output by offline and online recognition have been adopted within a preset number of times.

[0053] Based on the same inventive concept, embodiments of the present invention also provide an electronic device. Figure 5 This is a structural block diagram of an electronic device provided in an embodiment of the present invention. Figure 5 As shown, an embodiment of the present invention provides an electronic device including: one or more processors 101, a memory 102, and one or more I / O interfaces 103. The memory 102 stores one or more programs, which, when executed by the one or more processors, cause the one or more processors to implement any of the semantic arbitration methods described in the above embodiments; the one or more I / O interfaces 103 are connected between the processors and the memory, configured to enable information interaction between the processors and the memory.

[0054] The processor 101 is a device with data processing capabilities, including but not limited to a central processing unit (CPU); the memory 102 is a device with data storage capabilities, including but not limited to random access memory (RAM, more specifically SDRAM, DDR, etc.), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), and flash memory (FLASH); the I / O interface (read / write interface) 103 is connected between the processor 101 and the memory 102, and can realize information interaction between the processor 101 and the memory 102, including but not limited to a data bus (Bus).

[0055] In some embodiments, the processor 101, memory 102, and I / O interface 103 are interconnected via bus 104, and thus connected to other components of the computing device.

[0056] In some embodiments, the one or more processors 101 include a field-programmable gate array.

[0057] This invention also provides a computer-readable medium. The computer-readable medium stores a computer program, which, when executed by a processor, implements the steps of any of the semantic arbitration methods described in the above embodiments. The computer-readable storage medium may be volatile or non-volatile.

[0058] This invention also provides a computer program product, including computer-readable code, or a non-volatile computer-readable storage medium carrying computer-readable code. When the computer-readable code is run in the processor of an electronic device, the processor in the electronic device executes the above-described semantic arbitration method.

[0059] Those skilled in the art will understand that all or some of the steps, systems, and apparatuses disclosed above, and their functional modules / units, can be implemented as software, firmware, hardware, or suitable combinations thereof. In hardware implementations, the division between functional modules / units mentioned above does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed collaboratively by several physical components. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit (ASIC). Such software can be distributed on a computer-readable storage medium, which may include computer storage media (or non-transitory media) and communication media (or transient media).

[0060] As is known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information, such as computer-readable program instructions, data structures, program modules, or other data. Computer storage media includes, but is not limited to, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), static random access memory (SRAM), flash memory or other memory technologies, portable compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, it is known to those skilled in the art that communication media typically contain computer-readable program instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0061] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0062] The computer program instructions used to perform the operations of this invention may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing state information from the computer-readable program instructions. This electronic circuitry can execute the computer-readable program instructions to implement various aspects of the invention.

[0063] The computer program product described herein can be implemented specifically through hardware, software, or a combination thereof. In one alternative embodiment, the computer program product is specifically embodied in a computer storage medium; in another alternative embodiment, the computer program product is specifically embodied in a software product, such as a software development kit (SDK), etc.

[0064] Various aspects of the present invention are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0065] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processor of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner; thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0066] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0067] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction, which contains one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those shown in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0068] Example embodiments have been disclosed herein, and while specific terminology has been used, it is for illustrative purposes only and should be construed as such, and is not intended to be limiting. In some instances, it will be apparent to those skilled in the art that features, characteristics, and / or elements described in conjunction with particular embodiments may be used alone, or in combination with features, characteristics, and / or elements described in conjunction with other embodiments, unless otherwise expressly indicated. Therefore, those skilled in the art will understand that various changes in form and detail may be made without departing from the scope of the invention as set forth in the appended claims.

Claims

1. A semantic arbitration method, characterized in that, include: Offline recognition of user speech to obtain the first speech result; Online recognition of user speech to obtain a second speech result; The first voice result and the second voice result each include at least one user intent; Based on the contextual relevance and the number of successful arbitrations in the past, the first weight of the first speech result and the second weight of the second speech result are calculated, and the one with the higher weight value is adopted as the semantic result. The historical arbitration success count refers to the number of times the voice results output by offline and online recognition were adopted within a preset number of times.

2. The method according to claim 1, wherein, The step of calculating the first weight of the first speech result and the second weight of the second speech result based on contextual relevance and historical arbitration records includes: If the number of user intents included in the first voice result and the second voice result is greater than one, then for each user intent, a first weight of the first voice result and a second weight of the second voice result are calculated. The first weight and the second weight are obtained by weighted summation of the confidence level of the speech result, the contextual relevance, and the number of successful arbitrations in the past.

3. The method according to claim 2, wherein, The higher adoption weight value is used as the semantic result, including: If the number of user intents included in the first voice result and the second voice result is greater than one, then the importance of the user intents is ranked according to a preset priority. The semantic results are output in descending order of importance, with the highest weight values ​​being adopted in sequence.

4. The method according to claim 3, wherein, The user intent includes at least control intent, navigation intent, and entertainment intent; the preset priority includes traffic jam scenario priority and startup scenario priority; the importance of the traffic jam scenario priority, from highest to lowest, is control intent, entertainment intent, and navigation intent; The priority of the launch scenarios, from highest to lowest importance, is: control intent, navigation intent, and entertainment intent.

5. The method according to claim 1, wherein, The adoption weight value with the higher value is used as the semantic result, which previously included: The response timeout threshold for online identification is adjusted based on the network status; the response timeout threshold is dynamically adjusted by a preset baseline threshold, the current network latency, and the packet loss rate.

6. The method according to claim 1, wherein, The adoption weight value with the higher value is used as the semantic result, which previously included: Determine whether the difference between the first weight and the second weight is less than a preset adjustment threshold; If so, the first speech result is directly selected as the semantic result; If not, then proceed with the step of adopting the one with the higher weight value as the semantic result.

7. The method according to claim 1, wherein, The higher adoption weight value is taken as the semantic result, which is then followed by: Based on the semantic results, the user intent is executed via the vehicle control bus.

8. A semantic arbitration system, characterized in that, Capable of implementing the method as described in any one of claims 1 to 7, comprising: The edge recognition module is used to recognize user speech offline and obtain the first speech result; A cloud-based recognition module is used to recognize user speech online and obtain a second speech result; the first speech result and the second speech result include at least one user intent; An integrated arbitration module is used to calculate the first weight of the first speech result and the second weight of the second speech result based on the contextual relevance and the number of successful historical arbitrations, and to adopt the one with the higher weight value as the semantic result; the number of successful historical arbitrations refers to the number of times the speech results output by offline recognition and online recognition have been adopted within a preset number of times.

9. An electronic device, characterized in that, include: One or more processors; Memory, used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1 to 7.

10. A computer-readable medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 7.

Images