Vehicle-machine interface interaction method based on man-machine interaction

Abstract

The invention relates to the technical field of man-machine interaction, in particular to a man-machine interaction-based vehicle-machine interface interaction method, which comprises the following steps of: collecting information such as voice input and touch operation of a user, and performing quantitative expression by using a vector space model, so as to evaluate the association degree between the user and a vehicle-machine function and realize accurate identification of a user intention; according to the method, the identification result of the user intention and the state information of the in-vehicle infotainment system are combined, a Bayesian network is adopted to analyze an interaction scene, and an interaction response mechanism is dynamically adjusted to adapt to user requirements in different scenes; the weights of the function modules of the vehicle machine interface are determined by applying an analytic hierarchy process, and the interface layout is optimized, so that the convenience and efficiency of user interaction are improved; the Markov chain model is constructed according to the user feedback, the in-vehicle infotainment feedback strategy is adjusted, and the interaction performance and the user experience of the in-vehicle infotainment system are improved. The method and the device are used for solving the technical problem of poor user interaction experience in a vehicle-mounted information system.

Description

Technical Field

[0001] The present invention relates to the field of artificial interaction technology, and in particular to a vehicle-machine interface interaction method based on artificial interaction. Background Art

[0002] Artificial interaction technology is a professional field of computer science. It studies the interaction between people and computers, aiming to improve the naturalness, efficiency and satisfaction of this interaction. This field involves many core technologies such as interaction design, artificial intelligence, speech recognition and synthesis, and natural language processing.

[0003] The current vehicle information system management and control in the automotive industry still faces some challenges, including the following aspects: poor user interaction experience. The existing vehicle-machine interface interaction methods may have errors, resulting in the inability to accurately meet user needs, inconvenient user operation and poor experience; the interactive response mechanism is unreasonable and not flexible enough. The previous vehicle-machine interface interaction methods failed to fully consider the vehicle-machine system status information and user needs in different interaction scenarios, resulting in the interactive response mechanism being not personalized and flexible enough; the interface layout is not intuitive and has poor usability. The current vehicle-machine interface layout has not been reasonably optimized according to the importance of functional modules and user usage frequency, resulting in users having difficulty in quickly finding and using required functions; the adjustment strategy for user feedback lacks pertinence and effectiveness. Some vehicle-machine interface interaction methods may lack dynamic adjustment in feedback strategies based on user behavior feedback data, resulting in an unoptimized user experience; in response to these challenges, corresponding solutions and technical means need to be adopted to improve the interaction efficiency and user experience of the vehicle-machine system. Summary of the invention

[0004] The purpose of the present invention is to solve the problems in the background technology and to propose a vehicle-machine interface interaction method based on manual interaction.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions: A vehicle-machine interface interaction method based on manual interaction, comprising: Step 1: The car interface collects multi-source interaction information and uses the vector space model to quantify it to deeply identify user intentions; The multi-source interaction information includes user voice input or touch operation; Step 2: Combine the recognition result of user intention with the vehicle system status information, use Bayesian network to analyze the interaction scenario, and formulate the interaction response mechanism under the current interaction scenario; The vehicle system status information is the sensor data such as vehicle speed, vehicle environment (temperature, wind speed), current time, etc. These status information may affect the user's interaction needs and appropriate response methods. For example, when the vehicle is traveling at a high speed, some complex operations (such as long text input operations) may not be appropriate, and simple voice interaction and quick response operations are more preferred. Step 3: Based on the results of the interaction scenario analysis, the weight of each functional module is determined using the hierarchical analysis method for the functional modules included in the vehicle-machine interface, thereby optimizing the layout of the vehicle-machine interface; Among them, the vehicle computer interface includes several functional modules including navigation function, multimedia function, vehicle status display function, etc. The importance of these functional modules may vary in different interaction scenarios; the layout of the vehicle computer interface includes display size, position, color contrast, etc.; Step 4: Based on the user's reaction after the vehicle computer provides feedback, the user behavior feedback is modeled by constructing a Markov chain to adjust the feedback strategy of the vehicle computer.

[0006] It should be noted that the vehicle-machine interface interaction method based on manual interaction proposed in the present invention can be widely applicable to the management and control of in-vehicle information systems in the automotive industry. The method can be used to improve the interaction efficiency and user experience of the vehicle-machine interface. Specifically, it can be through vector space models, Bayesian networks, hierarchical analysis methods, Markov chain and other technologies to comprehensively analyze and evaluate the user interaction information, vehicle-machine system status information, and user behavior feedback data. This process covers multiple links such as user intention recognition, interaction scenario analysis, interface layout optimization, and feedback strategy adjustment; these links jointly ensure the accurate matching of user intentions, the rationality of the interactive response mechanism, the ease of use of the interface layout, and the effectiveness of the feedback strategy, thereby improving the intelligence level of the in-vehicle information system and user satisfaction.

[0007] Furthermore, the car interface collects multi-source interactive information and uses the vector space model to quantify it to deeply identify the user's intention. The process includes: When the user interacts with the vehicle interface, multi-source interaction information is collected: the user voice input signal is defined as ;in, represents the time variable, and is a multidimensional vector containing frequency, amplitude, and phase information; Mark user touch actions as ;in, Indicates the touch position, Indicates the touch strength. Indicates the time when the touch occurs; The vector space model is used to quantify the collected multi-source interaction information, including: For user voice input, it is converted into text form through voice recognition to obtain a text vector , where Represents the preset text feature quantization value, Indicates the number of pre-set features; For user touch operations, the pre-set functions and corresponding operation instructions of the vehicle interface are converted into touch vectors , where Indicates the quantized value of the corresponding feature in the pre-set vehicle function command; For example, the frequency of occurrence of a word or phrase in the text converted by speech recognition (if the word or phrase is in the vocabulary library related to the vehicle computer function), or the encoding of the screen area corresponding to the touch operation (according to the pre-divided screen area rules), etc. By calculating the cosine similarity between the text vector and the touch vector, the user intention is matched to the corresponding vehicle function to identify the user intention; the cosine similarity calculation formula is: , In the formula, represents the cosine similarity between the text vector and the touch vector; is the dot product of vectors, which represents the sum of the products of two vectors in each dimension; and They are text vectors and touch vector The modulus of , which is the length of the vector; Represents the feature index; It can be understood that the value of cosine similarity is between -1 and 1. If the value of cosine similarity is closer to 1, it means that the two vectors are more similar, that is, the user's interaction information is more matched with a certain functional instruction of the car computer; by calculating the cosine similarity between the two vectors, the degree of correlation between the user input (user voice input and touch operation) and the car computer function is evaluated, thereby completing the precise matching of user intentions and identifying user intentions.

[0008] Furthermore, the recognition result of user intention is combined with the vehicle system status information, and the process of using Bayesian network to analyze the interaction scenario includes: After determining the user's intention, the vehicle system status information is collected through sensors; Set the user intent to , It is the different status information of the vehicle system; among them, Indicates the number of different status information of the vehicle system; is a multi-valued variable. Can be the vehicle speed, It can be the ambient temperature in the car, etc. Get the currently observed user input or operation and mark it as ;in, is a discrete variable; The Bayesian formula is used to calculate the probability of user intention under different system states; the Bayesian formula is specifically: , where Represents the user input or action at a given observation The vehicle system status is Under the condition of The posterior probability of Indicates that given user intent Under the condition of The vehicle system status is The joint probability of Indicates user intent Prior probability of occurrence; Represents observed user input or action The vehicle system status is The joint probability of .

[0009] Furthermore, the process of formulating the interactive response mechanism in the current interactive scenario includes: Step B1: For each user intent Assign a prior probability ; Step B2: For each user intent and every possible observable user input or action And the vehicle system status , define the conditional probability and ; Step B3: For a given observed user input or operation and vehicle system status , using the Bayesian formula to calculate each user's intention The posterior probability ; Step B4: Select the user intent with the highest posterior probability As the intention of the current interaction scenario; Step B5: Based on the selected user intent , get the corresponding user intent estimate: , where represents an estimate of the user's intent, Indicates Maximum user intent The value of Step B6: predefine a response set; wherein the set includes all possible interactive response mechanisms, each of which corresponds to a specific behavior or operation, such as displaying a certain information, performing a certain action, changing the interface layout, etc.; Step B7: according to the estimated value of user intention obtained in step B5, searching for the response mechanism that best matches it in the predefined response set; when the response mechanism that best matches the estimated value of user intention is found, it is used as the interaction response mechanism in the current interaction scenario; Step B8: Execute the selected interactive response mechanism to respond to the user's input or operation.

[0010] Furthermore, based on the results of the interaction scenario analysis, the weight of each functional module is determined by using the hierarchical analysis method for several functional modules included in the vehicle computer interface, and the process of optimizing the layout of the vehicle computer interface includes: Assume that the vehicle interface exists Functional modules, build judgment matrix ;in, Indicates the vehicle interface function module For functional modules The importance of All are function module indexes. Indicates the number of functional modules; for example, if , it means that function module 1 is three times more important than function module 2 in the current interaction scenario; It can be understood that the judgment matrix is ​​a matrix used in the hierarchical analysis method to express the results of pairwise comparisons between factors at the same level with respect to the importance of a criterion at the previous level; Using the characteristic equation , calculate the judgment matrix The maximum eigenvalue of and its corresponding eigenvector , and solved by iterative method; It is understandable that in this context, features refer to features related to the functional modules of the vehicle-machine interface. These features reflect the relative importance of the functional modules. For example, for the navigation function and multimedia function in the vehicle-machine interface, their relative importance in different interaction scenarios (such as the urgent need to navigate to the destination while driving or the preference for multimedia functions when parking for leisure) is a feature; Quantified the vehicle interface functional modules Relative to functional modules The importance of the function, which may be based on factors such as the frequency and urgency of user use of the function in a specific scenario; In the feature vector When normalized, the eigenvector The normalized vector of ,in, is the eigenvector The Euclidean norm of is, where represents the number of eigenvectors; According to the normalized weight vector, the relative importance of each functional module in the interface layout is determined, and the functional modules in the interface layout are adjusted, including but not limited to the display size and position of the functional modules; for example, if the weight of a functional module is higher, its display area can be increased or its position can be changed. Specifically, the functional module with a larger weight can be designed to have a larger display size or a more conspicuous position to attract the user's attention, and the functional module with a smaller weight can be designed to have a smaller display size or a less conspicuous position.

[0011] Furthermore, the process of using the characteristic equation to calculate the maximum eigenvalue of the judgment matrix and its corresponding eigenvector and solving it by an iterative method includes: Step C1, rewrite the equation as ,Right now , where represents the identity matrix; by solving the characteristic polynomial , find the eigenvalue ,in, Represents eigenvalue quantity; Step C2: Select an initial vector ; Step C3: calculate , where Indicates the number of iterations; Step C4: Normalization ,Right now ; Step C5: Calculation ; Step C6, repeat steps C3-C5 until Converge to the preset value to obtain the maximum eigenvalue ; Among them, the choice of initial vector will affect the convergence speed of the iterative process and the final eigenvector. The unit vector is selected as the initial vector to ensure that the modulus of the vector remains in a reasonable range during the iteration process. The more iterations there are, the more likely the algorithm is to converge to accurate eigenvalues ​​and eigenvectors. In practical applications, you can determine whether to stop the iteration by setting an upper limit on the number of iterations or based on the change in the eigenvalue. During the iteration process, the eigenvector will gradually converge to the eigenvector corresponding to the maximum eigenvalue.

[0012] Furthermore, according to the user's reaction after the vehicle computer provides feedback, the process of modeling the user behavior feedback by constructing a Markov chain to adjust the feedback strategy of the vehicle computer includes: Assume the feedback state space is , A state set representing possible user behavior feedback (e.g. Express satisfaction, Express dissatisfaction, Indicates that further action is required, etc.), where: Indicates the number of possible behavior feedback states of the user; Building the transition probability matrix , and complete the initialization: the transition probability matrix is a The matrix represents the feedback state from user behavior Transfer to state The probability of transition probability matrix Elements Defined as: , that is, the user's current behavior feedback status Next, transfer to the state The probability of, where are all indexes of user behavior feedback status. Represents the probability value; in the initial stage, the transition probability matrix Initialize, for example, assuming that initially the probability of the user transferring from the "satisfied" feedback state to the "satisfied" state is high, and the probability of transferring from the "unsatisfied" feedback state to the "need further action" state is also high; Based on the interaction between the user and the vehicle system, the user's behavior feedback data is continuously collected, and the transition probability matrix is ​​updated according to the collected user behavior feedback data ; Getting the current transition probability matrix and the user's current behavior feedback status After that, predict the user's next possible behavior feedback state, that is, ; Adjust the vehicle computer's feedback strategy based on the predicted user's next possible behavior feedback; for example, if the user is more likely to be "unsatisfied", the feedback method can be changed, such as adjusting the tone of the voice prompt or re-optimizing the interface display; It can be understood that a Markov chain model is constructed by defining the feedback state space and the transition probability matrix; by initializing and updating the transition probability matrix, the model can reflect the changes in actual data; and by predicting the user's next behavior and adjusting the vehicle feedback strategy, the user experience is optimized.

[0013] Compared with the existing technology, the advantages of the vehicle-machine interface interaction method based on manual interaction provided by the present invention are: The present invention uses a vector space model to quantify the user's voice input and touch operation, and uses the calculation of cosine similarity to match the user's intention with the vehicle machine function, which can more accurately identify the user's intention and reduce the misrecognition rate, thereby improving the user experience. At the same time, considering both voice and touch interaction modes, the vehicle machine interface can interact with the user more naturally, in line with the user's daily operating habits, and improve the convenience and comfort of interaction. The present invention combines user intention and vehicle system status information, uses Bayesian network to analyze the interaction scene, and can formulate an interaction response mechanism that is more in line with the current scene and user needs, realize personalized interaction, and effectively improve user satisfaction; by using the hierarchical analysis method to determine the weight of each functional module, and adjusting the interface layout according to the weight, the important functional modules are more prominent, which is convenient for users to quickly find and use, thereby improving the interaction efficiency; The present invention constructs a Markov chain model based on user behavior feedback data and continuously updates the transition probability matrix, so as to predict the user's possible next behavior feedback, thereby dynamically adjusting the feedback strategy of the vehicle computer, improving the pertinence and effectiveness of the feedback, and by dynamically adjusting the feedback strategy, it can better meet user needs, reduce user dissatisfaction and complaints, and enhance user satisfaction and loyalty.

[0014] In summary, the present invention realizes accurate identification of user intentions, dynamic analysis of interaction scenarios, reasonable optimization of interface layout, and continuous adjustment of feedback strategies through the interrelated and mutually reinforcing steps above, which significantly improves the interactive performance and user experience of the vehicle-machine system and ensures the normal implementation of a subsequent vehicle-machine interface interaction method based on manual interaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 The present invention provides a flow chart of a vehicle-machine interface interaction method based on manual interaction. DETAILED DESCRIPTION

[0016] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the implementation regulations described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0017] Reference Figure 1 , a vehicle-machine interface interaction method based on manual interaction, comprising: Step 1: The car interface collects multi-source interaction information and uses the vector space model to quantify it to deeply identify user intentions; The multi-source interaction information includes user voice input or touch operation; Step 2: Combine the recognition result of user intention with the vehicle system status information, use Bayesian network to analyze the interaction scenario, and formulate the interaction response mechanism under the current interaction scenario; The vehicle system status information is the sensor data such as vehicle speed, vehicle environment (temperature, wind speed), current time, etc. These status information may affect the user's interaction needs and appropriate response methods. For example, when the vehicle is traveling at a high speed, some complex operations (such as long text input operations) may not be appropriate, and simple voice interaction and quick response operations are more preferred. Step 3: Based on the results of the interaction scenario analysis, the weight of each functional module is determined using the hierarchical analysis method for the functional modules included in the vehicle-machine interface, thereby optimizing the layout of the vehicle-machine interface; Among them, the vehicle computer interface includes several functional modules including navigation function, multimedia function, vehicle status display function, etc. The importance of these functional modules may vary in different interaction scenarios; the layout of the vehicle computer interface includes display size, position, color contrast, etc.; Step 4: Based on the user's reaction after the vehicle computer provides feedback, the user behavior feedback is modeled by constructing a Markov chain to adjust the feedback strategy of the vehicle computer.

[0018] It should be noted that the vehicle-machine interface interaction method based on manual interaction proposed in the present invention can be widely applicable to the management and control of in-vehicle information systems in the automotive industry. The method can be used to improve the interaction efficiency and user experience of the vehicle-machine interface. Specifically, it can be through vector space models, Bayesian networks, hierarchical analysis methods, Markov chain and other technologies to comprehensively analyze and evaluate the user interaction information, vehicle-machine system status information, and user behavior feedback data. This process covers multiple links such as user intention recognition, interaction scenario analysis, interface layout optimization, and feedback strategy adjustment; these links jointly ensure the accurate matching of user intentions, the rationality of the interactive response mechanism, the ease of use of the interface layout, and the effectiveness of the feedback strategy, thereby improving the intelligence level of the in-vehicle information system and user satisfaction.

[0019] See also Figure 1 The present invention provides a vehicle-machine interface interaction method based on manual interaction. In step 1, the vehicle-machine interface collects multi-source interaction information and uses a vector space model to quantify and represent it, so as to deeply identify the user's intention. The steps include: Step 101: When the user interacts with the vehicle computer interface, collect multi-source interaction information: Define the user voice input signal as ;in, represents the time variable, and is a multidimensional vector containing frequency, amplitude, and phase information; Mark user touch actions as ;in, Indicates the touch position, Indicates the touch strength. Indicates the time when the touch occurs; Step 102: using a vector space model to quantify the collected multi-source interaction information, specifically including: For user voice input, it is converted into text form through voice recognition to obtain a text vector , where Represents the preset text feature quantization value, Indicates the number of pre-set features; For user touch operations, the pre-set functions and corresponding operation instructions of the vehicle interface are converted into touch vectors , where Indicates the quantized value of the corresponding feature in the pre-set vehicle function command; For example, the frequency of occurrence of a word or phrase in the text converted by speech recognition (if the word or phrase is in the vocabulary library related to the vehicle computer function), or the encoding of the screen area corresponding to the touch operation (according to the pre-divided screen area rules), etc. Step 103: By calculating the cosine similarity between the text vector and the touch vector, the user intention is matched to the corresponding vehicle function to identify the user intention; wherein the cosine similarity calculation formula is: , In the formula, represents the cosine similarity between the text vector and the touch vector; is the dot product of vectors, which represents the sum of the products of two vectors in each dimension; and They are text vectors and touch vector The modulus of , which is the length of the vector; Represents the feature index; In step 103, the value of cosine similarity is between -1 and 1. If the value of cosine similarity is closer to 1, it means that the two vectors are more similar, that is, the user's interaction information matches a certain function instruction of the vehicle computer more closely. By calculating the cosine similarity between the two vectors, the correlation between the user input (user voice input and touch operation) and the vehicle computer function is evaluated, thereby completing the precise matching of user intentions and identifying user intentions.

[0020] See also Figure 1 The present invention provides a vehicle-machine interface interaction method based on manual interaction. In the step 2, the recognition result of the user's intention is combined with the vehicle-machine system status information, the interaction scene is analyzed by using a Bayesian network, and the step of formulating an interaction response mechanism under the current interaction scene includes: Step 201: After determining the user's intention, collect vehicle system status information through sensors; Step 202: Set the user intention to , It is the different status information of the vehicle system; among them, Indicates the number of different status information of the vehicle system; is a multi-valued variable. Can be the vehicle speed, It can be the ambient temperature in the car, etc. Step 203: Get the currently observed user input or operation and mark it as ;in, is a discrete variable; Step 204: Calculate the probability of the user's intention occurring under different system states using the Bayesian formula; wherein the Bayesian formula is specifically: , where Represents the user input or action at a given observation The vehicle system status is Under the condition of The posterior probability of Indicates that given user intent Under the condition of The vehicle system status is The joint probability of Indicates user intent Prior probability of occurrence; Represents observed user input or action The vehicle system status is The joint probability of Step 205: The steps of formulating the interaction response mechanism in the current interaction scenario are as follows: Step B1: For each user intent Assign a prior probability ; Step B2: For each user intent and every possible observable user input or action And the vehicle system status , define the conditional probability and ; Step B3: For a given observed user input or operation and vehicle system status , using the Bayesian formula to calculate each user's intention The posterior probability ; Step B4: Select the user intent with the highest posterior probability As the intention of the current interaction scenario; Step B5: Based on the selected user intent , get the corresponding user intent estimate: , where represents an estimate of the user's intent, Indicates Maximum user intent The value of Step B6: predefine a response set; In step B6, the set includes all possible interactive response mechanisms, each of which corresponds to a specific behavior or operation, such as displaying a certain information, performing a certain action, changing the interface layout, etc.; Step B7: according to the estimated value of user intention obtained in step B5, searching for the response mechanism that best matches it in the predefined response set; when the response mechanism that best matches the estimated value of user intention is found, it is used as the interaction response mechanism in the current interaction scenario; Step B8: Execute the selected interactive response mechanism to respond to the user's input or operation.

[0021] See also Figure 1 The present invention provides a vehicle-machine interface interaction method based on manual interaction. In step 3, according to the result of the interaction scenario analysis, the weight of each functional module is determined by using the hierarchical analysis method for several functional modules included in the vehicle-machine interface, and then the step of optimizing the layout of the vehicle-machine interface includes: Step 301: Assume that the vehicle interface exists Functional modules, build judgment matrix ;in, Indicates the vehicle interface function module For functional modules The importance of All are function module indexes. Indicates the number of functional modules; for example, if , it means that function module 1 is three times more important than function module 2 in the current interaction scenario; In step 301, the judgment matrix is ​​a matrix used in the hierarchical analysis method to represent the results of pairwise comparisons between factors at the same level with respect to the importance of a criterion at the previous level; Step 302: Use characteristic equation , calculate the judgment matrix The maximum eigenvalue of and its corresponding eigenvector , and solve it by iterative method. The specific steps are as follows: Step C1, rewrite the equation as ,Right now , where represents the identity matrix; by solving the characteristic polynomial , find the eigenvalue ,in, Represents eigenvalue quantity; Step C2: Select an initial vector ; Step C3: calculate , where Indicates the number of iterations; Step C4: Normalization ,Right now ; Step C5: Calculation ; Step C6, repeat steps C3-C5 until Converge to the preset value to obtain the maximum eigenvalue ; In step 302, the selection of the initial vector will affect the convergence speed of the iteration process and the eigenvector finally obtained. The unit vector is selected as the initial vector to ensure that the modulus of the vector remains within a reasonable range during the iteration process. The more iterations there are, the more likely the algorithm is to converge to the precise eigenvalue and eigenvector. In practical applications, whether to stop the iteration can be determined by setting an upper limit on the number of iterations or according to the change in the eigenvalue. During the iteration process, the eigenvector will gradually converge to the eigenvector corresponding to the maximum eigenvalue. In steps 301-302, in this context, features refer to features related to the vehicle-machine interface function modules. These features reflect the relative importance of the function modules. For example, for the navigation function and multimedia function in the vehicle-machine interface, their relative importance in different interaction scenarios (such as the urgent need to navigate to the destination while driving or the preference for multimedia functions when parking for leisure) is a feature; Quantified the vehicle interface functional modules Relative to functional modules The importance of the function, which may be based on factors such as the frequency and urgency of user use of the function in a specific scenario; Step 303: When normalized, the eigenvector The normalized vector of ,in, is the eigenvector The Euclidean norm of is, where represents the number of eigenvectors; Step 304: determine the relative importance of each functional module in the interface layout according to the normalized weight vector, and adjust the functional modules in the interface layout; In step 304, the functional modules in the interface layout are adjusted, including but not limited to the display size, position, etc. of the functional modules; for example, if the weight of a functional module is high, its display area can be increased or its position can be changed. Specifically, the functional module with a larger weight can be designed to have a larger display size or a more conspicuous position to attract the user's attention, and the functional module with a smaller weight can be designed to have a smaller display size or a less conspicuous position.

[0022] See also Figure 1The present invention provides a vehicle-machine interface interaction method based on manual interaction. In the step 4, according to the user's reaction after the vehicle-machine provides feedback, the user behavior feedback is modeled by constructing a Markov chain, thereby adjusting the feedback strategy of the vehicle-machine, including: Step 401: Assume the feedback state space is , A state set representing possible user behavior feedback (e.g. Express satisfaction, Express dissatisfaction, Indicates that further action is required, etc.), where: Indicates the number of possible behavior feedback states of the user; Step 402: Establish a transition probability matrix , and complete the initialization: the transition probability matrix is a The matrix represents the feedback state from user behavior Transfer to state The probability of transition probability matrix Elements Defined as: , that is, the user's current behavior feedback status Next, transfer to the state The probability of, where are all indexes of user behavior feedback status. Represents the probability value; in the initial stage, the transition probability matrix Initialize, for example, assuming that initially the probability of the user transferring from the "satisfied" feedback state to the "satisfied" state is high, and the probability of transferring from the "unsatisfied" feedback state to the "need further action" state is also high; Step 403: Based on the interaction between the user and the vehicle system, continuously collect the user's behavior feedback data, and update the transition probability matrix according to the collected user behavior feedback data. ; Step 404: Obtain the current transition probability matrix and the user's current behavior feedback status After that, predict the user's next possible behavior feedback state, that is, ; Step 405: Adjust the feedback strategy of the vehicle computer according to the predicted possible next behavior feedback of the user; for example, if the probability of the user being in a "dissatisfied" state is high, the feedback method can be changed, such as adjusting the tone of the voice prompt or re-optimizing the interface display; For example, suppose the feedback state space ,in, Express satisfaction, Indicates dissatisfaction; the initial transition probability matrix Defined as: , indicating that the probability of the user transferring from the "satisfied" state to the "satisfied" state is 0.8, and the probability of transferring to the "unsatisfied" state is 0.2; the probability of transferring from the "unsatisfied" state to the "satisfied" state is 0.3, and the probability of transferring to the "unsatisfied" state is 0.7; when the user's current behavior feedback state is (Unsatisfied), the user's next possible behavior feedback status is predicted as: , ;Based on this prediction result, the vehicle system can take corresponding measures to adjust the feedback strategy to improve user satisfaction; In steps 401-405, a Markov chain model is constructed by defining the feedback state space and the transition probability matrix; the model is enabled to reflect changes in actual data by initializing and updating the transition probability matrix; and the user experience is optimized by predicting the user's next behavior and adjusting the vehicle feedback strategy.

[0023] In the embodiment of the present invention, by collecting multi-source interaction information such as user voice input and touch operation, it is possible to more comprehensively understand the user's intention and improve the accuracy and flexibility of the interaction. By using the vector space model to quantify the interaction information, the degree of correlation between the user input and the vehicle computer function can be accurately evaluated to achieve accurate matching of the user's intention. By calculating the cosine similarity between the text vector and the touch vector, the user's intention can be accurately identified, the possibility of misoperation can be reduced, and the user experience can be improved. By combining the user intention recognition result with the vehicle computer system status information, the interaction scenario can be more accurately analyzed and a suitable response mechanism can be formulated. By using the Bayesian network to analyze the interaction scenario, the probability of user intention under different system states can be calculated to improve the accuracy of the response mechanism. The interaction response mechanism can be dynamically adjusted according to the analysis results to adapt to users in different scenarios. The invention can meet the needs of users and improve the convenience and efficiency of interaction. By using the hierarchical analysis method to determine the weights of each functional module, the importance of different functional modules in different interaction scenarios can be reflected, and a basis can be provided for optimizing the interface layout. By adjusting the interface layout according to the weights of the functional modules, the functional modules with higher weights can be highlighted, and the convenience and efficiency of user operation can be improved. By optimizing the interface layout, the interface can be made more intuitive and easy to use, thereby improving the user experience and satisfaction. By constructing a Markov chain to model the user behavior feedback, the user's next possible behavior feedback state can be predicted, and a basis can be provided for adjusting the vehicle machine feedback strategy. The vehicle machine feedback strategy can be adjusted according to the prediction results, and the user needs and feedback can be actively adapted to improve the intelligence and personalization level of interaction. By continuously collecting user behavior feedback data and updating the transition probability matrix, the interactive experience can be continuously optimized and the user's use effect can be improved. In summary, the examples of the present invention involve the decision-making of data processing, comprehensive analysis and intelligent adjustment, and solve the technical problem of poor user interaction experience in the vehicle information system. In actual situations, more data and context information may be needed to make specific decisions and optimization plans.

[0024] In addition, the formulas involved in the above are all calculated by removing dimensions and taking their numerical values. They are a formula that is closest to the actual situation obtained by collecting a large amount of data and performing software simulation. The proportional coefficient in the formula and the various preset thresholds in the analysis process are set by technical personnel in this field according to actual conditions or obtained by simulating a large amount of data; the size of the proportional coefficient is to quantify each parameter to obtain a specific value for subsequent comparison. The size of the proportional coefficient depends on the amount of sample data and the preliminary setting of the corresponding processing coefficient for each group of sample data by technical personnel in this field; as long as it does not affect the proportional relationship between the parameter and the quantized value.

[0025] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically based on the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0026] For the convenience of description, the above device is described in various units according to their functions. Of course, when implementing the present application, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0027] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0028] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0029] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0030] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0031] Secondly: In the drawings of the embodiments disclosed in the present invention, only the structures related to the embodiments disclosed in the present invention are involved, and other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of the present invention can be combined with each other; Finally: The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A vehicle-machine interface interaction method based on manual interaction, characterized in that: Step 1: The vehicle interface collects multi-source interaction information and uses a vector space model to quantify it in order to deeply identify the user's intention; the multi-source interaction information includes user voice input or touch operation; Step 2: Combine the recognition result of user intention with the vehicle system status information, use Bayesian network to analyze the interaction scenario, and formulate the interaction response mechanism under the current interaction scenario; Step 3: Based on the results of the interaction scenario analysis, the weight of each functional module is determined using the hierarchical analysis method for the functional modules included in the vehicle-machine interface, thereby optimizing the layout of the vehicle-machine interface; Step 4: Based on the user's reaction after the vehicle computer provides feedback, the user behavior feedback is modeled by constructing a Markov chain to adjust the feedback strategy of the vehicle computer.

2. The vehicle-machine interface interaction method based on manual interaction according to claim 1, characterized in that: In step 1, the vehicle interface collects multi-source interaction information and uses a vector space model to quantify the information so as to deeply identify the user's intention. The process includes: When the user interacts with the vehicle interface, multi-source interaction information is collected: Define the user voice input signal as ;in, represents the time variable, and is a multidimensional vector containing frequency, amplitude, and phase information; Mark user touch actions as ;in, Indicates the touch position, Indicates the touch strength. Indicates the time when the touch occurs; The vector space model is used to quantify the collected multi-source interaction information, including: For user voice input, it is converted into text form through voice recognition to obtain a text vector , where Represents the preset text feature quantization value, Indicates the number of pre-set features; For user touch operations, the pre-set functions and corresponding operation instructions of the vehicle interface are converted into touch vectors , where Indicates the quantized value of the corresponding feature in the pre-set vehicle function command; By calculating the cosine similarity between the text vector and the touch vector, the user intention is matched to the corresponding vehicle function to identify the user intention; the cosine similarity calculation formula is: ， In the formula, represents the cosine similarity between the text vector and the touch vector; is the dot product of vectors, which represents the sum of the products of two vectors in each dimension; and They are text vectors and touch vector The modulus of , which is the length of the vector; Represents a feature index.

3. The vehicle-machine interface interaction method based on manual interaction according to claim 1, characterized in that: In step 2, the process of combining the recognition result of the user intention with the vehicle system status information and using the Bayesian network to analyze the interaction scenario includes: After determining the user's intention, the vehicle system status information is collected through sensors; Set the user intent to , It is the different status information of the vehicle system; among them, Indicates the number of different status information of the vehicle system; Get the currently observed user input or operation and mark it as ; The Bayesian formula is used to calculate the probability of user intention under different system states; the Bayesian formula is specifically: , where Represents the user input or action at a given observation The vehicle system status is Under the condition of The posterior probability of Indicates that given user intent Under the condition of The vehicle system status is The joint probability of Indicates user intent Prior probability of occurrence; Represents observed user input or action The vehicle system status is The joint probability of .

4. The vehicle-machine interface interaction method based on manual interaction according to claim 3, characterized in that: In step 2, the process of formulating the interactive response mechanism in the current interactive scenario includes: Step B1: For each user intent Assign a prior probability ; Step B2: For each user intent and every possible observable user input or action And the vehicle system status , define the conditional probability and ; Step B3: For a given observed user input or operation and vehicle system status , using the Bayesian formula to calculate each user's intention The posterior probability ; Step B4: Select the user intent with the highest posterior probability As the intention of the current interaction scenario; Step B5: Based on the selected user intent , get the corresponding user intent estimate: , where represents an estimate of the user's intent, Indicates Maximum user intent The value of Step B6: predefine a response set; Step B7: according to the estimated value of user intention obtained in step B5, searching for the response mechanism that best matches it in the predefined response set; when the response mechanism that best matches the estimated value of user intention is found, it is used as the interaction response mechanism in the current interaction scenario; Step B8: Execute the selected interactive response mechanism to respond to the user's input or operation.

5. The vehicle-machine interface interaction method based on manual interaction according to claim 1, characterized in that: In step 3, based on the results of the interaction scenario analysis, the weight of each functional module is determined by using the hierarchical analysis method for several functional modules included in the vehicle computer interface, and the process of optimizing the layout of the vehicle computer interface includes: Assume that the vehicle interface exists Functional modules, build judgment matrix ;in, Indicates the vehicle interface function module For functional modules The importance of All are function module indexes. Indicates the number of functional modules; Using the characteristic equation , calculate the judgment matrix The maximum eigenvalue of and its corresponding eigenvector , and solved by iterative method; In the feature vector When normalized, the eigenvector The normalized vector of ,in, is the eigenvector The Euclidean norm of is, where represents the number of eigenvectors; According to the normalized weight vector, the relative importance of each functional module in the interface layout is determined, and the functional modules in the interface layout are adjusted.

6. The vehicle-machine interface interaction method based on manual interaction according to claim 5, characterized in that: The process of using the characteristic equation to calculate the maximum eigenvalue of the judgment matrix and its corresponding eigenvector and solving it by an iterative method includes: Step C1, rewrite the equation as ,Right now , where represents the identity matrix; by solving the characteristic polynomial , find the eigenvalue ,in, Represents eigenvalue quantity; Step C2: Select an initial vector ; Step C3: calculate , where Indicates the number of iterations; Step C4: Normalization ,Right now ; Step C5: Calculation ; Step C6, repeat steps C3-C5 until Converge to the preset value to obtain the maximum eigenvalue .

7. The vehicle-machine interface interaction method based on manual interaction according to claim 1, characterized in that: In step 4, according to the user's reaction after the vehicle computer provides feedback, the process of modeling the user behavior feedback by constructing a Markov chain to adjust the feedback strategy of the vehicle computer includes: Assume the feedback state space is , A state set representing possible user behavior feedback, where: Indicates the number of possible behavior feedback states of the user; Building the transition probability matrix , and complete the initialization, specifically: the transfer probability matrix is a The matrix represents the feedback state from user behavior Transfer to state The probability of transition probability matrix Elements Defined as: , that is, the user's current behavior feedback status Next, transfer to the state The probability of, where are all indexes of user behavior feedback status. Represents the probability value; in the initial stage, the transition probability matrix Initialize; Based on the interaction between the user and the vehicle system, the user's behavior feedback data is continuously collected, and the transition probability matrix is ​​updated according to the collected user behavior feedback data ; Getting the current transition probability matrix and the user's current behavior feedback status After that, predict the user's next possible behavior feedback state, that is, ; Adjust the vehicle computer’s feedback strategy based on the predicted user’s next possible behavior feedback.

Images