A method and system for constructing an evaluation model for an automobile human-computer interaction system

Abstract

The present invention provides a method and system for constructing an evaluation model of an automobile human-computer interaction system, and relates to the technical field of evaluation models. The present invention collects user's line of sight movement data when operating an interface through eye tracking technology, objectively reflects the user's attention distribution and interest points, and provides quantitative behavior data support for the evaluation model; secondly, combined with direct user feedback collected through a questionnaire survey, the evaluation dimension is further enriched from the perspective of user satisfaction and subjective evaluation, and the interest of the subject in each browsing interval in a free browsing interface sample is evaluated through a first evaluation index, and the interest of the subject in each human-computer interaction function in a user task test sample is evaluated through a second evaluation index. Finally, combined with the human-computer interaction evaluation model, an evaluation adjustment strategy can be generated for the current test data set, which fully considers the user's behavior data in the actual operation process and can truly reflect the user's interest and preference for different interfaces or functions.

Description

Technical Field

[0001] The present invention relates to the technical field of evaluation models, and in particular to a method and system for constructing an evaluation model for an automobile human-computer interaction system. Background Art

[0002] In the current automotive industry, human-machine interaction (HMI) systems have become one of the key indicators for evaluating the level of automotive intelligence. These systems are not only a display platform for vehicle information, but also a bridge for communication between the driver and the vehicle, covering a variety of interaction methods from touch screen operation, voice recognition to advanced driver assistance systems (ADAS); With the rapid development of automotive technology, especially the advancement of autonomous driving technology, the requirements for human-machine interaction systems are also constantly increasing, including but not limited to improving the system's response speed, accuracy, ease of use and meeting personalized needs;

[0003] However, although the existing human-computer interaction systems have made certain progress in technology, they still face many challenges in evaluating the effectiveness and user satisfaction of these systems. First, the existing evaluation methods often focus on qualitative analysis and lack quantitative and objective evaluation indicators, which makes the evaluation results greatly affected by subjective factors and difficult to fully and accurately reflect the actual performance of the system and user experience. Secondly, most evaluation models fail to fully consider the user's behavioral data during actual operation, such as eye tracking data, resulting in the evaluation results not being able to truly reflect the user's interest and preference for different interfaces or functions. In addition, most of the existing evaluation systems fail to effectively combine users' direct feedback with behavioral data, thus lacking insight into the deep-level factors of user experience, limiting the guiding value of the evaluation model in practical applications.

[0004] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0005] The purpose of the present invention is to provide a method and system for constructing an evaluation model for an automobile human-computer interaction system, so as to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solution: a method for constructing an evaluation model of an automobile human-computer interaction system, the specific steps of which include:

[0007] Step S1, obtaining and screening samples of the automobile human-computer interaction system to form a test data set, wherein the test data set includes free browsing interface samples and user task test samples, and determining corresponding test subjects and questionnaire scoring tables according to the test data sets, wherein the questionnaire scoring table is a visual preference scoring table for the test data sets;

[0008] Step S2, collecting the first eye movement data of the subject when watching the free browsing interface sample, and the second eye movement data of the subject when watching the user task test sample, and after the experiment, obtaining the questionnaire score sheet filled in by the subject about the test data set;

[0009] Step S3, obtaining a free browsing interface sample and evenly dividing it into a plurality of browsing intervals, obtaining first eye movement data for each browsing interval for analysis and processing to generate a first evaluation index, wherein the first evaluation index is used to evaluate the subject's interest in each browsing interval in the free browsing interface sample;

[0010] Step S4, obtaining user task test samples, and classifying and marking them according to human-computer interaction functions, and obtaining corresponding second eye movement data for each human-computer interaction function for analysis and processing to generate a second evaluation index, wherein the second evaluation index is used to evaluate the subject's interest in each human-computer interaction function in the user task test samples;

[0011] Step S5: Obtain the first evaluation index, the second evaluation index, and the questionnaire score sheet, perform analysis and processing, and establish a human-computer interaction evaluation model, and the human-computer interaction evaluation model can generate an evaluation adjustment strategy for the current test data set.

[0012] Furthermore, the free browsing interface sample includes M main static interfaces in the current automobile human-computer interaction system as picture materials;

[0013] The first eye movement data includes the AOI first fixation time and the AOI fixation rate, and the AOI first fixation time and the AOI fixation rate are calibrated as SZt and ZSl respectively;

[0014] The AOI first fixation time SZt is the time when the subject first focuses his / her sight on the area of ​​interest AOI;

[0015] The fixation rate ZSl is the frequency of the subject's fixation on the AOI, expressed as a percentage, and is calculated by the ratio of the total duration of the subject's fixation on the AOI to the total duration of the experiment;

[0016] The user task test sample includes the test subject operating various functions of the product according to the specified script;

[0017] The second eye movement data includes the second fixation time, the return time and the number of return times for each human-computer interaction function, and the second fixation time, the return time and the number of return times are marked as DRt, HSt and HSm respectively;

[0018] The second fixation time DRt is the time it takes for the subject to return to and fixate on a human-computer interaction function after first fixating on the function;

[0019] The return time HSt is the time interval for the subject to return to and gaze at a function after completing a human-computer interaction function.

[0020] The number of look-backs HSm refers to the number of times the subject returns to and looks at a function after completing a human-computer interaction function.

[0021] Furthermore, the method of obtaining the free browsing interface sample and evenly dividing it into a number of browsing intervals specifically includes the following contents:

[0022] Obtain M main static interfaces in the current automobile human-computer interaction system, and evenly divide each main static interface into grids, defining several browsing intervals as {1, 2, ..., i, ..., n};

[0023] The first evaluation index is defined as E1, and the calculation formula is as follows:

[0024]

[0025] Where n is the total number of browsing intervals, SZt′ i With ZSl′ i represents the AOI first fixation time and AOI fixation rate of the i-th browsing interval, and SZt′ i With ZSl′ i All are normalized and quantized to ensure that their values ​​are between 0 and 1, and w1 and w2 are weight coefficients;

[0026] The value range of the first evaluation index E1 is limited to [0,1]. When 0≤E1<0.33, it indicates low interest; when 0.33≤E1<0.66, it indicates medium interest; and when 0.66≤E1≤1, it indicates high interest.

[0027] Furthermore, the user task test samples are obtained and classified and labeled according to the human-computer interaction functions, specifically including the following contents:

[0028] The classification of human-computer interaction functions is defined as {1, 2, ..., j, ..., h};

[0029] The second evaluation index is defined as E2, and the calculation formula is as follows:

[0030]

[0031] Among them, DRt′ j , HSt′ j HSm′ j All said that after Quantized value after normalization, DRt′ j is the second fixation time of the jth human-computer interaction function, HSt′ j is the look-back time of the jth human-computer interaction function, HSm′ j is the number of glances back for the jth human-computer interaction function, α, β, γ are weight coefficients, which are used to adjust the contribution of the three eye movement data in the overall evaluation index to meet different evaluation requirements, and α+β+γ=1; h is the total number of human-computer interaction functions; Z is the normalization coefficient, which ensures that the value range of E2 is within [0,1], log(DRt′ j +1) is used to scale the effect of the second fixation time, making it more sensitive over a larger range of variations;

[0032] is the Sigmoid function, which is used to map the lookback time to the range of (0,1). It is used to reduce the linear influence of the number of lookbacks on the results and limit the value range of E2 to [0,1];

[0033] When 0≤E2<0.25, it indicates low interest; when 0.25≤E2<0.75, it indicates medium interest; when 0.75≤E2≤1, it indicates high interest.

[0034] Furthermore, the obtaining of the first evaluation index, the second evaluation index, and the questionnaire score sheet, and performing analysis and processing to establish a human-computer interaction evaluation model specifically includes the following contents:

[0035] The calculation formula for defining the human-computer interaction evaluation model is:

[0036]

[0037] Among them, P is the total evaluation index; U is the questionnaire score vector, which includes the normalized scores of interface design aesthetics, interface usability, information presentation clarity, interactive feedback satisfaction, personalization degree, and function discovery ease; Θ is the parameter of the logistic regression model; λ, μ, ν are weight coefficients used to adjust E1, E2, The contribution in the total evaluation P, and λ+μ+ν=1; the value range of P is set in [0,1], where

[0038] 0≤P<0.18 indicates low satisfaction, which means that users are less satisfied with the human-computer interaction interface and need to be improved;

[0039] 0.18≤P<0.48 indicates moderate satisfaction, where users are somewhat satisfied with the interface, but there is room for improvement;

[0040] 0.48≤P≤1 indicates high satisfaction, and users are very satisfied with the human-computer interaction interface, which shows that the interface design and interaction logic meet user expectations.

[0041] A system, the system is used to execute the method for constructing an evaluation model of an automobile human-computer interaction system, comprising:

[0042] Data preparation module: used to obtain and screen samples of the automobile human-computer interaction system to form a test data set, which includes free browsing interface samples and user task test samples, and determine corresponding test subjects and questionnaire scoring tables according to the test data sets, and the questionnaire scoring table is a visual preference score for the test data set;

[0043] Data collection module: used to collect the first eye movement data of the subjects when they watch the free browsing interface samples, and the second eye movement data of the subjects when they watch the user task test samples. After the experiment, the questionnaire score sheet filled out by the subjects on the test data set is obtained;

[0044] A first index generating module is used to obtain a free browsing interface sample and evenly divide it into a plurality of browsing intervals. For each browsing interval, first eye movement data is obtained for analysis and processing to generate a first evaluation index. The first evaluation index is used to evaluate the interest of the subject in each browsing interval in the free browsing interface sample.

[0045] The second index generation module is used to obtain user task test samples, classify and label them according to human-computer interaction functions, and obtain corresponding second eye movement data for each human-computer interaction function for analysis and processing to generate a second evaluation index, wherein the second evaluation index is used to evaluate the interest of the subject in each human-computer interaction function in the user task test samples;

[0046] Adjustment strategy generation module: used to obtain the first evaluation index, the second evaluation index, and the questionnaire score sheet, perform analysis and processing, and establish a human-computer interaction evaluation model, and the human-computer interaction evaluation model can generate an evaluation adjustment strategy for the current test data set.

[0047] Compared with the prior art, the beneficial effects of the present invention are: by collecting the user's line of sight movement data when operating the interface through eye tracking technology, the user's attention distribution and interest points are objectively reflected, and quantitative behavioral data support is provided for the evaluation model; secondly, combined with the direct feedback from users collected through questionnaire surveys, the evaluation dimension is further enriched from the perspective of user satisfaction and subjective evaluation, and the first evaluation index is used to evaluate the subject's interest in each browsing interval in the free browsing interface sample, and the second evaluation index is used to evaluate the subject's interest in each human-computer interaction function in the user task test sample. Finally, combined with the human-computer interaction evaluation model, an evaluation adjustment strategy can be generated for the current test data set, which fully considers the user's behavioral data in the actual operation process and can truly reflect the user's interest and preference for different interfaces or functions. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a schematic diagram of the overall method flow of the present invention;

[0049] Figure 2 is a system module block diagram of the present invention;

[0050] Figure 3 It is the flow chart of the eye movement test of the present invention. DETAILED DESCRIPTION

[0051] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0052] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0053] Embodiment 1:

[0054] See also Figure 1 , Figure 2 and Figure 3 , the present invention provides a technical solution:

[0055] A method for constructing an evaluation model for an automobile human-computer interaction system comprises the following steps:

[0056] Step S1, obtaining and screening samples of the automobile human-computer interaction system to form a test data set, wherein the test data set includes free browsing interface samples and user task test samples, and determining corresponding test subjects and questionnaire scoring tables according to the test data sets, wherein the questionnaire scoring table is a visual preference scoring table for the test data sets;

[0057] It should be noted that the test data set is formed by the following steps:

[0058] A1. Research and collate representative samples of automotive human-machine interaction systems on the market, process the samples in grayscale mode, and create a brightness histogram;

[0059] A2. Combining existing literature and eye movement indicators in the automotive human-computer interaction system sample, five representative eye movement indicators were selected, and experiments were carried out using eye tracking methods to collect and organize the data of various eye movement indicators and visual preference scores;

[0060] Step S2, collecting the first eye movement data of the subject when watching the free browsing interface sample, and the second eye movement data of the subject when watching the user task test sample, and after the experiment, obtaining the questionnaire score sheet filled in by the subject about the test data set;

[0061] Step S3, obtaining a free browsing interface sample and evenly dividing it into a plurality of browsing intervals, obtaining first eye movement data for each browsing interval for analysis and processing to generate a first evaluation index, wherein the first evaluation index is used to evaluate the subject's interest in each browsing interval in the free browsing interface sample;

[0062] Step S4, obtaining user task test samples, and classifying and marking them according to human-computer interaction functions, and obtaining corresponding second eye movement data for each human-computer interaction function for analysis and processing to generate a second evaluation index, wherein the second evaluation index is used to evaluate the subject's interest in each human-computer interaction function in the user task test samples;

[0063] Step S5: Obtain the first evaluation index, the second evaluation index, and the questionnaire score sheet, perform analysis and processing, and establish a human-computer interaction evaluation model, and the human-computer interaction evaluation model can generate an evaluation adjustment strategy for the current test data set.

[0064] Embodiment 2:

[0065] Further illustrating on the basis of the first embodiment, the free browsing interface sample includes M main static interfaces in the current automobile human-computer interaction system as picture materials;

[0066] The first eye movement data includes the AOI first fixation time and the AOI fixation rate, and the AOI first fixation time and the AOI fixation rate are calibrated as SZt and ZSl respectively;

[0067] The AOI first fixation time SZt is the time when the subject first focuses his / her sight on the AOI, in milliseconds ms. The first fixation time can reflect the subject's attention to the AOI and the reaction speed. A longer first fixation time usually indicates that the subject has a higher degree of attention to the area.

[0068] The fixation rate ZSl is the frequency of the subject's fixation on the AOI, expressed as a percentage, and is calculated as the ratio of the total duration of the subject's fixation on the AOI to the total duration of the experiment; a higher fixation rate means that the subject fixates on the area more frequently, indicating that the area is important or attractive in visual attention;

[0069] The user task test sample includes the test subject operating various functions of the product according to the specified script;

[0070] The second eye movement data includes the second fixation time, the return time and the number of return times for each human-computer interaction function, and the second fixation time, the return time and the number of return times are marked as DRt, HSt and HSm respectively;

[0071] The second fixation time DRt is the time when the subject returns to fixate on a human-computer interaction function after first fixating on the function, which indicates the degree and duration of the subject's renewed attention to the function;

[0072] The look-back time HSt is the time interval for the subject to return to and look at a function after completing a human-computer interaction function. The look-back time reflects the speed and frequency of the subject's attention to the function that he / she has interacted with before.

[0073] The number of glances HSm is the number of times the subject returns to and looks at a function after completing a human-computer interaction function. The number of glances can be used to evaluate the subject's repeated attention to a specific function and the attention allocation pattern.

[0074] Embodiment three:

[0075] Further explanation based on Example 2, the questionnaire score sheet is a visual preference score for the test data set, specifically including:

[0076] The aesthetics of the interface design indicates the subjects’ aesthetic satisfaction rating of the color matching, icon design, and layout rationality of the car’s human-computer interaction interface, with a satisfaction rating of 1 and a dissatisfaction rating of 0;

[0077] Interface usability, which indicates the user's evaluation of the intuitiveness of the interface operation, ease of learning, and whether the task can be completed quickly. The satisfaction score is 1, and the dissatisfaction score is 0;

[0078] Information presentation clarity, which indicates the user's evaluation of the clarity and readability of information displayed on the car's human-computer interaction interface, with a satisfactory score of 1 and an unsatisfactory score of 0;

[0079] Interaction feedback satisfaction, which indicates the user's satisfaction with the button click response and sliding smoothness during the interface interaction process, with a satisfaction score of 1 and a dissatisfaction score of 0;

[0080] The degree of personalization indicates the user's evaluation of the interface's ability to be customized according to personal preferences, with a satisfaction score of 1 and a dissatisfaction score of 0;

[0081] Function discovery ease, which indicates how easy it is for users to discover and understand the functions provided by the interface. A satisfactory score is 1, and an unsatisfactory score is 0.

[0082] Embodiment 4:

[0083] Further explanation based on the third embodiment, the method of obtaining the free browsing interface sample and evenly dividing it into a number of browsing intervals specifically includes the following contents:

[0084] Obtain M main static interfaces in the current automobile human-computer interaction system, and evenly divide each main static interface into grids, and define several browsing intervals as {1, 2, ..., i, ..., n};

[0085] The first evaluation index is defined as E1, and the calculation formula is as follows:

[0086]

[0087] Where n is the total number of browsing intervals, SZt i ′ and ZSl i ′ respectively represent the AOI first fixation time and AOI fixation rate in the i-th browsing interval, and SZt i ′ and ZSl i ′ are all normalized and quantized to ensure that their values ​​are between 0 and 1, w1 and w2 are weight coefficients, which are determined based on experimental data or expert advice, and w1+w2=1;

[0088] Among them, the normalization formula is:

[0089]

[0090] x′ represents the normalized value. In this embodiment, x′ is SZt i ′ and ZSli ′, max(X) and min(X) are the maximum and minimum values ​​of the corresponding parameters respectively;

[0091] The value range of the first evaluation index E1 is limited to [0,1].

[0092] When 0≤E1<0.33, it indicates low interest; the evaluation index within this range indicates that the subjects have low interest in the free browsing interface samples;

[0093] When 0.33≤E1<0.66, it indicates a moderate interest level; the evaluation index within this range indicates that the subjects have a certain interest in the interface sample, but it is not particularly strong;

[0094] When 0.66≤E1≤1, it indicates high interest; the evaluation index within this range shows that the subjects are very interested in the interface sample and attract more attention.

[0095] Embodiment five:

[0096] Further explanation based on the fourth embodiment, the acquisition of user task test samples and classification and labeling according to human-computer interaction functions specifically includes the following contents:

[0097] The classification of human-computer interaction functions is defined as {1, 2, ..., j, ..., h};

[0098] The second evaluation index is defined as E2, and the calculation formula is as follows:

[0099]

[0100] Among them, DRt j ′, HSt j ′, HSm j ' means after Quantized value after normalization, DRt j ′ is the second fixation time of the jth human-computer interaction function, indicating the degree and duration of the subject's renewed attention to a certain human-computer interaction function;

[0101] HSt j ′ is the return time of the jth human-computer interaction function, which indicates the time interval for the subject to return to the function after completing a function, and is used to reflect the speed and frequency of renewed attention;

[0102] HkDJ j ′ is the number of times the jth human-computer interaction function is looked back, indicating the degree of repeated attention paid by the subjects to a certain function;

[0103] α, β, γ are weight coefficients, which are used to adjust the contribution of the three types of eye movement data in the overall evaluation index to meet different evaluation requirements, and α+β+γ=1; h is the total number of human-computer interaction functions; Z is the normalization coefficient, which ensures that the value range of E2 is within [0,1], so as to ensure that the output of the formula is within the specified range;

[0104] log(DRt j ′+1) is used to scale the effect of the second fixation time, making it more sensitive in a larger range of changes;

[0105] is a Sigmoid function, which is used to map the lookback time to the range of (0,1) to keep its contribution to the result in a reasonable range;

[0106] It is used to reduce the linear effect of the number of review times on the results, so as to avoid excessive impact on the overall evaluation due to too many reviews;

[0107] Limit the value range of E2 to [0,1],

[0108] When 0≤E2<0.25, it indicates low interest; this indicates that the subjects have low interest in human-computer interaction functions. Within this range, it can be considered that users do not pay much attention to these functions, indicating that these functions are not attractive enough or do not meet the needs of users;

[0109] When 0.25≤E2<0.75, it indicates a medium interest level; it means that the subjects have a medium interest in the human-computer interaction function. The functions within this range are attractive to users, but there is still room for improvement.

[0110] When 0.75≤E2≤1, it indicates high interest; it means that the subjects are very interested in the human-computer interaction functions, which shows that these functions are very consistent with the interests and needs of users and are the highlights of human-computer interaction design.

[0111] Embodiment six:

[0112] Further explanation based on the fifth embodiment, the obtaining of the first evaluation index, the second evaluation index, and the questionnaire score sheet, performing analysis and processing, and establishing a human-computer interaction evaluation model specifically includes the following contents:

[0113] The calculation formula for defining the human-computer interaction evaluation model is:

[0114]

[0115] Among them, P is the total evaluation index; U is the questionnaire score vector, which includes the normalized scores of interface design aesthetics, interface usability, information presentation clarity, interactive feedback satisfaction, personalization degree, and function discovery ease; Θ is the parameter of the logistic regression model; λ, μ, ν are weight coefficients used to adjust E1, E2, Contribution to the total evaluation P, and λ+μ+ν=1;

[0116] A logistic regression model is used on the questionnaire survey score vector U to analyze the questionnaire survey score table, Θ is a parameter vector of the logistic regression model, which is learned through training data and is used to determine the degree of influence of each feature on the final classification result;

[0117] Θ T U is the dot product of Θ and U, which is the input of the logistic regression model; T Substitute the result of U into the Sigmoid function to obtain the satisfaction probability Q of the questionnaire survey score:

[0118]

[0119] Set the value range of P to [0,1], where

[0120] 0≤P<0.18 indicates low satisfaction, which means that users are less satisfied with the human-computer interaction interface and need to be improved;

[0121] 0.18≤P<0.48 indicates moderate satisfaction, where users are somewhat satisfied with the interface, but there is room for improvement;

[0122] 0.48≤P≤1 indicates high satisfaction, and users are very satisfied with the human-computer interaction interface, which shows that the interface design and interaction logic meet user expectations.

[0123] Embodiment seven:

[0124] Further explanation based on the sixth embodiment, the human-computer interaction evaluation model can generate an evaluation adjustment strategy for the current test data set, which specifically includes the following contents:

[0125] For weights λ,μ, set the aesthetics of interface design, ease of use of interface, clarity of information presentation, satisfaction with interactive feedback, degree of personalization, and ease of function discovery as the dimensions in the questionnaire survey score sheet. When the score of any dimension is the lowest, increase the weight corresponding to the dimension by 10%. The aesthetics of interface design and ease of use of interface correspond to weight E1, and the clarity of information presentation, satisfaction with interactive feedback, degree of personalization, and ease of function discovery correspond to weight E2 μ;

[0126] For the normalized AOI first fixation time SZt i′ and AOI gaze rate ZSl i ', set the threshold to the average value Y1,

[0127]

[0128] If SZt′ i or ZSl′ i If it is lower than the mean value Y1 and P < 0.48, it is considered that the meshing area of ​​the static interface needs to be improved to improve the attractiveness;

[0129] Furthermore, the threshold of the AOI first fixation time SZt was set to 250 ms, and a value lower than this indicated insufficient initial attraction;

[0130] The adjustment strategy of AOI first fixation time SZt is: SZt″=SZt+Δ SZt , where SZt is the first fixation time of the original AOI, Δ SZt is the change in the first fixation time of AOI, SZt″ is the optimized adjustment value;

[0131] If SZt is less than 250 milliseconds, Δ SZt =250-SZt+c1, it means that the interface is not attractive enough, and it is necessary to improve the user's initial attention by adjusting the current 30% layout or highlighting 20% ​​of the interface display, increasing 10% animation, and improving 10% color vividness to enhance the visibility and attractiveness of the interface, where c1 is the adjustment coefficient, Δ SZt For every increase of 10, c1 takes 5 as the base value and adds 5, for example, Δ SZt is 19, c1 is 15; Δ SZt Each increase is less than a multiple of 10, calculated as the current value, for example, when Δ SZt is 24, adjust c1 according to 30;

[0132] The AOI gaze rate ZSl threshold is set to 30%. A value lower than this indicates that the interface is not attractive enough;

[0133] The AOI gaze rate ZSl adjustment strategy is: ZSl″=ZSl+Δ ZSl , where ZSl is the original fixation rate, Δ ZSl is the change in fixation rate;

[0134] If ZSl is lower than 30%, it means the interface is not attractive enough. It is necessary to enhance the visibility and attractiveness of the interface by adjusting the layout by 35% or highlighting the content by 28% to achieve ZSl+Δ ZSl ≥30%;

[0135] For the normalized second fixation time DRt′ j , Look-back time HSt′j and the number of look-backs HSm′ j , set the threshold to the average value Y2,

[0136]

[0137] If DRt′ j or HSt′ j or HSm′ j Higher than the average Y2, and 0.48≤P≤1, indicating that users have a sustained interest in this function or are confused about its use, which requires further investigation;

[0138] Specifically, the second fixation time DRt threshold is set to 500 milliseconds. A value higher than this value indicates that the user is interested in the function or has questions about its use;

[0139] The second fixation time DRt adjustment strategy is: DRt″=DRt-Δ DRt , where DRt is the original second fixation time, DRt″ is the optimized adjustment value, Δ DRt is the amount of change;

[0140] If DRt>500ms, user testing is required for this function to confirm whether it is caused by high interest or unclear operation, and optimization is carried out according to the feedback. If it is caused by high interest, focus on analyzing and replicating the success factors; if it is caused by unclear operation, simplify the current interactive function by 20% and highlight the visibility of each operation script by 30% to achieve Δ DRt +DRt>500;

[0141] If DRt≤500ms, it means that users spend less time finding the functions they are interested in, or the operation is unclear. If the operation is unclear, simplify the current interactive function by 15% and highlight the visibility of each operation script by 20%;

[0142] The HSt threshold of the re-look time is set to 2000 milliseconds. A value shorter than this indicates fast re-look, which indicates good usability.

[0143] If HSt ≥ 2000 milliseconds, it is necessary to simplify the operation process by 20%, highlight the position of the function interface by 20%, increase the prompt information by 5%, and shorten the user's search time to improve usability;

[0144] The HSm threshold of the number of re-views is set to 3 times. Exceeding this value indicates that the user is highly interested in the function or is confused. If there is confusion, the difference in the operation process represented by the interactive operation script is increased by 25% and the difference in the position of the function interface is increased by 30%. The difference refers to the color between the operation process steps, the deformation difference of the function icon shape, and the deformation difference between adjacent function interfaces.

[0145] Embodiment eight:

[0146] Further explanation based on Example 7, this example aims to verify the effectiveness of the human-computer interaction evaluation model, especially the role of the evaluation adjustment strategy in improving user experience. To this end, a series of experiments are designed to analyze the impact of the improvement strategy by comparing the changes before and after the test data set;

[0147] The experiment is divided into two stages: initial evaluation and post-optimization evaluation. In the initial evaluation stage, 40 users were invited to participate, with a male-female ratio of 1:1. The sample age was not limited, and the naked eye vision or corrected vision was normal. The experimental subjects had driving experience for more than one year and were not related personnel of automotive interface design majors. The selected models had not been used by the subjects.

[0148] The experimental task includes two stages: free browsing interface and user task test. After the eye movement test, a questionnaire survey is used to collect users' scores on six dimensions: interface design aesthetics, interface usability, information presentation clarity, interactive feedback satisfaction, personalization, and ease of function discovery. At the same time, software tools are used to collect users' key interaction indicators such as first fixation time, fixation rate, second fixation time, return time, and return number.

[0149] Free browsing interface stage: Determine M as 6, and then select 6 main static interfaces as image materials. Present the images to the subjects in sequence and let them browse freely for 8 seconds. The time limit is based on the recommendations of the eye tracker engineer.

[0150] Task testing phase: Design tasks according to test requirements and let the subjects operate some functions of the product according to the script, and use the eye tracker to record the user's usage trajectory. The eye tracker is aSeeProF25 model produced by Beijing Qixin Yiwei Information Technology Co., Ltd., and the eye tracker can be replaced with an adaptive model according to the needs of the actual experiment;

[0151] Based on the collected data, the evaluation adjustment strategy described in the invention is applied to optimize the interface, and the optimization measures include but are not limited to adding animations, adjusting the layout, and highlighting specific elements, aiming to increase the weight of low-scoring dimensions, enhance the attractiveness of the interface, simplify the operation process, and improve the clarity of information;

[0152] In the post-optimization evaluation phase, users were invited to use the updated interface again and the same rating and interaction metric data were collected;

[0153] The following table contains 10 sets of example data, each set of data includes a comparison before and after optimization to reflect the innovation and advantages of the invention content;

[0154]

[0155] Table 1

[0156] Data Analysis:

[0157] By deeply analyzing the above data, we can clearly see the positive impact of optimization measures on improving user experience. The following are the specific analysis results:

[0158] Significant reduction in first fixation time: After optimization, the first fixation time of all interfaces was reduced, with an average reduction of 60 milliseconds. Interface A was reduced from 300 milliseconds to 230 milliseconds, an improvement of 23.3%. Interfaces B, C, D, and E improved by 7.1%, 22.6%, 7.4%, and 24.1%, respectively, which means that the optimized interface can attract users' attention faster, reflecting that the visual appeal of the interface has been effectively improved by adding animations or brighter colors.

[0159] Adjustment strategy: For interfaces with a large reduction in first fixation time, such as A and E, we should focus on analyzing and replicating the success factors, such as changes in the aesthetics of the interface design;

[0160] Improvement of fixation rate: The optimization measures increased fixation rate by about 7% on average, with Interface E seeing the most significant increase, from 28% to 37%. This shows that by adjusting the layout and highlighting important elements, the visibility and attractiveness of the interface or function are enhanced, thus making it more able to attract and maintain user attention.

[0161] Adjustment strategy: For interfaces with significantly improved gaze rates, it is recommended to analyze the visual appeal of elements and the clarity of information, and adjust the design of other interfaces based on this to improve the visibility and appeal of key information or functions;

[0162] Moderate reduction in second fixation time: The data showed that second fixation time was reduced in most cases, with an average reduction of about 20 milliseconds, indicating that users spent less time finding the features they were interested in, indicating that the features were presented more intuitively or that users were more interested in the interface;

[0163] Adjustment strategy: You can further reduce the user's search time by optimizing the spatial layout between elements and improving the recognition of elements;

[0164] Reduction in re-look time and number of re-looks: The average re-look time for all interfaces decreased by 150 milliseconds, and the average number of re-looks decreased by 2. These data points show that the optimized interface reduces the confusion encountered by users during use, improves usability, and maintains users' interest in the function better;

[0165] Adjustment strategy: Continue to optimize the interaction logic and visual guidance of elements, especially for interfaces with significantly reduced return times. Their successful interaction patterns and visual cues should be analyzed for replication in other interfaces.

[0166] A system, the system is used to execute the method for constructing an evaluation model of an automobile human-computer interaction system, comprising:

[0167] Data preparation module: used to obtain and screen samples of the automobile human-computer interaction system to form a test data set, which includes free browsing interface samples and user task test samples, and determine corresponding test subjects and questionnaire scoring tables according to the test data sets, and the questionnaire scoring table is a visual preference score for the test data set;

[0168] Data collection module: used to collect the first eye movement data of the subjects when they watch the free browsing interface samples, and the second eye movement data of the subjects when they watch the user task test samples. After the experiment, the questionnaire score sheet filled out by the subjects on the test data set is obtained;

[0169] A first index generating module is used to obtain a free browsing interface sample and evenly divide it into a plurality of browsing intervals. For each browsing interval, first eye movement data is obtained for analysis and processing to generate a first evaluation index. The first evaluation index is used to evaluate the interest of the subject in each browsing interval in the free browsing interface sample.

[0170] The second index generation module is used to obtain user task test samples, classify and label them according to human-computer interaction functions, and obtain corresponding second eye movement data for each human-computer interaction function for analysis and processing to generate a second evaluation index, wherein the second evaluation index is used to evaluate the interest of the subject in each human-computer interaction function in the user task test samples;

[0171] Adjustment strategy generation module: used to obtain the first evaluation index, the second evaluation index, and the questionnaire score sheet, perform analysis and processing, and establish a human-computer interaction evaluation model, and the human-computer interaction evaluation model can generate an evaluation adjustment strategy for the current test data set.

[0172] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0173] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product. Those skilled in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed by hardware or software methods depends on the specific application and design constraints of the technical solution.

[0174] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0175] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A method for constructing an evaluation model for an automobile human-computer interaction system, characterized in that: The specific steps include: Step S1, obtaining and screening samples of the automobile human-computer interaction system to form a test data set, wherein the test data set includes free browsing interface samples and user task test samples, and determining corresponding test subjects and questionnaire scoring tables according to the test data sets, wherein the questionnaire scoring table is a visual preference scoring table for the test data sets; Step S2, collecting the first eye movement data of the subject when watching the free browsing interface sample, and the second eye movement data of the subject when watching the user task test sample, and after the experiment, obtaining the questionnaire score sheet filled in by the subject about the test data set; Step S3, obtaining a free browsing interface sample and evenly dividing it into a plurality of browsing intervals, obtaining first eye movement data for each browsing interval for analysis and processing to generate a first evaluation index, wherein the first evaluation index is used to evaluate the subject's interest in each browsing interval in the free browsing interface sample; Obtain M main static interfaces in the current automobile human-computer interaction system, and evenly divide each main static interface into grids, defining several browsing intervals as {1, 2, ..., i, ..., n}; The first evaluation index is defined as E1, and the calculation formula is as follows: Where n is the total number of browsing intervals, SZt i ′ and ZSl i ′ respectively represent the AOI first fixation time and AOI fixation rate in the i-th browsing interval, and SZt i ′ and ZSl i ′ are all normalized and quantized to ensure that their values ​​are between 0 and 1, w1 and w2 are weight coefficients; The value range of the first evaluation index E1 is limited to [0,1]. When 0≤E1<0.33, it indicates low interest; when 0.33≤E1<0.66, it indicates medium interest; when 0.66≤E1≤1, it indicates high interest; Step S4, obtaining user task test samples, and classifying and marking them according to human-computer interaction functions, and obtaining corresponding second eye movement data for each human-computer interaction function for analysis and processing to generate a second evaluation index, wherein the second evaluation index is used to evaluate the subject's interest in each human-computer interaction function in the user task test samples; The classification of human-computer interaction functions is defined as {1, 2, ..., j, ..., h}; The second evaluation index is defined as E2, and the calculation formula is as follows: Among them, DRt j ′, HSt j ′, HSm j ' means after Quantized value after normalization, DRt j ′ is the second fixation time of the jth human-computer interaction function, HSt j ′ is the look-back time of the jth human-computer interaction function, HSm j ′ is the number of glances of the jth human-computer interaction function, α, β, γ are weight coefficients, which are used to adjust the contribution of the three eye movement data in the overall evaluation index to meet different evaluation requirements, and α+β+γ=1; h is the total number of human-computer interaction functions; Z is the normalization coefficient, which ensures that the value range of E2 is within [0,1], log(DRt j ′+1) is used to scale the effect of the second fixation time, making it more sensitive in a larger range of changes; is the Sigmoid function, which is used to map the lookback time to the range of (0,1). It is used to reduce the linear influence of the number of lookbacks on the results and limit the value range of E2 to [0,1]; When 0≤E2<0.25, it indicates low interest; when 0.25≤E2<0.75, it indicates medium interest; when 0.75≤E2≤1, it indicates high interest; Step S5, obtaining the first evaluation index, the second evaluation index, and the questionnaire score sheet, analyzing and processing them, and establishing a human-computer interaction evaluation model, and the human-computer interaction evaluation model can generate an evaluation adjustment strategy for the current test data set; The calculation formula for defining the human-computer interaction evaluation model is: Among them, P is the total evaluation index; U is the questionnaire score vector, which includes the normalized scores of interface design aesthetics, interface usability, information presentation clarity, interactive feedback satisfaction, personalization degree, and function discovery ease; Θ is the parameter of the logistic regression model; λ, μ, ν are weight coefficients used to adjust E1, E2, Contribution to the total evaluation P, and λ+μ+ν=1; set the value range of P to be in [0,1], where; 0≤P<0.18 indicates low satisfaction, which means that users are less satisfied with the human-computer interaction interface and need to be improved; 0.18≤P<0.48 indicates moderate satisfaction, where users are somewhat satisfied with the interface, but there is room for improvement; 0.48≤P≤1 indicates high satisfaction, and users are very satisfied with the human-computer interaction interface, which shows that the interface design and interaction logic meet user expectations.

2. The method for constructing an evaluation model for an automobile human-computer interaction system according to claim 1, characterized in that: The free browsing interface sample includes M main static interfaces in the current automobile human-computer interaction system as picture materials; The first eye movement data includes the AOI first fixation time and the AOI fixation rate, and the AOI first fixation time and the AOI fixation rate are calibrated as SZt and ZSl respectively; The AOI first fixation time SZt is the time when the subject first focuses his / her sight on the area of ​​interest AOI; The fixation rate ZSl is the frequency of the subject's fixation on the AOI, expressed as a percentage, and is calculated by the ratio of the total duration of the subject's fixation on the AOI to the total duration of the experiment; The user task test sample includes the test subject operating various functions of the product according to the specified script; The second eye movement data includes the second fixation time, the return time and the number of return times for each human-computer interaction function, and the second fixation time, the return time and the number of return times are marked as DRt, HSt and HSm respectively; The second fixation time DRt is the time it takes for the subject to return to and fixate on a human-computer interaction function after first fixating on the function; The return time HSt is the time interval for the subject to return to and gaze at a function after completing a human-computer interaction function. The number of look-backs HSm refers to the number of times the subject returns to and looks at a function after completing a human-computer interaction function.

3. The method for constructing an evaluation model for an automobile human-computer interaction system according to claim 2, characterized in that: The questionnaire score sheet is a visual preference score for the test data set, specifically including: The aesthetics of the interface design indicates the subjects’ aesthetic satisfaction rating of the color matching, icon design, and layout rationality of the car’s human-computer interaction interface, with a satisfaction rating of 1 and a dissatisfaction rating of 0; Interface usability, which indicates the user's evaluation of the intuitiveness of the interface operation, ease of learning, and whether the task can be completed quickly. The satisfaction score is 1, and the dissatisfaction score is 0; Information presentation clarity, which indicates the user's evaluation of the clarity and readability of information displayed on the car's human-computer interaction interface, with a satisfactory score of 1 and an unsatisfactory score of 0; Interaction feedback satisfaction, which indicates the user's satisfaction with the button click response and sliding smoothness during the interface interaction process, with a satisfaction score of 1 and a dissatisfaction score of 0; The degree of personalization indicates the user's evaluation of the interface's ability to be customized according to personal preferences, with a satisfaction score of 1 and a dissatisfaction score of 0; Function discovery ease, which indicates how easy it is for users to discover and understand the functions provided by the interface. A satisfactory score is 1, and an unsatisfactory score is 0.

4. The method for constructing an evaluation model for an automobile human-computer interaction system according to claim 3, characterized in that: The human-computer interaction evaluation model can generate an evaluation adjustment strategy for the current test data set, which specifically includes the following contents: For weights λ,μ, set the aesthetics of interface design, ease of use of interface, clarity of information presentation, satisfaction with interactive feedback, degree of personalization, and ease of function discovery as the dimensions in the questionnaire score sheet. When the score of any dimension is the lowest, increase the weight of the dimension by 10%; For the normalized AOI first fixation time SZt i ′ and AOI gaze rate ZSl i ', set the threshold to the average value Y1; If SZt i ′ or ZSl i ′ is lower than the mean Y1, and P < 0.48, it is considered that the meshing area of ​​the static interface needs to be improved to improve the attractiveness; The AOI first fixation time SZt threshold was set to 250 ms, a value lower than this indicated insufficient initial attraction; The adjustment strategy of AOI first fixation time SZt is: SZt″=SZt+Δ SZt , where SZt is the first fixation time of the original AOI, Δ SZt is the change in the first fixation time of AOI, SZt″ is the optimized adjustment value; If SZt is less than 250 milliseconds, Δ SZt The adjustment formula is as follows: Δ SZt =250-pcs+c1 It means that when SZt is lower than 250 milliseconds, the user's initial attention is improved by adjusting the current layout by 30% or highlighting the interface display by 20%, increasing the animation by 10%, and improving the color vividness by 10%, so as to enhance the visibility and attractiveness of the interface, where c1 is the adjustment coefficient; The AOI gaze rate ZSl threshold is set to 30%. A value lower than this indicates that the interface is not attractive enough; The AOI gaze rate ZSl adjustment strategy is: ZSl″=ZSl+Δ ZSl , where ZSl is the original fixation rate, Δ ZSl is the change in fixation rate; If ZSl is lower than 30%, adjust the layout by 35% or highlight the content by 28% to enhance the visibility and attractiveness of the interface to achieve ZSl+Δ ZSl ≥30%; For the normalized second fixation time DRt j ′, look-back time HSt j ′ and the number of times of looking back HSm j ', set the threshold to the average value Y2, If DRt j ′ or HSt j ′ or HSm j If any of ′ is higher than the average Y2 and 0.48≤P≤1, it indicates that users have a sustained interest in the function or are confused about its use, which requires further investigation; The second fixation time DRt threshold is set to 500 milliseconds. A value higher than this indicates that the user is interested in the function or has questions about its use. The second fixation time DRt adjustment strategy is: DRt″=DRt-Δ DRt , where DRt is the original second fixation time, DRt″ is the optimized adjustment value, Δ DRt is the amount of change; If DRt>500ms, user testing is required to confirm whether it is due to high interest or unclear operation, and optimization is carried out according to the feedback. If it is due to high interest, focus on analyzing and replicating the success factors; if it is due to unclear operation, simplify the current interactive function by 20% and highlight the visibility of each operation script by 30% to achieve Δ DRt +DRt>500; If DRt≤500ms, it means that users spend less time finding the functions they are interested in, or the operation is unclear. If the operation is unclear, simplify the current interactive function by 15% and highlight the visibility of each operation script by 20%; The HSt threshold of the re-look time is set to 2000 milliseconds. A value shorter than this indicates fast re-look, which indicates good usability. If HSt ≥ 2000 milliseconds, it is necessary to simplify the operation process by 20%, highlight the position of the function interface by 20%, increase the prompt information by 5%, and shorten the user's search time to improve usability; The return viewing count HSm threshold is set to 3 times. If it exceeds this value, it means that the user is highly interested in the function or is confused. If there is confusion, the difference in the operation process represented by the interactive operation script is increased by 25% and the difference in the position of the function interface is increased by 30%.

5. A system, characterized in that: The system is used to execute the method for constructing an evaluation model for an automobile human-computer interaction system according to any one of claims 1 to 4, comprising: Data preparation module: used to obtain and screen samples of the automobile human-computer interaction system to form a test data set, which includes free browsing interface samples and user task test samples, and determine corresponding test subjects and questionnaire scoring tables according to the test data sets, and the questionnaire scoring table is a visual preference score for the test data set; Data collection module: used to collect the first eye movement data of the subjects when they watch the free browsing interface samples, and the second eye movement data of the subjects when they watch the user task test samples. After the experiment, the questionnaire score sheet filled out by the subjects on the test data set is obtained; A first index generating module is used to obtain a free browsing interface sample and evenly divide it into a plurality of browsing intervals. For each browsing interval, first eye movement data is obtained for analysis and processing to generate a first evaluation index. The first evaluation index is used to evaluate the interest of the subject in each browsing interval in the free browsing interface sample. The second index generation module is used to obtain user task test samples, classify and label them according to human-computer interaction functions, and obtain corresponding second eye movement data for each human-computer interaction function for analysis and processing to generate a second evaluation index, wherein the second evaluation index is used to evaluate the interest of the subject in each human-computer interaction function in the user task test samples; Adjustment strategy generation module: used to obtain the first evaluation index, the second evaluation index, and the questionnaire score sheet, perform analysis and processing, and establish a human-computer interaction evaluation model, and the human-computer interaction evaluation model can generate an evaluation adjustment strategy for the current test data set.

Images