Method for evaluating man-machine interaction efficiency of central control screen of intelligent automobile

Abstract

The invention discloses a human-computer interaction efficiency evaluation method for a central control screen of an intelligent automobile, and belongs to the technical field of intelligent automobile interaction. Determining use requirements of the central control screen; the method comprises the following steps: acquiring indexes in three aspects of driver response behavior, vehicle control and user experience, performing fuzzy C-means clustering on three dimensions for each vehicle using scene to obtain clustering scores of each sample on different dimensions in different vehicle using scenes, calculating weights of all dimensions in different vehicle using scenes according to the clustering scores, and calculating the weights of all the dimensions in different vehicle using scenes; calculating average clustering scores of all the samples in all the vehicle use scenes, and judging the man-machine interaction efficiency level of the central control screen of the intelligent vehicle; in combination with a Fuzzy C-means clustering algorithm and an entropy weight method, clustering results of different dimensions in each scene and the weight of each dimension in different vehicle use scenes are adaptively calculated, and the objectivity and generalization ability of data are improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent vehicle interaction technology, specifically a method for evaluating the human-computer interaction efficiency of a central control screen in an intelligent vehicle. Background Technology

[0002] With the development of intelligent driving technology, the car's central control screen, as the core platform for interaction between the driver and the in-vehicle system, has gradually taken on multiple functions such as navigation, entertainment, communication, and in-vehicle information query. However, when performing these tasks, drivers often face operational pressure in different driving scenarios, and the efficiency of operation directly affects driving safety and user experience. Therefore, evaluating the interaction efficiency of the central control screen is particularly important.

[0003] In existing technologies, interaction efficiency is typically evaluated by observing single metrics such as reaction time, recovery time, and operation success rate. However, these methods neglect the combined effect of multiple factors in actual operation. Furthermore, existing methods fail to consider the differences between driver behavior and system response in different driving scenarios.

[0004] Currently, many studies employ data analysis and model evaluation methods to assess the interaction efficiency of central control screens. For example, some studies have used traditional statistical analysis methods to collect and analyze user behavior data, deriving evaluation models for operation response time and recovery time. However, most of these methods are based on simple linear models and do not fully consider the changing factors under different scenarios, as well as the interaction of multi-dimensional data.

[0005] Some machine learning-based research, such as support vector machines (SVM) and cluster analysis, has been able to evaluate from more complex perspectives, but these methods often lack scene adaptability and rely on subjective evaluation in dimension integration and weight allocation, resulting in limited objectivity and generalization ability of the results.

[0006] Based on this, the present invention aims to provide a more accurate and objective method for evaluating the interaction efficiency of automotive central control screens. Summary of the Invention

[0007] To address the shortcomings of existing technologies, this invention provides a method for evaluating the human-computer interaction efficiency of a smart car central control screen. This method utilizes a multi-dimensional (driver reaction behavior, vehicle control, user experience) evaluation system, combined with Fuzzy C-means clustering algorithm and entropy weighting method, to adaptively calculate the clustering results for different dimensions in each scenario and the weight of each dimension in different vehicle usage scenarios. This improves the objectivity and generalization ability of the data, and a comprehensive interaction efficiency score is derived based on these results.

[0008] To achieve the above objectives, the present invention provides the following technical solution: A method for evaluating the human-computer interaction efficiency of a central control screen in an intelligent vehicle includes the following steps: (1) Construct a vehicle usage scenario in a closed road environment and simulate the interaction between the driver and the intelligent vehicle central control screen to make task requests under different vehicle usage scenarios; (2) Arrange multiple drivers to drive in different usage scenarios. The drivers complete the task requirements of the intelligent vehicle central control screen and record data in three dimensions: driver reaction behavior, vehicle control, and user experience during this process. (3) For each car use scenario, perform fuzzy C-means clustering on the three dimensions respectively, calculate the cluster center of each sample in different dimensions under different car use scenarios, and the membership degree of each sample to each cluster in different dimensions under different car use scenarios, so as to calculate the clustering score of each sample in different dimensions under different car use scenarios. (4) Calculate the information entropy of each dimension in different scenarios by using the clustering scores of each sample in different dimensions under different car use scenarios, and then calculate the weight of each dimension in different car use scenarios. (5) Based on the weights of each dimension under different car use scenarios and the clustering scores of each sample under different dimensions in different scenarios, calculate the clustering score of each sample under different car use scenarios, and then calculate the average clustering score of all samples under different car use scenarios. (6) Calculate the average clustering score of all samples in all driving scenarios based on the average clustering score of all samples in different driving scenarios, and determine the human-computer interaction efficiency level of the central control screen of the intelligent vehicle accordingly.

[0009] In this invention, the vehicle usage scenarios include: The vehicle operates at a low speed of 0-20 km / h, simulating urban congestion or parking lot scenarios; The vehicle operates at a medium speed of 20-60 km / h, simulating urban expressways or suburban roads; The vehicle operates at a high speed of 60-120 km / h, simulating a highway or expressway.

[0010] In this invention, the task requirements of the intelligent vehicle central control screen include: Basic tasks include adjusting the seat, adjusting the air conditioning, making a phone call, and starting the navigation system.

[0011] Medium-complexity tasks: such as switching in-car music, adjusting rearview mirrors, turning on / off driver assistance systems, and viewing in-car information.

[0012] Highly complex tasks: such as vehicle self-diagnosis in emergency situations and rapid switching of driving modes.

[0013] In this invention, the indicators of driver reaction behavior include: operation reaction time, distraction time, and recovery time; Vehicle control metrics include: lane departure distance, speed difference, and operation success rate; User experience metrics include: satisfaction, ease of use, and cognitive load.

[0014] In this invention, fuzzy C-means clustering is performed on three dimensions for each vehicle usage scenario, and the data is then forwarded and standardized, including: set up For car usage scenarios Next, the Each sample in dimension The The original values ​​of each indicator; ,in The total number of samples for each indicator in each scenario. This represents responsiveness, vehicle control, and user experience. Using Min-Max standardization Processed as ; For positive indicators: (1); For negative indicators: (2); in, and Car usage scenarios Dimension Lowering the target exist The minimum and maximum values ​​in each sample.

[0015] In this invention, the clustering scores of each sample in different dimensions under different vehicle usage scenarios in step (3) are calculated as follows: (2.1) Integration of Dimensional Indicators Car usage scenarios Below, sample In dimensions Standardized indicators Combined into a single scalar representation : (3); Where K is the dimension The total number of indicators above; (2.2) FCM clustering For each car usage scenario and each dimension FCM is executed independently; the FCM algorithm iteratively optimizes the membership degree. and cluster center Minimize the objective function ; objective function : (4); in, For cluster index, , Represents the fuzzy factor. For car usage scenarios Lower sample In dimensions j The membership degree of the upper cluster, For car usage scenarios Lower sample The cluster center in dimension j; Car usage scenarios Lower sample i In dimensions j Cluster center on Iteration formula: (5); Car usage scenarios Lower sample i In dimensions j Membership degree of the upper cluster Iteration formula: (6); in, Summation of terms in the denominator The iteration index variable used in the process also has a range of values. This is used to iterate through all clusters; (2.3) Cluster score extraction Based on membership degree and cluster center Calculate vehicle usage scenarios Lower sample In dimensions Clustering scores : (7).

[0016] In this invention, the calculation process for the weights of each dimension under different vehicle usage scenarios is as follows: (3.1) Calculate the information entropy of each dimension First, regarding vehicle usage scenarios Lower sample i In dimensions j Clustering scores After normalization, we get : (8); Then, calculate the vehicle usage scenario. lower dimension Information entropy : (9); (3.2) Calculate the weight of each dimension under different vehicle usage scenarios Calculate vehicle usage scenarios based on information entropy. s lower dimension j weight : (10).

[0017] In this invention, a vehicle usage scenario is used. s lower dimension j weight and car usage scenarios Lower sample In dimensions Clustering scores Perform a weighted summation to obtain the vehicle usage scenario. Lower sample Clustering score : (11).

[0018] In this invention, the scene All sample scores below Take the average value to obtain the scene. Interaction efficiency score : (12); in, For car usage scenarios Lower sample Clustering score; , , These represent the average clustering scores of all samples in the three driving scenarios: low speed, medium speed, and high speed, respectively. The average clustering scores of all samples across the three scenarios are then averaged again to obtain the average clustering score for all vehicle usage scenarios. : (13).

[0019] In this invention, the average clustering score of all samples under different vehicle usage scenarios is used. The human-machine interaction efficiency levels of intelligent vehicle central control screens are classified according to the following criteria: when When the value is ≥0.8, the human-machine interaction efficiency of the central control screen of the tested vehicle is excellent; When 0.7≤ When the value is less than 0.8, the human-machine interaction efficiency of the central control screen of the tested vehicle is good; When 0.6≤ When the efficiency is less than 0.7, the human-machine interaction efficiency of the central control screen of the tested vehicle is qualified. when When the efficiency is less than 0.6, the human-machine interaction efficiency of the central control screen of the tested vehicle is unqualified.

[0020] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention combines multi-dimensional clustering with scene adaptation. It uses the Fuzzy C-means clustering method to analyze three dimensions: driver reaction behavior, vehicle control, and user experience. The weights of driver reaction behavior, vehicle control, and user experience are dynamically adjusted according to different driving scenarios. This can accurately reflect the actual needs and interaction efficiency in different driving environments. It fully considers the different contributions of different dimensions and scenarios to interaction efficiency, making the final interaction efficiency score more accurate.

[0021] 2. This invention comprehensively captures driver behavior, vehicle response, and system user experience through feature data from three dimensions: driver reaction behavior, vehicle control, and user experience. This provides a more comprehensive evaluation of interaction efficiency. Compared with single-dimensional evaluation, it avoids information loss and improves the accuracy of the evaluation.

[0022] 3. This invention uses the entropy weight method to calculate the weights of each dimension. The entropy weight method can allocate weights according to the distribution characteristics of the data itself, rather than relying on expert experience, thus avoiding the subjective bias that may be caused by manually setting weights and ensuring the objectivity and reliability of weight allocation. Attached Figure Description

[0023] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0024] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0025] like Figure 1 As shown, this invention discloses a method for evaluating the human-computer interaction efficiency of a central control screen in an intelligent vehicle, specifically including: Step 1: Build a vehicle usage scenario in a closed road environment, simulate the interaction between the driver and the intelligent vehicle central control screen to make task requests under different vehicle usage scenarios, and collect data on the driver's reaction behavior, vehicle control, and user experience during the interaction between the driver and the intelligent vehicle central control screen.

[0026] (1) Scene determination Based on daily car usage scenarios, it can be divided into: Low speed (0-20 km / h): Simulates urban congestion or parking lot scenarios.

[0027] Medium speed (20-60 km / h): Simulated on urban expressways or suburban roads.

[0028] High-speed conditions (60-120 km / h): Simulates highways or expressways.

[0029] (2) Task definition Design different combinations of tasks to ensure the tests reflect the actual needs and behaviors of drivers in various situations. Tasks can be designed from basic to complex: Basic tasks include adjusting the seat, adjusting the air conditioning, making a phone call, and starting the navigation system.

[0030] Medium-complexity tasks: such as switching in-car music, adjusting rearview mirrors, turning on / off driver assistance systems, and viewing in-car information.

[0031] Highly complex tasks: such as vehicle self-diagnosis in emergency situations, rapid switching of driving modes, and adjustment of various settings (e.g., seat + steering wheel + rearview mirror).

[0032] (3) Selection of indicators The selection of indicators takes into account three aspects: driver reaction behavior, vehicle control, and user experience.

[0033] (3.1) Driver's reaction behavior Operational reaction time: The time (s) from when the driver receives the task instruction to when the specific task is completed.

[0034] Distraction time: The time (in seconds) during which a driver's visual attention is diverted from the driving task.

[0035] Recovery time: The time (s) from when the driver completes the task to when they fully return to driving.

[0036] (3.2) Vehicle control Lane offset: The average distance (m) that a vehicle deviates from the centerline of the road during the execution of a task.

[0037] Speed ​​difference: The change in vehicle speed (km / h) before and after the mission.

[0038] Operation success rate: Whether the driver completes the operation task (categorical variable, 0 / 1).

[0039] (3.3) User experience (questionnaire survey).

[0040] Satisfaction: The satisfaction questionnaire was used to measure the satisfaction using the Likert scale (1-5 points).

[0041] Usability: Standardized System Usability Scale (SUS).

[0042] Cognitive Load: Standardized Cognitive Composite Scale, NASA-TLX.

[0043] Satisfaction Questionnaire

[0044] SUS System Usability Questionnaire

[0045] Cognitive Load Questionnaire (NASA-TLX)

[0046] Compare pairs and select the factors that have a greater workload during the task.

[0047] (4) Operational test (4.1) Select 30 drivers of different ages, genders and driving experience.

[0048] (4.2) Each driver drives in the test site and randomly selects 3 tasks of different complexity, and performs them once in each of the 3 scenarios.

[0049] (4.3) Collect sample indicator data under different scenarios.

[0050] Step 2: Extract features from the three dimensions of data using the Fuzzy C-means (FCM) clustering method, then integrate them to obtain the clustering result for each dimension, i.e., the evaluation result for each dimension. By weighted synthesis of these clustering results from different dimensions, the accuracy of the analysis is further enhanced, thus obtaining the final evaluation result.

[0051] (1) Data forwarding and standardization To eliminate the dimensional differences between different indicators, the collected raw data needs to be normalized and standardized. Let... For car usage scenarios Next, the Each sample in dimension The Individual indicators ( The original value on ); , representing low speed, medium speed, and high speed; ,in For the scene The total number of samples for the following indicator. This represents responsiveness, vehicle control, and user experience; Min-Max standardization is used to... Processed as .

[0052] For positive metrics (operation success rate, satisfaction, and ease of use): (1) For negative metrics (recovery time, reaction time, distraction time, lane departure distance, speed difference, and ease of use): (2) in, and Car usage scenarios Dimension Lowering the target exist The minimum and maximum values ​​in each sample.

[0053] (2) Multidimensional joint clustering To better understand the independent contribution of each dimension to interaction efficiency, and to avoid potential redundancy or interference between different dimensions during clustering, we first cluster the metrics for each dimension, and then integrate the clustering results of all dimensions. Data from all dimensions is clustered using FCM (Fuzzy Clustering Method). FCM is a classic fuzzy clustering algorithm whose core idea is to introduce the concept of "fuzzy membership" into traditional clustering, allowing a sample to belong to multiple clusters simultaneously.

[0054] (2.1) Integration of Dimensional Indicators Car usage scenarios s Below, sample In dimensions Three standardized indicators Combined into a single scalar representation : (3) (2.2) FCM clustering For each car usage scenario and each dimension FCM is executed independently. The FCM algorithm iteratively optimizes the membership degree. and cluster center Minimize the objective function .

[0055] objective function : (4); in, For car usage scenarios Lower sample i In dimensions j The membership degree of the upper cluster, For car usage scenarios Lower sample i The cluster center in dimension j; For cluster index, In this assessment method, The value should be preset according to the actual needs of the evaluation to ensure that the clustering results have a clear evaluation meaning. For example, it can be set to... To correspond to three efficiency levels: "poor," "medium," and "excellent," or to set... The four levels are "unqualified", "qualified", "good" and "excellent". The fuzzy factor is typically taken as... .

[0056] Cluster center The iterative formula is: (5); Membership degree The iterative formula is: (6); in, Summation of terms in the denominator The iteration index variable used in the process also has a range of values. This is used to traverse all clusters.

[0057] (2.3) Cluster score extraction After the FCM algorithm converges, the obtained membership degrees are used... and cluster center ,calculate Scene Samples In dimensions Clustering scores : (7); (3) Determining the weights of multiple indicators To obtain objective weights for clustering results across various dimensions under different vehicle usage scenarios, the entropy weighting method is used to assign weights to each dimension. The entropy weighting method is a weighting method commonly used in multi-indicator decision analysis. It aims to determine weights by calculating the information entropy of each indicator, thereby rationally allocating weights to each indicator. The entropy weighting method is objective, avoiding the influence of subjective human factors, and is widely used in the weighted calculation of clustering results and multi-dimensional evaluations.

[0058] (3.1) Calculate the information entropy of each dimension Information entropy reflects the degree of uncertainty or dispersion of a dimension in data. The higher the entropy, the more dispersed the information in that dimension is, and the lower its weight; the lower the entropy, the more concentrated the information in that dimension is, and the higher its weight.

[0059] First, regarding vehicle usage scenarios Lower sample i In dimensions j Clustering scores After normalization, we get : (8); Then, calculate the vehicle usage scenario. lower dimension Information entropy : (9); (3.2) Calculate the weight of each dimension under different vehicle usage scenarios Calculate vehicle usage scenarios based on information entropy. s lower dimension j weight : (10); (4) Calculate the final interaction efficiency score (4.1) Calculate the final score of the sample in each scenario.

[0060] Car usage scenarios s lower dimension j weight and car usage scenarios Lower sample In dimensions Clustering scores Perform a weighted summation to obtain the vehicle usage scenario. Lower sample Clustering score : (11); (4.2) Calculate the final score for each scene.

[0061] For car usage scenarios Lower sample Clustering score Take the average value to obtain the car usage scenario. The average cluster score of all samples : (12); From this, we can obtain the low-speed ( ), medium speed ( ) and high speed ( The average clustering score of all samples in the three scenarios.

[0062] (4.3) Calculate the overall interaction efficiency score The average clustering scores of all samples across the three scenarios are then averaged again to obtain the average clustering score of all samples across all vehicle usage scenarios. : (13); this The value is the final result of the evaluation and is used to classify the level of interaction efficiency.

[0063] Step 3: Classify the current central control screen system into different interaction levels according to the classification standards, and provide evaluation opinions.

[0064] The standards for classifying the human-machine interaction efficiency levels of intelligent vehicle central control screens are as follows: when When the value is ≥0.8, the human-machine interaction efficiency of the central control screen of the tested vehicle is excellent.

[0065] When 0.7≤ When the value is less than 0.8, the human-machine interaction efficiency of the central control screen of the tested vehicle is good.

[0066] When 0.6≤ When the efficiency is less than 0.7, the human-machine interaction efficiency of the central control screen of the tested vehicle is qualified and can be appropriately optimized.

[0067] when When the efficiency is less than 0.6, the human-machine interaction efficiency of the central control screen of the tested vehicle is unqualified and needs to be improved.

[0068] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for evaluating the human-computer interaction efficiency of a central control screen in an intelligent vehicle, characterized in that, Includes the following steps: (1) Construct a vehicle usage scenario in a closed road environment and simulate the interaction between the driver and the intelligent vehicle central control screen to make task requests under different vehicle usage scenarios; (2) Arrange multiple drivers to drive in different usage scenarios. The drivers complete the task requirements of the intelligent vehicle central control screen and record data in three dimensions: driver reaction behavior, vehicle control, and user experience during this process. (3) For each car use scenario, perform fuzzy C-means clustering on the three dimensions respectively, calculate the cluster center of each sample in different dimensions under different car use scenarios, and the membership degree of each sample to each cluster in different dimensions under different car use scenarios, so as to calculate the clustering score of each sample in different dimensions under different car use scenarios. (4) Calculate the information entropy of each dimension in different scenarios by using the clustering scores of each sample in different dimensions under different car use scenarios, and then calculate the weight of each dimension in different car use scenarios. (5) Based on the weights of each dimension under different car use scenarios and the clustering scores of each sample under different dimensions in different scenarios, calculate the clustering score of each sample under different car use scenarios, and then calculate the average clustering score of all samples under different car use scenarios. (6) Calculate the average clustering score of all samples in all driving scenarios based on the average clustering score of all samples in different driving scenarios, and determine the human-computer interaction efficiency level of the central control screen of the intelligent vehicle accordingly.

2. The method for evaluating the human-computer interaction efficiency of a central control screen in an intelligent vehicle according to claim 1, characterized in that, The vehicle usage scenarios include: The vehicle operates at a low speed of 0-20 km / h, simulating urban congestion or parking lot scenarios; The vehicle operates at a medium speed of 20-60 km / h, simulating urban expressways or suburban roads; The vehicle operates at a high speed of 60-120 km / h, simulating a highway or expressway.

3. The method for evaluating the human-computer interaction efficiency of a central control screen in an intelligent vehicle according to claim 1, characterized in that, The requirements for intelligent vehicle central control screens include: Basic tasks: such as adjusting the seat, adjusting the air conditioning, making a phone call, and starting the navigation; Medium-complexity tasks: such as switching in-car music, adjusting rearview mirrors, turning on / off driver assistance systems, and viewing in-car information; Highly complex tasks: such as vehicle self-diagnosis in emergency situations and rapid switching of driving modes.

4. The method for evaluating the human-computer interaction efficiency of a central control screen in an intelligent vehicle according to claim 1, characterized in that, Indicators of driver reaction behavior include: reaction time, distraction time, and recovery time; Vehicle control metrics include: lane departure distance, speed difference, and operation success rate; User experience metrics include: satisfaction, ease of use, and cognitive load.

5. The method for evaluating the human-computer interaction efficiency of a central control screen in an intelligent vehicle according to claim 1, characterized in that, For each vehicle usage scenario, fuzzy C-means clustering was performed across three dimensions, and the data underwent positive and standardization processing, including: set up For car usage scenarios Next, the Each sample in dimension The The original values ​​of each indicator; ,in The total number of samples for each indicator in each scenario. This represents responsiveness, vehicle control, and user experience. Using Min-Max standardization Processed as ; For positive indicators: (1)； For negative indicators: (2)； in, and Car usage scenarios Dimension Lowering the target exist The minimum and maximum values ​​in each sample.

6. The method for evaluating the human-computer interaction efficiency of a central control screen in an intelligent vehicle according to claim 1, characterized in that, In step (3), the clustering scores of each sample in different dimensions under different car usage scenarios are calculated as follows: (2.1) Integration of Dimensional Indicators Car usage scenarios Below, sample In dimensions Standardized indicators Combined into a single scalar representation : (3)； Where K is the dimension The total number of indicators above; (2.2) FCM clustering For each car usage scenario and each dimension FCM is executed independently; the FCM algorithm iteratively optimizes the membership degree. and cluster center Minimize the objective function ; objective function : (4)； in, For cluster index, , Represents the fuzzy factor. For car usage scenarios Lower sample In dimensions j The membership degree of the upper cluster, For car usage scenarios Lower sample The cluster center in dimension j; Car usage scenarios Lower sample i In dimensions j Cluster center on Iteration formula: (5)； Car usage scenarios Lower sample i In dimensions j Membership degree of the upper cluster Iteration formula: (6)； in, Summation of terms in the denominator The iteration index variable used in the process also has a range of values. This is used to iterate through all clusters; (2.3) Cluster score extraction Based on membership degree and cluster center Calculate vehicle usage scenarios Lower sample In dimensions Clustering scores : (7)。 7. The method for evaluating the human-computer interaction efficiency of a central control screen in an intelligent vehicle according to claim 1, characterized in that, The calculation process for the weights of each dimension under different vehicle usage scenarios is as follows: (3.1) Calculate the information entropy of each dimension First, regarding vehicle usage scenarios Lower sample i In dimensions j Clustering scores After normalization, we get : (8)； Then, calculate the vehicle usage scenario. lower dimension Information entropy : (9)； (3.2) Calculate the weight of each dimension under different vehicle usage scenarios Calculate vehicle usage scenarios based on information entropy. s lower dimension j weight : (10)。 8. The method for evaluating the human-computer interaction efficiency of a central control screen in an intelligent vehicle according to claim 1, characterized in that, Vehicle usage scenarios s lower dimension j weight and car usage scenarios Lower sample In dimensions Clustering scores Perform a weighted summation to obtain the vehicle usage scenario. Lower sample Clustering score : (11)。 9. The method for evaluating the human-computer interaction efficiency of a central control screen in an intelligent vehicle according to claim 1, characterized in that, For the scene All sample scores below Take the average value to obtain the scene. Interaction efficiency score : (12)； in, For car usage scenarios Lower sample Clustering score; , , These represent the average clustering scores of all samples in the three driving scenarios: low speed, medium speed, and high speed, respectively. The average clustering scores of all samples across the three scenarios are then averaged again to obtain the average clustering score for all vehicle usage scenarios. : (13)。 10. The method for evaluating the human-computer interaction efficiency of a central control screen in an intelligent vehicle according to claim 1, characterized in that, Based on the average clustering score of all samples under different car usage scenarios The human-machine interaction efficiency levels of intelligent vehicle central control screens are classified according to the following criteria: when When the value is ≥0.8, the human-machine interaction efficiency of the central control screen of the tested vehicle is excellent; When 0.7≤ When the value is less than 0.8, the human-machine interaction efficiency of the central control screen of the tested vehicle is good; When 0.6≤ When the efficiency is less than 0.7, the human-machine interaction efficiency of the central control screen of the tested vehicle is qualified. when When the efficiency is less than 0.6, the human-machine interaction efficiency of the central control screen of the tested vehicle is unqualified.

Images