Automobile product interactive design system based on multi-dimensional user experience data analysis

Abstract

The invention provides an automobile product interactive design system based on multi-dimensional user experience data analysis, and relates to the technical field of automobile engineering. The master control system comprises a data collection and preprocessing module, a user behavior analysis module, a user portrait generation module, an interactive design optimization module, a real-time feedback and adjustment module and a visualization and interactive feedback module. And the interactive interface is dynamically adjusted. The personalized design not only improves the driving comfort of the user, but also enhances the sense of belonging to the vehicle and the satisfaction degree of the user to the vehicle. In addition, the intelligent function of the system can learn the behavior pattern of the user, continuously optimize the interactive experience, and ensure that each driving is customized.

Description

Technical Field

[0001] The present invention relates to the field of automobile engineering technology, and in particular to an automobile product interaction design system based on multi-dimensional user experience data analysis. Background Art

[0002] In today's automotive industry, user experience has become one of the core elements of design and manufacturing. With the rapid development of intelligence and digitalization, many automakers have begun to use advanced technical means to enhance the user's interactive experience. Among them, user interaction design systems based on data analysis have gradually become a hot topic in the industry. These systems usually rely on big data, artificial intelligence and machine learning technologies to provide personalized functions and services by analyzing user behavior, preferences and feedback. For example, some high-end cars are equipped with intelligent voice assistants that can understand natural language and perform multiple operations according to the driver's instructions, thereby improving driving convenience and safety.

[0003] However, there are still many problems in the application of existing technologies:

[0004] 1. Most existing user interaction systems can only perform simple pattern recognition and cannot deeply understand the complex behaviors and emotional states of users. Many systems process user feedback slowly, causing users to feel delays during the operation, affecting the overall experience.

[0005] 2. Users' expectations for interactive experience continue to increase. Traditional design methods can no longer meet personalized and diversified needs. Many existing systems do not fully consider the user's driving experience in different environments, resulting in users not being able to obtain ideal support in certain situations. For example, drivers may need stronger visual support when driving at night, and more flexible navigation prompts in congested urban environments.

[0006] Therefore, there is an urgent need for an innovative automotive product interaction design system that can integrate multi-dimensional user experience data for in-depth analysis and real-time response. Summary of the invention

[0007] Technical issues solved

[0008] In view of the shortcomings of the prior art, the present invention provides an automobile product interaction design system based on multi-dimensional user experience data analysis, which solves the following problems:

[0009] 1. It solves the problem that although many systems can collect and process large amounts of data, they are still insufficient in real-time analysis and response to user needs.

[0010] 2. Solved the problem that many systems process user feedback slowly, causing users to feel delays during operations and affecting the overall experience.

[0011] 3. Solve the problem that users' expectations for interactive experience are constantly increasing, and traditional design methods can no longer meet personalized and diversified needs. Many existing systems do not fully consider the user's driving experience in different environments, resulting in users not being able to get ideal support in certain situations.

[0012] Technical Solution

[0013] To achieve the above objectives, the present invention is implemented through the following technical solutions: an automobile product interactive design system based on multi-dimensional user experience data analysis, including a master control system, wherein the master control system includes a data collection and preprocessing module, a user behavior analysis module, a user portrait generation module, an interactive design optimization module, a real-time feedback and adjustment module, and a visualization and interactive feedback module;

[0014] Sp1, the data collection and preprocessing module cleans and preprocesses user data through rule-based data cleaning, anomaly detection algorithm and data normalization technology, and extracts user features through principal component analysis, independent component analysis and feature selection technology to provide high-quality data for subsequent analysis;

[0015] Sp2, the user behavior analysis module uses K-means clustering, DBSCAN and hierarchical clustering algorithms to perform cluster analysis on user behavior, and combines Markov chain, hidden Markov model and sequence pattern mining technology to deeply explore the user's behavior patterns in different situations, so as to identify user needs;

[0016] Sp3, the user portrait generation module uses decision tree, support vector machine and random forest classification algorithms to generate user portraits, and combines graph embedding algorithms such as Node2Vec, GraphSAGE and relational network embedding to provide users with comprehensive personalized portraits;

[0017] Sp4, the interaction design optimization module optimizes the interaction design through collaborative filtering, content recommendation and model-based recommendation algorithms, and applies genetic algorithms, particle swarm optimization and constraint optimization techniques to improve user experience on the basis of meeting multi-objective requirements;

[0018] Sp5, the real-time feedback and adjustment module uses Q-learning, deep Q network and policy gradient method reinforcement learning algorithms to dynamically adjust the system response according to the user's real-time data, and combines ARI MA, LSTM and Prophet, time series prediction algorithms to identify changes in user needs in advance;

[0019] Sp6, the visualization and interactive feedback module generates data visualization technology through data aggregation, heat map analysis and dynamic interactive charts to intuitively display user behavior and feedback, and combines dictionary-based sentiment analysis, machine learning classifiers and deep learning models to provide real-time sentiment analysis of user feedback;

[0020] Sp7. The data collection and preprocessing module, user behavior analysis module, user portrait generation module, interactive design optimization module, real-time feedback and adjustment module, and visualization and interactive feedback module all interact with each other through standardized interfaces. The master control system stores and processes data through a cloud platform, supporting large-scale user data analysis and real-time response.

[0021] Preferably, the data collection and preprocessing module realizes real-time reception and processing of data through multiple data sources to enhance the comprehensiveness and effectiveness of the data. The data collection and preprocessing module obtains vehicle operation data in real time through vehicle-mounted sensors, wherein the operation data includes vehicle speed, acceleration and steering angle. When applied, the user's driving style (aggressive or smooth) is inferred through data changes of the acceleration sensor, providing a reference for subsequent interaction design. For users with aggressive driving styles, the interaction prompts for optimizing vehicle power response are optimized.

[0022] Preferably, the user behavior analysis module adopts a variety of clustering and sequence analysis algorithms to form an in-depth understanding of user behavior patterns and reveal potential changes in user needs. The K-means clustering algorithm in the user behavior analysis module groups users based on their driving operation habits, that is, acceleration and braking frequency and amplitude.

[0023] Preferably, the user portrait generation module adopts a dynamic update mechanism to dynamically update the user portrait to ensure that it continues to reflect the user's personalized characteristics and preferences. The user portrait generation module monitors the user's online behavior in real time. When the user uses the in-vehicle navigation system to search for a new destination while driving, the system immediately records this behavior and updates the user's interest preference tag according to the type of destination. It recommends nearby outdoor sports equipment stores or related activity information on the homepage of the in-vehicle display screen, and prioritizes displaying convenient routes to outdoor places in the navigation system.

[0024] Preferably, the interaction design optimization module adopts an intelligent integration mechanism to integrate user feedback and make intelligent adjustments to interactive elements to improve the quality of user experience. The interaction design optimization module not only considers users' direct evaluation of the vehicle's existing interactive functions, but also analyzes users' text descriptions of their vehicle usage experience on social media, online forums and other channels through natural language processing technology.

[0025] Preferably, the real-time feedback and adjustment module integrates an anomaly detection mechanism to identify abnormal patterns of user behavior and generate corresponding feedback information. The anomaly detection mechanism in the real-time feedback and adjustment module adopts a statistical model-based method to establish a statistical distribution model of normal driving behavior for the user's driving behavior data, including the mean and standard deviation of the vehicle speed, and the frequency distribution of the steering wheel rotation angle, and combines a machine learning algorithm for anomaly detection.

[0026] Preferably, the visualization and interactive feedback module supports a variety of data visualization forms, supports interactive dashboard visualization, and uses map visualization technology, combined with the vehicle navigation system, to display the user's driving trajectory, frequently visited places and surrounding points of interest information in the form of a map.

[0027] Preferably, the master control system combines machine learning algorithms to optimize data processing and analysis processes, uses deep learning-based image recognition algorithms to analyze image data captured by vehicle-mounted cameras, and automatically identifies different driving scenarios and traffic conditions on urban roads, highways, and rural roads.

[0028] Preferably, the master control system has flexible expansion capabilities and can quickly integrate new data sources and analysis functions according to new needs. The master control system adopts a microservice architecture, splitting the data collection and preprocessing module, user behavior analysis module, user portrait generation module, interactive design optimization module, real-time feedback and adjustment module, and visualization and interactive feedback module into independent microservices.

[0029] Preferably, the data collection and preprocessing module, user behavior analysis module, user portrait generation module, interaction design optimization module, real-time feedback and adjustment module and visualization and interaction feedback module all interact with each other through standardized interfaces, using a Web service interface based on a RESTful architecture.

[0030] Beneficial Effects

[0031] The present invention provides an automobile product interaction design system based on multi-dimensional user experience data analysis. It has the following beneficial effects:

[0032] 1. The present invention can identify user preferences and needs in real time and dynamically adjust the interactive interface through comprehensive analysis of multi-dimensional user experience data. This personalized design not only improves the user's driving comfort, but also enhances the user's sense of belonging and satisfaction with the vehicle. In addition, the system's intelligent function can learn the user's behavior patterns, continuously optimize the interactive experience, and ensure that every driving is tailored.

[0033] 2. The present invention uses a real-time feedback and adjustment module, so that the system can issue an alarm in time when the user's behavior is abnormal, effectively improving driving safety. This early warning mechanism not only pays attention to the user's operating behavior, but also makes intelligent decisions based on environmental changes to ensure that a high level of safety is always maintained during driving. In addition, the system's instant feedback function can quickly respond to user needs, provide users with timely guidance and help, and improve overall driving safety.

[0034] 3. The present invention can systematically analyze user feedback and promote continuous improvement of product design through the integrated user portrait generation and interactive design optimization module. By analyzing historical data, the system can not only identify potential design defects, but also predict future changes in user needs, thereby providing data support for automobile manufacturers to help them continuously optimize product design and functions in a highly competitive market. This data-driven approach enables automobile products to quickly adapt to market changes and improves the company's innovation capabilities and market competitiveness. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is a processing flow chart of the master control system of the present invention;

[0036] Figure 2 The hardware composition diagram of the master control system of the present invention;

[0037] Figure 3 This is a histogram of the user experience simulation analysis of the present invention. DETAILED DESCRIPTION

[0038] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. Specific embodiment one:

[0040] like Figure 1-3 As shown, the automotive product interaction design system based on multi-dimensional user experience data analysis includes a master control system, which includes a data collection and preprocessing module, a user behavior analysis module, a user portrait generation module, an interaction design optimization module, a real-time feedback and adjustment module, and a visualization and interaction feedback module;

[0041] Sp1, data collection and preprocessing module, cleans and preprocesses user data through rule-based data cleaning, anomaly detection algorithm and data normalization technology, and extracts user features through principal component analysis, independent component analysis and feature selection technology to provide high-quality data for subsequent analysis;

[0042] Sp2, the user behavior analysis module uses K-means clustering, DBSCAN and hierarchical clustering algorithms to perform cluster analysis on user behavior, and combines Markov chain, hidden Markov model and sequence pattern mining technology to deeply explore the user's behavior patterns in different situations, so as to identify user needs;

[0043] Sp3, the user portrait generation module uses decision trees, support vector machines and random forest classification algorithms to generate user portraits, and combines graph embedding algorithms such as Node2Vec, GraphSAGE and relational network embedding to provide users with comprehensive personalized portraits;

[0044] Sp4, the interactive design optimization module optimizes the interactive design through collaborative filtering, content recommendation and model-based recommendation algorithms, and applies genetic algorithms, particle swarm optimization and constraint optimization techniques to improve user experience on the basis of meeting multi-objective requirements;

[0045] Sp5, real-time feedback and adjustment module uses Q-learning, deep Q network and policy gradient method reinforcement learning algorithm to dynamically adjust the system response according to the user's real-time data, and combines ARI MA, LSTM and Prophet, time series prediction algorithms to identify changes in user needs in advance;

[0046] Sp6, Visualization and Interactive Feedback Module generates data visualization technology through data aggregation, heat map analysis and dynamic interactive charts to intuitively display user behavior and feedback, and combines dictionary-based sentiment analysis, machine learning classifiers and deep learning models to provide real-time sentiment analysis of user feedback;

[0047] Sp7, data collection and preprocessing module, user behavior analysis module, user portrait generation module, interactive design optimization module, real-time feedback and adjustment module and visualization and interactive feedback module all interact with each other through standardized interfaces. The master control system stores and processes data through the cloud platform, supporting large-scale user data analysis and real-time response.

[0048] The data collection and preprocessing module realizes real-time data reception and processing through multiple data sources to enhance the comprehensiveness and effectiveness of the data. The data collection and preprocessing module obtains vehicle operation data in real time through on-board sensors, where the operation data includes vehicle speed, acceleration and steering angle. When applied, the user's driving style (aggressive or smooth) is inferred through the data changes of the acceleration sensor, providing a reference for subsequent interactive design. For users with aggressive driving styles, the interactive prompts of the vehicle's dynamic response are optimized, and on-board infotainment data is integrated, including the user's frequency of use of music and navigation functions, operating habits and preference settings. At the same time, the vehicle is equipped with a telematics unit to receive vehicle remote diagnostic data and data on the user's interaction with the vehicle through mobile phone applications.

[0049] At the same time, data caching technology is used. During the peak period of data reception, data is first temporarily stored in the cache and then processed according to priority. For critical safety-related data (such as brake system fault signals), they are processed and analyzed immediately to ensure vehicle safety. For general data (such as temperature change records in the car), they can be processed when system resources are idle to improve data processing efficiency and the overall response performance of the system. Data fusion algorithms are used to integrate data from different sensors and data sources. For example, data such as vehicle speed, engine speed, and accelerator pedal opening are fused to calculate the real-time power and torque of the vehicle, so as to have a more comprehensive understanding of the vehicle's operating status and the user's driving needs.

[0050] The user behavior analysis module uses a variety of clustering and sequence analysis algorithms to form an in-depth understanding of user behavior patterns and reveal potential changes in user needs. The K-means clustering algorithm in the user behavior analysis module groups users based on their driving operation habits, that is, the frequency and amplitude of acceleration and braking. Users who frequently accelerate and brake suddenly are classified into one category, and their behavioral characteristics are analyzed. The DBSCAN (density clustering algorithm) is used to cluster behavior patterns under different road conditions and different time periods. At the same time, a hierarchical model of user behavior is constructed through a hierarchical clustering algorithm. The evolution of user behavior is analyzed from macro to micro. When applied, users are first divided into three categories according to their overall driving style: aggressive, robust and conservative. Then, each major category is further subdivided according to the frequency of turn signal use and parking habits.

[0051] The Markov chain analyzes the sequence and transition probability of user operation behaviors and combines the user's emotional state data (such as analyzing the user's facial expressions or voice intonation through the in-car camera to identify emotions) for comprehensive analysis. For example, if it is found that the user's operation behavior transition probability changes when he is tired or anxious, such as frequently switching music or frequently adjusting the air-conditioning temperature, the system can intelligently adjust the in-car environment settings, such as playing soothing music, automatically adjusting the air-conditioning temperature and wind speed to a suitable comfortable state, and provide warm voice prompts to remind users to take a rest or adjust their driving status to improve driving safety and comfort.

[0052] The user portrait generation module adopts a dynamic update mechanism to dynamically update the user portrait to ensure that it continues to reflect the user's personalized characteristics and preferences. The user portrait generation module monitors the user's online behavior in real time. When the user uses the in-vehicle navigation system to search for a new destination while driving, the system immediately records this behavior and updates the user's interest preference label according to the type of destination. It recommends nearby outdoor sports equipment stores or related activity information on the home page of the in-vehicle display screen. At the same time, the navigation system gives priority to displaying convenient routes to outdoor places, and provides detailed prompts for surrounding parking lots, rest areas and other related information. By monitoring the long-term change trend of the user's average vehicle speed, acceleration distribution and braking frequency data, habit analysis is performed. The user portrait generation module uses a decision tree algorithm to construct a user profile based on the user's age, gender, occupation, driving behavior data and the characteristics of the vehicle usage scenario. The user classification decision tree and support vector machine algorithm analyze the user's preference for vehicle appearance and interior style, collect the user's selection and evaluation data on vehicle color, seat material and dashboard style, as well as the user's likes and comments on the car's exterior design on social media, and comprehensively consider the user's consumption behavior and vehicle usage habits through the random forest classification algorithm, evaluate and classify the user's consumption potential and loyalty, analyze the user's vehicle maintenance frequency, whether the user has purchased original accessories or value-added services, and the participation in vehicle brand activities. For users with high consumption potential and high loyalty, they are marked as "high-quality customers" in the user portrait. For these users, the system provides exclusive after-sales service discounts. The user portrait generation module combines graph embedding algorithms such as Node2Vec, GraphSAGE and relational network embedding to mine social relationships and interest associations between users.

[0053] For example, for users between the ages of 30 and 40, who work as office workers and often drive during rush hour on weekdays, the decision tree may classify them as "urban commuters". For this user group, the user portrait records in detail their characteristics such as high attention to traffic information, certain requirements for vehicle comfort, and the need to conveniently answer calls and handle work-related information while driving. Based on these portrait characteristics, the system optimizes the real-time broadcast function of traffic conditions in the in-vehicle infotainment system to make it more accurate and timely; adjusts the automatic adjustment mode of seats and air conditioning to provide a more comfortable driving environment; and integrates an intelligent voice assistant to support users to answer calls, send text messages and query work-related information through voice commands, improving convenience and safety during driving.

[0054] Support vector machines can build a user aesthetic preference model. For example, if it is found that users have a high preference for simple and modern interior design and bright and vivid body colors, it will be marked as "modern and simple aesthetic preference" in the user portrait. When vehicle manufacturers launch new interior upgrade kits or special body color options, the system can push relevant information to these users in a targeted manner based on user portraits, and give priority to displaying solutions that meet their aesthetic preferences in the virtual showroom or vehicle configuration interface, thereby increasing users' satisfaction and participation in vehicle customization.

[0055] When mining the social relationships and interest associations between users, for example, if it is found that multiple users often travel together (through the vehicle's Bluetooth connection records or common travel destination analysis), and they have certain similarities in their preferences for the use of vehicle functions, the system can build a user social network graph and use the graph embedding algorithm to integrate the relationships and common interest characteristics of these users into the user portrait. Based on this social relationship and interest association, the system can launch social interaction functions, such as allowing users to share travel plans, route recommendations, and driving experience in the vehicle system; at the same time, based on the common interests of the user group, recommend relevant surrounding activities or preferential information from cooperating merchants for them, such as recommending nearby outdoor sports event tickets or related equipment discount information for user groups who like outdoor sports, further enriching the user's driving experience and life services, and enhancing the user's perception of the use value and satisfaction of the vehicle interaction system.

[0056] The interactive design optimization module adopts an intelligent integration mechanism to integrate user feedback and make intelligent adjustments to interactive elements to improve the quality of user experience. The interactive design optimization module not only considers users' direct evaluation of the vehicle's existing interactive functions, but also uses natural language processing technology to analyze users' text descriptions of their vehicle usage experience on social media, online forums and other channels.

[0057] For example, sentiment analysis algorithms are used to extract dissatisfaction with the smoothness of the vehicle's central control screen operation and specific improvement suggestions from user comments, such as "I hope the central control screen can respond faster and there will be no lag when switching menus." The system classifies and quantifies these feedback information and converts them into specific interaction design optimization indicators, such as setting targets for shortening the central control screen operation response time and increasing the menu switching frame rate, and then optimizing the screen display driver and the animation effects of the interactive interface in a targeted manner to improve the interaction smoothness of the central control screen and meet user expectations.

[0058] For intelligent adjustment of interactive elements, the module is personalized based on the user's usage scenarios and habits. For example, based on the user's habit of driving at night and feedback on the brightness of the dashboard, the system automatically adjusts the night display mode of the dashboard. At night, the overall brightness of the dashboard is reduced to avoid excessive light irritating the user's eyes. At the same time, the display contrast and font size of key information on the dashboard (such as vehicle speed, speed, fuel level, etc.) are optimized to make it clearer and easier to read.

[0059] The results of user behavior data analysis are used to optimize the interactive interface layout of the vehicle. If it is found that when users operate a certain function button, they need to frequently switch interfaces or click multiple times to complete the task, it means that the current interface layout may not be reasonable enough. For example, in the vehicle's multimedia playback interface, users often need to find and switch music playlists in multiple submenus. The system can redesign the layout of the multimedia playback interface based on this behavior analysis, set commonly used playlists as shortcut entries, and display them directly on the main interface, or use intelligent prediction algorithms to automatically recommend suitable playlists based on the user's historical playback records and current time, location and other information, and display them in a prominent position, reducing the user's operation steps and time, and improving interaction efficiency and user experience.

[0060] In terms of meeting safety goals, the interactive design optimization module is closely integrated with the vehicle's safety system. For example, when the vehicle's forward collision warning system detects a possible collision risk, it not only reminds the user through sound and visual alarms, but also optimizes the display method and content of the alarm. On the in-vehicle display screen, the source direction, distance and estimated time of the collision risk are explained with intuitive graphics and concise text, and the vehicle's braking system and throttle response are automatically adjusted to make it safer and avoid users from exacerbating dangers due to misoperation. In addition, according to the user's feedback and acceptance of different types of safety prompts, the method and intensity of safety prompts are adjusted individually. For example, for some users who are more sensitive to alarms, a milder but still effective prompt method can be provided, such as reminding users to pay attention to safety through seat vibration or slight flashing of lights. At the same time, in the vehicle settings, users are allowed to adjust the method of safety prompts according to their preferences to achieve the best balance between safety effect and user experience.

[0061] In terms of improving the comfort goal, the module focuses on the interactive design optimization of the vehicle's internal environment. According to the user's feedback on the temperature, humidity and air quality in the car, the operating parameters of the air conditioning system and the air purification system are automatically adjusted. For example, if the user manually sets the temperature in the car to a lower value during multiple driving processes, the system can infer that the user has a preference for lower temperatures. In subsequent driving, the air conditioning temperature preset value is automatically adjusted to the temperature range that the user is accustomed to, and the air outlet mode and wind speed of the air conditioner are optimized to avoid direct blowing on the user and causing discomfort. In addition, through cooperation with seat suppliers, the intelligent seat adjustment function is introduced to automatically adjust the shape and support strength of the seat according to the user's sitting posture and body pressure distribution data, providing a more comfortable driving posture and reducing fatigue from long-term driving. In addition, in terms of vehicle noise control, the user's feedback on the noise level in the car is used to optimize the vehicle's sound insulation materials and sealing design, and the active noise reduction technology is used to further reduce the noise in the car, creating a quieter and more comfortable driving environment for users.

[0062] In terms of enhancing the convenience goal, the interactive design optimization module is committed to simplifying the vehicle's operating procedures and providing convenient service access. For example, through deep integration with smartphones, the vehicle's keyless entry and start functions are seamlessly connected to the phone. Users only need to bring their phones close to the vehicle, and the system will automatically identify and unlock the door, while starting the vehicle's personalized settings, such as adjusting the seat position, playing the user's favorite music, etc. In the in-vehicle navigation system, a variety of third-party application services are integrated, such as online ordering, gas station payment, and parking reservation. Users can operate these applications directly through the in-vehicle display during driving without using a mobile phone, which improves the convenience and safety of operation. In addition, based on the user's daily travel habits and common destinations, intelligent recommendations for related services and information are made, such as automatically prompting nearby supermarket discount information or recommending dinner restaurant options when the user is approaching the end of get off work and driving towards home, providing more convenience for the user's life and improving the user's overall satisfaction and dependence on the vehicle interaction system.

[0063] In terms of meeting personalized demand goals, the module uses user portraits and user behavior data to provide each user with a customized interactive experience. For example, based on the user's music preferences, a personalized music playlist is automatically created for the user and highlighted on the homepage of the in-vehicle display. For users who like a sense of technology, a cooler vehicle information display interface and interactive driving data visualization charts are provided, such as real-time display of the vehicle's power output curve, energy recovery, etc. At the same time, users are allowed to customize the vehicle's interactive interface theme and color according to their preferences, and choose from a variety of preset themes or upload their favorite pictures as the background, so that the vehicle's interactive interface is more in line with the user's personality and aesthetics, and enhance the user's emotional connection and sense of belonging to the vehicle.

[0064] The real-time feedback and adjustment module integrates an anomaly detection mechanism to identify abnormal patterns of user behavior and generate corresponding feedback information. The anomaly detection mechanism in the real-time feedback and adjustment module adopts a statistical model-based method. For the user's driving behavior data, a statistical distribution model of normal driving behavior is established, including the mean and standard deviation of the vehicle speed, and the frequency distribution of the steering wheel rotation angle. At the same time, a machine learning algorithm is combined for anomaly detection. The machine learning algorithm chooses to use the isolation forest algorithm to divide the data by constructing a random binary tree, and isolates abnormal data points in the shallower layer of the tree. In the user behavior analysis, for the user's operation sequence and time interval data of various vehicle functions, abnormal operation behaviors that are significantly different from the normal pattern are identified.

[0065] By monitoring the user's driving data in real time, if the current vehicle speed exceeds the normal range by multiple standard deviations, or the steering wheel turns abnormally frequently and significantly in a short period of time, the system will determine that the driving behavior is abnormal. At this time, the module will immediately trigger the alarm system, send a warning message to the user, and record the abnormal data for further analysis. This statistical model-based anomaly detection method can quickly and accurately identify behaviors that clearly deviate from normal driving patterns, providing strong support for timely ensuring driving safety.

[0066] The visualization and interactive feedback module supports a variety of data visualization forms. The visualization and interactive feedback module supports interactive dashboard visualization and uses map visualization technology, combined with the in-vehicle navigation system, to display the user's driving trajectory, frequently visited places, and surrounding points of interest information in the form of a map.

[0067] The master control system combines machine learning algorithms to optimize data processing and analysis processes, and uses deep learning-based image recognition algorithms to analyze image data captured by on-board cameras, automatically identifying different driving scenarios and traffic conditions on urban roads, highways, and rural roads.

[0068] The master control system will classify and store the corresponding data, and adopt different preprocessing strategies for data in different scenarios. For data in urban road scenarios, it will focus on the density of pedestrians and vehicles and the changes in traffic lights, and perform more refined feature extraction and cleaning on these data to improve the accuracy of subsequent analysis. Through this intelligent data screening and classification, data resources can be used more effectively, unnecessary data processing workload can be reduced, and the quality and availability of data can be improved, providing more targeted data support for subsequent user behavior analysis and interaction design optimization.

[0069] In the user behavior analysis module, the reinforcement learning algorithm is used to continuously optimize the behavior analysis model. For example, through the deep reinforcement learning algorithm, the system can continuously learn and adjust its understanding and prediction of user behavior during the interaction with the user. According to the real-time feedback and behavior changes of the user, the system dynamically adjusts the parameters of the clustering algorithm and the structure of the sequence analysis model to better adapt to the personalized behavior patterns of different users. For example, when it is found that the user's driving habits have changed significantly over a period of time (such as from frequent long-distance driving to mainly short-distance driving in the city), the reinforcement learning algorithm will automatically trigger the retraining and optimization of the user behavior analysis model, adjust the clustering categories and feature weights, so that the model can more accurately capture the user's new behavior patterns and needs, thereby providing a more timely and accurate basis for interaction design optimization and improving the adaptability and personalization of the user experience.

[0070] In the interactive design optimization module, a recommendation system algorithm based on machine learning is used to achieve more accurate interactive design recommendations. For example, using collaborative filtering algorithms, according to the interactive design choices of user groups with similar behavior patterns and preferences, the appropriate interactive interface layout, function settings and operation methods are recommended for the current user. By analyzing a large number of user feedback and satisfaction data on different interactive design elements, the system can discover potential associations and preference patterns between users. When a new user uses the system, the system will recommend optimized interactive design solutions based on its initial behavioral characteristics and similarities with other users. At the same time, combined with the content recommendation algorithm, relevant functions and information are recommended based on the user's specific behavior and current situation. For example, when a user is driving on a highway, the system recommends appropriate cruise control settings, rest area information and related safety tips based on the user's driving speed, navigation destination and historical behavior, improving the pertinence and practicality of the interactive design and enhancing the convenience and satisfaction of the user's interaction with the vehicle.

[0071] Through the comprehensive application of machine learning algorithms, the system can achieve more automated and intelligent data analysis and decision-making. For example, in the real-time feedback and adjustment module, the system can automatically determine whether it is necessary to adjust the vehicle's performance parameters or provide specific prompt information based on the user's real-time driving data and behavior patterns without manual intervention. When the system detects that the user is in a state of fatigue driving, based on the judgment of the machine learning model, it automatically lowers the volume of the music in the car, adjusts the air conditioning temperature to a suitable state of wakefulness, displays a fatigue driving warning icon on the dashboard, and prompts the user to take a rest through voice. This intelligent automatic adjustment capability not only improves driving safety and comfort, but also reflects the system's real-time perception and accurate response to user needs, greatly improving the intelligence level and user experience of the entire interactive design system.

[0072] Innovatively introduce advanced machine learning technologies such as generative adversarial networks (GAN) to generate virtual user behavior data and scenarios to expand data sets and conduct simulation tests. Through GAN, the system can generate virtual user data with various characteristics and behavior patterns, which can be used to train and verify various machine learning models to improve the generalization ability and stability of the models. At the same time, using the generated virtual scene data, such as driving scenes under different traffic conditions and weather conditions, the system can simulate and evaluate the interactive design in a virtual environment, discover potential problems and optimization points in advance, without the need to conduct a large number of experiments in actual driving, saving time and cost, and also improving the innovation ability and reliability of interactive design. For example, by simulating driving scenes in bad weather, the interactive design effects of functions such as wiper control and automatic headlight adjustment of the vehicle are tested, and optimization is carried out based on the test results to ensure that users can operate these functions more conveniently and safely when encountering bad weather, further improving the intelligence level of vehicle interactive design and the ability to cope with complex scenarios.

[0073] Attachments that require explanation Figure 3 The user satisfaction comparison results can be clearly displayed, where the horizontal axis (X axis) represents the user number. The values ​​from 1 to 100 represent different individual users. Each number corresponds to a user, showing the satisfaction data of each user.

[0074] Vertical axis (Y axis): represents the satisfaction score, ranging from 0 to 1. Each user's satisfaction score is a value within this range. The closer to 1, the higher the user satisfaction, and the closer to 0, the lower the satisfaction. Specific embodiment 2:

[0076] like Figure 1-3 As shown, the key algorithm mentioned in Example 1 is analyzed in detail below, including its core mathematical formula and explanation as well as specific application description:

[0077] 1. Data collection and preprocessing module:

[0078] 1. Rule-based data cleaning:

[0079] When collecting user data, the system checks each data point. If a null value is found, the data point is deleted directly to ensure the accuracy of subsequent analysis. Invalid or incomplete user data is automatically cleared through set rules (such as time range, data type, etc.) to ensure the data quality of subsequent analysis.

[0080] 2. Anomaly detection algorithm (Z-score):

[0081] formula:

[0082] Z: standard score, X: data point to be tested, μ: mean of the data, σ: standard deviation of the data.

[0083] Application description: In the preprocessing stage, the Z-score of all data points is calculated. If the absolute value of the Z-score is greater than the set threshold, the data point is considered an outlier and processed. Statistical methods or machine learning algorithms (such as isolation forests) are used to identify and remove outliers to prevent them from affecting subsequent analysis and model training.

[0084] 3. Data Normalization (Min-Max Normalization):

[0085] formula:

[0086] X′: normalized data point, X: original data point, X min : The minimum value in the data set, X max : The maximum value in the data set.

[0087] Application Description: Reduce the dimension of data, simplify the data set by extracting the main features, improve the computational efficiency, while retaining most of the information for subsequent analysis.

[0088] 4. Principal Component Analysis (PCA)

[0089] step:

[0090] 1. Calculate the covariance matrix of the data 2. Calculate the eigenvalues ​​and eigenvectors of the covariance matrix; 3. Select the first k eigenvectors and construct the projection matrix; 4. Projection: Z = X·W, where W is the matrix composed of eigenvectors.

[0091] Application description: PCA is used to reduce the dimension of data and extract the most representative user features for subsequent analysis. It reduces the dimension of data and simplifies the data set by extracting the main features, improving computational efficiency while retaining most of the information for subsequent analysis.

[0092] 2. User behavior analysis module:

[0093] 1. K-means clustering:

[0094] Formula: Update cluster center:

[0095] C k : The center of the kth cluster, N k : The number of points in the kth cluster, S k : All points in the kth cluster.

[0096] Loss function:

[0097] Application Description: The system performs K-means clustering on user behavior data to identify different types of user groups, which is convenient for the design of subsequent personalized services. It groups users according to their behavioral characteristics (such as click-through rate, visit duration, etc.) to help identify different user groups and facilitate subsequent targeted design.

[0098] 2. DBSCAN:

[0099] Parameters: ∈\epsi lon∈ (neighborhood radius), MinPts (minimum number of points).

[0100] Application Description: Through density clustering, dense user behavior patterns and isolated user behaviors are identified to help the system understand the differences in user behavior in specific scenarios.

[0101] 3. Markov Chain:

[0102] Transition probability: P(X t+1 =j|X t =i)=p ij

[0103] X t : Current state, X t+1 : Next state, p ij : The probability of transitioning from state i to state j.

[0104] Application Description: Used to analyze the user's transition behavior between different operations, help identify the common paths users use when using the product, analyze the user's transition probability between different interaction steps, identify common behavior paths, and optimize the user interaction process.

[0105] 4. Hidden Markov Model (HMM):

[0106] Formula: P(O|λ)=∑ q P(O|q,λ)P(q|λ)

[0107] O: observation sequence, λ: model parameter, q: state sequence.

[0108] Application description: Infer the user's potential intentions and needs through HMM, predict the user's future behavior based on the observed behavior sequence, model the user's behavior sequence in a specific situation, identify potential state changes, and help understand user needs.

[0109] 3. User portrait generation module

[0110] 1. Decision Tree

[0111] Basic formula: Information gain: IG(D,A)=H(D)-H(D|A)

[0112] H(D): entropy of data set D, H(D|A): entropy under the condition of attribute A.

[0113] Application description: Analyze user characteristics through decision trees and generate user portraits, so that the system can make personalized recommendations based on user attributes. Generate a visual decision tree model based on user characteristics and behavior data to quickly identify user types and preferences.

[0114] 2. Support Vector Machine (SVM):

[0115] Formula: Optimization problem:

[0116] w: hyperplane normal vector, x i : Input features, y i : label (1 or -1), b: bias term.

[0117] Application description: Use SVM to classify user data and help generate user classification profiles to facilitate personalized services.

[0118] 3. Random Forest:

[0119] Formula: The output of each tree is:

[0120] f i (x): prediction of the i-th tree, N: number of trees.

[0121] Application description: Improve the accuracy and robustness of user portraits through the voting mechanism of multiple decision trees.

[0122] 4. Graph embedding algorithms (such as Node2Vec):

[0123] Formula: Optimization objective: max∑ u∈V Σ v∈N(u) logσ(u T v)

[0124] V: node set, N(u): neighbors of node u, σ: Sigmoid function.

[0125] Application description: Use graph embedding algorithms to map users and behaviors into low-dimensional space to generate richer personalized portraits for users.

[0126] 4. Interaction Design Optimization Module

[0127] 1. Collaborative filtering:

[0128] Formula: Prediction score:

[0129]

[0130] User u’s predicted rating for item i, N(i): a set of items similar to item i, w ij : The similarity between items i and j, r u,j : The actual rating given by user u to item j.

[0131] Application Description: Recommend appropriate interaction design solutions based on user historical behavior and ratings to improve user experience.

[0132] 2. Genetic Algorithm:

[0133] Formula: Fitness function: f(x) = Evaluate(x)

[0134] Application Description: Through evolutionary strategies, optimize multi-objective requirements in interaction design to achieve better user experience.

[0135] 3. Particle Swarm Optimization (PSO):

[0136] formula:

[0137] Update formula: v i =wv i +c1r1(pbest i -x i )+c2r2(gbest-x i )

[0138] v i : The velocity of particle i, x i : Particle position, pbest i : the best position of particle i, gbest: the global best position.

[0139] Application Description: Improve the overall effect of interactive design by optimizing design parameters.

[0140] 5. Real-time feedback and adjustment module

[0141] 1.Q-learning:

[0142] Update formula:

[0143]

[0144] Q(s,a): the value of state s and action a, r: the reward after performing the action in the current state, α: learning rate, which controls the impact of new information on old information, γ: discount factor, which evaluates the importance of future rewards, s′: the next state after performing action a.

[0145] Application Description: The system monitors user behavior in real time and updates strategies to optimize the user experience. For example, if a user performs poorly in a certain interactive design, the system can dynamically adjust the design to improve the experience.

[0146] 2. Deep Q Network (DQN)

[0147] Core formula: loss function:

[0148]

[0149] Where L: loss function value, used to measure the difference between the current Q network's prediction and the target.

[0150] E: Expected symbol, representing the average error for all possible state-action pairs.

[0151] r: The immediate reward value of the current action, which is the reward obtained by the system after taking action a in state s; γ: Discount factor, ranging from 0 to 1.

[0152] max a′ Q(s′,a′): Target Q value, which represents the maximum Q value when the optimal action a′ is selected in the next state s′, and is used to estimate future benefits.

[0153] Q(s,a): The Q value estimated by the current Q network for state s and action a.

[0154] (r+γmax a′ Q(s′,a′)-Q(s,a)) 2 : The square of the difference between the target and the prediction. The smaller this value is, the closer the prediction of the Q network is to the true optimal value.

[0155] 3. Policy Gradient Method:

[0156] Update formula:

[0157]

[0158] θ: strategy parameter, J(θ): performance indicator of the strategy.

[0159] Application description: Directly improve the user interaction experience by optimizing strategy parameters, suitable for the optimization of complex strategies.

[0160] 4. Time Series Forecasting (ARIMA)

[0161] Y t =c+φ1Y t-1 +φ2Y t-2 +θ1∈ t-1 +θ2∈ t-2 +∈ t

[0162] Y t : The observed value at the current moment, c: constant term, φ: autoregressive coefficient, θ: moving average coefficient, ∈: white noise.

[0163] Application Description: Analyze historical data of user behavior and predict future changes in user demand to adjust system responses in advance.

[0164] 5. Long Short-Term Memory Network (LSTM):

[0165] f t =σ(W f ·[h t-1 ,x t ]+b f )

[0166] Application Description: Processing time series data, capturing the long-term dependencies of user behaviors, suitable for predicting changes in user demand and making adjustments.

[0167] 6. Prophet:

[0168] Combine trend, seasonality and holiday effects to make forecasts:

[0169] y(t)=g(t)+s(t)+h(t)

[0170] g(t): trend part, s(t): seasonal part, h(t): holiday effect.

[0171] Application Description: Used to analyze and predict trend changes in user interactions, helping the system adjust the design based on the predicted results. Specific embodiment three:

[0173] like Figure 1-3 As shown below, the specific application logic steps of each module and algorithm in the automotive product interaction design system based on multi-dimensional user experience data analysis are described:

[0174] 1. Data collection and preprocessing module

[0175] Application logic steps:

[0176] Data collection: Collect user behavior data in real time from various sensors inside and outside the vehicle, user operation records, and other Internet of Vehicles data sources.

[0177] Data cleaning and anomaly detection: Use rule-based data cleaning and anomaly detection algorithms to remove redundant, inconsistent or abnormal data.

[0178] Rule-based cleaning: Define rules to remove invalid data, such as sensor error data points.

[0179] Anomaly detection: Applying statistical or machine learning methods to detect data points that deviate from normal behavior.

[0180] Data normalization: Use normalization techniques (such as Min-Max normalization) to unify the data to the same scale to reduce the impact of dimension.

[0181] Feature extraction: Use principal component analysis (PCA), independent component analysis (ICA) or feature selection algorithms to extract the most valuable features for user behavior analysis, reduce data dimensions and improve the efficiency of subsequent analysis.

[0182] 2. User behavior analysis module

[0183] Application logic steps:

[0184] User behavior clustering: Use K-means, DBSCAN, and hierarchical clustering algorithms to perform cluster analysis on user behavior and identify typical user behavior patterns.

[0185] K-means clustering: divides user data into several behavior categories by specifying the number of clusters.

[0186] DBSCAN clustering: does not specify the number of clusters and is suitable for discovering behavioral groups with different noise and density.

[0187] Hierarchical clustering: Build a hierarchical structure in user data so that different behavior patterns can be analyzed hierarchically.

[0188] Behavioral pattern mining: Analyze the temporal characteristics of user behavior using Markov chains, hidden Markov models (HMMs), and sequential pattern mining techniques.

[0189] Markov chain: Analyzes the transition probability of user behavior in different scenarios to help the system predict user behavior.

[0190] HMM: used to mine hidden user behavior patterns and reveal the underlying rules of user operations.

[0191] Sequential pattern mining: Analyze frequent sequential patterns of user behavior to provide reference for system design.

[0192] 3. User portrait generation module

[0193] Application logic steps:

[0194] User classification: Use classification algorithms such as decision trees, support vector machines (SVM), and random forests to generate user portraits and identify user characteristics.

[0195] Decision Tree: Build a rule tree to assign user characteristics based on user behavior, helping the system distinguish different types of users.

[0196] SVM: Classify user data in high-dimensional space and identify user preferences in certain behaviors.

[0197] Random Forest: Multiple decision tree models are used to vote on user classification results and generate accurate user portraits.

[0198] Graph embedding: Use graph embedding algorithms such as Node2Vec and GraphSAGE to model user relationship data and discover user group characteristics.

[0199] Node2Vec: Convert user relationships into vector representations to reveal the similarities between users.

[0200] GraphSAGE: Build user graph embeddings by combining neighbor node information to discover similar behavioral characteristics of users.

[0201] Relationship network embedding: Establish a social relationship graph, obtain the social characteristics between users through embedding algorithms, and improve the accuracy of portraits.

[0202] 4. Interaction design optimization module

[0203] Application logic steps: Based on user portraits and behavior patterns, use collaborative filtering, content recommendation, and model-based recommendation algorithms to provide personalized recommendations.

[0204] Collaborative filtering: Recommend suitable interaction design options based on the preferences of similar users.

[0205] Content recommendation: Recommend relevant content or functional modules based on user preference characteristics to improve the interactive friendliness of the system.

[0206] Model-based recommendations: Use machine learning models to predict features or interface layouts that users may be interested in.

[0207] Optimization algorithm: Design interactive schemes under multi-objective optimization through genetic algorithms, particle swarm optimization (PSO), constrained optimization and other technologies.

[0208] Genetic algorithm: Select the optimal interaction design through simulated evolution to ensure that the design meets multi-dimensional user needs.

[0209] PSO: Searches for the optimal parameter combination in the interactive design space to ensure that the design achieves the expected user experience.

[0210] Constrained Optimization: Optimize a design under certain conditions to ensure that the design meets physical or design constraints.

[0211] 5. Real-time feedback and adjustment module

[0212] Application logic steps:

[0213] Reinforcement learning: Based on real-time user feedback, Q-learning, deep Q network (DQN) and policy gradient algorithm are used to adjust system responses.

[0214] Q-learning: Update system behavior based on immediate feedback from users, allowing the system to gradually improve.

[0215] DQN: Predicting user response behavior in complex state spaces and optimizing interaction design strategies.

[0216] Policy Gradient Algorithm: Dynamically adjust the strategy according to the user's preference to improve the user's interactive experience.

[0217] Time series forecasting: Combine ARIMA, LSTM, Prophet and other time series models to predict future changes in user demand and make system responses in advance.

[0218] ARIMA: Analyze historical user behavior data and predict short-term user behavior trends.

[0219] LSTM: Processes long-term dependent behavioral data and predicts changes in user demand in complex scenarios.

[0220] Prophet: Combines seasonal variation characteristics to predict user interaction needs at different time points.

[0221] All modules in this system interact with each other through standardized interfaces. The master control system realizes data storage and processing through the cloud platform, and supports the system's analysis and real-time response to large-scale user data. Specific embodiment four:

[0223] like Figure 1-3 As shown in the figure, the core module hardware composition of the master control system is as follows:

[0224] 1. Data collection and preprocessing module hardware composition

[0225] Multimodal sensor array: including driver biometric sensors (such as heart rate sensors, skin conductivity sensors) and environmental sensors (such as temperature, humidity, light intensity, etc.). These sensors can collect data on the user's physiological state and surrounding environment in the vehicle in real time, ensuring that different user behaviors and scenarios can be covered.

[0226] Camera system: mainly includes the built-in driver facial camera and the external driving recording camera, which helps capture the user's facial expressions, head posture and driving behavior. The built-in camera is responsible for identifying the driver's gaze direction, emotional changes and other information, and the external camera can record external road conditions.

[0227] Driving behavior detector: including steering wheel pressure sensor, accelerator and brake pedal sensors, steering angle sensor and vehicle speed sensor. These hardware can capture the user's driving operation details and provide basic data for user behavior analysis.

[0228] Data acquisition terminal: used to collect data from sensors and cameras in real time. The acquisition terminal usually includes a microcontroller or embedded computing module (such as Raspberry Pi, NVIDIA Jetson, etc.), which is responsible for integrating and transmitting various types of data to the data processing module.

[0229] 2. Hardware composition of the user behavior analysis module: (1) High-performance computing server: used to process massive user data and support complex clustering analysis and time series pattern mining. The server can use high-performance hardware with multi-core processors and high memory. Typical configurations include Intel Xeon processors, NVIDIA Tesla GPUs, etc., which can be deployed in the cloud or locally to achieve fast data processing and behavior pattern recognition. (2) FPGA or ASIC hardware accelerator: supports large-scale matrix operations and parallel processing, and is used to execute clustering algorithms and computationally intensive algorithms such as Markov chains and hidden Markov models. FPGA hardware is suitable for accelerating matrix calculations, while ASIC can be used for specific deep learning model inference tasks to improve the computational efficiency of the model. (3) Embedded AI chip: Embedded chips with built-in AI accelerators (such as Google Coral TPU, Edge TPU) can be applied to some real-time behavior analysis needs, and can realize edge computing in the car, reduce data latency, and reduce dependence on cloud computing.

[0230] 3. Hardware composition of user portrait generation module

[0231] Graph processing server: This module requires a high-performance graph computing server, which is mainly used to execute graph embedding algorithms (such as Node2Vec, GraphSAGE, etc.) and process user relationship graphs and behavior data. The graph processing server can use a CPU with large memory capacity and a multi-GPU architecture. The typical hardware configuration includes a high-end NVIDIAA100 GPU, which is suitable for processing large-scale graph data.

[0232] Storage system: used to store user characteristics, behavior data and generated user profile information. The storage system uses high-performance SSD arrays or distributed storage (such as Hadoop distributed file system) to ensure that user data can be quickly read and written under large data volumes.

[0233] User feature database: manages and stores user portraits based on NoSQL databases (such as MongoDB, Cassandra, etc.), suitable for large-scale and diverse data storage needs, and provides fast query capabilities and data management.

[0234] 4. Interactive design optimization module hardware composition

[0235] Recommendation algorithm server: used to run collaborative filtering, content recommendation, and model-based recommendation algorithms. This server needs to have efficient parallel processing capabilities. Typical hardware options include multi-core CPUs and high-performance GPUs to support training and reasoning operations of recommendation models.

[0236] Multi-objective optimization processor: used to execute optimization algorithms such as genetic algorithms and particle swarm optimization (PSO). This module can handle multi-objective optimization through FPGA or GPU acceleration hardware to ensure the best results in interactive design. The main task of this processor is to accelerate complex optimization calculations so that it can adapt to user needs in real time.

[0237] Touch screen and display: As the terminal interface for system and user interaction, the central control screen with high-resolution touch display is an essential hardware. The screen provides the user with a selection interface and displays recommended content or personalized interactive design options. The screen is usually configured in the center console for easy user operation and feedback.

[0238] 5. Real-time feedback and adjustment module hardware composition

[0239] Real-time data feedback system: Combining the real-time data acquisition hardware and wireless communication modules (such as 5G modules) in the car can ensure the rapid transmission of data and facilitate the implementation of real-time feedback mechanism.

[0240] Deep learning acceleration hardware: used for deep learning inference processing of DQN and policy gradient algorithms. NVIDIA DriveAGX or similar deep learning accelerators can accelerate the real-time operation of reinforcement learning algorithms in the car, enabling dynamic prediction and rapid response to user needs.

[0241] Time series data processor: used for prediction needs of time series models such as ARIMA and LSTM. The time series data processor can choose edge computing devices with embedded AI chips, or implement customized acceleration based on FPGA to ensure efficient computing capabilities during the time series prediction process.

[0242] Communication module: including Wi-Fi, 4G / 5G and Bluetooth modules, used to communicate with cloud platforms or other devices to ensure fast data transmission and support remote control. The communication module can transmit real-time feedback from users to the cloud analysis system or control system for timely adjustment.

[0243] 6. Hardware composition of visualization and interactive feedback module

[0244] Data processing and rendering server: used to generate visualization content such as dynamic interactive charts and heat map analysis. Typical configurations include high-performance computing servers with multiple GPUs, which are suitable for real-time processing and rendering of large amounts of user data.

[0245] Touchscreen: A high-resolution display in the car, usually a touchscreen, used to display visual data of user behavior, heat maps, user feedback, etc. The screen should have anti-reflective properties and high brightness to ensure visibility under different lighting conditions.

[0246] Sentiment analysis acceleration module: Process the emotional content in user feedback through a sentiment analysis accelerator (such as a dedicated NLP chip or TPU module). This module can quickly analyze the emotional state in user feedback, thereby dynamically adjusting the user experience. You can use a GPU or Edge TPU module that supports sentiment analysis to analyze the sentiment of user text feedback in real time.

[0247] Audio output and input devices: used to receive user voice feedback and provide voice prompts. They include microphone and speaker systems, which can be used to collect user voice feedback and provide user sentiment analysis results or system interaction prompts through voice feedback.

[0248] Hardware composition of the master control system

[0249] Cloud platform server cluster: Based on cloud computing architecture, it provides data storage, processing and model training resources. Cloud platform server clusters can be configured with multiple high-performance computing servers and distributed storage devices (such as Amazon AWS, Google Cloud or Microsoft Azure) to support efficient data management and model training.

[0250] Data storage and backup system: used for secure storage and backup of all data in the system, including distributed storage devices and backup servers. Distributed storage devices can be based on Hadoop HDFS or AWS S3 to achieve long-term storage and backup of large-scale data.

[0251] Network communication module: ensures smooth data transmission between modules of the system. The communication module includes Ethernet, 5G base station or optical fiber network equipment to quickly transmit data between the cloud platform, vehicles and user terminals.

[0252] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "comprising a reference structure" do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0253] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An automotive product interaction design system based on multi-dimensional user experience data analysis, including a master control system, characterized in that: The master control system includes a data collection and preprocessing module, a user behavior analysis module, a user portrait generation module, an interactive design optimization module, a real-time feedback and adjustment module, and a visualization and interactive feedback module; Sp1, the data collection and preprocessing module cleans and preprocesses user data through rule-based data cleaning, anomaly detection algorithm and data normalization technology, and extracts user features through principal component analysis, independent component analysis and feature selection technology to provide high-quality data for subsequent analysis; Sp2, the user behavior analysis module uses K-means clustering, DBSCAN and hierarchical clustering algorithms to perform cluster analysis on user behavior, and combines Markov chain, hidden Markov model and sequence pattern mining technology to deeply explore the user's behavior patterns in different situations, so as to identify user needs; Sp3, the user portrait generation module uses decision tree, support vector machine and random forest classification algorithm to generate user portraits, and combines Node2Vec, GraphSAGE and relational network embedding graph embedding algorithm to provide users with comprehensive personalized portraits; Sp4, the interaction design optimization module optimizes the interaction design through collaborative filtering, content recommendation and model-based recommendation algorithms, and applies genetic algorithms, particle swarm optimization and constraint optimization techniques to improve user experience on the basis of meeting multi-objective requirements; Sp5, the real-time feedback and adjustment module uses Q-learning, deep Q network and policy gradient method reinforcement learning algorithms to dynamically adjust the system response according to the user's real-time data, and combines ARIMA, LSTM and Prophet, time series prediction algorithms to identify changes in user needs in advance; Sp6, the visualization and interactive feedback module generates data visualization technology through data aggregation, heat map analysis and dynamic interactive charts to intuitively display user behavior and feedback, and combines dictionary-based sentiment analysis, machine learning classifiers and deep learning models to provide real-time sentiment analysis of user feedback; Sp7. The data collection and preprocessing module, user behavior analysis module, user portrait generation module, interactive design optimization module, real-time feedback and adjustment module, and visualization and interactive feedback module all interact with each other through standardized interfaces. The master control system stores and processes data through a cloud platform, supporting large-scale user data analysis and real-time response.

2. The automobile product interaction design system based on multi-dimensional user experience data analysis according to claim 1, characterized in that: The data collection and preprocessing module realizes real-time data reception and processing through multiple data sources to enhance the comprehensiveness and effectiveness of the data. The data collection and preprocessing module obtains vehicle operation data in real time through vehicle-mounted sensors, wherein the operation data includes vehicle speed, acceleration and steering angle. When applied, the user's driving style (aggressive or smooth) is inferred through the data changes of the acceleration sensor, providing a reference for subsequent interactive design. For users with aggressive driving styles, the interactive prompts of vehicle power response are optimized, and the vehicle-mounted infotainment data is integrated, including the user's frequency of use of music and navigation functions, operating habits and preference settings. At the same time, the vehicle telematics processing unit is equipped with Through the (Telematics) connection, the vehicle remote diagnosis data and the data of the user's interaction with the vehicle through the mobile phone application are received. When the vehicle fails, the user's behavior record of querying the fault information through the mobile phone application is used to optimize the vehicle fault prompt and diagnosis interaction process. The data collection and preprocessing module adopts streaming data processing technology when receiving data in real time to process the data in a timely manner. When the data enters the system, it immediately performs preliminary screening and classification, allocates different types of data to corresponding processing queues, and gives priority to abnormal detection of vehicle operation data directly related to driving safety, so as to ensure that safety hazards during vehicle driving can be discovered and handled in a timely manner.

3. The automobile product interaction design system based on multi-dimensional user experience data analysis according to claims 1-3, characterized in that: The user behavior analysis module adopts a variety of clustering and sequence analysis algorithms to form an in-depth understanding of user behavior patterns and reveal potential changes in user needs. The K-means clustering algorithm in the user behavior analysis module groups users based on their driving operation habits, that is, acceleration and braking frequency and amplitude, and classifies users who frequently accelerate and brake suddenly into one category, analyzes their behavioral characteristics, and uses DBSCAN (density clustering algorithm) to cluster behavior patterns under different road conditions and different time periods. At the same time, a hierarchical model of user behavior is constructed through a hierarchical clustering algorithm to analyze the evolution of user behavior from macro to micro. When applied, users are first divided into three categories according to their overall driving style: aggressive, robust and conservative, and then each major category is further subdivided according to the frequency of turn signal use and parking habits.

4. The automobile product interaction design system based on multi-dimensional user experience data analysis according to claims 1-3, characterized in that: The user portrait generation module adopts a dynamic update mechanism to dynamically update the user portrait to ensure that it continues to reflect the user's personalized characteristics and preferences. The user portrait generation module monitors the user's online behavior in real time. When the user uses the in-vehicle navigation system to search for a new destination while driving, the system immediately records this behavior and updates the user's interest preference tag according to the type of destination. It recommends nearby outdoor sports equipment stores or related activity information on the homepage of the in-vehicle display screen. At the same time, the navigation system gives priority to displaying convenient routes to outdoor places, and provides detailed prompts for related information such as surrounding parking lots and rest areas. Habit analysis is performed by monitoring the long-term change trend of the user's average vehicle speed, acceleration distribution, and braking frequency data. The user portrait generation module uses a decision tree algorithm to construct a user portrait based on the user's age, gender, occupation, driving behavior data, and characteristics of the vehicle usage scenario. User classification decision tree, the support vector machine algorithm analyzes the user's preference for vehicle appearance and interior style, collects user selection and evaluation data on vehicle color, seat material and dashboard style, as well as user likes and comments on car appearance design on social media, and comprehensively considers the user's consumption behavior and vehicle usage habits through the random forest classification algorithm, evaluates and classifies the user's consumption potential and loyalty, analyzes the user's vehicle maintenance frequency, whether the user has purchased original accessories or value-added services, and the degree of participation in vehicle brand activities. For users with high consumption potential and high loyalty, they are marked as "high-quality customers" in the user portrait. For these users, the system provides exclusive after-sales service discounts. The user portrait generation module combines graph embedding algorithms such as Node2Vec, GraphSAGE and relational network embedding to mine social relationships and interest associations between users.

5. The automobile product interaction design system based on multi-dimensional user experience data analysis according to claims 1-4, characterized in that: The interaction design optimization module adopts an intelligent integration mechanism to integrate user feedback and make intelligent adjustments to interactive elements to improve the quality of user experience. The interaction design optimization module not only considers users' direct evaluation of the vehicle's existing interactive functions, but also uses natural language processing technology to analyze users' text descriptions of their vehicle usage experience on social media, online forums and other channels.

6. The automobile product interaction design system based on multi-dimensional user experience data analysis according to claims 1-5, characterized in that: The real-time feedback and adjustment module integrates an abnormality detection mechanism to identify abnormal patterns of user behavior and generate corresponding feedback information. The abnormality detection mechanism in the real-time feedback and adjustment module adopts a statistical model-based method. For the user's driving behavior data, a statistical distribution model of normal driving behavior is established, including the mean and standard deviation of the vehicle speed, and the frequency distribution of the steering wheel rotation angle. At the same time, anomaly detection is performed in combination with a machine learning algorithm. The machine learning algorithm chooses to use an isolation forest algorithm to divide the data by constructing a random binary tree, and isolates abnormal data points in the shallower layer of the tree. In the user behavior analysis, for the user's operation sequence and time interval data of various vehicle functions, abnormal operation behaviors that are significantly different from the normal pattern are identified.

7. The automobile product interaction design system based on multi-dimensional user experience data analysis according to claims 1-6, characterized in that: The visualization and interactive feedback module supports various data visualization forms. The visualization and interactive feedback module supports interactive dashboard visualization and adopts map visualization technology, combined with the vehicle navigation system, to display the user's driving trajectory, frequently visited places and surrounding points of interest information in the form of a map.

8. The automobile product interaction design system based on multi-dimensional user experience data analysis according to claims 1-7, characterized in that: The master control system combines machine learning algorithms to optimize data processing and analysis processes, uses deep learning-based image recognition algorithms to analyze image data captured by on-board cameras, and automatically identifies different driving scenarios and traffic conditions on urban roads, highways, and rural roads.

9. The automobile product interaction design system based on multi-dimensional user experience data analysis according to claims 1-8, characterized in that: The master control system has flexible expansion capabilities and can quickly integrate new data sources and analysis functions according to new needs. The master control system adopts a microservice architecture, splits the data collection and preprocessing module, user behavior analysis module, user portrait generation module, interactive design optimization module, real-time feedback and adjustment module, and visualization and interactive feedback module into independent microservices, and designs a plug-in expansion mechanism to allow third-party developers or suppliers to add new functions and analysis capabilities to the system by developing plug-ins.

10. The automobile product interaction design system based on multi-dimensional user experience data analysis according to claims 1-9, characterized in that: The data collection and preprocessing module, user behavior analysis module, user portrait generation module, interaction design optimization module, real-time feedback and adjustment module, and visualization and interaction feedback module all interact with each other through standardized interfaces, using a Web service interface based on a RESTful architecture.

Images