Intelligent cockpit personalized recommendation method combining environmental perception and passenger preference

Abstract

The invention discloses an intelligent cabin personalized recommendation method combining environmental perception and passenger preference, and relates to the technical field of intelligent automobiles. Comprising the following steps: data acquisition: based on environment feature big data, vehicle intelligent cabin passenger preference data and historical recommendation schemes, collecting environment data, passenger preference data and historical recommendation schemes to obtain environment data items, passenger preference data items and historical recommendation scheme data items; and data analysis: based on the environment data item, the passenger preference data item and the historical recommendation scheme data item, analyzing the environment data item, the passenger preference data item and the historical recommendation scheme data item to obtain an analysis result. Through environment feature recognition and emotion preference recognition, the current driving scene of the passenger can be accurately recognized, various factors such as the vehicle speed, the road condition information, the weather condition and the geographic position are considered, and driving scene recognition is more accurate and comprehensive.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent vehicles, and specifically to a personalized recommendation method for an intelligent cockpit that combines environmental perception and occupant preferences. Background Art

[0002] An intelligent cockpit refers to the interior space of a vehicle that integrates advanced information technology, artificial intelligence, Internet of Things and other technologies, aiming to provide occupants with a more personalized, comfortable, safe and convenient driving experience. It can not only automatically adjust the interior environment according to the preferences of the occupants, such as temperature, music, seat angle, etc., but also through environmental perception technology, analyze and adapt to external weather, road conditions and other conditions in real time to ensure driving safety and comfort. The intelligent cockpit continuously optimizes the recommended content by deeply learning the behavior habits and preferences of the occupants, making the driving experience more considerate and intelligent.

[0003] For example, a personalized intelligent recommendation system and its recommendation method with the patent publication number CN114372164A includes a picture information management module, a keyword search module and a push management module; the picture information management module is used to obtain the user's picture viewing permission and automatically extract the picture information when the user opens the picture; the keyword search module is used to identify the picture content according to the picture information and convert it into corresponding recommended information keywords; the push management module is used to retrieve push data according to the queried recommended information keywords and perform timed push on the user. The present invention automatically identifies the image content in the picture information when the user views the picture information, thereby finding the corresponding push keywords according to the image content, and then finding the push data according to the push keywords to complete the intelligent push work for the user.

[0004] When performing intelligent recommendations for users, since the preferences of users are complex and changeable and are affected by various factors, such as personal emotions, social environment, etc., the existing recommendation algorithms may not be able to comprehensively and accurately capture the preference changes of users, resulting in the recommended information and the adjustment of the interior environment not meeting the actual needs of users. At the same time, different driving scenarios have different requirements for the interior environment. In the long-distance driving scenario, users need a comfortable and relaxing environment and may pay more attention to relaxing music, appropriate temperature and comfortable seat adjustment to relieve driving fatigue; while in short-distance driving, especially in traffic jams, users may need timely and accurate road condition information, safety tips and convenient operation experience. In addition, in bad weather (such as heavy rain, heavy snow), the comfort and safety requirements for the interior environment of users will also change. However, the existing intelligent cockpit systems lack the intelligent recognition ability for different scenarios and cannot automatically adjust the interior environment and push appropriate information according to the specific driving scenario. Therefore, in view of this situation, we propose a more convenient and practical recommendation method to meet the usage requirements. Summary of the Invention

[0005] The object of the present invention is to provide an intelligent cockpit personalized recommendation method combining environmental perception and occupant preferences to solve the problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions: an intelligent cockpit personalized recommendation method combining environmental perception and occupant preferences, including:

[0007] Data collection, based on big data of environmental characteristics, vehicle intelligent cockpit occupant preference data, and historical recommendation schemes, collect environmental data, occupant preference data, and historical recommendation schemes to obtain environmental data items, occupant preference data items, and historical recommendation scheme data items;

[0008] Data analysis, based on environmental data items, occupant preference data items, and historical recommendation scheme data items, analyze the environmental data items, occupant preference data items, and historical recommendation scheme data items to obtain an analysis result;

[0009] Personalized recommendation generation, based on the analysis result, match the environmental feature analysis result with the occupant preference analysis result to generate a personalized recommendation scheme and obtain an initial recommendation result;

[0010] Scheme adjustment, based on the initial recommendation result, allow the occupant to adjust the scheme parameters through the in-vehicle infotainment system to obtain a recommendation result;

[0011] Effect evaluation, based on the recommendation result, implement the adjusted recommendation result in the intelligent cockpit system, and evaluate the recommendation scheme through the occupant's satisfaction feedback on the recommendation result and the analysis of the occupant's behavior to obtain an evaluation result;

[0012] The personalized recommendation generation method includes;

[0013] A1. Environmental recognition, based on the environmental feature analysis result, recognize the occupant's current driving scenario to obtain a driving scenario recognition result;

[0014] A2. Emotional preference recognition, based on the occupant preference analysis result, recognize the occupant's current behavior state and emotion, and compare them with historical behavior state and emotion data to obtain an occupant emotion result;

[0015] A3. Result matching, based on the driving scenario recognition result and the occupant emotion result, perform matching to obtain a personalized recommendation result;

[0016] A4. Result output, based on the personalized recommendation result, and display it to the occupant through the interface and voice of the intelligent cockpit system.

[0017] Furthermore, the environmental feature recognition method includes the following steps;

[0018] B1. Data integration: Based on in-vehicle sensors and external data sources, obtain various data information, including vehicle speed, road conditions, weather conditions, and geographical location, to obtain a data integration set;

[0019] B2. Environmental feature extraction: Based on the data integration set, extract key features from the data integration set to obtain a feature data set;

[0020] B3. Driving scenario recognition: Based on the feature data set, establish a driving scenario classification model to obtain an initial driving scenario recognition result;

[0021] B4. Result verification: Based on the initial driving scenario recognition result, use a verification method to verify the initial environment recognition result to obtain a driving scenario recognition result.

[0022] Furthermore, the emotion preference recognition method includes the following steps:

[0023] C1. Behavior state and emotion recognition: Based on the analysis result of occupant preferences, use an analysis method to analyze the current behavior state and emotion of the occupant to obtain initial emotion preference data;

[0024] C2. Emotion classification: Based on the initial emotion preference data, classify the initial emotion preference data to obtain an emotion classification result;

[0025] C3. Data comparison: Based on the emotion classification result, compare the emotion classification result, historical behavior state, and emotion data to obtain the occupant emotion result.

[0026] Furthermore, the personalized recommendation result includes multiple personalized recommendation results, and the method for obtaining the personalized recommendation result includes:

[0027] D1. Establish a solution library: The solution library contains personalized recommendation solutions under different combinations of driving scenarios and occupant emotions;

[0028] D2. Real-time matching: Based on the driving scenario recognition result and the occupant emotion result, perform matching within the solution library to obtain multiple personalized recommendation solutions;

[0029] D3. Recommendation result screening and ranking: Based on multiple personalized recommendation solutions, remove unreasonable and conflicting solutions, and rank the multiple personalized recommendation solutions according to the highest similarity through a similarity calculation method to obtain a personalized recommendation solution.

[0030] Furthermore, the solution adjustment includes the following steps:

[0031] E1. Parameter adjustment: Based on the initial recommendation result, generate adjustable parameter options through the in-vehicle infotainment system to obtain a parameter adjustment plan;

[0032] E2, Occupant Interaction Operation: Based on the parameter adjustment plan, modify the recommended parameters through the in-vehicle interaction system to obtain the initial recommended result;

[0033] E3, Conflict Detection and Correction: Based on the initial recommended result, perform logical verification on the parameters adjusted by the occupant to obtain the verification result;

[0034] E4, Result Confirmation: Based on the verification result, generate the recommended result by combining the parameters confirmed by the occupant.

[0035] Furthermore, the effect evaluation includes the following steps:

[0036] F1, Satisfaction Rating: Collect the satisfaction of the occupant with the recommended result through the in-vehicle questionnaire and voice feedback to obtain the satisfaction result;

[0037] F2, Behavior Analysis: Based on the satisfaction result, analyze the behavior changes of the occupant after the implementation of the recommended result to obtain the behavior analysis result;

[0038] F3, Comprehensive Evaluation: Based on the satisfaction result and the behavior analysis result, conduct a comprehensive evaluation of the satisfaction result and the behavior analysis result to obtain the initial evaluation result;

[0039] F4, Feedback Optimization: Based on the initial evaluation result, make adjustments and optimizations to obtain the evaluation result.

[0040] Furthermore, the verification method includes the following steps:

[0041] G1, Data Cross-Verification: Compare the initial driving scenario recognition result with similar data from different data sources and sensors to obtain the initial verification result;

[0042] G2, Data Comparison Verification: Based on the initial verification result, compare the initial driving scenario recognition result with the initial verification result to obtain the verification result;

[0043] G3, Real-Time Feedback Verification: Based on the verification result, provide real-time feedback through the in-vehicle interaction system and adjust the initial driving scenario recognition result to obtain the driving scenario recognition result.

[0044] Furthermore, the analysis method includes the following steps:

[0045] H1, Behavior Pattern Recognition: Based on the occupant preference analysis result, analyze the historical behavior data of the occupant to obtain the behavior pattern data;

[0046] H2, Emotional Feature Extraction: From the behavior pattern data, extract the emotional features of the member to obtain the emotional feature result;

[0047] H3. Association analysis: Based on behavioral pattern data, analyze environmental characteristics and historical behavioral pattern data to obtain initial emotional preference data.

[0048] Furthermore, the similarity calculation method includes the following steps:

[0049] M1. Feature extraction: Based on multiple personalized recommendation schemes, extract the feature vectors of each scheme to obtain a feature vector set;

[0050] M2. Similarity calculation: Based on the feature vector set, use the similarity calculation method to calculate the similarity between each feature vector to obtain a similarity matrix;

[0051] M3. Sorting and selection: Based on the similarity matrix, sort multiple personalized recommendation schemes in descending order of similarity to obtain personalized recommendation schemes.

[0052] Compared with the prior art, the beneficial effects of the present invention are:

[0053] This intelligent cockpit personalized recommendation method combining environmental perception and occupant preferences can accurately identify the current driving scenario of the occupant through environmental feature recognition and emotional preference recognition. Considering various factors such as vehicle speed, road conditions, weather conditions, and geographical location, it makes the driving scenario recognition more accurate and comprehensive. It can also carefully identify the current behavior state and emotion of the occupant and compare them with historical data, so as to more accurately grasp the emotional preferences of the occupant, make the recommendation scheme more in line with the real-time needs of the occupant, and obtain multiple personalized recommendation schemes through real-time matching, increasing the diversity and flexibility of the recommendation, and being able to meet the needs of different occupants in different scenarios, with strong practicality.

[0054] At the same time, through the in-vehicle infotainment system, the occupant can adjust the scheme parameters, so that the recommendation scheme can be dynamically adjusted according to the personalized needs of the occupant, improving the pertinence and satisfaction of the recommendation, and performing logical verification on the adjusted parameters of the occupant to ensure that the adjusted scheme is reasonable and feasible, avoiding conflicting and unreasonable recommendation results, and further improving the quality of the recommendation. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 is a flowchart of the method of the present invention;

[0056] Figure 2 is a schematic diagram of the principle of the personalized recommendation scheme of the present invention;

[0057] Figure 3 is a schematic diagram of the principle of environmental recognition of the present invention;

[0058] Figure 4 is a schematic diagram of the principle of emotional preference recognition of the present invention;

[0059] Figure 5 Schematic diagram of the principle of the similarity calculation method of the present invention;

[0060] Figure 6 Schematic diagram of the principle of the recommendation method of the present invention. Detailed implementation manners

[0061] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0062] During the use of an automobile, a recommendation solution needs to be used. The recommendation solution provided by the present invention is specifically used for the intelligent cockpit preference recommendation method during the use of an automobile. When using this recommendation method, it should be noted to ensure that the big data of environmental characteristics, the data of occupant preferences, and the data of historical recommendation solutions collected are accurate, complete, and updated in a timely manner. Since it involves a large amount of personal information and behavior data of the occupants, relevant laws and regulations must be strictly observed, and measures such as encrypted storage and access control should be taken to prevent data leakage and abuse. The recommendation solution should be appropriate to avoid causing information overload to the occupants or interfering with driving safety. For example, during driving, the recommended information should be presented in a concise and clear manner without affecting the occupants' attention to the road conditions.

[0063] Such as Figures 1-6As shown in the figure, the present invention provides a technical solution: an intelligent cockpit personalized recommendation method combining environmental perception and occupant preferences, including data collection, based on big data of environmental characteristics, vehicle intelligent cockpit occupant preference data, and historical recommendation schemes, collecting environmental data, occupant preference data, and historical recommendation schemes, to obtain environmental data items, occupant preference data items, and historical recommendation scheme data items; data analysis, based on the environmental data items, occupant preference data items, and historical recommendation scheme data items, analyzing the environmental data items, occupant preference data items, and historical recommendation scheme data items to obtain an analysis result; personalized recommendation generation, based on the analysis result, matching the environmental feature analysis result with the occupant preference analysis result to generate a personalized recommendation scheme and obtain an initial recommendation result; scheme adjustment, based on the initial recommendation result, allowing the occupant to adjust the scheme parameters through the in-vehicle infotainment system to obtain a recommendation result; effect evaluation, based on the recommendation result, implementing the adjusted recommendation result in the intelligent cockpit system, and evaluating the recommendation scheme through the occupant's satisfaction feedback on the recommendation result and the analysis of the occupant's behavior to obtain an evaluation result; the personalized recommendation generation method includes; environmental recognition, based on the environmental feature analysis result, identifying the occupant's current driving scenario to obtain a driving scenario recognition result; emotion preference recognition, based on the occupant preference analysis result, identifying the occupant's current behavior state and emotion, and comparing them with historical behavior state and emotion data to obtain an occupant emotion result; result matching, based on the driving scenario recognition result and the occupant emotion result, performing matching to obtain a personalized recommendation result; result output, based on the personalized recommendation result, and presenting it to the occupant through the interface and voice of the intelligent cockpit system.

[0064] It should be noted that data collection mainly collects and processes data through in-vehicle sensors, external data sources, occupant interaction history, in-vehicle infotainment system logs, and data clerks of different users, extracts the occupant's static preferences (such as preset seat and air-conditioning settings) and dynamic behavior data (such as music playback frequency and interaction response time), and analyzes them. At the same time, query the schemes in the scheme library that match the driving scenario and emotion, output multiple candidate schemes, eliminate unreasonable combinations, and preferentially recommend the scheme that is closest to the occupant's emotion preference.

[0065] Such as Figure 3As shown in the figure, the environmental feature recognition method includes the following steps: data integration, based on in-vehicle sensors and external data sources, obtaining various data information, including vehicle speed, road conditions information, weather conditions, and geographical location, to obtain a data integration set; environmental feature extraction, based on the data integration set, extracting key features in the data integration set to obtain a feature data set; driving scenario recognition; based on the feature data set, establishing a driving scenario classification model to obtain an initial driving scenario recognition result; result verification, based on the initial driving scenario recognition result, using a verification method to verify the initial environmental recognition result to obtain a driving scenario recognition result, and the verification method includes the following steps: data cross-verification, comparing the initial driving scenario recognition result with similar data from different data sources and sensors to obtain an initial verification result; data comparison verification, based on the initial verification result, comparing the initial driving scenario recognition result with the initial verification result to obtain a verification result; real-time feedback verification, based on the verification result, providing real-time feedback through the in-vehicle interaction system and adjusting the initial driving scenario recognition result to obtain a driving scenario recognition result.

[0066] It should be noted that first, vehicle speed, acceleration, gyroscope data, radar and lidar point clouds, camera images, etc. are collected in real time through in-vehicle sensors, and real-time weather, map data, traffic flow information, etc. are obtained through the API and preprocessed to extract features such as road curvature, following distance, weather impact, traffic sign recognition, etc. to form a feature data set. An integrated model is selected, a driving scenario classification model is established, and labeled multi-scenario data sets, such as road conditions information like city, highway, bad weather, construction area, emergency vehicle, and congestion information, etc. are imported into the model for training. The model infers and outputs the scenario category, compares the initial result with external data or historical similar scenarios, calculates the consistency score. If the confidence level is lower than the threshold and there is an external data conflict, it is marked as "to be verified", and the driver is confirmed about the suspicious scenario through the in-vehicle interaction system. If the driver feedbacks an error, the corrected data is added to the training set to trigger online learning to update the model parameters to ensure the correct judgment of the current driving scenario.

[0067] Such as Figure 4As shown, the emotional preference recognition method includes the following steps: behavior state and emotion recognition, based on the analysis results of occupant preferences, using an analysis method to analyze the current behavior state and emotion of the occupant to obtain initial emotional preference data; emotion classification, based on the initial emotional preference data, classifying the initial emotional preference data to obtain an emotion classification result; data comparison, based on the emotion classification result, comparing the emotion classification result, historical behavior state, and emotion data to obtain the occupant emotion result. The analysis method includes the following steps: behavior pattern recognition, based on the analysis results of occupant preferences, analyzing the historical behavior data of the occupant to obtain behavior pattern data; emotion feature extraction, from the behavior pattern data, extracting the emotion features of the member to obtain an emotion feature result; association analysis, based on the behavior pattern data, analyzing the environmental features and historical behavior pattern data to obtain the initial emotional preference data.

[0068] It should be noted that first, analyze the historical behavior data of the occupant, such as seat adjustment frequency, operating habits, etc., extract stable behavior pattern data, and extract emotion-related features from the behavior pattern data, such as "a faster speech rate may correspond to anxiety". Then classify the emotions and label them. The labels include excitement, calm, anxiety, etc. Compare the current emotion classification result with the historical data, identify pattern changes, and judge the emotion trend, such as a short-term increase in stress. Finally, output the occupant emotion result.

[0069] The personalized recommendation results include multiple personalized recommendation results. The method for obtaining the personalized recommendation results includes: establishing a solution library, which contains personalized recommendation solutions under different driving scenarios and occupant emotion combinations; real-time matching, based on the driving scenario recognition result and the occupant emotion result, performing matching within the solution library to obtain multiple personalized recommendation solutions; screening and ranking of recommendation results, based on multiple personalized recommendation solutions, removing unreasonable and conflicting solutions, and ranking the multiple personalized recommendation solutions according to the highest similarity through a similarity calculation method to obtain the personalized recommendation solution.

[0070] It should be noted that driving scenario dimensions are first defined, such as urban congestion, highway cruising, night driving, rainy-day driving, sharp turns, etc. The system then identifies the scenarios in real time through in-vehicle sensors or external data, and tags each scenario. At the same time, it judges the emotional state through in-vehicle cameras, voice analysis, physiological sensors, etc. It designs recommended solutions for each combination of driving scenario and emotion. The solution library is stored as a structured database, with fields including: scenario tag, emotion tag, recommended content, priority weight, and conflict rules. It identifies driving scenario and emotion preferences, and first looks for solutions that exactly match the current scenario and current emotion. If the matching results are insufficient, the conditions are relaxed, and candidate solutions are output at the same time. It checks for contradictions between the solutions and eliminates solutions that conflict with the current vehicle state. It calculates the similarity between the current scenario and emotion and historical preferences, and outputs them in descending order of similarity. Through accurate environmental and emotion recognition results, it matches in a rich solution library, filters and sorts to obtain personalized recommended solutions, removes unreasonable and conflicting solutions, and provides the most suitable recommendations for the occupants according to their current needs.

[0071] The solution adjustment includes the following steps: Parameter adjustment: Based on the initial recommended result, generate adjustable parameter options through the in-vehicle infotainment system to obtain a parameter adjustment plan; Occupant interaction operation: Based on the parameter adjustment plan, modify the recommended parameters through the in-vehicle infotainment system to obtain the initial recommended result; Conflict detection and correction: Based on the initial recommended result, perform logical verification on the parameters adjusted by the occupant to obtain the verification result; Result confirmation: Based on the verification result, generate a recommended result from the parameter combination confirmed by the occupant.

[0072] It should be noted that initially, the in-vehicle infotainment system outputs the recommended result to the occupant. The occupant modifies the parameters of the recommended solution through the in-vehicle infotainment system, such as parameters like air-conditioning temperature, volume, vehicle speed, and ambient light brightness, and verifies the modified solution parameters to remove conflicting data. If the verification result is passed, the in-vehicle system directly uses the parameter combination adjusted by the occupant as the final recommended result. When the verification result is a warning, in addition to displaying a warning sign, the in-vehicle system also explains in detail to the occupant the reason for the warning and the possible impacts. If the verification result is an error, the in-vehicle system immediately reverts to the initial recommended result, displays the detailed information of the recommended result again before the occupant confirms, and conducts a final confirmation. By verifying the parameters adjusted by the occupant, it avoids the appearance of unreasonable or non-conflicting solutions, ensures the feasibility and reasonableness of the recommended result, and prevents poor experiences caused by incorrect parameters.

[0073] The effect evaluation includes the following steps: Satisfaction scoring: Collect the satisfaction of the occupants with the recommendation results through in-vehicle questionnaires and voice feedback to obtain the satisfaction results; Behavior analysis: Based on the satisfaction results, analyze the behavior changes of the occupants after the implementation of the recommendation results to obtain the behavior analysis results; Comprehensive evaluation: Based on the satisfaction results and behavior analysis results, conduct a comprehensive evaluation of the satisfaction results and behavior analysis results to obtain the initial evaluation results; Feedback optimization: Based on the initial evaluation results, make adjustments and optimizations to obtain the evaluation results.

[0074] It should be noted that first, through the voice recognition system and screen of the in-vehicle unit, a satisfaction evaluation is sent to the occupants, and data related to the occupants' behavior is collected, including in-vehicle system logs, vehicle sensor data, and external data, etc. The proportion of whether the occupants click and execute the recommendation cannot be used to evaluate whether to use the recommendation plan. Instead, whether the occupants frequently manually adjust after the recommendation is used to evaluate whether to use the recommendation plan, and the recommendation results are fed back to the plan library. By continuously adjusting and optimizing the recommendation plan through the occupants' feedback and behavior analysis, it can adapt to the changes of different occupants and scenarios, and continuously improve the accuracy and quality of the recommendation.

[0075] As Figure 5 shown, the similarity calculation method includes the following steps: Feature extraction: Based on multiple personalized recommendation plans, extract the feature vectors of each plan to obtain a set of feature vectors; Similarity calculation: Based on the set of feature vectors, use the similarity calculation method to calculate the similarity between each feature vector to obtain a similarity matrix; Sorting and selection: Based on the similarity matrix, sort the multiple personalized recommendation plans from high to low in terms of similarity to obtain the personalized recommendation plans.

[0076] It should be noted that first, each recommendation plan is converted into a quantifiable feature vector, and the cosine similarity calculation method is used to calculate the similarity between the feature vectors of each plan. Then, the multiple personalized recommendation plans are sorted from high to low in terms of similarity to obtain a list of sorted personalized recommendation plans. Finally, the multiple personalized recommendation plans are recommended to the occupants, and the personalized recommendation results are displayed to the occupants through the interface and voice of the intelligent cockpit system, providing an intuitive and convenient interaction method, enabling the occupants to easily obtain the recommendation information and improving the user experience. During the process of plan adjustment, the adjusted results are timely displayed to the occupants, allowing the occupants to understand the changes of the recommendation plan in real time, enhancing the real-time and effectiveness of the interaction. By deeply mining the behavior patterns and emotional characteristics of the occupants, the similarity between multiple personalized recommendation plans is accurately calculated and sorted from high to low in terms of similarity, making the selection of the recommendation plan more in line with the preferences of the occupants.

[0077] Although embodiments of the present invention have been shown and described, those of ordinary skill in the art will appreciate that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A smart cockpit personalized recommendation method combining environmental perception and occupant preferences, including Data collection, based on the big data of environmental characteristics, the preference data of the occupants of the vehicle intelligent cockpit and the historical recommendation schemes, collects environmental data, occupant preference data and historical recommendation schemes to obtain environmental data items, occupant preference data items and historical recommendation scheme data items; Data analysis, based on the environmental data items, the occupant preference data items and the historical recommendation solution data items, the environmental data items, the occupant preference data items and the historical recommendation solution data items are analyzed to obtain analysis results; Features: Personalized recommendation generation: Based on the analysis results, the environmental feature analysis results are matched with the occupant preference analysis results to generate personalized recommendation solutions and obtain initial recommendation results; Solution adjustment: Based on the initial recommendation results, the vehicle-computer interaction system allows passengers to adjust solution parameters to obtain recommended results; Effect evaluation: Based on the recommendation results, the adjusted recommendation results are implemented in the intelligent cockpit system, and the recommendation scheme is evaluated through the passengers' satisfaction feedback on the recommendation results and the analysis of the passengers' behavior to obtain the evaluation results; Personalized recommendation generation methods include: A1. Environmental recognition: Based on the environmental feature analysis results, the current driving scene of the occupant is identified to obtain the driving scene recognition result; A2. Emotional preference recognition: Based on the occupant preference analysis results, the occupant's current behavior status and emotions are identified and compared with historical behavior status and emotional data to obtain the occupant's emotional results; A3, result matching, based on the driving scene recognition results and the passenger emotion results, matching is performed to obtain personalized recommendation results; A4. Result output: Based on personalized recommendation results, it is displayed to passengers through the interface and voice of the smart cockpit system.

2. The intelligent cockpit personalized recommendation method combining environmental perception and occupant preference according to claim 1, characterized in that: The environmental feature recognition method comprises the following steps: B1. Data integration: Based on vehicle sensors and external data sources, various data information is obtained, including vehicle speed, road conditions, weather conditions and geographic location, to obtain a data integration set; B2. Environmental feature extraction: Based on the data integration set, key features in the data integration set are extracted to obtain a feature data set; B3, driving scene recognition; Based on the feature data set, a driving scene classification model is established to obtain the initial driving scene recognition results; B4. Result verification: Based on the initial driving scene recognition result, the initial environment recognition result is verified using a verification method to obtain a driving scene recognition result.

3. The intelligent cockpit personalized recommendation method combining environmental perception and occupant preference according to claim 1, characterized in that: The emotion preference identification method comprises the following steps: C1. Behavior status and emotion recognition: Based on the occupant preference analysis results, the occupant's current behavior status and emotion are analyzed using analysis methods to obtain initial emotion preference data; C2, emotion classification, based on the initial emotion preference data, classify the initial emotion preference data to obtain the emotion classification result; C3. Data comparison: Based on the emotion classification results, the emotion classification results, historical behavior status and emotion data are compared to obtain the occupant emotion results.

4. The intelligent cockpit personalized recommendation method combining environmental perception and occupant preference according to claim 1, characterized in that: The personalized recommendation result includes multiple personalized recommendation results, and the method for obtaining the personalized recommendation result includes: D1. Establish a solution library, which contains personalized recommendation solutions for different driving scenarios and passenger emotions; D2, real-time matching, based on the driving scene recognition results and the occupant emotion results, matching is performed in the solution library to obtain multiple personalized recommendation solutions; D3. Recommendation result screening and sorting: Based on multiple personalized recommendation schemes, unreasonable and conflicting schemes are removed, and multiple personalized recommendation schemes are ranked according to the highest similarity through similarity calculation method to obtain personalized recommendation schemes.

5. The intelligent cockpit personalized recommendation method combining environmental perception and occupant preference according to claim 1, characterized in that: The scheme adjustment comprises the following steps: E1. Parameter adjustment: Based on the initial recommendation results, the vehicle-computer interaction system generates adjustable parameter options and obtains the parameter adjustment plan; E2, passenger interaction operation: based on the parameter adjustment plan, modify the recommended parameters through the vehicle-computer interaction system to obtain the initial recommendation results; E3, conflict detection and correction: Based on the initial recommendation results, the parameters adjusted by the occupant are logically checked to obtain the verification results; E4. Result confirmation: Based on the verification results, the parameters confirmed by the occupant are combined to generate a recommended result.

6. The intelligent cockpit personalized recommendation method combining environmental perception and occupant preference according to claim 1, characterized in that: The effect evaluation comprises the following steps: F1. Satisfaction score: Collect passengers’ satisfaction with the recommended results through vehicle questionnaires and voice feedback to obtain satisfaction results; F2. Behavior analysis: Based on the satisfaction results, analyze the behavioral changes of the passengers after the implementation of the recommended results to obtain the behavior analysis results; F3. Comprehensive evaluation: Based on the satisfaction results and behavior analysis results, conduct a comprehensive evaluation of the satisfaction results and behavior analysis results to obtain the initial evaluation results; F4, Feedback optimization: Based on the initial evaluation results, make adjustments and optimizations to obtain the evaluation results.

7. The intelligent cockpit personalized recommendation method combining environmental perception and occupant preference according to claim 2, characterized in that: The verification method comprises the following steps: G1, data cross-validation, compare the initial driving scene recognition results with similar data from different data sources and sensors to obtain initial validation results; G2, data comparison verification, based on the initial verification result, compare the initial driving scene recognition result with the initial verification result to obtain the verification result; G3, real-time feedback verification, based on the verification results, real-time feedback is provided through the vehicle-computer interaction system, and the initial driving scene recognition results are adjusted to obtain the driving scene recognition results.

8. The intelligent cockpit personalized recommendation method combining environmental perception and occupant preference according to claim 3 is characterized by: The analytical method comprises the following steps: H1, behavior pattern recognition, based on the analysis results of occupant preference, analyze the occupant's historical behavior data to obtain behavior pattern data; H2, emotional feature extraction, behavioral pattern data, extract members' emotional features and obtain emotional feature results; H3, association analysis, based on behavioral pattern data, analyzes environmental characteristics and historical behavioral pattern data to obtain initial emotional preference data.

9. The intelligent cockpit personalized recommendation method combining environmental perception and occupant preference according to claim 4, characterized in that: The similarity calculation method comprises the following steps: M1. Feature extraction: Based on multiple personalized recommendation schemes, extract the feature vector of each scheme to obtain a feature vector set; M2, similarity calculation, based on the feature vector set, use the similarity calculation method to calculate the similarity between each feature vector to obtain a similarity matrix; M3, sorting and selection, based on the similarity matrix, sort multiple personalized recommendation schemes from high to low according to similarity to obtain personalized recommendation schemes.

Images