Auxiliary driving method for performing data fitting based on deep learning and related equipment

Abstract

The invention provides an auxiliary driving method for performing data fitting based on deep learning and related equipment, and relates to the field of automatic control. According to the method, a pavement spatio-temporal data set is determined by collecting historical pavement information, time points and pavement position coordinates; and then, on the basis of the pavement spatio-temporal data set, a pavement information data set including pavement temperature, humidity, wear state, pothole condition and repair record is taken as input data, after a pavement condition category label is determined, a deep learning model is obtained through training, and the model can output a pavement condition category matched with the input data. Then, real-time road surface information is obtained through a sensor and monitoring equipment, and the deep learning model is input to determine the type of the real-time road surface condition; finally, the automatic driving strategy of the target vehicle is adjusted according to the weather forecast condition, the vehicle tire data, the road surface real-time condition category and a preset automatic driving behavior optimization model, and therefore safer and more intelligent driving is achieved.

Description

Technical Field

[0001] The present application relates to the field of automatic control, and in particular to an assisted driving method and related equipment based on deep learning for data fitting. Background Art

[0002] With the advancement of science and technology, autonomous driving technology has gradually become a research hotspot. The perception link in the autonomous driving system is crucial for accurate judgment of road conditions.

[0003] In the prior art, sensors and monitoring equipment are usually used to obtain real-time road information, such as cameras, radars, etc. These devices mainly detect some basic visible conditions of the road, such as obstacles, lane lines, etc., through vision and radar, and perform preliminary analysis on the acquired real-time information according to preset relatively simple rules and algorithms to determine the basic conditions of the road, such as whether there are obstacles, whether the lane is deviated, etc., to adjust the driving mode.

[0004] However, this approach has obvious flaws. Sensors and monitoring equipment often obtain real-time local information, making it difficult to fully understand the historical conditions and spatiotemporal changes of the road surface. Moreover, the preset relatively simple rules and algorithms are difficult to accurately judge the potential risks and complex conditions of the road surface. Therefore, the existing methods are not accurate enough in judging complex road conditions. Summary of the invention

[0005] The present application provides an assisted driving method and related equipment based on deep learning for data fitting, which is used to achieve a safer, smarter and more comfortable driving experience when the autonomous driving strategy is difficult to accurately adjust according to the real-time road conditions, weather conditions, in-vehicle passenger conditions and road safety conditions in the prior art.

[0006] In a first aspect, the present application provides an assisted driving method for data fitting based on deep learning, which is applied to a server, and the method includes: determining a road spatiotemporal data set based on various historical road surface information, collection time points and road surface position coordinates that have been collected; using a road surface information data group as input data of a deep learning model based on the road surface spatiotemporal data set, and the road surface information data group includes road surface temperature, road surface humidity, road surface wear state, road surface repair record and road surface pothole state; after determining a road surface condition category label that matches the input data, training a deep learning model, and the deep learning model is used to output a road surface condition category that matches the input data; after obtaining real-time road surface information based on sensors and monitoring equipment, inputting the real-time road surface information into the deep learning model to determine the real-time road surface condition category; obtaining weather forecast information through a meteorological terminal, and obtaining real-time tire data of a target vehicle through a sensor; adjusting the target vehicle's automatic driving strategy based on the weather forecast information, the real-time tire data, the real-time road surface condition category and a preset automatic driving behavior optimization model, and the automatic driving behavior optimization model is trained in advance based on different weather forecast information, the real-time tire data, the real-time road surface condition category and the corresponding braking distance.

[0007] By adopting the above technical solutions, applying deep learning models to predict road conditions, and training models with historical data, we can more accurately evaluate the real-time road conditions, integrate multi-source heterogeneous data, including historical road data, real-time road monitoring, weather forecasts, and tire sensor data, and perform correlation analysis to make the evaluation of road conditions more comprehensive. Then, we can adjust the autonomous driving strategy in real time according to the road conditions to achieve smarter and safer assisted driving and improve the driving experience.

[0008] In combination with some embodiments of the first aspect, in some embodiments, after obtaining real-time road surface information based on sensors and monitoring equipment, the real-time road surface information is input into the deep learning model, and after the step of determining the real-time road surface condition category, it also includes: real-time monitoring of the number of vehicles within a set range on the current lane through the monitoring equipment; comparing the number of vehicles with a preset congestion threshold to determine the congestion level of the current lane; real-time monitoring of the distribution of vehicles in lanes on both sides matching the current lane through the monitoring equipment; inputting the current lane congestion level, the distribution of vehicles in lanes on both sides and the traffic regulation information obtained by querying the map database into a preset autonomous driving strategy model, which is established through training based on multiple actual driving data and traffic rules; and obtaining autonomous driving control parameters under current road conditions.

[0009] By adopting the above technical solution, the server can monitor the number of vehicles in the current lane and the distribution of vehicles on both sides of the lane in real time, and use the automatic driving strategy model to determine the automatic driving control parameters under the current road conditions in combination with traffic regulation information. These technical features interact with each other, so that the automatic driving strategy can be adjusted according to the actual situation of the road, avoiding congested sections and improving driving efficiency. At the same time, through the coordination of monitoring equipment and automatic driving strategy models, it can better cope with complex traffic conditions and improve the safety and reliability of automatic driving.

[0010] In combination with some embodiments of the first aspect, in some embodiments, after the step of using the road surface information data group as input data of the deep learning model based on the road surface spatiotemporal data set, it also includes: after receiving the real-time weather data sent by the meteorological terminal, judging the current weather conditions according to the real-time weather data, and the current weather conditions at least include sunny, rainy or snowy days; if it is judged to be sunny, the road surface humidity is determined by monitoring the road moisture content through sensors; if it is judged to be rainy or snowy, the parameters of the road surface humidity are corrected in combination with the road moisture content and the real-time weather precipitation; according to the corrected road surface humidity, a matching automatic driving mode is determined, and the automatic driving mode includes at least a slippery mode, a normal mode and a snow mode.

[0011] By adopting the above technical solution, the server can judge the current weather conditions based on the real-time weather data sent by the meteorological terminal, and correct the parameters of road humidity in combination with the road moisture content. Then, the matching autonomous driving mode is determined based on the corrected road humidity. In this way, the autonomous driving strategy can be adjusted according to different weather and road conditions. For example, in slippery mode, the vehicle can adopt a more cautious driving strategy, which improves driving safety. At the same time, the interaction between real-time weather data and road humidity correction makes the switching of autonomous driving modes more accurate and timely, improving the adaptability of autonomous driving.

[0012] In combination with some embodiments of the first aspect, in some embodiments, after the step of adjusting the automatic driving strategy of the target vehicle according to the real-time road condition category and the preset automatic driving behavior optimization model, it also includes: determining the real-time image in the car by means of an image acquisition device arranged in the car; determining whether there is a child passenger in the car based on the real-time image in the car combined with an image processing algorithm; when it is determined that there is a child passenger in the car, adjusting the automatic driving strategy according to the preset child riding mode.

[0013] By adopting the above technical solution, the server can determine whether there are child passengers in the car through the image acquisition device, and adjust the automatic driving strategy according to the preset child riding mode. In this way, the automatic driving strategy can pay more attention to the safety of child passengers, such as adjusting the speed of the car, maintaining a safer distance between cars, etc. At the same time, the combination of image acquisition equipment and image processing algorithms can accurately identify child passengers, providing a reliable basis for the adjustment of automatic driving strategies, and improving the safety and humanization of automatic driving.

[0014] In combination with some embodiments of the first aspect, in some embodiments, when it is determined that there is a child passenger in the vehicle, after the step of adjusting the automatic driving strategy according to the preset child riding mode, it also includes: after receiving the real-time location data sent by the on-board navigation device, matching according to the accident-prone section information pre-stored in the map database, to determine whether the vehicle is within the set range of the accident-prone section, the accident-prone section refers to a road area where the number of traffic accidents that have occurred in the past set time exceeds a preset threshold; when the vehicle is within the set range of the accident-prone section, adjusting the corresponding optimized automatic driving strategy according to the historical accident characteristics corresponding to the accident-prone section.

[0015] By adopting the above technical solution, the server can adjust and optimize the automatic driving strategy according to the historical accident characteristics corresponding to the accident-prone section when the vehicle enters the set range of the accident-prone section. In this way, the automatic driving strategy can deal with accident-prone sections more cautiously. At the same time, the matching of the on-board navigation equipment, map database and accident-prone section information can timely and accurately determine the position of the vehicle, providing a timely basis for the adjustment of the automatic driving strategy and improving the safety of automatic driving.

[0016] In combination with some embodiments of the first aspect, in some embodiments, after obtaining real-time road surface information based on sensors and monitoring equipment, the real-time road surface information is input into the deep learning model, and after the step of determining the real-time road surface condition category, it also includes: determining the passenger status information and passenger vehicle demand information monitored by the monitoring equipment in the vehicle, the passenger status information at least includes the passenger fatigue status and rest status, and the passenger vehicle demand information at least includes the passenger travel needs and emergency needs; adjusting the automatic driving strategy of the target vehicle according to the real-time road surface condition category, the passenger status information and the passenger vehicle demand information, the automatic driving strategy including reducing the vehicle speed within a preset scenic spot range in combination with map navigation information after determining the travel needs.

[0017] By adopting the above technical solutions, the server can adjust the target vehicle's autonomous driving strategy according to the real-time road condition category, passenger status information and passenger vehicle demand information. For example, after determining the travel needs, reducing the vehicle speed within the preset scenic spot range in combination with map navigation information can provide passengers with a better sightseeing experience. These technical features interact with each other, enabling the autonomous driving strategy to better meet the needs of passengers and improve passenger satisfaction. At the same time, the consideration of passenger status information, such as passenger fatigue status, can also make the autonomous driving strategy more humane and ensure the safety and comfort of passengers.

[0018] In combination with some embodiments of the first aspect, in some embodiments, after the step of adjusting the autonomous driving strategy of the target vehicle according to the weather forecast information, the real-time tire data, the real-time road condition category and the preset autonomous driving behavior optimization model, it also includes: receiving user feedback information on the autonomous driving strategy, the feedback information including the user's satisfaction and specific opinions on the autonomous driving strategy; analyzing the feedback information to optimize the autonomous driving strategy.

[0019] By adopting the above technical solution, the server can receive user feedback on the autonomous driving strategy, analyze and optimize it. In this way, the autonomous driving strategy can be continuously improved and perfected to better meet user needs. User satisfaction and specific opinions can provide important references for the optimization of the autonomous driving strategy, so that the autonomous driving system can continuously improve performance and enhance user experience. At the same time, this feedback mechanism also promotes the development and progress of autonomous driving technology.

[0020] In a second aspect, the present application provides a server, comprising: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to cause the server to execute the method described in the first aspect and any possible implementation of the first aspect.

[0021] In a third aspect, the present application provides a computer-readable storage medium comprising instructions, which, when executed on a server, causes the server to execute the method described in the first aspect and any possible implementation of the first aspect.

[0022] In a fourth aspect, the present application provides a computer program product, which, when executed on a server, enables the server to execute the method described in the first aspect and any possible implementation manner of the first aspect.

[0023] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. Since the spatiotemporal data set of the road surface is determined based on historical road surface information, etc. Then, a data set containing various road surface information is used as the input of the deep learning model, and the road surface condition category label is determined and then the model is trained to output the road surface condition category. After obtaining the real-time road surface information, the input model is used to determine the real-time condition category. At the same time, weather forecast information is obtained through the meteorological terminal, and the sensor obtains the real-time tire data of the target vehicle. Finally, the target vehicle's automatic driving strategy is adjusted according to the weather forecast information, real-time tire data, real-time road surface condition category and the preset automatic driving behavior optimization model, which effectively improves the safety and stability of automatic driving.

[0024] 2. The use of technical means to determine whether there are child passengers in the car through image acquisition equipment and image processing algorithms, and adjust the automatic driving strategy according to the preset child riding mode, effectively solves the technical problem of insufficient attention to the needs of passengers in the car in the existing technology, and thus achieves the technical effect of improving the safety and humanization of automatic driving.

[0025] 3. By adopting the technical means of receiving user feedback on the autonomous driving strategy and analyzing and optimizing it, the technical problem that the autonomous driving strategy in the existing technology is difficult to meet the personalized needs of users is effectively solved, thereby achieving the technical effect of continuously improving the autonomous driving performance and user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is a flowchart of an assisted driving method for data fitting based on deep learning in an embodiment of the present application; Figure 2 is another flowchart of the assisted driving method for data fitting based on deep learning in an embodiment of the present application; Figure 3 It is a schematic diagram of a physical device structure of a server in an embodiment of the present application. DETAILED DESCRIPTION

[0027] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to be used as limitations to the present application. As used in the specification and appended claims of the present application, the singular expressions "one", "a kind of", "said", "above", "the" and "this" are intended to also include plural expressions, unless there is a clear indication to the contrary in the context. It should also be understood that the term "and / or" used in the present application refers to and includes any or all possible combinations of one or more listed items.

[0028] In the following, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as suggesting or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features, and in the description of the embodiments of the present application, unless otherwise specified, "plurality" means two or more.

[0029] For ease of understanding, the following is a description of the process of the method provided by this implementation. Figure 1 , which is a flow chart of an assisted driving method for data fitting based on deep learning in an embodiment of the present application.

[0030] S101, determining a road surface spatiotemporal data set according to each collected historical road surface information, collection time point and road surface position coordinates; When the server performs this step, it first needs to collect and organize a large amount of historical road surface data. Among them, historical road surface information is a comprehensive collection of various data, which reflects the various conditions and characteristics of the road in the past period of time, including road surface temperature, humidity, road surface wear status, road surface repair records, road surface material information and road surface potholes, etc. These can be detected in advance by equipment such as camera radar. The staff can use the camera to take an image of the road surface, and compare the color, texture and other features of the image with the existing road surface feature table to judge the road surface state. The road surface feature table contains the typical color feature value range, texture feature description and corresponding road surface state label under different road surface conditions. For example, the color feature value of a slippery road surface may be within a specific brightness and saturation range, and the texture feature may be a relatively uniform reflection without obvious granularity; the color of an icy road surface usually presents a lighter tone, and the texture feature is smooth and flat with almost no texture changes; the color of a pothole road surface may show a large color difference due to shadows and the mixture of different materials, and the texture feature is an irregular concave and convex shape and there may be scattered light and shadow changes around. By comparing the features obtained by image analysis with the features in the road surface feature table, the road surface state can be judged more accurately.

[0031] These data sources may also include: regular inspection reports from road management departments, real-time road condition information uploaded by vehicle sensors, satellite remote sensing images, historical weather data provided by meteorological departments, and road maintenance records from transportation departments. The server organizes these data in chronological order and geographical location to form a database containing multi-dimensional information. For example, for a specific road, the database may contain the following information: January 1, 2023, location coordinates (longitude: 116.3, latitude: 39.9), road surface temperature: 5°C, road surface humidity: 30%, road surface wear state: slight, last repair time: November 15, 2022; April 15, 2023, location coordinates (longitude: 116.3, latitude: 39.9), road surface temperature: 18°C, road surface humidity: 45%, road surface wear state: medium, last repair time: March 20, 2023; July 30, 2023, location coordinates (longitude: 116.3, latitude: 39.9), road surface temperature: 32°C, road surface humidity: 20%, road surface wear condition: severe, last repair time: March 20, 2023.

[0032] The server connects these data points to form a road condition curve that changes over time. This curve not only reflects the seasonal changes in road conditions, but also reflects the trend of gradual wear and tear of the road surface as it is used, as well as the improvement effect of road maintenance on road conditions.

[0033] S102, using a road surface information data set as input data for a deep learning model based on the road surface spatiotemporal data set, the road surface information data set including road surface temperature, road surface humidity, road surface wear status, road surface repair record, and road surface pothole status; In this step, the server deeply mines the information in the road spatiotemporal data set, extracts the road surface temperature, road surface humidity, road surface wear status, road surface repair records, road surface pothole status and other data closely related to the road surface conditions, and forms a road surface information data set. For example, the server will filter out the road surface temperature data of a certain section of road in a specific time period from the road spatiotemporal data set. If the road surface temperature of the section of road fluctuates greatly at different time points during this time period, the server will record these temperature data and the corresponding time points as part of the road surface information data set.

[0034] For road surface humidity, the server will also obtain relevant information from the data set. For example, in rainy or humid weather conditions, the road surface humidity of the road section will increase significantly, and the server will record the data of these humidity changes as well as the corresponding weather and time information. The wear state of the road surface is also an important consideration. The server will analyze the information about road surface flatness, scratches, cracks, etc. in the data set to assess the degree of wear of the road surface. If the road surface wear of a certain section is found to be more serious, the server will include the relevant wear data in the road surface information data set. The road surface repair record should not be ignored either. If a road section has been repaired, the server will obtain information such as the time of repair, the repaired part, and the condition after repair, and add these records to the road surface information data set.

[0035] S103, after determining the road condition category label that matches the input data, training a deep learning model, where the deep learning model is used to output the road condition category that matches the input data; When the server performs this step, it first needs to identify the category labels of the road conditions. These labels can be set according to actual needs and experience, for example, they can be divided into different categories such as good, general, poor, slippery, icy, etc. Each category label represents a specific road condition.

[0036] Next, the server will associate the previously sorted road information data set with the corresponding road condition category label. For example, a set of data with high road temperature, low road humidity, slight road wear and no recent repair records may be labeled as a "good" road condition category; while a set of data with low road temperature, high road humidity, severe road wear and recent repair records may be labeled as a "poor" road condition category. The server will then use these labeled data to train a deep learning model. Deep learning models usually have a multi-layer structure that can automatically learn features and patterns from input data. During the training process, the model will continuously adjust its parameters so that the output road condition category is as consistent as possible with the label that matches the input data. For example, the server can use a convolutional neural network (CNN) as a deep learning model. CNN can automatically extract useful features from the road information data set, such as the trend of road temperature, the distribution of road humidity, and the characteristics of road wear. Then, these features are combined and classified through the fully connected layer to finally output the road condition category.

[0037] During the training process, the server will input a large number of road information data sets into the model, and compare the output of the model with the actual label to calculate the loss function. Then, the back propagation algorithm is used to update the parameters of the model so that the loss function gradually decreases. Through continuous iterative training, the model can learn the complex relationship between different road information data sets and road condition categories. For better explanation, assume that the server has the following set of data: Data 1: The road surface temperature is 25°C, the road surface humidity is 40%, the road surface wear state is slight, there is no recent repair record, and the label is "good". Data 2: The road surface temperature is 10°C, the road surface humidity is 80%, the road surface wear state is moderate, there is a recent repair record, and the label is "general". Data 3: The road surface temperature is 0°C, the road surface humidity is 90%, the road surface wear state is severe, there is a recent repair record, and the label is "poor".

[0038] The server will input this data into the deep learning model for training multiple times. In the early stages of training, the model's output may be quite different from the actual label, but as training progresses, the model will gradually learn the characteristics and patterns of this data, and the output will become closer and closer to the actual label.

[0039] S104, after obtaining the real-time road surface information according to the sensor and the monitoring equipment, input the real-time road surface information into the deep learning model to determine the real-time road surface condition category; When the server performs this step, it will receive real-time road information from sensors and monitoring equipment. This information includes but is not limited to the current road surface temperature, road surface humidity, road surface wear status, etc. For example, the sensor may detect that the current road surface temperature is 20°C, the road surface humidity is 50%, the road surface wear status is slight, and there are no obvious signs of repair. The server will organize this real-time information into a data group and input it into the previously trained deep learning model. The deep learning model will analyze and judge the input real-time information based on the previously learned patterns and features. The model will automatically extract the key features of this information and compare and match them with the knowledge learned during previous training. For example, based on the input road surface temperature and humidity information, the model may judge that the current road surface is unlikely to be slippery or icy; based on the information on the road surface wear status, it may judge that the road surface has a good flatness. Combining these judgments, the model will output a real-time road condition category, such as "good".

[0040] In addition, the server will continue to receive new real-time information and input it into the model for judgment in a timely manner. In this way, the model can adjust the output category in time according to the changes in road conditions to ensure the timeliness and accuracy of the autonomous driving strategy. For example, during the driving process of the vehicle, the server may obtain real-time road information every once in a while (such as a few seconds) and input it into the model. If the vehicle drives from a dry road to a road that has just rained, the server will detect the change in road humidity in time and adjust the road condition category according to the judgment of the model, so as to adjust the autonomous driving strategy in time, such as reducing the speed and increasing the safety distance.

[0041] In some embodiments, the server may also receive data from in-vehicle monitoring equipment, which may include cameras or other sensors, etc. By analyzing this data, the server can determine the passenger's status information. Specifically: the server can determine whether the passenger is in a state of fatigue by analyzing the passenger's facial image captured by the camera, detecting indicators such as blinking frequency, yawning frequency, and preset fatigue state indicator thresholds. The server can also determine whether the passenger is in a resting state by analyzing the passenger's body posture, breathing rate and other data and rest state standard data sent by the detection device. For example, if the server detects that the passenger's blinking frequency has decreased significantly for 30 consecutive minutes and yawns many times, it may be determined that the passenger is in a state of fatigue.

[0042] The server can also analyze the passenger's behavior pattern, destination information, and interaction with the vehicle system to determine the passenger's vehicle needs. Specifically, it includes passenger travel needs, which can be determined by analyzing the destination set in the navigation system and whether the passenger frequently checks the information of scenic spots along the way; emergency needs, which can be determined by analyzing whether the passenger frequently checks the arrival time in the map data of the vehicle system, or obtained through the conversation information between the passenger and the voice equipment.

[0043] Based on the above information, the server will comprehensively consider the real-time road condition category, passenger status information and passenger vehicle demand information to adjust the target vehicle's autonomous driving strategy. Specific strategies may include: when it is determined to be a tourist demand, the server will combine the map navigation information to reduce the speed within the preset scenic spot range. When it is determined that the passenger has an urgent need, the server may choose a faster route and slightly increase the driving speed within the permitted range while ensuring safety. When it is detected that the passenger is resting, the server may suggest adjusting to a less noisy section of the road and further control the volume equipment in the car to reduce the noise in the car. In this way, the server can dynamically adjust the autonomous driving strategy according to real-time road conditions, passenger status and needs, which not only ensures driving safety, but also meets the personalized needs of passengers to the greatest extent, providing a more intelligent and humane riding experience.

[0044] S105, obtaining weather forecast information through a meteorological terminal, and obtaining real-time tire data of a target vehicle through a sensor; The server establishes a communication connection with the meteorological terminal. The meteorological terminal can be a service interface provided by a professional meteorological agency, or it can be the result of data interaction between the vehicle's own meteorological monitoring module and a wider meteorological network. The server sends a request to the meteorological terminal to obtain weather forecast information for the location of the target vehicle and the surrounding area. After receiving the request, the meteorological terminal determines the weather conditions in the relevant area based on the vehicle's positioning information (which can be obtained through the vehicle's GPS module and provided to the server). The weather forecast information includes but is not limited to weather type (such as sunny, cloudy, rainy, snowy, etc.), temperature range, humidity range, wind speed and wind direction, etc. In addition, the target vehicle is equipped with various tire sensors, which can monitor the relevant parameters of the tire in real time, including but not limited to tire pressure sensors and tire temperature sensors. The tire pressure sensor can measure the air pressure value inside the tire. If the tire pressure is too high or too low, it will affect the vehicle's handling and braking performance, and may also cause increased tire wear. The tire temperature sensor is used to monitor the temperature of the tire. During the driving process of the vehicle, the friction between the tire and the road will generate heat. Excessive tire temperature may reduce the service life of the tire and even cause safety hazards such as tire blowout. By monitoring the tire temperature in real time, abnormal conditions can be discovered in time and corresponding measures can be taken.

[0045] S106. Adjust the autonomous driving strategy of the target vehicle according to the weather forecast information, the real-time tire data, the real-time road condition category and a preset autonomous driving behavior optimization model. The autonomous driving behavior optimization model is trained in advance according to different weather forecast information, real-time tire data, real-time road condition categories and corresponding braking distances.

[0046] Specifically, in order to build an autonomous driving behavior optimization model that can adapt to various environments, a large amount of comprehensive data can be collected in advance, covering various weather types such as sunny days, rainy days, and snowy days in different weather conditions. For each weather, the relevant weather parameters are recorded in detail, such as temperature, humidity, wind speed, wind direction, and precipitation. At the same time, the driving data of the vehicle in these weather conditions are collected, including the vehicle's speed, acceleration, braking distance, and steering angular velocity. These data can reflect the actual driving status and performance of the vehicle in different weather environments. In terms of road conditions, data on different road types are collected, such as braking distance parameters for dry, wet, and icy roads. In addition, detailed data under different road condition categories are collected, including parameters such as road temperature, humidity, wear level, pothole degree, and road material. These road-related data can help the model understand the impact of different road conditions on vehicle driving. For tire conditions, relevant data including tire pressure parameters are collected. Tire pressure and temperature will affect the contact area and friction between the tire and the road, and thus affect the driving performance of the vehicle. By collecting vehicle driving data under different tire pressures, the impact of tire pressure on vehicle driving can be evaluated. The collected raw data needs to be preprocessed so that it can be better input into the deep neural network for training. For continuous numerical data such as vehicle speed, acceleration, braking distance and steering angular velocity, normalization is usually used. The convolution layer of the autonomous driving behavior optimization model can automatically extract local features in the data by sliding the convolution kernel on the input data. Taking the combined data of vehicle speed and weather parameters as an example, the convolution layer can extract the pattern characteristics of vehicle speed changes under different weather conditions. For example, the vehicle speed may decrease on rainy days. The convolution layer can capture the correlation characteristics between this speed change and the weather parameters of rainy days. Similarly, data such as vehicle tire temperature, tire pressure and braking distance under different road surfaces and weather conditions are extracted. All the previously extracted features are then integrated and comprehensively analyzed through the fully connected layer. It takes into account the relationship between various environmental factors and their combined impact on autonomous driving behavior. For example, it comprehensively considers factors such as road surface temperature, humidity, wear level, weather conditions and tire pressure to determine the final autonomous driving behavior adjustment strategy.

[0047] During the model training process, the model is repeatedly trained using a large amount of collected training data. For each set of input data (including various environmental parameters) and the corresponding expected output (correct autonomous driving behavior adjustment), the error between the model's predicted output and the expected output is calculated. Common error calculation methods include mean square error, etc., and then the weight parameters of the network are updated based on this error through the back propagation algorithm. The back propagation algorithm is an optimization algorithm based on gradient descent. It adjusts the weight of the network according to the gradient direction of the error so that the predicted output of the model gradually approaches the expected output. When the model receives different environmental inputs, it adjusts the vehicle speed according to the learned patterns and relationships. For example, in a rainy day and wet road environment, the model may reduce the vehicle speed based on the learned characteristics of increased braking distance and decreased tire grip when the road surface is wet on rainy days, as well as the corresponding vehicle speed adjustment strategy. The specific speed reduction may be fine-tuned according to actual conditions, such as precipitation intensity, road surface humidity and other factors. In snowy or icy road environments, due to the smaller friction of the road surface, the vehicle's braking distance increases significantly, and the model will significantly reduce the vehicle speed to ensure driving safety.

[0048] Through the above detailed technical means and processes, the autonomous driving behavior optimization model can continuously learn and adapt to different environmental inputs, thereby acquiring the ability to accurately adjust autonomous driving behaviors such as vehicle speed, steering, and braking according to different environmental inputs.

[0049] In some embodiments, the server monitors the number of vehicles within a set range on the current lane in real time through a monitoring device. The monitoring device may include sensors such as cameras or radars installed in advance at the set position of the target vehicle, or other devices with monitoring functions, which are not limited here. They can capture vehicle information on the lane in real time. For example, the camera can capture images on the lane and identify the number and position of vehicles through image processing technology; the radar can detect the presence and distance of the vehicle by emitting electromagnetic waves and receiving reflected waves. The server compares the number of vehicles in the current lane monitored with the preset congestion threshold to determine the congestion level of the current lane. Among them, the preset congestion threshold can be set based on experience and traffic rules. For example, if the number of vehicles on the lane per kilometer exceeds a certain number, the lane can be considered to be in a congested state. Assuming that the preset congestion threshold is 30 vehicles per kilometer, when the number of vehicles within the set range on the current lane is monitored to be 40 vehicles per kilometer, the server will determine that the current lane is highly congested.

[0050] At the same time, the server uses monitoring equipment to monitor the distribution of vehicles on both lanes that match the current lane in real time. These monitoring devices collect information such as the number, speed, and spacing of vehicles on both lanes. For example, the relative speed and distance of vehicles on both lanes can be detected by radar, so as to understand the distribution of vehicles. Next, the server inputs the current lane congestion, the distribution of vehicles on both lanes, and the traffic regulation information obtained by querying the map database into the preset autonomous driving strategy model. This autonomous driving strategy model is established by training based on multiple actual driving data and traffic rules. It can formulate the most appropriate autonomous driving strategy based on the input information and comprehensive consideration of various factors. For example, the map database may contain traffic regulation information such as speed limit information and lane division information of the road section. Suppose in a specific scenario, the server detects that there are a large number of vehicles on the current lane, the speed is slow, and the congestion is high; the distribution of vehicles on both lanes is relatively small and the speed is fast. The traffic regulation information in the map database shows that the road section allows vehicles to switch lanes safely. Based on these input information, the autonomous driving strategy model calculates the best autonomous driving control parameters, such as suggesting that the vehicle switch to the left lane at the right time and adjust the speed to adapt to the new lane conditions. In this way, the server can obtain the autonomous driving control parameters under current road conditions, thereby achieving more intelligent and safe autonomous driving.

[0051] In some practical examples, the server can also actively receive user feedback on the autonomous driving strategy. This feedback includes user satisfaction with the autonomous driving strategy and specific opinions. For example, users may evaluate the performance of the autonomous driving strategy in a certain journey through the interface of the vehicle system or related applications, and express their satisfaction with vehicle speed control, lane keeping, brake sensitivity, etc. At the same time, they may also put forward some specific opinions, such as hoping that the vehicle can accelerate more smoothly in certain situations, or decelerate earlier on specific sections of the road.

[0052] After receiving these feedback information, the server will use a series of technical means to analyze it. First, the server will use natural language processing technology to perform text analysis on the user's specific opinions and extract key information, such as the specific problems mentioned by the user and the expected direction of improvement. Then, the server will quantify the user's satisfaction data, for example, dividing the satisfaction into different levels and assigning corresponding scores to each level. Next, the server will optimize the autonomous driving strategy based on the analysis results. For example, if multiple users report that the vehicle accelerates too hastily under a certain type of road condition, the server will adjust the relevant control parameters so that the vehicle can accelerate more smoothly under similar road conditions. For another example, if users generally hope that the vehicle can maintain a higher cruising speed on the highway, the server will appropriately increase the default cruising speed on the highway under the premise of safety.

[0053] In addition, the server will use machine learning algorithms to analyze historical feedback information to predict users' possible needs and preferences. For example, by analyzing a large number of user feedback information in different seasons and under different road conditions, the server can discover certain patterns, such as users' high attention to the vehicle's performance in snow mode in winter, or users' specific expectations for autonomous driving strategies near tourist attractions during the peak tourist season. Based on these predictions, the server can optimize the autonomous driving strategy in advance to better meet user needs.

[0054] In the embodiments of the present application, since a deep learning model is used to learn and analyze historical road information and combined with multi-dimensional data such as real-time road conditions, weather conditions, in-vehicle passenger conditions, and road safety information, it is possible to more accurately judge the current road conditions and make corresponding driving strategy adjustments, effectively solving the problem in the prior art that it is difficult to accurately adjust the autonomous driving strategy according to complex road conditions, thereby achieving a safer, smarter and more comfortable autonomous driving experience.

[0055] In some scenarios, in some family travel scenarios, especially family self-driving trips or daily pick-ups with child passengers, or for some special vehicles for children's medical care, such as children's ambulances or shuttle buses for children's specialized hospitals, some autonomous driving modes are needed to improve the safety of child passengers.

[0056] After combining the above scenarios, the following is a more detailed description of the process of the method provided by this implementation. Figure 2 , is another flow chart of the assisted driving method for data fitting based on deep learning in an embodiment of the present application.

[0057] S201, determining a real-time image in the vehicle compartment by using an image acquisition device disposed in the vehicle; When the server performs this step, it will start the image acquisition devices installed in the car. These devices usually include high-definition cameras that can capture real-time images in the car from all directions. For example, cameras are installed in the front row, back row and every corner of the vehicle to ensure that the entire car space can be covered. The camera will continuously capture images at a certain frame rate and convert the real-time situation in the car into image data. These image data will be transmitted to the server in real time for processing. After receiving the image data, the server will perform preliminary processing and analysis on the data. First, the server will perform noise reduction and enhancement processing on the image to improve the quality and clarity of the image. For example, by removing noise points in the image and enhancing the contrast and brightness of the image, the objects and passengers in the car can be presented more clearly. Then, the server will segment and identify the image to distinguish different areas and objects in the car. For example, the seat area, passenger area, luggage area, etc. are divided so that passengers can be identified more accurately later. In practical applications, assuming that the vehicle is driving, the image acquisition device in the car will continuously capture real-time images in the car. The server will receive these image data and perform the above-mentioned processing.

[0058] S202, determining whether there is a child passenger in the carriage according to the real-time image in the carriage in combination with an image processing algorithm; The server uses object detection algorithms to detect whether there are human bodies in the image. These algorithms are usually based on deep learning techniques, such as convolutional neural networks, which can automatically extract features from the image and determine whether there are human outlines. If a human body is detected, the server will further analyze the features of these human bodies.

[0059] For each detected human body, the server extracts multiple features, such as height, body shape, head proportions, etc. For example, the height is estimated by analyzing the pixel height of the human body in the image, the body size is determined by analyzing the outline shape of the human body, and the characteristics of a child are determined by comparing the ratio of the head to the body. In addition, the server also considers other features, such as facial features, clothing, etc. When judging whether it is a child passenger, the server will make a judgment based on preset standards. For example, a human body with a height below a certain threshold, a large head proportion and a small body may be a child passenger. At the same time, the server will combine the analysis of facial features, such as children's unique facial contours and expressions, to improve the accuracy of the judgment.

[0060] To better illustrate, assume that the server receives a real-time image of the interior of a vehicle. After preprocessing, it uses a target detection algorithm to detect two human bodies in the image. Then, feature extraction is performed on the two human bodies, and it is found that one of the human bodies is shorter, has a larger head proportion, is smaller in size, and has facial features that indicate that it may be a child. The server will combine this information and determine that the human body may be a child passenger. At the same time, the other human body is analyzed and found to be taller and larger in size, which does not meet the characteristics of a child, and is determined to be an adult passenger. In this way, the server can accurately identify whether there are child passengers from the real-time images in the vehicle compartment, providing an important basis for subsequent adjustments to the autonomous driving strategy.

[0061] If it is determined that there is no child passenger in the carriage, step S203 is executed; If it is determined that there is a child passenger in the carriage, step S204 is executed.

[0062] S203: Do not adjust the autonomous driving strategy; When the server determines that there is no child passenger in the vehicle according to the judgment result of step S202, or determines in step S205 that the vehicle is not within the set range of the accident-prone road section, it will determine not to adjust the automatic driving strategy. In this case, the automatic driving strategy will maintain the previous settings to adapt to the current road conditions, traffic conditions and other relevant factors.

[0063] S204, adjusting the automatic driving strategy according to a preset children's riding mode; When the server performs this step, it will adjust the autonomous driving strategy accordingly according to the preset child riding mode. The preset child riding mode usually takes into account the special needs and safety factors of children to ensure the safety and comfort of children during the ride. For example, the server will reduce the speed of the vehicle. Because children are relatively fragile and their ability to react in emergencies is not as good as that of adults, a lower speed can reduce the degree of injury in the event of an accident. Assuming that the original vehicle's driving speed in normal mode is 60 kilometers per hour, when it is determined that there are child passengers in the car, the server will reduce the speed to 40 kilometers per hour. At the same time, the server will increase the safety distance between the vehicle and the vehicle in front. This can provide more braking time for the vehicle to avoid impact on children during emergency braking. For example, the safety distance will be increased from the original 50 meters to 80 meters.

[0064] In addition, the server will adjust the acceleration and braking sensitivity of the vehicle to make the acceleration and braking of the vehicle smoother and reduce the impact of inertia on children. For example, the acceleration will be slowed down to avoid children feeling uncomfortable during sudden acceleration; the braking sensitivity will be adjusted to make the braking gentler to avoid sudden braking that may cause children to fall. In terms of steering, the server will make the steering of the vehicle smoother. Avoid sharp turns and other drastic operations to ensure the stability of children in the car.

[0065] S205, after receiving the real-time location data sent by the vehicle-mounted navigation device, matching is performed according to the accident-prone road section information pre-stored in the map database to determine whether the vehicle is within a set range of the accident-prone road section, where the accident-prone road section refers to a road area where the number of traffic accidents that occurred within a set period of time in the past exceeds a preset threshold; When the server performs this step, it will receive the real-time location data sent by the vehicle navigation device in real time. These location data include the current latitude and longitude information of the vehicle. The server will match these real-time location data with the accident-prone sections information pre-stored in the map database. The map database pre-stores a large amount of road information, including historical traffic accident data of each section. Based on these data, the server will determine which sections have exceeded the preset threshold number of traffic accidents in the past set time, and thus mark these sections as accident-prone sections. When the server receives the real-time location data of the vehicle, it will determine whether the vehicle is within the set range of these accident-prone sections. For example, the set range of a certain accident-prone section is an area with a radius of 500 meters centered on the section. If the real-time location of the vehicle is within this range, the server will determine that the vehicle is within the set range of the accident-prone section.

[0066] If it is determined that the vehicle is within the set range of the accident-prone road section, step S206 is executed; If it is determined that the vehicle is not within the set range of the accident-prone road section, step S203 is executed.

[0067] S206. Adjust the corresponding optimized autonomous driving strategy according to the historical accident characteristics corresponding to the accident-prone road section.

[0068] When the server performs this step, it will deeply analyze the historical accident characteristics corresponding to the accident-prone section. These characteristics may include the time of the accident, weather conditions, accident type (such as collision, rear-end collision, etc.), accident cause (such as speeding, illegal lane change, etc.), and traffic flow at the time of the accident. For example, if the historical accident data shows that there are more accidents on this section of road during rainy days, and the main reason is that the road is slippery and causes the vehicle to lose control, the server will adjust the automatic driving strategy based on the real-time weather data if it is currently raining and the vehicle is driving on this section of road, such as reducing the speed, increasing the brake response sensitivity, adjusting the tire traction control parameters, etc., to improve the stability of the vehicle on slippery roads. For another example, if the historical accident characteristics show that the road section is prone to rear-end collisions when the traffic flow is heavy during peak hours, the server will appropriately increase the safe distance from the vehicle in front when the vehicle enters the peak hours of the section, and pay close attention to the dynamics of vehicles in the lanes on both sides to avoid frequent lane changes to reduce the risk of rear-end collisions. For example, if it is found that accidents on this section of road often occur at night and the cause is poor visibility, the server will adjust the vehicle's lighting settings, increase the brightness and illumination range of the headlights, and remind the driver to remain highly vigilant.

[0069] Through detailed analysis of historical accident characteristics on accident-prone sections and targeted adjustments to autonomous driving strategies, the server can effectively improve vehicle driving safety on these sections and reduce the probability of accidents.

[0070] In the embodiments of the present application, since the situation of passengers in the vehicle, vehicle location information and road safety conditions are comprehensively considered and the autonomous driving strategy is accurately adjusted based on this, the problem that the existing autonomous driving technology is difficult to take into account various factors to ensure driving safety and passenger comfort is effectively solved, thereby achieving the goal of improving the safety, humanization and adaptability of autonomous driving.

[0071] The following describes the server in the embodiment of the present invention from the perspective of hardware processing. Figure 3 , is a schematic diagram of a physical device structure of a server in an embodiment of the present application.

[0072] It should be noted that Figure 3 The structure of the server shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0073] like Figure 3As shown, the server includes a central processing unit (CPU) 301, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 302 or the program loaded from the storage part 308 to the random access memory (RAM) 303, such as executing the method described in the above embodiment. In RAM 303, various programs and data required for system operation are also stored. CPU 301, ROM 302 and RAM 303 are connected to each other through bus 304. Input / output (I / O) interface 305 is also connected to bus 304.

[0074] The following components are connected to the I / O interface 305: an input section 306 including an audio input device, a button switch, etc.; an output section 307 including a liquid crystal display (LCD) and an audio output device, an indicator light, etc.; a storage section 308 including a hard disk, etc.; and a communication section 309 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as needed. A removable medium 311, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 310 as needed so that a computer program read therefrom is installed into the storage section 308 as needed.

[0075] In particular, according to an embodiment of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication part 309, and / or installed from a removable medium 311. When the computer program is executed by the central processing unit (CPU) 301, various functions defined in the present invention are performed.

[0076] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, apparatus, or device.

[0077] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. Each box in the flowchart or block diagram may represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in an order different from that marked in the accompanying drawings.

[0078] Specifically, the server of this embodiment includes a processor and a memory, and a computer program is stored in the memory. When the computer program is executed by the processor, the assisted driving method for data fitting based on deep learning provided in the above embodiment is implemented.

[0079] As another aspect, the present invention further provides a computer-readable storage medium, which may be included in the server described in the above embodiment; or may exist independently without being assembled into the server. The above storage medium carries one or more computer programs, and when the above one or more computer programs are executed by a processor of the server, the server implements the assisted driving method based on deep learning for data fitting provided in the above embodiment.

[0080] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

[0081] As used in the above embodiments, the term "when..." may be interpreted to mean "if..." or "after..." or "in response to determining..." or "in response to detecting...", depending on the context. Similarly, the phrases "upon determining..." or "if (the stated condition or event) is detected" may be interpreted to mean "if determining..." or "in response to determining..." or "upon detecting (the stated condition or event)" or "in response to detecting (the stated condition or event)", depending on the context.

[0082] Those skilled in the art can understand that to implement all or part of the processes in the above-mentioned embodiments, the processes can be completed by computer programs to instruct related hardware, and the programs can be stored in computer-readable storage media. When the programs are executed, they can include the processes of the above-mentioned method embodiments. The aforementioned storage media include: ROM or random access memory RAM, magnetic disk or optical disk and other media that can store program codes.

Claims

1. An assisted driving method for data fitting based on deep learning, applied to a server, characterized in that: The method comprises: Determine the road surface spatiotemporal data set based on the collected historical road surface information, collection time points and road surface location coordinates; Based on the road spatiotemporal data set, a road surface information data set is used as input data of a deep learning model, wherein the road surface information data set includes road surface temperature, road surface humidity, road surface wear state, road surface repair record and road surface pothole state; After determining the road condition category label that matches the input data, training a deep learning model, wherein the deep learning model is used to output the road condition category that matches the input data; After obtaining real-time road surface information according to sensors and monitoring equipment, the real-time road surface information is input into the deep learning model to determine the real-time road surface condition category; Obtain weather forecast information through the meteorological terminal and obtain real-time tire data of the target vehicle through sensors; The autonomous driving strategy of the target vehicle is adjusted according to the weather forecast information, the real-time tire data, the real-time road condition category and a preset autonomous driving behavior optimization model, wherein the autonomous driving behavior optimization model is obtained by training in advance based on different weather forecast information, the real-time tire data, the real-time road condition category and the corresponding braking distance.

2. The method according to claim 1, characterized in that After obtaining the real-time road surface information according to the sensor and the monitoring equipment, the real-time road surface information is input into the deep learning model to determine the real-time road surface condition category, and the following steps are also included: Monitor the number of vehicles within the set range on the current lane in real time through monitoring equipment; Comparing the number of vehicles with a preset congestion threshold to determine the congestion level of the current lane; Monitoring the distribution of vehicles on the lanes on both sides matching the current lane in real time through monitoring equipment; Inputting the current lane congestion level, the vehicle distribution of the lanes on both sides, and the traffic regulation information obtained by querying a map database into a preset automatic driving strategy model, wherein the automatic driving strategy model is established by training based on a plurality of actual driving data and traffic rules; Obtain the automatic driving control parameters under the current road conditions.

3. The method according to claim 1, characterized in that After the step of using the road surface information data set as input data of the deep learning model based on the road surface spatiotemporal data set, the method further includes: After receiving the real-time weather data sent by the meteorological terminal, judging the current weather conditions according to the real-time weather data, the current weather conditions at least include sunny days, rainy days or snowy days; If it is determined to be a sunny day, the road surface humidity is determined by monitoring the road moisture content through a sensor; If it is determined to be the rainy day or the snowy day, the parameters of the road surface humidity are corrected in combination with the road moisture content and the real-time weather precipitation; A matching automatic driving mode is determined according to the corrected road surface humidity, and the automatic driving mode includes at least a slippery mode, a normal mode and a snow mode.

4. The method according to claim 1, characterized in that After the step of adjusting the automatic driving strategy of the target vehicle according to the weather forecast information, the real-time tire data, the real-time road condition category adjustment and the preset automatic driving behavior optimization model, the step further includes: Determine the real-time image in the vehicle compartment by using an image acquisition device arranged in the vehicle; Determine whether there is a child passenger in the carriage based on the real-time image in the carriage in combination with an image processing algorithm; When it is determined that there is a child passenger in the vehicle cabin, the automatic driving strategy is adjusted according to a preset child riding mode.

5. The method according to claim 4, characterized in that When it is determined that there is a child passenger in the vehicle compartment, after the step of adjusting the automatic driving strategy according to the preset child riding mode, the method further includes: After receiving the real-time location data sent by the vehicle navigation device, matching is performed according to the accident-prone road section information pre-stored in the map database to determine whether the vehicle is within the set range of the accident-prone road section, where the number of traffic accidents that occurred in the past set time exceeds a preset threshold; When the vehicle is within the set range of the accident-prone section, the corresponding optimized automatic driving strategy is adjusted according to the historical accident characteristics corresponding to the accident-prone section.

6. The method according to claim 1, characterized in that After obtaining the real-time road surface information according to the sensor and the monitoring equipment, the real-time road surface information is input into the deep learning model to determine the real-time road surface condition category, and the following steps are also included: Determine the passenger status information and the passenger vehicle demand information monitored by the monitoring device in the vehicle, wherein the passenger status information at least includes the passenger fatigue status and rest status, and the passenger vehicle demand information at least includes the passenger travel demand and emergency demand; The automatic driving strategy of the target vehicle is adjusted according to the real-time road condition category, the passenger status information and the passenger vehicle demand information. The automatic driving strategy includes reducing the vehicle speed within a preset scenic spot range in combination with map navigation information after determining tourism needs.

7. The method according to claim 1, characterized in that After the step of adjusting the automatic driving strategy of the target vehicle according to the weather forecast information, the real-time tire data, the real-time road condition category adjustment and the preset automatic driving behavior optimization model, the step further includes: receiving user feedback information on the autonomous driving strategy, wherein the feedback information includes user satisfaction and specific opinions on the autonomous driving strategy; The feedback information is analyzed to optimize the autonomous driving strategy.

8. A server, characterized in that: The server includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the server to execute the method described in any one of claims 1-7.

9. A computer-readable storage medium comprising instructions, characterized in that: When the instructions are executed on a server, the server is caused to execute the method according to any one of claims 1 to 7.

10. A computer program product, characterized in that When the computer program product is run on a server, the server is caused to execute the method according to any one of claims 1 to 7.

Images