An intelligent system for automatically determining the condition of a vehicle

Abstract

The application discloses an intelligent system for automatically judging the vehicle condition of an automobile, and relates to the technical field of the automobile industry. The application discloses an intelligent system for automatically judging the vehicle condition of an automobile, and relates to the technical field of the automobile industry. The application discloses an intelligent system for automatically judging the vehicle condition of an automobile, and relates to the technical field of the automobile industry. The application discloses an intelligent system for automatically judging the vehicle condition of an automobile, and relates to the technical field of the automobile industry. The application discloses an intelligent system for automatically judging the vehicle condition of an automobile, and relates to the technical field of the automobile industry. The application discloses an intelligent system for automatically judging the vehicle condition of an automobile, and relates to the technical field of the automobile industry. The application discloses an intelligent system for automatically judging the vehicle condition of an automobile, and relates to the technical field of the automobile industry. The application discloses an intelligent system for automatically judging the vehicle condition of an automobile, and relates to the technical field of the automobile industry. The application discloses an intelligent system for automatically judging the vehicle condition of

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of vehicle management, in particular to an intelligent system for automatically judging the condition of a car. BACKGROUND

[0002] Currently, the judgment of the condition of a car mainly relies on manual inspection and analysis, which is inefficient and prone to human error. With the increase in the number of vehicles and the accumulation of vehicle condition data, how to use intelligent technology to quickly and accurately determine the condition of a vehicle has become a problem to be solved. In the prior art, there is a lack of a system capable of effectively analyzing and learning from massive historical text data, resulting in a low level of automation and intelligence in vehicle condition judgment. SUMMARY

[0003] The present application aims to at least partially solve one of the above technical problems. To this end, the present application aims to provide an intelligent system for automatically judging the condition of a car, which uses natural language processing (NLP) technology and neural network models to realize the automatic analysis and determination of vehicle text data, solving the problems of low efficiency and error-prone in the prior art.

[0004] To achieve the above-mentioned purpose, the embodiments of the present application provide an intelligent system for automatically judging the condition of a car, comprising:

[0005] a collection module for collecting vehicle condition data and performing text conversion to obtain text data;

[0006] a first judgment module for inputting the text data into a pre-trained vehicle condition judgment model and outputting a vehicle condition determination result.

[0007] According to some embodiments of the present application, the collection module comprises:

[0008] a first construction module for:

[0009] determining the interval information between each data collection sensor for collecting the condition data of the car; the interval information includes distance information obtained by GPS positioning, communication delay and network hop count;

[0010] constructing a minimum distance matrix according to the interval information and a particle swarm algorithm, the minimum distance matrix being a square matrix, wherein each data collection node represents the minimum collection cost between the corresponding two data collection sensors as a data collection network;

[0011] a first determination module for:

[0012] obtaining the total bandwidth corresponding to the data collection network;

[0013] determine a first number of receiving paths and a second number of forwarding paths of each data collection node;

[0014] query a preset allocation data table according to the first number and the second number, and determine an importance coefficient of each data collection node in the minimum distance matrix;

[0015] an allocation module configured to allocate the total bandwidth according to the importance coefficient of each data collection node to obtain an allocation result, and the data collection network collects the vehicle condition data of the vehicle according to the allocation result;

[0016] a conversion module configured to perform text conversion on the vehicle data to obtain text data.

[0017] According to some embodiments of the present application, the various data collection sensors include an air flow sensor, an ABS sensor, a throttle position sensor, a crankshaft position sensor, a camshaft position sensor, a water temperature sensor, an oil pressure sensor, an oxygen sensor, a vehicle speed sensor, and a knock sensor.

[0018] According to some embodiments of the present application, the first determination module obtains the total bandwidth corresponding to the data collection network in real time based on Nagi os, Zabbix, PRTG Network Monitor network monitoring tools.

[0019] According to some embodiments of the present application, the allocation module allocates the total bandwidth according to the importance coefficient of each data collection node to obtain an allocation result, including:

[0020] determining proportion information according to the importance coefficient of each data collection node;

[0021] allocating the total bandwidth according to the proportion information to obtain an allocation result.

[0022] According to some embodiments of the present application, the first determination module obtains a vehicle condition determination model, including:

[0023] an acquisition module configured to acquire historical vehicle condition data of the vehicle and perform text conversion to obtain historical text data of the vehicle;

[0024] a data preprocessing module configured to clean and format the historical text data, remove noise and irrelevant information, standardize the data format, and obtain preprocessed data;

[0025] an NLP processing module configured to perform word segmentation processing on the preprocessed data using a professional word segmentation dictionary, disassemble the preprocessed data into independent word groups, classify the word groups, and obtain classified data;

[0026] The neural network training module is configured to train the combined model of the convolutional neural network and the recurrent neural network based on the classified data, and obtain a trained model.

[0027] The evaluation module is configured to perform model evaluation on the trained model based on test data, and obtain a vehicle condition judgment model when the evaluation is qualified.

[0028] According to some embodiments of the present application, the NLP processing module comprises:

[0029] The second construction module is configured to obtain vehicle component information, and construct a professional segmentation dictionary for vehicle components according to the vehicle component information.

[0030] The segmentation module is configured to perform segmentation processing on the preprocessed data based on the professional segmentation dictionary, and decompose the preprocessed data into independent word groups.

[0031] The classification module is configured to perform part-of-speech tagging on the word groups, and perform category classification on the word groups according to the tagging results, to obtain classified data.

[0032] According to some embodiments of the present application, the data preprocessing module comprises:

[0033] The cleaning module is configured to obtain type information of historical text data, query a preset data cleaning rule file according to the type information, obtain corresponding data cleaning rules, and perform data cleaning according to the corresponding data cleaning rules, to obtain cleaned data.

[0034] The formatting processing module is configured to perform data processing on the cleaned data according to an adaptive regular expression, to obtain preprocessed data.

[0035] According to some embodiments of the present application, the classification module comprises:

[0036] The second determination module is configured to determine a word vector of each word group, match the word vector with a standard word vector, and perform part-of-speech tagging on the word group according to a part of speech corresponding to the matched standard word vector.

[0037] The arrangement module is configured to perform category classification and arrangement on the word groups of the same part of speech, to obtain classified data.

[0038] According to some embodiments of the present application, the evaluation module comprises:

[0039] The calculation module is configured to obtain H test text data included in the test data and sample vehicle condition judgment results corresponding to the H test text data, input the H test text data into the trained model respectively to output H corresponding predicted vehicle condition judgment results, and calculate a matching degree Q of the sample vehicle condition judgment results corresponding to the H test text data and the H predicted vehicle condition judgment results.

[0040]

[0041] wherein, S a is the a-th predicted vehicle condition determination result output by the training model; S a is the sample vehicle condition determination result of the a-th test text data input into the training model;

[0042] The second judgment module is configured to judge whether the matching degree is greater than a preset matching degree, and when it is determined that the matching degree is greater than the preset matching degree, it indicates that the evaluation is qualified, and a vehicle condition judgment model is obtained; otherwise, it indicates that the evaluation is unqualified, and the training of the training model needs to be continued.

[0043] The present application provides an intelligent system for automatically judging the vehicle condition of an automobile, which utilizes natural language processing (NLP) technology and a neural network model to realize the automatic analysis and determination of vehicle text data, and solves the problems of low efficiency and error-prone in the prior art.

[0044] Other features and advantages of the present application will be described in the following specification, and some will become apparent from the specification, or will be understood through implementation of the present application. The objects and other advantages of the present application can be achieved and obtained by the structure specifically pointed out in the written specification and the accompanying drawings.

[0045] The technical solutions of the present application will be further described in detail below with the aid of the accompanying drawings and examples. BRIEF DESCRIPTION OF DRAWINGS

[0046] The accompanying drawings are used to provide a further understanding of the present application, and constitute a part of the specification, and are used to explain the present application together with the embodiments of the present application, and do not constitute a limitation on the present application. In the drawings:

[0047] Figure 1 is a block diagram of an intelligent system for automatically judging the vehicle condition of an automobile according to an embodiment of the present application;

[0048] Figure 2 is a block diagram of a collection module according to an embodiment of the present application;

[0049] Figure 3 is a block diagram of a first judgment module according to an embodiment of the present application. DETAILED DESCRIPTION

[0050] The preferred embodiments of the present application will be described below in conjunction with the accompanying drawings, and it should be understood that the preferred embodiments described herein are only used to illustrate and explain the present application, and do not constitute a limitation on the present application.

[0051] As Figure 1 shown, the embodiment of the present application provides an intelligent system for automatically judging the vehicle condition of an automobile, which comprises:

[0052] The collection module is configured to collect vehicle condition data of the automobile and perform text conversion to obtain text data.

[0053] The first determination module is configured to input the text data into a pre-trained vehicle condition determination model to output a vehicle condition determination result.

[0054] The working principle of the above technical solution is as follows: the collection module collects vehicle condition data of the automobile and performs text conversion to obtain text data; and the first determination module inputs the text data into a pre-trained vehicle condition determination model to output a vehicle condition determination result.

[0055] The above technical solution has the following beneficial effects: the natural language processing (NLP) technology and the neural network model are used to realize automatic analysis and determination of vehicle text data, and the problems of low efficiency and error-prone in the prior art are solved.

[0056] As shown in Figure 2 According to some embodiments of the present application, the collection module comprises:

[0057] The first construction module is configured to:

[0058] determine interval information between each data collection sensor for collecting vehicle condition data of the automobile; the interval information comprises distance information obtained by GPS positioning, communication delay, and network hop count;

[0059] construct a minimum distance matrix according to the interval information and a particle swarm algorithm; the minimum distance matrix is a square matrix, wherein each data collection node represents the minimum collection cost between corresponding two data collection sensors, and serves as a data collection network;

[0060] The first determination module is configured to:

[0061] obtain a total bandwidth corresponding to the data collection network;

[0062] determine a first number of receiving paths and a second number of forwarding paths of each data collection node;

[0063] query a preset allocation data table according to the first number and the second number to determine an importance coefficient of each data collection node in the minimum distance matrix;

[0064] The allocation module is configured to allocate the total bandwidth according to the importance coefficient of each data collection node to obtain an allocation result; and the data collection network collects the vehicle condition data of the automobile according to the allocation result.

[0065] The conversion module is configured to perform text conversion on the vehicle data to obtain text data.

[0066] The working principle of the above technical solution: in this embodiment, the distance information between each data sensor, the communication delay and the network hop count are determined based on the interval information; in order to reduce the communication delay, high-speed and low-delay communication protocols and technologies can be used. At the same time, optimizing the network layout and reducing network congestion are also effective methods to reduce the communication delay. In the vehicle network, the network hop count and the delay of data transmission can be reduced through multi-hop communication.

[0067] In this embodiment, the data acquisition node: each data acquisition sensor is regarded as a node. The acquisition cost: changes under the influence of the interval information. Particle swarm optimization algorithm: PSO is an optimization algorithm that finds the optimal solution to a problem by simulating the foraging behavior of a bird swarm. Each particle is regarded as a possible network configuration, and its position represents the connection mode between nodes, while its speed represents the adjustment direction of the connection mode. Step 1: initialization: node list: list all data acquisition sensors (nodes). Interval matrix: create a matrix, where element M[i][j] represents the physical distance between node i and node j (considering the weight of communication delay). Particle swarm: initialize a swarm of particles, each representing a possible network configuration (i.e. the connection mode between nodes). Step 2: evaluate fitness: calculate the total cost for each particle (i.e. each network configuration). This usually involves traversing the interval matrix and selecting appropriate paths and connections according to the configuration of the particle, and then accumulating the cost on these paths. Fitness function: take the total cost as the output of the fitness function. The goal is to find the configuration with the lowest total cost. Step 3: update particles: personal best position: for each particle, update its personal best position (i.e. the best configuration found so far). Global best position: find the global best position among all particles (i.e. the best configuration found in the entire particle swarm). Speed and position update: update the speed and position of each particle according to the rules of PSO. This usually involves moving a certain distance towards the personal best position and the global best position. Step 4: iteration: repeat steps 2 and 3 until a certain stopping condition is met (such as reaching the maximum number of iterations, the change in fitness value being less than a certain threshold, etc.). Step 5: build the minimum distance matrix: after the PSO algorithm converges, a globally optimal network configuration is obtained. According to this configuration, the minimum distance matrix is constructed. The "minimum distance matrix" here is actually a matrix representing the minimum cost (considering physical distance and communication delay) between nodes in the optimal network configuration. Each element D[i][j] in the matrix can represent the cost of the minimum cost path from node i to node j (in the optimal network configuration).

[0068] In the embodiment, the first determination module obtains the total bandwidth corresponding to the data collection network; the receiving path is a path for receiving data; and the forwarding path is a path for forwarding data. The first number of the receiving path and the second number of the forwarding path of each data collection node are determined; and the data table for distribution is a data table for querying the importance coefficient according to the first number and the second number. The importance coefficient of each data collection node in the minimum distance matrix is determined by querying the preset data table for distribution according to the first number and the second number, that is, the importance of each data collection node is determined. The distribution module is configured to distribute the total bandwidth according to the importance coefficient of each data collection node to obtain a distribution result; the data collection network collects the vehicle condition data of the vehicle according to the distribution result; and the conversion module is configured to perform text conversion on the vehicle data to obtain text data.

[0069] The technical scheme has the beneficial effects that a minimum distance matrix is constructed based on interval information and a particle swarm algorithm, an optimized data collection network is constructed, and a corresponding bandwidth is allocated to each data collection node in the data collection network, so that the network is further optimized, and the data collection rate and accuracy are improved.

[0070] According to some embodiments of the application, the various data collection sensors include an air flow sensor, an ABS sensor, a throttle position sensor, a crankshaft position sensor, a camshaft position sensor, a water temperature sensor, an oil pressure sensor, an oxygen sensor, a vehicle speed sensor, and a knock sensor.

[0071] The air flow sensor is used to measure the amount of air drawn into the engine and convert the information into an electrical signal input to the electronic control unit (ECU) as a reference signal for the injection time.

[0072] The ABS sensor is used to monitor the vehicle speed and feed back the wheel speed to the brake system during emergency braking, so that the brake system controls the rotation of the wheel to achieve the best braking effect.

[0073] The throttle position sensor is used to detect the opening degree of the throttle valve, measure the angle voltage signal of the throttle valve opening through a lever mechanism linked with the throttle valve, and transmit the signal to the ECU as a reference signal for the injection amount and ignition advance angle correction.

[0074] The crankshaft position sensor is used to detect the top dead center signal, the crankshaft angle signal, and the engine speed signal, and provide the signals to the ECU as reference signals for determining the ignition timing, the injection timing, and the working sequence.

[0075] The camshaft position sensor provides a crankshaft angle reference position signal as a main control signal for the injection timing control and the ignition timing control.

[0076] Water temperature sensor: installed on the cylinder block, used to detect the temperature information of the engine cooling water, and convert the information into an electrical signal to provide to the ECU.

[0077] Oil pressure sensor: used to detect the pressure of the oil, and send a warning signal when the pressure is not enough.

[0078] Oxygen sensor: installed on the exhaust pipe, used to detect the oxygen content in the engine exhaust, determine whether the gasoline and air are completely burned, and closed-loop control the fuel injection amount.

[0079] Speed sensor: detects the speed of the vehicle, provides a speed signal to the ECU for cruise control and speed limit fuel cut control, and is also the main control signal of the automatic transmission.

[0080] Knock sensor: detects the vibration of the engine cylinder to identify the engine knock condition for the electronic controller. When the engine knocks, the engine vibration is transmitted to the crystal through the mass in the sensor, and the piezoelectric crystal generates voltage on the two polar surfaces due to the pressure generated by the mass vibration, converting the vibration into a voltage signal output.

[0081] In addition, there are intake temperature sensor, fuel level sensor, coolant level sensor, battery level sensor, oil rail fuel pressure sensor, oil tank pressure sensor, transmission oil pressure sensor, collision sensor, ultrasonic ranging sensor, angular sonar (angular radar) sensor, brake pressure switch sensor, brake light switch sensor, distance sensor, intake manifold absolute pressure sensor, supercharging pressure sensor, cylinder combustion pressure sensor, tire pressure detection sensor, wheel speed sensor, deceleration sensor, electric seat sensor, headlight high-low beam control sensor and many other sensors, each of which undertakes different vehicle data collection tasks, and together ensures the safe, stable and efficient operation of the vehicle.

[0082] According to some embodiments of the present application, the first determination module is based on Nagios, Zabbix, PRTG NetworkMonitor network monitoring tools to obtain the total bandwidth corresponding to the data collection network in real time.

[0083] The working principle and beneficial effects of the above technical solution are as follows: Nagios is an open-source network and system monitoring tool that can monitor network traffic, server performance, application status, etc. It extends its functions through a plug-in system, including monitoring network bandwidth. Zabbix is a powerful enterprise-level open-source monitoring solution for monitoring various IT components, including network bandwidth. It can monitor network traffic, bandwidth utilization, and other indicators in real time, and provide rich chart and report functions. PRTG Network Monitor is a graphical network monitoring software that supports bandwidth monitoring, traffic analysis, and other functions. It provides an intuitive graphical interface to help administrators quickly understand the use of network bandwidth. Installation and configuration: Install Nagios, Zabbix, or PRTG Network Monitor on the key nodes of the data collection network. Configure the monitoring tool to connect to the network interface that needs to be monitored. Set monitoring indicators, including total bandwidth, bandwidth utilization, etc. Real-time monitoring: Use the real-time monitoring function of the monitoring tool to continuously collect network bandwidth data. The monitoring tool will automatically calculate and display the current total bandwidth usage. Based on the network monitoring tool, the total bandwidth corresponding to the data collection network is obtained in real time, improving the accuracy of data acquisition.

[0084] According to some embodiments of the application, the distribution module distributes the total bandwidth according to the importance coefficient of each data collection node to obtain a distribution result, including:

[0085] Determine the proportion information according to the importance coefficient of each data collection node;

[0086] Distribute the total bandwidth according to the proportion information to obtain a distribution result.

[0087] The working principle and beneficial effects of the above technical solution are as follows: Determine the proportion information according to the importance coefficient of each data collection node; distribute the total bandwidth according to the proportion information to obtain a distribution result, which realizes accurate bandwidth distribution for each data collection node and improves the data collection rate.

[0088] As shown in Figure 3 According to some embodiments of the application, the first judgment module obtains a vehicle condition judgment model, including:

[0089] The acquisition module is used to acquire historical vehicle condition data of the vehicle and perform text conversion to obtain historical text data of the vehicle;

[0090] The data preprocessing module is used to clean and format the historical text data, remove noise and irrelevant information, standardize the data format, and obtain preprocessed data;

[0091] An NLP processing module is configured to perform word segmentation processing on the preprocessed data using a professional word segmentation dictionary, to disassemble the preprocessed data into independent word groups, to classify the word groups, and to obtain classified data;

[0092] A neural network training module is configured to train a combined model of a convolutional neural network and a recurrent neural network based on the classified data, to obtain a trained model;

[0093] An evaluation module is configured to perform model evaluation on the trained model based on test data, and to obtain a vehicle condition judgment model when the evaluation is qualified.

[0094] The working principle and beneficial effects of the above technical solution are as follows: The acquisition module is responsible for acquiring historical vehicle condition data of a vehicle from various data sources (such as a database of a vehicle manufacturer, a vehicle OBD interface, a third-party data provider, etc.), and converting these data into a text format for subsequent processing. Data is acquired from data sources using API interfaces or data scraping techniques. Data may exist in binary, XML, JSON, or other formats and needs to be converted into a pure text format. During the conversion process, data may need to be parsed and organized to ensure that all necessary information is correctly extracted and formatted. The data preprocessing module cleans the historical text data, removes noise (such as irrelevant characters, duplicate information, etc.), and standardizes the data format to facilitate subsequent processing. Convert data types to ensure consistency and accuracy of data. The NLP processing module uses a professional word segmentation dictionary to perform word segmentation processing on the preprocessed text data, disassembles the text into independent word groups, and classifies these word groups. Use natural language processing (NLP) libraries (such as Python's NLTK, jieba, etc.) for word segmentation. Customize or use existing word segmentation dictionaries to ensure accurate recognition of vehicle-related professional terms. Use part-of-speech tagging, named entity recognition, and other techniques to further analyze word groups. Design a classification algorithm or rule to classify word groups into different categories (such as engine condition, tire condition, fuel consumption, etc.). The neural network training module trains a combined model of a convolutional neural network (CNN) and a recurrent neural network (RNN) based on the classified data to extract features and learn the relationship between vehicle conditions and text data. Design a hybrid model that combines CNN and RNN. CNN is used to capture local features of text (such as relationships between word groups), and RNN (especially LSTM or GRU) is used to handle long-term dependencies in sequential data. Divide the classified data into training and validation sets. Use a deep learning framework (such as TensorFlow, PyTorch) to build and train the model. Adjust model parameters such as number of layers, number of neurons, learning rate, etc. to optimize model performance. The evaluation module evaluates the trained model using test data to verify its accuracy and reliability. If the evaluation result is qualified, a vehicle condition judgment model that can be used for actual vehicle condition judgment is obtained. This facilitates obtaining an accurate vehicle condition judgment model.

[0095] According to some embodiments of the present application, the NLP processing module comprises:

[0096] A second construction module is configured to obtain vehicle component information and construct a professional segmentation dictionary for vehicle components based on the vehicle component information;

[0097] A segmentation module is configured to perform segmentation processing on the preprocessed data based on the professional segmentation dictionary, and to disassemble the preprocessed data into independent word groups;

[0098] A classification module is configured to perform part-of-speech tagging on the word groups, and to classify the word groups into categories based on the tagging results to obtain classified data.

[0099] The working principle and beneficial effects of the above technical solution are as follows: The second construction module is responsible for obtaining vehicle component information and constructing a professional segmentation dictionary for vehicle components based on the information. This dictionary will be used for subsequent segmentation processing to ensure accurate identification of vehicle-related professional terms. Implementation: Data collection: Collect vehicle component information from vehicle manufacturers, repair manuals, technical documents, or professional databases. Dictionary construction: Construct a professional segmentation dictionary based on the sorted component information. Each entry in the dictionary should be a vehicle component-related professional term. Dictionary update: As new models and technologies develop, update the segmentation dictionary regularly to ensure its accuracy and comprehensiveness. The segmentation module performs segmentation processing on the preprocessed text data based on the professional segmentation dictionary, disassembling the text into independent word groups. Implementation: Load dictionary: Load the constructed professional segmentation dictionary before segmentation processing. Segmentation algorithm: Use a segmentation algorithm (such as maximum forward matching, minimum segmentation, etc.) combined with the professional segmentation dictionary to segment the text. The algorithm should prioritize the entries in the dictionary to ensure accurate identification of professional terms. Result output: Output the segmentation results as a list of independent word groups for subsequent processing. The classification module performs part-of-speech tagging on the word groups output by the segmentation module and classifies the word groups into categories based on the tagging results to obtain classified data. Implementation: Part-of-speech tagging: Use a part-of-speech tagging tool (such as HanLP, jieba, etc.) to perform part-of-speech tagging on the segmented word groups. The tagging results will include the part-of-speech (such as noun, verb, adjective, etc.) of each word group. Category classification: Based on the part-of-speech tagging results and pre-set category rules, classify the word groups into categories. For example, noun category word groups can be classified into vehicle components, fault types, etc.; verb category word groups can be classified into operation behaviors, fault manifestations, etc. Rule formulation: The rules for category classification can be formulated according to actual needs, including but not limited to keyword-based matching, word-based screening, context-based understanding, etc. Result output: Output the classified data in a structured format, such as JSON, XML, or a database table, for subsequent processing and analysis.

[0100] According to some embodiments of the application, the data preprocessing module comprises:

[0101] The cleaning module is configured to obtain type information of historical text data, query a preset data cleaning rule file according to the type information, obtain corresponding data cleaning rules, and perform data cleaning according to the corresponding data cleaning rules to obtain cleaned data.

[0102] The formatting processing module is configured to perform data processing on the cleaned data according to an adaptive regular expression to obtain preprocessed data.

[0103] The working principle and beneficial effects of the above technical solution are as follows: The cleaning module obtains type information (such as log type, report type, etc.) of historical text data, then queries a preset data cleaning rule file according to the type information, obtains corresponding data cleaning rules, and applies the rules to clean data to remove noise, irrelevant information, and error data, and obtain cleaned data. Implementation: Type information acquisition: Extract type information from the metadata, file name, file path, or other identifiers of the text data. Rule file query: According to the obtained type information, find the corresponding cleaning rules in the preset data cleaning rule file. These rule files can be in XML, JSON, YAML, etc. format, which contains cleaning logic and parameters for different types of data. Rule application: Load the found cleaning rules and use these rules to process the original text data. The processing process may include but is not limited to: removing useless characters, spaces, and line breaks. Replacing incorrect data or inconsistent data. Deleting or correcting incomplete records. Identifying and deleting duplicate data. Result output: Save the cleaned data as a new file or data structure for subsequent modules. The formatting processing module uses the adapted regular expression to further process the cleaned data to ensure that the data format meets the requirements of subsequent analysis or model training. Formatting processing may include data format standardization, field extraction and reorganization, etc. Implementation: Regular expression design: According to the characteristics of the data and the needs of subsequent processing, design the appropriate regular expression. These expressions are used to match and extract specific information in the text, such as dates, times, and numerical values. Data processing: Use regular expressions to match and extract operations on cleaned data, separate the required information from the text. According to the needs, format conversion or standardization processing is performed on the extracted information, such as converting date strings to a unified date format, converting numerical value strings to floating point numbers or integers, etc. The data is reorganized into a new data structure according to the predetermined format, such as a JSON object, a database record, etc. Result output: The formatted data is output as preprocessed data for subsequent modules such as NLP processing module, neural network training module, etc. Through the cooperative work of the two modules, the system can effectively convert the original historical text data into clean, uniform format, and suitable for subsequent processing. The preprocessed data lays a solid foundation for subsequent data analysis and model training.

[0104] According to some embodiments of the application, the classification module comprises:

[0105] The second determination module is configured to determine a word vector of each word group, match the word vector with a standard word vector, and perform part-of-speech tagging on the word group according to a part of speech corresponding to the matched standard word vector;

[0106] The collation module is used for classifying and collating word groups of the same part of speech to obtain classified data.

[0107] The working principle and beneficial effects of the above technical solution are as follows: the second determination module determines the word vector for each word group and matches the word vector with the word vector in the standard word vector library. According to the matching result, the part of speech of the word group is annotated by using the part of speech corresponding to the matched standard word vector. Implementation: word vector determination: a word vector is determined for each word group obtained after word segmentation by using word embedding technology (such as Word2Vec, GloVe, BERT, etc.). The word vector is a numerical representation of the word group in the vector space and can capture the semantic relationship between word groups. Word vector matching: the calculated word vector is subjected to similarity calculation (such as cosine similarity) with the word vector in the pre-trained standard word vector library. The standard word vector library contains a large number of word groups of known parts of speech and the corresponding word vectors. Part of speech annotation: according to the similarity calculation result, the standard word vector most similar to the word group to be annotated is found, and the part of speech corresponding to the standard word vector is obtained. Then, the part of speech is taken as the part of speech annotation result of the word group to be annotated. The collation module is responsible for classifying and collating the word groups with the part of speech annotated by the second determination module. According to the part of speech of the word group and other possible context information, the word groups with the same or similar categories are classified together to form classified data. Implementation: category definition: the category system for classification needs to be defined. These categories can be divided based on vehicle components, fault types, operation behaviors, and other actual business scenarios. Classification rule making: according to the part of speech of the word group and possible context information (such as the part of speech of adjacent word groups, semantic relationship, etc.), the rules for classifying the word groups are made. These rules can be rule-based classification algorithms (such as decision trees, rule engines, etc.) or statistical-based classification algorithms (such as Naive Bayes, support vector machines, etc.). Word group classification: the word groups with the annotated part of speech are classified into the corresponding categories by applying the made classification rules. Result output: the classified data is arranged in a structured format (such as JSON, XML, database table, etc.) and output as part of the preprocessed data for use by subsequent modules. Through the cooperative work of the second determination module and the collation module, the system can realize the part of speech annotation and category classification and collation of the word groups after word segmentation, and provide a more accurate and useful data basis for subsequent data analysis and model training.

[0108] According to some embodiments of the present application, the evaluation module comprises:

[0109] The calculation module is configured to obtain H test text data included in the test data and sample vehicle condition determination results corresponding to the H test text data, input the H test text data into the trained model respectively to output H corresponding predicted vehicle condition determination results, and calculate matching degrees Q of the sample vehicle condition determination results corresponding to the H test text data and the H predicted vehicle condition determination results.

[0110]

[0111] wherein, S a is the a th predicted vehicle condition determination result output by the training model; S a is the sample vehicle condition determination result of the a th test text data input into the training model;

[0112] The second judging module is configured to judge whether the matching degree is greater than a preset matching degree. If it is determined that the matching degree is greater than the preset matching degree, it indicates that the evaluation is qualified, and a vehicle condition determination model is obtained. Otherwise, it indicates that the evaluation is unqualified, and the training of the training model needs to be continued.

[0113] The working principle and beneficial effects of the above technical solution are as follows: The H test text data and the sample vehicle condition determination results corresponding to the H test text data included in the test data are obtained based on the calculation module. The H test text data are input into the training model respectively to output H corresponding predicted vehicle condition determination results. The matching degrees of the sample vehicle condition determination results corresponding to the H test text data and the H predicted vehicle condition determination results are calculated. It is judged whether the matching degree is greater than a preset matching degree. If it is determined that the matching degree is greater than the preset matching degree, it indicates that the evaluation is qualified, and a vehicle condition determination model is obtained. Otherwise, it indicates that the evaluation is unqualified, and the training of the training model needs to be continued. The model obtained by training is accurately judged.

[0114] Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the claims of the present application and their equivalents, the present application also intends to include these modifications and variations.

Claims

1. An intelligent system for automated determination of the condition of a vehicle, characterized in that The application relates to a vehicle condition data collection method and device. The application comprises: A collection module is used for collecting vehicle condition data and performing text conversion to obtain text data; A first judgment module is used for inputting the text data into a pre-trained vehicle condition judgment model to output a vehicle condition judgment result; The collection module comprises: A first construction module is used for: Determining interval information between each data collection sensor for collecting vehicle condition data; the interval information comprises distance information obtained by GPS positioning, communication delay and network hop count; According to the interval information and a particle swarm algorithm, a minimum distance matrix is constructed, the minimum distance matrix is a square matrix, each data collection node in the minimum distance matrix represents the minimum collection cost between corresponding two data collection sensors, and the minimum distance matrix is used as a data collection network; A first determination module is used for: Obtaining a total bandwidth corresponding to the data collection network; Determining a first number of receiving paths and a second number of forwarding paths of each data collection node; According to the first number and the second number, a preset allocation data table is inquired to determine an importance coefficient of each data collection node in the minimum distance matrix; An allocation module is used for allocating the total bandwidth according to the importance coefficient of each data collection node to obtain an allocation result; the data collection network collects vehicle condition data according to the allocation result; A conversion module is used for performing text conversion on vehicle data to obtain text data; The first judgment module obtains a vehicle condition judgment model, which comprises: An acquisition module is used for acquiring historical vehicle condition data of a vehicle and performing text conversion to obtain historical text data of the vehicle; A data preprocessing module is used for cleaning and formatting the historical text data, removing noise and irrelevant information, standardizing the data format and obtaining preprocessed data; An NLP processing module is used for performing word segmentation processing on the preprocessed data by using a professional word segmentation dictionary, decomposing the preprocessed data into independent word groups, classifying the word groups and obtaining classified data; A neural network training module is used for training a combined model of a convolutional neural network and a recurrent neural network based on the classified data to obtain a training model; An evaluation module is used for performing model evaluation on the training model based on test data, and when the evaluation is qualified, a vehicle condition judgment model is obtained; The evaluation module comprises: wherein S a is the a-th predicted vehicle condition determination result output by the trained model; S a is the sample vehicle condition determination result of the a-th test text data input to the trained model. A calculation module is used for acquiring H test text data included in test data and sample vehicle condition judgment results corresponding to the H test text data, inputting the H test text data into the training model to output H corresponding predicted vehicle condition judgment results, and calculating matching degrees Q of the sample vehicle condition judgment results corresponding to the H test text data and the H predicted vehicle condition judgment results; A second judgment module is used for judging whether the matching degree is greater than a preset matching degree; when it is determined that the matching degree is greater than the preset matching degree, it is indicated that the evaluation is qualified, and a vehicle condition judgment model is obtained; otherwise, it is indicated that the evaluation is unqualified, and the training model needs to be continuously trained. According to the interval information and the particle swarm algorithm, a minimum distance matrix is constructed, including: step 1: initializing a node list: listing all data acquisition sensors, i.e. all nodes; initializing an interval matrix: creating a matrix, wherein element M[i][j] represents the physical distance between node i and node j; initializing a particle swarm: each particle represents a network configuration, i.e. the connection mode between nodes; step 2: evaluating fitness: calculating the total cost for each particle, traversing the interval matrix, and selecting paths and connections according to the configuration of the particle, and then accumulating the cost on these paths; taking the total cost as the output of the fitness function, and the goal is to find the configuration with the lowest total cost; step 3: updating particles: personal best position: updating the personal best position of each particle; global best position: finding the global best position among all particles; speed and position update: updating the speed and position of each particle according to the rules of PSO; step 4: repeating steps 2 and 3 until the maximum number of iterations is reached; step 5: after the PSO algorithm converges, a globally optimal network configuration is obtained, i.e. a minimum distance matrix is constructed.

2. The intelligent system for automated judgment of the condition of a vehicle of claim 1, wherein, The various data acquisition sensors include an air flow sensor, an ABS sensor, a throttle position sensor, a crankshaft position sensor, a camshaft position sensor, a water temperature sensor, an oil pressure sensor, an oxygen sensor, a vehicle speed sensor, and a knock sensor.

3. The intelligent system for automated judgment of the condition of a vehicle of claim 1, wherein, The first determination module obtains the total bandwidth corresponding to the data acquisition network in real time based on Nagios, Zabbix, PRTG Network Monitor network monitoring tools.

4. The intelligent system for automated judgment of the condition of a vehicle of claim 1, wherein, The distribution module distributes the total bandwidth according to the importance coefficient of each data acquisition node to obtain a distribution result, including: determining proportion information according to the importance coefficient of each data acquisition node; distributing the total bandwidth according to the proportion information to obtain a distribution result.

5. The intelligent system for automated judgment of the condition of a vehicle of claim 1, wherein, The NLP processing module includes: A second construction module is configured to obtain vehicle component information and construct a professional segmentation dictionary for vehicle components based on the vehicle component information. A segmentation module is configured to perform segmentation processing on the preprocessed data based on the professional segmentation dictionary, and to decompose the preprocessed data into independent word groups. A classification module is configured to perform part-of-speech tagging on the word groups, classify the word groups according to the tagging results, and obtain classified data.

6. The intelligent system of automated judgment of the car condition according to claim 1, characterized in that, The data preprocessing module includes: A cleaning module is configured to obtain type information of historical text data, query a preset data cleaning rule file according to the type information, obtain corresponding data cleaning rules, and perform data cleaning according to the corresponding data cleaning rules to obtain cleaned data. A formatting processing module is configured to perform data processing on the cleaned data according to an adaptive regular expression to obtain preprocessed data.

7. The intelligent system for automated judgment of the condition of a vehicle of claim 5, wherein, The classification module includes: A second determination module is configured to determine the word vector of each word group, match the word vector with a standard word vector, and perform part-of-speech tagging on the word group according to the part-of-speech corresponding to the matched standard word vector. An arrangement module is configured to classify and arrange word groups of the same part-of-speech to obtain classified data.

Images