Anti-theft starting battery control method and system based on multi-source data

Abstract

The invention relates to the field of vehicle battery control, in particular to an anti-theft starting battery control method and system based on multi-source data. The method comprises the following steps: acquiring real-time vehicle multi-source monitoring data; performing in-vehicle unmanned inference on the real-time vehicle multi-source monitoring data, generating an in-vehicle unmanned signal, and entering a vehicle unmanned mode; when the vehicle is in the vehicle unmanned mode, obtaining vehicle component voltage state data; performing minimum-voltage continuous power supply based on the vehicle component voltage state data, and acquiring vehicle real-time anti-theft state data; and vehicle starting voltage real-time monitoring is carried out according to the vehicle real-time anti-theft state data, sudden change recognition is carried out, and starting voltage sudden change data are extracted. Through multi-source data analysis, when it is monitored that the vehicle is stolen, battery output is cut off, power is continuously supplied in other time, the safety anti-theft performance of the vehicle is improved, and daily vehicle data are protected.

Description

Technical Field

[0001] The present invention relates to the field of vehicle battery control, and in particular to an anti-theft starting battery control method and system based on multi-source data. Background Art

[0002] With the continuous improvement of the level of automobile intelligence, automobile anti-theft technology is becoming more and more complicated. Traditional anti-theft methods based on manual intervention often have many limitations: once the owner leaves the vehicle, the power output needs to be manually turned off, which can easily lead to the accidental loss of vehicle data and systems; and the anti-theft effect is also very limited, and it is difficult to detect and prevent car theft in time. In actual automobile usage scenarios, the battery power supply system is the core power supply for vehicle-mounted electronic equipment, and its safety and reliability are crucial to the normal operation of the entire vehicle. However, most of the existing anti-theft measures only focus on cutting off the battery output and ignore the protection of vehicle-mounted data.

[0003] Therefore, it is urgent to study an intelligent anti-theft starting battery control method based on multi-source data to cope with the increasingly serious problem of car theft while taking into account the normal operation of the vehicle system. Summary of the invention

[0004] In order to solve the above technical problems, the present invention proposes an anti-theft starting battery control method and system based on multi-source data to solve at least one of the above technical problems.

[0005] To achieve the above object, the present invention provides an anti-theft starting battery control method based on multi-source data, comprising the following steps: Step S1: acquiring real-time vehicle multi-source monitoring data; performing inference on the real-time vehicle multi-source monitoring data that no one is in the vehicle, generating a vehicle no-one signal, and entering the vehicle no-one mode; Step S2: when the vehicle is in the vehicle unmanned mode, obtaining the voltage status data of the vehicle components; providing continuous power supply at the lowest voltage based on the voltage status data of the vehicle components, and obtaining the real-time anti-theft status data of the vehicle; Step S3: monitoring the vehicle starting voltage in real time according to the vehicle real-time anti-theft status data, identifying mutations, and extracting the starting voltage mutation data; Step S4: Acquire historical vehicle owner startup behavior data; perform personalized behavior habit analysis on the historical vehicle owner startup behavior data to construct a vehicle owner startup behavior model; Step S5: calculating the abnormal behavior probability of the starting voltage mutation data based on the vehicle owner's starting behavior model to obtain an abnormal evaluation result; Step S6: When the abnormal evaluation result is that the vehicle is abnormally started, the current output of the anti-theft starting battery is cut off, and a vehicle abnormal start warning signal is generated.

[0006] The present invention realizes real-time monitoring of vehicle status through multi-source monitoring data acquisition, which helps to find abnormal situations in real time. It realizes the inference that there is no one in the car, ensures the safety of the car, and avoids unnecessary safety hazards. It ensures that the minimum voltage required for the normal operation of vehicle components is continuously powered to ensure the normal operation of the vehicle system. Real-time anti-theft status data is obtained to provide basic data support for the subsequent monitoring of abnormal situations. Through real-time monitoring and mutation identification, the abnormal situation of the vehicle starting voltage is timely discovered, and the monitoring accuracy of the vehicle status is improved. The safety of the vehicle starting process is guaranteed to avoid the safety hazards caused by the starting problem. The owner's starting behavior model is constructed through historical data analysis to understand the owner's behavior habits and provide a basis for the identification of abnormal behavior. Personalized behavior habit analysis helps to predict the normal behavior pattern of the owner, so that abnormal behavior is easier to detect. The probability of abnormal behavior is calculated based on the owner's starting behavior model, the safety of vehicle starting is evaluated, and the early warning capability of abnormal situations is improved. Abnormal behavior is discovered in time, early warning measures are taken, and the safety of the vehicle and the owner is guaranteed. Once the vehicle is found to be abnormally started, the current output of the anti-theft starting battery is immediately cut off to ensure the safety of the vehicle. Generate abnormal start-up warning signals to alert vehicle owners and relevant managers to ensure that timely measures are taken to deal with abnormal situations and effectively prevent the risk of vehicle theft.

[0007] Preferably, step S1 comprises the following steps: Step S11: acquiring real-time vehicle multi-source monitoring data, wherein the real-time vehicle multi-source monitoring data includes sonar sensor data, seat pressure sensor data and air monitoring sensor data; Step S12: filtering and denoising the sonar sensor data to obtain filtered audio data; Step S13: filtering abnormal pressure values ​​of the seat pressure sensor data to generate abnormal value filtered pressure data; Step S14: performing parameter standardization on the air monitoring sensor data to obtain standardized air monitoring data; Step S15: Inferring that there is no one in the vehicle based on the filtered audio data, the outlier filtered pressure data, and the standardized air monitoring data, generating a signal that there is no one in the vehicle, and entering the vehicle unmanned mode.

[0008] The present invention realizes comprehensive monitoring of the vehicle's internal environment through the comprehensive application of sonar sensor data, seat pressure sensor data and air monitoring sensor data. Combining multi-source monitoring data helps to fully understand the internal situation of the vehicle and improve the accuracy of abnormal situation detection. Through filtering and noise reduction processing, the accuracy of sonar sensor data is improved, interference is reduced, and it is helpful to accurately identify sound signals and obtain filtered audio data, providing a clearer data basis for the analysis of the vehicle's internal environmental sound. Through outlier filtering, interference data is excluded, the accuracy of seat pressure sensor data is ensured, the monitoring effect of seat usage is improved, and outlier filtered pressure data is generated. It helps to detect abnormal seat usage and provide important clues about the status of occupants in the vehicle. Standardized air monitoring data makes data from different time points or different vehicles comparable, which facilitates data analysis and comparison. Obtaining standardized air monitoring data helps to assess the air quality in the vehicle and promptly discover air pollution problems. By inferring that there is no one in the vehicle, the vehicle enters unmanned mode to ensure the safety of the vehicle interior, prevent unauthorized entry into the vehicle, and strengthen the vehicle's anti-theft control. Combined with filtered audio data, outlier filtered pressure data and standardized air monitoring data, it can comprehensively judge the situation in the vehicle and improve the accuracy of detecting abnormal situations.

[0009] Preferably, the specific steps of step S15 are: Define the sliding time window monitoring length; The real-time vehicle multi-source monitoring data is divided into multiple time windows according to the monitoring length of the sliding time window, so as to obtain multi-source monitoring data of multiple time windows; Analyze the changes in the sound intensity in the car on the filtered audio data to obtain the sound intensity change data; The number of people in the car is determined based on the sound intensity change data, and the number of people in the car is estimated by the sound; Perform seat pressure detection on each seat for the outlier filtered pressure data. When the seat pressure is 0, obtain the empty seat data. Conduct breath detection analysis on standardized air monitoring data and extract breath detection data; Based on the breathing detection data, the number of people in the car is determined, and the number of people in the car is estimated by breathing; The number of people in the car estimated by sound, the number of empty seats, and the number of people in the car estimated by breathing are used to infer that there is no one in the car. When the number of people in the car estimated by sound is 0, the empty seats are all empty, and the number of people in the car estimated by breathing is 0, then a signal of no one in the car is obtained; The vehicle unmanned inference operation is repeated based on multi-source monitoring data from multiple time windows. When multi-source monitoring data from two consecutive time windows generates a vehicle unmanned signal, the vehicle enters unmanned mode.

[0010] The present invention realizes continuous monitoring and analysis of real-time data by setting a suitable time window length, thereby ensuring the continuity and integrity of the data. Dividing the data into multiple time windows is helpful to compare and analyze the data at different times and identify potential abnormal situations. Analyzing the change in sound intensity is helpful to identify the activities in the car and prompt the change in the number of people in the car. The number of people in the car is estimated by the change in sound intensity, thereby realizing real-time monitoring of the situation in the car. Detecting the seat pressure one by one and identifying the empty seats is helpful to determine the occupancy of the seats. Through the breathing detection data, the air quality in the car can be evaluated and the existing health risks can be found. The number of people in the car is judged based on the breathing detection data, thereby increasing the accuracy of the assessment of the number of passengers in the car. Inferring that there is no one in the car by integrating multiple data sources helps to accurately determine whether there is anyone in the car and strengthen the anti-theft control of the vehicle. By continuously monitoring multiple time window data, the continuous monitoring of the state in the car is strengthened, and the accuracy of the unmanned state of the vehicle is improved. When two consecutive time windows confirm that there is no one in the car, the vehicle unmanned mode is automatically entered to strengthen the safety protection measures of the vehicle and prevent unauthorized persons from entering the vehicle.

[0011] Preferably, step S2 specifically comprises the following steps: Step S21: when the vehicle is in the vehicle unmanned mode, obtaining the voltage status data of the vehicle components; Step S22: performing component cluster analysis on the vehicle component voltage state data to obtain a plurality of vehicle component type data; Step S23: performing minimum voltage requirement analysis on multiple vehicle component type data to obtain the minimum voltage requirement value of each component; Step S24: Based on the minimum voltage requirement value of each component, the anti-theft starting battery is used to continuously supply power at the minimum voltage, and real-time anti-theft status data of the vehicle is obtained.

[0012] The present invention helps to monitor the voltage conditions of various components of the vehicle by acquiring the voltage status data of vehicle components in the unmanned mode of the vehicle, and provides basic data for subsequent voltage demand analysis and continuous power supply. By performing component cluster analysis on the voltage status data of vehicle components, the vehicle components can be classified by type, which helps to conduct targeted analysis and processing of the voltage requirements of different types of components. By performing a minimum voltage requirement analysis on multiple vehicle component type data to obtain the minimum voltage requirement value of each component, the minimum voltage required for each component can be accurately determined, providing specific guidance for continuous battery power supply. Based on the minimum voltage requirement value of each component, the anti-theft starting battery is used to continuously supply the minimum voltage, and the real-time anti-theft status data of the vehicle is obtained, which can ensure that each component has sufficient voltage supply during operation, while ensuring the normal operation of the vehicle's anti-theft system.

[0013] Preferably, step S3 specifically comprises the following steps: Step S31: Perform real-time monitoring of the vehicle starting voltage according to the real-time anti-theft status data of the vehicle, and generate anti-theft status starting voltage monitoring data; Step S32: performing sliding average processing on the anti-theft state starting voltage monitoring data to obtain smoothed starting voltage monitoring data; Step S33: fitting the time series change distribution of the smoothed starting voltage monitoring data to construct a starting voltage change curve; Step S34: performing mutation identification on the starting voltage variation curve and extracting the starting voltage mutation data.

[0014] The present invention monitors the vehicle starting voltage according to real-time anti-theft status data, can detect abnormal conditions in time, ensure the safety of vehicle starting, obtain smooth starting voltage monitoring data through sliding average processing, help to remove noise and mutations, improve data stability and reliability, construct a starting voltage change curve through time series change distribution fitting, help understand the changing trend of the starting voltage, provide more intuitive data presentation for subsequent analysis, perform mutation identification on the starting voltage change curve, extract mutation data, help to identify existing problems or abnormal conditions, and further strengthen the monitoring and control of the vehicle starting voltage.

[0015] Preferably, the specific steps of step S4 are: Step S41: Obtaining historical vehicle owner startup behavior data; Step S42: extracting the timestamp of each startup based on the historical startup behavior data of the vehicle owner; Step S43: performing a time series startup behavior feature analysis on the historical startup behavior data of the vehicle owner based on the timestamp of each startup, thereby obtaining startup behavior time series feature data; Step S44: performing periodic start frequency distribution calculation on the start behavior time series characteristic data to obtain periodic start frequency distribution data; Step S45: Performing personalized behavior habit analysis on historical vehicle owner startup behavior data to generate vehicle owner startup behavior habit data; Step S46: mining the behavior characteristics of the periodic start frequency distribution data and the car owner's start behavior habit data to construct a car owner's start behavior model. The present invention helps to accumulate data for subsequent analysis and model building by acquiring historical car owners' startup behavior data. The timestamp of each startup is extracted based on historical data to provide a time reference for subsequent timing analysis and frequency calculation. Extracting the characteristic data of each startup through timing analysis helps to understand the startup behavior patterns and laws of the car owner. By calculating the periodic startup frequency distribution, the startup frequency law of the car owner can be revealed to provide data support for model building. Personalized behavior habit analysis of historical car owners' startup behavior data helps to understand the personalized operating habits of the car owner. Based on the periodic startup frequency and personalized behavior habit data, behavioral feature mining is performed to build a car owner startup behavior model, which can be used to predict and identify abnormal behavior and improve the intelligence level of the anti-theft starting battery control system.

[0016] Preferably, the specific steps of step S5 are: Step S51: calculating the mutation rate of the startup voltage mutation data to generate a startup voltage mutation rate parameter; Step S52: judging the starting voltage mutation rate parameter based on the preset starting power supply change threshold value, and when the starting voltage mutation rate parameter is greater than or equal to the preset starting power supply change threshold value, determining that the vehicle is in the starting state, and obtaining the current vehicle starting behavior data; Step S53: calculating the abnormal behavior probability of the current vehicle startup behavior data according to the vehicle owner startup behavior model to generate current startup abnormal probability data; Step S54: Perform a comprehensive abnormality assessment on the previous startup abnormality probability data to obtain an abnormality assessment result.

[0017] The present invention can monitor the voltage changes during the vehicle startup process by calculating the voltage mutation rate parameter, providing a data basis for subsequent analysis. By judging whether the voltage mutation rate parameter exceeds the preset threshold, it is possible to accurately determine whether the vehicle is in the startup state, and then obtain relevant startup behavior data. According to the owner's startup behavior model, the abnormal probability of the current vehicle startup behavior is calculated, which helps to detect unusual startup behavior and improve the perception of potential risks. By combining the above data, a comprehensive evaluation result of the current startup abnormality probability is obtained, further improving the ability to identify and evaluate abnormal behavior.

[0018] Preferably, the specific steps of step S6 are: Step S61: When the abnormal evaluation result is that the vehicle is started normally, start the vehicle normally; Step S62: When the abnormal evaluation result is abnormal vehicle start, the current output of the anti-theft start battery is cut off, and a vehicle abnormal start warning signal is generated; Step S63: Perform deep learning evolution of anti-theft startup behavior based on the abnormal vehicle startup warning signal to build a battery anti-theft control strategy.

[0019] When the abnormal evaluation result of the present invention is normal startup, the system can quickly execute the normal startup procedure to ensure that the vehicle can start smoothly, thereby improving user experience and vehicle reliability. When the abnormal evaluation result is abnormal startup, the system immediately cuts off the current output of the anti-theft starting battery to prevent the vehicle from continuing to start, and generates an early warning signal to remind relevant personnel to pay attention to the abnormal situation. The generation of the early warning signal can notify relevant personnel in real time so that they can take timely action to ensure vehicle safety. By performing deep learning evolution on the abnormal startup signal, the system can continuously optimize the anti-theft startup control strategy, improve the ability to respond to various abnormal situations, and further enhance the safety of the vehicle. Constructing the evolution process of the battery anti-theft control strategy can make the system more adaptive, and can flexibly adjust the anti-theft control strategy according to actual conditions, thereby improving the robustness and intelligence of the system.

[0020] In this specification, an anti-theft starting battery control system based on multi-source data is provided, which is used to execute the anti-theft starting battery control method based on multi-source data as described above, including: The unmanned inference module is used to obtain real-time vehicle multi-source monitoring data; perform inference on the real-time vehicle multi-source monitoring data to determine that there is no one in the vehicle, generate an unmanned signal in the vehicle, and enter the vehicle unmanned mode; A low-voltage power supply module is used to obtain the voltage status data of vehicle components when the vehicle is in the unmanned mode; provide continuous power supply at the lowest voltage based on the voltage status data of vehicle components, and obtain the real-time anti-theft status data of the vehicle; A mutation identification module is used to monitor the vehicle starting voltage in real time according to the vehicle's real-time anti-theft status data, identify mutations, and extract the starting voltage mutation data; The startup behavior module is used to obtain the historical startup behavior data of the car owner; conduct personalized behavior habit analysis on the historical startup behavior data of the car owner and build the startup behavior model of the car owner; An abnormality assessment module is used to calculate the abnormal behavior probability of the starting voltage mutation data based on the vehicle owner's starting behavior model to obtain an abnormality assessment result; The power-off anti-theft module is used to cut off the current output of the anti-theft starting battery and generate a warning signal for abnormal vehicle start when the abnormal evaluation result is abnormal vehicle start.

[0021] The present invention obtains real-time vehicle multi-source monitoring data, generates an in-vehicle no-man signal by inferring that there is no one in the vehicle, enters the vehicle no-man mode, realizes real-time monitoring and response to the situation in the vehicle, obtains vehicle component voltage status data, realizes monitoring of the vehicle power supply status, performs continuous power supply at the minimum voltage based on the voltage data, obtains the vehicle real-time anti-theft status data, ensures the real-time anti-theft status of the vehicle in the no-man mode, monitors the vehicle starting voltage in real time, identifies mutations and extracts relevant data, which is helpful to capture abnormal situations in the starting process, obtains historical car owner starting behavior data, performs personalized behavior habit analysis, constructs a car owner starting behavior model, provides a basis for the evaluation of subsequent starting behaviors, calculates the abnormal behavior probability of the starting voltage mutation data based on the car owner starting behavior model, obtains abnormal evaluation results, improves the recognition and response capabilities of abnormal situations, and when the abnormal evaluation result is abnormal vehicle startup, cuts off the current output of the anti-theft starting battery, generates an early warning signal, responds quickly to the abnormal situation, and ensures vehicle safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 A schematic flow chart of the steps of an anti-theft starting battery control method based on multi-source data of the present invention; Figure 2 Detailed implementation flow chart of step S1; Figure 3 Detailed implementation flow chart of step S2; Figure 4 Detailed implementation flow chart of step S3. DETAILED DESCRIPTION

[0023] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0024] The example of this application provides an anti-theft starting battery control method and system based on multi-source data. The execution subject of the anti-theft starting battery control method and system based on multi-source data includes but is not limited to: mechanical equipment, data processing platform, cloud server node, network upload device, etc. equipped with the system can be regarded as the general computing node of this application, and the data processing platform includes but is not limited to: at least one of an audio image management system, an information management system, and a cloud data management system.

[0025] See also Figures 1 to 4 The present invention provides an anti-theft starting battery control method based on multi-source data, the method comprising the following steps: Step S1: acquiring real-time vehicle multi-source monitoring data; performing inference on the real-time vehicle multi-source monitoring data that no one is in the vehicle, generating a vehicle no-one signal, and entering the vehicle no-one mode; Step S2: when the vehicle is in the vehicle unmanned mode, obtaining the voltage status data of the vehicle components; providing continuous power supply at the lowest voltage based on the voltage status data of the vehicle components, and obtaining the real-time anti-theft status data of the vehicle; Step S3: monitoring the vehicle starting voltage in real time according to the vehicle real-time anti-theft status data, identifying mutations, and extracting the starting voltage mutation data; Step S4: Acquire historical vehicle owner startup behavior data; perform personalized behavior habit analysis on the historical vehicle owner startup behavior data to construct a vehicle owner startup behavior model; Step S5: calculating the abnormal behavior probability of the starting voltage mutation data based on the vehicle owner's starting behavior model to obtain an abnormal evaluation result; Step S6: When the abnormal evaluation result is that the vehicle is abnormally started, the current output of the anti-theft starting battery is cut off, and a vehicle abnormal start warning signal is generated.

[0026] The present invention realizes real-time monitoring of vehicle status through multi-source monitoring data acquisition, which helps to find abnormal situations in real time. It realizes the inference that there is no one in the car, ensures the safety of the car, and avoids unnecessary safety hazards. It ensures that the minimum voltage required for the normal operation of vehicle components is continuously powered to ensure the normal operation of the vehicle system. Real-time anti-theft status data is obtained to provide basic data support for the subsequent monitoring of abnormal situations. Through real-time monitoring and mutation identification, the abnormal situation of the vehicle starting voltage is timely discovered, and the monitoring accuracy of the vehicle status is improved. The safety of the vehicle starting process is guaranteed to avoid the safety hazards caused by the starting problem. The owner's starting behavior model is constructed through historical data analysis to understand the owner's behavior habits and provide a basis for the identification of abnormal behavior. Personalized behavior habit analysis helps to predict the normal behavior pattern of the owner, so that abnormal behavior is easier to detect. The probability of abnormal behavior is calculated based on the owner's starting behavior model, the safety of vehicle starting is evaluated, and the early warning capability of abnormal situations is improved. Abnormal behavior is discovered in time, early warning measures are taken, and the safety of the vehicle and the owner is guaranteed. Once the vehicle is found to be abnormally started, the current output of the anti-theft starting battery is immediately cut off to ensure the safety of the vehicle. Generate abnormal start-up warning signals to alert vehicle owners and relevant managers to ensure that timely measures are taken to deal with abnormal situations and effectively prevent the risk of vehicle theft.

[0027] In the embodiment of the present invention, refer to Figure 1 , is a schematic flow chart of the steps of an anti-theft starting battery control method based on multi-source data of the present invention. In this example, the steps of the anti-theft starting battery control method based on multi-source data include: Step S1: acquiring real-time vehicle multi-source monitoring data; performing inference on the real-time vehicle multi-source monitoring data that no one is in the vehicle, generating a vehicle no-one signal, and entering the vehicle no-one mode; In this embodiment, a data acquisition module is designed, and appropriate hardware (such as a microcontroller or an on-board computer) and software (such as a data acquisition program) are used to acquire data in real time, ensuring that the module can be connected to each sensor through the car's CAN bus or other communication protocols, and the acquired sensor data is formatted and stored for subsequent analysis. A machine learning model or rule engine is designed to determine whether there is no one in the car. Logical judgment or training models (such as decision trees, support vector machines, etc.) can be used to set judgment conditions for the unmanned state, for example: there is no one in the driver's seat (seat_occupied is 0), the door is closed (door_closed is 1), and the interior environment meets the unmanned state (such as temperature and humidity are within a certain range). If the inference result is that there is no one, a corresponding signal or state identification is generated.

[0028] Step S2: when the vehicle is in the vehicle unmanned mode, obtaining the voltage status data of the vehicle components; providing continuous power supply at the lowest voltage based on the voltage status data of the vehicle components, and obtaining the real-time anti-theft status data of the vehicle; In this embodiment, ensure that the vehicle is equipped with appropriate voltage sensors that can monitor the voltage status of various key components (such as batteries, engines, and in-vehicle electronic devices) in real time. Develop a data acquisition program that regularly reads voltage data from voltage sensors and stores it in memory or data logs. Set a minimum voltage threshold according to the vehicle's electrical system specifications to ensure that key components can operate at this voltage. Assume that the minimum threshold is 11.5V. Write program logic to monitor the current voltage data and determine whether it is lower than the set minimum voltage threshold. If the voltage is lower than the threshold, execute the corresponding continuous power supply strategy to ensure that key components work properly. Implement a backup power supply activation mechanism to ensure that vehicle components can still be powered under low voltage conditions. Ensure that the vehicle is equipped with an anti-theft system that can monitor and feedback the anti-theft status of the vehicle in real time (such as whether the anti-theft is activated, sensor status, etc.). Develop a program to obtain real-time anti-theft status data and store it in memory.

[0029] Step S3: monitoring the vehicle starting voltage in real time according to the vehicle real-time anti-theft status data, identifying mutations, and extracting the starting voltage mutation data; In this embodiment, ensure that the vehicle is equipped with a voltage sensor that can measure the starting voltage (such as battery voltage) in real time. In addition, the system needs to be integrated with the anti-theft monitoring system to obtain the real-time anti-theft status. Develop a program to regularly read the starting voltage data from the voltage sensor and monitor it in combination with the anti-theft status. Determine the method for mutation identification. Simple threshold detection or more complex algorithms (such as change rate calculation, standard deviation judgment, etc.) can be used. Mutations are usually defined as significant changes in voltage in a short period of time. Set a voltage change threshold (such as 1V) to detect mutations. Each time the voltage is read, compare it with the previous voltage to determine whether it exceeds the set change threshold. Design a data structure (such as a list) to store the detected mutation data, including timestamps and mutation values. After the mutation is detected, call the storage function to record the mutation data.

[0030] while 1: start_voltage = read_start_voltage() security_status = get_security_status() print(f"Start voltage: {start_voltage:.2f}V, Anti-theft status: {security_status}") # Detect mutations is_mutation, change_value = detect_voltage_variation(start_voltage) if is_mutation: print(f"A sudden change was detected: voltage change {change_value:.2f}V") store_mutation_data(change_value)# Store mutation data time.sleep(1) Step S4: Acquire historical vehicle owner startup behavior data; perform personalized behavior habit analysis on the historical vehicle owner startup behavior data to construct a vehicle owner startup behavior model; In this embodiment, the storage location of the historical startup behavior data is determined, including the vehicle system, cloud database or local storage device. Export from database: Use SQL query or API call to extract the historical owner startup behavior data from the database. The data should include the startup time, owner ID, startup duration, etc. Extract meaningful features from the data, such as startup frequency, startup time distribution, and usage duration.

[0031] # Count the number of startups by owner group behavior_features = historical_data.groupby('owner_id').agg({ 'start_time': 'count', # number of starts 'duration': 'mean' # Average startup duration }).reset_index() behavior_features.columns = ['owner_id', 'startup_count', 'average_duration'] Analyze the startup behavior of car owners in different time periods (such as morning and evening). Select appropriate machine learning models based on behavioral characteristics, such as K-Means clustering, decision tree, or random forest. Prepare training data: Prepare the extracted feature data into model input format. Use the selected algorithm to train the model. Use the test data set to evaluate the accuracy and effect of the model.

[0032] Step S5: calculating the abnormal behavior probability of the starting voltage mutation data based on the vehicle owner's starting behavior model to obtain an abnormal evaluation result; In this embodiment, during the driving process of the vehicle, the data of the on-board voltage sensor is collected in real time, and the mutation characteristics of the voltage during the startup process are detected in real time. By comparing the voltage mutation data with the normal startup behavior model established in step 1, it is determined whether the current startup process is abnormal. The voltage mutation data monitored in real time is input into the owner's startup behavior model established in step 1, and the model is used to calculate the probability that the current startup process is an abnormal behavior. The calculation formula of the abnormal behavior probability is: P (abnormal|voltage mutation data) = 1-P (normal|voltage mutation data) where P (normal|voltage mutation data) can be directly calculated by the probability density function of the owner's startup behavior model. The calculated abnormal behavior probability is output as the final abnormal evaluation result, reflecting the probability of abnormality in the current startup process. A certain abnormal probability threshold can be set, and when the abnormal probability is detected to exceed the threshold, the subsequent anti-theft processing flow is triggered.

[0033] Step S6: When the abnormal evaluation result is that the vehicle is abnormally started, the current output of the anti-theft starting battery is cut off, and a vehicle abnormal start warning signal is generated.

[0034] In this embodiment, in the abnormal evaluation module, the abnormal behavior probability of the starting voltage mutation data is calculated according to the vehicle owner's starting behavior model, and the abnormal evaluation result is confirmed to be an abnormal start of the vehicle. Once the abnormal start of the vehicle is confirmed, the system will immediately execute the operation of cutting off the current output of the anti-theft starting battery, and send a signal to the relevant relay or circuit breaker through the electronic control unit (ECU) or a similar control module to cut off the battery current output. While cutting off the current output, the system generates an early warning signal of abnormal vehicle start to alert relevant personnel to the abnormal situation. Through the alarm module or communication module inside the system, the alarm signal is triggered and sent to the relevant monitoring equipment or the personnel's mobile phone. The system records the occurrence time, specific circumstances and measures taken of the abnormal start event for subsequent analysis and tracking. The system checks the current output cut-off and the early warning signal generation to ensure that the operation is successfully executed.

[0035] In this embodiment, refer to Figure 2 , is a flowchart of detailed implementation steps of step S1. In this embodiment, the detailed implementation steps of step S1 include: Step S11: acquiring real-time vehicle multi-source monitoring data, wherein the real-time vehicle multi-source monitoring data includes sonar sensor data, seat pressure sensor data and air monitoring sensor data; Step S12: filtering and denoising the sonar sensor data to obtain filtered audio data; Step S13: filtering abnormal pressure values ​​of the seat pressure sensor data to generate abnormal value filtered pressure data; Step S14: performing parameter standardization on the air monitoring sensor data to obtain standardized air monitoring data; Step S15: Inferring that there is no one in the vehicle based on the filtered audio data, the outlier filtered pressure data, and the standardized air monitoring data, generating a signal that there is no one in the vehicle, and entering the vehicle unmanned mode.

[0036] In this embodiment, the types of sensors that need to be installed are determined, including sonar sensors, seat pressure sensors, and air quality sensors. A suitable hardware platform (such as Arduino, Raspberry Pi, or a dedicated vehicle-mounted computer) is selected to integrate these sensors. The sonar sensors, seat pressure sensors, and air quality sensors are connected to the data acquisition system. The sensors are configured to ensure that they can collect data in real time and send the data to the processing unit through a serial port or a network interface. A suitable filtering algorithm, such as a low-pass filter or a median filter, is used to remove noise and retain valid signals. The filtering function in Python's SciPy library is used to process the audio data. A reasonable seat pressure range is determined to identify abnormal data, which can be defined by statistical analysis (such as mean ± 3σ). The seat pressure data is traversed to filter out abnormal values ​​that exceed the reasonable range. Z-score standardization or Min-Max standardization is used to convert the air monitoring data to a uniform scale. NumPy is used Or use Pandas to standardize the data, combine the processed sonar data, filtered pressure data, and standardized air data to form a comprehensive feature vector, and use machine learning models (such as random forests, logistic regression, etc.) to determine the unmanned state. The model can be trained with historical data. If the inference result is that there is no one, the vehicle state is switched to unmanned mode.

[0037] In this embodiment, the specific steps of step S15 are: Define the sliding time window monitoring length; The real-time vehicle multi-source monitoring data is divided into multiple time windows according to the monitoring length of the sliding time window, so as to obtain multi-source monitoring data of multiple time windows; Analyze the changes in the sound intensity in the car on the filtered audio data to obtain the sound intensity change data; The number of people in the car is determined based on the sound intensity change data, and the number of people in the car is estimated by the sound; Perform seat pressure detection on each seat for the outlier filtered pressure data. When the seat pressure is 0, obtain the empty seat data. Conduct breath detection analysis on standardized air monitoring data and extract breath detection data; Based on the breathing detection data, the number of people in the car is determined, and the number of people in the car is estimated by breathing; The number of people in the car estimated by sound, the number of empty seats, and the number of people in the car estimated by breathing are used to infer that there is no one in the car. When the number of people in the car estimated by sound is 0, the empty seats are all empty, and the number of people in the car estimated by breathing is 0, then a signal of no one in the car is obtained; The vehicle no-person inference operation is repeated based on multi-source monitoring data from multiple time windows. When the multi-source monitoring data from two consecutive time windows generates a vehicle no-person signal, the vehicle enters the vehicle no-person mode. In this embodiment, an appropriate sliding time window monitoring length is selected, such as 5 seconds, 10 seconds or 30 seconds, depending on the application environment and the frequency of sensor data acquisition. Real-time vehicle multi-source monitoring data is obtained, including sonar data, seat pressure data and air monitoring data. The real-time data is segmented using sliding window technology to generate data for multiple time windows. The sound intensity (such as RMS value or decibel value) of the filtered audio data in each time window is calculated. The sound intensity change of each window is recorded. A relationship model between the sound intensity and the number of people in the car is established. A simple threshold or regression model can be used. The pressure value of each seat is traversed to determine whether it is 0, and the empty seat is recorded. A suitable algorithm (such as peak detection or frequency analysis) is used to extract the breathing signal from the air monitoring data. The breathing signal in each time window is counted to infer the number of people in the car. No one is judged based on the number of people inferred by sound, the empty seat data and the number of people inferred by breathing. Check whether two consecutive time windows are both in the unmanned state.

[0038] def check_continuous_no_one(no_one_signals): count = 0 for signal in no_one_signals: if signal: count += 1 else: count = 0 if count>= 2:# two consecutive time windows return 1 return 0 if check_continuous_no_one(no_one_signals): print("Enter vehicle unmanned mode") # Execute the relevant code to enter the unmanned mode In this embodiment, refer to Figure 3 , is a flowchart of detailed implementation steps of step S2. In this embodiment, the detailed implementation steps of step S2 include: Step S21: when the vehicle is in the vehicle unmanned mode, obtaining the voltage status data of the vehicle components; Step S22: performing component cluster analysis on the vehicle component voltage state data to obtain a plurality of vehicle component type data; Step S23: performing minimum voltage requirement analysis on multiple vehicle component type data to obtain the minimum voltage requirement value of each component; Step S24: Based on the minimum voltage requirement value of each component, the anti-theft starting battery is used to continuously supply the minimum voltage, and the real-time anti-theft status data of the vehicle is obtained. In this embodiment, the sensors required for voltage monitoring are determined, which generally include voltage sensors and data acquisition modules. Connect the sensors to the vehicle's control unit and send data through a serial port or a network interface. Write data acquisition code to obtain the voltage status data of vehicle components in real time. Organize the collected vehicle component voltage data into a format suitable for cluster analysis, such as using a Pandas data frame. Use clustering algorithms such as K-means or DBSCAN to cluster the components. Establish the minimum voltage requirements for each component according to the vehicle manufacturer's specifications or industry standards. Calculate the minimum voltage requirement value for each component type. Print the minimum voltage requirement value for each component. Ensure that there is an anti-theft starting battery in the vehicle that can continuously supply power to the components with the lowest voltage requirements. Based on the minimum voltage requirement value, control the power supply of the anti-theft starting battery. Monitor and record anti-theft status data, such as whether any component receives additional power.

[0039] In this embodiment, refer to Figure 4 , is a flowchart of detailed implementation steps of step S3. In this embodiment, the detailed implementation steps of step S3 include: Step S31: Perform real-time monitoring of the vehicle starting voltage according to the real-time anti-theft status data of the vehicle, and generate anti-theft status starting voltage monitoring data; Step S32: performing sliding average processing on the anti-theft state starting voltage monitoring data to obtain smoothed starting voltage monitoring data; Step S33: fitting the time series change distribution of the smoothed starting voltage monitoring data to construct a starting voltage change curve; Step S34: identify the sudden change of the starting voltage change curve and extract the sudden change data of the starting voltage. In this embodiment, ensure that a real-time anti-theft status monitoring system is installed in the vehicle to obtain the starting voltage data. Connect to the anti-theft system through the vehicle network (such as CAN bus). Store the collected starting voltage data in a buffer for subsequent processing. A list or data frame structure can be used. Select the window length (n) of the sliding average, and the commonly used one is the simple sliding average (SMA). Apply the sliding average algorithm to the voltage monitoring data to generate smooth starting voltage monitoring data. Select a suitable fitting method according to the data characteristics, such as linear regression or polynomial fitting, and use SciPy or NumPy for data analysis. Use the fitting model to generate the starting voltage change curve.

[0040] from scipy import stats import numpy as np import matplotlib.pyplot as plt x = np.arange(len(smoothed_voltage)) slope, intercept, r_value, p_value, std_err = stats.linregress(x, smoothed_voltage) fitted_voltage = slope * x + intercept plt.plot(x, smoothed_voltage, label='Smoothed start voltage data') plt.plot(x, fitted_voltage, label='fitted curve', linestyle='--') plt.legend() plt.title('Starting voltage change curve') plt.xlabel('time point') plt.ylabel('Voltage(V)') plt.show() Determine the threshold of voltage mutation, for example, using the rate of change or absolute change value as the standard. Implement mutation identification algorithm: traverse the fitted voltage data and identify the mutation point. Output the detected mutation information.

[0041] In this embodiment, step S4 includes the following steps: Step S41: Obtaining historical vehicle owner startup behavior data; Step S42: extracting the timestamp of each startup based on the historical startup behavior data of the vehicle owner; Step S43: performing a time series startup behavior feature analysis on the historical startup behavior data of the vehicle owner based on the timestamp of each startup, thereby obtaining startup behavior time series feature data; Step S44: performing periodic start frequency distribution calculation on the start behavior time series characteristic data to obtain periodic start frequency distribution data; Step S45: Performing personalized behavior habit analysis on historical vehicle owner startup behavior data to generate vehicle owner startup behavior habit data; Step S46: Conduct behavioral feature mining on the periodic startup frequency distribution data and the owner's startup behavior habit data to construct the owner's startup behavior model. In this embodiment, determine the data storage location, such as a database, CSV file, or real-time data stream. Ensure that the data contains records of each startup, including timestamps, owner ID, and other relevant information. Use appropriate tools (such as Pandas) to load historical owner startup behavior data. Extract the timestamp of each startup from the data set, assuming that the timestamp field is called start_time. Copy Ensure that the timestamp format is correct and convert it to datetime type for subsequent analysis. Calculate the time interval between each startup to obtain startup interval data. Record the timestamp of each startup and its corresponding time interval. Count the number of startups in each time period, such as by hour, day, or week. Copy Convert the statistical results to DataFrame and name the columns. Determine the startup behavior habits of the owner, such as high-frequency startup time period, startup frequency, etc. Generate the startup behavior habit data of the owner by analyzing the startup frequency data. Combine the periodic startup frequency distribution data with the personalized behavior habit data to construct a feature matrix. Select an appropriate modeling method according to the goal, such as cluster analysis, classification model, or regression model.

[0042] Build models using machine learning libraries such as Scikit-learn. Evaluate model performance and check clustering results or classification accuracy.

[0043] In this embodiment, step S5 includes the following steps: Step S51: calculating the mutation rate of the startup voltage mutation data to generate a startup voltage mutation rate parameter; Step S52: judging the starting voltage mutation rate parameter based on the preset starting power supply change threshold value, and when the starting voltage mutation rate parameter is greater than or equal to the preset starting power supply change threshold value, determining that the vehicle is in the starting state, and obtaining the current vehicle starting behavior data; Step S53: calculating the abnormal behavior probability of the current vehicle startup behavior data according to the vehicle owner startup behavior model to generate current startup abnormal probability data; Step S54: Perform a comprehensive abnormality assessment on the previous startup abnormality probability data to obtain an abnormality assessment result.

[0044] In this embodiment, the calculation formula for determining the mutation rate is usually the ratio of the mutation value to the previous value. Record the calculated mutation rate data. Preset a reasonable startup power change threshold, such as 0.2 (20%). Traverse the mutation rate parameters to determine whether there is a value greater than or equal to the threshold.

[0045] def check_vehicle_start(mutation_rates, threshold): return any(rate>= threshold for rate in mutation_rates) vehicle_start_status = check_vehicle_start(mutation_rates, threshold) if vehicle_start_status: print("The vehicle has been determined to be in the start state") # Get the current vehicle startup behavior data current_startup_data = get_current_startup_data() Define a function to obtain the current startup behavior data, including timestamp, frequency and other features. Use the owner startup behavior model (such as logistic regression, decision tree, etc.) to calculate the probability of abnormal behavior. Clarify the evaluation criteria for abnormal probability, such as setting a threshold (such as 0.5) to determine whether there is abnormal behavior. Perform a comprehensive evaluation based on previous abnormal probability data.

[0046] def evaluate_abnormal_behavior(probability, threshold=0.5): if probability>= threshold: return "abnormal behavior exists" else: return "normal behavior" abnormal_assessment_result = evaluate_abnormal_behavior(abnormal_probability[0]) # Get the first probability value print("Abnormal assessment result:", abnormal_assessment_result) In this embodiment, step S6 includes the following steps: Step S61: When the abnormal evaluation result is that the vehicle is started normally, start the vehicle normally; Step S62: When the abnormal evaluation result is abnormal vehicle start, the current output of the anti-theft start battery is cut off, and a vehicle abnormal start warning signal is generated; Step S63: Perform deep learning evolution of anti-theft startup behavior based on the abnormal vehicle startup warning signal to build a battery anti-theft control strategy.

[0047] In this embodiment, check whether the abnormal evaluation result indicates that the vehicle is started normally. Call the vehicle control system to start the vehicle and send a command to start the vehicle. Determine whether the abnormal evaluation result indicates that the vehicle is started abnormally. Implement the logic of controlling the current output to ensure that the anti-theft starting battery is powered off. Record or trigger the system to generate an early warning signal for abnormal start-up for subsequent processing. Call the function that generates the early warning signal. Collect the relevant data of the abnormal start-up (such as time, frequency, and starting voltage) to prepare for subsequent model training.

[0048] abnormal_data = { 'timestamp': pd.Timestamp.now(), 'voltage': current_startup_data['frequency'], # Sample data 'status': "Abnormal startup" } Add the collected abnormal startup data to the historical data set to form a data set for deep learning training. Select a suitable deep learning model (such as LSTM, CNN, etc.), use the historical data set for training, and build a battery anti-theft control strategy.

[0049] from keras.models import Sequential from keras.layers import LSTM,Dense # Build LSTM model model = Sequential() model.add(LSTM(50, activation='relu', input_shape=(timesteps, features))) model.add(Dense(1))# Output layer model.compile(optimizer='adam', loss='mse') # Train the model model.fit(training_data, training_labels, epochs=100, batch_size=32) Make real-time decisions based on the trained model to control the battery's startup behavior.

[0050] def apply_control_strategy(new_data): prediction = model.predict(new_data) if prediction>threshold:# Set an appropriate threshold cut_off_power()# If the prediction is abnormal, cut off the current else: start_vehicle()# Otherwise start normally In this embodiment, an anti-theft starting battery control system based on multi-source data is provided, which is used to execute the anti-theft starting battery control method based on multi-source data as described above, including: The unmanned inference module is used to obtain real-time vehicle multi-source monitoring data; perform inference on the real-time vehicle multi-source monitoring data to determine that there is no one in the vehicle, generate an unmanned signal in the vehicle, and enter the vehicle unmanned mode; A low-voltage power supply module is used to obtain the voltage status data of vehicle components when the vehicle is in the unmanned mode; provide continuous power supply at the lowest voltage based on the voltage status data of vehicle components, and obtain the real-time anti-theft status data of the vehicle; A mutation identification module is used to monitor the vehicle starting voltage in real time according to the vehicle's real-time anti-theft status data, identify mutations, and extract the starting voltage mutation data; The startup behavior module is used to obtain the historical startup behavior data of the car owner; conduct personalized behavior habit analysis on the historical startup behavior data of the car owner and build the startup behavior model of the car owner; An abnormality assessment module is used to calculate the abnormal behavior probability of the starting voltage mutation data based on the vehicle owner's starting behavior model to obtain an abnormality assessment result; The power-off anti-theft module is used to cut off the current output of the anti-theft starting battery and generate an abnormal vehicle start warning signal when the abnormal evaluation result is abnormal vehicle start. The present invention obtains real-time vehicle multi-source monitoring data, generates an in-vehicle no-man signal by inferring that there is no one in the vehicle, enters the vehicle no-man mode, realizes real-time monitoring and response to the situation in the vehicle, obtains the voltage status data of vehicle components, realizes monitoring of the vehicle power supply status, performs continuous power supply at the minimum voltage based on the voltage data, obtains the real-time anti-theft status data of the vehicle, ensures the real-time anti-theft status of the vehicle in the no-man mode, monitors the vehicle starting voltage in real time, identifies mutations and extracts relevant data, which is helpful to capture abnormal situations in the starting process, obtains historical car owner starting behavior data, performs personalized behavior habit analysis, constructs a car owner starting behavior model, provides a basis for the evaluation of subsequent starting behaviors, calculates the probability of abnormal behavior for the starting voltage mutation data based on the car owner starting behavior model, obtains abnormal evaluation results, improves the ability to identify and respond to abnormal situations, and when the abnormal evaluation result is abnormal vehicle startup, cuts off the current output of the anti-theft starting battery, generates an early warning signal, responds quickly to abnormal situations, and ensures vehicle safety. Therefore, no matter from which point of view, the embodiments should be regarded as illustrative and non-restrictive, the scope of the present invention is defined by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and scope of equivalent elements of the application documents are included in the present invention. What has been described above is only a specific implementation mode of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined in this article can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown in this article, but will conform to the widest scope consistent with the principles and novel features invented in this article.

Claims

1. A method for controlling an anti-theft starting battery based on multi-source data, characterized in that: The following steps are involved: Step S1: Acquire real-time vehicle multi-source monitoring data; Inferring no one in the vehicle based on real-time multi-source vehicle monitoring data, generating no one in the vehicle signal, and entering the vehicle no-one mode; Step S2: when the vehicle is in the vehicle unmanned mode, obtaining the voltage status data of the vehicle components; providing continuous power supply at the lowest voltage based on the voltage status data of the vehicle components, and obtaining the real-time anti-theft status data of the vehicle; Step S3: monitoring the vehicle starting voltage in real time according to the vehicle real-time anti-theft status data, identifying mutations, and extracting the starting voltage mutation data; Step S4: Acquire historical vehicle owner startup behavior data; perform personalized behavior habit analysis on the historical vehicle owner startup behavior data to construct a vehicle owner startup behavior model; Step S5: calculating the abnormal behavior probability of the starting voltage mutation data based on the vehicle owner's starting behavior model to obtain an abnormal evaluation result; Step S6: When the abnormal evaluation result is that the vehicle is abnormally started, the current output of the anti-theft starting battery is cut off, and a vehicle abnormal start warning signal is generated.

2. The anti-theft starting battery control method based on multi-source data according to claim 1 is characterized in that: The specific steps of step S1 are: Step S11: acquiring real-time vehicle multi-source monitoring data, wherein the real-time vehicle multi-source monitoring data includes sonar sensor data, seat pressure sensor data and air monitoring sensor data; Step S12: filtering and denoising the sonar sensor data to obtain filtered audio data; Step S13: filtering abnormal pressure values ​​of the seat pressure sensor data to generate abnormal value filtered pressure data; Step S14: performing parameter standardization on the air monitoring sensor data to obtain standardized air monitoring data; Step S15: Inferring that there is no one in the vehicle based on the filtered audio data, the outlier filtered pressure data, and the standardized air monitoring data, generating a signal that there is no one in the vehicle, and entering the vehicle unmanned mode.

3. The anti-theft starting battery control method based on multi-source data according to claim 2 is characterized in that: The specific steps of step S15 are: Define the sliding time window monitoring length; The real-time vehicle multi-source monitoring data is divided into multiple time windows according to the monitoring length of the sliding time window, so as to obtain multi-source monitoring data of multiple time windows; Analyze the changes in the sound intensity in the car on the filtered audio data to obtain the sound intensity change data; The number of people in the car is determined based on the sound intensity change data, and the number of people in the car is estimated by the sound; Perform seat pressure detection on each seat for the outlier filtered pressure data. When the seat pressure is 0, obtain the empty seat data. Conduct breath detection analysis on standardized air monitoring data and extract breath detection data; Based on the breathing detection data, the number of people in the car is determined, and the number of people in the car is estimated by breathing; The number of people in the car estimated by sound, the number of empty seats, and the number of people in the car estimated by breathing are used to infer that there is no one in the car. When the number of people in the car estimated by sound is 0, the empty seats are all empty, and the number of people in the car estimated by breathing is 0, then a signal of no one in the car is obtained; The vehicle unmanned inference operation is repeated based on multi-source monitoring data from multiple time windows. When multi-source monitoring data from two consecutive time windows generates a vehicle unmanned signal, the vehicle enters unmanned mode.

4. The anti-theft starting battery control method based on multi-source data according to claim 1, characterized in that: The specific steps of step S2 are: Step S21: when the vehicle is in the vehicle unmanned mode, obtaining the voltage status data of the vehicle components; Step S22: performing component cluster analysis on the vehicle component voltage state data to obtain a plurality of vehicle component type data; Step S23: performing minimum voltage requirement analysis on multiple vehicle component type data to obtain the minimum voltage requirement value of each component; Step S24: Based on the minimum voltage requirement value of each component, the anti-theft starting battery is used to continuously supply the minimum voltage, and the real-time anti-theft status data of the vehicle is obtained.

5. The anti-theft starting battery control method based on multi-source data according to claim 1, characterized in that: The specific steps of step S3 are: Step S31: Perform real-time monitoring of the vehicle starting voltage according to the real-time anti-theft status data of the vehicle, and generate anti-theft status starting voltage monitoring data; Step S32: performing sliding average processing on the anti-theft state starting voltage monitoring data to obtain smoothed starting voltage monitoring data; Step S33: fitting the time series change distribution of the smoothed starting voltage monitoring data to construct a starting voltage change curve; Step S34: performing mutation identification on the starting voltage variation curve and extracting the starting voltage mutation data.

6. The anti-theft starting battery control method based on multi-source data according to claim 1, characterized in that: The specific steps of step S4 are: Step S41: Obtaining historical vehicle owner startup behavior data; Step S42: extracting the timestamp of each startup based on the historical startup behavior data of the vehicle owner; Step S43: performing a time series startup behavior feature analysis on the historical startup behavior data of the vehicle owner based on the timestamp of each startup, thereby obtaining startup behavior time series feature data; Step S44: performing periodic start frequency distribution calculation on the start behavior time series characteristic data to obtain periodic start frequency distribution data; Step S45: Performing personalized behavior habit analysis on historical vehicle owner startup behavior data to generate vehicle owner startup behavior habit data; Step S46: mining the behavior characteristics of the periodic start frequency distribution data and the car owner's start behavior habit data to construct a car owner's start behavior model.

7. The anti-theft starting battery control method based on multi-source data according to claim 1, characterized in that: The specific steps of step S5 are: Step S51: calculating the mutation rate of the startup voltage mutation data to generate a startup voltage mutation rate parameter; Step S52: judging the starting voltage mutation rate parameter based on the preset starting power supply change threshold value, and when the starting voltage mutation rate parameter is greater than or equal to the preset starting power supply change threshold value, determining that the vehicle is in the starting state, and obtaining the current vehicle starting behavior data; Step S53: calculating the abnormal behavior probability of the current vehicle startup behavior data according to the vehicle owner startup behavior model to generate current startup abnormal probability data; Step S54: Perform a comprehensive abnormality assessment on the previous startup abnormality probability data to obtain an abnormality assessment result.

8. The anti-theft starting battery control method based on multi-source data according to claim 1, characterized in that: The specific steps of step S6 are: Step S61: When the abnormal evaluation result is that the vehicle is started normally, start the vehicle normally; Step S62: When the abnormal evaluation result is abnormal vehicle start, the current output of the anti-theft start battery is cut off, and a vehicle abnormal start warning signal is generated; Step S63: Perform deep learning evolution of anti-theft startup behavior based on the abnormal vehicle startup warning signal to build a battery anti-theft control strategy.

9. An anti-theft starting battery control system based on multi-source data, characterized in that: The method for executing the anti-theft starting battery control method based on multi-source data as claimed in claim 1 comprises: The unmanned inference module is used to obtain real-time vehicle multi-source monitoring data; perform inference on the real-time vehicle multi-source monitoring data to determine that there is no one in the vehicle, generate an unmanned signal in the vehicle, and enter the vehicle unmanned mode; A low-voltage power supply module is used to obtain the voltage status data of vehicle components when the vehicle is in the unmanned mode; provide continuous power supply at the lowest voltage based on the voltage status data of vehicle components, and obtain the real-time anti-theft status data of the vehicle; A mutation identification module is used to monitor the vehicle starting voltage in real time according to the vehicle's real-time anti-theft status data, identify mutations, and extract the starting voltage mutation data; The startup behavior module is used to obtain the historical startup behavior data of the car owner; conduct personalized behavior habit analysis on the historical startup behavior data of the car owner and build the startup behavior model of the car owner; An abnormality assessment module is used to calculate the abnormal behavior probability of the starting voltage mutation data based on the vehicle owner's starting behavior model to obtain an abnormality assessment result; The power-off anti-theft module is used to cut off the current output of the anti-theft starting battery and generate a warning signal for abnormal vehicle start when the abnormal evaluation result is abnormal vehicle start.

Images