Vehicle-mounted cold chain temperature and humidity intelligent supervision and remote control method, system and device

Abstract

The invention relates to the technical field of data processing, and discloses a vehicle-mounted cold chain temperature and humidity intelligent supervision and remote control method, system and device. The method comprises the steps of collecting environment data through a temperature and humidity sensor and adding time codes, dividing priority data packets and encrypting and compressing the priority data packets, performing cleaning and abnormity correction on the data, dividing an early warning threshold value based on feature data and generating an early warning instruction, converting the early warning instruction into control parameters and issuing a remote regulation and control command, and finally, generating a monitoring report in combination with a terminal execution result to form an intelligent supervision scheme. According to the method, a multi-level data processing mechanism is introduced, cleaning, correction and interpolation compensation are carried out on the data in combination with an intelligent algorithm, and the accuracy and reliability of a monitoring result are remarkably improved.

Description

Technical Field

[0001] The present application relates to the field of data processing, and in particular to a method, system and device for intelligent monitoring and remote control of vehicle-mounted cold chain temperature and humidity. Background Art

[0002] In the existing cold chain transportation temperature and humidity monitoring technology, a distributed sensor network is usually used to monitor the temperature and humidity environment in the transport compartment in real time. These systems generally use wireless sensors to collect environmental data, and transmit the data to the monitoring center through a communication module to achieve remote supervision of the transportation process. The current monitoring solutions already have basic data collection, transmission and storage functions. Some systems also integrate simple threshold alarms and remote control mechanisms, which can issue alarms when the temperature and humidity exceed the preset range, and support remote control of refrigeration equipment. At the same time, some advanced systems have also introduced Internet of Things technology and cloud computing platforms to achieve centralized management and visual display of data.

[0003] However, the existing technical solutions have the following shortcomings: First, the raw data collected by sensors often have noise interference and data missing problems. Traditional data processing methods are difficult to effectively identify and correct outliers, affecting the accuracy of monitoring; second, most existing systems adopt a unified data transmission strategy, and fail to perform differentiated processing according to the importance of data, which easily causes waste of network bandwidth or delayed transmission of key data; third, the current early warning mechanism is mainly based on fixed threshold judgment, lacks dynamic analysis of temperature and humidity change trends, and is difficult to timely predict and prevent potential environmental anomalies; finally, existing remote control solutions generally adopt simple switch control or fixed parameter adjustment, lack adaptive regulation capabilities, and are difficult to flexibly adjust control strategies according to actual environmental changes. Summary of the invention

[0004] The present application provides a method, system and device for intelligent monitoring and remote control of vehicle-mounted cold chain temperature and humidity, which is used to solve the technical problem that data processing in existing cold chain transportation temperature and humidity monitoring systems is not intelligent enough. Specifically, it is difficult for the existing technology to effectively process outliers and missing values ​​in sensor collected data, and it is impossible to achieve intelligent cleaning and correction of data, resulting in insufficient accuracy and reliability of monitoring results. The present invention significantly improves data quality and monitoring accuracy by introducing a multi-level data processing mechanism and combining it with an intelligent algorithm to clean, correct and interpolate data.

[0005] In the first aspect, the present application provides a vehicle-mounted cold chain temperature and humidity intelligent supervision and remote control method, the vehicle-mounted cold chain temperature and humidity intelligent supervision and remote control method comprising: collecting and preliminarily processing environmental parameters through temperature and humidity sensors in the car, adding time codes and format identifiers to the data, and obtaining an initial data stream; dividing the initial data stream into priority transmission data packets and non-priority transmission data packets, and compressing and encrypting the priority transmission data packets and non-priority transmission data packets respectively to obtain encrypted monitoring data; performing data cleaning based on the encrypted monitoring data, correcting and calculating the identified abnormal values, performing interpolation compensation for data missing points, determining the change trend through statistical operations, and forming temperature and humidity characteristic data; dividing warning thresholds according to the temperature and humidity characteristic data, grading abnormal states and filtering redundant information according to the threshold judgment results, and generating warning instructions; converting the warning instructions into control parameters, and generating remote control commands, sending the remote control commands to the vehicle-mounted control terminal, and obtaining the terminal execution results;

[0006] Based on the terminal execution results and the temperature and humidity characteristic data, a remote monitoring report is generated, and combined with the execution records of the early warning instructions and remote control commands, an intelligent supervision plan for cold chain transportation is formed.

[0007] In the second aspect, the present application provides a vehicle-mounted cold chain temperature and humidity intelligent supervision and remote control system, the vehicle-mounted cold chain temperature and humidity intelligent supervision and remote control system comprising:

[0008] The acquisition module is used to collect and preliminarily process environmental parameters through the temperature and humidity sensors in the car, add time coding and format identification to the data, and obtain the initial data stream;

[0009] An encryption module is used to divide the initial data stream into priority transmission data packets and non-priority transmission data packets, and perform data compression and encryption encoding on the priority transmission data packets and non-priority transmission data packets respectively to obtain encrypted monitoring data;

[0010] A correction module is used to perform data cleaning based on the encrypted monitoring data, perform correction calculations on identified abnormal values, perform interpolation compensation for data missing points, determine the change trend through statistical calculations, and form temperature and humidity characteristic data;

[0011] A classification module, used to classify warning thresholds according to the temperature and humidity characteristic data, classify abnormal states according to the threshold determination results, filter redundant information, and generate warning instructions;

[0012] A generating module, used for converting the warning instruction into a control parameter, generating a remote control command, sending the remote control command to the vehicle control terminal, and obtaining the terminal execution result;

[0013] The supervision module is used to generate a remote monitoring report based on the terminal execution results and the temperature and humidity characteristic data, and to form an intelligent supervision plan for cold chain transportation in combination with the execution records of the early warning instructions and remote control commands.

[0014] The third aspect of the present application provides a vehicle-mounted cold chain temperature and humidity intelligent monitoring and remote control device, including: a vehicle-mounted cold chain temperature and humidity intelligent monitoring and remote control system.

[0015] In the technical solution provided by the present application, the environmental parameters are collected and preliminarily processed by the temperature and humidity sensors in the car, and the data are added with time codes and format identifiers to obtain the initial data stream, which effectively ensures the accuracy and integrity of data collection and provides a reliable data basis for subsequent analysis. By dividing the initial data stream into priority transmission data packets and non-priority transmission data packets, and performing data compression and encryption encoding respectively, the encrypted monitoring data is obtained, which significantly improves the data transmission efficiency and ensures the security of data transmission. Data cleaning is performed based on the encrypted monitoring data, the identified abnormal values ​​are corrected and calculated, interpolation compensation is performed for the data missing points, and the change trend is determined by statistical operations to form temperature and humidity characteristic data, which greatly improves the data quality and eliminates the influence of abnormal values ​​and data missing on the monitoring effect. The warning threshold is divided according to the temperature and humidity characteristic data, and the abnormal state is graded and redundant information is filtered according to the threshold judgment result, and the warning instruction is generated, which realizes accurate abnormal detection and graded warning, and avoids false alarms and missed alarms. The warning instruction is converted into a control parameter, and a remote control command is generated, and the remote control command is sent to the vehicle control terminal to obtain the terminal execution result, so that the temperature and humidity control is more accurate and timely, and the control effect is improved. Based on the terminal execution results and temperature and humidity characteristic data, a remote monitoring report is generated, and combined with the execution records of early warning instructions and remote control commands, an intelligent supervision plan for cold chain transportation is formed, which realizes intelligent supervision of the entire process and improves the overall quality and reliability of cold chain transportation. This method combines multi-level data processing and intelligent control strategies to achieve accurate monitoring and efficient adjustment of the temperature and humidity environment during cold chain transportation, significantly improving the safety and reliability of cold chain logistics. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0017] Figure 1 This is a schematic diagram of an embodiment of the vehicle-mounted cold chain temperature and humidity intelligent supervision and remote control method in the embodiment of the present application;

[0018] Figure 2 This is a schematic diagram of an embodiment of the vehicle-mounted cold chain temperature and humidity intelligent supervision and remote control system in the embodiment of the present application. DETAILED DESCRIPTION

[0019] Embodiments of the present application provide a method, system and device for intelligent supervision and remote control of temperature and humidity of a vehicle-mounted cold chain. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0020] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 In the embodiment of the present application, an embodiment of the vehicle-mounted cold chain temperature and humidity intelligent supervision and remote control method includes:

[0021] Step S101, collect and preliminarily process environmental parameters through the temperature and humidity sensors in the vehicle compartment, add time coding and format identification to the data, and obtain an initial data stream;

[0022] Step S102, dividing the initial data stream into priority transmission data packets and non-priority transmission data packets, respectively compressing and encrypting the priority transmission data packets and the non-priority transmission data packets to obtain encrypted monitoring data;

[0023] Step S103: perform data cleaning based on the encrypted monitoring data, perform correction calculation on the identified abnormal values, perform interpolation compensation for the data missing points, determine the change trend through statistical calculations, and form temperature and humidity characteristic data;

[0024] Step S104, dividing the warning threshold according to the temperature and humidity characteristic data, classifying the abnormal state according to the threshold determination result and filtering the redundant information, and generating a warning instruction;

[0025] Step S105: convert the warning instruction into a control parameter, generate a remote control command, send the remote control command to the vehicle control terminal, and obtain the terminal execution result;

[0026] Step S106: Generate a remote monitoring report based on the terminal execution results and temperature and humidity characteristic data, and combine the execution records of the early warning instructions and remote control commands to form an intelligent supervision plan for cold chain transportation.

[0027] It is understandable that the execution subject of this application can be an on-board cold chain temperature and humidity intelligent supervision and remote control system, or a terminal or a server, which is not limited here. The embodiment of this application is described by taking the server as the execution subject as an example.

[0028] Optionally, the environmental parameter data is collected in real time through the temperature and humidity sensor network installed in the carriage. High-precision temperature and humidity sensors are arranged on the top, middle and bottom of the carriage respectively. The DHT22 digital temperature and humidity sensor is used. The temperature measurement range is -40℃ to 80℃, the accuracy is ±0.5℃, the humidity measurement range is 0-100%RH, the accuracy is ±2%RH, and the sampling frequency is 10 seconds / time. The raw data collected by the sensor is screened by signal strength, and data points with signal strength lower than -85dBm are removed to ensure data quality. The screened data is then calibrated, and duplicate sampling points are removed by the data calibration algorithm. Timestamp codes and sensor location identifiers accurate to milliseconds are added to each data point, and finally a standardized initial data stream is generated. When the initial data stream is prioritized, the sliding time window method is used, and the window size is set to 5 minutes to process the data in segments. Based on the temperature and humidity change amplitude, the priority judgment standard is set: when the temperature change rate exceeds 2℃ / minute or the humidity change rate exceeds 5%RH / minute, the data in the time window is marked as a priority transmission data packet; the remaining data is divided into non-priority transmission data packets. Priority transmission data packets are quickly compressed using the LZ4 compression algorithm, with a compression ratio of up to 3:1, and non-priority transmission data packets are processed using the ZSTD algorithm with a high compression ratio of up to 5:1. The compressed data is securely encrypted using a hybrid encryption method of RSA asymmetric encryption and AES symmetric encryption, and a digital signature generated based on the SHA-256 algorithm is embedded to ensure data transmission security.

[0029] The data cleaning process first decrypts the received encrypted monitoring data, then uses the 3σ criterion to detect outliers, calculates the mean and standard deviation of the temperature and humidity data, and marks the data points that deviate from the mean by more than three times the standard deviation as outliers. For the detected outliers, the wavelet decomposition method is used for multi-scale denoising, and the db4 wavelet basis function is used to decompose the signal into four layers, and the signal is reconstructed after removing the high-frequency noise component. The missing data points are compensated by the Lagrange interpolation algorithm, and three valid data points before and after the missing points are selected to construct the interpolation polynomial. The Gaussian kernel function is used for data smoothing, and the kernel width parameter is set to 1.5 to map the mutation point to a reasonable interval. Finally, the temperature and humidity change trend is calculated by the exponentially weighted moving average method, and the weighting factor is set to 0.2 to obtain the smoothed temperature and humidity characteristic data. The cluster analysis method is used to divide the warning threshold, and the K-means clustering algorithm is used to group the temperature and humidity characteristic data. The number of cluster centers is set to 3, corresponding to the three levels of normal, warning and danger. The threshold range of each level is calculated by the interquartile range algorithm, and the upper and lower limits are Q3+1.5IQR and Q1-1.5IQR, respectively, where Q1 and Q3 are the first and third quartiles, and IQR is the interquartile range. The maximum entropy algorithm is used to quantify the risk level, and the entropy value threshold is set to 0.75. The state above this value is judged as high risk. Duplicate warnings are removed by calculating similarity, and the similarity threshold is set to 0.9. The distributed Kalman filter is used to filter the interference signal, and finally a warning instruction in a standard format is generated.

[0030] During the conversion of control parameters, the warning instructions are first parsed into specific temperature and humidity deviation values, and the adaptive PID algorithm is used to calculate the cooling power and ventilation intensity. The PID controller parameters are adaptively adjusted, with the range of the proportional coefficient Kp being 0.5-2.0, the range of the integral time Ti being 60-180 seconds, and the range of the differential time Td being 10-30 seconds. The calculated control parameters are encoded into binary instructions, CRC32 check bits and millisecond timestamps are added, and a secure remote control command is constructed, which is sent to the vehicle terminal device through the 4G network, and the terminal execution status is obtained in real time. Finally, the terminal execution results are analyzed in time and space with the temperature and humidity characteristic data, and the control effect is evaluated using a multi-layer neural network. The network structure is three-layer, the number of hidden layer neurons is 20, and the ReLU activation function is used. The temperature and humidity change trend is predicted in combination with the gray system theory, with a prediction step of 30 minutes. The system operation status is analyzed by Bayesian network reasoning, and a network model containing nodes such as temperature, humidity, and control parameters is constructed to calculate the state transition probability. The random forest algorithm was used to extract key performance indicators, the number of decision trees was set to 100, and the tree depth was limited to 10, ultimately forming a complete intelligent supervision solution for cold chain transportation.

[0031] For example, during the transportation of a cold chain transport vehicle, the top sensor collected a temperature data sequence of [2.5℃, 2.6℃, 2.8℃, 5.2℃, 2.7℃, 2.6℃]. According to the 3σ criterion, the data mean is 3.1℃, the standard deviation is 1.1℃, and 5.2℃ is identified as an abnormal value. After wavelet denoising and Gaussian smoothing, the correction value is 2.8℃. At the same time, the middle sensor has data missing during this period. Through Lagrange interpolation, the temperature value of the missing point is calculated to be 2.6℃ based on the temperature values ​​before and after [2.4℃, 2.5℃, 2.7℃, 2.6℃]. The system detects that the temperature is rising. According to the results of cluster analysis, the current temperature is close to the warning threshold (3.0℃), and the early warning mechanism is automatically triggered. The PID controller calculates the refrigeration power increase command, and after sending it to the vehicle terminal, the temperature gradually drops to 2.4℃ and returns to the normal range. During the entire process, the system ensures the timely transmission of key data through a priority mechanism, with a data compression ratio of 4:1, while maintaining the stability of the temperature and humidity environment.

[0032] In the embodiment of the present application, the environmental parameters are collected and preliminarily processed by the temperature and humidity sensor in the car, and the data is added with time code and format identification to obtain the initial data stream, which effectively ensures the accuracy and integrity of data collection and provides a reliable data basis for subsequent analysis. By dividing the initial data stream into priority transmission data packets and non-priority transmission data packets, and performing data compression and encryption coding respectively, the encrypted monitoring data is obtained, which significantly improves the data transmission efficiency and ensures the security of data transmission. Data cleaning is performed based on the encrypted monitoring data, the identified abnormal values ​​are corrected and calculated, interpolation compensation is performed for the data missing points, and the change trend is determined by statistical calculation to form temperature and humidity characteristic data, which greatly improves the data quality and eliminates the influence of abnormal values ​​and data missing on the monitoring effect. The warning threshold is divided according to the temperature and humidity characteristic data, and the abnormal state is graded and redundant information is filtered according to the threshold judgment result, and the warning instruction is generated, which realizes accurate abnormal detection and graded warning, and avoids false alarms and missed alarms. The warning instruction is converted into a control parameter, and a remote control command is generated, and the remote control command is sent to the vehicle control terminal to obtain the terminal execution result, so that the temperature and humidity control is more accurate and timely, and the control effect is improved. Based on the terminal execution results and temperature and humidity characteristic data, a remote monitoring report is generated, and combined with the execution records of early warning instructions and remote control commands, an intelligent supervision plan for cold chain transportation is formed, which realizes intelligent supervision of the entire process and improves the overall quality and reliability of cold chain transportation. This method combines multi-level data processing and intelligent control strategies to achieve accurate monitoring and efficient adjustment of the temperature and humidity environment during cold chain transportation, significantly improving the safety and reliability of cold chain logistics.

[0033] In an optional embodiment, the process of executing step S101 may specifically include the following steps:

[0034] (1) Collect environmental parameters through temperature and humidity sensors on the top, middle and bottom of the car, select effective sampling points based on signal strength, and generate raw sampling data;

[0035] (2) Sort the original sampling data according to the acquisition time, remove the duplicate sampling points through data calibration operation, and obtain the calibration data;

[0036] (3) Timestamp the calibration data and mark the sampling node information according to the data source location to form a data sequence with an identifier;

[0037] (4) Convert the identified data sequence into a standard data format and generate a standard data structure through data formatting;

[0038] (5) Perform integrity check on the standard data structure and perform linear interpolation on the detected data breakpoints to obtain complete data;

[0039] (6) The completed data is sorted in time series and the initial data stream is generated through data streaming processing.

[0040] Optionally, DHT22 digital temperature and humidity sensors are installed on the top, middle and bottom of the carriage respectively. The sensor has high measurement accuracy, with a temperature measurement range of -40℃ to 80℃, an accuracy of ±0.5℃, and a humidity measurement range of 0-100%RH, with an accuracy of ±2%RH. The sampling frequency of each sensor is set to 10 seconds / time, and data is transmitted in real time through the wireless communication module. In order to ensure the quality of the data, the received signal strength indication (RSSI) value is used as the signal strength screening standard, and the data points with RSSI values ​​lower than -85dBm are judged as weak signal data and removed, thereby ensuring the reliability of the sampled data. After obtaining the original sampled data, the data needs to be sorted in time sequence and repeated values ​​are removed. First, the data is sorted in ascending order according to the acquisition timestamp, and the timestamp is accurate to milliseconds. Then, the repeated sampling points are identified and removed through data calibration operations. The specific method is: set the time threshold to 100 milliseconds. For any two adjacent data points, if the difference between their timestamps is less than the threshold and comes from the same sensor, it is judged as repeated sampling, and only the data points with earlier timestamps are retained. In this way, data redundancy is effectively reduced and data processing efficiency is improved. The timestamp encoding of the data adopts the Unix timestamp format, which is accurate to the millisecond level, and location identification information is added to each data point. The location identification includes three levels: sensor number (8-bit string), installation location (top / middle / bottom) and spatial coordinates (distance relative to the front end of the car, in centimeters). This information is combined to form a unique node identifier, which is convenient for subsequent data tracking and analysis. The identified data sequence is organized in JSON format, and each data point contains fields such as timestamp, temperature value, humidity value, location identification, etc.

[0041] When converting the identified data sequence into a standard data format, a unified data structure definition is used. The standard data structure contains the following fields: record ID (unique identifier), timestamp (millisecond level), sensor information (including device ID, model, location), environmental parameters (temperature, humidity), signal quality (RSSI value), data status (valid / abnormal / compensated). Through data formatting, all fields are uniformly converted to the specified data type and precision to ensure the standardization and consistency of the data. The integrity check link adopts a multiple verification mechanism. First, the integrity of the data is verified by CRC32 checksum calculation. If a data breakpoint is detected, the linear interpolation process is started. Linear interpolation uses a piecewise linear interpolation algorithm. For data breakpoints with a time interval of more than 30 seconds, interpolation calculations are performed based on the valid data before and after the breakpoint. The interpolation calculation takes into account the continuity characteristics of temperature changes and generates reasonable transition data based on the temperature change trend before and after the breakpoint.

[0042] Finally, the completed data is sorted in time series, and the initial data stream is generated using data streaming processing. Data streaming processing includes three steps: data segmentation, time series alignment, and data merging. Data segmentation is divided into 5-minute units, and each data block contains data from different sensors in the same period. Time series alignment ensures that the data from different sensors are synchronized in the time dimension, and a sliding window with a time window of 1 second is used for alignment processing. Data merging integrates multiple sensor data in the same time window into a complete record.

[0043] For example, during the operation of a cold chain transport vehicle, sensors at three locations collected the following data from 10:00:00 to 10:00:30: top sensor (temperature: 2.5°C, humidity: 85%, RSSI: -75dBm, timestamp: 10:00:00), middle sensor (temperature: 2.3°C, humidity: 87%, RSSI: -82dBm, timestamp: 10:00:00), bottom sensor (temperature: 2.4°C, humidity: 86%, RSSI: -78dBm, timestamp: 10:00:00). At 10:00:10, the middle sensor lost data due to signal interference, while the top and bottom sensors recorded data normally (top: 2.6°C, 86%; bottom: 2.5°C, 85%). The system first screens valid data points based on RSSI values, then sorts the data by timestamp and removes duplicate sampling points (a duplicate record was found for the top sensor at 10:00:00, with RSSI values ​​of -75dBm and -77dBm, respectively, and the data with a stronger signal was retained). For the missing data at 10:00:10 for the middle sensor, a linear interpolation method was used to calculate the missing point data of 2.35℃, 86.5% based on the temperature and humidity data at 10:00:00 and 10:00:20 (2.3℃, 87% and 2.4℃, 86%). Finally, the processed data was organized into a data stream in a standard format, containing complete timing information and location identifiers, to form a continuous monitoring record.

[0044] In an optional embodiment, the process of executing step S102 may specifically include the following steps:

[0045] (1) Segment the initial data stream through time windows, divide the data blocks according to the data collection time interval, and generate time series data segments;

[0046] (2) Prioritize the time series data segments, classify the data according to the temperature and humidity change range, and obtain data groups with priorities;

[0047] (3) Splitting the data group with priority according to the priority level, separating the priority transmission data packets and the non-priority transmission data packets, and obtaining classified data packets;

[0048] (4) Perform dual compression of differential encoding and run-length encoding on the classified data packets, process repeated data segments through a block compression algorithm, and form a compressed data stream;

[0049] (5) Encrypt the compressed data stream in segments, perform dual encryption conversion of asymmetric key encryption and symmetric key encryption on the data segments through a hybrid encryption algorithm, and generate a hybrid encrypted data stream;

[0050] (6) Embed the digital signature and timestamp information into the hybrid encrypted data stream, generate a check code through an integrity check algorithm, and combine the check code with the hybrid encrypted data stream to encapsulate the encrypted monitoring data.

[0051] Optionally, the initial data stream is segmented by a sliding time window, the time window size is set to 5 minutes, and the window sliding step is 1 minute. The data collection time interval is 10 seconds, so each time window contains 30 sampling points. For the data in each time window, it is organized into independent data blocks according to the timestamp of the data, and each data block contains the complete temperature and humidity records in the time period, thereby forming a time series data segment. After obtaining the time series data segment, priority marking is performed based on the temperature and humidity change amplitude. Specifically, the change rate of temperature and humidity in each data block is calculated, and the priority judgment threshold is set: when the temperature change rate exceeds 2℃ / minute or the humidity change rate exceeds 5%RH / minute, the data block is marked as high priority; when the temperature change rate is between 1-2℃ / minute or the humidity change rate is between 3-5%RH / minute, it is marked as medium priority; the remaining data blocks are marked as low priority. In this way, a data group with a priority mark is formed.

[0052] Subsequently, the data groups with priorities are split according to the priority level. The high-priority and medium-priority data blocks are classified as priority transmission data packets, and the low-priority data blocks are classified as non-priority transmission data packets. This classification mechanism ensures the priority processing and transmission of key data, and improves the real-time performance and reliability of the system. The classified data packets are subjected to dual compression processing of differential encoding and run-length encoding. First, differential encoding takes the difference between adjacent data points as the encoding object, which effectively reduces the redundancy of the data. For example, the temperature sequence [2.5, 2.6, 2.8, 2.7, 2.6] becomes [2.5, 0.1, 0.2, -0.1, -0.1] after differential encoding. Subsequently, run-length encoding is applied to the differential encoding result, and the continuously repeated values ​​are represented by counting, such as [-0.1, -0.1] is encoded as [-0.1, 2], indicating that the value -0.1 is repeated twice. For repeated segments in the data, the LZ77 block compression algorithm is used to replace the repeated data sequences with references to historical data, thereby achieving a higher compression rate.

[0053] The encryption of compressed data streams uses a hybrid encryption scheme that combines the advantages of RSA asymmetric encryption and AES symmetric encryption. First, the session key is encrypted using the RSA-2048 algorithm, and then the actual data is encrypted using the AES-256 algorithm using the session key. Specifically, the compressed data stream is segmented into 1KB segments, each data segment is encrypted with a different session key using AES, and the session key itself is protected by RSA encryption. This dual encryption mechanism ensures the security of data transmission while maintaining a high processing efficiency. To ensure the integrity and non-repudiation of the data, it is necessary to embed a digital signature and timestamp information into the hybrid encrypted data stream. The digital signature is generated using the following formula:

[0054] in, Represents the final checksum. Indicates the content value of the data block. Represents the result of hash function calculation. represents the weight factor, represents a large prime number, Indicates the total number of data blocks. Calculate the SHA-256 hash value for each data block and generate a checksum in combination with the weight factor. Finally, encapsulate the checksum and encrypted data stream to form complete encrypted monitoring data.

[0055] For example, a cold chain transport vehicle collected 30 sets of temperature and humidity data in the 5-minute time window from 10:00:00 to 10:05:00. During the period from 10:02:00 to 10:03:00, the temperature rose rapidly from 2.5°C to 4.5°C, and the humidity dropped from 85% to 75%. The calculated temperature change rate was 2°C / minute, and the humidity change rate was 10%RH / minute, which exceeded the preset threshold. Therefore, the data in this period was marked as high priority. This segment of data is differentially encoded, and the temperature sequence [2.5, 3.0, 3.5,4.0, 4.5] is encoded as [2.5, 0.5, 0.5, 0.5, 0.5], and then processed by run-length encoding to [2.5, 0.5, 4], indicating 4 consecutive increments of 0.5 after the initial value of 2.5. During the encryption process, a 256-bit AES session key is first generated, the session key is encrypted with the RSA public key, and then the compressed data is encrypted with the session key. Finally, the SHA-256 hash value of the data is calculated, and a digital signature is generated in combination with the time window information. The signature and the encrypted data are packaged and transmitted together. This process not only ensures the security of data transmission, but also significantly improves the transmission efficiency.

[0056] In an optional embodiment, the process of executing step S103 may specifically include the following steps:

[0057] (1) Decrypt and restore the encrypted monitoring data, filter out data outliers using the 3σ criterion, and output the basic data stream;

[0058] (2) For the noise signal in the basic data stream, multi-scale denoising is performed in conjunction with wavelet decomposition to obtain a denoised data chain;

[0059] (3) Using time windows to identify data breakpoints from the denoised data chain, supplement missing values ​​using the Lagrangian interpolation algorithm, and construct a data continuum;

[0060] (4) Combine the Gaussian kernel function to map the mutation points in the data continuum to a reasonable interval to generate a smooth data matrix;

[0061] (5) Calculate the trend function of the smoothed data matrix using the exponential weighted average method, derive the characteristics of environmental parameter changes, and form a dynamic feature chain;

[0062] (6) Project the dynamic feature chain into the feature space, extract the temperature and humidity correlation factors, and summarize them into temperature and humidity feature data.

[0063] Optionally, the data is restored using a preset decryption key to obtain the original temperature and humidity data sequence. The 3σ criterion is then used to screen out data outliers. This criterion identifies outliers based on the standard deviation of the data. The specific method is: calculate the mean and standard deviation of the temperature and humidity data, mark the data points that deviate from the mean by more than three times the standard deviation as outliers and temporarily remove them, thereby obtaining the basic data stream. For the noise signal in the basic data stream, wavelet decomposition technology is used for multi-scale denoising. Specifically, the db4 wavelet basis function is used to decompose the signal into 4 layers to obtain wavelet coefficients in different frequency bands. The high-frequency noise coefficient is processed by setting a threshold, and the wavelet coefficient is shrunk by a soft threshold method. Finally, the denoised data chain is obtained by wavelet reconstruction. Wavelet denoising can effectively retain the mutation characteristics of the signal while removing the influence of random noise.

[0064] On the basis of the denoising data chain, a 5-minute sliding time window is used to identify data breakpoints, that is, the position where the sampling interval exceeds 30 seconds is marked as a data breakpoint. For the detected data breakpoints, the Lagrange interpolation algorithm is used to supplement the missing values. Specifically, three valid data points before and after the breakpoint are selected as interpolation reference points, the Lagrange interpolation polynomial is constructed, and the interpolation results of the breakpoint position are calculated. This interpolation method can maintain the continuity and smoothness of the data, so that the supplemented data is more in line with the actual change law. In order to further process the mutation points in the data, the Gaussian kernel function is used to smooth the data. The kernel width parameter of the Gaussian kernel function is set to 1.5, and the data points are locally weighted averaged, so that the value of the mutation point is close to the surrounding data points, while keeping the overall trend of the data unchanged. In this way, a smooth data matrix is ​​generated, in which each row represents the temperature and humidity state at a time point.

[0065] When calculating the trend function for the smoothed data matrix, the exponential weighted average method is used, and the calculation formula is:

[0066] in, express The trend value at the moment, represents the data value of the previous k moments, represents the time attenuation coefficient, represents the smoothing factor, represents the relative rate of change, represents the seasonal factor, represents the weight coefficient, and Respectively represent the adjustment parameters of the trend term and the seasonal term, n represents the size of the historical data window, and m represents the number of seasonal cycles. The change trend of the temperature and humidity parameters is calculated by this formula to form a dynamic feature chain.

[0067] Finally, the dynamic feature chain is projected into the two-dimensional feature space. The horizontal axis represents the temperature change characteristics, and the vertical axis represents the humidity change characteristics. By calculating the correlation coefficient between temperature and humidity changes, the correlation between temperature and humidity is extracted, and finally summarized into temperature and humidity feature data.

[0068] For example, the temperature data sequence collected by a cold chain transport vehicle over a period of time is [2.5℃, 2.6℃, 2.8℃, 5.2℃, 2.7℃, 2.6℃]. First, outlier detection is performed using the 3σ criterion, and the data mean is calculated to be 3.1℃, with a standard deviation of 1.1℃. It is found that 5.2℃ exceeds the range of three times the standard deviation of the mean (3.1±3.3℃), and it is marked as an outlier. The remaining data is decomposed by wavelet, and the db4 wavelet basis is used for 4-layer decomposition to obtain multi-scale wavelet coefficients. The signal is reconstructed after threshold processing to obtain the denoised data sequence [2.5℃, 2.6℃, 2.8℃, 2.7℃, 2.6℃]. It is then found that there is a 30-second data breakpoint in the data sequence. The Lagrange interpolation algorithm is used to calculate the interpolation result of 2.75℃ based on the temperature values ​​before and after the breakpoint. After smoothing with the Gaussian kernel function, the temperature change trend is calculated, and finally a dynamic feature chain reflecting the temperature change characteristics is obtained. This feature chain is correlated with the humidity feature data of the same period to extract the correlation factor between temperature and humidity, providing a basis for subsequent early warning judgment.

[0069] In an optional embodiment, the process of executing step S104 may specifically include the following steps:

[0070] (1) Calculate the coefficient of variation from the temperature and humidity characteristic data, distinguish the temperature and humidity fluctuation range through cluster analysis, and mark the data critical point;

[0071] (2) Use the data critical points to construct the boundary function, calculate the abnormal threshold range through the interquartile range algorithm, and output the threshold judgment matrix;

[0072] (3) Based on the threshold judgment matrix, identify abnormal points from the temperature and humidity change trend, quantify the risk level through the maximum entropy algorithm, and generate a risk state sequence;

[0073] (4) Similarity calculation is performed on the risk status sequence to eliminate duplicate warning items, and interference signals are removed through distributed filters to obtain core warning data;

[0074] (5) Convert the core warning data into a standard warning format, assign processing priorities based on the decision tree algorithm, and form a hierarchical warning sequence;

[0075] (6) Conduct spatiotemporal correlation analysis on the hierarchical warning sequence, integrate geographic location information, and construct warning instructions.

[0076] Optionally, the coefficient of variation is the ratio of the standard deviation to the mean, which is used to measure the degree of dispersion of the data. The coefficient of variation is calculated for the temperature and humidity data respectively, where the standard deviation reflects the fluctuation range of the data and the mean reflects the central tendency of the data. Then the K-means clustering algorithm is used to cluster the temperature and humidity data, and the number of clusters is set to 3 categories, corresponding to the normal interval, the warning interval and the danger interval, respectively. The clustering process divides similar data points into the same category through iterative optimization, and finally determines the boundary points of each interval, that is, the data critical points. Based on the obtained data critical points, a segmented boundary function is constructed to describe the allowable range of temperature and humidity changes. The interquartile range (IQR) algorithm is used to calculate the abnormal threshold range. First, the first quartile Q1 and the third quartile Q3 of the data are calculated. The interquartile range is IQR=Q3-Q1. Then [Q1-1.5IQR, Q3+1.5IQR] is set as the normal range, [Q1-3IQR, Q1-1.5IQR] and [Q3+1.5IQR, Q3+3IQR] are set as the warning range. Data beyond these ranges are considered abnormal. These threshold ranges are organized into a matrix form, with each row corresponding to a time point and each column corresponding to a threshold boundary.

[0077] Based on the threshold judgment matrix, the abnormal points of the temperature and humidity change trend are identified. When the data points fall into different intervals, the maximum entropy algorithm is used to calculate their risk level. The maximum entropy algorithm quantifies the risk level by calculating the uncertainty of data distribution. The algorithm considers multiple features such as the distance between the data point and the critical point and the rate of change, and outputs a risk value between 0 and 1. Risk values ​​greater than 0.8 are judged as high risk, between 0.5 and 0.8 are judged as medium risk, between 0.3 and 0.5 are judged as low risk, and less than 0.3 are judged as normal, thereby generating a state sequence containing risk level markers. The generated risk state sequence is similarity calculated, and the adjacent risk states are compared using the cosine similarity method. When the similarity exceeds 0.85, it is judged as a repeated warning, and only the warning information that appears for the first time is retained. Then, a distributed Kalman filter is used to filter the warning data for noise. The state vector of the filter contains temperature, humidity and its rate of change. Through the two steps of prediction and update, false warnings caused by sensor fluctuations are effectively removed, and the real core warning data are finally screened out.

[0078] The core warning data is converted into a predefined standard format, including warning time, location, type, level and other fields. The CART decision tree algorithm is used to assign a processing priority to each warning. The characteristics of the decision tree include risk level, impact range, duration and other factors. The leaf nodes of the decision tree correspond to different priority levels, from P0 (highest) to P4 (lowest), a total of 5 levels, and a hierarchical warning sequence with priority marks is generated based on this. Finally, a spatiotemporal correlation analysis is performed on the hierarchical warning sequence to associate the warning event with the vehicle's geographic location information. The vehicle location is obtained through the GPS module with an accuracy of meters. Through spatiotemporal correlation analysis, it is possible to identify whether the warning event is related to a specific geographical environment (such as direct sunlight and high temperature areas), and integrate this information into the warning instruction. The standard warning instruction contains trigger conditions, disposal suggestions, response time limits and other contents.

[0079] For example, during the operation of a cold chain transport vehicle, the temperature data sequence collected within one hour from 10 am to 11 am is [2.5℃, 2.6℃, 2.8℃, 3.2℃, 3.8℃, 4.5℃, 5.0℃, 5.2℃, 5.1℃, 4.9℃], and the sampling interval is 6 minutes. First, the coefficient of variation is calculated. The average value of the data is 3.96℃, the standard deviation is 1.07℃, and the coefficient of variation is 0.27. Through K-means cluster analysis, the data is divided into three categories: [2.5℃-3.2℃] is the normal range, [3.2℃-4.5℃] is the warning range, and [above 4.5℃] is the danger range. The interquartile range algorithm is used to calculate Q1=2.8℃, Q3=5.0℃, and IQR=2.2℃, and the abnormal threshold range is determined accordingly. After the maximum entropy algorithm analysis, the temperature points of 4.5℃ and above were judged as high risk (risk value>0.8), and a risk state sequence was generated. Similarity calculation shows that the similarity of the three consecutive data points of 5.0℃, 5.2℃ and 5.1℃ exceeds 0.85, and they are merged into one warning event. After the decision tree analysis, considering that the temperature exceeds the standard range and lasts for a long time, the warning event is assigned to the P1 level (the second highest priority). Combined with the GPS location information of the vehicle at the time (showing that the vehicle is in direct sunlight), a warning instruction is generated, including specific disposal suggestions such as "the temperature continues to exceed the standard, it is recommended to adjust the cooling power immediately, and the response time limit is 10 minutes".

[0080] In an optional embodiment, the process of executing step S105 may specifically include the following steps:

[0081] (1) Analyze the abnormal temperature and humidity status according to the early warning instructions, quantify the temperature and humidity deviation values ​​by the adaptive algorithm, and generate control benchmark data;

[0082] (2) Substitute the control benchmark data into the feedback control function, calculate the cooling power and ventilation intensity based on the PID algorithm, and form the equipment control parameters;

[0083] (3) Encode the device control parameters into binary instructions through the protocol converter, embed the check bit and timestamp, and construct the control instruction packet;

[0084] (4) Add priority identification and routing information to the control instruction packet, perform security encryption, and output the remote control command;

[0085] (5) Send the remote control command to the vehicle terminal via the communication link, obtain the terminal response data, record the command interaction process, and form an execution status flow;

[0086] (6) Analyze the execution state flow through the state parser, extract key execution information, and generate terminal execution results.

[0087] Optionally, the received warning instruction needs to be parsed first to extract the abnormal temperature and humidity status information. The warning instruction contains information such as the current value, target value and deviation range of temperature and humidity. The temperature and humidity deviation is quantified by an adaptive algorithm, which dynamically adjusts the quantification parameters according to the historical control effect. The quantification range of temperature deviation is -10℃ to +10℃, and the quantification range of humidity deviation is -20%RH to +20%RH, with quantification accuracy of 0.1℃ and 0.5%RH respectively. Based on the quantified deviation value and the change trend of environmental parameters, control benchmark data including target temperature and humidity, adjustment direction and adjustment step are generated. After the control benchmark data is input into the feedback control module, the PID (proportional-integral-differential) algorithm is used to calculate the specific control parameters. The proportional coefficient Kp of the PID algorithm ranges from 0.5 to 2.0, which is used to adjust the response speed; the integral time Ti ranges from 60 to 180 seconds, which is used to eliminate steady-state errors; the differential time Td ranges from 10 to 30 seconds, which is used to suppress overshoot. According to the size and change rate of the temperature and humidity deviation, the PID controller calculates the cooling power and ventilation intensity in real time. The adjustment range of cooling power is 0-100%, in steps of 5%; the ventilation intensity is divided into 1-10 levels, corresponding to different fan speeds.

[0088] The control parameters are encoded by the protocol converter, and the analog quantity is converted into binary instructions in the standard Modbus-RTU protocol format. During the protocol conversion process, the parameters such as cooling power and ventilation intensity are first packaged according to the predefined data type (16-bit unsigned integer), and then the function code (0x06 means writing a single register), device address (0x01-0xFF) and other information are added. To ensure the reliability of data transmission, the instruction packet is subjected to CRC16 cyclic redundancy check to generate a 2-byte check code. At the same time, a Unix timestamp accurate to milliseconds is embedded in the instruction packet for instruction timing control and timeout judgment.

[0089] Add a control priority identifier to the complete control instruction packet. The priority is divided into five levels from P0 (emergency) to P4 (normal). The priority information is encoded in 3-bit binary. At the same time, add routing information, including the source address (control center) and the destination address (vehicle terminal). The address uses a 6-byte MAC address format. Then the entire instruction packet is encrypted with AES-256, the key length is 256 bits, and the encryption mode is CBC to ensure the security of data transmission. The encrypted data constitutes a complete remote control command. The remote control command is sent to the vehicle terminal through the 4G / 5G wireless communication network. The communication uses the TCP / IP protocol to ensure the reliability of data transmission. After the instruction is issued, the response status of the terminal is monitored in real time, including instruction reception confirmation, execution start time, execution progress, execution completion time and other information. This information is recorded in chronological order to form a complete instruction execution status stream. Each record in the status stream contains fields such as timestamp, status code (0-normal, 1-warning, 2-error), and progress value (0-100%).

[0090] The state parser analyzes the execution state flow and extracts key execution information. First, check whether the execution timing of the instruction is normal, including whether the response delay is within the allowable range (usually less than 500ms) and whether the execution process is continuous. Then verify whether the execution result meets expectations, including whether the control parameters are set correctly and whether the device is operating normally. Finally, generate the terminal execution result in a standard format, including execution status (success / failure), completion degree, exception information (if any), etc.

[0091] For example, a cold chain transport vehicle detects that the compartment temperature rises to 5.2°C (target temperature range 2-4°C), triggering an early warning command. The adaptive algorithm calculates the temperature deviation value as +2.2°C (relative to the target temperature of 3°C), and records the temperature rise rate as 0.4°C / minute. Based on these parameters, the PID controller calculates that the cooling power needs to be increased to 85% (the original value is 60%) and the ventilation intensity is adjusted to 8 gears (the original value is 5 gears). These control parameters are encoded as binary instructions: function code 0x06, device address 0x01, cooling power register address 0x1000, data value 0x0055 (85%), CRC checksum 0xA1B2. The instruction packet is given P1 priority, the MAC address of the target terminal 00:1A:2B:3C:4D:5E is added, and it is sent through the 4G network after AES encryption. After receiving the instruction, the terminal returns a confirmation message within 50ms and completes the parameter setting within 2 seconds. The status flow shows that the execution process is normal and there are no abnormal events. Finally, the cabin temperature dropped to 3.8℃ within 10 minutes, achieving the control target.

[0092] In an optional embodiment, the process of executing step S106 may specifically include the following steps:

[0093] (1) Perform spatiotemporal registration of the terminal execution results with the temperature and humidity feature data, construct a multidimensional feature matrix with the help of correlation calculation, and generate a state evaluation sequence;

[0094] (2) Perform multi-layer neural network analysis on the state assessment sequence, combine the grey system theory to quantify the control effect, and output dynamic response data;

[0095] (3) Based on dynamic response data, the execution information of warning instructions and remote control commands is integrated, and the operation status chain is generated through Bayesian network reasoning;

[0096] (4) Conduct deep learning analysis on the operation status chain, extract key performance indicators through iterative optimization algorithms, and build a monitoring and evaluation system;

[0097] (5) Import the monitoring and evaluation system into the knowledge graph, use the graph computing engine to mine the system operation rules, and form a remote monitoring report;

[0098] (6) The remote monitoring reports are summarized in multiple dimensions using the random forest algorithm, and the control strategies are sorted out using time series data mining technology to build an intelligent supervision solution for cold chain transportation.

[0099] Optionally, the terminal execution result is spatiotemporally aligned with the temperature and humidity feature data. The alignment process is based on the timestamp, the alignment accuracy is set to 1 second, and the missing values ​​between data points are supplemented by the interpolation method. The correlation calculation adopts the grey correlation analysis method to calculate the correlation coefficient between different features. The correlation threshold is set to 0.8, and the features above this value are considered to have significant correlation. The multidimensional feature matrix contains dimensions such as temperature value, humidity value, cooling power, ventilation intensity, and control delay. Each dimension is standardized to the [0,1] interval to form a unified numerical scale. For the generated state evaluation sequence, a multilayer neural network consisting of an input layer, two hidden layers, and an output layer is constructed. The number of nodes in the input layer is the number of feature dimensions (usually 10-15 nodes), the first hidden layer has 20 nodes, the second hidden layer has 10 nodes, and the output layer has 4 nodes, which correspond to different aspects of the control effect. The ReLU activation function is used to train the network parameters through the back propagation algorithm. At the same time, the GM(1,1) model in the grey system theory is combined to predict the temperature and humidity change trend, with a prediction step of 30 minutes, and the prediction accuracy is evaluated by residual analysis.

[0100] The dynamic response data is used as the basis, and the execution information of the early warning instructions and remote control commands is integrated to build a Bayesian network model. The network nodes include temperature status, humidity status, control parameters, execution effect, etc. The conditional probability between nodes is obtained through historical data statistics. The Bayesian network evaluates the effects of different control strategies through probabilistic reasoning methods and predicts possible abnormal situations. The operation status chain records the complete control process, including the state parameters and conversion probabilities at each time point. The deep learning analysis of the operation status chain uses an LSTM (long short-term memory) network. The network structure contains an LSTM layer (128 neurons) and two fully connected layers (64 and 32 neurons). The input sequence length is 60 (corresponding to 10 minutes of data), and the sequence prediction is performed by sliding window. The iterative optimization uses the Adam optimizer, the learning rate is set to 0.001, and the training round is 100 rounds. This method extracts key performance indicators, including temperature control accuracy, response time, energy efficiency, etc.

[0101] The representation of the monitoring and evaluation system in the knowledge graph adopts an entity-relationship structure. Entities include equipment, parameters, states, events and other types, and relationships describe the mechanism of action between entities. The graph computing engine uses a depth-first search algorithm to mine the association pattern between the temperature and humidity change rules and the control effect. The analysis results form a remote monitoring report in a standard format, which includes performance evaluation, anomaly analysis, optimization suggestions and other contents. Finally, the random forest algorithm is used to summarize the monitoring report in multiple dimensions. The random forest contains 100 decision trees, each with a maximum depth of 8 and a feature sampling ratio of 0.8. The splitting criterion of the decision tree uses the Gini coefficient to evaluate the importance of the feature. The SAX (SymbolicAggregate approXimation) algorithm is used for time series data mining to convert continuous time series into discrete symbol sequences, which is convenient for discovering repeated patterns and abnormal behaviors.

[0102] For example, during a 4-hour transportation process of a cold chain transport vehicle, the temperature control target is 2-4℃. During this period, there was an abnormal temperature (rising to 5.2℃), and the temperature returned to 3.8℃ within 10 minutes after the control command was triggered. The spatiotemporal registration process aligns the temperature data per second (a total of 14,400 points) with the execution status of the control command, and the correlation between temperature and control power is calculated to be 0.85, indicating a good control response. Multi-layer neural network analysis shows that the comprehensive score of this control is 92 points (out of 100 points), including 95 points for response speed, 90 points for stability, and 91 points for energy efficiency. Bayesian network analysis shows that the success rate of this control strategy is 95% under similar conditions. The key indicators extracted from the time series data by the LSTM network include: average temperature deviation 0.3℃, maximum overshoot 0.5℃, and adjustment time 8.5 minutes. After analyzing these data with the random forest algorithm, the optimal control parameter combination is determined: the refrigeration power start threshold is set to 75%, and the ventilation intensity adjustment step is 2 gears. The final intelligent supervision plan recommends that when the temperature rise rate exceeds 0.3℃ / minute, regulation should be intervened in advance and the starting value of the cooling power should be appropriately increased to improve the response speed and reduce energy consumption.

[0103] The above describes the vehicle-mounted cold chain temperature and humidity intelligent supervision and remote control method in the embodiment of the present application. The following describes the vehicle-mounted cold chain temperature and humidity intelligent supervision and remote control system in the embodiment of the present application. Please refer to Figure 2 In the embodiment of the present application, one embodiment of the vehicle-mounted cold chain temperature and humidity intelligent supervision and remote control system includes:

[0104] The acquisition module is used to collect and preliminarily process environmental parameters through the temperature and humidity sensors in the car, add time coding and format identification to the data, and obtain the initial data stream;

[0105] An encryption module is used to divide the initial data stream into priority transmission data packets and non-priority transmission data packets, and perform data compression and encryption encoding on the priority transmission data packets and non-priority transmission data packets respectively to obtain encrypted monitoring data;

[0106] A correction module is used to perform data cleaning based on the encrypted monitoring data, perform correction calculations on identified abnormal values, perform interpolation compensation for data missing points, determine the change trend through statistical calculations, and form temperature and humidity characteristic data;

[0107] A classification module, used to classify warning thresholds according to the temperature and humidity characteristic data, classify abnormal states according to the threshold determination results, filter redundant information, and generate warning instructions;

[0108] A generating module, used for converting the warning instruction into a control parameter, generating a remote control command, sending the remote control command to the vehicle control terminal, and obtaining the terminal execution result;

[0109] The supervision module is used to generate a remote monitoring report based on the terminal execution results and the temperature and humidity characteristic data, and to form an intelligent supervision plan for cold chain transportation in combination with the execution records of the early warning instructions and remote control commands.

[0110] Through the cooperation of the above components, the environmental parameters are collected and preliminarily processed by the temperature and humidity sensors in the car, and the data are added with time codes and format identifiers to obtain the initial data stream, which effectively ensures the accuracy and integrity of data collection and provides a reliable data basis for subsequent analysis. By dividing the initial data stream into priority transmission data packets and non-priority transmission data packets, and compressing and encrypting them respectively, the encrypted monitoring data is obtained, which significantly improves the data transmission efficiency and ensures the security of data transmission. Data cleaning is performed based on the encrypted monitoring data, the identified abnormal values ​​are corrected and calculated, interpolation compensation is performed for the data missing points, and the change trend is determined by statistical operations to form temperature and humidity characteristic data, which greatly improves the data quality and eliminates the influence of abnormal values ​​and data missing on the monitoring effect. The warning threshold is divided according to the temperature and humidity characteristic data, and the abnormal state is graded and redundant information is filtered according to the threshold judgment result, and the warning instruction is generated, which realizes accurate abnormality detection and graded warning, avoiding false alarms and missed alarms. The warning instruction is converted into control parameters, and a remote control command is generated, which is sent to the vehicle control terminal to obtain the terminal execution result, making the temperature and humidity control more accurate and timely, and improving the control effect. Based on the terminal execution results and temperature and humidity characteristic data, a remote monitoring report is generated, and combined with the execution records of early warning instructions and remote control commands, an intelligent supervision plan for cold chain transportation is formed, which realizes intelligent supervision of the entire process and improves the overall quality and reliability of cold chain transportation. This method combines multi-level data processing and intelligent control strategies to achieve accurate monitoring and efficient adjustment of the temperature and humidity environment during cold chain transportation, significantly improving the safety and reliability of cold chain logistics.

[0111] The present application also provides a vehicle-mounted cold chain temperature and humidity intelligent monitoring and remote control device, including: a vehicle-mounted cold chain temperature and humidity intelligent monitoring and remote control system.

[0112] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0113] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A vehicle-mounted cold chain temperature and humidity intelligent supervision and remote control method, characterized in that: The vehicle-mounted cold chain temperature and humidity intelligent supervision and remote control method includes: The environmental parameters are collected and preliminarily processed through the temperature and humidity sensors in the car, and the data are added with time coding and format identification to obtain the initial data stream; Dividing the initial data stream into priority transmission data packets and non-priority transmission data packets, respectively compressing and encrypting the priority transmission data packets and non-priority transmission data packets to obtain encrypted monitoring data; Based on the encrypted monitoring data, data cleaning is performed, correction calculations are performed on the identified abnormal values, interpolation compensation is performed for data missing points, and the change trend is determined through statistical calculations to form temperature and humidity characteristic data; Divide the warning threshold according to the temperature and humidity characteristic data, classify the abnormal state according to the threshold determination result, filter the redundant information, and generate a warning instruction; Convert the warning instruction into a control parameter and generate a remote control command, send the remote control command to the vehicle control terminal, and obtain the terminal execution result; Based on the terminal execution results and the temperature and humidity characteristic data, a remote monitoring report is generated, and combined with the execution records of the early warning instructions and remote control commands, an intelligent supervision plan for cold chain transportation is formed.

2. The vehicle-mounted cold chain temperature and humidity intelligent supervision and remote control method according to claim 1 is characterized in that: The environmental parameters are collected and preliminarily processed by the temperature and humidity sensors in the vehicle compartment, and the data are added with time codes and format identifiers to obtain the initial data stream, including: Environmental parameters are collected through temperature and humidity sensors on the top, middle and bottom of the carriage, and effective sampling points are selected based on signal strength to generate raw sampling data; The original sampling data are sorted according to the acquisition time, and repeated sampling points are removed through data calibration operation to obtain calibration data; Timestamp encoding is performed on the calibration data, and sampling node information is marked according to the data source location to form a data sequence with identification; Convert the identified data sequence into a standard data format, and generate a standard data structure through data formatting; Performing integrity check on the standard data structure, performing linear interpolation processing on the detected data breakpoints, and obtaining supplementary data; The supplemented data is time-sequentially correlated and arranged, and the initial data stream is generated through data stream processing.

3. The vehicle-mounted cold chain temperature and humidity intelligent supervision and remote control method according to claim 1 is characterized in that: The step of dividing the initial data stream into priority transmission data packets and non-priority transmission data packets, and performing data compression and encryption encoding on the priority transmission data packets and non-priority transmission data packets respectively to obtain encrypted monitoring data comprises: Segmenting the initial data stream by time window, dividing the data blocks according to the data collection time interval, and generating time series data segments; Priority marking is performed on the time series data segments, and the data is graded according to the temperature and humidity change amplitude to obtain a data group with priority; Splitting the prioritized data group according to priority levels, separating the prioritized transmission data packets and the non-priority transmission data packets, and obtaining classified data packets; The classified data packets are double compressed by differential coding and run-length coding, and repeated data segments are processed by a block compression algorithm to form a compressed data stream; The compressed data stream is encrypted in segments, and the data segments are double-encrypted by asymmetric key encryption and symmetric key encryption using a hybrid encryption algorithm to generate a hybrid encrypted data stream; A digital signature and timestamp information are embedded in the mixed encrypted data stream, a check code is generated through an integrity check algorithm, and the check code is combined and encapsulated with the mixed encrypted data stream to obtain the encrypted monitoring data.

4. The vehicle-mounted cold chain temperature and humidity intelligent supervision and remote control method according to claim 1 is characterized in that: The method of performing data cleaning based on the encrypted monitoring data, performing correction calculation on the identified abnormal values, performing interpolation compensation on the data missing points, determining the change trend through statistical calculation, and forming temperature and humidity characteristic data includes: Decrypt and restore the encrypted monitoring data, filter out data outliers using the 3σ criterion, and output a basic data stream; For the noise signal in the basic data stream, multi-scale denoising is performed in conjunction with wavelet decomposition to obtain a denoised data chain; Using a time window to identify data breakpoints from the denoised data chain, supplementing missing values ​​using a Lagrangian interpolation algorithm, and constructing a data continuum; The mutation points in the data continuum are mapped to a reasonable interval by combining a Gaussian kernel function to generate a smooth data matrix; Calculating the trend function of the smoothed data matrix by exponential weighted average method, deriving the characteristics of environmental parameter changes, and forming a dynamic characteristic chain; The dynamic feature chain is projected into the feature space, and the temperature and humidity correlation factors are extracted and summarized as the temperature and humidity feature data.

5. The vehicle-mounted cold chain temperature and humidity intelligent supervision and remote control method according to claim 1 is characterized in that: The method of dividing the warning threshold according to the temperature and humidity characteristic data, grading the abnormal state and filtering the redundant information according to the threshold determination result, and generating the warning instruction includes: Calculate the data variation coefficient from the temperature and humidity characteristic data, distinguish the temperature and humidity fluctuation range through cluster analysis, and mark the data critical point; The boundary function is constructed using the data critical points, the abnormal threshold range is calculated using the interquartile range algorithm, and the threshold determination matrix is ​​output; Based on the threshold determination matrix, abnormal points are identified from the temperature and humidity change trend, and the risk level is quantified by the maximum entropy algorithm to generate a risk state sequence; The similarity calculation is called for the risk state sequence, repeated warning items are eliminated, interference signals are removed through a distributed filter, and core warning data is screened; Convert the core warning data into a standard warning format, assign processing priorities based on a decision tree algorithm, and form a hierarchical warning sequence; A spatiotemporal correlation analysis is performed on the hierarchical warning sequence, and geographical location information is integrated to construct the warning instruction.

6. The vehicle-mounted cold chain temperature and humidity intelligent supervision and remote control method according to claim 1 is characterized in that: The step of converting the warning instruction into a control parameter and generating a remote control command, sending the remote control command to the vehicle-mounted control terminal, and obtaining the terminal execution result includes: Analyze the abnormal temperature and humidity status according to the early warning instruction, quantify the temperature and humidity deviation value by the adaptive algorithm, and generate control benchmark data; Substituting the control benchmark data into the feedback control function, calculating the cooling power and ventilation intensity based on the PID algorithm, and forming the equipment control parameters; The device control parameters are encoded into binary instructions via a protocol converter, a check bit and a timestamp are embedded, and a control instruction packet is constructed; Adding a priority identifier and routing information to the control instruction packet, performing security encryption, and outputting a remote control command; Sending the remote control command to the vehicle terminal via the communication link, obtaining the terminal response data, recording the instruction interaction process, and forming an execution status flow; The execution state flow is analyzed by a state parser, key execution information is extracted, and the terminal execution result is generated.

7. The vehicle-mounted cold chain temperature and humidity intelligent supervision and remote control method according to claim 1 is characterized in that: The remote monitoring report is generated based on the terminal execution result and the temperature and humidity characteristic data, and the execution record of the early warning instruction and the remote control command is combined to form an intelligent supervision plan for cold chain transportation, including: Performing spatiotemporal registration of the terminal execution result and the temperature and humidity feature data, constructing a multidimensional feature matrix by means of correlation calculation, and generating a state evaluation sequence; Perform multi-layer neural network analysis on the state evaluation sequence, combine the grey system theory to quantify the control effect, and output dynamic response data; Based on the dynamic response data, the execution information of the early warning command and the remote control command is integrated, and an operation status chain is generated through Bayesian network reasoning; Perform deep learning analysis on the operation status chain, extract key performance indicators through iterative optimization algorithm, and build a monitoring and evaluation system; Import the monitoring and evaluation system into the knowledge graph, mine the system operation rules through the graph computing engine, and form a remote monitoring report; The remote monitoring report is summarized in multiple dimensions using the random forest algorithm, and the control strategy is sorted out using time series data mining technology to construct the intelligent supervision solution for cold chain transportation.

8. An on-vehicle cold chain temperature and humidity intelligent supervision and remote control system, used to implement the on-vehicle cold chain temperature and humidity intelligent supervision and remote control method as described in any one of claims 1-7, characterized in that: The vehicle-mounted cold chain temperature and humidity intelligent supervision and remote control system includes: The acquisition module is used to collect and preliminarily process environmental parameters through the temperature and humidity sensors in the car, add time coding and format identification to the data, and obtain the initial data stream; An encryption module is used to divide the initial data stream into priority transmission data packets and non-priority transmission data packets, and perform data compression and encryption encoding on the priority transmission data packets and non-priority transmission data packets respectively to obtain encrypted monitoring data; A correction module is used to perform data cleaning based on the encrypted monitoring data, perform correction calculations on identified abnormal values, perform interpolation compensation for data missing points, determine the change trend through statistical calculations, and form temperature and humidity characteristic data; A classification module, used to classify warning thresholds according to the temperature and humidity characteristic data, classify abnormal states according to the threshold determination results, filter redundant information, and generate warning instructions; A generating module, used for converting the warning instruction into a control parameter, generating a remote control command, sending the remote control command to the vehicle control terminal, and obtaining the terminal execution result; The supervision module is used to generate a remote monitoring report based on the terminal execution results and the temperature and humidity characteristic data, and to form an intelligent supervision plan for cold chain transportation in combination with the execution records of the early warning instructions and remote control commands.

9. A vehicle-mounted cold chain temperature and humidity intelligent monitoring and remote control device, characterized in that: include: An on-vehicle cold chain temperature and humidity intelligent monitoring and remote control system as described in claim 8.

Images