Flow state solidified soil monitoring data safety detection system

Abstract

The invention discloses a flow-state solidified soil monitoring data security detection system, which comprises a sensor node, an edge calculation module, a reinforcement learning module and a data transmission and storage module, and is characterized in that the sensor node is used for continuously acquiring multiple physical quantities in a target area; the edge calculation module is used for realizing real-time preprocessing, quality monitoring and security encryption of data, and sending the processed encrypted data to the data transmission and storage module; the reinforcement learning module dynamically optimizes the monitoring strategy of the whole system through interaction with the environment, so that the data precision, the energy consumption and the delay are optimally balanced; and the data transmission and storage module is used for transmitting the encrypted data to a central server for long-term storage and analysis. According to the invention, by deploying the edge computing module on each sensor node, local processing and preliminary analysis of data are realized, dependence on a central server is reduced, delay of data transmission and processing is significantly reduced, and response speed and real-time performance of the system are improved.

Description

Technical Field

[0001] This invention relates to the field of data monitoring technology, and more specifically, to a data security detection system for monitoring fluidized solidified soil. Background Technology

[0002] With the continuous development of soil engineering and environmental monitoring technologies, fluidized bed soil, as a new type of soil amendment material, is widely used in infrastructure construction, land remediation, and other fields. To ensure the stability and construction quality of fluidized bed soil, real-time monitoring of its physical state is crucial. Current technologies primarily rely on sensor arrays to monitor parameters such as temperature, humidity, pressure, and strain of the fluidized bed soil. The sensors transmit the collected data to a central processing system for subsequent analysis and evaluation. Through data analysis, engineers can obtain real-time soil condition information, thereby guiding construction and management work.

[0003] However, most existing monitoring systems adopt a centralized data processing architecture, which has the following drawbacks:

[0004] Latency Issues: Existing fluidized solidified soil monitoring systems typically employ a centralized architecture, meaning all data collected by sensors must be transmitted over a network to a remote server for centralized processing. This approach inevitably leads to data transmission delays, especially in large-scale sensor networks where the volume of data is enormous. The transmission and processing delays increase significantly, resulting in delayed feedback of monitoring results and impacting the system's response speed and decision-making efficiency.

[0005] Bandwidth and computational bottlenecks: Due to the massive amounts of data generated by sensors, centralized systems often face problems such as insufficient network bandwidth and limited computing power. Data transmission may experience packet loss, delays, or even errors, leading to incomplete or inaccurate monitoring data. Furthermore, the limited computing resources of data processing centers make it difficult to efficiently process large amounts of real-time monitoring data, further exacerbating the system's performance bottlenecks.

[0006] Insufficient data security and transmission stability: Current data transmission technologies largely rely on traditional open network protocols, lacking security protections during the data transmission process. This makes data vulnerable to network attacks, tampering, or leakage during transmission, making it difficult to guarantee the integrity and confidentiality of monitoring data.

[0007] Therefore, a new solution is needed to address the above problems. Summary of the Invention

[0008] The purpose of this invention is to overcome the shortcomings of the prior art and provide a secure detection system for monitoring data of fluidized solidified soil.

[0009] To achieve the above objectives, the present invention adopts the following technical solution:

[0010] A data security monitoring system for fluidized solidified soil includes sensor nodes, an edge computing module, a reinforcement learning module, and a data transmission and storage module. The sensor nodes continuously collect multiple physical quantities within the target area during the construction and solidification process of the fluidized solidified soil and transmit the raw collected data to the edge computing module. The edge computing module, integrated within the sensor nodes, performs real-time data preprocessing, quality monitoring, and security encryption, and sends the processed encrypted data to the data transmission and storage module. The edge computing module interacts bidirectionally with the reinforcement learning module, providing system status information and executing optimization strategies issued by the reinforcement learning module. The reinforcement learning module dynamically optimizes the monitoring strategy of the entire system through interaction with the environment, achieving an optimal balance between data accuracy, energy consumption, and latency. The reinforcement learning module feeds back the optimized strategies of the learned parameters to the sensor nodes and the edge computing module. The data transmission and storage module transmits the encrypted data to a central server for long-term storage and analysis.

[0011] Furthermore, the sensor node includes a sensing unit, a signal conditioning unit, an analog-to-digital conversion unit, and a data acquisition and control unit. The sensing unit uses built-in sensors to comprehensively acquire the state of the fluidized solidified soil and outputs an analog signal to the signal conditioning unit. The signal conditioning unit filters and amplifies the analog signal output by the sensing unit and then inputs it to the analog-to-digital conversion unit. The analog-to-digital conversion unit converts the analog signal into a digital signal and then inputs it to the data acquisition and control unit. The data acquisition and control unit is responsible for scheduling sampling, data buffering, and performing mean filtering.

[0012] Furthermore, the data acquisition and control unit includes a microprocessor and a storage module. The microprocessor samples different channels according to the sampling strategy and temporarily stores the sampled data in the storage module. The sampled data is organized in the form of a time series, and each group of data includes a timestamp and the corresponding sensing value.

[0013] Furthermore, the expression for the mean filtering performed by the data acquisition and control unit is:

[0014]

[0015] In the formula, For at any time The filtered output signal value; These are the original sampled values ​​from the first N times; The sampling interval; To smooth the window length; For the first Each sampling time point; .

[0016] Furthermore, the edge computing module includes a data processing unit, an anomaly detection unit, and an encryption unit. The data processing unit is used to preprocess the received raw collected data; the anomaly detection unit detects abnormal values ​​in the data in real time based on statistical thresholds and triggers an alarm; the encryption unit is used to encrypt the processed data.

[0017] Furthermore, the encryption unit employs a homomorphic encryption algorithm, and the encrypted data satisfies the following characteristics:

[0018]

[0019] In the formula, For encryption functions; and At time respectively Two different raw data sets were collected; For the first Each sampling time point.

[0020] Furthermore, the reinforcement learning module includes an environment model, an action space, a reward mechanism, a learning algorithm, and a policy update unit. The environment model constructs a state space represented by the physical parameters of the sensor nodes and their current working state. The action space defines executable operations. The reward mechanism is used to design a reward function to provide feedback on the system's operational performance based on data quality, energy efficiency, and latency. The learning algorithm uses the Q-learning algorithm to learn the optimal policy. The policy update unit updates and issues new control policies periodically based on the reinforcement learning results.

[0021] Furthermore, the expression for the reward function is:

[0022]

[0023] In the formula, , and These are weighting coefficients, representing the relative importance of data accuracy, energy efficiency, and latency, respectively. , and Representing time respectively Data accuracy, energy consumption, and latency.

[0024] Furthermore, the expression for the Q-learning algorithm is:

[0025]

[0026] In the formula, The Q-value of the current state-action pair; The learning rate; This is a discount factor, indicating the importance of future rewards; This is the current reward value; The maximum Q value for the next state.

[0027] Furthermore, the data transmission and storage module includes a data transmission unit, an encrypted transmission mechanism, a data storage unit, and a data backup management unit. The data transmission unit uses a compression algorithm to transmit the encrypted data generated by the edge computing module to the central server. The encrypted transmission mechanism uses symmetric encryption, asymmetric encryption, and digital signatures to encrypt the transmitted data. The data storage unit uses distributed and multi-level data storage and associates each dataset with a unique identifier for traceability and indexing. The data backup management unit is used for redundant data backup and log recording.

[0028] The beneficial effects of this invention are:

[0029] 1. This invention enables local data processing and preliminary analysis by deploying edge computing modules on each sensor node, reducing reliance on a central server, significantly reducing data transmission and processing latency, and improving system response speed and real-time performance.

[0030] 2. This invention introduces a reinforcement learning algorithm, enabling the system to automatically adjust the sampling frequency and data processing strategy based on environmental changes in the fluidized solidified soil. This adaptive adjustment capability not only optimizes the accuracy of data acquisition and processing but also reduces energy consumption and improves the overall performance of the system.

[0031] 3. This invention employs homomorphic encryption technology to encrypt data and ensures the confidentiality and integrity of data during collection, transmission, and storage through a secure encrypted transmission protocol. This encryption mechanism prevents data tampering or leakage, ensuring the security of the monitoring data. Attached Figure Description

[0032] Figure 1 This is a structural block diagram of the data security detection system for fluidized solidified soil monitoring in this embodiment;

[0033] Figure 2 This is a flowchart of data in the data security detection system for fluidized solidified soil monitoring in this embodiment;

[0034] Figure 3 This is a flowchart of a data security detection system for fluidized solidified soil monitoring in this embodiment.

[0035] Reference numerals: Sensor node 1, Sensing unit 11, Signal conditioning unit 12, Analog-to-digital conversion unit 13, Data acquisition and control unit 14, Edge computing module 2, Data processing unit 21, Anomaly detection unit 22, Encryption unit 23, Reinforcement learning module 3, Environment model 31, Action space 32, Learning algorithm 33, Reward mechanism 34, Policy update unit 35, Data transmission and storage module 4, Data transmission unit 41, Encrypted transmission mechanism 42, Data storage unit 43, Data backup management unit 44. Detailed Implementation

[0036] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0037] Example: A data security monitoring system for fluidized solidified soil, such as... Figures 1-3 As shown, it includes sensor node 1, edge computing module 2, reinforcement learning module 3, and data transmission and storage module 4.

[0038] Among them, sensor node 1 is used to continuously collect multiple physical quantities in the target area during the construction and solidification process of fluidized solidified soil, and connects to edge computing module 2 through communication interface unit to transmit the raw collected data to it, while its working parameters (such as sampling frequency) are controlled by reinforcement learning module 3.

[0039] Specifically, sensor node 1 includes a sensing unit 11, a signal conditioning unit 12, an analog-to-digital conversion unit 13, and a data acquisition and control unit 14. Each unit realizes data interaction and command transmission through an internal bus circuit.

[0040] The sensing unit 11 selects different types of sensors according to monitoring requirements, such as temperature sensors, humidity sensors, pore water pressure sensors, and strain sensors. The temperature sensor is used to acquire real-time temperature signals of the fluidized solidified soil; the humidity sensor is used to detect changes in moisture content during the solidification process; the pore water pressure sensor is used to monitor changes in pore water pressure over time; and the strain sensor is used to obtain the deformation of the solidified soil. The sensing unit 11 utilizes its built-in sensors to comprehensively acquire the state of the fluidized solidified soil and outputs analog signals to the signal conditioning unit 12.

[0041] The signal conditioning unit 12 is used to filter and amplify the analog signal output by the sensing unit 11 to ensure that the signal amplitude and frequency characteristics meet the input requirements of the analog-to-digital conversion unit 13.

[0042] The analog-to-digital converter 13 is used to convert analog signals into digital signals and sets the output resolution to 12 bits or 16 bits to ensure numerical accuracy. The analog-to-digital converter 13 is connected to the data acquisition and control unit 14 via the SPI bus. The sampling period is triggered by a timing signal generated by the system clock. In each sampling period, the analog-to-digital converter 13 completes the sampling operation of each sensing channel in a set order.

[0043] The data acquisition and control unit 14 is responsible for performing sampling scheduling and data caching operations. Internally, it includes a microprocessor and a storage module. The microprocessor samples different channels according to a preset sampling strategy and temporarily stores the sampling results in the storage module. The sampled data is organized in time-series format, with each set of data containing a timestamp and the corresponding sensor; for example, let's say at time... The collected parameters are temperature. ,humidity ,pressure ,strain Then the set of data from a single sampling is represented as:

[0044]

[0045] In the formula, The sampling sequence number; For the first Secondary sampling data.

[0046] To reduce the impact of environmental noise, the data acquisition and control unit 14 performs a mean filtering operation on each sampling channel before storage. The expression for mean filtering is:

[0047]

[0048] In the formula, For at any time The filtered output signal value; These are the original sampled values ​​from the first N times; The sampling interval; To smooth the window length; For the first Each sampling time point; By calculating this equation, measurement errors caused by short-term fluctuations can be effectively eliminated.

[0049] The filtered digital signal is transmitted to edge computing module 2 via the communication interface unit. The communication interface unit uses a UART or RS485 standard interface to realize data communication between nodes, and the communication rate can be set within the range of 9600 to 115200 baud rate. Each data frame includes a start bit, a data field, a check field, and an end bit. The check field is generated using the CRC16 algorithm and is used to verify the integrity of data transmission.

[0050] When the system is running, the data acquisition and control unit 14 adjusts the sampling frequency. Control sampling timing, In the formula, The sampling period is expressed in seconds. The sampling frequency is set during the initialization phase based on the monitoring target and soil solidification characteristics, and can be dynamically adjusted by subsequent reinforcement learning strategies. All collected data is written to a circular buffer in chronological order. When the buffer is full, an overwrite operation is performed to ensure continuous data updates.

[0051] The structure and acquisition process of sensor node 1 enable the synchronous acquisition and preliminary processing of multiple parameters of fluidized solidified soil, providing input data for the subsequent anomaly detection and encryption calculation of edge computing module 2.

[0052] Furthermore, such as Figure 1 As shown, the edge computing module 2 is integrated into each sensor node 1 to perform real-time data preprocessing, quality monitoring, and security encryption, and then sends the processed encrypted data to the data transmission and storage module 4. The edge computing module 2 not only significantly reduces data transmission latency but also monitors data quality in real time, automatically identifies and processes abnormal data, ensuring the efficient and secure operation of the system. Simultaneously, the edge computing module 2 interacts bidirectionally with the reinforcement learning module 3, providing system status information and executing optimization strategies issued by the reinforcement learning module 3.

[0053] The edge computing module 2 includes a data processing unit 21, an anomaly detection unit 22, and an encryption unit 23. The data processing unit 21 is used to preprocess the received raw collected data, specifically:

[0054] Each set of data from sensor node 1 Each includes output values ​​from multiple sensors, such as temperature. ,humidity ,pressure and strain The data processing unit 21 preprocesses these data, including denoising, normalization, smoothing, and feature extraction. For each parameter, denoising is performed first; common denoising methods include Kalman filtering and moving average filtering.

[0055] Specifically, for the data acquired by each sensor, a Kalman filter is used for dynamic estimation, assuming that at time... The sensor's sampling data is Furthermore, the data is affected by Gaussian noise. The Kalman filter, influenced by the current time step, recursively corrects the predicted value at the current moment. The filtered data... It can be calculated using the following recursive formula:

[0056]

[0057] In the formula, Kalman gain represents the degree to which the model corrects the error between the predicted data and the actual sampled data. This is the estimated value from the previous moment; This is the raw data at the current moment.

[0058] After noise reduction, the data is normalized. The data from each sensor is normalized based on its historical maximum and minimum values, using the following formula:

[0059]

[0060] In the formula, These are the original sampled values. and These are the minimum and maximum values ​​of historical data, and the normalized data. It will be converted to a value between 0 and 1 for easier subsequent processing.

[0061] After denoising and normalization, the data enters the anomaly detection unit 22. This unit is responsible for detecting any outliers in the sensor-acquired data. Anomaly detection is based on a set statistical threshold. If a certain collected value is different from the historical average If the difference exceeds this threshold, it is considered abnormal. The specific formula is as follows:

[0062]

[0063] In the formula, The difference between the current time and the historical average; if this difference exceeds a set threshold. If the data is not found at that time, it is considered abnormal, the system will trigger an alarm and mark the data at that moment as abnormal.

[0064] After anomaly detection, encryption unit 23 encrypts the processed data. Encryption unit 23 employs a homomorphic encryption algorithm to ensure that the encrypted data can still be used for addition and multiplication calculations during transmission. The encrypted data satisfies the following characteristics:

[0065]

[0066] In the formula, For encryption functions; and At time respectively Two different raw data sets were collected; For the first Each sampling time point.

[0067] This encryption method ensures that even if the data is intercepted, an unauthorized third party cannot decrypt it, thus effectively guaranteeing data security and privacy protection.

[0068] After completing the above data processing, edge computing module 2 sends the encrypted data to the host computer or central server through the communication interface unit. The data is protected during transmission using advanced encryption protocols (such as TLS or SSL) to ensure data security.

[0069] like Figure 2 and Figure 3 As shown, the overall data processing flow of edge computing module 2 is as follows: data reception → noise reduction → normalization → anomaly detection → encryption → data transmission. Edge computing module 2 effectively achieves real-time data processing and early warning, while ensuring the security of data transmission. Through the implementation of edge computing module 2, not only is the burden on the central server reduced, but data can also be initially processed and filtered locally, improving the system's response speed and real-time performance.

[0070] Furthermore, such as Figure 1 As shown, reinforcement learning module 3 is a key component of the system, designed to dynamically optimize the monitoring strategy of sensor node 1 based on environmental changes, thereby enhancing the system's adaptability and the accuracy of data acquisition. Through intelligent algorithms of reinforcement learning, the system can adjust the sampling frequency in real time, select the optimal monitoring parameters, and optimize the sampling methods of different sensors, thus achieving precise monitoring of fluidized solidified soil.

[0071] Reinforcement learning module 3 interacts with the system's actual environment, gradually adjusting parameter configurations and monitoring strategies to ensure efficient and accurate monitoring. Reinforcement learning module 3 includes an environment model 31, an action space 32, a reward mechanism 34, a learning algorithm 33, and a policy update unit 35.

[0072] During system operation, reinforcement learning module 3 receives data collected by sensor node 1 and system state information as input to the reinforcement learning model. Specifically, environment model 31 constructs a state space represented by the physical parameters of sensor node 1 (such as temperature, humidity, pressure, and strain) and the node's current operating state (such as battery level and signal strength). The state space is represented as follows:

[0073]

[0074] In the formula, For a moment The state vector contains temperature. ,humidity ,pressure ,strain Battery power and signal strength .

[0075] In reinforcement learning, the action space 32 is defined as the set of all operations that the system can perform. These operations include adjusting the sampling frequency, selecting different sensor sampling modes, configuring data processing algorithms, and adjusting data transmission strategies. For example, the action space for sampling frequency... It can be defined as:

[0076]

[0077] In the formula, For each possible sampling frequency value This corresponds to a specific time interval. Through action space 32, the system can select an appropriate sampling frequency based on the current state.

[0078] The goal of reinforcement learning is to maximize long-term cumulative reward. The reward mechanism is used to design the reward function, which provides feedback on the system's operational performance based on data quality, energy efficiency, and latency. The reward function is designed based on the system's efficiency and the accuracy of the monitoring data, and can be defined according to the following aspects:

[0079] 1. Data Quality: The accuracy and stability of the data. The system receives a higher reward if the data error is small.

[0080] 2. Energy Efficiency: Optimize node battery consumption and network bandwidth utilization while ensuring data quality. The system controls energy consumption by adjusting the sampling frequency; too high a frequency may lead to rapid battery depletion, while too low a frequency may affect data accuracy.

[0081] 3. Latency Optimization: Low latency is crucial for improving system real-time performance during data processing and transmission. Therefore, the system reduces data transmission latency by intelligently adjusting the sampling frequency and transmission strategy.

[0082] The expression for the reward function is:

[0083]

[0084] In the formula, , and These are weighting coefficients, representing the relative importance of data accuracy, energy efficiency, and latency, respectively. , and Representing time respectively Data accuracy, energy consumption, and latency.

[0085] In reinforcement learning, learning algorithm 33 utilizes the Q-learning algorithm to learn the optimal policy. The Q-learning algorithm learns the optimal policy by analyzing each state-action pair. Calculate the Q-value, which represents the expected cumulative reward obtained by taking a certain action in a given state. The expression for the Q-learning algorithm is:

[0086]

[0087] In the formula, The Q-value of the current state-action pair; The learning rate; This is a discount factor, indicating the importance of future rewards; This is the current reward value; The maximum Q value for the next state.

[0088] By continuously updating the Q-value, reinforcement learning module 3 can learn the optimal sampling strategy, enabling the system to maintain the best monitoring performance in changing environments.

[0089] In reinforcement learning, the policy update unit 35 periodically updates and issues new control policies based on the reinforcement learning results, such as periodically adjusting the sampling frequency, sampling method, and data transmission frequency to minimize energy consumption, reduce latency, and maximize data quality. Specifically, when the system detects that the solidification process has entered a relatively stable stage, the reinforcement learning module 3 may reduce the sampling frequency to reduce battery consumption; while when drastic changes occur in the soil, it may increase the sampling frequency to improve data accuracy.

[0090] Based on reinforcement learning module 3, the system can dynamically adjust the monitoring strategy according to the real-time changes of the fluidized solidified soil, thereby achieving precise control of the monitoring task, enhancing the system's intelligence and adaptability, and ensuring the efficiency and accuracy of the fluidized solidified soil monitoring process.

[0091] Furthermore, such as Figure 1 As shown, the data transmission and storage module 4 is used to realize data communication and storage management between sensor node 1 and the central server. This module ensures the secure and stable transmission of data during the monitoring of fluidized solidified soil and effectively stores the collected data for subsequent analysis and processing. The data transmission and storage module 4 includes a data transmission unit 41, an encrypted transmission mechanism 42, a data storage unit 43, and a data backup management unit 44. Specifically:

[0092] The data transmission unit 41 is responsible for transmitting the encrypted data generated by the edge computing module 2 to the central server. To ensure the stability and efficiency of data transmission, wireless communication technologies (such as Wi-Fi, LoRa, 5G, etc.) are used. The frequency of data transmission is dynamically adjusted according to the real-time requirements of the system and network bandwidth conditions. To reduce network bandwidth consumption, the data transmission unit 41 uses a compression algorithm to compress the data, and the compressed data packets are sent to the host computer or central server through the wireless communication module.

[0093] During data transmission, to ensure data confidentiality and integrity, the encryption transmission mechanism 42 performs encryption processing in the following ways: During the data acquisition and preprocessing stage, the encryption unit 23 of the edge computing module 2 encrypts the data, using either symmetric or asymmetric encryption methods for data protection. For large volumes of real-time monitoring data, AES (Advanced Encryption Standard) or RSA (Public Key Segregation Algorithm) is used for encryption.

[0094] The encryption process of symmetric encryption is as follows: Before data transmission, sensor node 1 generates a random symmetric key. The key is then used to encrypt data D; the encryption expression is:

[0095]

[0096] In the formula, This represents the ciphertext obtained by symmetric encryption of data D. ; The symmetric key used for encryption.

[0097] The asymmetric encryption process is as follows: For situations requiring authentication during communication, the system uses an asymmetric encryption algorithm for public-key encryption. The central server's public key... Symmetric key used for encrypting data transmission The encrypted key is transmitted to the receiving end through a secure channel, and the receiving end uses the private key. Decryption yields the symmetric key for data transmission. .

[0098] The encrypted data is sent to the server via data transmission unit 41. After receiving the encrypted data, the server restores the data using a decryption algorithm and verifies the integrity of the data to ensure that no data is lost or tampered with during transmission.

[0099] To further ensure data security and reliability, the system also employs digital signature technology to guarantee the integrity and authenticity of transmitted data. Before data transmission, the data acquisition unit generates a hash value for the data content and encrypts it using a private key, generating a digital signature. Upon receiving the data, the receiving end verifies it using the public key to ensure that the data has not been tampered with.

[0100] During data storage, data storage unit 43 adopts a distributed storage architecture and manages large-scale data through a multi-level storage structure. Data transmitted to the central server first enters the cache, and then, based on data type and storage strategy, is allocated to appropriate storage media (such as hard drives, solid-state drives, or cloud storage). Each dataset is associated with a unique identifier for traceability and indexing. Simultaneously, during database design, the system uses timestamps to mark each dataset, ensuring the integrity of the data in chronological order.

[0101] Data storage unit 43 employs a distributed database management system, such as Apache Cassandra or HBase, which can dynamically expand storage capacity as needed. Distributed storage not only improves data access efficiency but also ensures high data availability and fault tolerance. Furthermore, when a storage node fails, the system automatically migrates data to other nodes for backup, ensuring no data loss.

[0102] During data storage, the data backup management unit 44 employs a redundant backup mechanism, periodically backing up stored data to a remote server or cloud platform to prevent data loss. In the event of system failure or disaster, recovery can be achieved using the backup data. The backup cycle can be flexibly adjusted according to needs, typically daily, weekly, or on-demand.

[0103] To enhance data traceability and manageability, the data backup management unit 44 also features logging and auditing capabilities. Every data read, modification, and storage operation is recorded in a log file, including information such as the operation time, operator, and operation type. All operation logs are encrypted and accessible only to authorized personnel.

[0104] By designing the data transmission and storage module 4, the security, integrity, and reliability of the monitoring data of fluidized solidified soil can be effectively guaranteed, ensuring that the data is not tampered with or lost during the monitoring process, and providing efficient and reliable data support for subsequent data analysis.

[0105] The above description is merely a preferred embodiment of the present invention, and the scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principle of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A data security detection system for monitoring fluidized solidified soil, characterized in that, The system includes a sensor node (1), an edge computing module (2), a reinforcement learning module (3), and a data transmission and storage module (4). The sensor node (1) is used to continuously collect multiple physical quantities in the target area during the construction and solidification process of the fluidized solidified soil and transmit the raw collected data to the edge computing module (2). The edge computing module (2) is integrated into the sensor node (1) and is used to realize real-time preprocessing, quality monitoring and security encryption of the data, and send the processed encrypted data to the data transmission and storage module (4). The edge computing module (2) interacts bidirectionally with the reinforcement learning module (3), provides system status information to the reinforcement learning module (3), and executes the optimization strategy issued by the reinforcement learning module (3). The reinforcement learning module (3) dynamically optimizes the monitoring strategy of the entire system by interacting with the environment, so as to achieve the best balance between data accuracy, energy consumption and latency. The reinforcement learning module (3) feeds back the optimization strategy of the learned parameters to the sensor node (1) and the edge computing module (2). The data transmission and storage module (4) is used to transmit the encrypted data to the central server for long-term storage and analysis.

2. The data security detection system for monitoring fluidized solidified soil according to claim 1, characterized in that, The sensor node (1) includes a sensing unit (11), a signal conditioning unit (12), an analog-to-digital converter (13), and a data acquisition and control unit (14). The sensing unit (11) uses its built-in sensors to fully acquire the state of the fluidized solidified soil and outputs an analog signal to the signal conditioning unit (12). The signal conditioning unit (12) filters and amplifies the analog signal output by the sensing unit (11) and then inputs it to the analog-to-digital converter (13). The analog-to-digital converter (13) converts the analog signal into a digital signal and then inputs it to the data acquisition and control unit (14). The data acquisition and control unit (14) is responsible for scheduling sampling, data buffering, and performing mean filtering.

3. The data security detection system for monitoring fluidized solidified soil according to claim 2, characterized in that, The data acquisition control unit (14) contains a microprocessor and a storage module. The microprocessor samples different channels according to the sampling strategy and temporarily stores the sampled data in the storage module. The sampled data is organized in the form of a time series, and each group of data contains a timestamp and the corresponding sensing quantity.

4. The data security detection system for monitoring fluidized solidified soil according to claim 2, characterized in that, The expression for the mean filtering performed by the data acquisition control unit (14) is: In the formula, For at any time The filtered output signal value; These are the original sampled values ​​from the first N times; The sampling interval; To smooth the window length; For the first Each sampling time point; .

5. The data security detection system for monitoring fluidized solidified soil according to claim 1, characterized in that, The edge computing module (2) includes a data processing unit (21), an anomaly detection unit (22), and an encryption unit (23). The data processing unit (21) is used to preprocess the received raw collected data; the anomaly detection unit (22) detects abnormal values ​​in the data in real time based on statistical thresholds and triggers an alarm; the encryption unit (23) is used to encrypt the processed data.

6. The data security detection system for monitoring fluidized solidified soil according to claim 5, characterized in that, The encryption unit (23) uses a homomorphic encryption algorithm, and the encrypted data satisfies the following characteristics: In the formula, For encryption functions; and At time respectively Two different raw data sets were collected; For the first Each sampling time point.

7. The data security detection system for monitoring fluidized solidified soil according to claim 1, characterized in that, The reinforcement learning module (3) includes an environment model (31), an action space (32), a reward mechanism (34), a learning algorithm (33), and a policy update unit (35). The environment model (31) constructs a state space represented by the physical parameters of the sensor node (1) and the current working state of the node. The action space (32) defines executable operations. The reward mechanism (34) is used to design a reward function to provide feedback on the system operation effect based on data quality, energy efficiency, and latency. The learning algorithm (33) uses the Q-learning algorithm to learn the optimal policy. The policy update unit (35) updates and issues new control policies periodically based on the reinforcement learning results.

8. The data security detection system for monitoring fluidized solidified soil according to claim 7, characterized in that, The expression for the reward function is: In the formula, , and These are weighting coefficients, representing the relative importance of data accuracy, energy efficiency, and latency, respectively. , and Representing time respectively Data accuracy, energy consumption, and latency.

9. The data security detection system for monitoring fluidized solidified soil according to claim 7, characterized in that, The expression for the Q-learning algorithm is: In the formula, The Q-value of the current state-action pair; The learning rate; This is a discount factor, indicating the importance of future rewards; This is the current reward value; The maximum Q value for the next state.

10. The data security detection system for monitoring fluidized solidified soil according to claim 1, characterized in that, The data transmission and storage module (4) includes a data transmission unit (41), an encrypted transmission mechanism (42), a data storage unit (43), and a data backup management unit (44). The data transmission unit (41) uses a compression algorithm to transmit the encrypted data generated by the edge computing module (2) to the central server. The encrypted transmission mechanism (42) uses symmetric encryption, asymmetric encryption, and digital signature to encrypt the transmitted data. The data storage unit (43) uses distributed and multi-level data storage and associates each dataset with a unique identifier for traceability and indexing. The data backup management unit (44) is used for redundant backup of data and log recording.

Images