Networked system operating state estimation method and related apparatus

Abstract

Embodiments of the present application provide a networked system operation state estimation method and related device, in particular to the field of data processing technology. The method comprises: obtaining sensor measurement data of the networked system within a preset time window at the current time and before the current time, and obtaining a historical state estimation value at the starting point of the preset time window. Determine the threat type identifier corresponding to the sensor measurement data at each time based on the preset abnormal measurement model. Data preprocessing is performed on all sensor measurement data after threat type identification and classification to obtain target measurement data. Based on the target measurement data and the historical state estimation value, a rolling horizon optimization problem is constructed and solved to obtain an optimal state estimation value sequence at each time within the preset time window. The running state estimation result of the networked system at the current time is extracted from the optimal state estimation value sequence. The present application can improve the estimation accuracy of the running state of the networked system.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and more specifically, to a method and related apparatus for estimating the operating status of a networked system. Background Technology

[0002] Networked systems are complex systems that connect devices, platforms, and terminals through communication networks to achieve resource sharing, information exchange, and collaborative operation, such as intelligent transportation and the Industrial Internet. Networked systems typically rely on communication networks to enable data exchange between sensors and estimation centers, and the accuracy of their remote state estimation is crucial for the safe operation of the system.

[0003] Currently, the mainstream approach is based on Kalman filtering and its variants, using recursive updates of prediction and measurement information to achieve real-time estimation of the system state. However, when measurement data is simultaneously subjected to the combined effects of multiple network threats such as malicious tampering, communication delays, and packet loss, the above methods struggle to effectively identify and suppress the impact of abnormal data. This leads to a significant decrease in the accuracy of networked system state estimation, making it difficult to accurately estimate the system's operating state in real-time under complex network environments. Summary of the Invention

[0004] The present invention aims to, for example, provide a method and related apparatus for estimating the operating state of a networked system, which can at least partially solve the above-mentioned technical problems.

[0005] The embodiments of the present invention can be implemented as follows: In a first aspect, embodiments of the present invention provide a method for estimating the operating state of a networked system, the method comprising: The system acquires sensor measurement data of the networked system within a preset time window at the current time and before, and acquires the historical state estimate of the starting point of the preset time window, wherein the historical state estimate is the prediction estimate of the starting point state of the preset time window in the previous round. Based on a preset abnormal measurement model, the sensor measurement data is used to identify and classify threat types to determine the threat type identifier corresponding to the sensor measurement data at each time. All sensor measurement data after threat type identification and classification are preprocessed to obtain target measurement data; Based on the target measurement data and the historical state estimates, a rolling time-domain optimization problem is constructed and solved to obtain the optimal state estimate sequence for each moment within the preset time window. Extract the state estimate corresponding to the current moment from the optimal state estimate sequence, and use it as the operating state estimate result of the networked system at the current moment.

[0006] Optionally, the threat types include communication latency, packet loss, and data tampering. The step of identifying and classifying the threat types of the sensor measurement data based on a preset abnormal measurement model to determine the threat type identifier corresponding to the sensor measurement data at each time point includes: The preset anomaly measurement model is invoked, and the preset anomaly measurement model is as follows:

[0007] Where k is the current time, The first before the network transmission anomaly Constantly receiving normal sensor measurement data. , They are respectively in The sensor measurement data actually received at time k and time k that were affected by network anomalies. In order to be in Sensor measurement data at any given time. Y represents the number of communication delay steps. k Let k be the tampering offset value injected into the sensor measurement data at time k. , as well as All are binary variables, taking values ​​of 1 or 0. , , , , Υ k , , For the set of real numbers, It is an n-dimensional column vector. Determined by timestamp, Indicates no delay. Indicates a time delay. Determined by the data packet reception status. This indicates that the data packet was successfully received. Indicates packet loss. Determined by saturation method, This indicates that data tampering has occurred. This indicates that no data tampering has occurred; when =1、 When = 0, if If the value is 0, then the communication delay is determined to occur in the sensor measurement data at the current moment. when =1、 When = 0, if If the value is 0, it is determined that the sensor measurement data at the current moment has been lost. when =1、 When =1, if If the value is 1, then it is determined that the sensor measurement data at the current moment has been tampered with, and the tampering offset value is Y. k ; when =1、 When = 0, if If the value is 1, then it is determined that the sensor measurement data at the current moment has experienced both the communication delay and the data tampering, and the tampering offset value is Y. k .

[0008] Optionally, the step of preprocessing all sensor measurement data after threat type identification and classification to obtain target measurement data includes: Define a cached data set to store abnormal measurement data and historical state estimates within a preset time window period prior to the current time. The cached data set is represented as follows:

[0009] in, The cached data set at time k, The preset time window length, They represent The moment arrive State estimate at time 10:00 This represents the initial state prediction value at the start point of the preset time window. , ; Abnormal measurement data in the cached dataset is truncated based on the following truncation mechanism:

[0010] in, To truncate the processed sensor measurement data, , As a preset upper bound for the delay, , ; The truncated sensor measurement data is saturated to obtain the target measurement data.

[0011] Optionally, the saturation processing of the truncated sensor measurement data includes: Calculate the predicted measurement value corresponding to the sensor measurement data at each moment within the preset time window; Calculate the residual between the measurement data of each sensor and the corresponding measurement prediction value; For sensor measurement data where the Euclidean norm of the residual exceeds a preset residual threshold, the value of the sensor measurement data is compressed toward the corresponding measurement prediction value until the Euclidean norm of the residual equals the preset residual threshold, thus completing the saturation process.

[0012] Optionally, the step of constructing and solving a rolling time-domain optimization problem based on the target measurement data and the historical state estimates to obtain the optimal state estimate sequence at each moment within the preset time window includes: A preset cost function is constructed as the objective function of the rolling time-domain optimization problem; Using the sequence of state estimates at each moment within the preset time window as optimization variables, and the state recursive equation and measurement output equation of the networked system as constraints, the cost function is minimized to obtain the optimal sequence of state estimates at each moment within the preset time window.

[0013] Optionally, the preset cost function is:

[0014] in, For the cost function in the rolling time domain, The historical state estimate is given. For the target measurement data, The predicted measurement value for the target. To adjust the parameters, P1 and P2 are positive definite weight matrices. For weighted norm, , , P1 P2 , Represents the set of real matrices with n rows and n columns.

[0015] Optionally, the method further includes: The operational state estimation result is saved as the historical state estimate at the starting point of the preset time window in the next round of estimation, and the cached data set is slid forward by one step as the preset time window progresses. The recursive formula is as follows:

[0016] Wherein, A and B are the state matrix and input matrix of the networked system, respectively. for Time-based control input; The steps of performing threat type identification and classification, data preprocessing, constructing and solving a rolling time-domain optimization problem, and extracting state estimates on sensor measurement data acquired in a new preset time window are used to continuously estimate the operating state of the networked system in real time.

[0017] Secondly, embodiments of the present invention provide a networked system operating state estimation device, the networked system operating state estimation device comprising: The data acquisition unit is used to acquire sensor measurement data of the networked system at the current moment and within a preset time window, and to acquire the historical state estimate of the starting point of the preset time window, wherein the historical state estimate is the prediction estimate of the starting point state of the preset time window in the previous round. The threat type identification unit is used to identify and classify the threat type of the sensor measurement data based on a preset abnormal measurement model, so as to determine the threat type identification corresponding to the sensor measurement data at each time. The data preprocessing unit is used to preprocess all sensor measurement data after threat type identification and classification to obtain target measurement data. The optimal state estimate sequence determination unit is used to construct and solve a rolling time-domain optimization problem based on the target measurement data and the historical state estimates, and obtain the optimal state estimate sequence at each moment within the preset time window. The operation state estimation result determination unit is used to extract the state estimation value corresponding to the current moment from the optimal state estimation value sequence, and use it as the operation state estimation result of the networked system at the current moment.

[0018] Thirdly, embodiments of the present invention provide an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of any of the methods described above.

[0019] Fourthly, embodiments of the present invention provide a computer-readable storage medium, the computer-readable storage medium including a computer program, wherein the computer program, when executed, controls a server where the computer-readable storage medium is located to implement the steps of any of the methods described above.

[0020] The beneficial effects of the embodiments of the present invention include, for example: By acquiring sensor measurement data and historical state estimates from the starting point within a preset time window, and then identifying the threat type of the data at each moment based on an anomaly measurement model, all measurement data are preprocessed to obtain target measurement data. A rolling time-domain optimization problem is then constructed and solved to obtain the optimal state estimate sequence within the window. Finally, the current moment estimate is extracted as the operating state result. This enables the estimation center to effectively identify and suppress the impact of abnormal data caused by the coupling effects of various network threats such as communication delay, packet loss, and malicious tampering. As a result, the accuracy and reliability of state estimation are improved in complex network environments, achieving real-time and accurate estimation of the operating state of networked systems. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 A block diagram illustrating an electronic device according to an embodiment of the present invention; Figure 2 A flowchart illustrating the steps of a networked system operation state estimation method provided in this embodiment of the invention; Figure 3 This invention provides a flowchart of steps for preprocessing sensor measurement data. Figure 4 This is an architectural diagram of a networked system operation status estimation device provided in an embodiment of the present invention.

[0023] Icons: 100 - Electronic device; 110 - Memory; 120 - Processor; 130 - Communication module; 300 - Networked system operation status estimation device; 301 - Data acquisition unit; 302 - Threat type identification unit; 303 - Data preprocessing unit; 304 - Optimal state estimation value sequence determination unit; 305 - Operation status estimation result determination unit. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0025] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0026] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0027] Furthermore, the terms "first" and "second" are used only to distinguish descriptions and should not be interpreted as indicating or implying relative importance.

[0028] It should be noted that, where there is no conflict, the features in the embodiments of the present invention can be combined with each other.

[0029] Please refer to Figure 1 This is a block diagram of an electronic device 100 provided in this application. The electronic device 100 can be a device capable of data processing, and this embodiment does not limit this. The electronic device 100 includes a memory 110, a processor 120, and a communication module 130. The components are electrically connected directly or indirectly to achieve data transmission or interaction. For example, these components can be electrically connected to each other through one or more communication buses or signal lines.

[0030] The memory 110 is used to store programs or data. The memory 110 may be, but is not limited to, random access memory (RAM), read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), etc.

[0031] The processor 120 is used to read / write data or programs stored in the memory and to perform corresponding functions.

[0032] The communication module 130 is used to establish a communication connection between the server and other communication terminals through the network, and to send and receive data through the network.

[0033] It should be understood that, Figure 1 The structure shown is only a schematic diagram of the electronic device 100. The electronic device 100 may also include components that are larger than... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown. Figure 1 The components shown can be implemented using hardware, software, or a combination thereof. The electronic device 100 can be integrated into other devices or configured as a stand-alone device.

[0034] This invention provides a method for estimating the operating state of a networked system, which can be applied to an electronic device 100 for real-time estimation of the operating state of the networked system. The method includes, for example: Figure 2 The following steps are shown: Step S110: Obtain sensor measurement data of the networked system within the current time and the preset time window before it, and obtain the historical state estimate of the starting point of the preset time window. The historical state estimate is the prediction estimate of the starting point state of the preset time window in the previous round.

[0035] Step S120: Based on a preset abnormal measurement model, the sensor measurement data is subjected to threat type identification and classification to determine the threat type identifier corresponding to the sensor measurement data at each time.

[0036] Step S130: Perform data preprocessing on all sensor measurement data after threat type identification and classification to obtain target measurement data.

[0037] Step S140: Based on the target measurement data and the historical state estimates, construct and solve the rolling time domain optimization problem to obtain the optimal state estimate sequence for each moment within the preset time window.

[0038] Step S150: Extract the state estimate corresponding to the current moment from the optimal state estimate sequence, and use it as the operating state estimate result of the networked system at the current moment.

[0039] In step S110, sensor measurement data of the networked system within the current time and the preset time window are obtained, and the historical state estimate of the starting point of the preset time window is obtained. The historical state estimate is the prediction estimate of the starting point state of the preset time window in the previous round.

[0040] Sensor measurement data refers to the raw measurement signals collected in real time by various sensors (such as position sensors, inertial measurement units, temperature sensors, pressure sensors, etc.) deployed in a networked system during operation and transmitted to a remote state estimation center via a communication network. For example, in a vehicle motion tracking scenario, sensor measurement data may include the vehicle's longitudinal coordinates, lateral coordinates, and heading angle. In a drone navigation scenario, it may include GPS latitude and longitude coordinates, altitude, three-axis attitude angles, and angular velocity. In an industrial process monitoring scenario, it may include process variables such as reactor temperature, pipeline pressure, and flow meter readings.

[0041] When estimating the operational state of a networked system in real time, the system can simultaneously collect sensor measurement data at the current moment, as well as historical measurement data within a preset time window prior to the current moment. It can also extract the historical state estimate corresponding to the starting point of this preset time window. It's important to note that the historical state estimate is not an arbitrarily selected initial value, but rather derived from the optimized calculation results of the networked system state at the same moment in the previous estimation cycle. This result serves as prior information input to the current estimation cycle, providing a reliable initial reference benchmark for subsequent optimization. The length of the preset time window determines the amount of sensor measurement data involved in each optimization. Taking an "∞"-shaped trajectory tracking task as an example, in this step, the estimator needs to acquire the horizontal and vertical coordinates and orientation angle measurement data for the current and several past sampling moments (corresponding to the preset time window length), while simultaneously extracting the state estimate result obtained from the previous calculation at the window's starting point.

[0042] In step S120, based on a preset abnormal measurement model, the sensor measurement data is used to identify and classify threat types to determine the threat type identifier corresponding to the sensor measurement data at each time point.

[0043] After data acquisition is complete, a preset anomaly measurement model can be invoked to perform point-by-point analysis on all collected sensor measurement data. This anomaly measurement model can identify whether each sensor measurement data has been affected by network threats such as communication delay, packet loss, or data tampering during transmission based on the temporal and statistical characteristics of the data, and mark each sensor measurement data with a corresponding threat type identifier (if there is no network threat, it is marked as normal or not marked).

[0044] Optionally, the threat types include communication latency, packet loss, and data tampering. The step of identifying and classifying the threat types of the sensor measurement data based on a preset abnormal measurement model to determine the threat type identifier corresponding to the sensor measurement data at each time point includes: The preset anomaly measurement model is invoked, and the preset anomaly measurement model is as follows:

[0045] Where k is the current time, The first time before the network transmission anomaly occurred Constantly receiving normal sensor measurement data. , They are respectively in The sensor measurement data actually received at time k and time k that were affected by network anomalies. In order to be in Sensor measurement data at any given time. Y represents the number of communication delay steps. k Let k be the tampering offset value injected into the sensor measurement data at time k. , as well as All are binary variables, taking values ​​of 1 or 0. , , , , Υ k , , For the set of real numbers, It is an n-dimensional column vector. Determined by timestamp, Indicates no delay. Indicates that a delay has occurred. Determined by the data packet reception status. This indicates that the data packet was successfully received. Indicates packet loss. Determined by saturation method, This indicates that data tampering has occurred. This indicates that no data tampering has occurred; when =1、 When = 0, if If the value is 0, then the communication delay is determined to occur in the sensor measurement data at the current moment. when =1、 When = 0, if If the value is 0, it is determined that the sensor measurement data at the current moment has been lost. when =1、 When =1, if If the value is 1, then it is determined that the sensor measurement data at the current moment has been tampered with, and the tampering offset value is Y. k ; when =1、 When = 0, if If the value is 1, then it is determined that the sensor measurement data at the current moment has experienced both the communication delay and the data tampering, and the tampering offset value is Y. k .

[0046] As an optional implementation, the expression for the preset anomaly measurement model can be: .

[0047] In the actual identification process, through interpretation , as well as The different combinations of values ​​for these three binary variables can accurately determine the type of threat the measurement data is facing at the current moment. When =1、 When =0, the abnormal measurement model degenerates into If at this time =0 means that the measurement data at the current moment has experienced a communication delay, and the estimated center actually receives the data after the delay. Historical data of steps .when =1、 When =0, the model degenerates into If at this time If the value is 0, it indicates that a data packet loss has occurred at the current moment, and the estimation center will use the abnormal measurement value from the previous moment. To replace the lost data. When =1、 When =1, the model degenerates into If at this time =1 indicates that the measurement data at the current moment has been maliciously tampered with, and the tampering offset value injected by the attacker is Y. k Similarly, when =1、 When = 0, if =1 indicates that both communication delay and data tampering occurred simultaneously, with the tampering offset value being Y. k Since changes in binary variables are used to identify anomalies in sensor measurement data, the numerical values ​​of binary variables can also be used to identify the threat type of abnormal sensor measurement data.

[0048] In step S130, all sensor measurement data after threat type identification and classification are preprocessed to obtain target measurement data.

[0049] After obtaining the threat type identifiers corresponding to the measurement data at each time point, data preprocessing operations can be performed on all identified measurement data to suppress the adverse effects of anomalous data on the estimation process. Specifically, for measurement values ​​with extreme deviations due to tampering, their destructiveness can be mitigated by constraining their amplitudes within a reasonable and feasible range determined by the system's dynamic characteristics. For information loss or delays caused by communication latency or packet loss, effective historical information from adjacent time points is used for compensation and filling. This transforms the original measurement data, containing various anomalies, into target measurement data with higher reliability and continuity, laying a data foundation for subsequent optimization estimation.

[0050] Optionally, the step of preprocessing all sensor measurement data after threat type identification and classification to obtain target measurement data includes: Define a cached data set to store abnormal measurement data and historical state estimates within a preset time window period prior to the current time. The cached data set is represented as follows:

[0051] in, The cached data set at time k, The preset time window length, They represent The moment arrive State estimate at time 10:00 This represents the initial state prediction value at the start point of the preset time window. , ; Abnormal measurement data in the cached dataset is truncated based on the following truncation mechanism:

[0052] in, To truncate the processed sensor measurement data, , As a preset upper bound for the delay, , ; The truncated sensor measurement data is saturated to obtain the target measurement data.

[0053] The steps for preprocessing sensor measurement data can be as follows: Figure 3 As shown, since it is impractical to store all measurement data and historical state estimates simultaneously in a real system, the above-mentioned cached data set can be defined. After data caching is completed, a data truncation mechanism can be introduced to handle measurement points within a preset time window whose predicted values ​​cannot be calculated using the system dynamics equations due to excessive time delay. The specific expression of this truncation mechanism is as shown above. Traversal from arrive Every moment. According to this truncation mechanism, when the first... The interval between the time and the start time of the window Not less than the preset upper limit of delay When an anomaly is detected, the abnormal measurement value at that moment is directly retained. In other cases, the measurement values ​​received and cached in the previous moment are used for retention; specifically, this means that... As part of the measurement value at that moment, the continuity and integrity of the entire optimization sequence in time are ensured. After the truncation process is completed, the truncated measurement data is further saturated to suppress extreme outliers caused by data tampering. By compressing the measurement values ​​that exceed the reasonable range toward the predicted values, the target measurement data that can be directly used by the optimization solver is finally obtained.

[0054] Optionally, the saturation processing of the truncated sensor measurement data includes: Calculate the predicted measurement value corresponding to the sensor measurement data at each moment within the preset time window.

[0055] Calculate the residual between the measurement data of each sensor and the corresponding measurement prediction value.

[0056] For sensor measurement data where the Euclidean norm of the residual exceeds a preset residual threshold, the value of the sensor measurement data is compressed toward the corresponding measurement prediction value until the Euclidean norm of the residual equals the preset residual threshold, thus completing the saturation process.

[0057] After data truncation, saturation processing can be applied to the sensor measurement data. The predicted measurement value corresponding to the sensor measurement data at each moment within a preset time window is calculated. This prediction is based on the system state estimate from the previous time step. The physical meaning of the measurement obtained through the system measurement equation lies in characterizing the reasonable expected range that the sensor measurement should be within at that moment under the current dynamic evolution law of the system.

[0058] After obtaining the predicted measurement values ​​at each time point, the actual measurement data at each time point is calculated. and the corresponding measurement prediction value The residual vector between, the residual The magnitude of the Euclidean norm value directly reflects the degree to which the actual measurement deviates from the expected value. For residuals whose Euclidean norm value does not exceed a preset residual threshold... The measured data were considered to be within the normal noise range and were therefore retained unchanged. It is a real number greater than 0. For measurement data where the Euclidean norm of the residual exceeds a preset residual threshold, its value is compressed towards the corresponding measurement prediction value, i.e., through a saturation function. The data is forcibly mapped back to a feasible region centered on the predicted value and with a preset threshold as its radius, until the compressed residual modulus is exactly equal to the preset residual threshold, thus completing the saturation processing of the abnormal measurement data. Let be a saturation function, defined as: .

[0059] This process can be indicated by a signal. To mark which data is saturated, where || || represents the Euclidean norm. The preset residual threshold can be reasonably selected based on the statistical level of measurement noise, and is usually set to be slightly higher than the amplitude of normal measurement noise, so as to effectively suppress the impact of attacks while avoiding excessive compression of normal measurement information.

[0060] It is understandable that saturation processing of truncated sensor measurement data does not only apply to sensor measurement data that meets the truncation conditions and has been truncated, but rather to all sensor measurement data that has passed the truncation condition judgment (i.e., normal measurement data, communication delay data, tampered data, and packet loss data).

[0061] In step S140, based on the target measurement data and the historical state estimates, a rolling time-domain optimization problem is constructed and solved to obtain the optimal state estimate sequence for each moment within the preset time window.

[0062] After preprocessing the sensor measurement data, a rolling time-domain optimization problem can be constructed based on the preprocessed target measurement data and the historical state estimates at the start of the window. In this rolling time-domain optimization problem, the system state at each moment within the entire preset time window can be used as the optimization variable, the historical state estimates can be used as the prior constraint term at the start of the window, and the target measurement data can be used as the measurement constraint term at each moment within the window.

[0063] By minimizing a cost function that comprehensively measures the deviation between the predicted and actual state values, the optimal state estimate sequence corresponding to each moment within a preset time window can be obtained by solving this rolling time domain optimization problem.

[0064] Optionally, the step of constructing and solving a rolling time-domain optimization problem based on the target measurement data and the historical state estimates to obtain the optimal state estimate sequence at each moment within the preset time window includes: A preset cost function is constructed as the objective function of the rolling time-domain optimization problem; Using the sequence of state estimates at each moment within the preset time window as optimization variables, and the state recursive equation and measurement output equation of the networked system as constraints, the cost function is minimized to obtain the optimal sequence of state estimates at each moment within the preset time window.

[0065] Optionally, the preset cost function is:

[0066] in, For the cost function in the rolling time domain, The historical state estimate is given. For the target measurement data, The predicted measurement value for the target. To adjust the parameters, P1 and P2 are positive definite weight matrices. For weighted norm, , , P1 P2 , Represents the set of real matrices with n rows and n columns.

[0067] In one alternative implementation, the discrete-time state-space model of the networked system can be expressed as: (State recurrence equation) and (Measurement output equation), where Let k be the system state vector at time k. To control input and The measurement vector output by the sensor. Let be the system state matrix, describing the linear transition relationship between states at adjacent time points. The input matrix describes the effect of the control input on the state evolution. This is the measurement output matrix of a single sensor, describing the mapping relationship between the system state and the sensor measurements. and The system noise and measurement noise are independent zero-mean Gaussian noise. Let be the number of sensors in the sensor network, then the measurement aggregation matrix of all sensors and the aggregation output of the measurement sensors are defined as follows: and . For the set of real numbers, for 3D real column vector Represents a real matrix. For the weighted norm, its expression is: = , Similarly.

[0068] A predefined cost function is constructed as the objective function for the rolling time-domain optimization problem. The specific form of this cost function is shown above. This cost function consists of a weighted sum of two terms, the first of which... The optimal estimate of the starting point of the penalty window is used to determine the cost of reaching the penalty window. Compared to its historical forecasts The deviation. Second item. The cost of execution is used to penalize the measurement predictions at each time point within the window. Measurement of actual target The deviation between them.

[0069] After constructing the cost function, the sequence of state estimates at each time point within a preset time window is used. The optimization variables to be solved are the state recursive equation and measurement output equation of the networked system shown above, which serve as constraints. Specifically, the constraint form of the measurement output equation depends on the type of threat faced by each measurement point: for delay steps less than a preset upper limit of delay... The measurement points are constrained as follows:

[0070] This represents the actual delay steps, and satisfies... ;the remaining In some cases, The above constraints ensure that the solved state estimate sequence conforms to the physical dynamic characteristics of the system; that is, the state estimates at adjacent times must satisfy the system state equations, and the measurement predictions at each time point must satisfy the modified measurement output equations according to the corresponding threat type. By solving this constrained linear quadratic optimization problem, the optimal state estimate sequence at each time point within the preset time window can be obtained.

[0071] In step S150, the state estimate value corresponding to the current moment is extracted from the optimal state estimate value sequence and used as the operating state estimate result of the networked system at the current moment.

[0072] The operational state of a networked system refers to a set of parameters describing the system's current situation. These parameters are typically components of the state vector in the system model and represent the final output of the state estimation. For example, in the motion scenario of a vehicle or drone, the operational state can include the system's lateral and longitudinal coordinates in a plane, as well as its current orientation angle. These three parameters together determine the system's position and attitude in the plane. In industrial process monitoring, the operational state can be represented by process variables such as the temperature, pressure, and liquid level of a reactor. In this embodiment of the invention, the operational state refers to system characteristic parameters that cannot be directly and accurately measured and need to be estimated using data from remote sensors.

[0073] After obtaining the optimal state estimate sequence, the state estimate corresponding to the current time can be extracted from the optimal state estimate sequence obtained by solving, and the estimate can be output as the operating state estimate result of the networked system at the current time.

[0074] Optionally, the method further includes: The operational state estimation result is saved as the historical state estimate at the starting point of the preset time window in the next round of estimation, and the cached data set is slid forward by one step as the preset time window progresses. The recursive formula is as follows:

[0075] Wherein, A and B are the state matrix and input matrix of the networked system, respectively. for Time-based control input; The steps of performing threat type identification and classification, data preprocessing, constructing and solving a rolling time-domain optimization problem, and extracting state estimates on sensor measurement data acquired in a new preset time window are used to continuously estimate the operating state of the networked system in real time.

[0076] In one optional implementation, after obtaining the current state estimation result, the result can be stored in a cached data set and used as the historical state estimate of the preset time window starting point in the next round of estimation. Simultaneously, the cached data set slides forward one step as time progresses, meaning the new preset time window becomes the one starting from... At time k+1, the oldest existing data point is removed from the cache, while newly arrived measurement data is added to the cache. After completing the above update and sliding operations, all steps, including threat type identification and classification, data preprocessing, constructing and solving the rolling time-domain optimization problem, and extracting state estimates, are re-executed for all sensor measurement data acquired in the new preset time window. Through this continuously iterative recursive mechanism, reliable state estimates for each time step are continuously output over time, thereby meeting the continuous and real-time requirements of state monitoring in practical networked systems.

[0077] Based on the same inventive concept, such as Figure 4 As shown in the figure, an embodiment of the present invention provides a networked system operating state estimation device 300, comprising: The data acquisition unit 301 is used to acquire sensor measurement data of the networked system at the current time and within a preset time window, and to acquire the historical state estimate of the starting point of the preset time window, wherein the historical state estimate is the prediction estimate result of the starting point state of the preset time window in the previous round.

[0078] Threat type identification unit 302 is used to identify and classify the threat type of the sensor measurement data based on a preset abnormal measurement model, so as to determine the threat type identification corresponding to the sensor measurement data at each time.

[0079] The data preprocessing unit 303 is used to preprocess all sensor measurement data after threat type identification and classification to obtain target measurement data.

[0080] The optimal state estimate sequence determination unit 304 is used to construct and solve a rolling time domain optimization problem based on the target measurement data and the historical state estimates, so as to obtain the optimal state estimate sequence at each time within the preset time window.

[0081] The operation state estimation result determination unit 305 is used to extract the state estimation value corresponding to the current moment from the optimal state estimation value sequence, as the operation state estimation result of the networked system at the current moment.

[0082] Regarding the aforementioned networked system operation state estimation device 300, the specific functions of each unit have been described in detail in the embodiments of the networked system operation state estimation method provided in this specification, and will not be elaborated here.

[0083] Based on the same inventive concept, embodiments of this invention provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any of the methods in the aforementioned networked system operating state estimation method.

[0084] The present invention has at least the following beneficial effects: The networked system operation status estimation method provided by this invention acquires sensor measurement data and historical state estimates from the starting point within a preset time window. Then, it identifies the threat type of the data at each moment based on an abnormal measurement model. Subsequently, it preprocesses all measurement data to obtain target measurement data, constructs and solves a rolling time-domain optimization problem to obtain the optimal state estimate sequence within the window, and finally extracts the current moment estimate as the operation status result. This enables the estimation center to effectively identify and suppress the impact of abnormal data caused by the coupling effect of various network threats such as communication delay, data packet loss, and malicious tampering. As a result, it improves the accuracy and reliability of state estimation in complex network environments and achieves real-time and accurate estimation of the networked system operation status.

[0085] In the several embodiments provided by this invention, it should be understood that the disclosed apparatus and methods can also be implemented in other ways. The apparatus embodiments described above are merely illustrative; for example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of the invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0086] In addition, the functional modules in the various embodiments of the present invention can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0087] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0088] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for estimating the operating state of a networked system, characterized in that, The method includes: The system acquires sensor measurement data of the networked system within a preset time window at the current time and before, and acquires the historical state estimate of the starting point of the preset time window, wherein the historical state estimate is the prediction estimate of the starting point state of the preset time window in the previous round. Based on a preset abnormal measurement model, the sensor measurement data is used to identify and classify threat types to determine the threat type identifier corresponding to the sensor measurement data at each time. All sensor measurement data after threat type identification and classification are preprocessed to obtain target measurement data; Based on the target measurement data and the historical state estimates, a rolling time-domain optimization problem is constructed and solved to obtain the optimal state estimate sequence for each moment within the preset time window. Extract the state estimate corresponding to the current moment from the optimal state estimate sequence, and use it as the operating state estimate result of the networked system at the current moment.

2. The networked system operating state estimation method as described in claim 1, characterized in that, The threat types include communication latency, packet loss, and data tampering. The step of identifying and classifying the threat types of the sensor measurement data based on a preset abnormal measurement model to determine the threat type identifier corresponding to the sensor measurement data at each time point includes: The preset anomaly measurement model is invoked, and the preset anomaly measurement model is as follows: Where k is the current time, The first before the network transmission anomaly Constantly receiving normal sensor measurement data. , They are respectively in The sensor measurement data actually received at time k and time k that were affected by network anomalies. In order to be in Sensor measurement data at any given time. Y represents the number of communication delay steps. k Let k be the tampering offset value injected into the sensor measurement data at time k. , as well as All are binary variables, taking values ​​of 1 or 0. , , , , Υ k , , For the set of real numbers, It is an n-dimensional column vector. Determined by timestamp, Indicates no delay. Indicates a time delay. Determined by the data packet reception status. This indicates that the data packet was successfully received. Indicates packet loss. Determined by saturation method, This indicates that data tampering has occurred. This indicates that no data tampering has occurred; when =1、 When = 0, if If the value is 0, then the communication delay is determined to occur in the sensor measurement data at the current moment. when =1、 When = 0, if If the value is 0, it is determined that the sensor measurement data at the current moment has been lost. when =1、 When =1, if If the value is 1, then it is determined that the sensor measurement data at the current moment has been tampered with, and the tampering offset value is Y. k ; when =1、 When = 0, if If the value is 1, then it is determined that the sensor measurement data at the current moment has experienced both the communication delay and the data tampering, and the tampering offset value is Y. k .

3. The networked system operating state estimation method as described in claim 2, characterized in that, The process of preprocessing all sensor measurement data after threat type identification and classification to obtain target measurement data includes: Define a cached data set to store abnormal measurement data and historical state estimates within a preset time window period prior to the current time. The cached data set is represented as follows: in, The cached data set at time k, The preset time window length, They represent The moment arrive State estimate at time 10:00 This represents the initial state prediction value at the start point of the preset time window. , ; Abnormal measurement data in the cached dataset is truncated based on the following truncation mechanism: in, To truncate the processed sensor measurement data, , As a preset upper bound for the delay, , ; The truncated sensor measurement data is saturated to obtain the target measurement data.

4. The networked system operating state estimation method as described in claim 3, characterized in that, The saturation processing of the truncated sensor measurement data includes: Calculate the predicted measurement value corresponding to the sensor measurement data at each moment within the preset time window; Calculate the residual between the measurement data of each sensor and the corresponding measurement prediction value; For sensor measurement data where the Euclidean norm of the residual exceeds a preset residual threshold, the value of the sensor measurement data is compressed toward the corresponding measurement prediction value until the Euclidean norm of the residual equals the preset residual threshold, thus completing the saturation process.

5. The networked system operating state estimation method as described in claim 1, characterized in that, The step of constructing and solving a rolling time-domain optimization problem based on the target measurement data and the historical state estimates to obtain the optimal state estimate sequence at each moment within the preset time window includes: A preset cost function is constructed as the objective function of the rolling time-domain optimization problem; Using the sequence of state estimates at each moment within the preset time window as optimization variables, and the state recursive equation and measurement output equation of the networked system as constraints, the cost function is minimized to obtain the optimal sequence of state estimates at each moment within the preset time window.

6. The networked system operating state estimation method as described in claim 5, characterized in that, The preset cost function is: in, For the cost function in the rolling time domain, The historical state estimate is... For the target measurement data, The predicted measurement value for the target. To adjust the parameters, P1 and P2 are positive definite weight matrices. For weighted norm, , , P1 P2 , Represents the set of real matrices with n rows and n columns.

7. The networked system operating state estimation method as described in claim 3, characterized in that, The method further includes: The operational state estimation result is saved as the historical state estimate at the starting point of the preset time window in the next round of estimation, and the cached data set is slid forward by one step as the preset time window progresses. The recursive formula is as follows: Where A and B are the state matrix and input matrix of the networked system, respectively. for Time-based control input; The steps of performing threat type identification and classification, data preprocessing, constructing and solving a rolling time-domain optimization problem, and extracting state estimates on sensor measurement data acquired in a new preset time window are used to continuously estimate the operating state of the networked system in real time.

8. A networked system operation state estimation device, characterized in that, The networked system operation status estimation device includes: The data acquisition unit is used to acquire sensor measurement data of the networked system at the current moment and within a preset time window, and to acquire the historical state estimate of the starting point of the preset time window, wherein the historical state estimate is the prediction estimate of the starting point state of the preset time window in the previous round. The threat type identification unit is used to identify and classify the threat type of the sensor measurement data based on a preset abnormal measurement model, so as to determine the threat type identification corresponding to the sensor measurement data at each time. The data preprocessing unit is used to preprocess all sensor measurement data after threat type identification and classification to obtain target measurement data. The optimal state estimate sequence determination unit is used to construct and solve a rolling time domain optimization problem based on the target measurement data and the historical state estimates, and obtain the optimal state estimate sequence at each moment within the preset time window. The operation state estimation result determination unit is used to extract the state estimation value corresponding to the current moment from the optimal state estimation value sequence, and use it as the operation state estimation result of the networked system at the current moment.

9. An electronic device, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a computer program, which, when executed, controls the server where the computer-readable storage medium is located to implement the steps of the method according to any one of claims 1 to 7.

Images