Intelligent compensation method for dynamic response delay of sensor

Abstract

The invention relates to the technical field of data compensation, and discloses an intelligent compensation method for dynamic response delay of a sensor, which comprises the following steps: judging whether a target sensor has dynamic response delay, collecting regional data, analyzing the distribution condition of the sensors in a regional set, judging whether the dynamic response delay is related to the environment, and if yes, judging whether the dynamic response delay is related to the environment. According to the intelligent compensation method for the dynamic response delay of the sensor, the efficiency of the whole process is improved through a batch processing and efficient data analysis method, analysis and compensation of a large number of sensors can be completed in a short time, and the intelligent compensation method for the dynamic response delay of the sensor is suitable for large-scale popularization and application. Through quantitative analysis and accurate environmental impact assessment, the accuracy and reliability of sensor measurement are improved, and the overall performance of the system is enhanced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data compensation, in particular to an intelligent compensation method for dynamic response delay of a sensor. BACKGROUND

[0002] With the rapid development of industrial automation, Internet of Things, intelligent manufacturing and other fields, sensors, as important equipment for obtaining and perceiving external environmental information, are widely used in various fields such as measurement of physical quantities such as temperature, pressure, flow, acceleration and attitude. The working principle and structural mode of the sensor are various, for example, thermal resistance, strain gauge, photoelectric sensor, etc. The sensor converts physical quantities into electrical signals, and then performs data processing and analysis. However, due to the physical properties of the sensor itself, the circuit structure and the influence of the environment, the sensor usually has a certain dynamic response delay when responding to changes in the external environment, such as step response delay and frequency response characteristics. This response delay can cause measurement errors, control system performance degradation and information processing delay in subsequent data processing, control system and decision-making.

[0003] The existing intelligent compensation method for dynamic response delay of a sensor cannot analyze and determine whether the same type of sensor in the region also has a dynamic response delay according to the dynamic response delay of a sensor, cannot determine whether the dynamic response delay is related to the environment according to the overall dynamic response delay of the same type of sensor in the region, cannot determine whether the dynamic response delay is related to the environment according to the distribution of the same type of sensor in the region, especially the distribution of the same type of sensor with dynamic response delay in the region, and cannot determine the influence of the equipment on the environment around the sensor according to the environmental conditions and the distribution of the mechanical equipment and sensors in the region, and adjust the damping coefficient of multiple sensors or the control signal of a single sensor accordingly. It is prone to cause the situation that data compensation is not timely and affects the operation of the equipment. The practicability has certain limitations. SUMMARY

[0004] The present application provides an intelligent compensation method for dynamic response delay of a sensor to promote the solution to the problems in the background art.

[0005] The present application provides the following technical solution: an intelligent compensation method for dynamic response delay of a sensor, comprising:

[0006] acquiring target sensor data;

[0007] determining whether the target sensor has a dynamic response delay;

[0008] If it is determined that the target sensor does not have a dynamic response delay, the target sensor data is continuously collected, and it is determined whether the target sensor has a dynamic response delay;

[0009] If the target sensor is determined to have a dynamic response delay, then collect data for the area.

[0010] Analyze the distribution of sensors within the region set;

[0011] If it is determined to be related to the environment, the environmental impact is analyzed, and the parameters of sensors with dynamic response delays in the area are adjusted.

[0012] If it is determined that the data is unrelated to the environment, data compensation is performed to adjust the data of the target sensor.

[0013] As an optional solution to the intelligent compensation method for the dynamic response delay of the sensor described in this invention, the method involves: collecting area data, specifically:

[0014] Obtain the unique identifier and location coordinates of the target sensor. ,type and model ;

[0015] A region is defined with the target sensor as the center and R as the radius.

[0016] For each sensor in the database Its position coordinates are , type Model number ;

[0017] Search for all sensors of the same type and model within this area;

[0018] Computational Sensors Distance to the target sensor : ;

[0019] Sensors that perfectly match the target sensor type and model are selected from the query results to form a region set. : ;

[0020] For the set of regions Each sensor in Collect its output data at high frequency And record precise timestamps;

[0021] For the set of regions Each sensor in Collect actual output data : ;

[0022] For the set of regions Each sensor in , the same dynamic signal as the target sensor is inputted into the theoretical model, and the theoretical output data is calculated : ;

[0023] For each sampling time , the time difference between the actual output data and the theoretical output data is calculated: ; wherein, is the actual sampling time, is the time corresponding to in the theoretical model;

[0024] For each sampling time , the phase difference between the actual output data and the theoretical output data is calculated: ; wherein, and are the phases of the actual and theoretical outputs, respectively;

[0025] By using a delay determination function, it is determined whether the sensor has a dynamic response delay:

[0026] ; wherein, is the time difference threshold, is the phase difference threshold; if , it is determined that the sensor has a dynamic response delay; if , it is determined that the sensor does not have a dynamic response delay.

[0027] As an optional solution of the intelligent compensation method for the dynamic response delay of the sensor, wherein: the distribution of the sensors in the region set is analyzed, specifically:

[0028] A sensor set with a dynamic response delay in the region set is obtained :

[0029] ;

[0030] A region analysis function is defined to determine whether there is an association between the sensors with a dynamic response delay in the region: ;

[0031] If , it is determined that there is no association between the sensors with a dynamic response delay in the region;

[0032] artificial measures are taken;

[0033] If , it is determined that there is an association between the sensors with dynamic response delay in the region;

[0034] If , the position coordinates of the target sensor are obtained

[0035] For each sensor in the set , the distance between it and the target sensor is calculated:

[0036] ;

[0037] The position information of all sensors in the region set is obtained, and a sensor distribution map is constructed;

[0038] The direction vector of the straight line path between the sensor and the target sensor is calculated:

[0039] ;

[0040] The parametric equation of the straight line path between the sensor and the target sensor is calculated:

[0041] ; where is a dimensionless parameter, used to represent the position ratio along the straight line path from to the target sensor;

[0042] Iterate through all other sensors, for each other sensor , calculate its distance to the straight line path: ; where is the vector from to another sensor , and represents the cross product of the vector;

[0043] Define a path analysis function to determine whether there are other sensors on the straight line path between the sensor and the target sensor: ; where is the distance threshold;

[0044] If , it is determined that there are no other sensors on the straight line path between the sensor and the target sensor;

[0045] If , it is determined that there are other sensors on the straight line path between the sensor and the target sensor;

[0046] Then calculate the analysis parameter of other sensors existing on the straight line path between sensor and target sensor: ;

[0047] Set a line segment analysis function to determine whether other sensors exist on the line segment between sensor and target sensor: ;

[0048] If , it is determined that other sensors exist on the line segment between sensor and target sensor;

[0049] If , it is determined that other sensors do not exist on the line segment between sensor and target sensor;

[0050] Define a penetration analysis function to determine whether other sensors exist between sensor and target sensor: ;

[0051] If , it is determined that other sensors exist between sensor and target sensor;

[0052] If , it is determined that other sensors do not exist between sensor and target sensor;

[0053] Calculate the concentration degree of sensors with dynamic response delay existing in the region: ;

[0054] Define an impact analysis function to determine whether the dynamic response delay is related to the environment:

[0055] ; wherein, is the concentration degree threshold, is the majority surrounding sensors with dynamic response delay;

[0056] If , it is determined that it is related to the environment;

[0057] If , it is determined that it is not related to the environment.

[0058] As an optional solution of the intelligent compensation method for the dynamic response delay of the sensor described in the present application, wherein: the environmental impact is analyzed, and the parameters of the sensors with dynamic response delay existing in the region are adjusted, specifically:

[0059] Obtain temperature and humidity data in the target area: ;

[0060] Obtain a set of sensors in the area that have dynamic response delay ;

[0061] Extract the original damping coefficient of each sensor in the sensor set , denoted as ;

[0062] Obtain the concentration of sensors in the area that have dynamic response delay and the distribution of mechanical equipment in the area ;

[0063] For each sensor , calculate the degree of influence of the mechanical equipment in the area on the temperature and humidity of the environment around each sensor that has dynamic response delay:

[0064] ;

[0065] wherein, , , is a weight coefficient, is the distance weight between the sensor and the mechanical equipment, is the density of mechanical equipment around the sensor ;

[0066] According to the degree of influence and the current environmental temperature and humidity , find the corresponding influence temperature and humidity in the influence model ;

[0067] According to , extract the corresponding damping coefficient difference value in the influence model ;

[0068] Calculate the final damping coefficient : ;

[0069] Adjust the damping coefficient of the sensor to the final damping coefficient .

[0070] As an optional solution of the intelligent compensation method for the dynamic response delay of the sensor described in the present application, wherein the generation of the influence model is specifically:

[0071] Distribute the mechanical equipment at different distances and orientations to form a device set:

[0072] ;

[0073] wherein the position coordinates of each device are ;

[0074] The sensors are distributed at different distances and orientations to form a sensor set:

[0075] ;

[0076] wherein the position coordinates of each sensor are and the initial damping coefficient is ;

[0077] The device set is combined with the sensor set in different combinations to form a simulation set: ;

[0078] wherein each element is ;

[0079] Obtain several different temperature and humidity conditions: ;

[0080] For each simulation element and each temperature and humidity condition , collect the temperature and humidity data at the location of the sensor , and define the influence of temperature and humidity: ;

[0081] For each simulation element and each temperature and humidity condition , according to its corresponding influence of temperature and humidity , calculate the influence of the mechanical device on the temperature and humidity of the environment around each sensor under the simulation element: ;

[0082] Simulate the sensors with different data collection and different damping coefficients to form an initial simulation set: ; wherein is the data collected by the sensor , and is the corresponding damping coefficient;

[0083] For each element in the initial simulation set, under each influence of temperature and humidity , collect the data of the sensor again, denoted as ;

[0084] Each re-collected sensor Data Each element in the initial simulation set corresponds one-to-one with another to form a data change set: ;

[0085] For each element in the data change set Adjust the damping coefficient , making With the initial simulation set Consistent;

[0086] Calculate the difference in damping coefficients: .

[0087] As an optional solution to the intelligent compensation method for sensor dynamic response delay described in this invention, the following is included: performing data compensation and adjusting the data of the target sensor, specifically:

[0088] Let the dynamic model of the target sensor be:

[0089] ;

[0090] ;

[0091] in, This represents the state of the system at discrete time step k. This represents the actual state vector of the system at discrete time step k-1. Control input, For process noise, It is the sensor's observation of the system state at discrete time step k. To measure noise, For the system matrix, For the control matrix, For measurement matrix;

[0092] Set the initial state of the Kalman filter :

[0093] ;

[0094] in, This indicates the calculation of the mean or expected value of a random variable. Represents the initial state vector Expected value;

[0095] Set the initial error covariance matrix of the Kalman filter. :

[0096] ;

[0097] in, This represents the error covariance matrix of the initial state estimate;

[0098] Predict the current state of the target sensor:

[0099] ;

[0100] Predict the error covariance of the target sensor in its current state:

[0101] ;

[0102] in, The process noise covariance matrix;

[0103] Calculate the Kalman gain:

[0104] ;

[0105] in, To measure the noise covariance matrix;

[0106] The predicted state is corrected using actual measurements and Kalman gain, and the state estimate is updated.

[0107] ;

[0108] Update error covariance:

[0109] ;

[0110] Based on the updated state estimate Adjust control signals in advance .

[0111] As an optional solution to the intelligent compensation method for sensor dynamic response delay described in this invention, wherein: performing data compensation and adjusting the data of the target sensor can also be:

[0112] Collect historical data from the sensors, including input signals, output signals, and environmental factors, and normalize the data.

[0113] Choose and determine the neural network structure;

[0114] Let the input of the neural network be... The output is ;

[0115] Perform computations from the input layer to the hidden layer in a neural network:

[0116] ;

[0117] in, a weight matrix for the input layer to the hidden layer, a bias vector for the hidden layer, an activation function;

[0118] performing a calculation of the neural network from the hidden layer to the output layer:

[0119] ;

[0120] wherein, a weight matrix for the hidden layer to the output layer, a bias vector for the output layer;

[0121] calculating a loss function:

[0122] ;

[0123] wherein, an actual output, a predicted output, a number of samples;

[0124] updating the weights and biases of the adjustment network:

[0125] ;

[0126] ;

[0127] ;

[0128] ;

[0129] wherein, a learning rate;

[0130] adjusting the control signal in advance according to the predicted output of the neural network .

[0131] As an optional solution of the intelligent compensation method for the dynamic response delay of the sensor described in the application, wherein: judging whether the target sensor has a dynamic response delay, specifically:

[0132] According to the working principle and mathematical model of the target sensor, a theoretical model is established.

[0133] For each sampling time , the actual output data is collected:

[0134] ;

[0135] For each sampling time , the same dynamic signal as the target sensor is input to the theoretical model The theoretical output data is calculated as follows:

[0136] ;

[0137] Calculate the actual output data Compared with theoretical output data Time difference between them:

[0138] ;

[0139] in, , which is the actual sampling time. For the corresponding theoretical model Time;

[0140] Calculate the phase difference between the actual output data and the theoretical output data:

[0141] ;

[0142] in, and These represent the phases of the actual and theoretical outputs, respectively.

[0143] Define a delay determination function to determine whether the target sensor has a dynamic response delay:

[0144] ;

[0145] in, The time difference threshold, The phase difference threshold;

[0146] like If so, it is determined that the target sensor has a dynamic response delay;

[0147] like If so, it is determined that the target sensor does not have dynamic response delay.

[0148] As an optional solution to the intelligent compensation method for sensor dynamic response delay described in this invention, the acquisition of target sensor data specifically includes:

[0149] Initialize the data acquisition system and set the sampling frequency to [value]. ;

[0150] Calculate the sampling period :

[0151] ;

[0152] Start the data acquisition system and begin collecting output data from the target sensor. ;

[0153] every other The output data of the sensor is recorded once every time interval:

[0154] ;

[0155] wherein, , represents the sampling number;

[0156] For each sampling data The accurate time stamp is recorded:

[0157] ;

[0158] The collected data And the time stamp Is stored in the data buffer :

[0159] .

[0160] The present application has the following beneficial effects:

[0161] 1. The sensor dynamic response delay intelligent compensation method, by analyzing the data of the target sensor, judges whether the target sensor has dynamic response delay, calculates the time difference and phase difference and other parameters, realizes the quantitative analysis of the dynamic response delay, improves the accuracy and objectivity of the judgment, can timely find the dynamic response delay problem of the sensor, and provides early warning for taking further measures.

[0162] 2. The sensor dynamic response delay intelligent compensation method, according to the dynamic response delay of a sensor, analyzes and judges whether the same type of sensor in the region also has dynamic response delay, obtains all the same type and model sensors in the region set, expands the analysis range, helps to find the common problems that may exist in the region, provides data support for subsequent group analysis, and helps to judge the locality or universality of the problem.

[0163] 3. The intelligent compensation method for sensor dynamic response delay analyzes the distribution of the same sensors in the area, especially the concentration of the same sensors with dynamic response delay in the area, judges the overall dynamic response delay of the same sensors in the area, judges whether the dynamic response delay is related to the environment according to the overall dynamic response delay of the same sensors in the area, realizes batch judgment of multiple sensors in the area, improves the analysis efficiency, helps to quickly locate the sensors with dynamic response delay in the area, provides a basis for subsequent concentrated processing, analyzes the distribution of the sensors, can effectively judge whether the dynamic response delay is related to environmental factors, provides a direction for subsequent environmental impact analysis, calculating the concentration helps to quantitatively evaluate the distribution density of the dynamic response delay sensors, and provides a quantitative basis for subsequent environmental impact analysis.

[0164] 4. The intelligent compensation method for sensor dynamic response delay, when it is determined that the dynamic response delay is related to environmental factors, simulates different positions of the mechanical equipment and the sensor, simulates different damping coefficients of the sensor, simulates different environmental temperature and humidity, generates an impact set, determines the influence degree of the mechanical equipment on the temperature and humidity of the environment around each sensor under different conditions, and determines the difference value of the corresponding adjusted damping coefficient, determines the influence degree of the mechanical equipment on the temperature and humidity of the environment around each sensor under the condition, and determines the difference value of the corresponding damping coefficient, so as to determine the damping coefficient that needs to be adjusted, calculates the influence degree of the mechanical equipment on the temperature and humidity of the environment, and extracts the difference value of the damping coefficient, which can accurately adjust the damping coefficient of the sensor, improve the adaptability and accuracy of the sensor under different environmental conditions, realize effective compensation of the influence of environmental factors, and improve the measurement precision and reliability of the sensor in complex environment.

[0165] 5. The intelligent compensation method for sensor dynamic response delay, when it is determined that the dynamic response delay is not related to environmental factors, through a dynamic compensation algorithm, the response of the sensor is predicted in advance and the control signal is adjusted, the dynamic response delay is reduced, the real-time and response speed of the system are improved, the performance of the sensor and the entire measurement control system is improved, and the stability and reliability of the system in dynamic environment are ensured. BRIEF DESCRIPTION OF DRAWINGS

[0166] Figure 1 The flowchart of the intelligent compensation method for sensor dynamic response delay of the application. DETAILED DESCRIPTION

[0167] With reference to the accompanying drawings, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all the other embodiments obtained by those skilled in the art without creative work fall within the scope of the present application.

[0168] Embodiment one, a method for intelligent compensation of dynamic response delay of a sensor, referring to Figure 1 , comprising:

[0169] acquiring target sensor data;

[0170] determining whether the target sensor has dynamic response delay;

[0171] if it is determined that the target sensor has no dynamic response delay, continuously collecting target sensor data, and determining whether the target sensor has dynamic response delay;

[0172] if it is determined that the target sensor has dynamic response delay, collecting regional data;

[0173] analyzing the distribution of sensors in the region set;

[0174] if it is determined that it is related to the environment, analyzing the environmental influence, and adjusting parameters of the sensors in the region, including the target sensor, which have dynamic response delay;

[0175] if it is determined that it is not related to the environment, performing data compensation, and adjusting data of the target sensor.

[0176] The embodiment further provides that the target sensor data is acquired, specifically:

[0177] initializing a data collection system, setting a sampling frequency as ;

[0178] calculating a sampling period :

[0179] ;

[0180] starting the data collection system, and starting to collect output data of the target sensor; ;

[0181] recording output data of the sensor once every time interval:

[0182] ;

[0183] wherein, , represents a sampling serial number;

[0184] For each sampling data Record the accurate time stamp:

[0185] ;

[0186] Store the collected data and time stamp into the data buffer:

[0187] .

[0188] The embodiment also provides that whether the target sensor has dynamic response delay is judged, specifically:

[0189] According to the working principle and mathematical model of the target sensor, a theoretical model is established

[0190] For a first-order sensor, its transfer function can be expressed as ;

[0191] Wherein, is the gain, is the time constant;

[0192] For a second-order sensor, its transfer function can be expressed as ;

[0193] Wherein, is the natural frequency, is the damping ratio;

[0194] For each sampling time , the actual output data is collected:

[0195] ;

[0196] For each sampling time , the same dynamic signal as the target sensor is input into the theoretical model , and the theoretical output data is calculated:

[0197] ;

[0198] The time difference between the actual output data and the theoretical output data is calculated:

[0199] ;

[0200] Wherein, is the actual sampling time, is the time corresponding to in the theoretical model;​

[0201] Calculate the phase difference between the actual output data and the theoretical output data:

[0202] ;

[0203] Where, and are the phases of the actual and theoretical outputs, respectively;

[0204] Define a delay judgment function to determine whether the target sensor has a dynamic response delay:

[0205] ;

[0206] Where, is the time difference threshold, is the phase difference threshold, used to determine whether the target sensor has a dynamic response delay;

[0207] If , it is determined that the target sensor has a dynamic response delay;

[0208] If , it is determined that the target sensor does not have a dynamic response delay.

[0209] Through the above method, through batch processing and efficient data analysis method, the efficiency of the whole process is improved, a large number of sensors can be analyzed and compensated in a short time, through quantitative analysis and accurate environmental impact evaluation, the accuracy and reliability of sensor measurement are improved, the overall performance of the system is enhanced, the problem of sensor dynamic response delay can be solved systematically, from data acquisition, delay judgment, environmental impact analysis to dynamic compensation, a complete solution is formed, which has strong universality and adaptability.

[0210] Example two, this embodiment is improved on the basis of example one, the sensor dynamic response delay intelligent compensation method, the acquisition area data, specifically:

[0211] Get the unique identifier (ID) of the target sensor, position coordinates , type and model ;

[0212] Take the target sensor as the center, R as the radius, and form the area range;

[0213] For each sensor in the database , the position coordinates are , the type is , and the model is ;

[0214] Retrieve all sensors of the same type and model within the area through sensor network or database query interface;

[0215] Computational Sensors Distance to the target sensor :

[0216] ;

[0217] Sensors that perfectly match the target sensor type and model are selected from the query results to form a region set. :

[0218] ;

[0219] For the set of regions Each sensor in Collect its output data at high frequency And record precise timestamps;

[0220] For the set of regions Each sensor in Collect actual output data :

[0221] ;

[0222] For the set of regions Each sensor in By inputting the same dynamic signal as the target sensor into the theoretical model, the theoretical output data is calculated. :

[0223] ;

[0224] For each sampling time Calculate the actual output data Compared with theoretical output data Time difference between them:

[0225] ;

[0226] in, , which is the actual sampling time. For the corresponding theoretical model Time;

[0227] For each sampling time Calculate the actual output data Compared with theoretical output data Phase difference between them:

[0228] ;

[0229] wherein, and are the phase of the actual and theoretical output respectively;

[0230] By means of a delay decision function, it is determined whether the sensor has a dynamic response delay:

[0231] ;

[0232] wherein, is a time difference threshold value, is a phase difference threshold value for determining whether the sensor has a dynamic response delay;

[0233] If , it is determined that the sensor has a dynamic response delay;

[0234] If , it is determined that the sensor does not have a dynamic response delay.

[0235] The present embodiment further provides that the distribution of the sensors within the region set is analyzed, in particular:

[0236] A sensor set with a dynamic response delay within the region set is obtained :

[0237] ;

[0238] A region analysis function is defined to determine whether there is an association between the sensors with a dynamic response delay within the region. If there are many sensors with a dynamic response delay within the region set, it is indicated that there can be a certain association between the sensors with a dynamic response delay within the region, and then the target sensor and the sensors within the region are analyzed and timely adjusted in parameters so that the data collected by them is restored to the data as it should be. If there are few sensors with a dynamic response delay within the region set, it is indicated that there is a small possibility of an association between the sensors with a dynamic response delay within the region, and then the sensors are manually checked, repaired or replaced:

[0239] ;

[0240] If , it is determined that there is no association between the sensors with a dynamic response delay within the region;

[0241] Manual measures are taken, i.e. the sensors are manually checked, repaired or replaced;

[0242] If if the distance is less than a threshold value, then determine that there is an association between the sensor with dynamic response delay and the target sensor within the region;

[0243] then obtain the position coordinates of the target sensor ;

[0244] For each sensor in the set , calculate the distance between it and the target sensor:

[0245] ;

[0246] Obtain the position information of all sensors in the set of regions, including sensors with and without dynamic response delay, and construct a sensor distribution map;

[0247] Calculate the direction vector of the straight line path between the sensor and the target sensor:

[0248] ;

[0249] Calculate the parametric equation of the straight line path between the sensor and the target sensor:

[0250] ;

[0251] where, is a dimensionless parameter representing the proportion of the position along the straight line path from to the target sensor, when t = 0, the position point is located at , when t = 1, the position point is located at the target sensor, and when the parameter value is in the range [0, 1], it represents the position between the straight line path and the target sensor;

[0252] Iterate through all other sensors, for each other sensor , calculate its distance to the straight line path:

[0253] ;

[0254] where, is the vector from to another sensor , used to calculate the perpendicular distance from to the straight line path, which helps to determine whether is near the straight line path, represents the cross product of the vector;

[0255] Define a path analysis function to determine whether the sensor ​whether there is another sensor on the straight line path between the sensor and the target sensor:

[0256] ;

[0257] wherein, is a distance threshold value for determining whether there is another sensor on the straight line path between the sensor and the target sensor;

[0258] If , it is determined that there is no other sensor on the straight line path between the sensor and the target sensor;

[0259] If , it is determined that there is another sensor on the straight line path between the sensor and the target sensor;

[0260] The analysis parameter of the other sensor existing on the straight line path between the sensor and the target sensor is calculated, the analysis parameter being a scalar value obtained by a dot product operation, for determining whether the sensor is located on the line segment of the straight line path to the target sensor, indicating the proportion of the projection position on the straight line path relative to If the proportion is within the range of [0, 1], it means that the projection of the sensor is on the line segment; otherwise, the projection is on the extension line of the line segment:

[0261] ;

[0262] A line segment analysis function is set to determine whether there is another sensor on the line segment between the sensor and the target sensor:

[0263] ;

[0264] If , it is determined that there is another sensor on the line segment between the sensor and the target sensor;

[0265] If , it is determined that there is no other sensor on the line segment between the sensor and the target sensor;

[0266] A penetration analysis function is defined to determine whether there is another sensor between the sensor and the target sensor:

[0267] ​​​​​​​​​​​​​ ;

[0268] If , it is determined that there are other sensors between the sensor and the target sensor;

[0269] If , it is determined that there are no other sensors between the sensor and the target sensor;

[0270] Calculate the concentration degree of sensors with dynamic response delay in the region:

[0271] ;

[0272] Define an impact analysis function to determine whether the dynamic response delay is related to the environment:

[0273] ;

[0274] Wherein, is the concentration degree threshold value for determining whether the dynamic response delay is related to the environment, is the majority surrounding the sensor with dynamic response delay, if the sensor has other sensors with dynamic response delay around it, and the number of such sensors is large, then output , otherwise output to determine whether the sensors with dynamic response delay in the region are concentrated, thereby determining whether the dynamic response delay is related to the environment. If the sensors with dynamic response delay in the region are concentrated, it is very rare for sensor equipment to be damaged at the same time or for data collection to have problems, and these sensors with dynamic response delay are distributed in a concentrated manner. Therefore, it is speculated that the possibility of environmental impact is greater, and it is determined that the dynamic response delay is related to the environment;

[0275] If , it is determined that it is related to the environment;

[0276] If , it is determined that it is not related to the environment.

[0277] The embodiment also provides that the environmental impact is analyzed, and the parameters of the sensors with dynamic response delay in the region are adjusted, specifically:

[0278] Obtain temperature and humidity data in the target region through environmental monitoring equipment:

[0279] ;

[0280] Obtaining a set of sensors with dynamic response delay in the area ;

[0281] Extracting the original damping coefficient of each sensor in the set of sensors , denoted as ;

[0282] Obtaining the concentration of sensors with dynamic response delay in the area and the distribution of mechanical equipment in the area ;

[0283] For each sensor , calculate the degree of influence of the mechanical equipment in the area on the temperature and humidity of the environment around each sensor with dynamic response delay:

[0284] ;

[0285] wherein , , is a weight coefficient satisfying , used to adjust the relative contribution of each factor to the degree of influence, is the distance weight between the sensor and the mechanical equipment, representing the change of the influence of the mechanical equipment on the sensor with distance, specifically:

[0286] ;

[0287] wherein is the distance from the th mechanical equipment to the sensor , is a very small positive number used to avoid the case where the denominator is zero;

[0288] is the density of mechanical equipment around the sensor , representing the number of mechanical equipment per unit volume or unit area, specifically:

[0289] ;

[0290] wherein is the total number of mechanical equipment in the area, is the area of the region around the sensor ;

[0291] According to the degree of influence and the current environmental temperature and humidity , find the corresponding influence temperature and humidity in the influence model;

[0292] According to , the corresponding damping coefficient difference value is extracted in the influence model ;

[0293] The final damping coefficient is calculated :

[0294] ;

[0295] The damping coefficient adjustment of the sensor is set to the final damping coefficient .

[0296] Wherein, the generation of the influence model is specific to:

[0297] The mechanical equipment is distributed at different distances and orientations to form a device set:

[0298] ;

[0299] Wherein, the position coordinates of each device are ;

[0300] The sensors are distributed at different distances and orientations to form a sensor set:

[0301] ;

[0302] Wherein, the position coordinates of each sensor are , and the initial damping coefficient is ;

[0303] The device set and the sensor set are combined in different ways to form a simulation set:

[0304] ;

[0305] Wherein, each element is ;

[0306] Obtain several different temperature and humidity conditions:

[0307] ;

[0308] For each simulation element and each temperature and humidity condition , collect the temperature and humidity data at the location of the sensor , and define the influence temperature and humidity:

[0309] ;

[0310] for each simulation element and each temperature and humidity condition , according to its corresponding influence temperature and humidity , calculate the degree of influence of the mechanical equipment on the temperature and humidity of the environment around each sensor under the simulation element:

[0311] ;

[0312] Simulate the sensor with different data collection and different damping coefficients to form an initial simulation set:

[0313] ;

[0314] wherein, is the data collected by the sensor , and is the corresponding damping coefficient;

[0315] For each element in the initial simulation set , under each influence temperature and humidity , collect the data of the sensor again, denoted as ;

[0316] Each of the re-collected data of the sensor corresponds to each element in the initial simulation set to form a data change set:

[0317] ;

[0318] For each element in the data change set , adjust the damping coefficient so that is consistent with in the initial simulation set;

[0319] Calculate the damping coefficient difference value:

[0320] .

[0321] The embodiment also provides that the data compensation is performed, the target sensor is adjusted through a dynamic compensation algorithm based on Kalman filtering, the Kalman filtering estimates the response of the target sensor in real time through prediction and update steps, so as to compensate the dynamic response delay, and specifically:

[0322] Suppose the dynamic model of the target sensor is:

[0323] ;

[0324] ​ ;

[0325] where, xk represents the state of the system at discrete time step k, containing all necessary information to describe the behavior of the system, such as position, velocity, acceleration, etc., xk-1 represents the actual state vector of the system at discrete time step k-1, containing the true state information of the system at that time step, such as position, velocity, temperature, etc., uk represents the control input, indicating the known external control signal or input applied to the system during the discrete time step k-1 to k, wk is the process noise, representing the random noise or uncertainty affecting the system during state transition, whose characteristics are usually assumed to be zero-mean Gaussian white noise, the process noise will cause the uncertainty of the system state to increase, for example, due to unmodeled dynamic effects, environmental disturbances, etc. yk represents the observation of the system state by the sensor at discrete time step k, vk is the measurement noise, representing the random noise in the sensor measurement process, which is also usually assumed to be zero-mean Gaussian white noise, the measurement noise will cause the deviation between the sensor output and the true state, for example, due to the limitation of sensor accuracy, environmental disturbances, etc. Ak is the system matrix, used to describe the transition relationship of the system state, i.e. how to get the predicted value of the current state from the previous state xk-1, the system matrix is a square matrix, whose dimension is determined by the dimension of the state vector, Bk is the control matrix, describing how the control input affects the state transition, i.e. how to convert the control input into the contribution to the state, whose dimension is determined by the dimension of the state vector and the dimension of the control input vector, Ck is the measurement matrix, used to convert the state vector into the measurement space, i.e. to describe how to get the sensor measurement value from the system state, whose dimension is determined by the dimension of the measurement value vector and the dimension of the state vector;

[0326] Set the initial state of the Kalman filter :

[0327] ;

[0328] where, E[.] represents the calculation of the mean or expected value of a random variable, in probability theory and statistics, the expected value reflects the average behavior of the random variable, E[x0] represents the expected value of the initial state vector , i.e. the best estimate of the initial state of the system;

[0329] Set the initial error covariance matrix of the Kalman filter. :

[0330] ;

[0331] in, The error covariance matrix represents the initial state estimate. It measures the uncertainty of the initial state estimate. Each element in the matrix represents the covariance between different state variables, reflecting the correlation between the estimated value and the true value.

[0332] Predict the current state of the target sensor:

[0333] ;

[0334] in, This indicates that at time step k-1, based on all measurement data up to time k-1, i.e., y1, y2, ..., y... k－1， Regarding system status The optimal estimate, calculated during the Kalman filter's update step, combines information from the predicted state and actual measurements. It represents the filter's best guess of the system state at time k-1. For example, if we are tracking the position and velocity of a moving object, then... It is the most accurate estimate of the position and velocity of the object at time k-1 by combining previous measurement data and system dynamic model;

[0335] Predict the error covariance of the target sensor in its current state:

[0336] ;

[0337] in, This represents the error covariance matrix of the system state estimate at time step k-1, based on all measurement data up to time k-1. It measures the error of the estimate. The uncertainty reflects the covariance between the estimated value and the true state. Each element in the matrix represents the covariance between different state variables. The larger the covariance, the higher the uncertainty of the estimated value. This matrix is ​​also calculated and updated in the update step of the Kalman filter to represent the accuracy of the estimate, for example, in a scenario of tracking a moving object. It can tell us the accuracy of the estimated position and velocity, as well as the correlation between the position and velocity estimates;

[0338] in, The process noise covariance matrix;

[0339] Calculate the Kalman gain:

[0340] ;

[0341] wherein, is the measurement noise covariance matrix;

[0342] The predicted state is corrected using the actual measurement and the Kalman gain to update the state estimate:

[0343] ;

[0344] The error covariance is updated:

[0345] ;

[0346] According to the updated state estimate , the control signal is adjusted in advance.

[0347] The embodiment also provides that data compensation is performed, and the target sensor is adjusted in data by a dynamic compensation algorithm based on a neural network. The neural network learns historical data and reference input of the sensor, predicts the response of the target sensor, and adjusts the control signal according to the response. In addition, the neural network can also be used for:

[0348] Collecting historical data of the sensor, including input signals, output signals, and environmental factors, and normalizing the data for neural network training;

[0349] Selecting a neural network structure, and the neural network structure is a multi-layer perception MLP;

[0350] Supposing that the input of the neural network is , and the output is ;

[0351] Calculating the neural network from the input layer to the hidden layer:

[0352] ;

[0353] wherein, is a weight matrix from the input layer to the hidden layer, is a bias vector of the hidden layer, is an activation function, such as a Sigmoid function, a ReLU function, or the like;

[0354] Calculating the neural network from the hidden layer to the output layer:

[0355] ;

[0356] wherein, is a weight matrix from the hidden layer to the output layer, is a bias vector of the output layer;

[0357] The loss function is calculated as:

[0358]

[0359] wherein, is the actual output, is the predicted output, is the number of samples;

[0360] The weights and biases of the adjustment network are updated to make the network output as close as possible to the actual sensor output:

[0361]

[0362]

[0363]

[0364]

[0365] wherein, is the learning rate;

[0366] According to the predicted output of the neural network , the control signal is adjusted in advance.

[0367] In this embodiment, the efficiency of the entire process is improved through batch processing and efficient data analysis methods, and a large number of sensors can be analyzed and compensated in a short time. Through quantitative analysis and accurate environmental impact assessment, the accuracy and reliability of sensor measurement are improved, and the overall performance of the system is enhanced. The system can systematically solve the problem of sensor dynamic response delay, from data acquisition, delay judgment, environmental impact analysis to dynamic compensation, forming a complete solution with strong universality and adaptability.​​​​​

Claims

1. A smart compensation method for sensor dynamic response delay, characterized in that: include: Acquire target sensor data; Determine if the target sensor has a dynamic response delay; If it is determined that the target sensor does not have dynamic response delay, then the target sensor data will be continuously collected, and the target sensor will be judged to have dynamic response delay. If the target sensor is determined to have a dynamic response delay, then collect data for the area. Analyze the distribution of sensors within the region set; If it is determined to be related to the environment, the environmental impact is analyzed, and the parameters of sensors with dynamic response delays in the area are adjusted. If it is determined that the data is unrelated to the environment, data compensation is performed to adjust the data of the target sensor.

2. The intelligent compensation method for sensor dynamic response delay according to claim 1, characterized in that: Data to be collected from the area, specifically: Obtain the unique identifier and location coordinates of the target sensor. ,type and model ; A region is defined with the target sensor as the center and R as the radius. For each sensor in the database Its position coordinates are , type Model number ; Search for all sensors of the same type and model within this area; Computational Sensors Distance to the target sensor : ； Sensors that perfectly match the target sensor type and model are selected from the query results to form a region set. : ； For the set of regions Each sensor in Collect its output data at high frequency And record precise timestamps; For the set of regions Each sensor in Collect actual output data : ； For the set of regions Each sensor in By inputting the same dynamic signal as the target sensor into the theoretical model, the theoretical output data is calculated. : ； For each sampling time Calculate the actual output data Compared with theoretical output data Time difference between them: ； in, , which is the actual sampling time. For the corresponding theoretical model Time; For each sampling time Calculate the actual output data Compared with theoretical output data Phase difference between them: ； in, and These represent the phases of the actual and theoretical outputs, respectively. Determine the sensor using the delay decision function. Is there a dynamic response delay? ； in, The time difference threshold, The phase difference threshold; like Then determine the sensor There is a dynamic response delay; like Then determine the sensor There is no dynamic response delay.

3. The intelligent compensation method for sensor dynamic response delay according to claim 1, characterized in that: The distribution of sensors within the analysis region set is as follows: Get the set of sensors with dynamic response delay within the region set. : ； Define a region analysis function to determine whether there is a correlation between sensors with dynamic response delays within a region: ； like If the sensor has a dynamic response delay, then there is no correlation between the sensors in the region. Then artificial measures will be taken; like If the sensors in the region exhibit dynamic response delays, then a correlation is determined between them. Then obtain the position coordinates of the target sensor. ; For sets Each sensor in Calculate the distance between it and the target sensor: ； Obtain the location information of all sensors within the region set and construct a sensor distribution map; Computational Sensors The direction vector of the straight path between the target sensor and the target sensor: ； Computational Sensors The parametric equation of the straight path between the target sensor and the target sensor: ； in, is a dimensionless parameter used to represent the path along a straight line from The ratio of the position to the target sensor; Iterate through all other sensors, for each other sensor Calculate its distance to the straight path: ； in, , indicating from To another sensor The vector, Represents the cross product of vectors; Define a path analysis function to determine the sensor Are there other sensors on the straight path between the target sensor and the target sensor? ； in, Distance threshold; like Then determine the sensor There are no other sensors on the straight path between the target sensor and the target sensor; like Then determine the sensor There are other sensors on the straight path between the target sensor and the target sensor; Then calculate the sensor Analysis parameters of other sensors present along the straight path between the target sensor and the target sensor: ； Set up a line segment analysis function to determine the sensor Are there other sensors on the line segment between the target sensor and the target sensor? ； like Then determine the sensor There are other sensors on the line segment between the target sensor and the target sensor; like Then determine the sensor There are no other sensors on the line segment between the target sensor and the target sensor; Define an interleaving analysis function to determine the sensor Are there any other sensors between the target sensor and the target sensor? ； like Then determine the sensor There are other sensors between the target sensor and the target sensor; like Then determine the sensor There are no other sensors between the target sensor and the target sensor; The concentration of sensors with dynamic response delays within the computational region: ； Define an impact analysis function to determine whether dynamic response delay is environment-dependent: ； in, The concentration threshold, The majority A sign indicating the presence of sensors with dynamic response delays in the vicinity; like If so, it is determined to be related to the environment; like If so, it is determined to be unrelated to the environment.

4. The intelligent compensation method for sensor dynamic response delay according to claim 1, characterized in that: Analyze the environmental impact and adjust the parameters of sensors with dynamic response delays within the area, specifically: Obtain temperature and humidity data for the target area: ； Acquire the set of sensors in the region that have dynamic response delay. ; Extract each sensor from the sensor set The original damping coefficient is denoted as ; The concentration of sensors with dynamic response delays within the acquisition area. Distribution of mechanical equipment within the region ; For each sensor The calculation determines the impact of mechanical equipment within the calculation area on the temperature and humidity of the surrounding environment of each sensor with dynamic response delay. ； in, , , These are weighting coefficients. It is a sensor Distance weight between the equipment and the machinery It is a sensor The density of surrounding machinery and equipment; Based on the degree of impact and current ambient temperature and humidity Find the corresponding temperature and humidity influences in the influence model. ; according to Extract the corresponding damping coefficient difference from the influence model. ; Calculate the final damping coefficient : ； sensor The damping coefficient adjustment is set to the final damping coefficient. .

5. The intelligent compensation method for sensor dynamic response delay according to claim 4, characterized in that: The factors affecting model generation are as follows: The mechanical equipment is distributed at different distances and in different directions to form an equipment collection: ； Each device The position coordinates are ; Sensors are distributed at different distances and orientations to form a sensor ensemble: ； Each sensor The position coordinates are The initial damping coefficient is ; Assembly of devices With sensor set By combining different elements, a simulated set can be formed: ； Each element is ; Obtain several different temperature and humidity conditions: ； For each simulation element and each temperature and humidity condition Acquisition sensor The temperature and humidity data of the location are defined as factors affecting temperature and humidity. ； For each simulation element and each temperature and humidity condition According to its corresponding influence on temperature and humidity Calculate the degree of influence of the mechanical equipment on the temperature and humidity of the environment surrounding each sensor under this simulated element: ； The sensor is simulated with different acquired data and different damping coefficients to form an initial simulation set: ； in, For sensors The collected data This is the corresponding damping coefficient; For each element in the initial simulation set In each factor affecting temperature and humidity Next, the sensor will collect data again. The data is denoted as ; Each sensor that collects data again Data Each element in the initial simulation set corresponds one-to-one with another to form a data change set: ； For each element in the data change set Adjust the damping coefficient , making With the initial simulation set Consistent; Calculate the difference in damping coefficients: 。 6. The intelligent compensation method for sensor dynamic response delay according to claim 1, characterized in that: Perform data compensation to adjust the data of the target sensor, specifically as follows: Let the dynamic model of the target sensor be: ； ； in, This represents the state of the system at discrete time step k. This represents the actual state vector of the system at discrete time step k-1. Control input, For process noise, It is the sensor's observation of the system state at discrete time step k. To measure noise, For the system matrix, For the control matrix, For measurement matrix; Set the initial state of the Kalman filter : ； in, This indicates the calculation of the mean or expected value of a random variable. Represents the initial state vector Expected value; Set the initial error covariance matrix of the Kalman filter. : ； in, This represents the error covariance matrix of the initial state estimate; Predict the current state of the target sensor: ； Predict the error covariance of the target sensor in its current state: ； in, The process noise covariance matrix; Calculate the Kalman gain: ； in, To measure the noise covariance matrix; The predicted state is corrected using actual measurements and Kalman gain, and the state estimate is updated. ； Update error covariance: ； Based on the updated state estimate Adjust control signals in advance .

7. The intelligent compensation method for sensor dynamic response delay according to claim 1, characterized in that: Performing data compensation, adjusting the data of the target sensor, can also be done as follows: Collect historical data from the sensors, including input signals, output signals, and environmental factors, and normalize the data. Choose and determine the neural network structure; Let the input of the neural network be... The output is ; Perform computations from the input layer to the hidden layer in a neural network: ； in, This is the weight matrix from the input layer to the hidden layer. The bias vector of the hidden layer. For activation functions; Perform calculations from hidden layers to the output layer of the neural network: ； in, This is the weight matrix from the hidden layer to the output layer. This is the bias vector for the output layer; Calculate the loss function: ； in, For the actual output, To predict the output, The number of samples; Update and adjust the network's weights and biases: ； ； ； ； in, The learning rate; Based on the predicted output of the neural network Adjust the control signals in advance.

8. The intelligent compensation method for sensor dynamic response delay according to claim 1, characterized in that: To determine whether the target sensor has a dynamic response delay, the specific steps are as follows: Based on the working principle and mathematical model of the target sensor, its theoretical model is established. For each sampling time Collect actual output data: ； For each sampling time Input the same dynamic signal as the target sensor into the theoretical model. The theoretical output data is calculated as follows: ； Calculate the actual output data Compared with theoretical output data Time difference between them: ； in, , which is the actual sampling time. For the corresponding theoretical model Time; Calculate the phase difference between the actual output data and the theoretical output data: ； in, and These represent the phases of the actual and theoretical outputs, respectively. Define a delay determination function to determine whether the target sensor has a dynamic response delay: ； in, The time difference threshold, The phase difference threshold; like If so, it is determined that the target sensor has a dynamic response delay; like If so, it is determined that the target sensor does not have dynamic response delay.

9. The intelligent compensation method for sensor dynamic response delay according to claim 1, characterized in that: Acquire target sensor data, specifically: Initialize the data acquisition system and set the sampling frequency to [value]. ; Calculate the sampling period : ； Start the data acquisition system and begin collecting output data from the target sensor. ; Every Record the sensor's output data once at time intervals: ； in, , indicating the sampling sequence number; For each sampled data Record precise timestamps: ； Collected data and timestamp Stored in data buffer middle: 。

Images