A data recovery method and apparatus for data sensing

Abstract

The present invention proposes a data recovery method and device for data perception, which monitors the data transmission dynamics of each perception node in a perception layer, identifies faulty nodes in the perception nodes according to the data transmission dynamics, and determines whether there are associated nodes in the perception layer that are adjacent to the faulty nodes and meet preset conditions. The preset conditions include the distance of the associated nodes and the type of perception data. When there are associated nodes in the perception layer that are adjacent to the faulty nodes and meet the preset conditions, the perception data transmitted in real time by the associated nodes are obtained, and the perception data of the faulty nodes are restored based on the perception data transmitted in real time by the associated nodes, which can effectively restore the perception data of unavailable nodes in the perception layer.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a data recovery method and device applied to data perception. Background Art

[0002] With the development of Internet of Things technology, data perception technology based on various types of sensors has extended information and intelligence to a wider range of aspects of people's lives. The data perception network composed of a large number of sensors as perception nodes is the basis for the realization of various Internet of Things business scenarios. However, the variety of business scenarios means that the deployment environment of sensors is also complex and diverse. Therefore, it is inevitable that some perception nodes will be unavailable due to damage to the perception nodes, network interruption or power supply anomalies, resulting in the loss of some perception data in the perception network. Summary of the invention

[0003] Based on the above problems, the present invention proposes a data recovery method and device applied to data perception, which can effectively recover the perception data of unavailable nodes in the perception layer.

[0004] In view of this, a first aspect of the present invention proposes a data recovery method applied to data perception, comprising:

[0005] Monitor the data transmission dynamics of each sensing node in the perception layer;

[0006] dynamically identifying a faulty node among the sensing nodes based on the data transmission;

[0007] Determine whether there is an associated node in the perception layer that is adjacent to the faulty node and meets preset conditions, wherein the preset conditions include a distance between the associated nodes and a type of perception data;

[0008] When there is an associated node in the perception layer that is adjacent to the faulty node and meets a preset condition, acquiring perception data transmitted in real time by the associated node;

[0009] The perception data of the faulty node is restored based on the perception data transmitted in real time by the associated node.

[0010] Further, after the step of dynamically identifying a faulty node in the sensing node according to the data transmission, the method further includes:

[0011] Determine a second target node in the perception nodes of the perception layer, the second target node is a non-faulty node, and adjacent nodes of the second target node in the perception layer are all non-faulty nodes;

[0012] Performing data association analysis on the second target node and the adjacent nodes of the second target node to determine whether there is data association between the perception data of the second target node and the adjacent nodes of the second target node according to the result of the association analysis;

[0013] When the perception data of the second target node and the neighboring nodes of the second target node have data association, constructing a training sample data set using the perception data of the second target node and the neighboring nodes of the second target node having data association;

[0014] The perception data generation model of the second target node is trained using the training sample data set.

[0015] Furthermore, the step of restoring the perception data of the faulty node based on the perception data transmitted in real time by the associated node specifically includes:

[0016] The perception data transmitted in real time by the associated node is used as input data and input into the perception data generation model to generate the perception data of the second target node.

[0017] Furthermore, the step of performing correlation analysis on the second target node and the adjacent nodes of the second target node specifically includes:

[0018] Constructing the first perception data sequence of the second target node:

[0019] TX={tx1,tx2,…,tx m},

[0020] And the second perception data sequence of the adjacent node:

[0021] RX i ={rx1,rx2,…,rx m},

[0022] Wherein i is a positive integer between 1 and n, n is the number of neighboring nodes of the second target node, and m is the number of pre-configured sampling points;

[0023] The first perception data sequence TX and the second perception data sequence RX i Perform hysteresis relationship analysis and spectrum analysis;

[0024] Whether the perception data of the second target node and the adjacent nodes of the second target node have data correlation is determined according to the results of the hysteresis relationship analysis and the spectrum analysis.

[0025] Further, the first perception data sequence TX and the second perception data sequence RX iThe steps to perform hysteresis relationship analysis include:

[0026] Calculate the first mean μ of the first perception data sequence TX T And the second perception data sequence RX i The second mean μ R :

[0027]

[0028] Where t is a positive integer between 1 and m;

[0029] Calculate the first standard deviation σ of the first sensory data sequence TX T And the second perception data sequence RX i The second standard deviation σ R :

[0030]

[0031] Constructing the first perception data sequence TX and the second perception data sequence RX i The cross-correlation function is:

[0032]

[0033] Where k is the hysteresis variable;

[0034] Determine the first perceptual data sequence TX and the second perceptual data sequence RX based on the cross-correlation function i hysteresis relationship.

[0035] Further, the first perception data sequence TX and the second perception data sequence RX are determined based on the cross-correlation function. i The steps of the hysteresis relationship specifically include:

[0036] Calculate the cross-correlation function value CCF between the hysteresis variable k from -m to m i (k);

[0037] Construct the cross-correlation function value CCF i (k) a cross-correlation curve between -m and m with respect to the hysteresis variable k;

[0038] The cross-correlation peak is determined on the cross-correlation curve:

[0039]

[0040] Determine satisfaction The peak hysteresis k max ;

[0041] Determine the peak hysteresis k max Is it 0?

[0042] When k max =0, determine the first perception data sequence TX and the second perception data sequence RX i There is no hysteresis relationship.

[0043] Further, the first perception data sequence TX and the second perception data sequence RX i The steps to perform spectrum analysis include:

[0044] The first perception data sequence TX and the second perception data sequence RX are respectively i Perform discrete Fourier transform to obtain the third perception data sequence FTX and the fourth perception data sequence FRX i :

[0045]

[0046] Where f is a frequency variable and for f∈[1,m] it satisfies:

[0047]

[0048] Calculate the third perception data sequence FTX and the fourth perception data sequence FRX i The cross power spectral density is:

[0049] CPSD(f)=ftx f frx* f ,

[0050] Where f is a positive integer between 1 and m, frx* f for frx f The complex conjugate of ;

[0051] Calculate the phase difference of the cross power spectral density CPSD(f) at frequency f:

[0052]

[0053] Wherein Re[CPSD(f)] represents the real part of the cross power spectrum density CPSD(f), and Im[CPSD(f)] represents the imaginary part of the cross power spectrum density CPSD(f);

[0054] Determine the first perception data sequence TX and the second perception data sequence RX based on the phase difference φ(f) i spectral relationship.

[0055] Further, the first perception data sequence TX and the second perception data sequence RX are determined based on the phase difference φ(f). i The steps of determining the spectrum relationship specifically include:

[0056] Calculating the phase difference φ(f) between the frequency variable f from 1 to m;

[0057] Determining whether the phase difference φ(f) is a constant;

[0058] When the phase difference φ(f) is a constant, the first perception data sequence TX and the second perception data sequence RX are determined. i There is a displacement relationship;

[0059] When the phase difference φ(f) is not a constant, determining whether the phase difference φ(f) is linearly related to the frequency variable f;

[0060] When the phase difference is linearly related to the frequency variable f, the first perception data sequence TX and the second perception data sequence RX are determined. i Has a hysteresis relationship.

[0061] Furthermore, the step of judging whether there is data correlation between the perception data of the second target node and the adjacent nodes of the second target node according to the results of the hysteresis relationship analysis and the spectrum analysis specifically includes:

[0062] When the first sensing data sequence TX and the second sensing data sequence RX i When there is a hysteresis relationship or a displacement relationship, it is determined that there is data correlation between the perception data of the second target node and the i-th adjacent node of the second target node.

[0063] The second aspect of the present invention proposes a data recovery device for data perception, comprising a processor and a memory, wherein the processor executes a computer program stored in the memory to implement the data recovery method for data perception described in any one of the first aspect of the present invention.

[0064] The present invention proposes a data recovery method and device for data perception, which monitors the data transmission dynamics of each perception node in a perception layer, identifies faulty nodes in the perception nodes according to the data transmission dynamics, and determines whether there are associated nodes in the perception layer that are adjacent to the faulty nodes and meet preset conditions. The preset conditions include the distance of the associated nodes and the type of perception data. When there are associated nodes in the perception layer that are adjacent to the faulty nodes and meet the preset conditions, the perception data transmitted in real time by the associated nodes are obtained, and the perception data of the faulty nodes are restored based on the perception data transmitted in real time by the associated nodes, which can effectively restore the perception data of unavailable nodes in the perception layer. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 is a schematic diagram of a data recovery system applied to data perception provided by an embodiment of the present invention;

[0066] Figure 2 It is a flow chart of a data recovery method applied to data perception provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0067] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0068] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0069] In the description of the present invention, the term "multiple" refers to two or more. Unless otherwise clearly defined, the orientation or positional relationship indicated by the terms "upper", "lower", etc. is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present invention and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation of the present invention. The terms "connection", "installation", "fixation", etc. should be understood in a broad sense. For example, "connection" can be a fixed connection, a detachable connection, or an integral connection; it can be directly connected or indirectly connected through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to the specific circumstances. In addition, the terms "first", "second", etc. are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first", "second", etc. can explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, the meaning of "multiple" is two or more.

[0070] In the description of this specification, the description of the terms "one embodiment", "some implementations", "specific embodiments", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0071] A data recovery method and device for data perception provided according to some embodiments of the present invention will be described below with reference to the accompanying drawings.

[0072] like Figure 1As shown, a data recovery method and device for data perception proposed by the present invention is applied to a data recovery system, and the data recovery system includes a perception layer, a business layer and a data layer. The perception layer includes perception nodes arranged at various monitoring positions for obtaining perception data, each perception node is composed of one or more sensing devices, and the monitoring position is a geographical location where the perception node needs to be arranged in advance in order to realize the corresponding business scenario. It should be known that the monitoring position of the perception node is relatively fixed, that is, the monitoring position of the perception node will not change without human movement. The business layer includes a business server for running a business program that realizes a business service function according to the perception data. The business server can be an independent server device or a server cluster containing multiple server devices for jointly realizing the business service function. The business service function includes but is not limited to background management, remote control, data statistics, data analysis, data display, cloud storage and cloud computing functions provided by front-end service programs and back-end service programs. The data layer includes a data storage device for storing the perception data and business data and a data recovery device for recovering the perception data of unavailable nodes. The business data includes but is not limited to user data, authority data, perception node location data, perception node configuration data, statistical analysis data, etc. The data storage device may be a dedicated or general database server, a personal computer or a workstation, etc. Similarly, the data recovery device may be a dedicated or general cloud computing server, a personal computer or a workstation, etc.

[0073] like Figure 2 As shown, the first aspect of the present invention proposes a data recovery method applied to data perception, comprising:

[0074] Monitor the data transmission dynamics of each sensing node in the perception layer;

[0075] dynamically identifying a faulty node among the sensing nodes based on the data transmission;

[0076] Determine whether there is an associated node in the perception layer that is adjacent to the faulty node and meets preset conditions, wherein the preset conditions include a distance between the associated nodes and a type of perception data;

[0077] When there is an associated node in the perception layer that is adjacent to the faulty node and meets a preset condition, acquiring perception data transmitted in real time by the associated node;

[0078] The perception data of the faulty node is restored based on the perception data transmitted in real time by the associated node.

[0079] Specifically, the data transmission dynamics is the real-time state of the perception node transmitting data to the background, that is, the business layer and / or the data layer. When the perception node is in a normal working state, that is, when the perception node is in a non-fault state, it will obtain the perception data in real time and transmit the perception data to the background. According to different business implementation architectures, the transmission mode of the perception data will have different transmission modes. For example, in some embodiments of the present invention, the perception node, the data recovery device and the data storage device are respectively connected to the business server in communication, and the perception node of the perception layer transmits the acquired real-time perception data to the business server of the business layer through the communication network, and then transmits it to the data recovery device and / or data storage device of the data layer through the business server. In the technical solutions of other embodiments of the present invention, the perception node is directly connected to the data recovery device in communication, and the business server is connected to the data storage device and the data recovery device in communication, and the perception node of the perception layer directly transmits the acquired real-time perception data to the data recovery device of the data layer through the communication network, and the data recovery device provides the perception data to the data storage device for persistent storage as needed, or provides the perception data to the business server according to business needs to realize the corresponding business function.

[0080] When any of the sensing nodes is damaged, has a network interruption, or has abnormal power supply, its data transmission is dynamically reflected in the service server or data recovery device connected thereto as receiving real-time sensing data transmitted by an inappropriate sensing node, or the received real-time sensing data of the sensing node is abnormal data, etc. In this case, in the step of dynamically identifying the faulty node among the sensing nodes according to the data transmission, the corresponding sensing node is determined as the faulty node.

[0081] In the technical solutions of some implementations of the present invention, the step of determining whether there is an associated node in the perception layer that is adjacent to the faulty node and meets a preset condition specifically includes:

[0082] Determine the faulty node as a first target node;

[0083] Determine, in the perception layer, adjacent nodes of the first target node, the adjacent nodes being perception nodes whose distance from the first target node is less than a preset distance threshold;

[0084] Obtaining the data transmission dynamics and perceived data type of the adjacent node;

[0085] dynamically identifying a non-faulty node among the adjacent nodes based on the data transmission;

[0086] A non-faulty node among the adjacent nodes whose sensed data type is the same as that of the first target node is determined as the associated node.

[0087] In order to avoid definition conflicts or confusion, in this embodiment, the faulty node for which the associated node needs to be determined is referred to as the first target node. In the step of determining whether there are associated nodes in the perception layer that are adjacent to the faulty node and meet preset conditions, the surrounding nodes should be non-faulty nodes. As mentioned above, each of the perception nodes is composed of one or more sensing devices, that is, each perception node can generally collect one or more types of perception data. For example, when the perception node is a weather data perception node, it includes at least a temperature sensor, a humidity sensor, a wind speed sensor, etc., and its perception data types include at least temperature data, humidity data, and wind speed data.

[0088] Further, after the step of dynamically identifying a faulty node in the sensing node according to the data transmission, the method further includes:

[0089] Determine a second target node in the perception nodes of the perception layer, the second target node is a non-faulty node, and adjacent nodes of the second target node in the perception layer are all non-faulty nodes;

[0090] Performing data association analysis on the second target node and the adjacent nodes of the second target node to determine whether there is data association between the perception data of the second target node and the adjacent nodes of the second target node according to the result of the association analysis;

[0091] When the perception data of the second target node and the neighboring nodes of the second target node have data association, constructing a training sample data set using the perception data of the second target node and the neighboring nodes of the second target node having data association;

[0092] The perception data generation model of the second target node is trained using the training sample data set.

[0093] In the technical solutions of some embodiments of the present invention, the step of determining the second target node in the perception node of the perception layer specifically includes:

[0094] After dynamically identifying any sensing node as a non-faulty node according to the data transmission, determining the non-faulty node as a candidate target node;

[0095] Obtaining a list of adjacent nodes of the candidate target node, wherein the list of adjacent nodes includes all adjacent nodes of the candidate target node;

[0096] Traversing the adjacent node list to determine whether the adjacent node list contains a faulty node;

[0097] When the adjacent node list does not include a faulty node, the candidate target node is determined as the second target node.

[0098] It should be known that, although it is very common for sensor nodes to fail, this is reflected on the basis of a huge number of sensor nodes. In addition, due to regular maintenance by maintenance personnel, most sensor nodes are in normal working condition on a daily basis, that is, each sensor node is likely to be determined as the second target node, thereby generating its corresponding sensory data generation model. Even if a sensor node does not meet the conditions for being determined as the second target node at a specific time point, after waiting for a period of time, after maintenance or replacement of the sensor by maintenance personnel, it will still be determined as the second target node in the next processing cycle to generate its corresponding sensory data generation model. Of course, if the sensor node does not have any adjacent nodes due to special geographical location or other reasons, it does not fall within the scope of application of the data recovery method of the present invention.

[0099] Furthermore, the step of restoring the perception data of the faulty node based on the perception data transmitted in real time by the associated node specifically includes:

[0100] The perception data transmitted in real time by the associated node is used as input data and input into the perception data generation model to generate the perception data of the second target node.

[0101] Specifically, the perception data generation model of the second target node can be a Markov model, a recurrent neural network model, a variational autoencoder model or other machine learning models for generating data sequences, which takes the time series of the perception data of one or more associated nodes as input data and outputs the time series of the perception data of the second target node.

[0102] Furthermore, the step of performing correlation analysis on the second target node and the adjacent nodes of the second target node specifically includes:

[0103] Constructing the first perception data sequence of the second target node:

[0104] TX={tx1,tx2,…,tx m},

[0105] And the second perception data sequence of the adjacent node:

[0106] RX i ={rx1,rx2,…,rx m},

[0107] Wherein i is a positive integer between 1 and n, n is the number of neighboring nodes of the second target node, and m is the number of pre-configured sampling points;

[0108] The first perception data sequence TX and the second perception data sequence RX i Perform hysteresis relationship analysis and spectrum analysis;

[0109] Whether the perception data of the second target node and the adjacent nodes of the second target node have data correlation is determined according to the results of the hysteresis relationship analysis and the spectrum analysis.

[0110] Specifically, according to different deployment situations of sensing nodes at different monitoring locations, the second target node may have no adjacent nodes, one adjacent node, or multiple adjacent nodes. When the adjacent nodes are multiple, the second sensing data sequences are also multiple, and the number of the second sensing data sequences is the same as the number of the adjacent nodes. In the step of performing hysteresis relationship analysis and spectrum analysis on the first sensing data sequence and the second sensing data sequence, specifically, hysteresis relationship analysis and spectrum analysis are performed on the first sensing data sequence and the second sensing data sequence corresponding to each adjacent phase point.

[0111] As mentioned above, the perception nodes of the perception layer will acquire the perception data in real time and transmit the perception data to the background under normal working conditions, and construct the first perception data sequence TX of the second target node and the second perception data sequence RX of the adjacent node. i In the step of sampling the sensing data transmitted by the second target node and its adjacent nodes at a preset sampling period to obtain the first sensing data sequence TX and the second sensing data sequence RX i , wherein the first perception data sequence TX and the second perception data sequence RX i is sampled from the sensing data transmitted by the second target node and its adjacent nodes in the same time period, and the first sensing data sequence TX and the second sensing data sequence RX i There is a one-to-one correspondence, that is, the sampling time points of tx1 and rx1 are the same, the sampling time points of tx2 and rx2 are the same, and so on.

[0112] In the technical solution of the present invention, m is the number of sampling points for data sampling from the perception data transmitted by the second target node and its neighboring nodes. The value of m is usually an empirical value. In the actual implementation process, the value of m can be configured according to actual needs. It should be known that when the value of m is too small, the first perception data sequence TX and the second perception data sequence RX iThe steps of performing hysteresis relationship analysis and spectrum analysis cannot accurately reflect the correlation between the two data sequences. When the value of m is too large, it will bring unnecessary calculations, resulting in a long analysis process and reduced analysis efficiency. In the technical solution of the above implementation, t is a counting variable less than m and t is a positive integer, that is, the counting variable t does not represent a fixed parameter with a specific meaning, and it is only used for iterative counting between 1 and m during the calculation process.

[0113] Further, the first perception data sequence TX and the second perception data sequence RX i The steps to perform hysteresis relationship analysis include:

[0114] Calculate the first mean μ of the first perception data sequence TX T And the second perception data sequence RX i The second mean μ R :

[0115]

[0116] Where t is a positive integer between 1 and m;

[0117] Calculate the first standard deviation σ of the first sensory data sequence TX T And the second perception data sequence RX i The second standard deviation σ R :

[0118]

[0119] Constructing the first perception data sequence TX and the second perception data sequence RX i The cross-correlation function is:

[0120]

[0121] Where k is the hysteresis variable;

[0122] Determine the first perceptual data sequence TX and the second perceptual data sequence RX based on the cross-correlation function i hysteresis relationship.

[0123] Further, the first perception data sequence TX and the second perception data sequence RX are determined based on the cross-correlation function. i The steps of the hysteresis relationship specifically include:

[0124] Calculate the cross-correlation function value CCF between the hysteresis variable k from -m to m i (k);

[0125] Construct the cross-correlation function value CCFi (k) a cross-correlation curve between -m and m with respect to the hysteresis variable k;

[0126] The cross-correlation peak is determined on the cross-correlation curve:

[0127]

[0128] Determine satisfaction The peak hysteresis k max ;

[0129] Determine the peak hysteresis k max Is it 0?

[0130] When k max =0, determine the first perception data sequence TX and the second perception data sequence RX i There is no hysteresis relationship.

[0131] The above embodiment determines the first perception data sequence TX and the second perception data sequence RX by analyzing the change of the hysteresis variable k reflected in the cross-correlation curve. i When there is no hysteresis relationship between the two, the cross-correlation peak on the cross-correlation curve falls on k max =0.

[0132] In the technical solutions of some embodiments of the present invention, when determining the peak hysteresis value k max After the step of whether it is 0, it also includes:

[0133] When k max ≠0, determine the first perception data sequence TX and the second perception data sequence RX i With a hysteresis relationship, the second sensing data sequence RX i The delay time relative to the first sensing data sequence TX is:

[0134] T=k max ·t sample ,

[0135] where t sample is a sampling period for sampling the sensing data transmitted by the second target node and its adjacent nodes.

[0136] Further, the first perception data sequence TX and the second perception data sequence RX i The steps to perform spectrum analysis include:

[0137] The first perception data sequence TX and the second perception data sequence RX are respectively iPerform discrete Fourier transform to obtain the third perception data sequence FTX and the fourth perception data sequence FRX i :

[0138]

[0139] Where f is a frequency variable and for f∈[1,m] it satisfies:

[0140]

[0141] Calculate the third perception data sequence FTX and the fourth perception data sequence FRX i The cross power spectral density is:

[0142] CPSD(f)=ftx f frx* f ,

[0143] Where f is a positive integer between 1 and m, frx* f for frx f The complex conjugate of ;

[0144] Calculate the phase difference of the cross power spectrum density GPSD(f) at frequency f:

[0145]

[0146] Wherein Re[CPSD(f)] represents the real part of the cross power spectrum density CPSD(f), and Im[CPSD(f)] represents the imaginary part of the cross power spectrum density CPSD(f);

[0147] Determine the first perception data sequence TX and the second perception data sequence RX based on the phase difference φ(f) i spectral relationship.

[0148] Further, the first perception data sequence TX and the second perception data sequence RX are determined based on the phase difference φ(f). i The steps of determining the spectrum relationship specifically include:

[0149] Calculating the phase difference φ(f) between the frequency variable f from 1 to m;

[0150] Determining whether the phase difference φ(f) is a constant;

[0151] When the phase difference φ(f) is a constant, the first perception data sequence TX and the second perception data sequence RX are determined. i There is a displacement relationship;

[0152] When the phase difference φ(f) is not a constant, determining whether the phase difference φ(f) is linearly related to the frequency variable f;

[0153] When the phase difference is linearly related to the frequency variable f, the first perception data sequence TX and the second perception data sequence RX are determined. i Has a hysteresis relationship.

[0154] It should be known that due to data errors or external interference factors, the phase difference φ(f) will still fluctuate within a certain range. Therefore, the step of determining whether the phase difference φ(f) is a constant is specifically to determine whether the phase difference φ(f) fluctuates within a certain range around a specific constant. Similarly, the step of determining whether the phase difference is linearly related to the frequency variable f is specifically to determine whether the phase difference φ(f) fluctuates within a certain range around a specific straight line.

[0155] In the technical solutions of some embodiments of the present invention, the step of determining whether the phase difference φ(f) is a constant comprises:

[0156] Calculate the phase difference mean of the phase difference φ(f) in the frequency range of 1 to m:

[0157]

[0158] Determine the phase difference φ(f) relative to the mean phase difference in the frequency range of 1 to m The first maximum phase difference fluctuation amplitude:

[0159]

[0160] Determine the first maximum phase difference fluctuation amplitude φ1 max Whether it is greater than the preset fluctuation amplitude threshold;

[0161] When the first maximum phase difference fluctuation amplitude φ1 max When the phase difference φ(f) is less than or equal to a preset fluctuation amplitude threshold, the phase difference φ(f) is determined to be a constant.

[0162] In the technical solutions of some implementations of the present invention, the step of determining whether the phase difference φ(f) is linearly related to the frequency variable f comprises:

[0163] Fitting the phase difference function using the least squares method A trend line function φ'(f) of which the trend line function φ'(f) is a straight line function;

[0164] Determine the second maximum phase difference fluctuation amplitude of the phase difference φ(f) relative to the trend line function φ'(f) in the frequency range of 1 to m:

[0165]

[0166] Determine the second maximum phase difference fluctuation amplitude φ2 max Whether it is greater than the preset fluctuation amplitude threshold;

[0167] When the second maximum phase difference fluctuation amplitude φ2 max When the phase difference φ(f) is less than or equal to a preset fluctuation amplitude threshold, it is determined that the phase difference φ(f) is linearly related to the frequency variable f.

[0168] Furthermore, the step of judging whether there is data correlation between the perception data of the second target node and the adjacent nodes of the second target node according to the results of the hysteresis relationship analysis and the spectrum analysis specifically includes:

[0169] When the first sensing data sequence TX and the second sensing data sequence RX i When there is a hysteresis relationship or a displacement relationship, it is determined that there is data correlation between the perception data of the second target node and the i-th adjacent node of the second target node.

[0170] The second aspect of the present invention proposes a data recovery device for data perception, comprising a processor and a memory, wherein the processor executes a computer program stored in the memory to implement the data recovery method for data perception described in any one of the first aspect of the present invention.

[0171] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0172] According to the embodiments of the present invention as described above, these embodiments do not describe all the details in detail, nor do they limit the invention to the specific embodiments described. Obviously, many modifications and changes can be made based on the above description. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can make good use of the present invention and the modified use based on the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A data recovery method applied to data perception, characterized in that: include: Monitor the data transmission dynamics of each sensing node in the perception layer; dynamically identifying a faulty node among the sensing nodes based on the data transmission; Determine whether there is an associated node in the perception layer that is adjacent to the faulty node and meets preset conditions, wherein the preset conditions include a distance between the associated nodes and a type of perception data; When there is an associated node in the perception layer that is adjacent to the faulty node and meets a preset condition, acquiring perception data transmitted in real time by the associated node; Restoring the perception data of the faulty node based on the perception data transmitted in real time by the associated node; After the step of dynamically identifying a faulty node in the sensing node according to the data transmission, the method further includes: Determine a second target node in the perception nodes of the perception layer, the second target node is a non-faulty node, and adjacent nodes of the second target node in the perception layer are all non-faulty nodes; Performing data association analysis on the second target node and the adjacent nodes of the second target node to determine whether there is data association between the perception data of the second target node and the adjacent nodes of the second target node according to the result of the association analysis; When the perception data of the second target node and the neighboring nodes of the second target node have data association, constructing a training sample data set using the perception data of the second target node and the neighboring nodes of the second target node having data association; Using the training sample data set to train a perception data generation model of the second target node; The step of restoring the perception data of the faulty node based on the perception data transmitted in real time by the associated node specifically includes: The perception data transmitted in real time by the associated node is used as input data and input into the perception data generation model to generate the perception data of the second target node.

2. The data recovery method applied to data perception according to claim 1, characterized in that: The step of performing association analysis on the second target node and the adjacent nodes of the second target node specifically includes: Constructing the first perception data sequence of the second target node: TX={tx1,tx2,…,tx m }, And the second perception data sequence of the adjacent node: RX i ={rx1,rx2,…,rx m }, Wherein i is a positive integer between 1 and n, n is the number of neighboring nodes of the second target node, and m is the number of pre-configured sampling points; The first perception data sequence TX and the second perception data sequence RX i Perform hysteresis relationship analysis and spectrum analysis; Whether the perception data of the second target node and the adjacent nodes of the second target node have data correlation is determined according to the results of the hysteresis relationship analysis and the spectrum analysis.

3. The data recovery method applied to data perception according to claim 2, characterized in that: The first perception data sequence TX and the second perception data sequence RX i The steps to perform hysteresis relationship analysis include: Calculate the first mean μ of the first perception data sequence TX T And the second perception data sequence RX i The second mean μ R : Where t is a positive integer between 1 and m; Calculate the first standard deviation σ of the first sensory data sequence TX T And the second perception data sequence RX i The second standard deviation σ R : Constructing the first perception data sequence TX and the second perception data sequence RX i The cross-correlation function is: Where k is the hysteresis variable; Determine the first perceptual data sequence TX and the second perceptual data sequence RX based on the cross-correlation function i hysteresis relationship.

4. The data recovery method applied to data perception according to claim 3, characterized in that: Determine the first perceptual data sequence TX and the second perceptual data sequence RX based on the cross-correlation function i The steps of the hysteresis relationship specifically include: Calculate the cross-correlation function value CCF between the hysteresis variable k from -m to m i (k); Construct the cross-correlation function value CCF i (k) a cross-correlation curve between -m and m with respect to the hysteresis variable k; The cross-correlation peak is determined on the cross-correlation curve: Determine satisfaction The peak hysteresis k max ; Determine the peak hysteresis k max Is it 0? When k max =0, determine the first perception data sequence TX and the second perception data sequence RX i There is no hysteresis relationship.

5. The data recovery method applied to data perception according to claim 2, characterized in that: The first perception data sequence TX and the second perception data sequence RX i The steps to perform spectrum analysis include: The first perception data sequence TX and the second perception data sequence RX are respectively i Perform discrete Fourier transform to obtain the third perception data sequence FTX and the fourth perception data sequence FRX i : Where f is a frequency variable and for f∈[1,m] it satisfies: Calculate the third perception data sequence FTX and the fourth perception data sequence FRX i The cross power spectral density is: CPSD(f)=ftx f ·frx* f , Where f is a positive integer between 1 and m, frx* f for frx f The complex conjugate of ; Calculate the phase difference of the cross power spectral density CPSD(f) at frequency f: Wherein Re[CPSD(f)] represents the real part of the cross power spectrum density CPSD(f), and Im[CPSD(f)] represents the imaginary part of the cross power spectrum density CPSD(f); Determine the first perception data sequence TX and the second perception data sequence RX based on the phase difference φ(f) i spectral relationship.

6. The data recovery method applied to data perception according to claim 5, characterized in that: Determine the first perception data sequence TX and the second perception data sequence RX based on the phase difference φ(f) i The steps of determining the spectrum relationship specifically include: Calculating the phase difference φ(f) between the frequency variable f from 1 to m; Determining whether the phase difference φ(f) is a constant; When the phase difference φ(f) is a constant, the first perception data sequence TX and the second perception data sequence RX are determined. i There is a displacement relationship; When the phase difference φ(f) is not a constant, determining whether the phase difference φ(f) is linearly related to the frequency variable f; When the phase difference is linearly related to the frequency variable f, the first perception data sequence TX and the second perception data sequence RX are determined. i Has a hysteresis relationship.

7. The data recovery method applied to data perception according to claim 2, characterized in that: The step of judging whether the sensing data of the second target node and the adjacent nodes of the second target node have data correlation according to the results of the hysteresis relationship analysis and the spectrum analysis specifically includes: When the first sensing data sequence TX and the second sensing data sequence RX i When there is a hysteresis relationship or a displacement relationship, it is determined that there is data correlation between the perception data of the second target node and the i-th adjacent node of the second target node.

8. A data recovery device applied to data perception, characterized in that: It comprises a processor and a memory, wherein the processor executes a computer program stored in the memory to implement the data recovery method applied to data perception as described in any one of claims 1 to 7.

Images