Security distributed detection method and device for cognitive wireless sensor network

Abstract

The application provides a security distributed detection method and device for a cognitive wireless sensor network. The method comprises the following steps: a part of nodes in the sensor network generate fake data based on their own observation results, and send the fake data to a fusion center to obtain a mean value of global detection statistics on a side of an eavesdropping center under a hypothesis H0, a mean value under a hypothesis H1, and a variance under the hypothesis H1; a correction deviation coefficient on the side of the eavesdropping center is determined according to the mean value under the hypothesis H0, the mean value under the hypothesis H1, and the variance under the hypothesis H1; a value range of a correction deviation coefficient on a side of the fusion center is determined according to the correction deviation coefficient on the side of the eavesdropping center; and the optimal parameter of the security distributed detection for the cognitive wireless sensor network is determined according to the value range of the correction deviation coefficient on the side of the fusion center. The method provided by the application solves the problem that the network is subjected to eavesdropping attack from the perspective of a physical layer, the method does not need a complex algorithm with a large amount of calculation, and the power consumption of the wireless sensor network is reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of network security, in particular to a security distributed detection method and device for cognitive wireless sensor networks. BACKGROUND

[0002] Cognitive wireless sensor networks (CWSN) are composed of a large number of low-cost, low-power, mobile micro-sensor. In recent years, landslides, mudslides, forest fires and other geological disasters in China caused many loss of life and property, communication base stations were destroyed. There is an urgent need for cognitive wireless sensor networks that can quickly, flexible layout, not dependent on fixed communication frequency self-organizing network technology, through the monitoring of disasters, to achieve effective early warning and rescue. In addition, wireless sensor networks provide a new technical approach for spatial intrusion target detection, such as the layout around the airport, to achieve effective detection of illegal unmanned aerial vehicle black flight. Wireless sensor network perception data can contain a lot of sensitive information, such as military and medical applications of wireless sensor networks, the network data accuracy and privacy requirements are very high, however, due to the sensor network is often deployed in harsh or dangerous monitoring areas, sensor nodes are easily captured by attackers, attackers through active data forgery or eavesdropping to attack the security of wireless sensor networks. Therefore, the distributed detection of wireless sensor networks has an urgent need and important significance for network security.

[0003] Among them, there are two types of eavesdropping attacks, passive and active. Passive eavesdroppers detect information by eavesdropping on data transmission between local sensors and legitimate users; active eavesdroppers disguise themselves as friendly nodes and send queries to some local sensors. Since passive eavesdropping is the basis for active eavesdropping and is difficult to detect and defend, the present application takes passive eavesdropping as a starting point.

[0004] And passive eavesdropping currently uses encryption algorithms to prevent eavesdropping and protect the confidentiality of the system. However, encryption algorithms involve very complex algorithms, and for long-term operation of WSN, complex algorithms consume more power. SUMMARY

[0005] The present application aims to solve the defects of the prior art, and provides a security distributed detection method and device for cognitive wireless sensor networks.

[0006] A security distributed detection method for cognitive wireless sensor networks, the method comprising the following steps:

[0007] obtaining a mean value of the global test statistic on the side of the eavesdropping center under a hypothesis H0, a mean value under a hypothesis H1, and a variance under the hypothesis H1; wherein the H0 represents that a target detected by the sensor node does not exist; and the H1 represents that the target detected by the sensor node exists;

[0008] determining a correction bias coefficient on the side of the eavesdropping center according to the mean value under the H0, the mean value under the H1, and the variance under the H1;

[0009] determining a correction bias coefficient on the side of the fusion center according to the correction bias coefficient on the side of the eavesdropping center;

[0010] determining the security distributed detection mode for the cognitive wireless sensor network according to the correction bias coefficient on the side of the fusion center.

[0011] Further, the security distributed detection method for the cognitive wireless sensor network as described above, the obtaining of the mean value of the global test statistic on the side of the eavesdropping center under the hypothesis H0, the mean value under the hypothesis H1, and the variance under the hypothesis H1 comprises:

[0012] obtaining target measurement data sent by the sensor node and eavesdropped by the eavesdropping center;

[0013] determining a probability of distribution of the eavesdropping center under the hypothesis H0 and a probability of distribution under the hypothesis H1 according to the target measurement data;

[0014] determining a mean value and a variance under the hypothesis H0 and a mean value and a variance under the hypothesis H1 of the target measurement data according to the probability of distribution of the eavesdropping center under the hypothesis H0 and the probability of distribution under the hypothesis H1;

[0015] determining the mean value of the global test statistic on the side of the eavesdropping center under the hypothesis H0, the mean value under the hypothesis H1, and the variance under the hypothesis H1 based on a linear weighted fusion method according to the mean value and the variance under the hypothesis H0 and the mean value and the variance under the hypothesis H1 of the target measurement data.

[0016] Further, the security distributed detection method for the cognitive wireless sensor network as described above, the mean value of the global test statistic on the side of the eavesdropping center under the hypothesis H0 is obtained by using the following formula:

[0017]

[0018] wherein the w i represents a weighted coefficient corresponding to the sensor node i, M represents a number of local detection samples of the sensor node i, represents a mean value of the measurement data of the sensor i under the hypothesis H0; K represents a number of the cheating nodes; Pf represents the false alarm probability corresponding to the eavesdropping center; a represents the proportion of the deceptive nodes in the network to all nodes; A i is the average signal-to-noise ratio of the signal received by the i th sensor node.

[0019] Further, in the security distributed detection method for cognitive wireless sensor networks as described above, the mean value of the global test statistic on the eavesdropping center side under the assumption of H1 is obtained by using the following formula:

[0020]

[0021] Further, in the security distributed detection method for cognitive wireless sensor networks as described above, the variance of the global test statistic on the eavesdropping center side under the assumption of H1 is obtained by using the following formula:

[0022]

[0023] wherein the represents the variance of the measurement data of sensor i under the assumption of H1; represents the mean value of the measurement data of sensor i under the assumption of H0; P d represents the detection probability corresponding to the eavesdropping center.

[0024] Further, in the security distributed detection method for cognitive wireless sensor networks as described above, the correction bias coefficient is obtained by using the following formula:

[0025]

[0026] wherein T represents the global test statistic, E(T|H1) represents the mean value of the global test statistic under the assumption of H1; E(T|H0) represents the mean value of the global test statistic under the assumption of H0, and Var(T|H1) represents the variance of the global test statistic under the assumption of H1.

[0027] Further, in the security distributed detection method for cognitive wireless sensor networks as described above, the correction bias coefficient of the fusion center is determined according to the correction bias coefficient on the eavesdropping center side, which includes:

[0028] in the case that the correction bias coefficient on the eavesdropping center side is 0, the proportion a of the deceptive nodes in the network to all nodes is determined;

[0029] the correction bias coefficient of the fusion center is determined according to the proportion a of the deceptive nodes to all nodes.

[0030] Further, the security distributed detection method for the cognitive wireless sensor network according to the above, the security distributed detection method for the cognitive wireless sensor network is determined according to the modified bias coefficient of the fusion center, and includes:

[0031] The modified bias coefficient of the fusion center is obtained by the following formula:

[0032]

[0033] Wherein, The modified bias coefficient of the fusion center corresponding to the global test statistic Z of all honest nodes; The modified bias coefficient of the fusion center corresponding to the global test statistic Z of all honest nodes.

[0034] The security distributed detection method for the cognitive wireless sensor network is determined according to the value range of the modified bias coefficient of the fusion center.

[0035] Further, the security distributed detection method for the cognitive wireless sensor network according to the above, in the case that the modified bias coefficient of the fusion center takes the maximum value, the security distributed detection method for the cognitive wireless sensor network is the most secure.

[0036] A security distributed detection device for the cognitive wireless sensor network, the device includes:

[0037] An acquisition unit is configured to acquire the mean value of the global test statistic under the assumption H0, the mean value under the assumption H1 and the variance under the assumption H1 on the eavesdropping center side; wherein, the H0 represents that the target detected by the sensor node does not exist; the H1 represents that the target detected by the sensor node exists;

[0038] A determination unit is configured to determine the modified bias coefficient of the eavesdropping center side according to the mean value under the assumption H0, the mean value under the assumption H1 and the variance under the assumption H1.

[0039] The determination unit is further configured to determine the value range of the modified bias coefficient of the fusion center side according to the modified bias coefficient of the eavesdropping center side.

[0040] The determination unit is further configured to determine the security distributed detection method for the cognitive wireless sensor network according to the value range of the modified bias coefficient of the fusion center.

[0041] An electronic device includes a memory, a processor and a computer program stored on the memory and executable on the processor, and the processor executes the program to realize the security distributed detection method for the cognitive wireless sensor network according to the above.

[0042] The secure distributed detection method for cognitive wireless sensor networks provided by this invention determines the range of values ​​for the correction deviation coefficient of the fusion center by obtaining the correction deviation coefficient of the eavesdropping center. The number of sensor nodes is determined by the detection probability corresponding to maximizing the correction deviation coefficient of the fusion center. Finally, secure distributed detection of the cognitive wireless sensor network is performed based on the number of sensor nodes. This method solves the problem of network being subjected to eavesdropping attacks from the physical layer perspective. This approach does not require computationally complex algorithms and reduces the power consumption of the wireless sensor network. Attached Figure Description

[0043] Figure 1 This invention provides a secure distributed detection method for cognitive wireless sensor networks.

[0044] Figure 2 This is a schematic diagram of distributed detection in a cognitive wireless sensor network;

[0045] Figure 3 This is the curve showing how FC's detection performance changes with the probability of local detection by a malicious attack node;

[0046] Figure 4 This is a comparison of the ROC curves of Eve and FC under different detection probabilities;

[0047] Figure 5 This is a comparison of the ROC curves of Eve and FC under different numbers of sensors;

[0048] Figure 6 A schematic diagram of a secure distributed detection device for cognitive wireless sensor networks;

[0049] Figure 7 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention are described clearly and completely below. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0051] Figure 1 The secure distributed detection method for cognitive wireless sensor networks provided by this invention includes the following steps:

[0052] Step 101: Obtain the mean of the global test statistic at the eavesdropping center under hypothesis H0, the mean under hypothesis H1, and the variance under hypothesis H1; wherein, H0 indicates that the target detected by the sensor node does not exist; and H1 indicates that the target detected by the sensor node exists.

[0053] Specifically, Figure 2 This is a schematic diagram of distributed detection in a cognitive wireless sensor network, such as... Figure 2 As shown, assume H0 represents the absence of the target detected by the sensor node, and H1 represents the presence of the target detected by the sensor node. The CWSN consists of N sensors, some of which are spoofed nodes (RPNodes) and some are honest nodes (NRNodes). The spoofed nodes send opposite sensing data to the fusion center (FC). Data from all sensor nodes is connected to the FC through a set of parallel channels. x = [x1, x2, ..., x...] N ] represents the sensor's observation, where x i These are random variables or random vectors. The data outputs of all sensor nodes are eavesdropped on by the eavesdropping center through a set of parallel eavesdropping channels. It is assumed that the eavesdropping center has the same information about the detection algorithm as the fusion center (FC), including the sensor observation model, sensor decision rules, channel states, and hypothetical prior probabilities. It is also assumed that the sensor observations are conditionally independent and identically distributed; therefore, the detection problem for the i-th sensor node can be represented by the following binary detection model:

[0054]

[0055] Where s i Observation signal, The noise is additive white Gaussian noise in the sensing channel of the i-th sensor. Local detection by the sensor uses energy detection. After sampling M times locally, sensor node i obtains the target measurement data as follows:

[0056] The following section details the process of obtaining the mean, mean, and variance of the global test statistic at the eavesdropping center under hypothesis H0, hypothesis H1, and hypothesis H1 respectively:

[0057] First, the target measurement data sent by the sensor nodes is obtained by the eavesdropping center; second, the probability of the distribution of the eavesdropping center under the assumption H0 and the probability of the distribution under the assumption H1 are determined according to the target measurement data; third, the mean and variance of the target measurement data under the assumption H0 and the mean and variance under the assumption H1 are determined according to the probability of the distribution of the eavesdropping center under the assumption H0 and the probability of the distribution under the assumption H1; and finally, the mean under the assumption H0, the mean under the assumption H1 and the variance under the assumption H1 of the global test statistic of the eavesdropping center are determined based on the linear weighted fusion method according to the mean and variance of the target measurement data under the assumption H0 and the mean and variance under the assumption H1.

[0058] Specifically, according to the local detection statistic y i , the probability distribution thereof is derived to obtain the mean and variance expressions of the statistic y i under the assumptions H0 and H1. According to the central limit theorem, when M is large enough, y

[0059]

[0060] where μ i,0 represents the mean of the target measurement data of the sensor i under the assumption H0, represents the variance of the target measurement data of the sensor i under the assumption H0; μ i,1 represents the mean of the target measurement data of the sensor i under the assumption H1, represents the variance of the target measurement data of the sensor i under the assumption H1, which can be respectively expressed as:

[0061]

[0062]

[0063]

[0064]

[0065] where A i is the average signal-to-noise ratio of the received signal of the i-th sensor node.

[0066] The following will construct a model against eavesdropping attack:

[0067] It is assumed that the proportion of the cheating nodes in the network is α, the fusion center FC knows the identity of each node, and the eavesdropping center does not know. The honest nodes will directly send the target measurement data y iThe sending to FC, and for the cheating node, first use likelihood ratio test (LRT) to make local decision on the existence of target, and the LRT decision rule is expressed as:

[0068]

[0069] The decision threshold is λ i When the left formula is greater than the decision threshold, it is determined as H1, and when the left formula is less than the decision threshold, it is determined as H0. The cheating node first makes a local decision through the formula, reverses the decision result, and sends the distribution data conforming to the opposite result, so as to resist the eavesdropping attack. Since the observation values of the sensor nodes are independent and identically distributed, the decision threshold λ i of each sensor is expressed as λ = λ i . j When the local decision is H1, the target measurement data conforming to the H0 distribution is generated and sent to the FC, and when the local decision is H0, the target measurement data conforming to the H1 distribution is generated and sent to the FC. The attack model can be expressed as

[0070]

[0071] Where y M,j is the local measurement data sent by the jth cheating node to the FC.

[0072] For the data received by the eavesdropping center The probability distribution of the data received by the eavesdropping center under the assumption of H0 and the decision of H1 is as follows:

[0073]

[0074] For the data received by the eavesdropping center The probability distribution of the data received by the eavesdropping center under the assumption of H0 and the decision of H0 is as follows:

[0075]

[0076] For the data received by the eavesdropping center The probability distribution of the data received by the eavesdropping center under the assumption of H1 and the decision of H0 is as follows:

[0077]

[0078] For the data received by the eavesdropping center The probability distribution of the data received by the eavesdropping center under the assumption of H1 and the decision of H1 is as follows:

[0079]

[0080] Target measurement data Mean under hypothesis H0 Target measurement data Mean under hypothesis H1 Respectively as follows:

[0081]

[0082]

[0083] Target measurement data Variance under hypothesis H0 Target measurement data Variance under hypothesis H1 As follows:

[0084]

[0085]

[0086] The fusion center FC and the eavesdropping center both use a linear weighted combination fusion method, and the global test statistic of the eavesdropping center is:

[0087]

[0088] where w i represents a weighting coefficient. According to the central limit theorem, when N is large enough, approximately normally distributed. It is assumed that the first K sensor nodes are cheating nodes, and the remaining N-K nodes are honest nodes. The global test statistic of the eavesdropping center Mean under hypothesis H0 and the global test statistic of the eavesdropping center Mean under hypothesis H1 can be expressed as:

[0089]

[0090]

[0091] Similarly, the global test statistic of the eavesdropping center Variance under hypothesis H0 and the global test statistic of the eavesdropping center Variance under hypothesis H1 can be expressed as:

[0092]

[0093]

[0094] The following introduces how to obtain the global test statistic Z of all nodes at the fusion center FC. Since the fusion center FC knows the identity of the nodes, unlike the eavesdropping center which does not know the identity of the nodes, the fusion center FC needs to divide the mean and variance of the global statistic Z under H0 and H1 into two categories of cheating nodes and honest nodes for calculation.

[0095] Specifically, at the fusion center FC, the mean and variance of the measurement data received by the honest nodes can be calculated as follows:

[0096] The mean of the target measurement data of the honest nodes under the H0 assumption i,0 The mean of the target measurement data of the honest nodes under the H1 assumption i,1 is as follows, respectively:

[0097]

[0098]

[0099] Similarly, at the fusion center FC, the variance of the target measurement data of the honest nodes under the H0 assumption The variance of the target measurement data of the honest nodes under the H1 assumption is as follows, respectively:

[0100]

[0101]

[0102] According to formulas (21)-(24), at the fusion center FC, the mean E(Z|H0) of the global test statistic Z of all honest nodes under the assumption H0, and the mean E(Z|H1) under H1 are as follows:

[0103]

[0104]

[0105] Similarly, the variance Var(Z|H0) of the global test statistic Z of all honest nodes under the assumption H0, and the variance Var(Z|H1) under H1 are as follows:

[0106]

[0107]

[0108] At the fusion center FC, the mean and variance of the measurement data received by the cheating nodes can be calculated as follows:

[0109] The mean of the target measurement data of the cheating nodes under the H0 assumption Mean of target measurement data under H1 hypothesis Respectively as follows:

[0110]

[0111]

[0112] Similarly, the variance of the target measurement data of the cheating node under H0 hypothesis Variance of target measurement data under H1 hypothesis Respectively as follows:

[0113]

[0114]

[0115] According to formulas (29)-(32), at the fusion center FC, the global test statistic of all cheating nodes Mean under H0 hypothesis And mean under H1 hypothesis As follows:

[0116]

[0117]

[0118] Similarly, at the fusion center FC, the global test statistic of all cheating nodes Variance under H0 hypothesis And variance under H1 hypothesis As follows:

[0119]

[0120]

[0121] Step 102: According to the mean under H0 hypothesis, the mean under H1 hypothesis and the variance under H1 hypothesis, determine the correction bias coefficient on the eavesdropping center side.

[0122] Step 103: According to the correction bias coefficient on the eavesdropping center side, determine the correction bias coefficient value range of the fusion center FC.

[0123] Specifically, the application adopts a correction bias coefficient to measure the detection performance of the fusion center FC and the eavesdropping center, and the expression of the correction bias coefficient is as follows:

[0124]

[0125] Wherein, T represents the global detection statistic, d 2 is the correction bias coefficient.

[0126] Since when the correction deviation coefficient d 2 The smaller, the lower the detection performance; on the contrary, the higher the detection performance. Therefore, in order to achieve the purpose of defending eavesdropping attack, the detection performance of the fusion center FC is as large as possible, and the detection performance of the eavesdropping center is as small as possible, and based on this, the following security constraint distributed detection problem model is established:

[0127]

[0128]

[0129] According to formula (37), (17), (18), (20) is substituted and simplified, the deviation coefficient of the eavesdropping center is as follows:

[0130]

[0131] Let the weight w i =1 / N, since the sensors are independent and identically distributed, so A i =A, σ i =σ, then

[0132]

[0133] In order to make It is to make

[0134] Then:

[0135]

[0136] That is, when the proportion of the cheating node reaches , the eavesdropping center has no detection performance, and the FC will reach complete secrecy.

[0137] Step 104: determining the security distributed detection mode of the cognitive wireless sensor network according to the value range of the correction deviation coefficient of the fusion center FC.

[0138] Specifically, the case of maximizing the detection performance of the fusion center FC under the security constraint condition is analyzed below. Since the target measurement data of the fusion center FC is composed of two parts of the cheating node and the honest node, the corresponding correction deviation coefficient is also composed of two parts, finally, the detection performance of the fusion center FC can be expressed as:

[0139]

[0140] Wherein, Indicates the correction deviation coefficient of the cheating node, Let represent the correction deviation coefficient of honest nodes. Since the proportion of deceitful nodes to all nodes is α, the total correction deviation coefficient of the fusion center FC is given by formula (42).

[0141] According to formula (42), the overall detection performance of the fusion center FC can be obtained as follows:

[0142]

[0143] Let the weight w i =1 / N, A i =A,σ i =σ, then formula (43) can be simplified to:

[0144]

[0145] Among them, P f P represents the local false alarm probability (the probability that the fusion center FC will determine the target as H1 under the assumption of H0). d For the local detection probability (assuming H1, the probability that the fusion center FC will determine the target as H1), from It can be seen from this that α, P f All are P d The function, therefore, find the one that makes The largest P d The value of P maximizes the detection performance of the fusion center FC, while P d The local operating point of the sensor can be obtained by setting its operating point to achieve the corresponding local detection probability.

[0146] The performance analysis of the safety constraint distribution method of the present invention is as follows:

[0147] To verify the effectiveness of the method provided by this invention, 1000 Monte Carlo experiments were conducted, with the signal SNR set to 0dB, M=20, N=30, and the noise set to Gaussian white noise with a mean of 0 and a variance of 1. Figure 3 yes With P d From the changes, we can see that when 0 ≤ P d Within the range of ≤1, It is a convex function, but under these simulation conditions, when P d When Pd is less than or equal to 0.16, the proportion of deceitful nodes α is greater than or equal to 1. Therefore, the closer Pd is to 1, the better the detection performance of FC.

[0148] Figure 4 This is a comparison chart of the ROC curves at the fusion center FC and Eve, and... Figure 2 The simulation conditions are the same, and it can be seen from the figure that when P d=0.2, 0.4, 0.8, the eavesdropping center has almost no detection performance, and the detection performance of the FC increases with the increase of P d , so that the safe detection is realized under the given algorithm, and the eavesdropping attack is effectively resisted.

[0149] Figure 5 When P d =0.6, the ROC curves of Eve and the FC change with the number of sensor nodes N.

[0150] In summary, the present application proposes a method for resisting eavesdropping attack by using malicious attack nodes from the perspective of physical layer signal processing in the sensor network. d The method is based on the premise that the fusion center FC knows the properties of each node, the eavesdropping center does not know, the honest nodes send correct observation information, and the malicious nodes send false observation information. d The weighted fusion is performed at the eavesdropping center and the fusion center FC, the modified bias coefficient is used to measure the detection performance of the FC and Eve, and the safe distributed detection under the complete privacy constraint is realized by setting the local working point of the sensor node, i.e.

[0151] The present application also provides a safe distributed detection device for cognitive wireless sensor network, Figure 6 As shown in the structure schematic diagram of the safe distributed detection device for cognitive wireless sensor network, Figure 6 The device comprises:

[0152] An acquisition unit 601 is configured to acquire the global inspection statistical quantity on the eavesdropping center side The mean value under the assumption H0, the mean value under the assumption H1, and the variance under the assumption H1; wherein the H0 represents that the target detected by the sensor node does not exist; and the H1 represents that the target detected by the sensor node exists.

[0153] A determination unit 602 is configured to determine the modified bias coefficient on the eavesdropping center side according to the mean value under the assumption H0, the mean value under the assumption H1, and the variance under the assumption H1.

[0154] The determination unit 602 is further configured to determine the modified bias coefficient of the fusion center according to the modified bias coefficient on the eavesdropping center side.

[0155] The determining unit 602 is further configured to determine the secure distributed detection method for cognitive wireless sensor networks based on the correction deviation coefficient of the fusion center.

[0156] Figure 7 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 7 As shown, the electronic device may include a processor 710, a communications interface 720, a memory 730, and a communication bus 740. The processor 710, communications interface 720, and memory 730 communicate with each other via the communication bus 740. The processor 710 can call logical instructions in the memory 730 to execute a secure distributed detection method for cognitive wireless sensor networks. This method includes: acquiring global inspection statistics at the eavesdropping center. The mean under assumption H0, the mean under assumption H1, and the variance under assumption H1; where H0 indicates that the target detected by the sensor node does not exist; and H1 indicates that the target detected by the sensor node exists.

[0157] Based on the mean under H0, the mean under H1, and the variance under H1, determine the correction deviation coefficient for the eavesdropping center side;

[0158] The range of values ​​for the correction deviation coefficient of the fusion center is determined based on the correction deviation coefficient of the eavesdropping center side;

[0159] The secure distributed detection method for cognitive wireless sensor networks is determined based on the range of values ​​obtained from the correction deviation coefficient of the fusion center.

[0160] Furthermore, the logical instructions in the aforementioned memory 730 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0161] The device embodiments described above are merely illustrative, wherein the units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, i.e., can be located in one place, or can be distributed to multiple network units. Part or all of the modules can be selected to achieve the purposes of the embodiments according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0162] Through the description of the above embodiments, those skilled in the art can clearly understand that the embodiments can be realized by means of software and the necessary general hardware platform, and of course can also be realized by hardware. Based on such understanding, the above technical solutions can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods described in each embodiment or some parts of the embodiments.

[0163] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement to part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A secure distributed detection method for cognitive wireless sensor networks, characterized in that, The method includes the following steps: Obtain the mean, mean, and variance of the global test statistic at the eavesdropping center under hypothesis H0, hypothesis H1, and hypothesis H1 respectively; wherein H0 indicates that the target detected by the sensor node does not exist; and H1 indicates that the target detected by the sensor node exists. Based on the mean under H0, the mean under H1, and the variance under H1, determine the correction deviation coefficient for the eavesdropping center side; The range of values ​​for the correction deviation coefficient on the fusion center side is determined based on the correction deviation coefficient on the eavesdropping center side. The secure distributed detection method for cognitive wireless sensor networks is determined based on the range of values ​​of the correction deviation coefficient on the fusion center side. The acquisition of the mean, mean, and variance of the global test statistic at the eavesdropping center under hypothesis H0, hypothesis H1, and hypothesis H1 includes: Obtain target measurement data sent by sensor nodes intercepted by the eavesdropping center; Based on the target measurement data, it is determined that the eavesdropping center is in the assumption... H0 under The probability of the distribution, and the sum under the assumption that H1 The probability of the distribution; Based on the aforementioned eavesdropping center, it is assumed that... H0 under The probability of the distribution, and the sum under the assumption that H1 The probability of the distribution under the assumption is determined by the distribution of the target measurement data. H0 The mean and variance under the assumption that, and under the assumption that H1 The mean and variance of the given values; Based on the linear weighted fusion method, according to the target measurement data under the assumption H0 The mean and variance under the assumption that, and under the assumption that H1 The mean and variance of the global test statistic for the eavesdropping center are used to determine the mean under hypothesis H0, the mean under hypothesis H1, and the variance under hypothesis H1, respectively. The mean of the global test statistic at the eavesdropping center under hypothesis H0 is obtained using the following formula: ; Among them, the Representative sensor node i The corresponding weighting coefficients, where M represents the sensor node. i The number of times local testing samples were collected. Representative sensor i exist H The mean of the measured data under the null hypothesis; K represents the number of spoofed nodes; This represents the false alarm probability corresponding to the eavesdropping center; This indicates the proportion of deceptive nodes in the network. It is the first i The average signal-to-noise ratio of the signals received by each sensor node; The mean of the global test statistic at the eavesdropping center, under hypothesis H1, is obtained using the following formula: ; The variance of the global test statistic obtained at the eavesdropping center, under the assumption H1, is obtained using the following formula: ; Among them, the Indicates sensor i exist H 1. The variance of the measurement data under the assumption; Indicates sensor i exist H The mean of the measured data under the assumption of zero; This indicates the detection probability corresponding to the eavesdropping center; The correction deviation coefficient is obtained using the following formula: ; in, T This represents the global detection statistics. This indicates the global detection statistic under the assumption The mean of the following; This indicates the global detection statistic under the assumption The mean of the following, This indicates that the global test statistic is in accordance with the hypothesis. The variance below; The range of values ​​for the correction deviation coefficient of the fusion center, determined based on the correction deviation coefficient at the eavesdropping center, includes: With the correction bias coefficient at the eavesdropping center being 0, determine the proportion of deceitful nodes among all nodes in the network. ; Based on the proportion of deceptive nodes among all nodes Determine the range of values ​​for the correction deviation coefficient of the fusion center; The step of determining the secure distributed detection method for cognitive wireless sensor networks based on the range of the correction deviation coefficient of the fusion center includes: The correction deviation coefficient of the fusion center is obtained using the following formula: ; in, The corrected bias coefficient for the global test statistic of all deceptive nodes at the fusion center; The corrected bias coefficient for the global test statistic of all honest nodes at the fusion center; The step of determining the secure distributed detection method for cognitive wireless sensor networks based on the value range of the correction deviation coefficient of the fusion center includes: Determine the number of sensor nodes corresponding to the maximum value of the correction deviation coefficient at the fusion center; Secure distributed detection is performed on the cognitive wireless sensor network based on the number of sensor nodes.

2. An apparatus for implementing the secure distributed detection method for cognitive wireless sensor networks as described in claim 1, characterized in that, The device includes: The acquisition unit is used to acquire the mean, mean, and variance of the global test statistic at the eavesdropping center under hypothesis H0, hypothesis H1, and hypothesis H1; wherein H0 indicates that the target detected by the sensor node does not exist; and H1 indicates that the target detected by the sensor node exists. The determining unit is used to determine the correction deviation coefficient of the eavesdropping center side based on the mean under H0, the mean under H1, and the variance under H1. The determining unit is further configured to determine the range of values ​​for the correction deviation coefficient on the fusion center side based on the correction deviation coefficient on the eavesdropping center side. The determining unit is further configured to determine the secure distributed detection method for cognitive wireless sensor networks based on the range of the correction deviation coefficient of the fusion center.

3. An electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the secure distributed detection method for cognitive wireless sensor networks as described in claim 1.

Images