A method and device for quantizing a key of an underwater acoustic physical layer

Abstract

The present application belongs to the technical field of underwater acoustic physical layer, and particularly relates to a kind of underwater acoustic physical layer key quantization method and device, to be used in the condition of underwater acoustic channel under high signal-to-noise ratio Quantization and key generation, legal communication parties complete channel sounding in channel sounding stage using OFDM signal, and extract the observation sequence containing channel frequency response characteristics;Communication parties use clustering method to complete clustering to obtain cluster center, then build quantization protection interval through membership discrimination, and remove the fuzzy boundary observation point in the protection interval.The present application can solve the problem of quantization error caused by the inconsistency of observation sequence due to local noise difference under high signal-to-noise ratio, and by using the idea of clustering and protection interval to build irregular quantization protection interval, the quantization boundary point is removed, and under the condition of underwater acoustic environment and / or high signal-to-noise ratio, a lower initial key mismatch rate can be obtained, which is convenient for practical scene application.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of underwater acoustic physical layer security communication, and particularly relates to a method and device for quantizing underwater acoustic physical layer keys. BACKGROUND

[0002] As the main means of underwater wireless information transmission, the importance of underwater acoustic communication is increasingly prominent. However, the inherent broadcast nature of the underwater acoustic channel poses a security risk to the rapidly developing underwater acoustic communication technology. Meanwhile, due to the problems such as the easy loss of underwater acoustic key management facilities and the limited operation resources of the equipment, the key distribution and management technology in the traditional communication security protection mechanism is difficult to apply to the underwater acoustic environment. Physical layer security technology can complete key generation and distribution at the same time by utilizing the inherent characteristics of the channel, and can effectively solve the key distribution problem in underwater acoustic security communication, and is attracting more and more attention. Maurer pointed out that the wireless channels measured by the two parties at adjacent times have reciprocity, randomness and location specificity, and demonstrated the possibility of extracting keys by utilizing the randomness of public channels. However, in practical applications, due to problems such as hardware differences and environmental noise, the reciprocity of the observation sequences obtained by the two parties is impaired, resulting in a decrease in the consistency rate of the key sequences generated by the two parties.

[0003] To solve this problem, a smoothing filter is used to preprocess the observation sequence, and by selecting a suitable filter, the influence of the environment on the reciprocity can be effectively improved. However, filtering reduces the randomness of the observation sequence, and when the environmental noise changes, different filters need to be used for preprocessing to achieve good performance. The feedback-based quantization method utilizes the transmission of auxiliary sequences to enhance the reciprocity of the observation sequences. Through the transmission of auxiliary sequences, the two parties can obtain compensation factors such as amplitude and phase difference between the observation sequences, and then realize compensation quantization to reduce the key mismatch rate. The feedback-based quantization method can greatly improve the key consistency rate, but it often needs to feed back the compensation factors of all observation sequences, which results in a large communication overhead, making it difficult to apply in resource-limited underwater acoustic environments. SUMMARY

[0004] Therefore, the present application provides a method and device for quantizing underwater acoustic physical layer keys, which solves the problem of quantization error caused by the inconsistency of observation sequences due to local noise differences under high signal-to-noise ratio conditions.

[0005] According to the design scheme provided by the present application, a method for quantizing underwater acoustic physical layer keys is provided, which comprises:

[0006] The two parties in the legal underwater acoustic node obtain the observation sequence of the corresponding channel frequency response by sending OFDM pilot signals for channel sounding;

[0007] The observation sequence is clustered to obtain a cluster center, and the membership between each observation point in the observation sequence and the cluster center is evaluated using a membership degree;

[0008] A quantization protection interval is constructed according to the membership between the observation point and the cluster center, and a quantization region of the communication parties is generated, so that a physical layer key is generated by encoding the quantization region.

[0009] As the water acoustic physical layer key quantization method, further, the communication parties obtain an observation sequence corresponding to a channel frequency response by sending an OFDM pilot signal for channel sounding, which includes:

[0010] First, the communication parties perform bidirectional channel sounding within a relevant time and obtain a receiving sequence respectively;

[0011] Then, for the receiving sequence, the communication parties of the legal water acoustic node obtain an observation sequence of the channel frequency response using an OFDM pilot signal according to channel decorrelation, wherein the OFDM pilot signal is a uniformly distributed pseudo-random sequence.

[0012] As the water acoustic physical layer key quantization method, further, the observation sequence is clustered to obtain a cluster center of each quantization bit number, and the membership between each observation point in the observation sequence and the cluster center is evaluated using a membership degree, which includes:

[0013] First, the cluster center corresponding to each quantization bit number in the observation sequence is obtained according to the quantization bit number agreed by the communication parties and using a K-means algorithm;

[0014] Then, a membership degree matrix of the observation point belonging to the corresponding class is calculated using a membership function, so that the membership between the observation point and each cluster center is evaluated according to the corresponding membership degree.

[0015] As the water acoustic physical layer key quantization method, further, the membership function is represented as: Wherein, u ji represents the membership degree of the i th observation point to the j th cluster center, K is the quantization bit number agreed by the communication parties, R A (f i ) is the symbol carried by the i th observation point in the observation sequence received by the communication party A, C j is the j th cluster center, and μ represents a weighted index.

[0016] As the water acoustic physical layer key quantization method, further, a quantization protection interval is constructed according to the membership between the observation point and the cluster center, and a quantization region of the communication parties is generated, which includes:

[0017] Firstly, the membership degree discrimination rule is set according to the maximum membership degree of the observation point, the membership degree discrimination rule is used to judge the membership degree of the observation point and the cluster center one by one, and the discrimination sequence of the corresponding observation point is obtained according to the judgment result;

[0018] Then, one of the two communication parties sends the cluster center and the discrimination sequence to the other party as an auxiliary sequence, and the other party receives the auxiliary sequence, removes the observation point in the quantization boundary area by using the discrimination sequence, and calculates the Euclidean distance between the remaining observation points and each cluster center;

[0019] Then, according to the calculated Euclidean distance between the observation points and each cluster center, each observation point is divided into the corresponding cluster according to the principle of minimum Euclidean distance, so as to generate the quantization area by classifying the observation sequence into each cluster center.

[0020] As the underwater acoustic physical layer key quantization method of the application, further, the membership degree discrimination rule is represented as: v i =sgn(max(u ji )-η), wherein u ji represents the membership degree of the i th observation point to the j th cluster center, sgn() represents the sign operator, η is a preset threshold, and v i represents the membership degree of the i th observation point.

[0021] As the underwater acoustic physical layer key quantization method of the application, further, the membership degree discrimination rule is used to judge the membership degree of the observation point and the cluster center one by one, and the discrimination sequence of the corresponding observation point is obtained according to the judgment result, which further comprises: calculating the membership degree of the observation point according to the membership degree discrimination rule, and judging whether the maximum membership degree of the calculated observation point is less than or equal to the preset threshold, if yes, it is determined that the observation point is located in the quantization boundary area with fuzzy membership relationship, and the membership degree of the corresponding observation point is set to a preset fuzzy membership value.

[0022] As the underwater acoustic physical layer key quantization method of the application, further, when the corresponding observation point is removed by using the discrimination sequence, the corresponding observation point with the membership degree value being the preset fuzzy membership value is removed.

[0023] As the underwater acoustic physical layer key quantization method of the application, further, the calculation process of the Euclidean distance between the observation point and each cluster center is represented as: d(R A (f i ),C k )=(Re(R A (f i ))-Re(C k )) 2 +(Im(R A (f i ))-Im(C k )) 2wherein, R A (f i ) is the observation sequence of the communication party A, f i is the i-th observation point corresponding sub-carrier, C k is the k-th cluster center, Re and Im are the real part and the imaginary part taking operation respectively.

[0024] Further, the application also provides a kind of underwater acoustic physical layer key quantization device, comprising: data acquisition module, data analysis module and data quantization module, wherein,

[0025] Data acquisition module, for the communication parties in legal underwater acoustic node to obtain the observation sequence of corresponding channel frequency response by sending OFDM pilot signal for channel sounding;

[0026] Data analysis module is used to cluster and obtain cluster center to the observation sequence, and the membership relationship between each observation point in observation sequence and cluster center is evaluated using membership;

[0027] Data quantization module is used to construct quantization protection interval according to the membership between observation point and cluster center and generate the quantization area of communication parties, to generate physical layer key by encoding quantization area.

[0028] The beneficial effects of the application are as follows:

[0029] The application aims at the problem that the traditional quantization protection interval scheme in key quantization algorithm is difficult to apply to irregular quantization boundary, and constructs irregular quantization protection interval through clustering algorithm and membership, reduces key mismatch rate, can obtain lower initial key mismatch rate in underwater acoustic environment, and can realize quantization and key generation under the condition of high signal-to-noise ratio of underwater acoustic channel. And further through simulation results, the scheme can effectively improve the key mismatch rate performance on the basis of losing a small amount of key generation rate performance, and is more suitable for practical scene application. BRIEF DESCRIPTION OF DRAWINGS

[0030] Figure 1 It is the underwater acoustic physical layer key quantization process schematic in embodiment;

[0031] Figure 2 It is the underwater acoustic physical layer key quantization algorithm principle schematic in embodiment;

[0032] Figure 3 It is the typical physical layer key generation scene schematic in embodiment;

[0033] Figure 4 It is the simulation environment schematic in embodiment;

[0034] Figure 5 It is the sound speed profile schematic in embodiment;

[0035] Figure 6 Key mismatch rate performance diagram of the method in the embodiment;

[0036] Figure 7 Key generation rate performance diagram of the method in the embodiment. DETAILED DESCRIPTION

[0037] In order to make the objects, technical solutions and advantages of the present application clearer and more apparent, the present application will be further described in detail below with reference to the drawings and technical solutions.

[0038] The traditional physical layer key generation technology utilizes channel time variation, reciprocity and spatial decorrelation to realize key distribution and rapid update. However, the underwater acoustic channel environment is complex, and the noise influence is often strong, which leads to poor consistency of the observation sequences obtained by the two communication parties from channel detection, and further causes the increase of the initial key mismatch rate. Therefore, the embodiment of the present application, as shown in the figure, provides an underwater acoustic physical layer key quantization method, which comprises: Figure 1

[0039] S101, the two communication parties in the legal underwater acoustic node obtain the observation sequence of the corresponding channel frequency response by sending the OFDM pilot signal for channel detection.

[0040] S102, the observation sequence is clustered and the cluster center is obtained, and the membership relationship between each observation point in the observation sequence and the cluster center is evaluated by using the membership degree.

[0041] S103, the quantization protection interval is constructed according to the membership degree between the observation point and the cluster center, and the quantization region of the two communication parties is generated, so as to generate the physical layer key by encoding the quantization region.

[0042] Quantization, as a key step to realize key generation, affects the key generation rate and the key mismatch rate. By designing the quantization protection interval, the irregular quantization boundary and other methods, the quantizer can adapt to various application scenarios to reduce the quantization error and reduce the key mismatch rate. When the observation sequence has good aggregation and can be classified, the quantizer with irregular quantization boundary can also obtain better classification effect, thereby reducing the key mismatch rate. And by deleting the values near the quantization boundary, the quantization protection interval can further reduce the key mismatch rate under high signal-to-noise ratio. When the observation sequence is randomly distributed and the noise is large, it is difficult to realize effective classification of the observation sequence. There are often more observation sequences near the quantization boundary, and the quantization protection interval and the irregular quantization boundary have limited effect. In the present scheme, for the observation sequence containing the channel frequency response characteristics, the two legal communication parties utilize the clustering method to complete clustering and obtain the cluster center, and obtain the membership matrix by membership degree function calculation; and the quantization protection interval is constructed according to the membership relationship, the observation points in the protection interval are removed, and the observation points in the quantization region are mapped to the cluster center to generate the initial key.​

[0043] As a preferred embodiment, further, the communication parties obtain the observation sequence of the channel frequency response by sending OFDM pilot signals for channel sounding, which can be designed to include the following contents:

[0044] First, the communication parties perform bidirectional channel sounding in a relevant time and obtain the receiving sequence respectively;

[0045] Then, for the receiving sequence, according to the channel decorrelation, the communication parties of the legitimate underwater acoustic node obtain the observation sequence of the channel frequency response by using the OFDM pilot signal, wherein the OFDM pilot signal is a uniformly distributed pseudo-random sequence.

[0046] As shown in Figure 3 , Alice and Bob are legitimate underwater acoustic nodes, and can realize time division half duplex communication by using OFDM sequence. Eve is a passive eavesdropping node, which can obtain all public communication information of the network but cannot actively interfere. In the case of good hydrological environment, Alice and Bob can complete bidirectional channel sounding in a coherence time and obtain the receiving sequence respectively:

[0047]

[0048] Wherein, Y A and Y B are the receiving sequence, X is the pilot sequence, N is the number of subcarriers, f i is the i-th subcarrier, H AB and H BA are the channel frequency response and have H AB = H BA = H according to the channel reciprocity, N A and N B are local Gaussian noise sequences. According to the channel decorrelation, when the distance between the eavesdropping party and the legitimate party is greater than half the wavelength, the legitimate channel and the eavesdropping channel are not correlated. At this time, Alice and Bob can obtain the observation sequence of the channel frequency response by using the pilot sequence:

[0049]

[0050] Wherein, R A and R B are the observation sequence, N A ' and N B ' are local Gaussian noise sequences, and the variances are σ A and σ B

[0051] The clustering-based method can divide irregular quantization boundaries according to data distribution, and is suitable for data with random distribution; and the protection interval-based method can specifically process boundary points prone to quantization errors, and effectively reduce the key mismatch rate. In the embodiments of the present case, a selection quantification based membership (SQM) algorithm can be proposed by combining the advantages of the two methods, and the basic process is as shown in Figure 2 The observation sequence is first used to obtain the cluster center by using the K-means algorithm, and then the membership matrix is calculated by using the membership function to evaluate the membership relationship between the observation point and each cluster center. Since the observation points with fuzzy membership are mostly near the quantization boundary. Therefore, in the scheme of the present case, the quantization protection interval can be constructed by judging the relationship between the maximum membership value of the observation point and the preset threshold, and the observation points in the protection interval are the boundary points, and further the observation points in the protection interval are proposed to achieve the purpose of reducing the initial key mismatch rate.

[0052] Further, in the embodiments of the present case, the observation sequence is clustered and the cluster center of each quantization bit is obtained, the membership relationship between each observation point in the observation sequence and the cluster center is evaluated by using the membership, which can be designed to include the following contents:

[0053] First, the cluster center corresponding to each quantization bit in the observation sequence is obtained according to the quantization bit number agreed by the two parties in communication and by using the K-means algorithm;

[0054] Then, the membership matrix of the observation point belonging to the corresponding class is calculated by using the membership function, so as to evaluate the membership relationship between the observation point and each cluster center by using the corresponding membership.

[0055] The division of quantization region is the key of quantization algorithm, and the clustering algorithm can divide irregular quantization region according to the data distribution, and is widely concerned. In the clustering algorithm, the K-means algorithm is a hard division based clustering algorithm, which considers that an observation point only belongs to one cluster center. In the quantization process, the observation points near the boundary often cause classification errors due to environmental noise, thereby generating quantization errors. As a typical soft division clustering algorithm, fuzzy C-means algorithm can realize flexible division of observation sequence. Through the membership function, the algorithm can calculate the membership degree of each observation point to each class. The membership degree is in the interval [0, 1], and the value close to 1 indicates that the object belongs to the class to a higher degree. The fuzzy C-means algorithm can use the membership matrix to represent the relationship between the observation points and the cluster centers. However, in the clustering process, the fuzzy C-means clustering needs to calculate the membership degree of each observation point to each cluster center in each iteration, which has high computational complexity and long running time, and is not suitable for the water acoustic environment with limited computing resources. In the scheme, the advantages of K-means algorithm and fuzzy C-means algorithm are combined to propose the SQM algorithm. First, the K-means algorithm with lower complexity is used to obtain the cluster centers, and then the membership matrix is calculated using the cluster centers to complete the clustering analysis. Subsequently, the quantization protection interval is generated by removing the observation points with fuzzy membership relationship, and the quantization region is divided according to the membership matrix to complete the key quantization.

[0056] As a preferred embodiment, further, in the embodiment, the quantization protection interval is constructed according to the membership between the observation points and the cluster centers, and the quantization regions of the communication parties are generated, which can be designed to include the following contents:

[0057] First, the membership degree discrimination rule is set according to the maximum membership degree of the observation points, the membership degree discrimination rule is used to judge the membership degree of the observation points and the cluster centers one by one, and the judgment sequence of the corresponding observation points is obtained according to the judgment result;

[0058] Then, one of the communication parties sends the cluster centers and the judgment sequence as auxiliary sequences to the other party, and the other party receives the auxiliary sequences and removes the observation points in the quantization boundary region using the judgment sequence and calculates the Euclidean distance between the remaining observation points and each cluster center;

[0059] Then, according to the calculated Euclidean distance between the observation points and each cluster center, each observation point is divided into the corresponding cluster according to the principle of minimum Euclidean distance, so as to generate the quantization region by classifying the observation sequence to each cluster center.

[0060] In the method, the membership degree of each observation point and the clustering center is determined one by one by using a membership degree determination rule, and a determination sequence of the corresponding observation point is obtained according to the determination result, and the method further comprises: calculating the membership degree of the observation point according to the membership degree determination rule, and determining whether the maximum membership degree of the calculated observation point is less than or equal to a preset threshold value, if yes, it is determined that the observation point is located in a quantization boundary region with fuzzy membership relationship, and the membership degree of the corresponding observation point is set to a preset fuzzy membership value. When the corresponding observation point is removed by using the determination sequence, the corresponding observation point with the membership value being the preset fuzzy membership value can be removed.

[0061] Referring to Figure 2 and 3 , the content of the key quantization algorithm can be designed as follows:

[0062] Step 1: Alice and Bob send OFDM pilot signals in turn to complete channel sounding, and obtain N-point observation sequences R A and R B , respectively.

[0063] Step 2: According to the agreed quantization bit number K, Alice uses the K-means algorithm to complete clustering analysis on the observation sequence R A , and obtains clustering centers C1, C2, L, C K . The membership value of each observation point to each clustering center is obtained by using a membership function:

[0064]

[0065] Wherein, u ji represents the membership degree of the i-th observation point to the j-th clustering center, and μ represents a weighted index.

[0066] Step 3: In order to reduce the key mismatch rate, the algorithm deletes the observation point with fuzzy membership by setting a quantization protection interval. Alice determines the observation point one by one by using a membership degree determination formula and obtains a determination sequence V:

[0067] v i = sgn(max(u ji )-η) (4)

[0068] Wherein, sgn() represents a sign operator, and η is a preset threshold value, which can adjust the range of the quantization protection interval. When the maximum membership degree of the observation point R A (f i ) is less than or equal to η, R A (f i ) is located near the quantization boundary, the membership relationship is fuzzy, and v i is set to-1.

[0069] Step 4: Alice sends the cluster centers and the discriminant sequence to Bob as an auxiliary sequence;

[0070] Step 5: After receiving the sequence, Bob sends an acknowledgement frame to Alice. Then, Bob discards part of the observation points using the discriminant sequence, and then calculates the Euclidean distance between the remaining observation points and each cluster center:

[0071] d(R A (f i ),C k )=(Re(R A (f i ))-Re(C k )) 2 +(Im(R A (f i ))-Im(C k )) 2 (5)

[0072] Where Re and Im are the real part and the imaginary part operations, respectively. After obtaining the Euclidean distance between each observation point and each cluster center, Alice divides each observation point into K classes according to the principle of minimum Euclidean distance;

[0073] Step 6: Alice and Bob classify the observation sequence to each cluster center according to the principle of minimum Euclidean distance, generate a quantization region, and complete quantization.

[0074] Subsequently, the communication parties use Gray coding to complete coding on the quantization region, and map the observation points in the quantization region to the cluster centers to generate an initial key.

[0075] When dividing the quantization protection interval, the preset threshold η can determine the size of the quantization protection interval, thereby affecting the initial key mismatch rate and the key generation rate. Since the membership degree represents the probability of an observation point belonging to a certain class, the threshold η has a value range of [1 / K, 1). When η = 1 / K, the protection interval range is the quantization boundary, and only the points on the quantization boundary are removed. When η = 1, all observation points are within the protection interval range, and at this time the quantizer has no output, so η < 1. When the signal-to-noise ratio is high, the quantization protection interval range needs to be as small as possible to avoid affecting the key generation rate. When the signal-to-noise ratio is low, the quantization protection interval range needs to increase as the signal-to-noise ratio decreases, balancing the key generation rate and the key mismatch rate.

[0076] Further, based on the above method, the embodiment of the present application also provides an underwater acoustic physical layer key quantization device, comprising: a data acquisition module, a data analysis module and a data quantization module, wherein,

[0077] A data acquisition module configured to acquire an observation sequence of a corresponding channel frequency response by sending an OFDM pilot signal for channel sounding by two communication parties in a legal underwater acoustic node;

[0078] A data analysis module configured to cluster the observation sequence and obtain a cluster center, and evaluate a membership relationship between each observation point in the observation sequence and the cluster center by using a membership degree;

[0079] A data quantization module configured to construct a quantization guard interval according to the membership degree between the observation point and the cluster center, and generate a quantization region of the two communication parties, so as to generate a physical layer key by encoding the quantization region.

[0080] To verify the effectiveness of the scheme, the following experimental data are further explained and described:

[0081] As shown in Figure 4 , Alice and Bob are transceiver transducer arrangements arranged underwater at a distance of 3000m. It is assumed that the relative positions of Alice and Bob do not change during the channel sounding process and the forward channel and the reverse channel are reciprocal. In the embodiment, the underwater acoustic channel simulation software bellhop is used to generate the channel impulse response. The latitude and longitude coordinates of the two communication parties are set to (115.5°N, 19.5°N), the sound speed profile information and the ocean depth information of the corresponding region are obtained from the Argo database, and the sound speed profile is shown in Figure 5 .

[0082] The performance of the two quantization methods in terms of key mismatch rate, key generation rate and interaction symbol number is discussed respectively. To compare the performance of the algorithm, the detection method still uses the classical point-to-point detection protocol, and is compared with the channel quantization with guardband (CQG), the K-means-based quantization algorithm and the compensation K-means quantization algorithm (CKQ).

[0083] When the quantization bit number is 2, Figure 6 The relationship between the key mismatch rate of the algorithm and the signal-to-noise ratio is shown. As can be seen from the figure, the SQM is obviously superior to the CQG quantization algorithm and the K-means-based quantization algorithm. Compared with the CQG quantization algorithm, the algorithm uses irregular quantization guard intervals, and the boundary division is based on actual data, which is more suitable for randomly distributed observation sequences. Compared with the K-means-based quantization algorithm, the algorithm uses the design of the guard interval, which eliminates the boundary points susceptible to noise interference, effectively improving the initial key mismatch rate.

[0084] Figure 7The relationship between the key generation rate and the signal-to-noise ratio of the algorithm is shown. As can be seen from the figure, since K-means does not remove observation points, its key generation rate is higher; and the CQG algorithm and the SQM algorithm use the design of the quantization protection interval to remove part of the observation points to reduce the initial key mismatch rate, and its key generation rate is lower than that of the K-means quantization algorithm. At a signal-to-noise ratio of 30 dB, compared with the K-means quantization algorithm, the key generation rate of the algorithm is reduced by about 8 bits / probe, but the initial key mismatch rate performance is improved by about an order of magnitude. Compared with the CQG quantization algorithm, the SQM algorithm proposed in the present application has greater improvement in key generation rate and initial key mismatch rate performance. Therefore, the SQM algorithm proposed in the present application balances the initial key mismatch rate and the key generation rate performance of the system, and is more suitable for practical environment.

[0085] For the SQM algorithm proposed in the present application, when the threshold η is close to 1, the algorithm protection interval increases, the initial key mismatch rate performance improves, but the key generation rate performance gradually decreases. When η = 0.9, the initial key mismatch rate performance of the SQM algorithm is optimal, but its key generation rate performance decreases significantly. Therefore, the SQM algorithm in the present application can reasonably select the threshold according to the actual application scenario to balance the algorithm performance and meet the actual needs.

[0086] Unless specifically stated otherwise, the relative steps, numerical expressions, and numerical values of the components and steps set forth in these embodiments do not limit the scope of the present application.

[0087] The various embodiments in the specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other. For the system disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple, and the relevant parts are referred to the method part description.

[0088] The units and method steps of each example described in combination with the embodiments disclosed herein can be realized in electronic hardware, computer software or a combination of both. In order to clearly illustrate the interchangeability of hardware and software, the components and steps of each example are generally described in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation does not exceed the scope of the present application.

[0089] Those skilled in the art can understand that all or part of the steps in the foregoing method can be instructed by programs to the relevant hardware to complete, and the programs can be stored in a computer readable storage medium, such as a read-only memory, a magnetic disk or an optical disk. Alternatively, all or part of the steps of the foregoing embodiments can also be implemented using one or more integrated circuits, and accordingly, each module / unit in the foregoing embodiments can be implemented in the form of hardware or in the form of a software function module. The present application is not limited to any specific form of combination of hardware and software.

[0090] Finally, it should be noted that the above-described embodiments are merely specific implementations of the present application, which are used to illustrate the technical solutions of the present application, rather than limit the same. The protection scope of the present application is not limited thereto. Although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can make modifications or easily think of changes to the technical solutions recorded in the foregoing embodiments, or make equivalent replacements to some of the technical features, within the technical scope disclosed by the present application. Such modifications, changes or replacements do not cause the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A key quantization method for underwater acoustic physical layer, characterized in that, Include: In a legitimate underwater acoustic node, the communicating parties obtain the observation sequence of the corresponding channel frequency response by sending OFDM pilot signals for channel sounding. Cluster the observation sequence and obtain cluster centers. Use membership degree to evaluate the membership relationship between each observation point in the observation sequence and the cluster center. Based on the membership degree between the observation point and the cluster center, a quantization protection interval is constructed and a quantization region for both communicating parties is generated, so as to generate a physical layer key by encoding the quantization region; Specifically, the quantization protection interval is constructed based on the membership degree between the observation point and the cluster center, and the quantization region for both communicating parties is generated, including: First, membership discrimination rules are set based on the maximum membership degree of the observation points. The membership degree of each observation point and the cluster center is determined by the membership degree discrimination rules. The discrimination sequence of the corresponding observation points is obtained based on the judgment results. Next, one of the communicating parties sends the cluster centers and discriminant sequence as an auxiliary sequence to the other party. After receiving the auxiliary sequence, the other party uses the discriminant sequence to remove observation points in the quantization boundary region and calculates the Euclidean distance between the remaining observation points and each cluster center. Then, based on the calculated Euclidean distance between the observation points and each cluster center, each observation point is assigned to its corresponding cluster according to the principle of minimum Euclidean distance, so as to generate quantization regions by classifying the observation sequences to each cluster center.

2. The underwater acoustic physical layer key quantization method according to claim 1, characterized in that, The communicating parties acquire the observation sequence of the corresponding channel frequency response by sending OFDM pilot signals for channel sounding, including: First, the two communicating parties perform bidirectional channel probing within a relevant time period and obtain the received sequence respectively; Then, for the received sequence, based on the channel decorrelation, the two communicating parties of the legitimate underwater acoustic nodes use OFDM pilot signals to obtain the observed sequence of channel frequency response, where the OFDM pilot signals are uniformly distributed pseudo-random sequences.

3. The underwater acoustic physical layer key quantization method according to claim 1, characterized in that, Cluster the observation sequences and obtain cluster centers for each quantization bit. Utilize membership degrees to evaluate the membership relationship between each observation point and its cluster center, including: First, according to the quantization bit depth agreed upon by both communicating parties, the cluster center corresponding to each quantization bit depth in the observation sequence is obtained using the K-means algorithm; Then, the membership function is used to calculate the membership matrix of the observation point to the corresponding class, so as to evaluate the membership relationship between the observation point and each cluster center through the corresponding membership degree.

4. The underwater acoustic physical layer key quantization method according to claim 3, characterized in that, The membership function is expressed as: ,in, Indicates the first The observation point for the first The membership degree of each cluster center, where K is the quantization bit depth agreed upon by both communicating parties. The first observation in the observation sequence received by communication party A The symbols carried by each observation point For the first There are 1 cluster center, where μ represents the weighting index.

5. The underwater acoustic physical layer key quantization method according to claim 1, characterized in that, The membership degree discrimination rule is expressed as: ,in, Indicates the first The observation point for the first Membership degree of each cluster center Represents symbolic operators, v is a preset threshold. i Indicates the first Membership degree of each observation point.

6. The underwater acoustic physical layer key quantization method according to claim 1 or 5, characterized in that, The membership degree discrimination rule is used to determine the membership degree between each observation point and the cluster center. Based on the discrimination result, the discrimination sequence of the corresponding observation point is obtained. It also includes: calculating the membership degree of the observation point according to the membership degree discrimination rule, and determining whether the calculated maximum membership degree of the observation point is less than or equal to a preset threshold. If so, the observation point is determined to be located in the quantization boundary region of fuzzy membership relationship, and the membership degree of the corresponding observation point is set to the preset fuzzy membership relationship value.

7. The underwater acoustic physical layer key quantization method according to claim 6, characterized in that, When using the discriminant sequence to remove corresponding observation points, the corresponding observation points whose membership degree value is the preset fuzzy membership value are removed.

8. The underwater acoustic physical layer key quantization method according to claim 1, characterized in that, The calculation process for the Euclidean distance between the observation point and each cluster center is expressed as follows: ,in, For the observation sequence of communicating party A, For the first Each observation point corresponds to a subcarrier, C k Let Re be the k-th cluster center, and let Re and Im be the operations of taking the real part and imaginary part, respectively.

9. A key quantization device for underwater acoustic physical layer, characterized in that, It includes: a data acquisition module, a data analysis module, and a data quantification module, among which, The data acquisition module is used by the two communicating parties in a legitimate underwater acoustic node to obtain the observation sequence of the corresponding channel frequency response by sending OFDM pilot signals for channel sounding. The data analysis module is used to cluster the observation sequence and obtain the cluster centers, and to evaluate the membership relationship between each observation point in the observation sequence and the cluster center using membership degree. The data quantization module is used to construct a quantization protection interval based on the membership degree between the observation point and the cluster center and to generate quantization regions for both communicating parties, so as to generate physical layer keys by encoding the quantization regions; Specifically, the quantization protection interval is constructed based on the membership degree between the observation point and the cluster center, and the quantization region for both communicating parties is generated, including: First, membership discrimination rules are set based on the maximum membership degree of the observation points. The membership degree of each observation point and the cluster center is determined by the membership degree discrimination rules. The discrimination sequence of the corresponding observation points is obtained based on the judgment results. Next, one of the communicating parties sends the cluster centers and discriminant sequence as an auxiliary sequence to the other party. After receiving the auxiliary sequence, the other party uses the discriminant sequence to remove observation points in the quantization boundary region and calculates the Euclidean distance between the remaining observation points and each cluster center. Then, based on the calculated Euclidean distance between the observation points and each cluster center, each observation point is assigned to its corresponding cluster according to the principle of minimum Euclidean distance, so as to generate quantization regions by classifying the observation sequences to each cluster center.

Images