A Physical Layer Detection Method for Ultraviolet Spread Spectrum Communication Systems

Abstract

This invention discloses a physical layer detection method for ultraviolet (UV) spread spectrum communication systems. Based on the characteristics of UV communication, corresponding spreading and despreading schemes are designed. The detectability of the system is verified for both eavesdropper models with and without prior knowledge. An experimental platform is designed, using two UV LEDs as a noise source and a signal source at the transmitting end, and a PMT (Programmable Module) at the receiving end to receive the signal. The arrival time interval of the received photons is analyzed to verify the detectability of the communication. This method addresses the problems of incomplete coverage of current covert communication applications and the relatively singular nature of current detection methods. It verifies the detectability of UV communication and provides a powerful supplement to covert communication in different scenarios.

Description

Technical Field

[0001] This invention relates to the field of ultraviolet light communication technology, and in particular to a physical layer detection method for an ultraviolet light spread spectrum communication system. Background Technology

[0002] To address the shortage of spectrum resources in wireless communication bands, research has turned to higher frequency bands, such as millimeter-wave and terahertz communication, which have begun to attract attention. However, with the increase in frequency, extremely stringent requirements are placed on the fabrication of radio frequency antennas and devices. As a possible complement to microwave communication, wireless optical communication technology uses light waves as carriers and has an absolute advantage in bandwidth resources compared to the microwave bands used in traditional wireless communication. Ultraviolet (UV) communication is a scattering communication technology based on ultraviolet light. After the light signal is emitted from the signal source, it can be received by the receiver through various scattering paths due to the scattering effect in the atmosphere, thus enabling the acquisition of light signals transmitted from non-line-of-sight directions. This communication mechanism has significant advantages in terms of resistance to solar interference, high security performance, and non-line-of-sight transmission, making it a promising technology for future communication.

[0003] With the advent of the big data era, massive amounts of private data will be transmitted through wireless communication systems. The openness of wireless channels poses a challenge to the security of this data. Therefore, information security has become a critical issue in wireless networks. Besides other security technologies, covert communication, due to its high level of security, has become a potential solution for wireless network security. In covert communication networks, the transmitter introduces randomness to hide the transmitted signal in environmental or artificial noise to avoid detection by eavesdroppers. By confusing eavesdroppers' judgment about whether a signal is being transmitted, covert communication can maintain information security more robustly than other secure transmission technologies. Due to its excellent security protection performance, covert communication has been successfully applied in various wireless communication scenarios. However, covert communication in ultraviolet light communication remains a research gap, therefore, it is necessary to explore the detectability of this type of communication. Summary of the Invention

[0004] The purpose of this invention is to provide a physical layer detection method based on ultraviolet spread spectrum communication system. This method can solve the problems of incomplete coverage of current covert communication application scenarios and relatively simple detection methods. It verifies the detectability of ultraviolet communication and is a powerful supplement to covert communication in different scenarios.

[0005] The objective of this invention is achieved through the following technical solution:

[0006] A physical layer detection method for an ultraviolet spread spectrum communication system, the method comprising:

[0007] Step 1: Based on the characteristics of ultraviolet light communication, design corresponding spreading and despreading schemes;

[0008] Step 2: For the eavesdropper model with and without prior knowledge, verify the detectability of the system, that is, verify whether the system can be intercepted by an eavesdropper.

[0009] Step 3: Design an experimental platform. At the transmitting end, use two ultraviolet LEDs as noise and signal sources, respectively. At the receiving end, use a PMT to receive the signal. Analyze the time interval between the arrival of the received photons to verify the detectability of the communication.

[0010] As can be seen from the technical solution provided by the present invention, the above method can solve the problems of incomplete coverage of current covert communication application scenarios and relatively simple detection methods. It has verified the detectability of ultraviolet light communication and is a powerful supplement to covert communication in different scenarios. Attached Figure Description

[0011] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a schematic diagram of the physical layer detection method for an ultraviolet spread spectrum communication system provided in an embodiment of the present invention;

[0013] Figure 2 The following is an example of the time slice update process using a quantization method when synchronous search and positioning is used to determine the start position of the time slice in the example of this invention.

[0014] Figure 3 This is a schematic diagram of the bit error rate curve after despreading under double spread spectrum conditions using synchronous correlation peak search, as described in the example of the present invention, and compared with the initial transmission sequence number.

[0015] Figure 4 This is a schematic diagram showing the bit error rate curve after despreading and comparing the initial transmission sequence number under 16x spread spectrum conditions using synchronous correlation peak search.

[0016] Figure 5 This is a schematic diagram of the bit error rate curve comparing the despreading and the initial transmission sequence number under the blind synchronization condition and double spread spectrum condition in the example of the present invention.

[0017] Figure 6 This is a schematic diagram showing the bit error rate curve after despreading under blind synchronization conditions and with the initial transmission sequence number compared to the original sequence number under 16x spread spectrum conditions.

[0018] Figure 7 This is a schematic diagram illustrating the bit error rate analysis of a single experiment in the example given in this invention;

[0019] Figure 8 This is a fitted histogram of the photon arrival time interval at different positions under condition number 01 in an embodiment of the present invention.

[0020] Figure 9 This is a fitted histogram of the photon arrival time interval at different positions under condition number 02 in an embodiment of the present invention;

[0021] Figure 10 This is a schematic diagram of the exponential distribution fitting of the photon arrival time interval under the condition of a bit error rate of 0.2833, as described in the example of this invention.

[0022] Figure 11 This is a schematic diagram of the exponential distribution fitting of the photon arrival time interval under a bit error rate of 0.0500.

[0023] Figure 12 This is a schematic diagram of the exponential distribution fitting of the photon arrival time interval under the condition of 0 bit error rate. Detailed Implementation

[0024] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments, and do not constitute a limitation of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of the present invention.

[0025] like Figure 1 The diagram shown is a schematic flowchart of a physical layer detection method for an ultraviolet spread spectrum communication system provided by an embodiment of the present invention. The method includes:

[0026] Step 1: Based on the characteristics of ultraviolet light communication, design corresponding spreading and despreading schemes;

[0027] In this step, the spreading scheme at the transmitting end uses m-sequences for spreading processing to increase the undetectability of the communication; the specific process of using m-sequences for spreading processing is as follows:

[0028] Suppose the original sequence is X = [x1, x2, ..., x...]. p The m-sequence used for spreading is M = [m1, m2, ..., m]. q The spreading process involves XORing each symbol in the original sequence X with the entire m-sequence to form a new sequence of length k = p × q. This new sequence Y is the synchronization sequence mentioned below;

[0029] The despreading scheme at the receiver is based on the maximum a posteriori probability criterion, according to the decision criterion. To make a judgment; where N i Let i ∈ {1, 2, ..., k} represent the number of photons corresponding to the i-th symbol, satisfying the parameter λ. b Or (λ) b +λ s The Poisson distribution of λ b and λ s The probability formulas for the number of photons arriving per unit time period, corresponding to the transmission of only noise and only signal, are as follows:

[0030]

[0031]

[0032] P(N1, N2, ..., N) k |1) and P(N1, N2, ..., N k |0) is the joint probability density function. Since the number of photons arriving at each symbol is independent, the joint probability formula is: The judgment rules are then simplified, resulting in the following simplified formula for the judgment criteria:

[0033] Where S1 = {k: symbol 1, ..., 1} represents the index of the 1 symbol in the sequence obtained after spreading with m sequence when the initial sequence is 1; S0 = {k: symbol 0, ..., 0} represents the index of the 1 symbol in the sequence obtained after spreading with m sequence when the initial sequence is 0.

[0034] Step 2: For the eavesdropper model with and without prior knowledge, verify the detectability of the system, that is, verify whether the system can be intercepted by an eavesdropper.

[0035] In this step, for the listener model with prior knowledge, that is, the known symbol period, the starting position of the signal is located according to the synchronization correlation peak search and the K-means algorithm under blind synchronization conditions, and despreading is performed according to the despreading scheme designed in step 1.

[0036] In specific implementation, the process of searching and locating the synchronization correlation peak is as follows: quantizing the start time of the sequence within a certain time range, and determining the start position of the synchronization sequence by solving for the peak value of the correlation peak and the peak value of the silhouette coefficient of the K-means algorithm. Specifically:

[0037] To achieve a relatively high transmission rate, the photon arrival time needs to be quantized in the actual receiver circuit design to determine the final photon arrival time, which is the correct photon slice mentioned below. The specific quantization process is as follows: the time length T corresponding to each symbol is divided into Q time slices, and the time length corresponding to each time slice is Q. In the estimation process, it can be assumed that the generation starts from 0. Since the deviation was found to be symmetrical in the preliminary experiments, [-T, T] is ultimately chosen as the estimation interval. t0∈[-T, T] is the arrival time of the entire sequence, i.e., the frame header. Then, the time values ​​from t0 to t0+T all belong to the photon arrival time corresponding to the first symbol, and the number of time values ​​is equal to the number of photons n1. The time values ​​from t0+T to t0+2T all belong to the photon arrival time corresponding to the second symbol, and the number of time values ​​is equal to the number of photons n2; and so on. This allows us to calculate n1, n2, ... n corresponding to the starting point of the time. N ; via t0+nT chip For n∈0,1,…,Q, the starting time value can be updated by sliding to obtain the corresponding photon number ni, such as Figure 2 The diagram illustrates the time slice update process using a quantization method when synchronous search and positioning are used to determine the start position of the time slice in an example of the present invention.

[0038] The synchronization correlation peak method determines the starting position of the synchronization sequence by solving for the peak value of the correlation peak. When using the synchronization correlation peak method, the synchronization sequence Y must first be preprocessed to obtain the preprocessed sequence Y. i The preprocessing process is as follows:

[0039]

[0040] For the synchronous correlation peak method, synchronous estimation is to find the formula... The maximum corresponding start time t0 is also called the synchronization correlation peak estimate for the start arrival time σ of the synchronization sequence.

[0041] The process of the K-means algorithm under the blind synchronization condition is as follows:

[0042] 1) Define different time slices and obtain the number of arriving photons N for each symbol at different time starting points. i ;

[0043] 2) Count the number of photons arriving for each symbol in each experiment, and calculate the average of these photon arrival numbers over multiple experiments to obtain an estimated value for the number of photons arriving for each symbol. This estimate is used to obtain the parameter estimate of the Poisson stochastic process corresponding to each symbol.

[0044] Similarly, we obtain the symbol corresponding to each symbol of the synchronization sequence Y.

[0045] 3) To Cluster analysis is performed, assuming the clusters can be divided into two classes. The clustering effect of the K-means algorithm is evaluated by calculating the silhouette coefficient. A larger silhouette coefficient indicates a better clustering effect. The silhouette coefficient is used to determine the starting position of the synchronization sequence. Specifically:

[0046] The silhouette coefficient is chosen as an explanation and test of the effectiveness of clustering. The following is an explanation of the silhouette coefficient: the silhouette coefficient sil(i) of sample i is calculated based on the intra-cluster dissimilarity a of the samples. i Cluster dissimilarity v i To define:

[0047]

[0048]

[0049] It can be seen that the closer sil(i) is to 1, the better the clustering effect of sample i is, and sample i belongs to this cluster; the closer sil(i) is to -1, the worse the clustering effect of sample i is, and sample i should be assigned to another cluster; the closer sil(i) is to 0, the more likely sample i is to be located on the boundary between two clusters.

[0050] Intra-cluster dissimilarity a of samples i This refers to calculating the average distance 'a' from sample i to all other sample points within the cluster. i ′, for all sample points a within that cluster i The mean value is the intra-cluster dissimilarity 'a' of the cluster. i ; Inter-cluster dissimilarity b of the samples i This refers to calculating the distances from sample i to all other clusters C except the cluster containing sample i. j The distances to all other sample points within the range are then averaged to obtain the average distance b. ij b ij Sample i is referred to as other clusters C j Inter-cluster dissimilarity;

[0051] For k clustering types, the inter-cluster dissimilarity of sample i is:

[0052] b i =min{b i1 b i2 , ..., b ik}

[0053] For intra-cluster dissimilarity a i In other words, a i The smaller the value, the better the clustering effect; for inter-cluster dissimilarity b i In other words, b i The larger the value, the farther the sample is from other clusters, the less likely it is to belong to other clusters, and the better the clustering effect.

[0054] For example, given prior knowledge, i.e., the symbol period is known (let's say the symbol period T = 1), an experiment is conducted using the synchronization correlation peak search method to find the start position of the sequence. Then, based on this, despreading is performed, and the bit error rate is observed. Figure 3 The figure shown is a schematic diagram of the bit error rate curve after despreading under double spread spectrum conditions using synchronous correlation peak search, compared with the initial transmission sequence number in the example of this invention. Figure 4 The figure shows a schematic diagram of the bit error rate curve after despreading under 16x spread spectrum conditions using synchronous correlation peak search, compared with the initial transmission sequence number. Figure 3 and 4 λ s =5, which shows a very low bit error rate and high communication performance.

[0055] Given prior knowledge, i.e., the symbol period is known (assuming the symbol period T = 1), the K-means algorithm is used to conduct experiments to find the starting position of the sequence. Then, based on this, the despreading operation is performed, and the bit error rate is observed. Figure 5 This is a schematic diagram illustrating the bit error rate curves under blind synchronization conditions and double spread spectrum conditions, comparing the despreading sequence number with the initial transmission sequence number in the example given in this invention. Figure 6 The figure shows a schematic diagram of the bit error rate curve after despreading under blind synchronization conditions with 16x spread spectrum and the initial transmission sequence number. Figure 5 and 6 λ s =5, indicating a very low bit error rate and high communication performance. Observing the results of 2x spread spectrum, it can be found that the blind synchronization algorithm performs better than the synchronization correlation peak method.

[0056] In addition, for cases where there is no prior knowledge, i.e., the symbol period is unknown, the KS Test method is used to determine whether a signal is transmitted based on a binary hypothesis. Specifically:

[0057] Based on the two-dimensional mixed Poisson distribution model, the following binary assumptions are made:

[0058] H0: indicates that the sample comes from a mixture of two Poisson distributions;

[0059] H1: Indicates that the sample does not come from a mixture of two Poisson distributions;

[0060] Compare the empirical distribution function F1 and the cumulative distribution function F0 under the null hypothesis, respectively, and the test statistic (KS statistic) is the maximum value of the absolute difference between the two. max F1, F0 and D max The solution process is as follows:

[0061] For F1, the known photon number samples N1, N2, ..., N k ,but in For indicator functions;

[0062] For F0, if we assume H0 to be true, then the number of arriving photons follows a mixed Poisson distribution.

[0063] For D max In other words, it is the maximum difference between F1 and F0, that is:

[0064]

[0065] Then calculate the corresponding confidence level α and compare it with D. max By comparing the results, we can determine whether the signal has been transmitted. The process of calculating α and making the determination is as follows:

[0066]

[0067] in,

[0068]

[0069] if Then we reject the H0 hypothesis at confidence level α.

[0070] Without prior knowledge, i.e., with the symbol period unknown, the KS test method is used for hypothesis testing. The hypothesis distribution is assumed to be a mixture of two Poisson distributions, i.e., a mixture Poisson distribution. In actual experiments, the probability density function is:

[0071] p(n=k)=0.5p(n=k|λ b )+0.5p(n=k|λ b +λ s )

[0072] KS test is performed according to this probability distribution formula.

[0073] The experimental conditions are: assuming the known noise rate is 2 photons propagated per unit, that is, λ b = 2 signals per unit time. The initial signal rate set for the experiment was λ. s=8 frames per unit time, but this rate is unknown at the receiving end. Therefore, the receiving end uses different time periods and different λ values. s Experiments were conducted to compare D under different conditions. max And the critical value under a confidence level of 0.1. If D max If the value is less than the critical value, the null hypothesis is accepted, and the distribution can be considered as a mixture of two Poisson distributions, meaning that signal transmission has been detected; otherwise, the null hypothesis is rejected and the alternative hypothesis is selected, meaning that signal transmission has not been detected.

[0074] Experiments revealed that only λ s When D is near 8, max If the value is less than the critical value, then signal transmission is detected; otherwise, it can be considered that no signal transmission has been detected.

[0075] Step 3: Design an experimental platform. At the transmitting end, use two ultraviolet LEDs as noise and signal sources, respectively. At the receiving end, use a PMT to receive the signal. Analyze the time interval between the arrival of the received photons to verify the detectability of the communication.

[0076] In this step, when designing the experimental platform, the symbol period at the transmitting end can be 10μs, and the oscilloscope can acquire 1000 data points at a time; λ is controlled by changing the deflection angles of the background noise source and the signal source. b and λ s Size, λ b and λ s The number of photons arriving per unit time period corresponds to the number of photons transmitted when only noise is transmitted and the number of photons transmitted when only signal is transmitted; the cases with only background noise and background noise + signal are compared, and the number of received photons is analyzed.

[0077] Simultaneously, the received signal is despread according to the designed despreading scheme;

[0078] Since the sequence captured by the oscilloscope each time is random, a synchronization window of a certain length is defined, and then the window length is evenly divided and quantized; the window is slid within the window according to a certain step size to find the moment when synchronization can be achieved, thereby verifying the detectability of communication.

[0079] For example, choose any set of data measured by an experimental platform, such as... Figure 7 The diagram shown illustrates the bit error rate analysis of a single experiment in this invention. The minimum value is considered to be the bit error rate at the alignment time. The bit error rate performance of several sets of data measured during the experiment is then calculated, and the results are analyzed.

[0080] Analysis 1: Given that the time interval of the Poisson distribution follows an exponential distribution, the time interval of the collected photons can be analyzed. By controlling different positions, i.e., different deflection angles of the transmitter relative to the receiver, as variables, data from three positions at different deflection angles are extracted. For example... Figure 8 The figure shown is a fitted histogram of the photon arrival time interval at different positions under the condition corresponding to serial number 01 in the embodiment of the present invention, as follows: Figure 9 The figure shows the fitted histogram of the photon arrival time interval at different positions under the condition corresponding to serial number 02;

[0081] Table 1

[0082] 01 12.9380 02 0.4640 Position 1 -2.9460 0.2833 Position 1 3.1700 0 Position 2 -2.7140 0.2500 Position 2 3.1580 0 Position 3 -2.5280 0.2667 Position 3 2.1620 0

[0083] Note: λ in Table 1 s This estimation is based on a defined symbol period, assuming that the probability of both 0 and 1 symbols is 0.5. If the total number of received photons and the number of photons in the background noise are known, then λ can be roughly calculated. s .observe Figure 8 It can be observed that when λ s When the signal is very small, the difference between the distribution histogram under transmitted and untransmitted conditions is very small. If despreading is performed at this time, more bit errors will occur.

[0084] Analysis 2: Calculate the fit of the exponential distribution under different bit error rates for both transmitted and untransmitted signals, such as... Figure 10 The diagram shown is a schematic representation of the exponential distribution fitting of the photon arrival time interval under a bit error rate of 0.2833, as illustrated in the example provided in this invention. Figure 11 The figure shown is a schematic diagram of the exponential distribution fitting of the photon arrival time interval under the condition of a bit error rate of 0.0500. Figure 12 The figure shows a schematic diagram of the exponential distribution fitting of the photon arrival time interval under the condition of 0 bit error rate. Figures 10-12 The comparison shows that the lower the bit error rate, the greater the difference in the above distributions.

[0085] The above method provides a verification approach for the detectability performance of ultraviolet spread spectrum communication. It not only proves the limit performance of ultraviolet spread spectrum communication, but also verifies the detectability of communication, which is a strong proof for the realization of covert communication in ultraviolet communication systems.

[0086] It is worth noting that the contents not described in detail in the embodiments of the present invention belong to the prior art known to those skilled in the art.

[0087] Furthermore, those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware, and the corresponding program can be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk.

[0088] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims. The information disclosed in the background section is intended only to enhance the understanding of the overall background technology of the present invention and should not be construed as an admission or implication in any way that such information constitutes prior art known to those skilled in the art.

Claims

1. A physical layer detection method for an ultraviolet spread spectrum communication system, characterized in that, The method includes: Step 1: Based on the characteristics of ultraviolet light communication, design corresponding spreading and despreading schemes; Step 2: For the eavesdropper model with and without prior knowledge, verify the detectability of the system, that is, verify whether the system can be intercepted by an eavesdropper. In step 2, for the case of no prior knowledge, i.e., the unknown symbol period, the KS Test method is used to determine whether to transmit a signal based on binary assumptions. Specifically: Based on the two-dimensional mixed Poisson distribution model, the following binary assumptions are made: : Indicates that the sample comes from a mixture of two Poisson distributions; : Indicates that the sample does not come from a mixture of two Poisson distributions; Solve for the empirical distribution function of the observed values ​​respectively. and cumulative distribution function under the null hypothesis To compare the two, the test statistic is the maximum absolute value of their differences. ; , and The solution process is as follows: for For example, a sample with a known number of photons ,but ,in For indicator functions; for For example, if we assume If this holds true, then the number of arriving photons in the sample follows a mixed Poisson distribution. for In short, it is and The maximum difference between them, that is: ； Then calculate the corresponding confidence level. , and By comparing the data, a determination can be made as to whether a signal has been transmitted. The solution and judgment process are as follows: ； in, ； if Then at the confidence level Refusal Assumption; Step 3: Design an experimental platform. At the transmitting end, use two ultraviolet LEDs as noise and signal sources, respectively. At the receiving end, use a PMT to receive the signal. Analyze the time interval between the arrival of the received photons to verify the detectability of the communication.

2. The physical layer detection method for an ultraviolet spread spectrum communication system according to claim 1, characterized in that, In step 1, the spreading scheme at the transmitting end uses an m-sequence for spreading; the specific process of using an m-sequence for spreading is as follows: Assume the original sequence is The m-sequence used for spread spectrum is The spreading process involves XORing each symbol in the original sequence X with the entire m sequence to form a sequence of length m. new sequence of This new sequence Y is the synchronization sequence mentioned below; The despreading scheme at the receiver is based on the maximum a posteriori probability criterion, according to the decision criterion. To make a judgment; in, Represents the number of photons corresponding to the i-th symbol, satisfying the parameter as follows or The Poisson distribution, and The probability formulas for the number of photons arriving per unit time period, corresponding to the transmission of only noise and only signal, are as follows: ； and Let be the joint probability density function. Since the number of photons arriving at each symbol is independent, the joint probability formula is: The judgment rules are then simplified, resulting in the following simplified formula for the judgment criteria: ; in, This represents the index of the 1 symbol in the sequence obtained after spreading with the m-sequence when the initial sequence is 1; This represents the index of the 1 symbol in the sequence obtained after spreading with the m sequence when the initial sequence is 0.

3. The physical layer detection method for an ultraviolet spread spectrum communication system according to claim 1, characterized in that, In step 2, for the listener model with prior knowledge, that is, the known symbol period, the starting position of the signal is located according to the synchronization correlation peak search and the K-means algorithm under blind synchronization conditions, and despreading is performed according to the despreading scheme designed in step 1.

4. The physical layer detection method for an ultraviolet spread spectrum communication system according to claim 3, characterized in that, The process of searching and locating the synchronization correlation peak is as follows: the start time of the sequence is quantized within a certain time range, and the start position of the synchronization sequence is determined by solving the peak value of the correlation peak and the peak value of the silhouette coefficient of the K-means algorithm. The process of the K-means algorithm under the blind synchronization condition is as follows: 1) Define different time slices and, at different time starting points, obtain the number of arriving photons corresponding to each symbol. ; 2) Count the number of photons arriving for each symbol in each experiment, and calculate the average of these photon arrival numbers over multiple experiments to obtain an estimated value for the number of photons arriving for each symbol. This estimate is used to obtain the parameter estimate of the Poisson stochastic process corresponding to each symbol. ; Similarly, we obtain the symbol corresponding to each symbol of the synchronization sequence Y. ; 3) To Cluster analysis is performed, assuming the clusters can be divided into two classes. The clustering effect of the K-means algorithm is evaluated by calculating the silhouette coefficient. A larger silhouette coefficient indicates a better clustering effect. The silhouette coefficient is used to determine the starting position of the synchronization sequence. Specifically: The silhouette coefficient was chosen as an explanation and test of the effectiveness of clustering. The silhouette coefficient of sample i... Based on the intra-cluster dissimilarity of the samples and inter-cluster dissimilarity To define: ； ； The closer the value is to 1, the better the clustering effect of sample i is, and sample i belongs to that cluster. The closer the value is to -1, the worse the clustering effect of sample i is. At this time, sample i will be assigned to a different cluster. The closer the value is to 0, the more likely sample i is to be located on the boundary between the two clusters; Intra-cluster dissimilarity of samples This refers to calculating the average distance from sample i to all other sample points within the cluster. For all sample points within that cluster The mean value is the intra-cluster dissimilarity of the cluster. ; Inter-cluster dissimilarity of samples This refers to calculating the distances from sample i to all other clusters except the cluster containing sample i. The distances to all other sample points within the range are then averaged to obtain the mean distance. , Sample i is referred to as other clusters Inter-cluster dissimilarity; For k clustering types, the inter-cluster dissimilarity of sample i is: ； For intra-cluster dissimilarity In other words, The smaller the value, the better the clustering effect; for inter-cluster dissimilarity... In other words, The larger the value, the farther the sample is from other clusters, the less likely it is to belong to other clusters, and the better the clustering effect.

5. The physical layer detection method for an ultraviolet spread spectrum communication system according to claim 1, characterized in that, In step 3, the deflection angles of the background noise source and the signal source are changed to control... and Size, and The number of photons arriving per unit time period corresponds to the number of photons transmitted when only noise is transmitted and the number of photons transmitted when only signal is transmitted; the cases with only background noise and background noise + signal are compared, and the number of received photons is analyzed. Simultaneously, the received signal is despread according to the designed despreading scheme; Since the sequence captured by the oscilloscope each time is random, a synchronization window of a certain length is defined, and then the window length is evenly divided and quantized; the window is slid within the window according to a certain step size to find the moment when synchronization can be achieved, thereby verifying the detectability of communication.

Images