An underwater optical communication received signal detection method

Abstract

The application relates to an underwater optical communication receiving signal detection method. The method comprises the following steps: pre-processing a receiving signal disturbed by turbulence by using secondary sampling to obtain a small parameter signal; using a Kalman filtering algorithm to obtain the envelope of the receiving signal disturbed by turbulence by processing the small parameter signal; processing the envelope of the receiving signal disturbed by turbulence according to a signal equalization formula; establishing an adaptive stochastic resonance system model, and taking the processed signal as the input signal of the adaptive stochastic resonance system model; using a multi-strategy fusion particle algorithm to optimize the system parameters in the stochastic resonance system model by taking the system output signal-to-noise ratio as a target function; judging whether the iteration of the multi-strategy fusion particle algorithm is terminated, and if the iteration is terminated, outputting the optimal system parameters corresponding to the maximum value of the target function; and inputting the optimal system parameters into the adaptive stochastic resonance system model. The application can realize adaptive stochastic resonance detection.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of communication signal detection, and particularly relates to an underwater optical communication received signal detection method. BACKGROUND

[0002] With the continuous innovation of ocean resource utilization and the growing demand for the development of precious resources, underwater communication technology has gradually become the focus of academic and industrial research. As a communication technology with fast data rate, low delay, wide frequency band, low power consumption and good privacy, underwater wireless optical communication (UWOC) technology has won widespread attention and attention from the scientific community, the industrial community and even the national defense field. However, the optical signal still faces the challenge of seawater optical properties during transmission, including absorption, scattering and turbulence effect, which will seriously interfere with the communication link and communication quality. Absorption and scattering cause signal attenuation, and ocean turbulence causes light intensity flicker, both of which will make the received signal extremely weak, and the receiving end is difficult to extract useful information from the background noise, thereby affecting the performance of the communication system. In order to realize long-distance communication, the problem of signal detection under low signal-to-noise ratio must be solved. Most of the traditional signal processing methods improve the signal-to-noise ratio by suppressing noise, but it is difficult to obtain good performance under low signal-to-noise ratio.

[0003] In related technologies, in order to reduce the received signal intensity or the light beam deviates from the transmission direction caused by absorption and scattering, an adaptive stochastic resonance is generally used for weak light signal detection. However, in addition to absorption and scattering, ocean turbulence is also an important and unpredictable factor when the optical signal is transmitted underwater, which will cause light intensity flicker, signal attenuation and distortion, thereby causing signal quality degradation and communication efficiency reduction. At this time, the effect of simply using the adaptive stochastic resonance weak light signal detection method is not ideal.

[0004] Therefore, it is necessary to provide a new technical solution to improve one or more problems in the above solutions.

[0005] It should be noted that the information disclosed in the above background section is only used to strengthen the understanding of the background of the present application, and therefore can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0006] The purpose of the present application is to provide an underwater optical communication received signal detection method, which at least partially overcomes one or more problems caused by the limitations and defects of related technologies.

[0007] According to an underwater optical communication received signal detection method provided by an embodiment of the present application, the method comprises:

[0008] The received signal disturbed by the turbulence is preprocessed by using secondary sampling to obtain a small parameter signal;

[0009] The envelope of the received signal disturbed by the turbulence is calculated by using a Kalman filtering algorithm on the small parameter signal;

[0010] The envelope of the received signal disturbed by the turbulence is processed according to a signal equalization formula to obtain a processed signal; wherein the signal equalization formula is as follows:

[0011] The processed signal = 2 x (small parameter signal / envelope of the received signal disturbed by the turbulence) (1);

[0012] An adaptive stochastic resonance system model is established, and the processed signal is used as an input signal of the adaptive stochastic resonance system model;

[0013] The system output signal-to-noise ratio is used as an objective function, and a multi-strategy fusion particle algorithm is used to optimize the system parameters in the stochastic resonance system model;

[0014] It is judged whether the iteration of the multi-strategy fusion particle algorithm is terminated, and if the iteration is terminated, the optimal system parameters corresponding to the maximum value of the objective function are output;

[0015] The optimal system parameters are input into the adaptive stochastic resonance system model, and a fourth-order Runge-Kutta algorithm is used to solve an output detection signal.

[0016] In the embodiment of the application, the envelope of the received signal disturbed by the turbulence is calculated by using a Kalman filtering algorithm on the small parameter signal, comprising:

[0017] The small parameter signal is extracted to obtain a turbulence characteristic signal, comprising:

[0018] The variables R_c and S_m are initialized, and each element in the small parameter signal is traversed, if the current element is greater than S_m, S_m is replaced; otherwise, S_m is updated using (S_m*R_c) / (R_c+1).

[0019] In the embodiment of the application, the envelope of the received signal disturbed by the turbulence is calculated by using a Kalman filtering algorithm on the small parameter signal, further comprising:

[0020] The turbulence characteristic signal is used as the target state at the initial time of the Kalman filtering algorithm, and the filter parameters, the state transition matrix, the noise driving matrix, the observation matrix, the covariance of the noise sequence input into the system, and the covariance of the Gaussian measurement noise sequence are initialized.

[0021] The envelope of the received signal disturbed by the turbulence is calculated by using the Kalman filtering algorithm.

[0022] The predicted value of the target state at the k time, the predicted value of the observation signal and the information are calculated by using the state equation, and the parameter prediction error covariance at the k time, the Kalman filtering gain and the estimation error covariance are calculated, and then the parameter estimation value at the k time is obtained.

[0023] The expression of the state equation is:

[0024] X(k+1)=ΦX(k)+ΓW(k) (2);

[0025] Z(k+1)=HX(k+1)+V(k+1) (3);

[0026] In the formula, X(k+1) represents the target state at the k+1 time, Z(k+1) represents the observation signal at the k+1 time, Φ represents the state transition matrix, Γ represents the noise driving matrix, H represents the observation matrix, in the linear filtering process, it is assumed that the noise is Gaussian white noise, W(k) represents the noise sequence input into the system, g represents the covariance of the noise sequence input into the system, V(k+1) represents the measurement noise sequence, and R represents the covariance of the Gaussian measurement noise sequence.

[0027] The corresponding expressions of the predicted value of the target state at the k time The predicted value of the observation signal and the information ε(k) are respectively:

[0028]

[0029] The corresponding expressions of the parameter prediction error covariance P(k|k-1) at the k time, the Kalman filtering gain K(k) and the state covariance matrix P(k|k) are respectively:

[0030] P(k|k-1)=ΦP(k-1|k-1)Φ T +Q k-1 (7);

[0031] K(k)=P(k|k-1)H T [HP(k|k-1)H T +R] -1 (8);

[0032] P(k|k)=[I n -K(k)H]P(k|k-1) (9);

[0033] In the formula, I n is the unit matrix, and T represents the transpose of the matrix.

[0034] wherein the parameter estimation value at k time The expression is:

[0035]

[0036] In the embodiment of the application, the envelope of the received signal disturbed by the turbulence is calculated by using the Kalman filtering algorithm on the small parameter signal, and the envelope further comprises:

[0037] According to the parameter estimation value at k time, the parameter estimation value at k+1 time is calculated, and when the preset condition k+1 is greater than the length N of the received signal disturbed by the turbulence, the envelope of the received signal disturbed by the turbulence processed by the Kalman filtering algorithm is output.

[0038] In the embodiment of the application, the adaptive stochastic resonance system model is established, and the adaptive stochastic resonance system model comprises:

[0039] The Langevin equation is used to describe the adaptive stochastic resonance system model, and the relationship among the signal, the noise and the nonlinear system is established, and the expression of the Langevin equation is as follows:

[0040]

[0041] In the formula, a and b represent the system parameters of the adaptive stochastic resonance system model, x(t) represents the system output signal of the adaptive stochastic resonance system model, s(t) represents the system input signal of the adaptive stochastic resonance system model, and n(t) represents the noise.

[0042] In the embodiment of the application, the system output signal-to-noise ratio is used as the target function, and the multi-strategy fusion particle algorithm is used to optimize the system parameters in the stochastic resonance system model, and the multi-strategy fusion particle algorithm comprises:

[0043] The parameters of the multi-strategy fusion particle swarm algorithm are initialized, the population number M, the maximum iteration number N itermax , the maximum inertia weight w max , the minimum inertia weight w min , the acceleration coefficient c1, the acceleration coefficient c2, the same number of continuous optimal solutions I max and the value range of the system parameters (a, b) are set;

[0044] The particle population position x i =(x i1 , x i2 ,... x id ) and the speed v i =(v i1 , v i2 ,... v id), i = 1, 2,..., M, judging whether a preset condition is met, and each position corresponds to a set of system parameters, and a target function value is calculated according to the system parameters.

[0045] In the embodiments of the application, the system output signal-to-noise ratio is used as the target function, and a multi-strategy fusion particle algorithm is used to optimize the system parameters in the stochastic resonance system model, and the method further comprises the following steps.

[0046] The fitness of each particle f i is calculated, and the individual optimal P i of each particle is obtained. g ;

[0047] The velocity and position of the particle are updated, comprising:

[0048] A random number P ∈ [0, 1] is generated, and a neighborhood search radius coefficient R k is calculated. k The size relationship between P and R

[0049] When P < R k , a new global optimal position P l is calculated, and the velocity and position formula under this condition is used for updating.

[0050] When P ≥ R k , the relationship between the counter T count and I max is compared, comprising:

[0051] When T count ≥ I max , a selection probability P θ is calculated.

[0052] If a random number R between 0 and 1 is less than P θ , the optimal solution is updated according to the reverse learning strategy formula.

[0053] Otherwise, the optimal solution is disturbed by mutation according to the Cauchy mutation strategy formula to generate a new solution.

[0054] When T count < I max , the velocity and position formula under this condition is used for updating.

[0055] Wherein, the expression of R k is:

[0056]

[0057] In the formula, t represents the current iteration number.

[0058] Wherein, P < Rk P l and the velocity and position update formulas are:

[0059] P l = P g + (-1) t · rand() · r k (13) ;

[0060]

[0061] v i (t+1) = wv i (t) + c1r1(P i (t) - x i (t)) + c2r2(P l - x i (t)) (15) ;

[0062] x i (t+1) = x i (t) + v i (t+1) (16) ;

[0063] wherein r k denotes a search radius, r k = R k · P g ; P g denotes a global optimal position; w denotes an inertia weight; c1 and c2 respectively denote acceleration coefficients for weighting P i (t) and P g (t); r1 and r2 respectively denote two random values in the range of [0, 1]; P i (t) and P g (t) respectively denote an individual optimal position and a global optimal position of the i-th particle in the t-th iteration, and rand() denotes a random number between (0, 1) ;

[0064] wherein P≥ R k and T count ≥ I max correspond to the reverse learning strategy formula:

[0065]

[0066] wherein P g '(t) denotes a reverse solution of the t-th generation optimal solution, x max and x min are respectively upper and lower bounds of the position x, r is a random number matrix subject to a (0, 1) standard uniform distribution, b1 denotes an information exchange control parameter, denotes XOR, and xi (t+1) represents the position of the particle in the t+1 iteration process.

[0067] wherein, P≥R k and T count ≥I max The corresponding Cauchy variation strategy formula is:

[0068]

[0069]

[0070] In the formula, cauchy(0, 1) is a standard Cauchy distribution, η is related to ζ,

[0071] wherein, P≥R k and T count <I max The corresponding velocity and position update formula is:

[0072] v i (t+1)=wv i (t)+c1r1(P i (t)-x i (t))+c2r2(P g (t)-x i (t)) (22);

[0073] x i (t+1)=x i (t)+v i (t+1) (23)。

[0074] In the embodiment of the application, the system output signal-to-noise ratio is used as the target function, and the multi-strategy fusion particle algorithm is used to optimize the system parameters in the stochastic resonance system model, which further comprises:

[0075] The fitness value of each particle is evaluated and updated to obtain the individual historical optimal value and the global optimal value, including:

[0076] The fitness value f i of each particle x i is calculated, and the size relationship between f i and the individual historical optimal value P i is compared, if f i is better, it is assigned to P i .

[0077] Otherwise, no update is performed.

[0078] The global optimal value P g is operated accordingly.

[0079] In the embodiments of the present application, the iteration of the multi-strategy fusion particle algorithm is determined, and if the iteration is terminated, the optimal system parameter corresponding to the maximum value of the target function is output, comprising:

[0080] When the target function converges to its maximum value and the number of iterations reaches the maximum number of iterations, the iteration is stopped, and the optimal system parameter corresponding to the maximum value of the target function is output.

[0081] The technical scheme provided by the embodiments of the present application can include the following beneficial effects:

[0082] In an embodiment of the present application, by using the above method, the Kalman filtering algorithm is used to realize tracking and estimation of the received signal disturbed by the turbulence at the receiving end, to adaptively estimate the channel characteristics and realize signal equalization. Then, the equalized signal is input into the adaptive stochastic resonance system model, the signal-to-noise ratio is taken as the target function, and the improved multi-strategy fusion particle swarm algorithm is used to determine the system parameters of the adaptive stochastic resonance model to realize the detection of the received signal. The embodiments of the present application can realize adaptive stochastic resonance detection, improve the output signal-to-noise ratio, and reduce the system bit error rate, and can be used for detection of the received signal of underwater optical communication.

[0083] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present application. BRIEF DESCRIPTION OF DRAWINGS

[0084] The drawings herein are incorporated into the specification and form part of the specification, show embodiments consistent with the present application, and together with the specification serve to explain the principles of the present application. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0085] Figure 1 A flow chart schematically showing the steps of a method for detecting a received signal of underwater optical communication in an exemplary embodiment of the present application;

[0086] Figure 2 A convergence curve diagram schematically showing the output signal-to-noise ratio using the multi-strategy fusion particle swarm algorithm when the input signal-to-noise ratio is 4dB in an exemplary embodiment of the present application;

[0087] Figure 3 A schematic diagram of an original signal sent by a sending end in an exemplary embodiment of the present application;

[0088] Figure 4 A schematic diagram of a signal after logarithmic intensity variance is (corresponding to the flicker index An overall diagram of the received signal affected by turbulence after passing through a turbulent channel;

[0089] Figure 5 schematically shown Figure 4 The overall image of the received signal subjected to turbulence disturbance after Kalman filtering and signal equalization;

[0090] Figure 6 schematically shown Figure 5 A magnified view of a portion of the image;

[0091] Figure 7 schematically shown Figure 5 A magnified view of the results after bipolar transformation;

[0092] Figure 8 schematically showing the process Figure 7 The result of processing the received signal affected by turbulence using the adaptive stochastic resonance signal detection method based on multi-strategy fusion particle swarm algorithm is shown in the figure.

[0093] Figure 9 schematically showing the process Figure 7 The recovered signal diagram after processing;

[0094] Figure 10 This illustration shows a comparison of the system bit error rate curves for each detection method under different input signal-to-noise ratios in an exemplary embodiment of this application. Detailed Implementation

[0095] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided to make this application more comprehensive and complete, and to fully convey the concept of the exemplary embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0096] Furthermore, the accompanying drawings are merely illustrative of this application and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore, repeated descriptions of them will be omitted.

[0097] The seawater channel is complex, and the optical signal is affected by absorption, scattering and turbulence during transmission, which causes the optical signal power to be severely attenuated. In particular, when the laser beam propagates through seawater, the ocean turbulence caused by the random changes in the optical refractive index due to temperature and salinity causes beam drift, beam jitter, and light intensity flicker, which degrades the quality of the optical signal and severely limits the performance of the UWOC system.

[0098] Based on this, the underwater optical communication receiving signal detection method is provided in the example embodiment. Referring to Figure 1 The method can include steps S101 to S107.

[0099] In step S101, the receiving signal disturbed by the turbulence is preprocessed using quadratic sampling to obtain a small parameter signal.

[0100] In step S102, the envelope of the receiving signal disturbed by the turbulence is calculated using the Kalman filter algorithm.

[0101] In step S103, the envelope of the receiving signal disturbed by the turbulence is processed according to the signal equalization formula to obtain a processed signal; wherein the signal equalization formula is as follows:

[0102] The processed signal = 2 x (small parameter signal / envelope of the receiving signal disturbed by the turbulence) (1).

[0103] In step S104, an adaptive stochastic resonance system model is established, and the processed signal is used as the input signal of the adaptive stochastic resonance system model.

[0104] In step S105, the system output signal-to-noise ratio is used as the objective function, and the multi-strategy fusion particle algorithm is used to optimize the system parameters in the stochastic resonance system model.

[0105] In step S106, it is determined whether the iteration of the multi-strategy fusion particle algorithm is terminated, and if the iteration is terminated, the optimal system parameters corresponding to the maximum value of the objective function are output.

[0106] In step S107, the optimal system parameters are input into the adaptive stochastic resonance system model, and the fourth-order Runge-Kutta algorithm is used to solve the output detection signal.

[0107] In an embodiment of the present application, by using the above method, the Kalman filter algorithm is used to realize the tracking and estimation of the receiving signal disturbed by the turbulence at the receiving end, the channel characteristics are adaptively estimated, and signal equalization is realized. Then, the equalized signal is input into the adaptive stochastic resonance system model, the output signal-to-noise ratio is used as the objective function, the improved multi-strategy fusion particle swarm algorithm is used to determine the system parameters of the adaptive stochastic resonance model, and the receiving signal detection is realized. The adaptive stochastic resonance detection can be realized in the embodiment of the present application, the output signal-to-noise ratio is improved, the system bit error rate is reduced, and the method is used for detecting the underwater optical communication receiving signal.

[0108] In the following, the above method in the example embodiment will be described in more detail. Figure 1 The above method in the example embodiment will be described in more detail.

[0109] In step S101, the received signal disturbed by the turbulence is preprocessed by using the subsampling to obtain a small parameter signal.

[0110] Specifically, the transmission frequency and the sampling frequency of the optical signal in the actual deep sea are large, which cannot meet the adiabatic approximation condition of the stochastic resonance. The received signal disturbed by the turbulence is preprocessed by using the subsampling to obtain a small parameter signal.

[0111] The subsampling is a process of converting a high-frequency signal into a low-frequency signal. By setting a variable scale coefficient, the frequency and time scale of the signal is changed to obtain a small parameter signal. According to the frequency of the signal, R is selected as the variable scale coefficient, f s is the signal sampling frequency, f sr is the variable scale sampling frequency, and f sr =f s / R. After the variable scale preprocessing, the small parameter signal can meet the adiabatic approximation condition of the stochastic resonance.

[0112] In step S102, the envelope of the received signal disturbed by the turbulence is calculated by using the Kalman filter algorithm on the small parameter signal.

[0113] It should be understood that the Kalman filter algorithm tracks the received signal disturbed by the turbulence, and the small parameter signal is used as the input signal of the subsequent adaptive stochastic resonance model. Since the propagation of the optical signal in the sea water is disturbed by the turbulence, the amplitude and phase of the received signal are unknown, and the direct use of the classical Kalman filter algorithm (KF) will introduce a large error. Therefore, the turbulence characteristic signal needs to be extracted first, and then the turbulence characteristic signal is used as the target state at the initial time of the Kalman filter algorithm (KF). The original state information input and the observation data output are constantly used to realize the state estimation and update, so as to calculate the envelope of the received signal disturbed by the turbulence.

[0114] In step S1021, the turbulence characteristic signal is extracted from the small parameter signal, including:

[0115] The variable R_c and the variable S_m are initialized, and each element in the small parameter signal is traversed. If the current element is greater than S_m, S_m is replaced; otherwise, S_m is updated by using (S_m*R_c) / (R_c+1).

[0116] It should be understood that in this process, the variable S_m is constantly updated to achieve an effect similar to the exponential weighted moving average, so that the signal is smoother and the noise is reduced. Specifically, if the current observation value is small, the change speed of S_m will be relatively slow; otherwise, the change speed of S_m will be relatively fast.

[0117] In step S1022, the turbulence characteristic signal is taken as the target state at the initial time of the Kalman filtering algorithm, and the filter parameters, the state transition matrix, the noise driving matrix, the observation matrix, the covariance of the noise sequence input into the system, and the covariance of the Gaussian measurement noise sequence are initialized.

[0118] In step S1023, the predicted value of the target state at time k, the predicted value of the observation signal, and the information are calculated by using the state equation, and the parameter prediction error covariance at time k, the Kalman filtering gain, and the estimation error covariance are calculated, and then the parameter estimation value at time k is obtained.

[0119] The expression of the state equation is:

[0120] X(k+1) = ΦX(k) + ΓW(k) (2);

[0121] Z(k+1) = HX(k+1) + V(k+1) (3);

[0122] In the formula, X(k+1) represents the target state at time k+1, Z(k+1) represents the observation signal at time k+1, Φ represents the state transition matrix, Γ represents the noise driving matrix, H represents the observation matrix, in the linear filtering process, it is assumed that the noise is Gaussian white noise, W(k) represents the noise sequence input into the system, Q represents the covariance of the noise sequence input into the system, V(k+1) represents the measurement noise sequence, and R represents the covariance of the Gaussian measurement noise sequence;

[0123] The predicted value of the target state at time k is The predicted value of the observation signal is And the corresponding expressions of the information ε(k) are respectively:

[0124]

[0125] The corresponding expressions of the parameter prediction error covariance P(k|k-1) at time k, the Kalman filtering gain K(k), and the state covariance matrix P(k|k) are respectively:

[0126] P(k|k-1) = ΦP(k-1|k-1) Φ T + Q k-1 (7);

[0127] K(k) = P(k|k-1) H T [HP(k|k-1) H T + R] 1 (8);

[0128] P(k|k) = [I n -K(k) H]P(k|k-1) (9);

[0129] wherein I n is a unit matrix, T represents the transpose of the matrix;

[0130] wherein the parameter estimation value at the k time is derived according to the projection rule to obtain the recursive equation of the state, and the parameter estimation value at the k time is obtained by using the deviation of the predicted state to correct the state estimation. The expression is:

[0131]

[0132] In step S1024, the parameter estimation value at the k+1 time is calculated according to the parameter estimation value at the k time, and when the preset condition k+1 is greater than the length N of the received signal disturbed by the turbulence, the envelope of the received signal disturbed by the turbulence processed by the Kalman filtering algorithm (KF) is output.

[0133] It should be understood that by continuously inputting the original state information and outputting the observation data, the state estimation and updating are realized, and when the parameter estimation value at the k+1 time is greater than the length N of the received signal disturbed by the turbulence, the loop is stopped, and the final result, i.e. the envelope of the received signal disturbed by the turbulence, is output.

[0134] In step S103, the envelope of the received signal disturbed by the turbulence is processed according to the signal equalization formula to obtain the processed signal; wherein the signal equalization formula is as follows:

[0135] The processed signal = 2 x (small parameter signal / envelope of the received signal disturbed by the turbulence) (1).

[0136] It should be understood that the transmitted signal is a random code sequence with an amplitude of 2.

[0137] In step S104, an adaptive stochastic resonance system model is established, and the processed signal is used as the input signal of the adaptive stochastic resonance system model.

[0138] It should be understood that the Langevin equation is used to describe the adaptive stochastic resonance system model, and the relationship between the signal, the noise and the nonlinear system is established, and the expression of the Langevin equation is as follows:

[0139]

[0140] wherein a and b represent the system parameters of the adaptive stochastic resonance system model, x(t) represents the system output signal of the adaptive stochastic resonance system model, s(t) represents the system input signal of the adaptive stochastic resonance system model, and n(t) represents the noise.

[0141] In step S105, the system output signal-to-noise ratio is taken as a target function, and a multi-strategy fusion particle algorithm is used to optimize the system parameters in the stochastic resonance system model.

[0142] In step S1051, the parameters of the multi-strategy fusion particle swarm algorithm are initialized, the population number M, the maximum iteration number N itermax , the inertia weight w max , the inertia weight w min , the acceleration coefficient c1, the acceleration coefficient c2, the same number of continuous optimal solutions I max , and the value range of the system parameters (a, b) are set.

[0143] The particle population position x i =(x i1 ,x i2 ,...x id ) and the speed v i =(v i1 ,v i2 ,...v id ) are initialized, i=1,2,...,M, whether the preset condition is met is judged, and each position corresponds to a set of system parameters, and the target function value is calculated according to the system parameters.

[0144] It should be understood that M=100, N itermax =100, w max =0.9, w min =0.4, c1=2, c2=2, I max =10, a∈[0,5], b∈[0,5].

[0145] In step S1052, the fitness f i of each particle is calculated, which is taken as the individual optimal P i , and the initial population particle individual optimal is taken as the global optimal P g .

[0146] The speed and position of the particle are updated, including:

[0147] A random number P∈[0,1] is generated, the neighborhood search radius coefficient R k is calculated, and the size relationship between P and R k is compared to realize the corresponding speed and position update.

[0148] When P k , the new global optimal position P l is calculated, and the speed and position formula under this condition is updated;

[0149] When P k , the counter Tcount with I max Relationships include:

[0150] When T count ≥I max When, calculate the selection probability P. θ ;

[0151] If the random number R between (0, 1) is less than P θ The optimal solution is updated according to the reverse learning strategy formula;

[0152] Otherwise, the optimal solution is perturbed by mutation according to the Cauchy mutation strategy formula to generate a new solution;

[0153] When T count <I max At that time, the speed and position formulas under these conditions will be updated accordingly;

[0154] Among them, R k The expression is:

[0155]

[0156] In the formula, t represents the current iteration number;

[0157] Where P < R k P corresponding to time l The formulas for updating velocity and position are as follows:

[0158] P l =P g +(-1) t rand() r k (13);

[0159]

[0160] v i (t+1)=wv i (t)+c1r1(P i (t)-x i (t))+c2r2(P l -x i (t)) (15);

[0161] x i (t+1)=x i (t)+v i (t+1) (16);

[0162] In the formula, r k Represents the search radius, r k =R k ·P g ;Pg Represents the global optimal position; w represents the inertia weight; c1 and c2 represent the acceleration coefficients, and represent the individual learning factor and the social learning factor, respectively, used to adjust P. i (t) and P g (t) weighting; r1 and r2 represent two random values ​​in the range [0,1]; P i (t), P g (t) represents the individual optimal position and the global optimal position of the i-th particle in the t-th iteration, respectively, and rand() represents generating a random number between (0,1);

[0163] Where P≥R k And T count ≥I max The corresponding reverse learning strategy formula is:

[0164]

[0165]

[0166] In the formula, P g '(t) represents the inverse solution of the optimal solution in generation t, x max x min These are the upper and lower bounds of position x, respectively; r is a random number matrix following a standard uniform distribution (0,1); and b1 represents the information exchange control parameters. XOR represents the unique OR operation. i (t+1) represents the position of the particle during the (t+1)th iteration.

[0167] Where P≥R k And T count ≥I max The corresponding Cauchy mutation strategy formula is:

[0168]

[0169]

[0170] In the formula, cauchy(0,1) is the standard Cauchy distribution, and the value of η is related to ζ.

[0171] Where P≥R k And T count <I max The corresponding velocity and position update formulas are:

[0172] v i (t+1)=wv i (t)+c1r1(P i (t)-x i(t))+c2r2(P g (t)-x i (t)) (22);

[0173] x i (t+1)=x i (t)+v i (t+1) (23).

[0174] In step S1053, the fitness value of each particle is evaluated and updated to obtain the individual historical best value and the global best value, including;

[0175] Calculate x for each particle i fitness value f i Compare f i With the individual's historical best value P i The size relationship, if f i It is better to assign it to P. i ;

[0176] Otherwise, no update will be performed;

[0177] For the global optimum P g Perform similar operations.

[0178] In step S106, it is determined whether the multi-strategy fusion particle algorithm iteration has terminated. If the iteration terminates, the optimal system parameters corresponding to the maximum value of the objective function are output.

[0179] It is important to understand that when the objective function converges to its maximum value and the number of iterations reaches the maximum number of iterations, the iteration stops, and the optimal system parameters corresponding to the maximum value of the objective function are output.

[0180] In step S107, the optimal system parameters are input into the adaptive stochastic resonance system model, and the fourth-order Runge-Kutta algorithm is used to solve for the output detection signal.

[0181] It is important to understand that since the Langevin equation does not have an exact analytical expression, the fourth-order Runge-Kutta algorithm is used for simulation and solution. The calculation process is as follows:

[0182]

[0183] In the formula, p n It is the nth sampling point of the system input signal s(t)+n(t), x n It is the nth sampling point of the system output signal x(t), h = 1 / f s f is the sampling step size. s The sampling frequency.

[0184] The following simulation experiment further illustrates this application.

[0185] (1) Simulation parameter settings: The state transition matrix Φ, noise driving matrix Γ, and observation matrix H of the Kalman filter algorithm are all identity matrices. The covariance of the noise sequence W(k) of the input system is Q = 0.0001, and the covariance of the Gaussian measurement noise sequence V(k+1) is R = 0.2. The population size M of the multi-strategy fusion particle swarm algorithm is set to 100, and the maximum number of iterations N is set to... itermax =100, maximum inertia weight w max =0.9, minimum inertia weight w min =0.4, acceleration coefficient c1=2, acceleration coefficient c2=2, consecutive optimal solutions of the same degree I max =10, the search range of the random resonance system parameters is a∈[0,5], b∈[0,5]. The input signal parameters are set as: a random symbol sequence with an amplitude of 2, and 50 symbols are selected for adaptive parameter optimization. 10 5 Error rate statistics are performed on each code element, with 50 sampling points for each code element.

[0186] (2) Simulation details: The proposed adaptive random resonance underwater optical communication receiving signal detection method based on Kalman filtering algorithm is studied under different signal-to-noise ratios.

[0187] Figure 2 This is a convergence curve of the output signal-to-noise ratio (SNR) using a multi-strategy fusion particle swarm optimization algorithm when the input SNR is 4dB in the implementation example of this application. The horizontal axis represents the number of iterations, and the vertical axis represents the output SNR, i.e., the objective function value.

[0188] Figures 3 to 9 This describes the process of processing the received signal results affected by turbulence using the Kalman filter-based adaptive stochastic resonance (KF-MFPSO-SR) underwater optical communication signal detection method when the input signal-to-noise ratio is 4dB, as described in the implementation example of this application. Figure 3 This represents the original signal sent by the transmitting end; Figure 4 This indicates that the logarithmic strength variance is (corresponding flicker index) An overall diagram of the received signal affected by turbulence after passing through a turbulent channel; Figure 5 express Figure 4 The overall image of the received signal subjected to turbulence disturbance after Kalman filtering and signal equalization; Figure 6 express Figure 5 A magnified view of a portion of the image; Figure 7 express Figure 5 A magnified view of the results after bipolar transformation; Figure 8 Indicates the process Figure 7The result of processing the received signal affected by turbulence using the adaptive stochastic resonance signal detection method based on multi-strategy fusion particle swarm algorithm is shown in the figure. Figure 9 Indicates the process Figure 7 The recovered signal obtained after processing.

[0189] Figure 10 This is a comparison chart of the system bit error rate curves for each detection method under different signal-to-noise ratios in the implementation examples of this application. The horizontal axis is SNR. out The vertical axis represents the "input signal-to-noise ratio" (SNR), and the vertical axis represents the "bit error rate" (BER). Under the simulation conditions described above, the BER of each detection method under different inputs and SNRs is compared. Figure 10 Simulation results show that the proposed Kalman filter-based adaptive random resonance underwater optical communication received signal detection method (KF-MFPSO-SR) significantly improves bit error rate performance compared to the two detection methods without random resonance (NO-SR) and adaptive random resonance (MFPSO-SR), and the proposed method has no bit error rate limit.

[0190] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein.

Claims

1. A method for detecting a received signal in an optical underwater communication, characterized in that, The method includes: The received signal subjected to turbulence disturbance is preprocessed by using secondary sampling to obtain a small parameter signal; The envelope of the received signal affected by turbulence is obtained by using the Kalman filter algorithm on the small parameter signal; The step of using the Kalman filter algorithm to obtain the envelope of the received signal affected by turbulence from the small parameter signal includes: Extracting turbulence feature signals from the small parameter signals includes: Initialize variable R_c and variable S_m and iterate over each element in the small parameter signal, if the current element is greater than S_m replace S_m ; otherwise, update S_m using ( S_m * R_c ) / ( R_c +1) The envelope of the received signal subjected to turbulence disturbance is processed according to the signal equalization formula to obtain the processed signal; wherein the signal equalization formula is as follows: The processed signal = 2 × (small parameter signal / envelope of the received signal disturbed by turbulence) (1); An adaptive stochastic resonance system model is established, and the processed signal is used as the input signal of the adaptive stochastic resonance system model. Using the system output signal-to-noise ratio as the objective function, a multi-strategy fusion particle algorithm is used to optimize the system parameters in the stochastic resonance system model; The establishment of the adaptive stochastic resonance system model includes: The Langevin equation is used to describe the adaptive stochastic resonance system model, establishing the relationship between the signal, noise, and nonlinear system. The expression of the Langevin equation is as follows: （11）； wherein a , b denotes a system parameter of the adaptive stochastic resonance system model, denotes a system output signal of the adaptive stochastic resonance system model, denotes a system input signal of the adaptive stochastic resonance system model, denotes a noise; The optimization of system parameters in the stochastic resonance system model using a multi-strategy fusion particle algorithm with the system output signal-to-noise ratio as the objective function includes: Initialize the parameters of the multi-strategy fusion particle algorithm, set the population size , the maximum number of iterations , the inertia weight , the inertia weight , the acceleration coefficient , the acceleration coefficient , the same number of continuous optimal solutions and the value range of system parameters a , b ; and initialize the particle population positions , velocities , , determine whether a preset condition is met, and each position corresponds to a set of system parameters, and a target function value is calculated according to the system parameters; Compute fitness of each particle , as individual best , and set initial population particle individual best as global best ; Update the particle's velocity and position, including: Generating random numbers ∈ [0,1], calculate neighborhood search radius coefficient , and compare and size relationship; When < the new global optimal position is calculated and updated according to the velocity and position formula under this condition; When ≥ the relationship of the comparison counter and is continued, comprising: When ≥ the selection probability is calculated ; If the random number between (0, 1) is less than 0.5, update the optimal solution according to the forward learning strategy formula; otherwise, update the optimal solution according to the reverse learning strategy formula. , update the optimal solution according to the reverse learning strategy formula;​ Otherwise, the optimal solution is perturbed by mutation according to the Cauchy mutation strategy formula to generate a new solution; When < the velocity and position are updated according to the velocity and position formula under this condition. wherein The expression for is: （12）； In the formula, denotes the current iteration number; wherein, corresponding to and the velocity, position update equations are:​ （13）； （14）； （15）； （16）； wherein, r k denotes the search radius, ; denotes the global optimum position; denotes the inertia weight; , denote acceleration coefficients to weight and ; r 1, r 2 denote two random values in the range [0,1]; , denote the individual optimum position of the i th particle in the t th iteration and the global optimum position, denotes the generation of a random number between (0,1). wherein, ≥ and ≥ The corresponding reverse learning strategy formula is: （17）； （18）； （19）； wherein, denotes the i-th inverse solution of the i-th , are the upper and lower bounds of the position , r is a random number matrix subject to (0, 1) standard uniform distribution, denotes the information exchange control parameter, denotes XOR, denotes the i-th t the position of the particle in the i-th iteration process; wherein ≥ and ≥ The corresponding Cauchy variation policy formula is: （20）； （21）； wherein cauchy (0,1) is a standard Cauchy distribution, values and relate to ; wherein ≥ and < The corresponding velocity, position update formula is: （22）； （23）； Evaluate and update the fitness value of each particle to obtain its individual historical best and global best, including: Calculate each particle fitness value ,Compare Compared with the individual's historical best value The size relationship, if Better, assign it to ; Otherwise, no update will be performed; to a global optimum performing the corresponding operation; Determine whether the multi-strategy fusion particle algorithm iteration has terminated. If the iteration terminates, output the optimal system parameters corresponding to the maximum value of the objective function. The step of determining whether the multi-strategy fusion particle algorithm iteration has terminated, and if the iteration has terminated, outputting the optimal system parameters corresponding to the maximum value of the objective function, includes: When the objective function converges to its maximum value and the number of iterations reaches the maximum number of iterations, the iteration stops and the optimal system parameters corresponding to the objective function reaching its maximum value are output. The optimal system parameters are input into the adaptive stochastic resonance system model, and the fourth-order Runge-Kutta algorithm is used to solve for the output detection signal.

2. The method of claim 1, wherein the step of detecting the received signal is performed by a photomultiplier tube. The step of using the Kalman filter algorithm to obtain the envelope of the received signal affected by turbulence from the small parameter signal further includes: The turbulence characteristic signal is used as the target state at the initial moment of the Kalman filter algorithm, and the filter parameters, state transition matrix, noise driving matrix, observation matrix, covariance of the noise sequence of the input system, and covariance of the Gaussian measurement noise sequence are initialized.

3. The method for detecting a received signal of an underwater optical communication according to claim 2, wherein The step of using the Kalman filter algorithm to obtain the envelope of the received signal affected by turbulence from the small parameter signal further includes: The state equation is used to calculate the predicted value of the target state at time t, the predicted value of the observation signal, and the information The predicted value of the target state at time t, the predicted value of the observation signal, and the information The parameter prediction error covariance at time t, the Kalman filter gain, and the estimation error covariance are calculated, and the parameter estimation value at time t is obtained The parameter estimation value at time t is obtained The expression for the state equation is: （2）； （3）； In the formula, represents k the target state at time +1, represents k the observation signal at time +1, represents a state transition matrix, represents a noise driving matrix, represents an observation matrix, in a linear filtering process, it is assumed that all noises are Gaussian white noises, represents a noise sequence input into the system, represents a Gaussian measurement noise sequence; wherein the prediction value of the target state at time k the prediction value of the observation signal and the corresponding expression of the information are respectively （4）； （5）； （6）； where the corresponding expressions for the parameter prediction error covariance , the Kalman filter gain , and the state covariance matrix at time k are given by （7）； （8）； （9）； wherein denotes the covariance of the noise sequence of the input system, denotes the covariance of the Gaussian measurement noise sequence, is the identity matrix, and T denotes the transpose of a matrix. wherein k parameter estimates at the time instant The expression is: （10）。 4. The method for detecting a received signal of an underwater optical communication according to claim 3, wherein The step of using the Kalman filter algorithm to obtain the envelope of the received signal affected by turbulence from the small parameter signal further includes: According to k the parameter estimation value at the time point, the parameter estimation value at the time point is calculated k +1, when a preset condition is satisfied k +1 is greater than the length of the received signal disturbed by the turbulence N , the envelope of the received signal disturbed by the turbulence processed by the Kalman filtering algorithm is output.

Images