Joint blind equalization method and device, electronic equipment, medium and program product

Abstract

The invention discloses a joint blind equalization method and device, electronic equipment, a medium and a program product, and solves the problem that the existing blind equalization method is poor in effect in a low signal-to-noise ratio environment by combining an MCMA algorithm improved based on simulated annealing and a DDLMS algorithm of which the step length is adjusted by an inverse hyperbolic sine function. Specifically, the improved MCMA algorithm is utilized to adjust the tap coefficient of the equalizer at the initial stage of communication, preliminary equalization processing is carried out to quickly achieve a stable equalization effect, and then the DDLMS algorithm with a step length optimization mechanism is switched to, so that the steady-state error performance is further improved at the subsequent stage. According to the method, the dependence of traditional adaptive equalization on a training sequence can be avoided, channel resource consumption is effectively reduced, the equalization effect in a complex transmission environment is improved through step length optimization, and the equalization requirement in a low signal-to-noise ratio scene is met.

Description

Technical Field

[0001] The present invention relates to the field of signal processing, and in particular to a joint blind equalization method, device, electronic equipment, medium and program product. Background Art

[0002] RF fingerprint technology uses the inconsistency of device hardware to generate a unique identity, and has broad application prospects in security authentication. However, since RF fingerprint is sensitive to channel characteristics, complex transmission environments can affect its stability, resulting in inaccurate device feature extraction. Adaptive equalization technology is usually used to offset channel interference, reduce bit error rate and improve the accuracy of RF fingerprint extraction. Traditional adaptive equalizers work in training mode and decision-oriented mode, measuring and compensating for channel effects by sending training sequences, but training sequences need to be sent frequently in fast time-varying channels, consuming bandwidth and increasing complexity.

[0003] In contrast, blind equalization does not require a training sequence, but adjusts the weight of the equalizer through the probability and statistical characteristics of the received signal, which can effectively reduce the waste of channel resources and simplify system design. Blind equalization technology is used in wireless communications to estimate and compensate for channel effects despite the lack of prior channel knowledge. Blind equalization technologies in existing technologies, such as CMA (Constant Modulus Algorithm), MCMA (Modified Constant Modulus Algorithm), DDLMS (Decision-Directed Least Mean Squares), etc., have poor equalization effects under low signal-to-noise ratio conditions. Summary of the invention

[0004] The purpose of the present invention is to provide a joint blind equalization method, device, electronic device, medium and program product, which can avoid the dependence of traditional adaptive equalization on training sequences, effectively reduce channel resource consumption, and improve the equalization effect in complex transmission environments through step size optimization, and adapt to the equalization needs in low signal-to-noise ratio scenarios.

[0005] In order to solve the above technical problems, the present invention provides a joint blind equalization method, comprising:

[0006] For the received signal obtained by the receiving end, the tap coefficient of the equalizer is adjusted by using the MCMA algorithm improved based on simulated annealing, so as to perform equalization processing on the received signal according to the current tap coefficient of the equalizer to obtain a first target signal;

[0007] Determining whether the first target signal achieves a preset equalization effect;

[0008] When the first target signal reaches the preset equalization effect, adjusting the step size of the DDLMS algorithm;

[0009] The tap coefficients of the equalizer are adjusted according to the step size of the DDLMS algorithm, so as to perform equalization processing on the first target signal according to the current tap coefficients of the equalizer to obtain a second target signal.

[0010] Optionally, the probability formula for receiving a new solution in the MCMA algorithm improved based on simulated annealing is:

[0011] , where P is the probability of accepting a new solution, E t+1 is the cost function value of the MCMA algorithm at time t+1, E t is the cost function value of the MCMA algorithm at time t, T t is the temperature at time t, T t Negatively correlated with the number of iterations.

[0012] Optionally, determining whether the first target signal achieves a preset equalization effect includes:

[0013] Calculating a first distance from the first target signal to the nearest symbol point in the constellation diagram;

[0014] Determine whether the first distance is less than a preset distance, where the preset distance is the product of a distance between two nearest symbol points in the constellation diagram and a preset parameter, where 0<the preset parameter<1;

[0015] If the first distance is smaller than the preset distance, it is determined that the preset balancing effect is satisfied; otherwise, it is determined that the preset balancing effect is not satisfied.

[0016] Optionally, calculating a first distance from the first target signal to a nearest symbol point in the constellation diagram includes:

[0017] Acquire signals outputted by the equalizer k times in succession, where the signals outputted k times in succession include the first target signal and signals outputted by the equalizer k-1 times in succession before outputting the first target signal, where k is an integer greater than 1;

[0018] Calculate the average distance between the signal outputted k times in succession and the nearest symbol point in the constellation diagram, and use the average distance as the first distance;

[0019] Determining whether the first distance is less than the preset distance includes:

[0020] Determining whether the average distance is less than the preset distance;

[0021] If the average distance is smaller than the preset distance, it is determined that the preset balancing effect is satisfied; otherwise, it is determined that the preset balancing effect is not satisfied.

[0022] Optionally, the average distance of the signal outputted for k consecutive times to the nearest symbol point in the constellation diagram is calculated, including:

[0023] according to Calculate the average distance between the signal output for k consecutive times and the nearest symbol point in the constellation diagram;

[0024] Among them, d MCMASA (n) is the distance from the signal equalized by the MCMA algorithm improved by simulated annealing to the nearest symbol point in the constellation diagram in the nth iteration, is the average distance, and n is the total number of iterations.

[0025] Optionally, determining whether the average distance is less than the preset distance includes:

[0026] According to |W(n) T X(n)-a(n)|<d×g determines whether the average distance is less than the preset distance;

[0027] Among them, |W(n) T X(n)-a(n)| is the average distance, W(n) T is a vector representing the weight values ​​of each tap of the equalizer in the nth iteration, X(n) is a vector consisting of samples input to the equalizer before the start of the nth iteration, a(n) is the original signal sent by the transmitter, d is the distance between the two nearest symbol points in the constellation diagram, and g is the preset parameter.

[0028] Optionally, the tap coefficients of the equalizer are adjusted using an improved MCMA algorithm based on simulated annealing, including:

[0029] use adjusting the tap coefficients of the equalizer;

[0030] Wherein, W(n+1) represents the tap coefficient of the equalizer at the n+1th iteration, W(n) represents the tap coefficient of the equalizer at the nth iteration, is the step size of the MCMA algorithm improved based on simulated annealing, y(n) is the signal output by the equalizer in the nth iteration, e1(n) represents the error between the modulus value of the signal output by the equalizer in the nth iteration and the constant modulus value when executing the MCMA algorithm improved based on simulated annealing, x(n) is the received signal, It is represented by taking the complex conjugate of x(n), and the constant modulus is a constant calculated based on the original signal sent by the transmitting end.

[0031] Optionally, adjusting the tap coefficient of the equalizer according to the step size of the DDLMS algorithm includes:

[0032] use adjusting the tap coefficients of the equalizer;

[0033] Wherein, W(n+1) represents the tap coefficient of the equalizer at the n+1th iteration, W(n) represents the tap coefficient of the equalizer at the nth iteration, is the step size of the DDLMS algorithm, y(n) is the signal output by the equalizer in the nth iteration, e2(n) represents the error between the signal output by the equalizer in the nth iteration and the signal after the decision when executing the DDLMS algorithm, x(n) is the received signal, It is represented by taking the complex conjugate of x(n), and the signal after judgment is a signal obtained after the signal output by the equalizer in the nth iteration is judged by the judge.

[0034] In order to solve the above technical problems, the present invention provides a joint blind equalization device, comprising:

[0035] A first equalization unit is used to adjust the tap coefficient of the equalizer for the received signal obtained by the receiving end by using the modified normal mode blind equalization MCMA algorithm based on simulated annealing, so as to perform equalization processing on the received signal according to the current tap coefficient of the equalizer to obtain a first target signal;

[0036] A judging unit, configured to judge whether the first target signal achieves a preset equalization effect;

[0037] An adjustment unit, configured to adjust the step size of a decision-directed least mean square (DDLMS) algorithm when the first target signal reaches the preset equalization effect;

[0038] The second equalizing unit is used to adjust the tap coefficient of the equalizer according to the step size of the DDLMS algorithm, so as to perform equalization processing on the first target signal according to the current tap coefficient of the equalizer to obtain a second target signal.

[0039] In order to solve the above technical problems, the present invention provides a joint blind equalization electronic device, comprising:

[0040] Memory for storing computer programs;

[0041] The processor is used to implement the steps of the joint blind equalization method as described above when executing a computer program.

[0042] In order to solve the above technical problem, the present invention provides a non-volatile storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the joint blind equalization method as described above are implemented.

[0043] In order to solve the above technical problem, the present invention provides a computer program product, including a computer program / instruction, which implements the steps of the above joint blind equalization method when executed by a processor.

[0044] The present invention provides a joint blind equalization method, device, electronic device, medium and program product, which solves the problem that the existing blind equalization method has poor effect in a low signal-to-noise ratio environment by combining an improved MCMA algorithm based on simulated annealing and a DDLMS algorithm whose step size is adjusted by an inverse hyperbolic sine function. Specifically, in the early stage of communication, the tap coefficients of the equalizer are adjusted using the improved MCMA algorithm, and preliminary equalization processing is performed to quickly achieve a stable equalization effect, and then switch to the DDLMS algorithm with a step size optimization mechanism, so as to further improve the steady-state error performance in subsequent stages. It can avoid the dependence of traditional adaptive equalization on training sequences, effectively reduce channel resource consumption, and improve the equalization effect in complex transmission environments through step size optimization, so as to adapt to the equalization requirements in low signal-to-noise ratio scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0046] Figure 1 A flow chart of a joint blind equalization method provided by the present invention;

[0047] Figure 2 A system flow chart of a joint blind equalization method provided by the present invention;

[0048] Figure 3 A schematic diagram of a joint blind equalization electronic device provided by the present invention. DETAILED DESCRIPTION

[0049] The core of the present invention is to provide a joint blind equalization method, device, electronic device, medium and program product, which can avoid the dependence of traditional adaptive equalization on training sequences, effectively reduce channel resource consumption, and improve the equalization effect in complex transmission environments through step size optimization, so that device feature extraction is more accurate and stable, and adapt to the needs of RF fingerprint extraction in low signal-to-noise ratio scenarios.

[0050] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0051] The earliest proposed blind equalization algorithm is the constant mode blind equalization algorithm (CMA), which uses the characteristic that the signal amplitude remains basically unchanged to enable the equalizer to converge stably, but the CMA algorithm has deficiencies in phase compensation. In order to further improve the phase recovery capability, the modified normal mode blind equalization algorithm (MCMA) was proposed, but it did not show significant improvements in convergence speed and steady-state error. In contrast, the decision-directed least mean square algorithm (DDLMS) performs better in steady-state error, but its correct convergence depends on the full opening of the eye diagram, which limits its application in certain scenarios. However, the equalization effect of the existing equalization methods in the prior art will be significantly affected in the case of low signal-to-noise ratio.

[0052] In order to solve the above technical problems, the present invention provides a joint blind equalization method, comprising:

[0053] S11: for the received signal obtained by the receiving end, use the MCMA algorithm improved based on simulated annealing (i.e., modified normal mode blind equalization algorithm improved based on simulated annealing) to adjust the tap coefficient of the equalizer, so as to perform equalization processing on the received signal according to the current tap coefficient of the equalizer to obtain a first target signal;

[0054] In this step, the joint blind equalization method adjusts the tap coefficients of the equalizer by using the MCMA algorithm improved based on simulated annealing to effectively compensate for the channel influence during the equalization processing of the received signal. Simulated annealing is an optimization algorithm that draws on the process in which the thermodynamic system gradually reaches a low energy state during the physical annealing process. It allows a certain amount of de-disturbance in the early stage to avoid falling into the local optimal solution, and gradually reduces the probability of disturbance, and finally finds the global optimal solution. The MCMA algorithm uses this feature and applies the simulated annealing algorithm to the tap coefficient adjustment process of the equalizer. By gradually optimizing the equalization effect of the received signal, the equalizer can better cope with complex channel conditions, especially in low signal-to-noise ratio and dynamic transmission environment. Maintain stable equalization performance. The MCMA algorithm itself is an improvement on the traditional CMA. It mainly adjusts the tap coefficients of the equalizer through the characteristic of constant amplitude, so that the equalizer can more accurately identify the constant mode characteristics of the signal amplitude. After the introduction of the simulated annealing algorithm, the equalizer can avoid falling into the local minimum during the adjustment process, thereby further improving the equalization effect, making the received signal tend to the preset target characteristics, and finally obtaining a clearer and more stable first target signal.

[0055] S12: Determine whether the first target signal achieves a preset equalization effect;

[0056] In this step, the initial success of the equalization processing is evaluated by judging whether the first target signal has achieved the preset equalization effect. The core principle of this step is that after the received signal is equalized, it is first necessary to confirm whether the quality of the signal meets the system requirements, such as whether the equalization error of the signal has been reduced to an acceptable range. Specifically, the first target signal is obtained by adjusting the tap coefficients of the equalizer, and the preset equalization effect can be a specific error tolerance, such as the equalization error is lower than a certain set value, or the phase, amplitude and other parameters of the signal reach a certain stability or accuracy. If the equalization effect of the first target signal has reached the preset standard after adjustment by the MCMA algorithm, it means that the preliminary equalization has been completed and the next stage of processing can be entered; otherwise, the equalization process continues to optimize until the required signal quality is achieved.

[0057] In this process, by dynamically evaluating the equalization error of the signal, it is ensured that the quality of the signal can meet the requirements of further optimization in subsequent steps, thereby providing a higher quality input signal for the next stage of signal optimization using the DDLMS algorithm.

[0058] S13: When the first target signal reaches a preset equalization effect, adjusting the step size of the DDLMS algorithm;

[0059] In this step, the step size of the DDLMS (Decision-Directed Least Mean Squares) algorithm is adjusted to further optimize the adjustment efficiency and signal quality of the equalizer (specifically, but not limited to, adjusting the step size of the DDLMS algorithm through an inverse hyperbolic sine function). The principle of this step is that the step size of the equalization algorithm has a significant impact on the equalization effect and convergence speed: a larger step size can accelerate the convergence of the equalizer, but may cause an increase in the steady-state error, while a smaller step size can reduce the steady-state error, but the convergence speed will slow down. In an optional embodiment, when the inverse hyperbolic sine function is used to adjust the step size, the needs of both can be taken into account in the adjustment of the step size. Through its gradually decreasing characteristic, the step size is larger at the beginning of the equalization to quickly adjust the error of the received signal and gradually approach the ideal equalization state; as the equalization process progresses, when the error gradually converges to near the ideal value, the step size gradually decreases to maintain a higher accuracy in the steady-state stage and reduce the residual error. This nonlinear step size adjustment method of the inverse hyperbolic sine function can adaptively change the step size, allowing the DDLMS algorithm to dynamically balance the convergence speed and steady-state error during the entire equalization process, effectively improving the equalization accuracy and system stability. Ultimately, through this step size adjustment mechanism, it is ensured that the equalizer can output signals with higher accuracy, providing optimized signal quality for subsequent signal processing.

[0060] S14: adjusting the tap coefficients of the equalizer according to the step size of the DDLMS algorithm, so as to perform equalization processing on the first target signal according to the current tap coefficients of the equalizer to obtain a second target signal.

[0061] In this step, based on the DDLMS algorithm step size adjusted in the previous step, the tap coefficient of the equalizer is optimized again to further improve the accuracy of the equalization processing and finally obtain the second target signal. The principle of this step is that after the initial equalization (MCMA algorithm) and the step size adjustment optimization (S13), the DDLMS algorithm uses the adjusted step size to perform fine equalization correction on the signal. That is, the tap coefficient of the equalizer is gradually optimized according to the step size adjusted by the DDLMS algorithm to minimize the error between the received signal and the original signal. The DDLMS algorithm fine-tunes the tap coefficient of the equalizer by feeding back the error information of the signal in real time, so that the error of the output signal is more minimized. The key in this process is that the DDLMS algorithm can use the flexible adjustment of the step size in the process of adjusting the equalizer, and further compensate for the noise and distortion in combination with the characteristics of the received signal, thereby improving the accuracy of the details of the equalization. Finally, through the precise adjustment of the DDLMS algorithm, the equalizer can output a second target signal that is closer to the ideal signal (that is, the original signal sent by the transmitter).

[0062] It is important to understand that the step size controls how the tap coefficients of the equalizer are adjusted in each iteration. In the DDLMS algorithm, the step size determines the size of the step when the tap coefficients are updated each time. If the step size is large, the update will be more drastic, which may accelerate convergence, but may also cause instability; if the step size is small, the update will be more gradual, and the convergence process will be more stable, but it may be slower.

[0063] In summary, the joint blind equalization method provided by the present invention uses the improved MCMA algorithm to adjust the tap coefficient of the equalizer in the early stage of communication, performs preliminary equalization processing, so as to quickly achieve a stable equalization effect, and then switches to the DDLMS algorithm with a step size optimization mechanism, so as to further improve the steady-state error performance in the subsequent stage. It can avoid the dependence of traditional adaptive equalization on training sequences, effectively reduce channel resource consumption, and improve the equalization effect in complex transmission environments through step size optimization, so that the device feature extraction is more accurate and stable, and meets the needs of RF fingerprint extraction in low signal-to-noise ratio scenarios.

[0064] Based on the above embodiments:

[0065] As a preferred embodiment, the probability formula for receiving a new solution in the MCMA algorithm improved based on simulated annealing is: , where P is the probability of accepting a new solution, E t+1 is the cost function value of the MCMA algorithm at time t+1, E t is the cost function value of the MCMA algorithm at time t, T t is the temperature at time t, T t Negatively correlated with the number of iterations.

[0066] In this preferred embodiment, the MCMA (Modified Constant Modulus Algorithm) algorithm based on simulated annealing sets a new solution acceptance probability formula to balance the algorithm's exploration and convergence, thereby improving the adaptability and convergence performance of the equalizer in a complex channel environment. The core idea of ​​this formula is to introduce a simulated annealing process to dynamically control the solution acceptance probability at the time of the MCMA algorithm update. Specifically, the probability formula is set as follows: When the new cost function value E t+1 ≥E t When , the new solution is accepted with a certain probability, where T t represents the current "temperature" parameter. As the number of iterations increases, the temperature gradually decreases, making it easier for the algorithm to accept poor solutions in the early stages, thus avoiding falling into local minima. t+1 <E t , the new solution has a better cost function value, so we directly set P = 1, that is, always accept the new solution. In the traditional method, if E t+1 ≥Et , then P = 0, that is, new solutions are rejected, which will limit the exploration space of the algorithm and make it easy to fall into the local optimum.

[0067] The introduction of this simulated annealing method allows for the acceptance of imperfect solutions at higher temperatures (i.e., early iteration stages), thereby increasing the possibility of the algorithm jumping out of the local optimal solution and enhancing its global search capability; as the temperature gradually decreases, the algorithm will be more inclined to select a better solution to accelerate convergence. Therefore, the improved MCMA algorithm balances the search range and convergence speed through dynamic acceptance probability, making the equalizer more adaptable to dynamic changes in the channel.

[0068] As a preferred embodiment, determining whether the first target signal achieves a preset equalization effect includes: calculating a first distance from the first target signal to the nearest symbol point in the constellation diagram; determining whether the first distance is less than a preset distance, the preset distance being the product of the distance between the two nearest symbol points in the constellation diagram and a preset parameter, 0<preset parameter<1; if the first distance is less than the preset distance, determining that the preset equalization effect is met, otherwise determining that the preset equalization effect is not met.

[0069] In this embodiment, the specific method of judging whether the first target signal achieves the preset equalization effect may be to calculate the distance between the signal and the nearest symbol point on the constellation diagram, and to determine whether the equalization effect meets the set requirements by judging the distance. Specifically, the first target signal is the signal output by the equalizer, and the constellation diagram contains a set of predefined symbol points, each of which corresponds to a possible transmission symbol. First, the distance between the first target signal and the symbol point closest to it in the constellation diagram is calculated (referred to as the "first distance"). The core of this step is to quantitatively measure the degree of deviation of the received signal from the ideal position by calculating the geometric distance. Then, the equalization effect is judged by comparing this distance with the preset distance (the product of the distance between the two nearest symbol points in the constellation diagram and the preset parameter, where the preferred value of the preset parameter can be but is not limited to 0.5). The distance between two symbol points on the constellation diagram reflects the minimum distinction between different symbols. The product of this distance and the preset parameter is used as a threshold because when the distance between the signal point and the nearest symbol point is less than this value, it means that the signal falls within the "tolerance area" of the symbol point, and it can be considered that the received signal basically achieves the preset equalization effect. This method provides an objective standard for measuring the quality of equalization by judging whether the signal deviation is within the allowable range.

[0070] As a preferred embodiment, calculating the first distance from the first target signal to the nearest symbol point in the constellation diagram includes: obtaining a signal output by the equalizer k times in succession, the signal output k times in succession includes the first target signal and a signal output by the equalizer k-1 times in succession before outputting the first target signal, where k is an integer greater than 1; calculating the average distance of the signal output k times in succession to the nearest symbol point in the constellation diagram, and taking the average distance as the first distance; judging whether the first distance is less than a preset distance, including: judging whether the average distance is less than the preset distance; if the average distance is less than the preset distance, determining that the preset equalization effect is met, otherwise determining that the preset equalization effect is not met.

[0071] As a preferred embodiment, calculating the average distance of the signal outputted k times continuously to the nearest symbol point in the constellation diagram comprises: Calculate the average distance between the signal output for k consecutive times and the nearest symbol point in the constellation diagram;

[0072] Among them, d MCMASA (n) is the distance from the signal equalized by the MCMA algorithm improved by simulated annealing to the nearest symbol point in the constellation diagram in the nth iteration, is the average distance, and n is the total number of iterations.

[0073] In a further preferred embodiment, the principle of calculating the first distance from the first target signal to the nearest symbol point in the constellation diagram adopts the average distance of the continuous output of the signal to improve the stability and accuracy of the judgment. Specifically, the method first obtains the signal output by the equalizer for k consecutive times, including the first target signal and the signal output for the first k-1 consecutive times. The introduction of the continuous signal sequence is to smooth the noise or error caused by accidental factors by accumulating the performance of multiple signals, and then obtain a more representative balanced state. Then, the average distance between the distances between the continuous k output signals and the nearest symbol point in the constellation diagram is calculated, and this average distance is used as a measure of the equalization effect. Since a single output signal may be affected by sudden noise or interference during the communication process, the interference of these factors can be reduced by introducing the distance mean of multiple continuous signals. Finally, it is determined whether the preset equalization effect is met by judging whether this average distance is less than half of the distance between the two nearest symbol points in the constellation diagram. Such an improvement makes the judgment of the equalization effect more reliable because it eliminates occasional errors and ensures that the judgment of the equalization effect is based on the overall performance over a period of time, rather than the deviation at a single moment.

[0074] As a preferred embodiment, judging whether the average distance is less than a preset distance includes: according to |W(n) T X(n)-a(n)|<d×g determines whether the average distance is less than the preset distance; where |W(n) T|X(n) - a(n)| is the average distance, and W(n) T represents the vector composed of the weight values of each tap of the equalizer in the nth iteration. X(n) is the vector composed of the samples input to the equalizer before the nth iteration starts, a(n) is the original signal sent by the transmitting end, d is the distance between the two closest symbol points in the constellation diagram, and g is a preset parameter.

[0075] In this embodiment, by judging the formula |W(n) T X(n) - a(n)| < d / 2 to determine whether the average distance is less than half of the distance between the two closest symbol points in the constellation diagram, and this is used as the judgment criterion for the equalization effect. Specifically, first, W(n) T represents the vector composed of the weight values of each tap of the equalizer in the nth iteration. This matrix is iteratively adjusted by the algorithm during the equalization process, aiming to compensate for the distortion effect suffered by the signal during transmission in the channel, thereby optimizing the quality of the received signal. Here, X(n) represents the matrix composed of multiple consecutive received signals, usually composed of the signal values observed by the receiving end within a period of time. These signal values are gradually adjusted during the equalization process to make them closer to the expected ideal signal.

[0076] a(n) in the formula represents the original signal sent by the transmitting end, which is an ideal and undisturbed reference signal. During the judgment process, the system will compare the distance deviation between the equalized signal and the original symbol signal, that is, |W(n) T X(n) - a(n)|. The purpose of this calculation is to obtain the deviation between the current signal and the target symbol point. The absolute value operation here ensures that the obtained is the magnitude of the signal point offset, regardless of the direction of the offset.

[0077] Finally, the comparison relationship between |W(n) T X(n) - a(n)| and d / 2 is the core judgment criterion of this method, where d represents the distance between two adjacent symbol points in the constellation diagram. Since the distance between adjacent symbol points on the constellation diagram represents the discrimination degree of different symbols, taking half of the distance (i.e., d / 2) as the threshold can effectively distinguish whether the received signal is close enough to the target symbol point. When |W(n) T X(n) - a(n)| < d / 2, it can be considered that the received signal has achieved the expected equalization effect, that is, it is close enough to the ideal symbol point and falls within the "tolerance region" of the symbol point; if it is greater than d / 2, it means that the signal deviation is too large and the equalization effect is not ideal. Therefore, this formula combines factors such as the received signal matrix, equalization weight, and symbol point spacing in the constellation diagram, and through a reasonable mathematical relationship, effectively judges whether the current equalization state meets the preset equalization effect.

[0078] As a preferred embodiment, the tap coefficient of the equalizer is adjusted by using the MCMA algorithm improved by simulated annealing, including: using Adjusting the tap coefficients of the equalizer; adjusting the tap coefficients of the equalizer according to the step size of the DDLMS algorithm, including: using Adjust the tap coefficients of the equalizer; wherein W(n+1) represents the tap coefficients of the equalizer for the n+1th iteration, and W(n) represents the tap coefficients of the equalizer for the nth iteration. is the step size of the MCMA algorithm improved based on simulated annealing, is the step size of the DDLMS algorithm, y(n) is the signal output by the equalizer in the nth iteration, e1(n) represents the error between the modulus value of the signal output by the equalizer in the nth iteration and the constant modulus value when executing the MCMA algorithm improved by simulated annealing, e2(n) represents the error between the signal output by the equalizer in the nth iteration and the signal after the decision when executing the DDLMS algorithm, x(n) is the received signal, It is expressed as the complex conjugate of x(n), the constant modulus is a constant calculated based on the original signal sent by the transmitter (it should be understood that the constant modulus usually represents the expected modulus or constant amplitude of the original signal sent by the transmitter), and the decided signal is the signal output by the equalizer in the nth iteration after being decided by the decider.

[0079] First of all, it should be noted that the signal after judgment is the result of the judgment step output in the DDLMS algorithm. In the DDLMS algorithm, the signal after equalization must be judged by the judge, and the error e2(n) is calculated using the signal after judgment and the signal after equalization, and the tap coefficient is updated using the error e2(n). The judgment process is to select the symbol of the closest ideal signal. It should be understood that in this application, the method of calculating the errors e1(n) and e2(n) is the same as the method of calculating the errors in the traditional DDLMS algorithm and MCMA algorithm, and this application will not repeat them here.

[0080] In this embodiment, the tap coefficients of the equalizer are adjusted by the MCMA algorithm improved based on simulated annealing, with the purpose of effectively correcting the signal error in the initial stage of equalization to quickly converge to a more ideal equilibrium state. When it is detected that the average distance of the received signal is greater than or equal to the set threshold, that is, half of the distance between the symbol points in the constellation diagram, the algorithm will execute the update of the tap parameters of the equalizer to reduce the error and improve the equalization effect. At this time, the MCMA algorithm adjustment method based on simulated annealing is adopted. This method controls the step size of the weight update through the simulated annealing mechanism of gradual cooling, so that the tap coefficients of the equalizer have a larger adjustment range at higher errors, and as the error gradually decreases, the step size also decreases to avoid the divergence risk caused by over-adjustment. In this way, the equalizer can effectively cope with high noise and strong interference conditions, so that the output signal gradually tends to the transmitted signal, ensuring a relatively stable convergence effect in the initial stage, and laying the foundation for subsequent refinement adjustment.

[0081] When the equalization process enters the later stage and the average distance is less than half of the symbol point spacing of the constellation diagram, the algorithm switches to the DDLMS-based update mode. This conversion is to further refine the adjustment of the tap coefficients when approaching the ideal equalization state to ensure the high accuracy of the final equalization result. The characteristic of the DDLMS algorithm is that it realizes gradual fine-tuning through decision-oriented error feedback, so that the output signal is more accurately aligned with the transmitted signal, especially with excellent stability and accuracy under low error conditions. At this stage, the step size of the update rule is smaller, which can more delicately control the tap coefficient adjustment of the equalizer and avoid distortion problems caused by large fluctuations. In addition, the DDLMS algorithm relies on gradually decreasing error feedback to adaptively converge to a steady-state equalization result, effectively ensuring the quality of the output signal. By dynamically switching between the MCMA and DDLMS algorithms under different error conditions, not only the convergence speed of the equalizer is improved, but also the equalization accuracy is significantly improved, so that the system can still stably output accurate signals under complex channel conditions.

[0082] In a specific embodiment, Figure 2 As shown, the implementation process is as follows:

[0083] Initialize the tap coefficient w(1), set the step size parameter μ1 of the MCMA algorithm improved by simulated annealing, and the initial temperature T first , final temperature T final , attenuation parameter λ, number of iterations M at the same temperature; set the step size control parameters a, b, c of the adaptive step size; set the distance d between the two nearest points in the constellation diagram of the transmitter signal.

[0084] The received signal at the receiving end is equalized using the MCMA algorithm (MCMASA) improved based on simulated annealing to achieve pre-convergence; specifically, the equalizer received signal of the nth iteration is x(n), the tap coefficient is w(n), N is the length of the equalizer, and the output signal y(n) of the equalizer can be calculated according to the following formula: .

[0085] The cost function J(n) in the MCMA algorithm is used as the objective function E of the simulated annealing algorithm at the current moment t , to obtain the global optimal equalizer tap coefficient, where the cost function J(n) can be calculated according to the following formula: , where J(n) is the cost function, y R (n) is the real part statistic of y(n), y I (n) is the imaginary part statistic of y(n), R R is the real part statistic of the original signal, R I is the imaginary part statistic of the original signal.

[0086] The Metropolis acceptance criterion is used in the simulated annealing algorithm to determine whether to accept a new solution. The specific acceptance probability is calculated according to the following formula:

[0087] ; where T t is the temperature at the current moment. The temperature decays with the number of iterations. The number of decays is determined by the parameter control, The value range is 0 to 1, and the calculation formula is as follows: ;

[0088] The pseudo code of improving the MCMA algorithm using simulated annealing is as follows:

[0089] Input: The target received signal sequence X(n), iteration step , blind equalization tap coefficient length N, initial temperature T first , attenuation parameter , final temperature T final , the number of iterations M at the same temperature.

[0090] Return: signal after blind equalization

[0091] Let , k=N

[0092] While T < T final do

[0093] Let j = 0;

[0094] while k≤length{x(n)} and j<M do

[0095] calculate the real and imaginary part of signal X R and X I

[0096] calculate y(n)

[0097] calculate R R and R I

[0098] calculate J(n) as the objective function

[0099] calculate e1(n)

[0100] calculate W(n+1)

[0101] if J n-1 >J n

[0102] W=W(n+1)

[0103] Else

[0104] generate a random number

[0105] calculate P

[0106] if

[0107] W=W(n+1)

[0108] Else

[0109] W = W(n)

[0110] k = k + 1

[0111] j = j + 1

[0112] end while

[0113] update the value of T

[0114] Secondly, according to the equalization results of the MCMA algorithm improved by simulated annealing, the average distance from the signal to the nearest symbol point in the constellation diagram after k consecutive equalizations is calculated. , specifically including: calculating the switching basis of the nth iteration by the following formula :

[0115] ; In the formula, d MCMASA (n) represents the distance from the signal to the nearest symbol point in the constellation diagram after equalization by the MCMA algorithm improved by simulated annealing in the nth iteration. The symbol point in the constellation diagram is related to the signal modulation method.

[0116] Finally, according to , determine whether it is less than half of the distance between the two nearest symbol points in the constellation diagram. If it is less than, switch to the DDLMS algorithm with adaptive step size, specifically including: The conditions for opening the eye diagram are:

[0117] |W(n) T X(n)-a(n)| <d / 2, for all n

[0118] Where d is the distance between the two nearest points in the transmitter signal constellation diagram, and a(n) is the original signal sent by the transmitter.

[0119] The step size of the DDLMS algorithm is adjusted using the inverse hyperbolic sine function. The function expression of the step size μ2(n) changing with the error signal e2(n) is:

[0120] =a×arcsinh[b|e2(n)| c ], where a, b, c are parameters for dynamically controlling the step size, and arcsinh is the inverse hyperbolic sine function.

[0121] In the joint blind equalization algorithm, the update equation of the equalizer tap coefficient is:

[0122] ;

[0123] Where W(n+1) and W(n) represent the equalizer tap coefficients of the n+1th iteration and the nth iteration, respectively. and are the step sizes of the MCMA algorithm improved by simulated annealing and the DDLMS algorithm with adaptive step size, respectively, x(n) * It is expressed as x(n) taking the complex conjugate.

[0124] In order to solve the above technical problems, the present invention provides a joint blind equalization device, comprising:

[0125] A first equalization unit is used to adjust the tap coefficient of the equalizer for the received signal obtained by the receiving end by using the modified normal mode blind equalization MCMA algorithm based on simulated annealing, so as to perform equalization processing on the received signal according to the current tap coefficient of the equalizer to obtain a first target signal;

[0126] A judging unit, used to judge whether the first target signal achieves a preset equalization effect;

[0127] An adjustment unit, used for adjusting the step size of the decision-directed least mean square DDLMS algorithm when the first target signal reaches a preset equalization effect;

[0128] The second equalizing unit is used to adjust the tap coefficient of the equalizer according to the step size of the DDLMS algorithm, so as to perform equalization processing on the first target signal according to the current tap coefficient of the equalizer to obtain the second target signal.

[0129] To solve the above technical problems, Figure 3 As shown, the present invention provides a joint blind equalization electronic device, comprising:

[0130] Memory for storing computer programs;

[0131] The processor is used to implement the steps of the joint blind equalization method as described above when executing the computer program. For other introductions to the joint blind equalization device, please refer to the above embodiments, which will not be described in detail in this application.

[0132] To solve the above technical problems, the present invention provides a non-volatile storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned joint blind equalization method are implemented. For other introductions to the non-volatile storage medium, please refer to the above-mentioned embodiments, and this application will not repeat them here.

[0133] To solve the above technical problems, the present invention provides a computer program product, including a computer program / instruction, which implements the steps of the above joint blind equalization method when executed by a processor. For other introductions to the computer program product, please refer to the above embodiments, which will not be repeated in this application.

[0134] It should also be noted that, in this specification, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprises", "comprising" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the statement "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device including the element.

[0135] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A joint blind equalization method, characterized in that: include: For the received signal obtained by the receiving end, the tap coefficient of the equalizer is adjusted by using the modified normal mode blind equalization MCMA algorithm based on simulated annealing, so as to perform equalization processing on the received signal according to the current tap coefficient of the equalizer to obtain a first target signal; Determining whether the first target signal achieves a preset equalization effect; When the first target signal reaches the preset equalization effect, adjusting the step size of the decision-directed least mean square DDLMS algorithm; The tap coefficients of the equalizer are adjusted according to the step size of the DDLMS algorithm, so as to perform equalization processing on the first target signal according to the current tap coefficients of the equalizer to obtain a second target signal.

2. The joint blind equalization method according to claim 1, characterized in that: The probability formula for receiving a new solution in the MCMA algorithm improved based on simulated annealing is: , where P is the probability of accepting a new solution, E t+1 is the cost function value of the MCMA algorithm at time t+1, E t is the cost function value of the MCMA algorithm at time t, T t is the temperature at time t, T t Negatively correlated with the number of iterations.

3. The joint blind equalization method according to claim 1, characterized in that: Determining whether the first target signal achieves a preset equalization effect includes: Calculating a first distance from the first target signal to the nearest symbol point in the constellation diagram; Determine whether the first distance is less than a preset distance, where the preset distance is the product of a distance between two nearest symbol points in the constellation diagram and a preset parameter, where 0<the preset parameter<1; If the first distance is smaller than the preset distance, it is determined that the preset balancing effect is satisfied; otherwise, it is determined that the preset balancing effect is not satisfied.

4. The joint blind equalization method according to claim 3, characterized in that: Calculating a first distance from the first target signal to the nearest symbol point in the constellation diagram includes: Acquire signals outputted by the equalizer k times in succession, where the signals outputted k times in succession include the first target signal and signals outputted by the equalizer k-1 times in succession before outputting the first target signal, where k is an integer greater than 1; Calculate the average distance between the signal outputted k times in succession and the nearest symbol point in the constellation diagram, and use the average distance as the first distance; Determining whether the first distance is less than a preset distance includes: Determining whether the average distance is less than the preset distance; If the average distance is smaller than the preset distance, it is determined that the preset balancing effect is satisfied; otherwise, it is determined that the preset balancing effect is not satisfied.

5. The joint blind equalization method according to claim 4, characterized in that: Calculate the average distance of the signal output for k consecutive times to the nearest symbol point in the constellation diagram, including: according to Calculate the average distance between the signal output for k consecutive times and the nearest symbol point in the constellation diagram; Among them, d MCMASA (n) is the distance from the signal equalized by the MCMA algorithm improved by simulated annealing to the nearest symbol point in the constellation diagram in the nth iteration, is the average distance, and n is the total number of iterations.

6. The joint blind equalization method according to claim 4, characterized in that: Determining whether the average distance is less than the preset distance includes: According to |W(n) T X(n)-a(n)|<d×g determines whether the average distance is less than the preset distance; Among them, |W(n) T X(n)-a(n)| is the average distance, W(n) T represents a vector composed of the weight values ​​of each tap of the equalizer in the nth iteration, X(n) is the data vector input to the equalizer in the nth iteration, a(n) is the original signal sent by the transmitter, d is the distance between the two nearest symbol points in the constellation diagram, and g is the preset parameter.

7. The joint blind equalization method according to any one of claims 1 to 6, characterized in that: The tap coefficients of the equalizer are adjusted using the improved MCMA algorithm based on simulated annealing, including: use adjusting the tap coefficients of the equalizer; Wherein, W(n+1) represents the tap coefficient of the equalizer at the n+1th iteration, W(n) represents the tap coefficient of the equalizer at the nth iteration, is the step size of the MCMA algorithm improved based on simulated annealing, y(n) is the signal output by the equalizer in the nth iteration, e1(n) represents the error between the modulus value of the signal output by the equalizer in the nth iteration and the constant modulus value when executing the MCMA algorithm improved based on simulated annealing, x(n) is the received signal, It is represented by taking the complex conjugate of x(n), and the constant modulus is a constant calculated based on the original signal sent by the transmitting end.

8. The joint blind equalization method according to any one of claims 1 to 6, characterized in that: Adjusting the tap coefficient of the equalizer according to the step size of the DDLMS algorithm includes: use adjusting the tap coefficients of the equalizer; Wherein, W(n+1) represents the tap coefficient of the equalizer at the n+1th iteration, W(n) represents the tap coefficient of the equalizer at the nth iteration, is the step size of the DDLMS algorithm, y(n) is the signal output by the equalizer in the nth iteration, e2(n) represents the error between the signal output by the equalizer in the nth iteration and the signal after the decision when executing the DDLMS algorithm, x(n) is the received signal, It is represented by taking the complex conjugate of x(n), and the signal after judgment is a signal obtained after the signal output by the equalizer in the nth iteration is judged by the judge.

9. A joint blind equalization device, characterized in that: include: A first equalization unit is used to adjust the tap coefficient of the equalizer for the received signal obtained by the receiving end by using the modified normal mode blind equalization MCMA algorithm based on simulated annealing, so as to perform equalization processing on the received signal according to the current tap coefficient of the equalizer to obtain a first target signal; A judging unit, configured to judge whether the first target signal achieves a preset equalization effect; an adjusting unit, configured to adjust the step size of a decision-directed least mean square DDLMS algorithm when the first target signal reaches the preset equalization effect; The second equalizing unit is used to adjust the tap coefficient of the equalizer according to the step size of the DDLMS algorithm, so as to perform equalization processing on the first target signal according to the current tap coefficient of the equalizer to obtain a second target signal.

10. A joint blind equalization electronic device, characterized in that: include: Memory for storing computer programs; A processor, configured to implement the steps of the joint blind equalization method according to any one of claims 1 to 8 when executing a computer program.

11. A non-volatile storage medium, characterized in that: The non-volatile storage medium stores a computer program, and when the computer program is executed by the processor, the steps of the joint blind equalization method according to any one of claims 1 to 8 are implemented.

12. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instruction is executed by a processor, the steps of the joint blind equalization method according to any one of claims 1 to 8 are implemented.

Images