A variable step-size dual-mode blind equalization method for QAM signals, a storage medium and an equipment

Abstract

The application discloses a variable step-size dual-mode blind equalization method for QAM signals, a storage medium and equipment, and belongs to the technical field of blind receiving and demodulation of wireless communication systems. The application solves the problem that the traditional hard switching mechanism is easy to cause system transient oscillation and lock loss. The application adjusts the iteration step size by constructing an error exponential function with a transient error signal as an independent variable, automatically gives the system a larger step size in the initial stage of the algorithm to accelerate convergence, and adaptively attenuates when the error tends to zero to suppress steady-state fluctuations, effectively overcoming the inherent contradiction between convergence speed and steady-state accuracy of the traditional fixed step-size algorithm; a soft switching mechanism based on a continuous confidence factor is used to replace the traditional hard switching decision, the global convergence ability of the MMA algorithm and the extremely high steady-state accuracy of the DD algorithm are smoothly and non-oscillatingly fused through a nonlinear fusion factor, transient oscillation and error propagation effect caused by mode switching are significantly suppressed, and system lock loss is avoided. The application can be applied to blind equalization of wireless communication systems.

Description

Technical Field

[0001] This invention belongs to the field of blind reception and demodulation technology of wireless communication systems, specifically relating to a variable step-size dual-mode blind equalization method, storage medium, and device for QAM signals. Background Technology

[0002] In high-speed broadband wireless communication systems, multipath fading channels can cause severe inter-symbol interference (ISI), leading to received signal distortion and increased system bit error rate. To eliminate ISI, equalization techniques are typically introduced at the receiver. Traditional adaptive equalization techniques require transmitting known training sequences, which consumes channel bandwidth and reduces system communication efficiency. Blind equalization techniques, on the other hand, do not require transmitting training sequences and can achieve channel equalization and feature recovery solely by utilizing the statistical characteristics of the received signal. This offers significant advantages in improving spectrum utilization and is particularly suitable for high-speed communication scenarios in complex, time-varying channel environments.

[0003] Although blind equalization technology has shown many advantages in communication systems, the widely used classic blind equalization algorithms mainly include Bussgang-type algorithms (such as CMA and MMA) and decision-oriented (DD) algorithms, which still have significant performance bottlenecks in practical engineering deployments.

[0004] 1. The Inherent Contradiction of Fixed Step Size Mechanisms: Traditional Bussgang-type and multi-mode blind equalization algorithms typically use a fixed empirical constant as the iteration step size. While a larger step size can improve the initial convergence rate and dynamic tracking capability, it significantly amplifies stochastic gradient noise in the steady-state phase, leading to excessively high residual mean square error. Conversely, choosing a smaller step size to achieve ideal steady-state convergence accuracy severely delays the acquisition process, making it difficult to meet the rapid response requirements in complex time-varying channel environments.

[0005] 2. Limitations of Single Algorithm Modes: While the CMA algorithm exhibits good robustness and fast initial convergence, it suffers from phase rotation ambiguity and high steady-state residuals. The MMA algorithm overcomes the phase recovery problem and improves the equalization accuracy for high-order QAM signals by introducing an orthogonal decoupling structure, but its steady-state residual error remains relatively large due to the inherent limitations of the cost function. Conversely, the DD algorithm possesses extremely high steady-state convergence accuracy and can approach the ideal performance of minimum mean square error (MMSE), but it is prone to getting trapped in local minima and causing error propagation effects when the initial bit error rate of the system is high.

[0006] 3. Defects of Traditional Dual-Mode Hard Handover Mechanisms: To balance global convergence and steady-state accuracy, dual-mode architectures are often adopted in engineering, but currently, hard handover mechanisms are the primary reliance. Hard handover relies on a fixed error decision threshold for state transitions. Due to its inherent threshold sensitivity, the optimal handover threshold is difficult to determine in non-stationary channels. The instantaneous occurrence of hard handover causes discontinuous changes in the gradient of weight coefficient updates, easily triggering transient oscillations in the system. Furthermore, when the system error fluctuates slightly around the threshold, the equalizer will frequently switch between the two modes, producing a "ping-pong effect," which in extreme cases may even lead to the entire equalization system losing lock.

[0007] In summary, in order to overcome the technical limitations of the traditional fixed step size mechanism and hard switching dual-mode architecture, it is urgent to design a more advanced blind equalization mechanism. Summary of the Invention

[0008] This invention addresses the problem of transient oscillations and loss of lock-up that are easily caused by traditional dual-mode hard switching mechanisms. It proposes a variable step-size dual-mode blind equalization method, storage medium, and device for QAM signals.

[0009] The technical solution adopted by this invention to solve the above-mentioned technical problems is: a variable step-size dual-mode blind equalization method for QAM signals, the method specifically including the following steps:

[0010] Step 1: Initialize equalizer tap coefficients And let the fusion error ;

[0011] Step 2: Initialization ;

[0012] Step 3: Receive the input. Data symbols Using an equalizer The process is performed to obtain the equalizer output value. ;

[0013] Step 4: Based on the equalizer output value With judgment value Calculate the local mean square normalized error. ;

[0014] Step 5: Based on the local mean square normalization error Calculate the variable step size factor With nonlinear fusion factor ;

[0015] Step Six, according to Calculate the multi-mode algorithm error separately Error with decision-oriented algorithm Using nonlinear fusion factors Multi-mode algorithm error Error with decision-oriented algorithm By performing a smooth linear convex combination, the joint error function value of the two modes is obtained. ;

[0016] Step 7: Utilize the variable step size factor Joint error function with dual modes Update the equalizer tap coefficients;

[0017] Step 8: Determine if there are still any unprocessed data symbols. If there are, then... Then return to step three and continue until all data symbols have been sent.

[0018] Furthermore, the specific process of step four is as follows:

[0019] (1)

[0020] In the formula, For the first Data symbols The judgment value, For the equalizer to the first Data symbols The processed output value To take the absolute value, To evaluate the window length.

[0021] Furthermore, the variable step size factor The calculation method is as follows:

[0022] Constructing based on instantaneous fusion error The error exponential function of the independent variable:

[0023] (2)

[0024] In the formula, and These are the positive adjustment parameters for controlling the range and shape of the control error exponential function, respectively. is the base of the natural logarithm;

[0025] Variable step size factor for:

[0026] (3)

[0027] In the formula, To fix the step size factor, This is a preset threshold.

[0028] Furthermore, the nonlinear fusion factor Determined using a continuous piecewise cosine function:

[0029] (4)

[0030] In the formula, , The preset threshold value is used. .

[0031] Furthermore, the dual-mode joint error function value for:

[0032] (5).

[0033] Furthermore, the specific process of step seven is as follows:

[0034] (6)

[0035] in, for conjugate, This refers to the updated equalizer tap coefficients.

[0036] A computer storage medium storing at least one instruction, which is loaded and executed by a processor to implement the aforementioned variable step-size dual-mode blind equalization method for QAM signals.

[0037] A variable step-size dual-mode blind equalization device for QAM signals, the device comprising a processor and a memory, the memory storing at least one instruction, the at least one instruction being loaded and executed by the processor to implement the variable step-size dual-mode blind equalization method for QAM signals.

[0038] The beneficial effects of this invention are:

[0039] This invention dynamically adjusts the iteration step size by constructing an error exponential function with the instantaneous error signal as the independent variable. This automatically assigns a larger step size to the system in the early stages of the algorithm to accelerate convergence, and adaptively decays the step size as the error approaches zero to suppress steady-state fluctuations. This effectively overcomes the inherent contradiction between convergence speed and steady-state accuracy in traditional fixed-step-size algorithms. Simultaneously, it replaces the traditional binary threshold hard-switching decision with a soft-switching mechanism based on a continuous confidence factor. Through a nonlinear fusion factor, it smoothly and oscillatorlessly fuses the global convergence capability of the MMA algorithm with the extremely high steady-state accuracy of the DD algorithm, significantly suppressing transient oscillations and error propagation effects caused by mode switching and preventing system lock-out. Furthermore, while completely eliminating inter-symbol interference and improving the system bit error rate, this invention also significantly enhances the dynamic tracking and rapid self-recovery robustness of high-order QAM signals in complex non-stationary environments such as channel step abrupt changes, providing an efficient and reliable engineering solution for blind demodulation of high-speed broadband communication signals. Attached Figure Description

[0040] Figure 1 This is a flowchart of a variable step-size dual-mode blind equalization method for QAM signals according to the present invention.

[0041] Figure 2 The constellation diagram received before equalization;

[0042] Figure 3 This is the signal constellation diagram output after equalization using the VSS-MMA-DD algorithm in this invention;

[0043] Figure 4 This is a comparison of the mean square error convergence curves of the VSS-MMA-DD algorithm of the present invention and the traditional fixed step size algorithm under static non-stationary channels;

[0044] Figure 5 The equivalent channel impulse response diagram of the cascaded system after the VSS-MMA-DD algorithm of the present invention reaches steady state after equalization;

[0045] Figure 6 This is a comparison of the bit error rate (BER) performance curves before and after applying the VSS-MMA-DD algorithm in this invention.

[0046] Figure 7 This is a comparison diagram of the step change response and mean square error convergence trajectory of the VSS-MMA-DD algorithm of the present invention under a dynamic time-varying channel environment. Detailed Implementation

[0047] Specific implementation method one: Combining Figure 1 This embodiment describes a variable step-size dual-mode blind equalization method for QAM signals. The QAM signal of this invention is a 16QAM high-order coherent modulation signal. The method specifically includes the following steps:

[0048] Step 1: Initialize equalizer tap coefficients And let the fusion error ;

[0049] Step 2: Initialization ;

[0050] Step 3: Receive the input. Data symbols Using an equalizer The process is performed to obtain the equalizer output value. ;

[0051] Step 4: Based on the equalizer output value With judgment value Calculate the local mean square normalized error. ;

[0052] (1)

[0053] In the formula, For the first Data symbols The judgment value, For the equalizer to the first Data symbols The processed output value To take the absolute value, To evaluate window length;

[0054] Step 5: Based on the local mean square normalization error Calculate the variable step size factor With nonlinear fusion factor ;

[0055] Specifically,

[0056] Constructing based on instantaneous fusion error The error exponential function of the independent variable:

[0057] (2)

[0058] In the formula, and These are the positive adjustment parameters for controlling the range and shape of the control error exponential function, respectively. is the base of the natural logarithm;

[0059] The variable step size factor under the dual-mode architecture for:

[0060] (3)

[0061] In the formula, This is the fixed step size factor used after the system reaches steady state. The preset threshold;

[0062] Nonlinear fusion factor Determined using a continuous piecewise cosine function:

[0063] (4)

[0064] In the formula, , The preset threshold value is used. ;

[0065] Step Six, according to Calculate the multi-mode algorithm error separately Error with decision-oriented algorithm Using nonlinear fusion factors Multi-mode algorithm error Error with decision-oriented algorithm By performing a smooth linear convex combination, the joint error function of the two modes is obtained. :

[0066] (5)

[0067] Step 7: Utilize the variable step size factor Joint error function with dual modes Update the equalizer tap coefficients;

[0068] Specifically,

[0069] (6)

[0070] in, for conjugate, This refers to the updated equalizer tap coefficients.

[0071] Step 8: Determine if there are still any unprocessed data symbols. If there are, then... Then return to step three and continue until all data symbols have been sent.

[0072] Specific Implementation Method Two: This implementation method is a computer storage medium that stores at least one instruction. The at least one instruction is loaded and executed by a processor to implement the variable step size dual-mode blind equalization method for QAM signals.

[0073] It should be understood that the instructions include computer program products, software, or computerized methods corresponding to any method described in this invention; the instructions can be used to program computer systems or other electronic devices. Computer storage media may include readable media on which instructions are stored, and may include, but are not limited to, magnetic storage media, optical storage media; magneto-optical storage media include read-only memory (ROM), random access memory (RAM), erasable programmable memory (e.g., EPROM and EEPROM), and flash memory layers, or other types of media suitable for storing electronic instructions.

[0074] Specific Implementation Method 3: This implementation method is a variable step size dual-mode blind equalization device for QAM signals. The device includes a processor and a memory. It should be understood that this includes any device described in this invention that includes a processor and a memory. The device may also include other units and modules that perform display, interaction, processing, control, and other functions through signals or instructions.

[0075] The memory stores at least one instruction, which is loaded and executed by the processor to implement the variable step size dual-mode blind equalization method for QAM signals.

[0076] Those skilled in the art will understand that at least one stored instruction constitutes a computer program product corresponding to a method or system. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of this application can be implemented using various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0077] This application is described with reference to flowchart illustrations and / or block diagrams of methods, systems, and computer program products according to embodiments of this application, and can also be used with corresponding devices. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0078] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0079] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0080] Experimental Section

[0081] The system signal-to-noise ratio is set to 25dB, the equalizer adopts a transverse finite impulse response filter structure, and the number of filter taps is set to 8.

[0082] The core parameters of this invention are configured as follows:

[0083] During the system initialization phase, set the initial values ​​of the equalizer tap coefficients. And let the initial dual-mode joint error As data symbols are input one by one, the system calculates the current filter's output value in real time. and the corresponding local mean square error normalized value .

[0084] To balance the resource utilization of the underlying hardware implementation with the algorithm's timing performance, strict engineering constraints were imposed on the core parameters of variable step size and soft switching. In the variable step size factor update strategy, to avoid the resource-intensive floating-point multiplier, the positive adjustment parameter of the error exponential function was... Set as Shape adjustment parameters Set to 10. When the instantaneous error is large, the step size factor is adjusted. Fast approximation of the maximum value To accelerate the initial convergence of the algorithm; once the system reaches steady state, the step size factor is fixed. Set as The power of 2 parameters mentioned above can be directly mapped to arithmetic shift register operations during hardware logic synthesis in FPGAs and other applications, significantly saving DSP processing units.

[0085] In the dual-mode soft handover mechanism, a high threshold is set. Set the low threshold to 0.0316 (equivalent to MSE of -15dB). The threshold is 0.01 (equivalent to an MSE of -20 dB). The system dynamically outputs the fusion factor based on these two thresholds and a continuous piecewise cosine function. This allows for a smooth transition between the Multi-Model Algorithm (MMA) and the Decision-Driven Algorithm (DD), and enables the calculation of the joint error function. And complete the tap coefficient Adaptive updates.

[0086] Combination Figures 2 to 7 The actual demodulation performance and anti-interference effect of the variable step size dual-mode blind equalization algorithm proposed in this embodiment are analyzed in detail:

[0087] Depend on Figure 2 and Figure 3 A comparison of the constellation diagrams reveals that, due to the small level intervals in the 16QAM signal transmission, the constellation diagram of the received signal before equalization exhibits a completely closed and discrete state under severe multipath distortion channels, resulting in significant phase rotation. After processing by the VSS-MMA-DD algorithm of this invention, the constellation diagram of the output signal after equalization can clearly converge to 16 standard aggregation points, proving that the algorithm of this invention successfully achieves amplitude and phase joint equalization and eliminates phase rotation ambiguity.

[0088] Depend on Figure 4 A comparison of the error convergence curves reveals that, under the same static non-stationary channel conditions, traditional fixed-step-size algorithms suffer from slow convergence speed or large steady-state errors. In contrast, the VSS-MMA-DD algorithm of this invention, relying on a large step size and the dominance of the MMA algorithm in the initial stage, can quickly enter the convergence state. In the later stages of iteration, the system smoothly switches to DD mode with a small step size, greatly reducing residual errors. Ultimately, the steady-state mean square error converges to the -23 dB level, demonstrating a significantly better overall equalization effect than traditional algorithms.

[0089] Depend on Figure 5 As can be seen from the equivalent channel impulse response diagram, after the equalizer reaches steady state, the impulse response of the cascaded channel exhibits a very sharp main peak at the main tap position, while the tap coefficients at other positions are greatly suppressed and approach zero. This indicates that through adaptive updating of the 8-tap transverse filter, the dispersion effect of the multipath channel has been canceled, and the inter-symbol interference of the channel has been fundamentally eliminated.

[0090] Depend on Figure 6As shown in the system bit error rate (BER) performance curve, the system was in a high BER state before equalization. After adopting the algorithm of this invention, in the low to medium signal-to-noise ratio range, the BER curve decreases by an order of magnitude with the increase of signal-to-noise ratio, and highly approximates the theoretical curve under interference-free conditions; under the high signal-to-noise ratio condition of 25dB, relying on the high-precision steady-state tracking capability of DD mode, the system maintains an extremely low BER flat bottom, which fully meets the demodulation threshold requirements of high-order signals.

[0091] To verify the robustness of the algorithm, a forced channel step jump was injected at the 10000th and 20000th iterations of the weights. Figure 7 As can be seen from the dynamic channel change response curve, the error surges at the instant the change occurs. The soft handover mechanism of this invention can quickly sense the change by... Adaptive zeroing reactivates the global search capability of the MMA mode, while the variable step size function automatically outputs a large step size. The system achieves reconvergence within an extremely short symbol period, demonstrating excellent dynamic tracking capability and robustness against loss of lock, making it highly suitable for complex and ever-changing broadband communication scenarios.

[0092] The above examples of the present invention are merely illustrative of the computational model and process of the present invention, and are not intended to limit the implementation of the present invention. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is impossible to exhaustively list all possible implementations here. Any obvious variations or modifications derived from the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A variable step-size dual-mode blind equalization method for QAM signals, characterized in that, The method specifically includes the following steps: Step 1: Initialize equalizer tap coefficients And let the fusion error ; Step 2: Initialization ; Step 3: Receive the input. Data symbols Using an equalizer The process is performed to obtain the equalizer output value. ; Step 4: Based on the equalizer output value With judgment value Calculate the local mean square normalized error. ; Step 5: Based on the local mean square normalization error Calculate the variable step size factor With nonlinear fusion factor ; Step Six, according to Calculate the multi-mode algorithm error separately Error with decision-oriented algorithm Using nonlinear fusion factors Multi-mode algorithm error Error with decision-oriented algorithm By performing a smooth linear convex combination, the joint error function value of the two modes is obtained. ; Step 7: Utilize the variable step size factor Joint error function with dual modes Update the equalizer tap coefficients; Step 8: Determine if there are still any unprocessed data symbols. If there are, then... Then return to step three and continue until all data symbols have been sent.

2. The variable step-size dual-mode blind equalization method for QAM signals according to claim 1, characterized in that, The specific process of step four is as follows: (1) In the formula, For the first Data symbols The judgment value, For the equalizer to the first Data symbols The processed output value To take the absolute value, To evaluate the window length.

3. The variable step-size dual-mode blind equalization method for QAM signals according to claim 2, characterized in that, The variable step size factor The calculation method is as follows: Constructing based on instantaneous fusion error The error exponential function of the independent variable: (2) In the formula, and These are the positive adjustment parameters for controlling the range and shape of the control error exponential function, respectively. is the base of the natural logarithm; Variable step size factor for: (3) In the formula, To use a fixed step size factor, This is a preset threshold.

4. The variable step-size dual-mode blind equalization method for QAM signals according to claim 3, characterized in that, The nonlinear fusion factor Determined using a continuous piecewise cosine function: (4) In the formula, , The preset threshold value is used. .

5. A variable step-size dual-mode blind equalization method for QAM signals according to claim 4, characterized in that, The dual-mode joint error function value for: (5)。 6. A variable step-size dual-mode blind equalization method for QAM signals according to claim 5, characterized in that, The specific process of step seven is as follows: (6) in, for conjugate, This refers to the updated equalizer tap coefficients.

7. A computer storage medium, characterized in that, The storage medium stores at least one instruction, which is loaded and executed by a processor to implement a variable step-size dual-mode blind equalization method for QAM signals as described in any one of claims 1 to 6.

8. A variable step-size dual-mode blind equalization device for QAM signals, characterized in that, The device includes a processor and a memory, the memory storing at least one instruction, which is loaded and executed by the processor to implement a variable step-size dual-mode blind equalization method for QAM signals as described in any one of claims 1 to 6.

Images