Mapsk blind equalization method and system suitable for low snr complex channel environment

Abstract

This invention discloses a MAPSK blind equalization method and system suitable for complex channel environments with low signal-to-noise ratios (SNR). The system includes: a transverse filter with multiple taps that receives a MAPSK signal sequence after symbol synchronization processing; an error calculator that calculates the amplitude and / or phase error of the signal based on the output of the transverse filter; and a weight vector updater that updates the weights of the transverse filter taps based on the calculated error. The blind equalization method of this invention can converge quickly in low SNR and complex channel environments, significantly improving the communication quality and reliability of satellite communications.

Description

Technical Field

[0001] This invention relates to the field of satellite communications, and particularly to satellite communication technology using the MAPSK modulation scheme. Background Technology

[0002] In practical communication systems, signals are affected by various factors during transmission, such as atmospheric interference, rain attenuation, and multipath effects. These factors can lead to inter-symbol interference (ISI), causing signal distortion. To eliminate ISI caused by channel distortion or multipath effects, equalization techniques are widely used in communication systems. Blind equalization is a special equalization technique that can achieve channel equalization without a training sequence, thereby effectively reducing the bit error rate and improving communication quality.

[0003] However, most existing blind equalization algorithms are optimized for MQAM modulation schemes, while research on MAPSK (M-ary Amplitude Phase Shift Keying) modulation schemes is relatively limited. Among existing blind equalization algorithms, the amplitude-based RDE (Radius Directed Equalization) blind equalization algorithm is considered the most suitable for MAPSK modulation schemes, but it has significant problems in practical applications. For example, it experiences a plateau during convergence, i.e., the convergence speed suddenly decreases, resulting in slow changes in residual ISI. This phenomenon is particularly severe in low signal-to-noise ratio (SNR) environments, greatly limiting the application of MAPSK modulation schemes in scenarios with low SNR, complex channel environments, and high computational speed requirements, such as satellite communications. Therefore, developing an efficient blind equalization method that can converge quickly under low SNR conditions and adapt to complex channel environments is of great significance for improving the communication quality and reliability of satellite communications. Summary of the Invention

[0004] In view of the shortcomings of the existing technology, the purpose of this invention is to propose an efficient blind equalization method that can converge quickly under low signal-to-noise ratio conditions and adapt to complex channel environments, especially satellite communication, which can significantly improve the communication quality and reliability of satellite communication.

[0005] The technical solution of the present invention is as follows:

[0006] A MAPSK blind equalization system suitable for complex channel environments with low signal-to-noise ratio includes: a filter that receives a MAPSK signal sequence after symbol synchronization processing; a weight vector updater that receives the output of the filter and feeds it back to the filter; and an error calculator that receives the output of the filter and feeds it back to the weight vector updater. The filter is a transverse filter with multiple taps. The error calculator calculates the amplitude and / or phase error of the signal based on the output of the filter, and the weight vector updater updates the weights of the taps of the transverse filter based on the error for filtering operations at the next time step.

[0007] According to some preferred embodiments of the present invention, the MAPSK blind equalization system further includes a register connected to the filter for receiving sampled data of the MAPSK signal sequence after symbol synchronization processing.

[0008] This invention further provides a MAPSK blind equalization method applying the above-mentioned blind equalization system, which includes:

[0009] S1 performs down-conversion processing on the MAPSK satellite signals received by the antenna, converting them into baseband signals;

[0010] S2 performs symbol synchronization on the baseband signal to obtain a synchronization signal x(t);

[0011] S3 initializes the weight vector W(n) of the taps of the filter to obtain a weight vector-initialized filter; the synchronization signal x(t) is filtered by the weight vector-initialized filter to obtain the filtered signal y(t);

[0012] S4 calculates the error e(t) between the filtered signal y(t) and the desired signal using the following error function:

[0013]

[0014] Among them, R 2 The statistical magnitude of the synchronization signal is as follows:

[0015] Where E represents the expectation, and || represents the modulus.

[0016] λ is the set signal threshold;

[0017] S5 updates the weight vector W(n) of the filter taps using the following weight vector update model until the signal reaches steady-state convergence:

[0018] w i (n+1)=w i (n)+μe(t)x *(t)i=1,2,...,N(2),

[0019] Among them, w i (n+1) represents the weight of the i-th tap of the filter at time n+1, w i (n) represents the weight of the i-th tap of the filter at the current time (i.e., time n), μ represents the step size iteration factor, e(t) represents the error function, and x * (t) represents the conjugate of the synchronization signal x(t), and N represents the number of filter taps.

[0020] According to some preferred embodiments of the present invention, the initialization method is as follows: the initial value of the weight of the tap located in the middle position in the filter is set to 1, and the weights of other positions are set to 0.

[0021] According to some preferred embodiments of the present invention, obtaining the filtered signal y(t) further includes: setting a register connected to the filter to receive N consecutive sampled data of the synchronization signal x(t), performing correlation operation between the filter initialized by the weight vector and the corresponding sampled data in the register to obtain the filtered signal y(t), wherein N is equal to the number of taps of the filter.

[0022] According to some preferred embodiments of the present invention, λ is set as the average of the amplitudes of the outer ring (the circle with the largest radius of signal point distribution in the MAPSK constellation diagram) and the second outer ring (the circle with the second largest radius of signal point distribution).

[0023] The beneficial effects of this invention include:

[0024] The blind equalization method of this invention only needs to utilize the error calculated by the outer loop, and can still converge quickly under low signal-to-noise ratio conditions, making it particularly suitable for satellite communication.

[0025] This invention significantly improves the convergence speed and stability of MAPSK signals in low signal-to-noise ratio environments by using a specific error function model and filter coefficient update strategy. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of the blind equalization structure used in this invention;

[0027] Figure 2 This is the constellation diagram of the received signal after down-conversion and timing synchronization obtained in Example 1;

[0028] Figure 3 The remaining ISI curves of the method and RDE algorithm of the present invention in Example 1 under the same multipath conditions are shown.

[0029] Figure 4The graph shows the MSE convergence curves of the method and RDE algorithm of the present invention in Example 1 under the same multipath conditions. Detailed Implementation

[0030] The present invention will now be described in detail with reference to embodiments and accompanying drawings. However, it should be understood that the embodiments and drawings are for illustrative purposes only and do not constitute any limitation on the scope of protection of the present invention. All reasonable modifications and combinations included within the inventive spirit of the present invention fall within the scope of protection of the present invention.

[0031] See attached document Figure 1 The blind equalization structure used in the MAPSK blind equalization method of the present invention, which is applicable to complex channel environments with low signal-to-noise ratio, includes: a filter that receives the signal sequence after symbol synchronization processing, a weight vector updater that receives the filter output and feeds it back to the filter, and an error calculator that receives the filter output and feeds it back to the weight vector updater.

[0032] The filter is a transverse filter with multiple taps, and the tap weights of the transverse filter are updated by the weight vector updater.

[0033] Based on the above blind equalization structure, the present invention can use the output y(t) of the filter as the input of the error calculator to calculate the amplitude and phase error e(t). e(t) is used as the input of the weight vector updater to update the filter weight vector. The updated weight vector W(n) is then fed back to the filter for filtering operation at the next time step.

[0034] According to some preferred embodiments of the present invention, the blind equalization structure further includes a register connected to the filter for receiving N sampled data of the signal sequence after symbol synchronization processing.

[0035] Under the above-mentioned preferred blind equalization structure, the blind equalization method of the present invention includes:

[0036] S1 performs down-conversion processing on the MAPSK satellite signals received by the antenna, converting them into baseband signals.

[0037] S2 performs symbol synchronization on the baseband signal to obtain the synchronization signal x(t).

[0038] S3 initializes the weight vector W(n) of the filter taps to obtain the weight vector-initialized filter. The weight vector-initialized filter is then correlated with the corresponding sampled data in the register to obtain the filtered signal y(t).

[0039] In some specific implementations, the initialization method is as follows: set the initial value of the tap located in the middle position of the filter to 1, and the other positions to 0.

[0040] For example, in one specific embodiment, if the number of filter taps N = 7, then the initial weight vector W(n) = [0,0,1,0,0]. T .

[0041] In some specific implementations, the related operations include multiplying each filter tap with the sampled data at the corresponding position stored in the register, and then summing all the multiplication results to obtain the filtered signal.

[0042] S4 calculates the error e(t) between the filtered signal y(t) and the desired signal using the following error function:

[0043]

[0044] Among them, R 2 The statistical magnitude of the signal is as follows:

[0045] Where E represents the expectation, and || represents the modulus.

[0046] λ is the set signal threshold.

[0047] In some specific implementations, the signal threshold can be set as the average of the amplitudes of the outer ring (the circle with the largest signal point distribution radius) and the second outer ring (the circle with the second largest signal point distribution radius) in the MAPSK constellation diagram.

[0048] S5 updates the weight vector W(n) of the filter taps using the following weight vector update model until the signal reaches steady-state convergence:

[0049] w i (n+1)=w i (n)+μe(t)x * (t)i=1,2,...,N(2),

[0050] Among them, w i (n+1) represents the weight of the i-th tap of the filter at time n+1, w i (n) represents the weight of the i-th tap of the filter at the current time (i.e., time n), μ represents the step size iteration factor, e(t) represents the error function, and x * (t) represents the conjugate of the synchronization signal x(t), and N represents the number of filter taps.

[0051] Example 1

[0052] The present invention uses a blind equalization method and a traditional amplitude-based RDE (Radius Directed Equalization) blind equalization algorithm to simulate the blind equalization of MAPSK signals in satellite communication.

[0053] The simulated satellite signal used had a signal-to-noise ratio of 20dB and channel parameters of [0.7632, -0.2567, ...

[0054] [-0.1343,0.0592,0.0267,0.0098], the signal modulation scheme is 64APSK, the constellation point configuration is 8+16+20+20, the number of filter taps is 15, and the register can store 15 consecutive sampled data of the signal.

[0055] The error function used in the RDE algorithm is as follows:

[0056] e RDE (t)=y(t)(R k 2 -|y(t)| 2 )|y(t)|∈λ k

[0057] For the statistical modulus values ​​of different ring regions, λ k The signal range is defined by k, which represents the ring region number. In this embodiment, 64APSK has 4 rings, so k = 1, 2, 3, 4.

[0058] Calculations show that R in this embodiment... k ∈{1,0.6923,0.4231,0.1923}.

[0059] Meanwhile, the remaining ISI and MSE of the two algorithms are recorded according to the following formulas:

[0060]

[0061] The constellation diagram of the synchronization signal x(t) obtained by the blind equalization method of this invention is attached. Figure 2 As shown, the received signal constellation points are overlapping, making it impossible to make an effective decision.

[0062] The residual ISI curves of the blind equalization method and the RDE algorithm of this invention are attached. Figure 3 As shown, the residual ISI of the blind equalization algorithm of this invention (OCMA curve shown) is at the 6th × 10th... 4 The algorithm reaches stability after several iterations, while the residual ISI of the traditional RDE algorithm (RDE curve shown in the figure) is at the 1.4 × 10⁻⁶ level. 5 The algorithm reached stability after several iterations, indicating that the blind equalization algorithm of this invention has a faster convergence speed than the traditional RDE algorithm.

[0063] The MSE convergence curve is attached. Figure 4As shown, although the MSE of the blind equalization algorithm of the present invention is larger than that of the traditional RDE algorithm in the initial stage, the MSE quickly decreases to less than that of the traditional RDE algorithm after the equalization converges. This indicates that the blind equalization algorithm of the present invention has a smaller MSE than the traditional RDE algorithm, and the equalization output is closer to the ideal constellation diagram.

[0064] The above embodiments are merely preferred embodiments of the present invention, and the scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A MAPSK blind equalization method suitable for low SNR complex channel environment, characterized in that, The MAPSK blind equalization system used in the MAPSK blind equalization method includes: a filter that receives a MAPSK signal sequence after symbol synchronization processing; a weight vector updater that receives the output of the filter and feeds it back to the filter; and an error calculator that receives the output of the filter and feeds it back to the weight vector updater. The filter is a transverse filter with multiple taps. The error calculator calculates the amplitude and / or phase error of the signal based on the output of the filter, and the weight vector updater updates the weights of the taps of the transverse filter based on the error for filtering operations at the next time step. The method includes the following steps: S1 performs down-conversion processing on the MAPSK satellite signals received by the antenna, converting them into baseband signals; S2 performs symbol synchronization on the baseband signal to obtain a synchronization signal x(t); S3 Initialize the weight vector W(n) of the taps of the filter to obtain the weight vector-initialized filter; use the weight vector-initialized filter to filter the synchronization signal x(t) to obtain the filtered signal y(t); S4 calculates the error e(t) between the filtered signal y(t) and the desired signal using the following error function: （1） wherein The modulus value is calculated for the synchronization signal as follows: where E denotes expectation and || denotes the modulus, a set signal threshold value; S5 updates the weight vector W(n) of the filter taps using the following weight vector update model until the signal reaches steady-state convergence: （2） Among them, w i (n+1) represents the weight of the i-th tap of the filter at time n+1, w i (n) represents the weight of the i-th tap of the filter at the current time, i.e., time n. Let x represent the step size iteration factor, e(t) represent the error function, and x represent the step size iteration factor. * (t) represents the conjugate of the synchronization signal x(t), and N represents the number of filter taps.

2. The MAPSK blind equalization method of claim 1, wherein, The MAPSK blind equalization system also includes a register connected to the filter that receives sampled data of the MAPSK signal sequence after symbol synchronization processing.

3. The MAPSK blind equalization method according to claim 1, characterized in that, The initialization method is as follows: set the initial value of the tap located in the middle position of the filter to 1, and the weight of other positions to 0.

4. The method of claim 1, wherein, The process of obtaining the filtered signal y(t) further includes: setting up a register connected to the filter to receive N consecutive sampled data of the synchronization signal x(t), performing correlation operations between the filter initialized by the weight vector and the corresponding sampled data in the register to obtain the filtered signal y(t), where N is equal to the number of taps of the filter.

5. The method for MAPSK blind equalization of claim 1, wherein, The The average of the amplitudes of the outer ring, which is the circle set to the largest radius of signal point distribution in the MAPSK signal constellation, and the secondary outer ring, which is the circle set to the second largest radius of signal point distribution, is set.

Images