Joint equalization method for signals under fading channel

Abstract

The invention provides a joint equalization method for signals under a fading channel, and the method comprises the steps: obtaining a cost function of a CMA equalizer and a cost function of a DD equalizer according to IQ data to be equalized; obtaining a cost function of a joint equalizer according to the cost function of the CMA equalizer and the cost function of the DD equalizer; and performing joint equalization processing on the to-be-equalized IQ data to obtain equalized IQ data. The method and the device have high practicability and robustness, and loss can be compensated even if strong channel fading exists in the signal, so that the signal can be demodulated.

Description

Technical Field

[0001] The present application relates to the technical field of signal processing, and in particular to a joint equalization method for signals in a fading channel. Background Art

[0002] Since its invention, radio communication has played an important role in many fields and is now inseparable from human society. More and more scientific and technological personnel are engaged in the research and development and optimization of radio communication technology. In digital communication, data is often affected by external factors such as white noise, group delay and multipath effect when passing through the channel, so the channel presents a complex and changeable characteristic, resulting in signal attenuation, distortion and distortion. According to the Nyquist no-inter-code interference criterion, this interference needs to be equalized and compensated for in order to facilitate subsequent demodulation processing.

[0003] In recent years, equalization technology has been increasingly developed, and commonly used single equalizers have begun to be widely used in signal demodulation equipment. However, in the face of complex and changeable channel conditions, single equalizers cannot fully meet the needs. Therefore, the need for updating and iteration of equalization technology is imminent, and how to combine various equalizers has become a trend. In radio signal technology, channel equalization plays a key role in demodulation. How to reduce convergence errors and how to accurately compensate channel parameters under multipath channel conditions have become the key. Summary of the invention

[0004] In view of this, the present application provides a joint equalization method for signals in fading channels, and uses a joint equalizer algorithm to perform signal recovery according to the severe situation of channel fading in radio propagation.

[0005] The present application discloses a joint equalization method for signals in a fading channel, which includes:

[0006] Step 1: According to the IQ data before equalization, the cost function of the constant modulus algorithm (CMA) equalizer and the cost function of the decision-directed (DD) equalizer are obtained;

[0007] Step 2: According to the cost function of the CMA equalizer and the cost function of the DD equalizer, the cost function of the joint equalizer is obtained;

[0008] Step 3: Perform joint equalization processing on the IQ data before equalization to obtain the IQ data after equalization.

[0009] Furthermore, the step 1 comprises:

[0010] Input the IQ data before equalization and arrange them head to tail to obtain the first data;

[0011] The first data is input into the CMA equalizer to obtain the equalization error function of the CMA equalizer. y CMA (n) and They are respectively the square of the statistical modulus value of the CMA equalizer equalized output signal and the square of the modulus value of the modulation type;

[0012] According to the equalization error function of the CMA equalizer, the cost function of the CMA equalizer is calculated, and the cost function is the error function e CMA The mean square error of (n)

[0013] Furthermore, the step 1 comprises:

[0014] Input the IQ data before equalization and arrange them head to tail to obtain the first data;

[0015] The first data is input into the DD equalizer to obtain the equalization error function of the DD equalizer. y DD (n) and They are the standard constellation mapping symbol points of the DD equalizer equalized output signal and modulation type respectively;

[0016] According to the equalization error function of the DD equalizer, the cost function of the DD equalizer is calculated, and the cost function is the error function e DD The mean square error of (n)

[0017] Furthermore, the step 2 comprises:

[0018] A nonlinear cascade is used to nonlinearly cascade the cost function of the CMA equalizer and the cost function of the DD equalizer, and the coupling is formed into the cost function of the joint equalizer;

[0019] The key parameters of the joint equalizer are set, wherein the key parameters are the number of filter taps, the iteration step size and the number of cycles.

[0020] Furthermore, the method for obtaining the cost function of the joint equalizer includes:

[0021] Based on the cost function JDD of the DD equalizer, the coupling coefficient is obtained using the hyperbolic tangent function:

[0022]

[0023] According to the coupling coefficient λ, the cost function of the DD equalizer and the cost function of the CMA equalizer, the cost function of the joint equalizer is obtained:

[0024] J=(1-λ)*J DD +λ*J CMA

[0025] Among them, J DD is the cost function of the DD equalizer, J CMA is the cost function of the CMA equalizer, and J is the cost function of the joint equalizer.

[0026] Furthermore, the cost function of the CMA equalizer is:

[0027]

[0028] Among them, y CMA (n) and are the statistical modulus square of the CMA equalizer output signal and the modulus square of the modulation type; J CMA is the cost function of the CMA equalizer; E is the mean square error function.

[0029] Furthermore, the cost function of the DD equalizer is:

[0030]

[0031] Among them, y DD (n) and are the standard constellation mapping symbol points of the DD equalizer output signal and modulation type respectively; J DD is the cost function of the DD equalizer; E is the mean square error function.

[0032] Furthermore, the step 3 comprises:

[0033] The IQ data before equalization is subjected to joint equalization processing to determine whether the IQ data after equalization meets the convergence condition; if the convergence condition is not met, the equalization is continued, otherwise a converged constellation diagram is obtained, that is, the equalized IQ data is obtained.

[0034] Further, if the vector error between the equalized constellation and the standard constellation of the corresponding modulation mode is less than 10 within a signal-to-noise ratio of 20 dB, it is determined that the equalized IQ data meets the convergence condition.

[0035] Furthermore, the calculation formula of the vector error is:

[0036]

[0037] Among them, EVM RMS is the vector error, Ik is the I data after equalization, Qk is the Q data after equalization, For the standard constellation diagram reference I data, is the reference Q data of the standard constellation diagram, and N is the IQ data length.

[0038] Due to the adoption of the above technical solution, the present application has the following advantages:

[0039] 1. The joint equalization method of the present application has high accuracy. The commonly used CMA equalizer has the problems of slow convergence speed and false convergence, and when the CMA algorithm is used to equalize the signal, the cost function is only related to the amplitude information of the equalized received signal, and has nothing to do with its phase information. Finally, the equalized output signal will have a phase shift; similarly, DD equalization also has certain problems, such as the poor restart ability of the DD algorithm, which causes the constellation diagram to diverge and the eye diagram cannot be opened. The joint equalization method of the present invention can avoid these two problems and ensure the accuracy of the signal.

[0040] 2. In the process of signal joint equalization processing, the present application can still correctly converge the signal even under the interference of deep multipath fading channels. The longer the duration of the signal, the more symbols can be used, and the smaller the steady-state error after joint equalization. This method has strong robustness.

[0041] 3. Compared with the prior art, the present application has stronger practicality and robustness. Even if there is strong channel fading in the signal, the loss can be compensated so that the signal can be demodulated and processed.

[0042] 4. This application nonlinearly cascades two commonly used equalizers so that the distorted signal can be restored more accurately. It has strong robustness and can perform the equalization task excellently even in the face of harsh channel conditions.

[0043] 5. The joint equalization method designed in this application for signals in fading channels can effectively meet the convergence conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 A flowchart of a joint equalization method for signals in a fading channel according to an embodiment of the present application is provided.

[0045] Figure 2 This is a schematic diagram of the adaptive linear equalization principle of CMA equalization and DD equalization in an embodiment of the present application.

[0046] Figure 3 This is a flow chart of an embodiment of the present application of using a nonlinear cascade to couple the CMA equalization cost function and the DD equalization cost function into a joint equalization cost function.

[0047] Figure 4 This is a diagram showing the effect of using a single CMA equalization signal to restore the constellation in an embodiment of the present application.

[0048] Figure 5 This is a diagram showing the effect of restoring a constellation using a single DD equalized signal in an embodiment of the present application.

[0049] Figure 6This is a diagram showing the constellation effect of the joint equalization signal recovery according to an embodiment of the present application. DETAILED DESCRIPTION

[0050] The present application is further described in conjunction with the accompanying drawings and embodiments, and the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by ordinary technicians in this field should fall within the scope of protection of the embodiments of the present application.

[0051] See also Figure 1 The present application provides an embodiment of a joint equalization method for signals in a fading channel, which includes:

[0052] Step 1: According to the IQ data before equalization, the cost function of the constant modulus algorithm (CMA) equalizer and the cost function of the decision-directed (DD) equalizer are obtained;

[0053] Step 2: According to the cost function of the CMA equalizer and the cost function of the DD equalizer, the cost function of the joint equalizer is obtained;

[0054] Step 3: Perform joint equalization processing on the IQ data before equalization to obtain the IQ data after equalization.

[0055] This application nonlinearly cascades two commonly used equalizers, the CMA equalizer and the DD equalizer, so that the distorted signal can be restored more accurately and has strong robustness. Even in the face of harsh channel conditions, it can also excellently complete the equalization task.

[0056] This application nonlinearly cascades two commonly used equalizers so that the distorted signal can be restored more accurately and has strong robustness. Even in the face of harsh channel conditions, it can also excellently complete the equalization task.

[0057] Optionally, step 1 includes:

[0058] like Figure 2 As shown, x(n) is the input data, y(n) is the output data, is the expected data, e(n) is the error data, is the training data, T is the signal period, w0,w1,w2,...,w M-1 It is the filter tap data, using subscripts to distinguish CMA and DD;

[0059] Input the IQ data before equalization and arrange them head to tail to obtain the first data;

[0060] The first data is input into the CMA equalizer to obtain the equalization error function of the CMA equalizer. y CMA (n) and They are respectively the square of the statistical modulus value of the CMA equalizer equalized output signal and the square of the modulus value of the modulation type;

[0061] According to the equalization error function of the CMA equalizer, the cost function of the CMA equalizer is calculated, and the cost function is the error function e CMA The mean square error of (n)

[0062] Optionally, step 1 includes:

[0063] Input the IQ data before equalization and arrange them head to tail to obtain the first data;

[0064] The first data is input into the DD equalizer to obtain the equalization error function of the DD equalizer. y DD (n) and They are the standard constellation mapping symbol points of the DD equalizer equalized output signal and modulation type respectively;

[0065] According to the equalization error function of the DD equalizer, the cost function of the DD equalizer is calculated, and the cost function is the error function e DD The mean square error of (n)

[0066] Optionally, the step 2 includes:

[0067] A nonlinear cascade is used to nonlinearly cascade the cost function of the CMA equalizer and the cost function of the DD equalizer, and the coupling is formed into the cost function of the joint equalizer;

[0068] The key parameters of the joint equalizer are set, wherein the key parameters are the number of filter taps, the iteration step size and the number of cycles.

[0069] Optionally, the method for obtaining the cost function of the joint equalizer includes:

[0070] Based on the cost function JDD of the DD equalizer, the coupling coefficient is obtained using the hyperbolic tangent function:

[0071]

[0072] According to the coupling coefficient λ, the cost function of the DD equalizer and the cost function of the CMA equalizer, the cost function of the joint equalizer is obtained:

[0073] J=(1-λ)*J DD +λ*J CMA

[0074] Among them, J DD is the cost function of the DD equalizer, J CMAis the cost function of the CMA equalizer, and J is the cost function of the joint equalizer.

[0075] Optionally, the cost function of the CMA equalizer is:

[0076]

[0077] Among them, y CMA (n) and are the statistical modulus square of the CMA equalizer output signal and the modulus square of the modulation type; J CMA is the cost function of the CMA equalizer; E is the mean square error function.

[0078] Optionally, the cost function of the DD equalizer is:

[0079]

[0080] Among them, y DD (n) and are the standard constellation mapping symbol points of the DD equalizer output signal and modulation type respectively; J DD is the cost function of the DD equalizer; E is the mean square error function.

[0081] Optionally, step 3 includes:

[0082] The IQ data before equalization is subjected to joint equalization processing to determine whether the IQ data after equalization meets the convergence condition; if the convergence condition is not met, the equalization is continued, otherwise a converged constellation diagram is obtained, that is, the equalized IQ data is obtained.

[0083] Optionally, if a vector error between the equalized constellation diagram and a standard constellation diagram shape of a corresponding modulation mode is less than 10 within a signal-to-noise ratio of 20 dB, it is determined that the equalized IQ data meets the convergence condition.

[0084] Optionally, the calculation formula of the vector error is:

[0085]

[0086] Among them, EVM RMS is the vector error, Ik is the I data after equalization, Qk is the Q data after equalization, For the standard constellation diagram reference I data, is the reference Q data of the standard constellation diagram, and N is the IQ data length.

[0087] In this specific embodiment, Figure 1 As shown, the present application first inputs the IQ data before equalization and arranges them head to tail; then obtains the CMA equalization cost function, which has a poor effect when used alone, such as Figure 4As shown; secondly, the DD equilibrium cost function is obtained, which is not effective when used alone, such as Figure 5 As shown; then the two equilibrium cost functions are coupled using a nonlinear cascade, as Figure 3 As shown; then determine the number of equalization filter taps, iteration step size and number of cycles and other parameters; finally, perform joint equalization recovery processing on the distorted signal to obtain a converged constellation diagram, as shown Figure 6 shown.

[0088] The cost function calculation structure of CMA equilibrium and DD equilibrium is as follows Figure 2 As shown in the figure, the feedback link structures of the two are the same, and the difference lies in the error calculation method of the cost function.

[0089] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present application rather than to limit it. Although the present application has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present application can still be modified or replaced by equivalents, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present application should be included in the scope of protection of the claims of the present application.

Claims

1. A joint equalization method for signals in fading channels, characterized in that: include: Step 1: According to the IQ data before equalization, the cost function of the CMA equalizer and the cost function of the DD equalizer are obtained; Step 2: Obtaining a cost function of a joint equalizer according to the cost function of the CMA equalizer and the cost function of the DD equalizer; Step 3: Perform joint equalization processing on the IQ data before equalization to obtain IQ data after equalization.

2. The joint equalization method for signals in fading channels according to claim 1, characterized in that: The step 1 comprises: Input the IQ data before equalization and arrange them head to tail to obtain the first data; The first data is input into the CMA equalizer to obtain an equalization error function of the CMA equalizer. y CMA (n) and The square of the statistical modulus value of the equalized output signal of the CMA equalizer and the square of the modulus value of the modulation type are respectively; According to the equalization error function of the CMA equalizer, the cost function of the CMA equalizer is calculated, and the cost function is the error function e CMA The mean square error of (n) 3. The joint equalization method for signals in fading channels according to claim 1, characterized in that: The step 1 comprises: Input the IQ data before equalization and arrange them head to tail to obtain the first data; The first data is input into the DD equalizer to obtain the equalization error function of the DD equalizer. y DD (n) and The standard constellation mapping symbol points for the DD equalizer equalized output signal and modulation type respectively; According to the equalization error function of the DD equalizer, the cost function of the DD equalizer is calculated, and the cost function is the error function e DD The mean square error of (n) 4. The joint equalization method for signals in fading channels according to claim 1, characterized in that: The step 2 comprises: Using a nonlinear cascade to nonlinearly cascade the cost function of the CMA equalizer and the cost function of the DD equalizer, and coupling them into a cost function of a joint equalizer; The key parameters of the joint equalizer are set, wherein the key parameters are the number of filter taps, the iteration step size and the number of cycles.

5. The joint equalization method for signals in fading channels according to claim 1, characterized in that: The method for obtaining the cost function of the joint equalizer includes: Based on the cost function JDD of the DD equalizer, the coupling coefficient is obtained using the hyperbolic tangent function: According to the coupling coefficient λ, the cost function of the DD equalizer and the cost function of the CMA equalizer, the cost function of the joint equalizer is obtained: J=(1-λ)*J DD +λ*J CMA Among them, J DD is the cost function of the DD equalizer, J CMA is the cost function of the CMA equalizer, and J is the cost function of the joint equalizer.

6. The joint equalization method for signals in fading channels according to claim 5, characterized in that: The cost function of the CMA equalizer is: Among them, y CMA (n) and are the statistical modulus square of the CMA equalizer output signal and the modulus square of the modulation type; J CMA is the cost function of the CMA equalizer; E is the mean square error function.

7. The joint equalization method for signals in fading channels according to claim 5, characterized in that: The cost function of the DD equalizer is: Among them, y DD (n) and are the standard constellation mapping symbol points of the DD equalizer output signal and modulation type respectively; J DD is the cost function of the DD equalizer; E is the mean square error function.

8. The joint equalization method for signals in fading channels according to claim 1, characterized in that: The step 3 comprises: The IQ data before being equalized is subjected to joint equalization processing to determine whether the equalized IQ data meets the convergence condition; if the convergence condition is not met, the equalization is continued, otherwise a converged constellation diagram is obtained, that is, the equalized IQ data is obtained.

9. The joint equalization method for signals in fading channels according to claim 8, characterized in that: If the vector error between the equalized constellation diagram and the standard constellation diagram shape of the corresponding modulation mode is less than 10 within a signal-to-noise ratio of 20 dB, it is determined that the equalized IQ data meets the convergence condition.

10. The joint equalization method for signals in fading channels according to claim 9, characterized in that: The calculation formula of the vector error is: Among them, EVM RMS is the vector error, Ik is the I data after equalization, Qk is the Q data after equalization, For the standard constellation diagram reference I data, is the reference Q data of the standard constellation diagram, and N is the IQ data length.

Images