Adaptive estimation method, system, electronic device and storage medium of channel

Abstract

The application relates to the field of communication, in particular to a channel adaptive estimation method and system, an electronic device and a storage medium. The channel adaptive estimation method comprises the following steps: obtaining an error between an output signal of a preset constrained adaptive filter and an output signal of a channel to be estimated as an error signal; performing q-Rényi form transformation on the error signal to obtain a transformed error signal; calculating a to-be-updated tap coefficient vector of the constrained adaptive filter according to the transformed error signal; and if the to-be-updated tap coefficient vector meets a preset result output condition, taking the to-be-updated tap coefficient vector as a channel coefficient vector of the channel to be estimated. The purpose is to guarantee the robustness and convergence of the channel estimation process when there is a heavy-tailed impulse noise in the channel.

Description

Technical Field

[0001] This application relates to the field of communications, and in particular to an adaptive estimation method, system, electronic device, and storage medium for a channel. Background Technology

[0002] Adaptive filtering algorithms have been widely applied in adaptive channel estimation and have achieved significant progress. However, when the channel weight vector to be estimated is subject to a linear constraint, traditional adaptive filtering methods are no longer suitable for estimating such channels. Simultaneously, constrained adaptive filtering methods have emerged. Constrained channel adaptive estimation, based on the Lagrangian function, utilizes the difference between the desired channel output signal and the filtered output signal, along with the gradient descent method, to update the channel weight vector. Constrained adaptive filters have found widespread application in estimating linear constraint parameter vectors.

[0003] However, the constrained minimum mean square algorithm based on the minimum mean square criterion, which uses gradient descent to iteratively calculate the filter tap coefficients to make the filtered signal as close as possible to the channel's output signal, only shows good performance when the noise is Gaussian distributed. The more mature constrained maximum entropy algorithm uses an upper function that serves as a nonlinear local similarity measure between the channel's output signal and the filter's output signal as a cost function to calculate the filter tap coefficients. It also uses the gradient method, but its performance deteriorates when the noise is heavy-tailed impulse noise. Since background noise is non-Gaussian distributed in many practical applications, especially heavy-tailed impulse noise, it severely affects the robustness and convergence of the above algorithms. Summary of the Invention

[0004] The main objective of this application is to propose an adaptive channel estimation method, system, electronic device, and storage medium, which aims to ensure the robustness and convergence of the channel estimation process even when heavy-tailed impulse noise exists in the channel, so that the channel can still be estimated quickly using an adaptive filter even when heavy-tailed impulse noise exists in the channel.

[0005] To achieve the above objectives, embodiments of this application provide an adaptive channel estimation method, comprising: obtaining the error between the output signal of a preset constrained adaptive filter and the output signal of the channel to be estimated as an error signal; performing a q-Rényi transformation on the error signal to obtain a transformed error signal; calculating the tap coefficient vector to be updated of the constrained adaptive filter based on the transformed error signal; and if the tap coefficient vector to be updated satisfies a preset result output condition, using the tap coefficient vector to be updated as the channel coefficient vector of the channel to be estimated.

[0006] To achieve the above objectives, this application also proposes an adaptive channel estimation system, comprising: an acquisition module for acquiring the error between the output signal of a preset constrained adaptive filter and the output signal of the channel to be estimated as an error signal; a transformation module for performing a q-Rényi transformation on the error signal to acquire the transformed error signal; an iteration module for calculating the tap coefficient vector to be updated of the constrained adaptive filter based on the transformed error signal; and a result determination module for using the tap coefficient vector to be updated as the channel coefficient vector of the channel to be estimated if the tap coefficient vector to be updated satisfies a preset result output condition.

[0007] To achieve the above objectives, embodiments of this application also propose an electronic device, the device comprising:

[0008] At least one processor; and,

[0009] A memory communicatively connected to the at least one processor; wherein,

[0010] The memory stores instructions that can be executed by the at least one processor, which enables the at least one processor to perform the adaptive channel estimation method described above.

[0011] To achieve the above objectives, embodiments of this application also propose a computer-readable storage medium storing a computer program that, when executed by a processor, implements the adaptive channel estimation method described above.

[0012] Compared with the prior art, the embodiments proposed in this application obtain the output signal of the preset constrained adaptive filter and the output signal of the channel to be estimated, and then obtain the error signal based on the two output signals, thereby enabling the error signal to be transformed in the q-Rényi form. Since the q-Rényi form transformation actually compresses the large error signal and has little impact on the small error signal, it can achieve both good anti-impulse noise capability and good convergence speed. Attached Figure Description

[0013] One or more embodiments are illustrated by way of example with reference to the accompanying drawings, and these illustrative descriptions do not constitute a limitation on the embodiments.

[0014] Figure 1 This is the flowchart of the adaptive channel estimation method provided in the first embodiment of this application. Figure 1 ;

[0015] Figure 2 This is the flowchart of the adaptive channel estimation method provided in the first embodiment of this application. Figure 2 ;

[0016] Figure 3 This is a distribution diagram of the channel coefficient estimation results of the adaptive channel estimation method provided in the first embodiment of this application;

[0017] Figure 4 This is a comparison of the learning curves of the adaptive channel estimation method provided in the first embodiment of this application under impulse noise.

[0018] Figure 5 This is a comparison of the learning curves of the adaptive channel estimation method provided in the first embodiment of this application under Laplace noise;

[0019] Figure 6 This is a flowchart of the adaptive channel estimation method provided in the second embodiment of this application;

[0020] Figure 7 This is a schematic diagram of the adaptive estimation system for the channel provided in the third embodiment of this application;

[0021] Figure 8 This is a schematic diagram of the structure of the electronic device provided in the fourth embodiment of this application. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the various embodiments of this application will be described in detail below with reference to the accompanying drawings. However, those skilled in the art will understand that many technical details have been provided in the various embodiments of this application to help readers better understand this application. However, the technical solutions claimed in this application can be implemented even without these technical details and various changes and modifications based on the following embodiments. The division of the various embodiments below is for the convenience of description and should not constitute any limitation on the specific implementation of this application. The various embodiments can be combined with and referenced by each other without contradiction.

[0023] The first embodiment of this application relates to an adaptive channel estimation method, applied in the process of estimating the channel, such as... Figure 1 As shown, it specifically includes:

[0024] Step 101: Obtain the error between the output signal of the preset constrained adaptive filter and the output signal of the channel to be estimated as the error signal.

[0025] Specifically, after acquiring the output signal of the preset constrained adaptive filter and the output signal of the channel to be estimated each time, the output signal of the channel to be estimated is used as the expected output in the channel estimation process. Then, the error signal e(n) of the filter relative to the channel at the current time n is: e(n) = d(n) - y(n), where d(n) is the channel output signal at the current time n and y(n) is the filter output signal at the current time n.

[0026] It should be noted that the constrained adaptive filter is used to indicate that a constrained adaptive method is employed to adjust the filter in order to estimate the channel using the filter, not to impose restrictions on the filter. Furthermore, the filter needs to be initialized before acquiring the output signal for the first time. Specifically, the initial value of the filter's tap coefficient vector is: h(0) = C(C T C) -1 f, where C is a matrix of size M×K (M>K), and f is a vector of length K. T Let C be the transpose of matrix C, (C T C) -1 For matrix C T The inverse matrix of C, where M represents the channel length and the number of filter taps. Since the filter can achieve similar filtering effects by setting different numbers of taps and specific tap coefficients, this implementation does not limit the specific initialization process; it can be randomly selected or set based on experience.

[0027] Step 102: Perform a q-Rényi transformation on the error signal to obtain the transformed error signal.

[0028] Specifically, the transformation of the q-Rényi form is achieved through the following expression:

[0029]

[0030] Where g(e(n)) is the transformed error signal, e(n) is the error signal, and q and σ are preset transformation parameters.

[0031] Further, the transformation parameter σ is a kernel width parameter, which can affect the compression degree of the error signal by the q-Rényi form transformation. The greater the compression degree of the error signal, the smaller the adverse effect of short-time and drastic changes caused by external influences such as pulses of heavy-tailed impulse responses on the iteration, so that the iteration result remains convergent, and the channel estimation result can be obtained faster and better. At the same time, the transformation parameter q will also produce a similar effect on the above basis: when the parameter q approaches 1, the q-Rényi distribution is close to the Gaussian distribution; when 1 / 3<q<1, the peak of the q-Rényi distribution is steeper than that of the Gaussian distribution; but when q>1, the q-Rényi distribution tends to a uniform distribution. Therefore, we prefer or only consider the case of 1 / 3<q<1, because in this case, the peak of the q-Rényi distribution is steeper than that of the Gaussian distribution, which makes the anti-noise performance of the present application far better than the constrained maximum entropy criterion method.

[0032] Step 103: Calculate the tap coefficient vector to be updated of the constrained adaptive filter according to the transformed error signal.

[0033] Specifically, the tap coefficient vector to be updated of the constrained adaptive filter is calculated by the following expression:

[0034] h(n+1)=P[h(n)+μg(e(n))e(n)x(n)]+z

[0035] Wherein h(n+1) is the currently obtained tap coefficient vector to be updated, h(n) is the tap coefficient vector of the current constrained adaptive filter, g(e(n)) is the transformed error signal, e(n) is the error signal, x(n) is the recorded input signal of the current constrained adaptive filter and the channel to be estimated, P, μ, and z are preset iteration parameters. More specifically, the parameter μ is a step size parameter, the vector z=C(C T C) -1 f, the matrix P=I-C(C T C) -1 C T .

[0036] It should be noted that the expression involved in step 103 is a specific application of the gradient ascent method. Since the error signal obtained after the q-Rényi form transformation based on step 102 is used therein, large error signals caused by non-Gaussian noise or impulse noise are effectively compressed, so that good performance can be obtained in a non-Gaussian noise environment. When the impulse noise is very large, g(e(n)) approaches 0, and the weight vector is basically not updated; when there is no impulse noise interference in the system, the error signal is small, and g(e(n)) is large, which is equivalent to the step size becoming larger, so that the system weight vector can be updated quickly, and the convergence of the filter is faster.

[0037] Step 104: If the tap coefficient vector to be updated meets the preset result output conditions, the tap coefficients are used as the channel coefficient vector of the channel to be estimated.

[0038] Specifically, channel estimation is actually channel coefficient estimation. Channel coefficient estimation is achieved by continuously adjusting the tap coefficients of the adaptive filter under constraint, so that the output signal after filtering is as close as possible to the output signal of the channel to be estimated. When the filter reaches stability, we can assume that the transmission characteristics of the two are consistent, and thus we can regard the tap coefficients of the filter as an estimate of the channel coefficients.

[0039] More specifically, the output signal of the channel at time n can be represented as d(n) = x(n). T h0+v(n), where h(0) is the true channel coefficient of the channel to be estimated, and v(n) is the system noise. The output signal of the filter at time n can be expressed as y(n)=x(n). T h(n) is the filter tap coefficient vector at time n. The channel estimation process is to continuously iterate h(n) to make it approach h0 until the preset output conditions are met, and the current h(n) is taken as the result of the channel estimation.

[0040] It should be noted that the tap coefficients to be updated may not meet the preset output conditions. In this case, such as... Figure 2 As shown, it includes:

[0041] Step 201: Obtain the error between the output signal of the preset constrained adaptive filter and the output signal of the channel to be estimated as the error signal.

[0042] Specifically, step 201 is largely the same as step 101 above, so I will not go into detail here.

[0043] Step 202: Perform a q-Rényi transformation on the error signal to obtain the transformed error signal.

[0044] Specifically, step 202 is largely the same as step 102 above, so I won't go into detail here.

[0045] Step 203: Calculate the tap coefficient vector to be updated for the constrained adaptive filter based on the transformed error signal.

[0046] Specifically, step 203 is largely the same as step 103 above, so I will not go into detail here.

[0047] Step 204: Check whether the tap coefficient to be updated meets the preset result output conditions. If yes, proceed to step 205; otherwise, proceed to step 206.

[0048] Specifically, the output condition can be that the error signals of the filter and the channel reach a steady state under the condition that the tap coefficients to be updated in the filter are met, or it can be that the error of the filter and the channel is within a certain value, such as 10. -4 The result output conditions can be other constraints, which will not be elaborated here.

[0049] Step 205: Use the tap coefficients as the channel coefficient vector of the channel to be estimated.

[0050] Specifically, step 205 is largely the same as step 104 above, so I won't go into detail here.

[0051] Step 206: Adjust the constraint adaptive filter based on the currently acquired tap coefficients.

[0052] Specifically, if the conditions are not met, the currently obtained better tap coefficients should be used as the tap coefficients of the filter before the error signal is obtained in the next iteration, and the channel estimation result that meets the requirements should be continuously approximated through iteration.

[0053] Compared with the prior art, the embodiments proposed in this application obtain the output signal of the preset constrained adaptive filter and the output signal of the channel to be estimated, and then obtain the error signal based on the two output signals, thereby enabling the error signal to be transformed in the q-Rényi form. Since the q-Rényi form transformation actually compresses the large error signal and has little impact on the small error signal, it can achieve both good anti-impulse noise capability and good convergence speed.

[0054] It should be noted that the simulation experiment was conducted according to the first embodiment, and references 1 "On the mean-square performance of the constrained LMS algorithm" (R. Arablouei, K. Dogancay, Signal Processing, vol. 117, pp. 192-197, 2015) and 2 "Constrained maximum correntropy adaptive filtering" (Siyuan Peng, Badong Chen and Lei Sun, Signal Processing, vol. 140, pp. 116-126, 2017) were used as the comparison group. The simulation experiment used a channel with actual channel coefficients and a constrained adaptive filter length of 31, and the input signal had a covariance matrix R = E[x(n)x(n)]. TThe zero-mean multivariate Gaussian signal was obtained, with parameters R, C, and f selected in the same way as in Reference 1. The background noise was the impulse noise and Laplace noise simulated by the Bernoulli-Gaussian process, both of which are heavy-tailed non-Gaussian noise. The kernel width parameter in Reference 2 and the method of this invention is 5, while the parameter q in this invention is set to 0.5, resulting in the following... Figure 3 , Figure 4 and Figure 5 The results shown are as follows: Figure 3 This represents the actual weighting coefficients of the obtained channel; Figure 4 This is a comparison of the learning curves of the constrained adaptive filtering algorithm under impulse noise. The variance of the impulse noise is 4, and its probability of occurrence is 0.3. The variance of the background Gaussian noise is 0.01. To achieve the same initial convergence speed, the step sizes of the methods in Reference 1, Reference 2, and this invention are 0.006, 0.005, and 0.05, respectively. Figure 4 It can be seen that the steady-state mean square error of the present invention is smaller than that of references 1 and 2. Under strong impulse noise environment, the steady-state mean square error of the present invention is even lower than that of the method in reference 2, which has better performance. Figure 5 This is a comparison of the learning curves of the constrained adaptive filtering algorithm under Laplace noise with a variance of 4. The step sizes of the methods in Reference 1, Reference 2, and the method of this invention are 0.0022, 0.0022, and 0.03, respectively. Figure 5 It can be seen that the steady-state mean square error of the present invention is smaller than that of references 1 and 2, and the effect of the present invention is better than that of the method in reference 2 under Laplace noise environment.

[0055] The second embodiment of this application relates to an adaptive channel estimation method. This embodiment is largely the same as the first embodiment, except that the signal input process is further refined. The specific process is as follows: Figure 6 As shown:

[0056] Step 601: Randomly generate an input signal that meets preset conditions.

[0057] Specifically, the preset conditions can be conditions set according to the actual situation, such as the input signal satisfying a Gaussian distribution, or they can be generated randomly without setting any conditions. This implementation does not limit the input signal. In the actual implementation process, any kind of constraint condition on the input signal can be used.

[0058] It should be noted that, as shown in the second embodiment, during the channel estimation process, steps 301-305 may need to be executed multiple times if the tap coefficients to be updated do not meet the preset output conditions in order to achieve channel estimation. In each implementation, the input signal is randomly generated, thus avoiding the randomness caused by iterating with the same signal.

[0059] Step 602: Input the input signal into the constrained adaptive filter and the channel to be estimated, respectively.

[0060] Step 603: Obtain the error between the output signal of the preset constrained adaptive filter and the output signal of the channel to be estimated as the error signal.

[0061] Specifically, step 603 is largely the same as step 101 above, so I won't go into detail here.

[0062] Step 604: Perform a q-Rényi transformation on the error signal to obtain the transformed error signal.

[0063] Specifically, step 604 is largely the same as step 102 above, so I won't go into detail here.

[0064] Step 605: Calculate the tap coefficient vector to be updated for the constrained adaptive filter based on the transformed error signal.

[0065] Specifically, step 605 is largely the same as step 103 above, so I won't go into detail here.

[0066] Step 606: If the tap coefficient vector to be updated meets the preset result output conditions, the tap coefficients are used as the channel coefficient vector of the channel to be estimated.

[0067] Specifically, step 606 is largely the same as step 104 above, so I won't go into detail here.

[0068] Compared with the prior art, this embodiment, based on the first embodiment, uses a randomly generated signal as the input signal for the filter and the channel to be estimated, which can estimate the channel under different conditions as much as possible and improve the accuracy of the channel estimation results.

[0069] Furthermore, it should be understood that the step divisions of the various methods described above are only for clarity. In practice, they can be combined into one step or some steps can be split into multiple steps. As long as they include the same logical relationship, they are all within the scope of protection of this patent. Adding insignificant modifications or introducing insignificant designs to the algorithm or process, but without changing the core design of the algorithm and process, are also within the scope of protection of this patent.

[0070] The third embodiment of the present invention relates to an adaptive estimation system for a channel, such as... Figure 7 As shown, it includes:

[0071] The acquisition module 701 is used to acquire the error between the output signal of the preset constrained adaptive filter and the output signal of the channel to be estimated as an error signal.

[0072] The transformation module 702 is used to perform a q-Rényi transformation on the error signal to obtain the transformed error signal.

[0073] Iteration module 703 is used to calculate the tap coefficient vector to be updated for the constrained adaptive filter based on the transformed error signal.

[0074] The result determination module 704 is used to take the tap coefficient vector to be updated as the channel coefficient vector of the channel to be estimated if the tap coefficient vector to be updated meets the preset result output conditions.

[0075] It is not difficult to see that this embodiment is a device embodiment corresponding to the first embodiment, and this embodiment can be implemented in conjunction with the first embodiment. The relevant technical details mentioned in the first embodiment are still valid in this embodiment, and will not be repeated here to reduce repetition. Accordingly, the relevant technical details mentioned in this embodiment can also be applied to the first embodiment.

[0076] It is worth mentioning that all modules involved in this embodiment are logical modules. In practical applications, a logical unit can be a physical unit, a part of a physical unit, or a combination of multiple physical units. Furthermore, to highlight the innovative aspects of this invention, this embodiment does not introduce units that are not closely related to solving the technical problem proposed by this invention; however, this does not mean that other units are absent from this embodiment.

[0077] The fourth embodiment of this application relates to an electronic device, such as... Figure 8 As shown, it includes: at least one processor 801; and a memory 802 communicatively connected to at least one processor 801; wherein the memory 802 stores instructions executable by at least one processor 801, which are executed by at least one processor 801 to enable at least one processor 801 to perform the adaptive estimation method of the channel described in any of the above method embodiments.

[0078] The memory 802 and processor 801 are connected via a bus, which can include any number of interconnecting buses and bridges. The bus connects various circuits of one or more processors 801 and memory 802 together. The bus can also connect various other circuits, such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. A bus interface provides an interface between the bus and the transceiver. The transceiver can be a single element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by processor 801 is transmitted over a wireless medium via an antenna, which further receives data and transmits it to processor 801.

[0079] The processor 801 is responsible for managing the bus and general processing, and can also provide various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. The memory 802 can be used to store data used by the processor 801 during operation.

[0080] The fifth embodiment of the present invention relates to a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the above-described method embodiments.

[0081] That is, those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0082] Those skilled in the art will understand that the above embodiments are specific embodiments for implementing this application, and in practical applications, various changes can be made to them in form and detail without departing from the spirit and scope of this application.

Claims

1. An adaptive estimation method for a channel, characterized in that, include: The error signal is obtained by comparing the output signal of the preset constrained adaptive filter with the output signal of the channel to be estimated. The error signal is transformed using the q-Rényi form to obtain the transformed error signal; The vector of tap coefficients to be updated for the constrained adaptive filter is calculated based on the transformed error signal. If the tap coefficient vector to be updated meets the preset result output conditions, the tap coefficient vector to be updated is used as the channel coefficient vector of the channel to be estimated.

2. The method according to claim 1, characterized in that, The step of performing a q-Rényi transformation on the error signal to obtain the transformed error signal includes: The error signal is transformed in q-Rényi form using the following expression: Where g(e(n)) is the transformed error signal, e(n) is the error signal, and q and σ are preset transformation parameters.

3. The method according to claim 2, characterized in that, The preset transformation parameter q is located in the interval Inside.

4. The method according to claim 1, characterized in that, Also includes: Obtain the initial tap coefficient vector of the constrained adaptive filter, wherein the initial tap coefficient vector h(0) = C(C T C) -1 f is a matrix of size M×K in C, and f is a vector of length K. T Let C be the transpose of matrix C, (C T C) -1 Let M be the inverse of matrix CTC, M be the channel length and the number of filter taps, and K be a constant satisfying M>K. The constrained adaptive filter is initialized based on the initial tap coefficient vector.

5. The method according to claim 4, characterized in that, The step of calculating the tap coefficient vector to be updated for the constrained adaptive filter based on the transformed error signal is specifically performed using the following expression: h(n+1)=P[h(n)+μg(e(n))e(n)x(n)]+z Where h(n+1) is the currently acquired tap coefficient vector to be updated, h(n) is the tap coefficient vector of the current constrained adaptive filter, g(e(n)) is the transformed error signal, e(n) is the error signal, x(n) is the recorded input signal of the current constrained adaptive filter and the channel to be estimated, μ is the preset step size parameter, and vector z = C(C T C) -1 f, matrix P = IC(C T C) -1 C T , where I is the identity matrix.

6. The method according to claim 1, characterized in that, Before obtaining the error signal between the output signal of the preset constrained adaptive filter and the output signal of the channel to be estimated, the method further includes: Randomly generate an input signal that meets preset conditions; The input signal is input to the constrained adaptive filter and the channel to be estimated, respectively.

7. The method according to claim 1, characterized in that, Before obtaining the error signal between the output signal of the preset constrained adaptive filter and the output signal of the channel to be estimated, the method further includes: If the tap coefficient to be updated does not meet the preset result output conditions, the constraint adaptive filter is adjusted according to the currently acquired tap coefficient, and the error between the output signal of the preset constraint adaptive filter and the output signal of the channel to be estimated is used as the error signal in the next acquisition.

8. An adaptive estimation system for a channel, characterized in that, include: The acquisition module is used to acquire the error between the output signal of the preset constrained adaptive filter and the output signal of the channel to be estimated as an error signal; The transformation module is used to perform a q-Rényi transformation on the error signal to obtain the transformed error signal; An iterative module is used to calculate the tap coefficient vector to be updated of the constrained adaptive filter based on the transformed error signal. The result determination module is used to take the tap coefficient vector to be updated as the channel coefficient vector of the channel to be estimated if the tap coefficient vector to be updated meets the preset result output conditions.

9. An electronic device, characterized in that, include: At least one processor; as well as, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the adaptive estimation method for the channel as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the adaptive estimation method for the channel according to any one of claims 1 to 7.

Images