Filter coefficient storage method and system

Abstract

The application provides a filter coefficient storage method and system, wherein the storage method comprises the following steps: constructing a filter sharing coefficient for the MMSE filter coefficient of the same passband center frequency point in a filtering device according to the MMSE filter coefficient of the same passband center frequency point; fitting a plurality of filter fitting coefficients according to the filter sharing coefficient and the signal correlation matrix of each MMSE filter coefficient, wherein the filter fitting coefficients correspond to the MMSE filter coefficients one by one; and storing the filter sharing coefficient and the filter fitting coefficient as the filter coefficient of the filtering device in the SFN scene. The application can reduce the storage space of the system in various filtering scenes.

Description

Technical Field

[0001] This invention relates to the field of wireless communication technology, and in particular to a method and system for storing filter coefficients. Background Technology

[0002] The 3GPP (Third Generation Partnership Project) defines the HST-SFN (High Speed ​​Train - Single Frequency Network) scenario, where a UE (User Equipment) on a train traveling on a high-speed track can simultaneously receive the same signal from four base stations. Its channel power spectrum has significant characteristics, namely, it can be divided into two parts: the main spectrum, whose frequency domain position (doppler) changes with the UE's position; and the edge spectrum, with relatively lower power. The performance of channel estimation directly affects the demodulation performance of the receiver.

[0003] In existing technologies, to reduce computational complexity, channel estimation is typically divided into three stages: first, an initial estimate of the reference signal is obtained; then, the reference signal is filtered; and finally, the data channel estimate is obtained through interpolation. Filtering the reference signal is a crucial step in channel estimation, and MMSE (Minimum Mean Square Error) filter coefficients are usually employed. The calculation of MMSE filter coefficients can be reduced using the Singular Value Decomposition (SVD) algorithm, saving computational load and storage space.

[0004] However, for HST-SFN scenarios, the MMSE filter coefficients are not only related to resource distribution and channel parameters, but also to the relative positions of the UE and the base station on a train traveling on a high-speed track. Existing technologies either sacrifice performance or increase computational load or system storage space to achieve the estimation of the reference signal.

[0005] Therefore, how to further reduce the amount of computation and save system storage space has become an urgent problem to be solved. Summary of the Invention

[0006] To address the aforementioned problems, the filter coefficient storage method and system provided by this invention can save system storage space.

[0007] In a first aspect, the present invention provides a method for storing filter coefficients, the method comprising:

[0008] Based on the MMSE filter coefficients with the same passband center frequency in the filtering device, construct a filter sharing coefficient for each MMSE filter coefficient with the same passband center frequency.

[0009] Based on the filter sharing coefficients and the signal correlation matrix of each MMSE filter coefficient, multiple filter fitting coefficients are fitted, and the filter fitting coefficients correspond one-to-one with the MMSE filter coefficients.

[0010] The filter sharing coefficients and filter fitting coefficients are stored as filter coefficients of the filtering device in the SFN scenario.

[0011] Optionally, the step of constructing a filter-shared coefficient for each MMSE filter coefficient with the same passband center frequency based on the MMSE filter coefficients with the same passband center frequency in the filtering device includes:

[0012] Among the MMSE filter coefficients with the same passband center frequency, the unitary matrix obtained by SVD decomposition of the signal correlation matrix corresponding to the MMSE filter coefficient with the smallest passband bandwidth is used as the filter shared coefficient.

[0013] Optionally, the step of fitting multiple filter fitting coefficients based on the filter shared coefficients and the signal correlation matrix of each MMSE filter coefficient includes:

[0014] Among the MMSE filter coefficients with the same passband center frequency, the filter-shared coefficients corresponding to the MMSE filter coefficients with the smallest passband bandwidth are fitted to the signal correlation matrices of the MMSE filter coefficients with other passband bandwidths, so as to fit the corresponding filter fitting coefficients for the MMSE filter coefficients with other passband bandwidths.

[0015] The signal correlation matrix corresponding to the MMSE filter coefficients with the smallest passband bandwidth is obtained by SVD decomposition, and the resulting diagonal matrix is ​​the filter fitting coefficient corresponding to the MMSE filter coefficients with the smallest passband bandwidth.

[0016] Optionally, the step of fitting multiple filter fitting coefficients based on the filter shared coefficients and the signal correlation matrix of each MMSE filter coefficient includes:

[0017] Iterate through each filter shared coefficient and fit a set of pre-selected fitting coefficients for each MMSE filter coefficient based on each filter shared coefficient and the corresponding signal correlation matrix. Each set of pre-selected fitting coefficients includes multiple pre-selected fitting coefficients, and the pre-selected fitting coefficient set corresponds one-to-one with the filter shared coefficient.

[0018] In each group of pre-selected fitting coefficients, the pre-selected fitting coefficient with the best fit is selected as the corresponding filter fitting coefficient.

[0019] The method also includes storing indices of the optimal prefit coefficients and the corresponding filter-shared coefficients.

[0020] Optionally, the non-SFN MMSE filter coefficients include: non-SFN overlapping filter coefficients and non-SFN new filter coefficients; the non-SFN MMSE filter coefficients are the MMSE filter coefficients of the filter device in the non-SFN scenario; the non-SFN overlapping filter coefficients are the same coefficients as the MMSE filter coefficients of the filter device in the SFN scenario, and the non-SFN new filter coefficients are coefficients that are different from the MMSE filter coefficients of the filter device in the SFN scenario; the SFN scenario includes the high-speed rail single-frequency network channel scenario, and the non-SFN scenario includes: EPA channel scenario, EVA channel scenario, ETU channel scenario, TDL channel scenario, and CDL channel scenario;

[0021] Storage methods also include:

[0022] Iterate through each filter shared coefficient and fit a set of pre-selected non-SFN fitting coefficients for each non-SFN new filter coefficient based on each filter shared coefficient. Each set of pre-selected non-SFN fitting coefficients includes multiple pre-selected non-SFN fitting coefficients, and each pre-selected non-SFN fitting coefficient in each set corresponds one-to-one with the filter shared coefficient.

[0023] In each group of pre-selected non-SFN fitting coefficients, the pre-selected non-SFN fitting coefficient with the best fit is selected as the corresponding new non-SFN fitting coefficient.

[0024] The norms corresponding to multiple non-SFN new fitted coefficients are compared with the threshold. The signal correlation matrix of the non-SFN new filter coefficients with norms greater than the threshold is decomposed by SVD. The resulting unitary matrix is ​​used as the non-SFN filter shared coefficients corresponding to the non-SFN new filter coefficients. The threshold is determined by the largest norm among all the norms corresponding to the MMSE filter coefficients that the filtering device responds to in the SFN scenario.

[0025] The signal correlation matrix of the non-SFN new filter coefficients with the corresponding norm greater than the threshold is decomposed by SVD, and the resulting diagonal matrix is ​​used as the non-SFN filter fitting coefficients corresponding to the non-SFN new filter coefficients with the corresponding norm greater than the threshold.

[0026] The non-SFN new fitting coefficients with a corresponding norm less than or equal to the threshold are taken as the non-SFN filter fitting coefficients corresponding to the non-SFN new filter coefficients with a corresponding norm less than or equal to the threshold.

[0027] Stores the non-SFN filter shared coefficients, non-SFN filter fitting coefficients, and indices of non-SFN MMSE filter coefficients and their corresponding filter shared coefficients or corresponding non-SFN filter shared coefficients, as well as indices of filter fitting coefficients corresponding to non-SFN overlapping filter coefficients.

[0028] Secondly, the present invention provides a filter coefficient storage system, the storage system comprising:

[0029] The construction module is configured to construct a filter-shared coefficient for each MMSE filter coefficient with the same passband center frequency, based on the MMSE filter coefficients with the same passband center frequency in the filtering device.

[0030] The fitting module is configured to fit multiple filter fitting coefficients based on the filter shared coefficients and the signal correlation matrix of each MMSE filter coefficient, with each filter fitting coefficient corresponding to a MMSE filter coefficient.

[0031] The storage module is configured to store the filter sharing coefficients and filter fitting coefficients as filter coefficients of the filtering device.

[0032] Optionally, the building module is also configured to use the unitary matrix obtained by SVD decomposition of the signal correlation matrix corresponding to the MMSE filter coefficient with the smallest passband bandwidth among the MMSE filter coefficients at the same passband center frequency as the filter shared coefficients.

[0033] Optionally, the fitting module includes:

[0034] The fitting submodule is configured to fit the filter-shared coefficients corresponding to the MMSE filter coefficients with the smallest passband bandwidth to the signal correlation matrix of the MMSE filter coefficients with other passband bandwidths among the MMSE filter coefficients with the same passband center frequency, thereby fitting the corresponding filter fitting coefficients for the MMSE filter coefficients with other passband bandwidths.

[0035] The decomposition submodule is configured such that the signal correlation matrix corresponding to the MMSE filter coefficients with the smallest passband bandwidth is decomposed by SVD, and the resulting diagonal matrix is ​​the filter fitting coefficient corresponding to the MMSE filter coefficients with the smallest passband bandwidth.

[0036] Optionally, the fitting module also includes:

[0037] The traversal submodule is configured to traverse each filter shared coefficient and fit a set of pre-selected fitting coefficients for each MMSE filter coefficient based on each filter shared coefficient and the corresponding signal correlation matrix. Each set of pre-selected fitting coefficients includes multiple pre-selected fitting coefficients, and the pre-selected fitting coefficient set corresponds one-to-one with the filter shared coefficient.

[0038] The selection submodule is configured to select the best-fitting pre-fit coefficient from each group of pre-fit coefficients as the corresponding filter fit coefficient.

[0039] The storage module is also configured to store indices of the optimal pre-fit coefficients and the corresponding filter-shared coefficients.

[0040] Optionally, the non-SFN MMSE filter coefficients include: non-SFN overlapping filter coefficients and non-SFN new filter coefficients; the non-SFN MMSE filter coefficients are the MMSE filter coefficients of the filter device in the non-SFN scenario; the non-SFN overlapping filter coefficients are the same coefficients as the MMSE filter coefficients of the filter device in the SFN scenario, and the non-SFN new filter coefficients are coefficients that are different from the MMSE filter coefficients of the filter device in the SFN scenario; the SFN scenario includes the high-speed rail single-frequency network channel scenario, and the non-SFN scenario includes: EPA channel scenario, EVA channel scenario, ETU channel scenario, TDL channel scenario, and CDL channel scenario;

[0041] The storage system also includes:

[0042] The traversal module is configured to traverse each filter shared coefficient and fit a set of pre-selected non-SFN fitting coefficients for each non-SFN new filter coefficient based on each filter shared coefficient. Each set of pre-selected non-SFN fitting coefficients includes multiple pre-selected non-SFN fitting coefficients, and each pre-selected non-SFN fitting coefficient in each set corresponds one-to-one with the filter shared coefficient.

[0043] The selection module is configured to select the best-fitting pre-selected non-SFN fitting coefficient from each group of pre-selected non-SFN fitting coefficients as the corresponding new non-SFN fitting coefficient.

[0044] The comparison module is configured to compare the norms of multiple non-SFN new fitted coefficients with a threshold. The signal correlation matrix of the non-SFN new filter coefficients with norms greater than the threshold is decomposed by SVD, and the resulting unitary matrix is ​​used as the non-SFN filter shared coefficients corresponding to the non-SFN new filter coefficients. The threshold is determined by the largest norm among all the norms of the MMSE filter coefficients that the filtering device responds to in the SFN scenario.

[0045] The first corresponding module is configured to perform SVD decomposition on the signal correlation matrix of the non-SFN new filter coefficients with the corresponding norm greater than the threshold, and the resulting diagonal matrix is ​​used as the non-SFN filter fitting coefficients corresponding to the non-SFN new filter coefficients with the corresponding norm greater than the threshold.

[0046] The second corresponding module is configured to use the non-SFN new fitting coefficients with a corresponding norm less than or equal to the threshold as the non-SFN filter fitting coefficients corresponding to the non-SFN new filter coefficients with a corresponding norm less than or equal to the threshold.

[0047] The storage module is also configured to store the non-SFN filter shared coefficients, non-SFN filter fitting coefficients corresponding to the non-SFN new filter coefficients, and the indices of the non-SFN MMSE filter coefficients and their corresponding filter shared coefficients or their corresponding non-SFN filter shared coefficients, as well as the indices of the filter fitting coefficients corresponding to the non-SFN overlapping filter coefficients.

[0048] Thirdly, the present invention provides an electronic device, the electronic device comprising:

[0049] At least one processor; and

[0050] A memory that is communicatively connected to at least one processor; wherein,

[0051] The memory stores instructions that can be executed by at least one processor, such that the at least one processor is able to perform the method of any of the first aspects.

[0052] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions that, when executed by a processor, implement the method as described in any of the first aspects.

[0053] The filter coefficient storage method and system provided in this invention construct filter shared coefficients and filter fitting coefficients, and store the filter shared coefficients and filter fitting coefficients as filter coefficients of the filtering device in the SFN scenario, thereby greatly reducing the occupation of system storage space without losing filtering performance. Attached Figure Description

[0054] To more clearly illustrate the technical solutions in the embodiments of this application or the conventional technology, the drawings used in the description of the embodiments or the conventional technology will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0055] Figure 1This is a schematic diagram of the response of each filter in an ideal state of a reference device according to an embodiment of this application;

[0056] Figure 2 This is a schematic diagram of a high-speed train traveling at a quasi-constant speed from RRH1 to RHH2 according to an embodiment of this application;

[0057] Figures 3 to 10 This is a power spectrum diagram of a user equipment relative to a base station at different locations according to an embodiment of this application;

[0058] Figure 11 This is a schematic diagram of the response of each filter in an ideal state of a reference device according to an embodiment of this application;

[0059] Figure 12 This is a schematic flowchart illustrating a method for storing filter coefficients according to an embodiment of this application. Detailed Implementation

[0060] To facilitate understanding of this application, a more complete description will be provided below with reference to the accompanying drawings, which illustrate embodiments of the present application. However, the present application can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that the disclosure of this application will be thorough and complete.

[0061] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application.

[0062] When used herein, the singular forms of “a,” “an,” and “the” may also include the plural forms unless the context clearly indicates otherwise. It should also be understood that the terms “comprising,” “including,” or “having,” etc., specify the presence of the stated feature, whole, step, operation, component, part, or combination thereof, but do not preclude the possibility of the presence or addition of one or more other features, wholes, steps, operations, components, parts, or combinations thereof.

[0063] In a first aspect, an embodiment of the present invention provides a reference signal estimation method, which is applied to the user equipment (UE) side and includes steps S101 to S102.

[0064] Step S101: Obtain the reference signal sent by the base station.

[0065] Step S102: Based on the absolute value of the frequency domain position of the main spectrum in the power spectrum of the reference signal and the maximum Doppler frequency shift Fd in the current scene, select a target filter coefficient in the filtering device and apply it to the filtering device to estimate the reference signal.

[0066] Among them, the main spectrum is the spectrum with the highest power among multiple power spectra.

[0067] The filtering device includes multiple MMSE filter coefficients, which are divided into n groups based on the different maximum Doppler frequency shifts (fd) of the reference signal in multiple preset scenarios, where n is a positive integer greater than 1. It is understandable that for a user equipment (UE) in a mobile state, the received reference signal will change as the UE moves, affecting the signal power, Doppler frequency shift, and time delay of signals transmitted from various base stations to the UE. The maximum Doppler frequency shift is the Doppler frequency shift of the reference signal with the greatest degree of frequency shift in a given scenario.

[0068] This invention, by setting n sets of filter coefficients in the filtering device, enables the estimation method to be applicable to various scenarios. In this embodiment, the estimation method is applied to a scenario where a high-speed train is moving, that is, a scenario where the user equipment is located on a high-speed train. Fd varies depending on the speed of the high-speed train.

[0069] For n groups of filter coefficients, each group of filter coefficients includes at least one MMSE filter coefficient. The passband bandwidth of the MMSE filter coefficients in the same group is different from each other. The MMSE filter coefficients with the same maximum Doppler frequency shift are in the same group. The center frequency of the passband of the MMSE filter coefficients in the same group decreases as the passband bandwidth increases.

[0070] The larger the value of m in the m-th filter coefficient group, the smaller the maximum Doppler frequency shift in the preset scene corresponding to the m-th filter coefficient group. The target filter coefficient is the MMSE filter coefficient in the i-th filter coefficient group whose passband includes the main spectrum and the corresponding maximum Doppler frequency shift, and whose passband bandwidth is the smallest. The maximum Doppler frequency shift fd corresponding to the i-th filter coefficient group is not less than the maximum Doppler frequency shift Fd in the current scene, and the maximum Doppler frequency shift fd corresponding to the (i+1)-th filter coefficient group is less than the maximum Doppler frequency shift Fd in the current scene. The passband edge corresponding to the target filter coefficient is greater than the maximum Doppler frequency shift Fd in the current scene.

[0071] The reference signal estimation method provided in this embodiment estimates the reference signal by selecting target filter coefficients in the filtering device whose passband contains the main spectrum and has the smallest passband bandwidth, based on the relationship between the absolute value of the frequency domain position of the main spectrum in the power spectrum of the reference signal and the magnitude of the maximum Doppler frequency shift in the current scene. This ensures that the main spectrum is always within the filter response passband and that the passband bandwidth is as narrow as possible, thereby achieving greater noise suppression and improving the system's performance in estimating the reference signal. In particular, this invention simplifies the filter response, making the system's filtering more robust.

[0072] Furthermore, the maximum Doppler frequency shift fd1 corresponding to the first group of filter coefficients is less than or equal to x. x is an integer greater than 1. It should be noted that... Designed based on the maximum Doppler frequency shift fd1 corresponding to the first group of filter coefficients, ensuring x ≥fd1. The distribution of the most-band edges of each filter coefficient group can be... 4 , 3.5 , 3 , 2.5 , 2 , 1.5 Alternatively, the design can be tailored to the specific application scenario, considering both the maximum and minimum flow deflection (Fd) values, to balance performance and complexity. For example, in a simplified filter design, the distribution of the maximum band edges of each filter coefficient group could be... 4 , 3 , 2 , 1.5 , of which 4 ≥fd1. The distribution of the most-band edges of each filter coefficient group in a finely designed filter device is as follows: 4 , 3.5 , 3 , 2.5 , 2 , 1.5 , 1 , of which 5 ≥fd1.

[0073] It should be noted that the largest Fd in the application scenario is the maximum Doppler frequency shift corresponding to the high-speed train moving at its maximum speed, and the smallest Fd is the maximum Doppler frequency shift corresponding to the high-speed train moving at its minimum speed.

[0074] In this implementation, x is 4, n is 6, and the MMSE filter coefficients are real-coefficient MMSE filter coefficients. For the six groups of filter coefficients, the maximum edges of groups 1 to 6 are respectively... 4 , 3.5 , 3 , 2.5 , 2 , 1.5 Its frequency domain response is as follows Figure 1 As shown, 15 MMSE coefficients are required; the minimum passband bandwidth in each filter coefficient group is [value missing]. In each group of filter coefficients, the passband bandwidth of all filters is... The tolerance is gradually increased until it is as close as possible to or equal to the corresponding maximum Doppler frequency shift fd. For example, in the filter coefficient groups of groups 1, 3, and 5, the maximum passband bandwidth is equal to the corresponding maximum Doppler frequency shift fd, and in the filter coefficient groups of groups 2, 4, and 6, the maximum passband bandwidth is as close as possible to the corresponding maximum Doppler frequency shift fd.

[0075] The number of MMSE filter coefficients in each filter coefficient group is determined by the maximum Doppler shift fd corresponding to the filter coefficient group and... The quotient is determined. Among them, the maximum Doppler frequency shift corresponding to the filter coefficient group and... When the quotient is an integer, the quotient obtained is the number of MMSE filter coefficients in the filter coefficient group; the maximum Doppler frequency shift corresponding to the filter coefficient group and When the quotient is not an integer, the integer quotient corresponding to the obtained quotient, or the integer quotient corresponding to the obtained quotient plus 1, is the number of MMSE filter coefficients in the filter coefficient group.

[0076] In this embodiment, the maximum Doppler frequency shift fd corresponding to the filter coefficient group and When the quotient is not an integer, the integer quotient corresponding to the obtained quotient is the number of MMSE filter coefficients in the filter coefficient set. Combined with... Figure 1 For example, the maximum Doppler frequency shift fd corresponds to the first group of filter coefficients. In the second group of filter coefficients, fd equals 3.5. , and If the integer quotient is 3, then the number of filters in the second group of filter coefficients is 3, namely W. 21 W 32 and W 43 .

[0077] Specifically, in Figure 1 In the first group of filter coefficients, the filters are W... 11 W 22 W 33 and W 44 The filters in the third group of filter coefficients are W... 31 W 42 and W 53 The filters in the fourth group of filter coefficients are W... 41 and W 52 The filters in the fifth group of filter coefficients are W... 51 and W 62 The filters in the sixth group of filter coefficients are W... 61 .

[0078] Understandably, the step of selecting a target filter coefficient in the filtering device to estimate the reference signal based on the absolute value of the frequency domain position of the main spectrum in the power spectrum of the reference signal and the maximum Doppler frequency shift in the current scene includes determining the filter coefficient group in which the target filter coefficient is located based on the maximum Doppler frequency shift in the current scene, and then determining the target filter coefficient in the selected filter coefficient group based on the absolute value of the frequency domain position of the main spectrum in the power spectrum of the reference signal.

[0079] Furthermore, the step of estimating the reference signal by selecting a target filter coefficient in the filtering device based on the absolute value of the frequency domain position of the main spectrum in the power spectrum of the reference signal and the maximum Doppler frequency shift in the current scene includes steps S1021 and S1022.

[0080] Step S1021: Form a range interval between the maximum Doppler frequency shift and 0 Hz corresponding to each group of filter coefficients, and divide the corresponding range interval into at least one selection interval according to the number of MMSE filter coefficients in each group of filter coefficients.

[0081] Among them, the range of the selected interval is no greater than The selected interval corresponds one-to-one with the MMSE filter coefficients in the corresponding filter coefficient group; in each filter coefficient group, the larger the passband bandwidth of the MMSE filter coefficient, the larger the minimum value in the corresponding selected interval.

[0082] It should be noted that within the range, it cannot be... When divisible, the maximum value in each selection interval, the range of the largest selection interval is less than... The range of the remaining selection intervals is... .

[0083] Step S1022: Compare the absolute value of the frequency domain position of the main spectrum in the power spectrum of the reference signal with the selection interval corresponding to the i-th group of filter coefficients, and select the MMSE filter coefficients corresponding to the selection interval containing the absolute value of the frequency domain position of the main spectrum as the target filter coefficients.

[0084] In some filter coefficient groups, there are no MMSE filter coefficients with a passband bandwidth including 0 Hz. That is, some filter coefficient groups do not have MMSE filter coefficients with a passband bandwidth equal to the maximum Doppler frequency shift corresponding to the m-th filter coefficient group. When it is necessary to select target filter coefficients in this filter coefficient group, it is necessary to select MMSE filter coefficients with a passband bandwidth including 0 Hz from other groups for use.

[0085] In a further optional embodiment of this embodiment, when there is no MMSE filter coefficient in the m-th filter coefficient group with a passband bandwidth equal to the maximum Doppler shift corresponding to the m-th filter coefficient group, the passband bandwidth of the ms-th filter coefficient group, including the MMSE filter coefficient with a passband bandwidth equal to the maximum Doppler shift corresponding to the ms-th filter coefficient group, is also used as an MMSE filter coefficient in the m-th filter coefficient group for the selection of target filter coefficients; s is a positive integer, and the ms-th filter coefficient group is a filter coefficient group adjacent to the m-th filter coefficient group that has a passband bandwidth equal to the maximum Doppler shift corresponding to the m-th filter coefficient group.

[0086] Combination Figure 1 In this embodiment, s is 1, but it is not limited to this. For example, there is no MMSE filter coefficient in the fourth group of filter coefficients whose passband bandwidth is equal to the maximum Doppler shift corresponding to the fourth group of filter coefficients. In this case, the passband bandwidth of both the third and first groups of filter coefficients includes MMSE filter coefficients whose passband bandwidth is equal to the corresponding maximum Doppler shift. However, since the third group of filter coefficients and the fourth group of filter coefficients are adjacent, W in the third group of filter coefficients is... 53 The target filter coefficient is selected as one of the MMSE filter coefficients in the fourth group of filter coefficients.

[0087] Similarly, if the target filter coefficients are selected in the second group of filter coefficients, W in the first group of filter coefficients... 44This will be used as an MMSE filter coefficient in the second group of filter coefficients for selecting the target filter coefficient. If the target filter coefficient is selected in the sixth group of filter coefficients, W in the seventh group of filter coefficients... 44 This will be used as one of the MMSE filter coefficients in the 6th group of filter coefficients for the selection of target filter coefficients.

[0088] Furthermore, the difference in the maximum Doppler frequency shift corresponding to adjacent filter coefficient groups is equal, such that the difference in the maximum Doppler frequency shift corresponding to adjacent filter coefficient groups is less than or equal to... .

[0089] In this embodiment, the maximum Doppler frequency shift fd corresponding to adjacent filter coefficient groups differs by 0.5 units. The difference is And for For two sets of filter coefficients that are integer multiples of each other, there exists only one set of filter coefficients corresponding to the maximum Doppler frequency shift, and the maximum Doppler frequency shift corresponding to it is not... Therefore, s is 1 in this embodiment, which is an integer multiple of s.

[0090] The passband center frequency of the MMSE filter coefficients is determined by the difference between the maximum Doppler frequency shift corresponding to the filter coefficient group to which the MMSE filter coefficients belong and half the passband bandwidth of the MMSE filter coefficients. That is, the passband center frequency fm of the MMSE filter coefficients satisfies the following formula:

[0091] fm=fd-fw / 2

[0092] Where fw is the passband bandwidth of the MMSE filter coefficients.

[0093] In this embodiment, the specific selection process for the target filter coefficients is as follows:

[0094] When the high-speed train where the UE is located has the highest speed, then 4 ≥Fd>3.5 The MMSE filter coefficients corresponding to different frequency domain responses of the UE and the base station at different relative positions are the first group of filter coefficients, which are W... 11 W 22 W 33 and W 44 .

[0095] Specifically, when the UE is located between two base stations, and the distance between the UE and the two base stations is approximately the same, that is, the absolute value of the frequency domain position of the main spectrum is ≥3... When using W 11As the target filter coefficient; when the UE is located between two base stations, and is relatively close to one of the two base stations, i.e., 3 The absolute value of the frequency domain position of the main spectrum is ≥2. When using W 22 As the target filter coefficient; when the UE is located between two base stations, and is very close to one of the two base stations, i.e., 2 The absolute value of the frequency domain position of the main spectrum is ≥1. When using W 33 As the target filter coefficient; when the absolute value of the frequency domain position of the main spectrum under a certain base station of the UE is <1 When using W 44 As target filter coefficients.

[0096] In other words, during the selection of target filter coefficients, the filter coefficient group containing the target filter coefficient is first selected based on the maximum Doppler frequency shift Fd in the current scenario. Then, based on the absolute value of the frequency domain position of the main spectrum, i.e., the position change of the UE relative to the base station, the target filter coefficient is selected from the selected filter coefficient group to estimate the reference signal. If there is only one filter coefficient in the selected filter coefficient group, that filter coefficient is directly selected to estimate the reference signal.

[0097] More specifically, assume the UE application scenario has the following characteristics: the distance between adjacent base stations is 700 meters, the closest distance between the UE and a base station is 150 meters, the high-speed train travels at a speed of 500 kilometers per hour, and the maximum Doppler frequency shift in the current scenario is 870 Hz. Combined with... Figure 2 As the high-speed train travels at a quasi-constant speed from RRH (Remote Radio Head) 1 to RHH2, its power spectrum changes as follows: Figures 3 to 10 As shown, where Figures 3 to 10 In this context, the spectrum with the highest peak value is the dominant spectrum. In this scenario, the design... =225Hz, obviously 4 =900Hz≥fd=870Hz.

[0098] The power spectrum is obtained by power spectrum estimation of the reference signal. Figure 3 At that time, the frequency domain position of the main spectrum was -74.9847Hz, and the absolute value of the frequency domain position of the main spectrum was <1. At this point, use W 44 Estimate the reference signal;

[0099] The power spectrum is obtained by power spectrum estimation of the reference signal. Figure 4 At that time, the frequency domain position of the main spectrum was -299.986Hz, 2 The absolute value of the frequency domain position of the main spectrum is ≥1. At this point, use W 33 Estimate the reference signal;

[0100] The power spectrum is obtained by power spectrum estimation of the reference signal. Figure 5 At that time, the frequency domain position of the main spectrum was -589.987Hz, 3 The absolute value of the frequency domain position of the main spectrum is ≥2. At this point, use W 22 Estimate the reference signal;

[0101] The power spectrum is obtained by power spectrum estimation of the reference signal. Figure 6 At that time, the frequency domain position of the main spectrum was -719.988Hz, 4 The absolute value of the frequency domain position of the main spectrum is ≥3. At this point, use W 11 Estimate the reference signal;

[0102] The power spectrum is obtained by power spectrum estimation of the reference signal. Figure 7 At that time, the frequency domain position of the main spectrum was 785.019Hz, 4 The absolute value of the frequency domain position of the main spectrum is ≥3. At this point, use W 11 Estimate the reference signal;

[0103] The power spectrum is obtained by power spectrum estimation of the reference signal. Figure 8 At that time, the frequency domain position of the main spectrum was 572.5185Hz, 3 The absolute value of the frequency domain position of the main spectrum is ≥2. At this point, use W 22 Estimate the reference signal;

[0104] The power spectrum is obtained by power spectrum estimation of the reference signal. Figure 9 At that time, the frequency domain position of the main spectrum was 322.517Hz, 2 The absolute value of the frequency domain position of the main spectrum is ≥1. At this point, use W 33 Estimate the reference signal;

[0105] The power spectrum is obtained by power spectrum estimation of the reference signal. Figure 10 At that time, the frequency domain position of the main spectrum was -42.4846Hz, and the absolute value of the frequency domain position of the main spectrum was <1. At this point, use W 44 Estimate the reference signal.

[0106] Similarly, when the speed of the high-speed train where the UE is located reaches 3.5... ≥Fd>3 At that time, the MMSE filter coefficients of different frequency domain responses corresponding to different relative positions of the UE and the base station are the second group of filter coefficients, which are W 21 W 32 and W 43 Meanwhile, since the second group of filter coefficients does not contain any MMSE filter coefficients with a passband bandwidth including 0 Hz, the MMSE filter coefficients W from the first group of filter coefficients with a passband bandwidth including 0 Hz are then... 44 , which are selected as filters in the second set of filter coefficients.

[0107] Specifically, when the UE is located between two base stations and at approximately the same distance from both base stations, the absolute value of the frequency domain position of the main spectrum is ≥3. At this time, use W 21 Estimate the reference signal; when the UE is located between two base stations, and is relatively close to one of the two base stations, 3 The absolute value of the frequency domain position of the main spectrum is ≥2. At this time, use W 32 Estimate the reference signal; when the UE is located between two base stations and is very close to one of the two base stations, 2 The absolute value of the frequency domain position of the main spectrum is ≥1. At this time, use W 43 Estimate the reference signal; when the UE is under a certain base station, the absolute value of the frequency domain position of the main spectrum is <1. At this time, use W 44 Estimate the reference signal.

[0108] When the speed of the high-speed train where UE is located makes 3 ≥Fd>2.5 The MMSE filter coefficients for different frequency domain responses corresponding to different relative positions of the UE and the base station are respectively the third group of filter coefficients, namely W 31 W 42 and W 53 .

[0109] Specifically, when the UE is located between two base stations and at approximately the same distance from both base stations, the absolute value of the frequency domain position of the main spectrum is ≥2. At this time, use W 31 Estimate the reference signal; when the UE is located between two base stations, and is relatively close to one of the two base stations, 2 The absolute value of the frequency domain position of the main spectrum ≥ At this time, use W 42 Estimate the reference signal; when the UE is under a certain base station, the absolute value of the frequency domain position of the main spectrum is < At this time, use W53 Estimate the reference signal.

[0110] Similarly, when the speed of the high-speed train where the UE is located makes 2.5 ≥Fd>2 At that time, the MMSE filter coefficients for different frequency domain responses corresponding to different relative positions of the UE and the base station are W 41 W 52 and W 53 ;

[0111] When the UE is on a high-speed train, it makes 2 ≥fd>1.5 At that time, the MMSE filter coefficients for different frequency domain responses corresponding to different relative positions of the UE and the base station are W 51 and W 62 When the UE is on a high-speed train, 1.5 When ≥fd, the MMSE filter coefficients for different frequency domain responses corresponding to different relative positions of the UE and the base station are respectively W 61 and W 62 This embodiment will not elaborate on the selection of specific target processors.

[0112] The reference signal estimation method provided in this embodiment employs filters with different frequency domain responses for different frequency ranges (Fd) and different relative positions of the UE and the base station. This effectively suppresses out-of-band noise, making the corresponding system more robust to channels with interference or distortion. Theoretically, there is some performance loss for ideal channels. The closer the UE is to the RRH, the greater the performance loss, but the closer the UE is to the RRH, the better the signal quality. Furthermore, the responses of each filter in this embodiment are symmetrical and can be implemented using real coefficients, thus saving space occupied by coefficient storage and filters, and consequently saving hardware resources.

[0113] Secondly, an embodiment of the present invention provides a reference signal estimation method, which differs from the method in the first aspect in that, in the second aspect, the maximum Doppler frequency shift fd corresponding to the filter coefficient group and... When the quotient is not an integer, the number of MMSE filter coefficients in the filter coefficient group is determined by rounding.

[0114] In this embodiment, the maximum Doppler frequency shift fd corresponding to the filter coefficient group and When the quotient is not an integer, the corresponding integer quotient plus 1 is the number of MMSE filter coefficients in the filter coefficient group.

[0115] Specifically, in this embodiment, fd1 is also equal to 4. A filtering device is constructed, and the passband bandwidth variation patterns of filters in the same group within the filtering device in this embodiment, as well as the variation patterns of the maximum Doppler frequency shift difference between adjacent filter coefficient groups, are consistent with those of the filtering device in the first aspect. Combined with... Figure 11 The filtering device used in this embodiment has three more filters than the filtering device in the first aspect, hereinafter referred to as 'complementary filters', i.e., W. 71 W 81 W 91 .

[0116] Among them, W 71 This refers to the MMSE filter coefficients in the second group of filter coefficients, whose passband bandwidth includes 0 Hz and whose passband bandwidth is 3.5 Hz. The passband center frequency is 1.75 GHz. W 81 This refers to the MMSE filter coefficients in the fourth group of filter coefficients, whose passband bandwidth includes 0 Hz and has a passband bandwidth of 2.5 Hz. The passband center frequency is 1.25. W 91 This refers to the MMSE filter coefficients in the 6th filter coefficient group, whose passband bandwidth includes 0 Hz and whose passband bandwidth is 1.5 Hz. The passband center frequency is 0.75. .

[0117] In the specific target filter coefficient selection process, when the UE is located between two base stations and the distance between the UE and the two base stations is approximately the same, that is, the absolute value of the frequency domain position of the main spectrum is ≥3... At this time, use W 11 Estimate the reference signal; when the UE is located between two base stations, and is relatively close to one of the two base stations, 3 The absolute value of the frequency domain position of the main spectrum is ≥2. At this time, use W 22 Estimate the reference signal; when the UE is located between two base stations and is very close to one of the two base stations, 2 The absolute value of the frequency domain position of the main spectrum is ≥1. When using W 33 Estimate the reference signal; when the UE is under a certain base station, the absolute value of the frequency domain position of the main spectrum is <1. At this time, use W 44 Estimate the reference signal.

[0118] Similarly, when the speed of the high-speed train where the UE is located reaches 3.5... ≥fd>3 At that time, the MMSE filter coefficients for different frequency domain responses corresponding to different relative positions of the UE and the base station are W 21 W 32 W 43 and W 71 When the UE is on the high-speed train, 3 ≥fd>2.5 The MMSE filter coefficients for different frequency domain responses corresponding to different relative positions of the UE and the base station are W, respectively. 31 W 42 and W 53 When the UE is on a high-speed train, 2.5 ≥fd>2 The MMSE filter coefficients for different frequency domain responses corresponding to different relative positions of the UE and the base station are W, respectively. 41 W 52 and W 81 When the UE is on the high-speed train, 2 ≥fd>1.5 The MMSE filter coefficients for different frequency domain responses corresponding to different relative positions of the UE and the base station are W, respectively. 51 and W 62 When the UE is on a high-speed train, 1.5 ≥fd, the MMSE filter coefficients for different frequency domain responses corresponding to different relative positions of the UE and the base station are W respectively. 61 and W 91 This embodiment does not elaborate on the specific process of determining the target filter coefficients in the filter coefficient group.

[0119] In a third aspect, the present invention provides a filter coefficient storage method, which is applied in an apparatus for MMSE filtering scenarios, including but not limited to filtering apparatuses as described in the first or second aspect.

[0120] In the two filtering devices described above, filters with the same passband center frequency are placed in one column, and filters with the same passband bandwidth are placed in another column to form a filter matrix. Specifically, the supplementary filters in the filtering device of the second aspect are all filters from the first column. It can be understood that... Figure 1 or Figure 11 Medium filter W ij The label in the lower right corner indicates its position in the filter matrix, i.e., W. ij This represents the filter in the i-th row and j-th column.

[0121] It should be noted that the MMSE filter coefficients corresponding to the filter in the i-th row and j-th column... for ,in, Let be the noise power corresponding to the filter in the i-th row and j-th column. For W ij The signal correlation matrix, Obtained from SVD decomposition ,in, It is a diagonal matrix; the diagonal matrix corresponding to the filter in the 1st row and 1st column. For example:

[0122]

[0123] in, for The corresponding response , for The corresponding response , Let be the time-domain interval between signal i and signal j, and obviously . K is determined by the number of filters in the first group of the filtering device. For example, in the filtering device of the first aspect, K is 4.

[0124] The diagonal matrix corresponding to the filter in the first row and first column For example:

[0125] .

[0126] Will Substitute We can obtain,

[0127]

[0128]

[0129]

[0130] ,

[0131] Similarly,

[0132] ...

[0133]

[0134] Therefore, under the same resource distribution, the system containing the filter device in the first aspect requires storage. There are a total of 30 matrices, namely The 15 matrices formed The 15 vectors formed.

[0135] In this invention, combined with Figure 12 The storage method provided in this embodiment includes steps 301 to 303.

[0136] Step 301: Based on the MMSE filter coefficients with the same passband center frequency in the filtering device, construct a filter sharing coefficient for each MMSE filter coefficient with the same passband center frequency.

[0137] In a further optional embodiment of this embodiment, the step of constructing a filter sharing coefficient for each MMSE filter coefficient with the same passband center frequency based on the MMSE filter coefficients with the same passband center frequency in the filtering device includes: taking the unitary matrix obtained by decomposing the signal correlation matrix corresponding to the MMSE filter coefficient with the smallest passband bandwidth among the MMSE filter coefficients with the same passband center frequency as the filter sharing coefficient.

[0138] Taking the filtering device in the first aspect as an example, there are six filter shared coefficients, namely W 11 W 21 W 31 W 41 W 51 and W 61 The signal correlation matrix is ​​obtained as a unitary matrix through SVD decomposition.

[0139] Step 302: Fit multiple filter fitting coefficients based on the filter shared coefficients and the signal correlation matrix of each MMSE filter coefficient.

[0140] In this context, the filter fitting coefficients correspond one-to-one with the MMSE filter coefficients. Taking the filtering device in the first aspect as an example, the optimal filter fitting coefficients can be obtained by fitting the unitary matrix obtained through SVD decomposition of the signal correlation matrix. . This makes Its performance is equal to or close to that of other products. Performance.

[0141] In an optional embodiment, the step of fitting multiple filter fitting coefficients based on the filter shared coefficients and the signal correlation matrix of each MMSE filter coefficient includes:

[0142] Among the MMSE filter coefficients with the same passband center frequency, the filter-shared coefficients corresponding to the MMSE filter coefficients with the smallest passband bandwidth are fitted to the signal correlation matrices of the MMSE filter coefficients with other passband bandwidths, so as to fit the corresponding filter fitting coefficients for the MMSE filter coefficients with other passband bandwidths.

[0143] The signal correlation matrix corresponding to the MMSE filter coefficients with the smallest passband bandwidth is obtained by SVD decomposition, and the resulting diagonal matrix is ​​the filter fitting coefficient corresponding to the MMSE filter coefficients with the smallest passband bandwidth.

[0144] Specifically, in combination Figure 1 For the MMSE filter coefficients in the first column, the corresponding filter fitting coefficients are the diagonal matrices obtained by performing SVD decomposition on the signal correlation matrix corresponding to the filter coefficient itself. For the MMSE filter coefficients in other columns, the corresponding filter fitting coefficients are obtained by fitting the shared coefficients corresponding to the MMSE filter coefficient with the smallest passband bandwidth in the row containing that MMSE filter coefficient, and the signal correlation matrix corresponding to that MMSE filter coefficient with the smallest passband bandwidth.

[0145] As in Figure 1 In the table, the filter fitting coefficients for all MMSE filter coefficients in the third row are as follows:

[0146] W 31 The corresponding filter fitting coefficient is W. 31 The signal correlation matrix is ​​obtained as a diagonal matrix through SVD decomposition; W 32 and W 33 The corresponding filter fitting coefficients are respectively derived from W 31 The corresponding filter sharing coefficients are respectively related to W 32 and W 33 The corresponding signal correlation matrix is ​​obtained by fitting.

[0147] The specific fitting method can be implemented using the following formula, but it is not limited to this.

[0148] .

[0149] In a further optional embodiment of this embodiment, the step of fitting multiple filter fitting coefficients based on the filter shared coefficients and the signal correlation matrix of each MMSE filter coefficient includes:

[0150] Each filter shared coefficient is iterated, and a set of pre-selected fitting coefficients is fitted for each MMSE filter coefficient based on each filter shared coefficient and the corresponding signal correlation matrix. Each set of pre-selected fitting coefficients includes multiple pre-selected fitting coefficients, and the pre-selected fitting coefficient set corresponds one-to-one with the filter shared coefficient. It can be understood that in this invention, "one-to-one correspondence" is used to represent the correspondence between the former and the latter. For example, if A and B correspond one-to-one, it means that each A corresponds to one B. However, for each B, it may correspond to only one A, or it may correspond to multiple A, depending on the actual situation.

[0151] In each group of pre-selected fitting coefficients, the pre-selected fitting coefficient with the best fit is selected as the corresponding filter fitting coefficient.

[0152] It is understandable that the pre-fit coefficient with the best fit is the pre-fit coefficient corresponding to the smallest norm.

[0153] Specifically as follows:

[0154] .

[0155] in, It is represented as one of the unitary matrices corresponding to the six MMSE filter coefficients.

[0156] It should be noted that the norm of two matrices is used to measure the difference between the two matrices. In this invention, each coefficient corresponds to a norm. The norm includes, but is not limited to, the induced norm and the element norm, etc., which will not be elaborated in this embodiment.

[0157] Therefore, under the same resource distribution, the system containing the filter device in the first aspect requires storage. There are a total of 21 matrices, namely The six matrices formed This results in 15 vectors. While maintaining almost no loss in filtering performance, it significantly reduces storage space, from storing 15 U matrices to storing 6 U matrices. Furthermore, such as W... 62 W 53 W 44 It can also be used in non-SFN scenarios, further reducing the storage space of systems integrating multiple scenarios. The filter coefficients that need to be stored correspond to... .

[0158] Step 303: Store the filter sharing coefficients and filter fitting coefficients as filter coefficients of the filtering device in the SFN scenario.

[0159] The method further includes storing indices of the optimal pre-selected fitting coefficients and the corresponding filter shared coefficients. By storing the indices, when estimating a signal using a certain filter shared coefficient, the corresponding filter fitting coefficients can be quickly identified to jointly complete the signal estimation. This embodiment does not limit the specific process of signal estimation.

[0160] In a further optional embodiment of this example, the storage method also includes storing the filter coefficients of the filtering device in non-SFN scenarios. It is understood that the MMSE filter coefficients (hereinafter referred to as SFN filter coefficients) required for the storage filtering device in SFN scenarios may overlap with the MMSE filter coefficients (hereinafter referred to as non-SFN filter coefficients) required for the storage filtering device in non-SFN scenarios.

[0161] Specifically, the non-SFN MMSE filter coefficients include: non-SFN overlapping filter coefficients and non-SFN new filter coefficients. The non-SFN MMSE filter coefficients are the MMSE filter coefficients of the filtering device in non-SFN scenarios; the non-SFN overlapping filter coefficients are the same as the MMSE filter coefficients of the filtering device in SFN scenarios; and the non-SFN new filter coefficients are coefficients different from the MMSE filter coefficients of the filtering device in SFN scenarios. The SFN scenarios include the high-speed rail single-frequency network channel scenario, and the non-SFN scenarios include: EPA channel scenario, EVA channel scenario, ETU channel scenario, TDL channel scenario, and CDL channel scenario.

[0162] Furthermore, the storage method also includes steps 304 to 309.

[0163] Step 304: Iterate through each filter shared coefficient and fit a set of pre-selected non-SFN fitting coefficients for each non-SFN new filter coefficient based on each filter shared coefficient.

[0164] Each group of pre-selected non-SFN fitting coefficients includes multiple pre-selected non-SFN fitting coefficients, and each pre-selected non-SFN fitting coefficient in each group corresponds one-to-one with the filter shared coefficients.

[0165] It is understandable that by iterating through each filter's shared coefficients and based on the signal correlation matrix corresponding to each filter's shared coefficients and the non-SFN new filter coefficients, a set of pre-selected non-SFN fitting coefficients is fitted for each non-SFN new filter coefficient.

[0166] Step 305: Select the best-fitting pre-selected non-SFN fitting coefficient from each group of pre-selected non-SFN fitting coefficients as the corresponding new non-SFN fitting coefficients.

[0167] It is understandable that the selection of the pre-selected non-SFN fitting coefficients with the best fit can be achieved by referring to the shared argmin (argument of the minimum, which obtains the fitting result by "traversing the independent variables and selecting the independent variables that make the norm in the parentheses reach the minimum value"), which will not be elaborated in this embodiment.

[0168] Step 306: Compare the norms and thresholds corresponding to multiple non-SFN new fitted coefficients. Perform SVD decomposition on the signal correlation matrix of the non-SFN new filter coefficients corresponding to the norms greater than the threshold. The resulting unitary matrix is ​​used as the non-SFN filter shared coefficients corresponding to the non-SFN new filter coefficients.

[0169] The threshold is determined by the largest norm among all the norms corresponding to the MMSE filter coefficients that the filtering device responds to in the SFN scenario.

[0170] Step 307: Perform SVD decomposition on the signal correlation matrix of the non-SFN new filter coefficients with the corresponding norm greater than the threshold, and use the resulting diagonal matrix as the non-SFN filter fitting coefficients corresponding to the non-SFN new filter coefficients with the corresponding norm greater than the threshold.

[0171] It is understandable that if the non-SFN new filter coefficients have a corresponding norm greater than the threshold, it means that the non-SFN filter shared coefficients corresponding to the corresponding non-SFN new filter coefficients are significantly different from all the filter shared coefficients. In this case, the non-SFN new filter coefficients cannot use the filter shared coefficients corresponding to the SFN filter coefficients to estimate the signal in the non-SFN scenario, and the corresponding non-SFN filter shared coefficients need to be re-determined and stored.

[0172] For non-SFN new filter coefficients whose norm is less than or equal to the threshold, it means that there are coefficients similar to the non-SFN filter shared coefficients corresponding to the non-SFN new filter coefficients among all the filter shared coefficients. In this case, it is only necessary to store the filter shared coefficients corresponding to the similar SFN filter coefficients, the indices of the non-SFN filter fitting coefficients corresponding to the corresponding SFN filter coefficients, and the corresponding non-SFN filter fitting coefficients. In this way, when using the non-SFN new filter coefficients for signal estimation in non-SFN scenarios, signal estimation can be completed simply by using the non-SFN filter fitting coefficients corresponding to the non-SFN new filter coefficients and the filter shared coefficients corresponding to the similar SFN filter coefficients through the index.

[0173] Step 308: Fit the nonSFN filter fitting coefficients corresponding to the nonSFN new filter coefficients with the corresponding norm less than or equal to the threshold, and the nonSFN shared coefficient group consisting of the nonSFN filter shared coefficients, the filter shared coefficients corresponding to the nonSFN new filter coefficients with the corresponding norm less than the threshold, and the filter shared coefficients corresponding to the nonSFN coincident filter coefficients, to obtain the nonSFN filter fitting coefficients corresponding to the nonSFN new filter coefficients with the corresponding norm less than or equal to the threshold.

[0174] Step 309: Store the non-SFN filter shared coefficients, non-SFN filter fitting coefficients, and the indices of the non-SFN MMSE filter coefficients and their corresponding filter shared coefficients or corresponding non-SFN filter shared coefficients, as well as the indices of the filter fitting coefficients corresponding to the non-SFN coincident filter coefficients.

[0175] Where the threshold is l M, where l ranges from 1 to 10, such as 3, 6 or 8, etc. This embodiment does not make a specific limitation on this.

[0176] Understandably, the specific method for selecting the pre-selected non-SFN fitting coefficients with the best fit as the corresponding new non-SFN fitting coefficients can be by selecting the norm corresponding to the pre-selected non-SFN fitting coefficients. The smaller the norm, the better the fit of the corresponding pre-selected non-SFN fitting coefficients.

[0177] For example, the SFN filter has 15 coefficients, such as Figure 1 The 15 MMSE filter coefficients in the data correspond to the filter-sharing coefficients and filter-fitting coefficients that need to be stored. There are 21 matrices provided. The number of coefficients for the non-SFN MMSE filter is also 10, namely B0, B1, B2, B3, B4, B5, B6, B7, B8, and B9, where B5 and B6 are respectively related to W... 42 and W 53 If they are the same, then B5 and B6 are the non-SFN coincident filter coefficients, and the other eight non-SFN MMSE filter coefficients are the non-SFN new filter coefficients. In this case, the non-SFN filter shared coefficients corresponding to B5 and B6 are... The corresponding non-SFN filter fitting coefficients are... Therefore, it is no longer necessary to separately store the non-SFN filter shared coefficients and non-SFN filter fitting coefficients corresponding to B5 and B6. However, it is still necessary to store the coefficients of B5 and B6 respectively with... The index, and The index.

[0178] Now, by constructing the corresponding filter sharing coefficients and filter fitting coefficients for B0, B1, B2, B3, B4, B7, B8, and B9 respectively, the filtering device can be applied to both SFN and non-SFN scenarios.

[0179] In the process of constructing the corresponding filter shared coefficients for B0, B1, B2, B3, B4, B7, B8, and B9 respectively, by traversing... The six filters share coefficients. The differences between these six shared coefficients and the shared coefficients obtained from the SVD decomposition of the signal correlation matrices of filters B0, B1, B2, B3, B4, B7, B8, and B9 are then determined. If... If the shared coefficients of the six filters have a minimum difference from one or more of the non-SFN filter shared coefficients of B0, B1, B2, B3, B4, B7, B8, and B9, and the pre-selected non-SFN fitted coefficient with the smallest difference from the non-SFN filter shared coefficients of B0, B1, B2, B3, and B4 is still greater than the threshold (i.e., the norm corresponding to the new non-SFN fitted coefficient is still greater than the threshold), then it indicates that B0, B1, B2, B3, and B4 are dissimilar to each SFN filter coefficient. In this case, it is necessary to construct and store the corresponding non-SFN filter shared coefficients separately for B0, B1, B2, B3, and B4. The method for constructing the non-SFN filter shared coefficients is the same as that for constructing the filter shared coefficients, and will not be elaborated here.

[0180] If the preselected non-SFN fitting coefficient with the smallest difference corresponding to the shared coefficients of non-SFN filters B7, B8, and B9 is less than or equal to the threshold, then it indicates that... There are filter-shared coefficients similar to those of B7, B8, and B9 that are not SFN filters. Therefore, it is not necessary to construct and store the non-SFN filter-shared coefficients of B7, B8, and B9 here.

[0181] The fitting coefficients for non-SFN filters B0, B1, B2, B3, B4, B7, B8, and B9 still need to be fitted and stored. The fitting method for the non-SFN filter fitting coefficients is the same as that for the filter fitting coefficients, and will not be elaborated here.

[0182] It is understandable that, regardless of whether it is an SFN MMSE filter coefficient or a non-SFN MMSE filter coefficient, each SFN MMSE filter coefficient corresponds to a signal correlation matrix, a filter shared coefficient, and a filter fitting coefficient; similarly, each non-SFN MMSE filter coefficient also corresponds to a signal correlation matrix, a non-SFN filter shared coefficient, and a non-SFN filter fitting coefficient. Furthermore, the non-SFN filter shared coefficients corresponding to non-SFN MMSE filter coefficients with norms less than or equal to the threshold are also filter shared coefficients.

[0183] Fourthly, one embodiment of the present invention provides a reference signal estimation system, which is applied to the user equipment side, and the system includes:

[0184] The acquisition module is configured to acquire the reference signal sent by the base station;

[0185] The selection filter module is configured to select a target filter coefficient in the filtering device and apply it to the filtering device to estimate the reference signal based on the relationship between the absolute value of the frequency domain position of the main spectrum in the power spectrum of the reference signal and the magnitude of the maximum Doppler frequency shift in the current scene.

[0186] Among them, the main spectrum is the spectrum with the largest power among multiple power spectra; the target filter coefficient is the MMSE filter coefficient with the smallest passband bandwidth in the i-th filter coefficient group, which includes the main spectrum and the corresponding maximum Doppler frequency shift; the maximum Doppler frequency shift corresponding to the i-th filter coefficient group is not less than the maximum Doppler frequency shift in the current scene, and the maximum Doppler frequency shift corresponding to the (i+1)-th filter coefficient group is less than the maximum Doppler frequency shift in the current scene.

[0187] In this embodiment, the filtering device used can be the filtering device in the first aspect or the filtering device in the second aspect, which will not be described in detail here.

[0188] In a further optional embodiment of this embodiment, the selection filtering module includes:

[0189] The partitioning unit is configured to divide the maximum Doppler frequency shift corresponding to each group of filter coefficients into a range interval with 0, and to divide the corresponding range interval into at least one selection interval according to the number of MMSE filter coefficients in each group of filter coefficients.

[0190] The selection interval corresponds one-to-one with the MMSE filter coefficients in the corresponding filter coefficient group; in each filter coefficient group, the larger the passband bandwidth of the MMSE filter coefficient, the larger the minimum value in the corresponding selection interval.

[0191] The comparison selection unit is configured to compare the absolute value of the frequency domain position of the main spectrum in the power spectrum of the reference signal with the selection interval corresponding to the i-th group of filter coefficients, and select the MMSE filter coefficients corresponding to the selection interval containing the absolute value of the frequency domain position of the main spectrum as the target filter coefficients.

[0192] Fifthly, one embodiment of the present invention provides a storage system for filter coefficients, the storage system being applied to a filtering device as described in the first or second aspect;

[0193] The storage system includes:

[0194] The construction module is configured to construct a filter-shared coefficient for each MMSE filter coefficient with the same passband center frequency, based on the MMSE filter coefficients with the same passband center frequency in the filtering device.

[0195] The fitting module is configured to fit multiple filter fitting coefficients based on the filter shared coefficients and the signal correlation matrix of each MMSE filter coefficient, with each filter fitting coefficient corresponding to a MMSE filter coefficient.

[0196] The storage module is configured to store the filter sharing coefficients and filter fitting coefficients as filter coefficients of the filtering device.

[0197] In a further optional embodiment of this embodiment, the construction module is further configured to use the unitary matrix obtained by SVD decomposition of the signal correlation matrix corresponding to the MMSE filter coefficient with the smallest passband bandwidth among the MMSE filter coefficients at the same passband center frequency as the filter shared coefficient.

[0198] In a further optional embodiment of this example, the fitting module includes:

[0199] The fitting submodule is configured to fit the filter-shared coefficients corresponding to the MMSE filter coefficients with the smallest passband bandwidth to the signal correlation matrix of the MMSE filter coefficients with other passband bandwidths among the MMSE filter coefficients with the same passband center frequency, thereby fitting the corresponding filter fitting coefficients for the MMSE filter coefficients with other passband bandwidths.

[0200] The decomposition submodule is configured such that the signal correlation matrix corresponding to the MMSE filter coefficients with the smallest passband bandwidth is decomposed by SVD, and the resulting diagonal matrix is ​​the filter fitting coefficient corresponding to the MMSE filter coefficients with the smallest passband bandwidth.

[0201] In a further optional embodiment of this example, the fitting module further includes:

[0202] The traversal submodule is configured to traverse each filter shared coefficient and fit a set of pre-selected fitting coefficients for each MMSE filter coefficient based on each filter shared coefficient and the corresponding signal correlation matrix. Each set of pre-selected fitting coefficients includes multiple pre-selected fitting coefficients, and the pre-selected fitting coefficient set corresponds one-to-one with the filter shared coefficient.

[0203] The selection submodule is configured to select the best-fitting pre-fit coefficient from each group of pre-fit coefficients as the corresponding filter fit coefficient.

[0204] The storage module is also configured to store indices of the optimal pre-fit coefficients and the corresponding filter-shared coefficients.

[0205] In a further optional embodiment of this example, the non-SFN MMSE filter coefficients include: non-SFN overlapping filter coefficients and non-SFN new filter coefficients; the non-SFN MMSE filter coefficients are the MMSE filter coefficients of the filtering device in the non-SFN scenario; the non-SFN overlapping filter coefficients are the same coefficients as the MMSE filter coefficients of the filtering device in the SFN scenario, and the non-SFN new filter coefficients are coefficients that are different from the MMSE filter coefficients of the filtering device in the SFN scenario; the SFN scenario includes the high-speed rail single-frequency network channel scenario, and the non-SFN scenario includes: EPA channel scenario, EVA channel scenario, ETU channel scenario, TDL channel scenario, and CDL channel scenario;

[0206] The storage system also includes:

[0207] The traversal module is configured to traverse each filter shared coefficient and fit a set of pre-selected non-SFN fitting coefficients for each non-SFN new filter coefficient based on each filter shared coefficient. Each set of pre-selected non-SFN fitting coefficients includes multiple pre-selected non-SFN fitting coefficients, and each pre-selected non-SFN fitting coefficient in each set corresponds one-to-one with the filter shared coefficient.

[0208] The selection module is configured to select the best-fitting pre-selected non-SFN fitting coefficient from each group of pre-selected non-SFN fitting coefficients as the corresponding new non-SFN fitting coefficient.

[0209] The comparison module is configured to compare the norms of multiple non-SFN new fitted coefficients with a threshold. The signal correlation matrix of the non-SFN new filter coefficients with norms greater than the threshold is decomposed by SVD, and the resulting unitary matrix is ​​used as the non-SFN filter shared coefficients corresponding to the non-SFN new filter coefficients. The threshold is determined by the largest norm among all the norms of the MMSE filter coefficients that the filtering device responds to in the SFN scenario.

[0210] The first corresponding module is configured to perform SVD decomposition on the signal correlation matrix of the non-SFN new filter coefficients with the corresponding norm greater than the threshold, and the resulting diagonal matrix is ​​used as the non-SFN filter fitting coefficients corresponding to the non-SFN new filter coefficients with the corresponding norm greater than the threshold.

[0211] The second corresponding module is configured to use the non-SFN new fitting coefficients with a corresponding norm less than or equal to the threshold as the non-SFN filter fitting coefficients corresponding to the non-SFN new filter coefficients with a corresponding norm less than or equal to the threshold.

[0212] The storage module is also configured to store the non-SFN filter shared coefficients, non-SFN filter fitting coefficients corresponding to the non-SFN new filter coefficients, and the indices of the non-SFN MMSE filter coefficients and their corresponding filter shared coefficients or their corresponding non-SFN filter shared coefficients, as well as the indices of the filter fitting coefficients corresponding to the non-SFN overlapping filter coefficients.

[0213] Sixthly, the present invention provides an electronic device comprising:

[0214] At least one processor; and

[0215] A memory that is communicatively connected to at least one processor; wherein,

[0216] The memory stores instructions that can be executed by at least one processor, such that the at least one processor is able to perform the methods of the first aspect, the second aspect, or the third aspect.

[0217] In a seventh aspect, the present invention provides a computer-readable storage medium storing computer instructions that, when executed by a processor, implement the methods as described in the first, second, or third aspects.

[0218] In the description of this specification, the references to terms such as "some embodiments," "other embodiments," "ideal embodiments," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example that are included in at least one embodiment or example of this application. In this specification, the illustrative descriptions of the above terms do not necessarily refer to the same embodiments or examples.

[0219] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0220] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for storing filter coefficients, characterized in that, The storage method includes: Based on the MMSE filter coefficients with the same passband center frequency in the filtering device, construct a filter sharing coefficient for each MMSE filter coefficient with the same passband center frequency. Based on the shared coefficients of the filters and the signal correlation matrix of each MMSE filter coefficient, multiple filter fitting coefficients are fitted, and the filter fitting coefficients correspond one-to-one with the MMSE filter coefficients. The filter sharing coefficients and the filter fitting coefficients are stored as the filter coefficients of the filtering device in the SFN scenario.

2. The storage method according to claim 1, characterized in that, The step of constructing a filter-shared coefficient for each MMSE filter coefficient with the same passband center frequency based on the MMSE filter coefficients with the same passband center frequency in the filtering device includes: Among the MMSE filter coefficients with the same passband center frequency, the unitary matrix obtained by SVD decomposition of the signal correlation matrix corresponding to the MMSE filter coefficient with the smallest passband bandwidth is used as the filter shared coefficient.

3. The storage method according to claim 1, characterized in that, The step of fitting multiple filter fitting coefficients based on the filter shared coefficients and the signal correlation matrix of each MMSE filter coefficient includes: Among the MMSE filter coefficients with the same passband center frequency, the filter-shared coefficients corresponding to the MMSE filter coefficients with the smallest passband bandwidth are fitted to the signal correlation matrices of the MMSE filter coefficients with other passband bandwidths, so as to fit the corresponding filter fitting coefficients for the MMSE filter coefficients with other passband bandwidths. The signal correlation matrix corresponding to the MMSE filter coefficients with the smallest passband bandwidth is obtained by SVD decomposition, and the resulting diagonal matrix is ​​the filter fitting coefficient corresponding to the MMSE filter coefficients with the smallest passband bandwidth.

4. The storage method according to claim 1, characterized in that, The step of fitting multiple filter fitting coefficients based on the filter shared coefficients and the signal correlation matrix of each MMSE filter coefficient includes: Iterate through each of the filter shared coefficients, and fit a set of pre-selected fitting coefficients for each MMSE filter coefficient according to each filter shared coefficient and the corresponding signal correlation matrix. Each set of pre-selected fitting coefficients includes multiple pre-selected fitting coefficients, and the pre-selected fitting coefficient set corresponds one-to-one with the filter shared coefficient. In each group of pre-selected fitting coefficients, the pre-selected fitting coefficient with the best fit is selected as the corresponding filter fitting coefficient. The method further includes storing an index of the optimal pre-fit coefficients and the corresponding filter-shared coefficients.

5. The storage method according to any one of claims 1 to 3, characterized in that, The non-SFN MMSE filter coefficients include: non-SFN overlapping filter coefficients and non-SFN new filter coefficients; the non-SFN MMSE filter coefficients are the MMSE filter coefficients of the filter device in response to non-SFN scenarios; the non-SFN overlapping filter coefficients are the same as the MMSE filter coefficients of the filter device in response to SFN scenarios, and the non-SFN new filter coefficients are the different from the MMSE filter coefficients of the filter device in response to SFN scenarios; the SFN scenarios include the high-speed rail single-frequency network channel scenario, and the non-SFN scenarios include: EPA channel scenario, EVA channel scenario, ETU channel scenario, TDL channel scenario, and CDL channel scenario; The storage method further includes: Iterate through each of the filter shared coefficients, and fit a set of pre-selected non-SFN fitting coefficients for each of the non-SFN new filter coefficients based on each of the filter shared coefficients. Each set of pre-selected non-SFN fitting coefficients includes multiple pre-selected non-SFN fitting coefficients, and each pre-selected non-SFN fitting coefficient in each set corresponds one-to-one with the filter shared coefficient. In each group of pre-selected non-SFN fitting coefficients, the pre-selected non-SFN fitting coefficient with the best fit is selected as the corresponding new non-SFN fitting coefficient. The norms corresponding to multiple non-SFN new fitting coefficients are compared with a threshold. The signal correlation matrix of the non-SFN new filter coefficients corresponding to the norms greater than the threshold is decomposed by SVD. The resulting unitary matrix is ​​used as the non-SFN filter shared coefficients corresponding to the non-SFN new filter coefficients. The threshold is determined by the largest norm among all the norms corresponding to the MMSE filter coefficients that the filtering device responds to in the SFN scenario. The signal correlation matrix of the non-SFN new filter coefficients with the corresponding norm greater than the threshold is decomposed by SVD, and the resulting diagonal matrix is ​​used as the non-SFN filter fitting coefficients corresponding to the non-SFN new filter coefficients with the corresponding norm greater than the threshold. The non-SFN new fitting coefficients with a corresponding norm less than or equal to the threshold are taken as the non-SFN filter fitting coefficients corresponding to the non-SFN new filter coefficients with a corresponding norm less than or equal to the threshold. Store the non-SFN filter shared coefficients and non-SFN filter fitting coefficients corresponding to the non-SFN new filter coefficients, as well as the indices of the non-SFN MMSE filter coefficients and their corresponding filter shared coefficients or corresponding non-SFN filter shared coefficients, and the indices of the filter fitting coefficients corresponding to the non-SFN coincident filter coefficients.

6. A filter coefficient storage system, characterized in that, The storage system includes: The construction module is configured to construct a filter-shared coefficient for each MMSE filter coefficient with the same passband center frequency, based on the MMSE filter coefficients with the same passband center frequency in the filtering device. The fitting module is configured to fit multiple filter fitting coefficients based on the filter shared coefficients and the signal correlation matrix of each MMSE filter coefficient, wherein the filter fitting coefficients correspond one-to-one with the MMSE filter coefficients. The storage module is configured to store the filter shared coefficients and the filter fitting coefficients as filter coefficients of the filtering device.

7. The storage system according to claim 6, characterized in that, The building module is further configured to use the unitary matrix obtained by SVD decomposition of the signal correlation matrix corresponding to the MMSE filter coefficient with the smallest passband bandwidth among the MMSE filter coefficients at the same passband center frequency as the filter shared coefficient.

8. The storage system according to claim 6, characterized in that, The fitting module includes: The fitting submodule is configured to fit the filter-shared coefficients corresponding to the MMSE filter coefficients with the smallest passband bandwidth to the signal correlation matrix of the MMSE filter coefficients with other passband bandwidths among the MMSE filter coefficients with the same passband center frequency, thereby fitting the corresponding filter fitting coefficients for the MMSE filter coefficients with other passband bandwidths. The decomposition submodule is configured such that the signal correlation matrix corresponding to the MMSE filter coefficients with the smallest passband bandwidth is decomposed by SVD, and the resulting diagonal matrix is ​​the filter fitting coefficient corresponding to the MMSE filter coefficients with the smallest passband bandwidth.

9. The storage system according to claim 6, characterized in that, The fitting module further includes: The traversal submodule is configured to traverse each of the filter shared coefficients and fit a set of pre-selected fitting coefficients for each MMSE filter coefficient based on each filter shared coefficient and the corresponding signal correlation matrix. Each set of pre-selected fitting coefficients includes multiple pre-selected fitting coefficients, and the pre-selected fitting coefficient set corresponds one-to-one with the filter shared coefficient. The selection submodule is configured to select the best-fitting pre-fit coefficient from each group of pre-fit coefficients as the corresponding filter fit coefficient. The storage module is also configured to store indices of the optimal pre-fit coefficients and the corresponding filter-shared coefficients.

10. The storage system according to any one of claims 6 to 9, characterized in that, The non-SFN MMSE filter coefficients include: non-SFN overlapping filter coefficients and non-SFN new filter coefficients; the non-SFN MMSE filter coefficients are the MMSE filter coefficients of the filter device in response to non-SFN scenarios; the non-SFN overlapping filter coefficients are the same as the MMSE filter coefficients of the filter device in response to SFN scenarios, and the non-SFN new filter coefficients are the different from the MMSE filter coefficients of the filter device in response to SFN scenarios; the SFN scenarios include the high-speed rail single-frequency network channel scenario, and the non-SFN scenarios include: EPA channel scenario, EVA channel scenario, ETU channel scenario, TDL channel scenario, and CDL channel scenario; The storage system also includes: The traversal module is configured to traverse each of the filter shared coefficients and fit a set of pre-selected non-SFN fitting coefficients for each of the non-SFN new filter coefficients based on each of the filter shared coefficients. Each set of pre-selected non-SFN fitting coefficients includes multiple pre-selected non-SFN fitting coefficients, and each pre-selected non-SFN fitting coefficient in each set corresponds one-to-one with the filter shared coefficient. The selection module is configured to select the best-fitting pre-selected non-SFN fitting coefficient from each group of pre-selected non-SFN fitting coefficients as the corresponding new non-SFN fitting coefficient. The comparison module is configured to compare the norms corresponding to multiple non-SFN new fitted coefficients with a threshold. The signal correlation matrix of the non-SFN new filter coefficients with norms greater than the threshold is decomposed by SVD, and the resulting unitary matrix is ​​used as the non-SFN filter shared coefficients corresponding to the non-SFN new filter coefficients. The threshold is determined by the largest norm among all the norms corresponding to the MMSE filter coefficients that the filtering device responds to in the SFN scenario. The first corresponding module is configured to perform SVD decomposition on the signal correlation matrix of the non-SFN new filter coefficients with the corresponding norm greater than the threshold, and the resulting diagonal matrix is ​​used as the non-SFN filter fitting coefficients corresponding to the non-SFN new filter coefficients with the corresponding norm greater than the threshold. The second corresponding module is configured to use the non-SFN new fitting coefficients with a corresponding norm less than or equal to the threshold as the non-SFN filter fitting coefficients corresponding to the non-SFN new filter coefficients with a corresponding norm less than or equal to the threshold. The storage module is further configured to store the non-SFN filter shared coefficients, non-SFN filter fitting coefficients corresponding to the non-SFN new filter coefficients, the indexes of the non-SFN MMSE filter coefficients and their corresponding filter shared coefficients or their corresponding non-SFN filter shared coefficients, and the indexes of the filter fitting coefficients corresponding to the non-SFN coincident filter coefficients.

11. An electronic device, characterized in that, The electronic device includes: At least one processor; and 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 method of any one of claims 1 to 5.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed by a processor, implement the method as described in any one of claims 1 to 5.

Images