PDP resampling shift algorithm, PDP filtering method and device, filtering equipment and storage medium

Abstract

The invention discloses a PDP resampling shift algorithm, a PDP filtering method, a PDP filtering device, filtering equipment and a storage medium, according to the PDP resampling shift algorithm, a relation between a power delay spectrum PDP before and after timing adjustment and a timing adjustment amount delta and a resampling coefficient are obtained, and when the sampling point offset of the current subframe and the sampling point offset of the later subframe are not an integer, the timing adjustment amount delta is adjusted. The timing adjustment amount delta can be not an integer, so that cyclic shift alignment can be performed through the timing adjustment amount delta, the timing positions of the front and rear subframes are matched, and the applicability is higher; according to the method, the historical PDP is resampled, and the historical PDP can comprise all important path information after being filtered, so that the real channel state is reflected and is closer to the real PDP of the current sub-frame, and the accuracy of channel estimation is improved.

Description

Technical Field

[0001] The present invention relates to the field of communication technology, and in particular to a PDP resampling shift algorithm and a PDP filtering method, a device computer device and a storage medium. Background Art

[0002] In wireless communication systems, signals may experience direct radiation, reflection, scattering, diffraction, etc. during propagation, resulting in multipath, movement, noise, and interference in the signals received by the receiving end. The receiver needs to perform time-frequency offset adjustment and channel estimation on the received signal before demodulation and decoding. Estimation of time offset, frequency offset, channel delay spread, and Doppler spread is collectively referred to as parameter estimation. Parameter estimation is a parameter that reflects channel characteristics, which determines the accuracy of time-frequency offset adjustment and the performance of channel estimation, thereby affecting the performance of the receiver.

[0003] The PDP spectrum is not only a direct input for time offset estimation, delay spread estimation, and frequency domain correlation estimation, but also determines the path selection position for frequency offset estimation and Doppler estimation, and also determines the filter coefficients in DFT channel estimation. In the PDP estimation process, the PDPs between different subframes need to be merged to improve the accuracy of PDP estimation. However, if the time offset is fixed, after the time offset is adjusted at the front end, the timing positions of the PDP of the current subframe and the historical PDP may not be aligned, resulting in the inability to directly merge the two. If the time offset adjusted by the front end is converted to an integer sampling point at the current resolution, the timing positions of the two can be aligned by cyclically shifting the historical PDP, but if it is not an integer sampling point, the two cannot be aligned by cyclic shifting. In the usual solution, the problem faced by this scenario cannot be solved. Summary of the invention

[0004] The present invention aims to overcome at least one defect of the above-mentioned prior art and provide a PDP resampling shift algorithm and a PDP filtering method, apparatus, filtering device and storage medium to solve the technical problem of solving the problem of timing misalignment before and after timing adjustment in the prior art.

[0005] The present invention provides a PDP resampling shift algorithm, comprising:

[0006] S100, let the channel time domain impulse response CIR be h, the frequency domain response CFR be H, and find the frequency domain response H of the kth subcarrier k ;

[0007] S101, cyclically shift the channel time domain impulse response h and the channel frequency domain response H, the timing adjustment amount of the cyclic shift is Δ, and the results are respectively expressed as and Indicates that the frequency domain response of the kth subcarrier is obtained Thus, the time domain impulse response of the nth path is obtained

[0008] S102, performing fast Fourier transform FFT on the time domain impulse response h and the frequency domain response H to obtain a cyclically shifted time domain impulse response and frequency domain response Fast Fourier Transform form;

[0009] S103, according to Perform point operation on the Fourier transform FFT result to obtain the power delay spectrum PDP;

[0010] S104: Obtain a time domain impulse response according to the power delay profile PDP. The nth element of the corresponding PDP;

[0011] S105: according to the time domain impulse response The nth element of the corresponding PDP obtains the relationship between the power delay profile PDP before and after the timing adjustment and the timing adjustment amount Δ;

[0012] S106: Obtain a resampling coefficient coef according to the relationship between the power delay profile PDP before and after the timing adjustment and the timing adjustment amount Δ nl :

[0013] coef nl =|a nl | 2 =|sinc(nl-Δ)| 2

[0014] Where n and l are the serial indexes of the path of the time domain impulse response h, 0≤n,l <N。

[0015] The H of the k-th subcarrier in S100 can be expressed as H k :

[0016]

[0017] Among them, h n is the time domain impulse response of the nth path, N is the length of the IFFT transform, 0≤n <N,0≤k<N。

[0018] The k-th subcarrier in S101 It can be expressed as

[0019]

[0020] The time domain impulse response of the nth path is for:

[0021]

[0022] By further calculation:

[0023]

[0024] Where m is the number of the path of the time domain impulse response h, 0≤m <N。

[0025] In order to facilitate the subsequent formula derivation, the time domain impulse response CIR and the frequency domain response CFR are converted into a fast Fourier transform FFT form. The time domain impulse response CIR and the frequency domain response CFR can be written in a matrix form:

[0026] H=Fh

[0027] Wherein, F is the fast Fourier transform (FFT) matrix.

[0028] According to the matrix form of the time domain impulse response CIR and the frequency domain response CFR, the frequency domain response It can be expressed as:

[0029]

[0030] Where E is represented by:

[0031]

[0032] The corresponding time domain impulse response It can be expressed as:

[0033]

[0034] Among them, F -1 is the inverse of the matrix F.

[0035] Then, the power delay spectrum PDP is The diagonal elements of :

[0036]

[0037] Among them, X H represents the conjugate transposed matrix of matrix X, Λ=hh H is the unit matrix, A=F -1 ·E·F.

[0038] The step of obtaining the power delay profile PDP in S103 further includes: making the multipaths of the time domain impulse response uncorrelated, then:

[0039]

[0040] Where Λ=hhH is the identity matrix, h H represents the conjugate transposed matrix of matrix h, |h0| 2 represents the power gain of the channel on the 0th path to the signal, |h N-1 | 2 It represents the power gain of the channel to the signal on the N-1th path.

[0041] The time domain impulse response is obtained as described in S104 The corresponding nth element of PDP is expressed as:

[0042]

[0043] Among them, A H It represents the conjugate transposed matrix of matrix A, n and l are the serial index of the path of the time domain impulse response h, 0≤n,l <N。

[0044]

[0045] Further derivation of the above formula gives:

[0046]

[0047] The relationship between the power delay profile PDP before and after the timing adjustment in S105 and the timing adjustment amount Δ is expressed as:

[0048]

[0049] During the PDP estimation process, the PDPs between different subframes need to be merged to improve the accuracy of PDP estimation. However, if the time offset is fixed, after the front end has been adjusted for the time offset, the timing positions of the PDP of the current subframe and the historical PDP may not be aligned, resulting in the inability to directly merge the two. If the time offset adjusted by the front end is converted to an integer sampling point at the current resolution, the timing positions of the two can be aligned by cyclically shifting the historical PDP, but if it is not an integer sampling point, the two cannot be aligned by cyclic shifting. The historical PDP and the current instantaneous PDP are aligned by setting the timing adjustment amount Δ, and the timing adjustment amount Δ may be an integer or may not be an integer. The cyclic shift Δ in S101 is determined based on the timing adjustment amount of the timing position sampling points of the first and last two subframes in the channel.

[0050] Therefore, the PDP resampling function is added before PDP filtering, and the resampling coefficient is:

[0051] coef nl =|a nl | 2 =|sinc(nl-Δ)| 2

[0052] The present invention also provides a PDP filtering method, comprising the above-mentioned PDP resampling shift algorithm, and further comprising:

[0053] S1. receiving subframe signal data from a wireless communication system;

[0054] S2. Extracting a pilot signal from the received subframe signal data, performing least squares channel estimation LS and inverse Fourier transform FFT, and obtaining a channel frequency domain response CFR and a time domain impulse response CIR;

[0055] S3, calculating the power of the time domain impulse response CIR to obtain a historical PDP;

[0056] S4, calculate the obtained channel frequency domain response CFR and time domain impulse response CIR according to steps S100-S106 to obtain instantaneous PDP and resampling coefficient coef nl ;

[0057] S5, filtering the historical PDP;

[0058] S6, resampling the historical PDP so that the historical PDP is aligned with the current PDP timing position and then filtering the two together to obtain a filtered PDP;

[0059] S7. De-noise the filtered PDP to obtain a de-noised PDP.

[0060] Since the historical PDP has no sampling deviation relative to the current instantaneous PDP, it only shows one main path, and this main path corresponds to the strongest path in signal propagation. When the historical PDP and the current instantaneous PDP are directly filtered, the filtered PDP may only reflect one of the main paths, usually the stronger one, due to the two main paths of the current subframe PDP, and thus cannot accurately reflect the actual channel state of the current subframe because it ignores another important path that appears due to the timing offset. Therefore, it is necessary to resample the historical PDP according to the timing adjustment amount Δ to match the timing position of the previous and next subframes.

[0061] After resampling the historical PDP, the historical PDP has two main paths, which are close to the instantaneous PDP and can match the timing positions of the previous and next subframes, so that the filtering process can more accurately reflect the channel state of the current subframe, including all important path information. The historical PDP filtered after resampling will be closer to the real PDP of the current subframe, thereby improving the accuracy of channel estimation and system performance.

[0062] The present invention also provides a PDP filtering device comprising:

[0063] Acquisition module: used for receiving signal data from a wireless communication system;

[0064] Conversion module: obtains the channel frequency domain response CFR and time domain impulse response CIR through least squares channel estimation LS and inverse Fourier transform FFT;

[0065] Calculation module: Calculate the power according to the time domain impulse response CIR, and obtain the historical PDP, instantaneous PDP and resampling coefficient coef nl ;

[0066] Filtering module: filtering the historical PDP;

[0067] Resampling module: resampling the historical PDP so that the historical PDP is aligned with the instantaneous PDP timing position;

[0068] Processing module: filtering and denoising the resampled historical PDP and instantaneous PDP.

[0069] The present invention also provides a filtering device, comprising: at least one processor; and a memory in communication connection with the at least one processor;

[0070] The memory stores instructions executable by the at least one processor, and the instructions are configured to execute any one of the above-mentioned PDP resampling shift algorithm and PDP filtering method.

[0071] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, wherein when the computer program is processed and executed, the steps of the PDP filtering method as described in any one of the PDP resampling shift algorithm and the PDP filtering method are implemented.

[0072] The beneficial effects of the present invention are as follows: the PDP resampling shift algorithm and the PDP filtering method, device, filtering equipment and storage medium, through the PDP resampling shift algorithm, obtain the relationship between the power delay spectrum PDP before and after the timing adjustment and the timing adjustment amount Δ and the resampling coefficient. When the sampling point offset of the current and subsequent subframes is not an integer, since the timing adjustment amount Δ may not be an integer, the timing adjustment amount Δ may be used to perform cyclic shift alignment to match the timing positions of the previous and subsequent subframes, which is more applicable; by resampling the historical PDP, the historical PDP can include all important path information after filtering, thereby reflecting the real channel state, being closer to the real PDP of the current subframe, and improving the accuracy of channel estimation. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0074] Figure 1 Schematic diagram of the steps of the PDP resampling shift algorithm.

[0075] Figure 2 It is a step diagram of the PDP filtering method.

[0076] Figure 3 is a flow chart of the PDP filtering method.

[0077] Figure 4 This is a schematic diagram of the non-resampling PDP filtering before and after the third embodiment.

[0078] Figure 5 Schematic diagram of PDP filtering before and after resampling in Example 3. DETAILED DESCRIPTION

[0079] In order to make the purpose, technical solution and advantages of the present invention more clear, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings:

[0080] The embodiment of the present invention provides an embodiment of a PDP resampling shift algorithm and a PDP filtering method, device, and storage medium. Although the logical order is shown in the flow chart, under certain data, the steps shown or described can be completed in an order different from that here.

[0081] Example 1: Reference Figure 1

[0082] like Figure 1 As shown in the schematic diagram of the steps of the PDP resampling shift algorithm, the present invention provides a PDP resampling shift algorithm, including:

[0083] S100, let the channel time domain impulse response CIR be h, the frequency domain response CFR be H, and find the frequency domain response H of the kth subcarrier k ;

[0084] S101, cyclically shift the channel time domain impulse response h and the channel frequency domain response H, the timing adjustment amount of the cyclic shift is Δ, and the results are respectively expressed as and Indicates that the frequency domain response of the kth subcarrier is obtained Thus, the time domain impulse response of the nth path is obtained

[0085] S102, performing fast Fourier transform FFT on the time domain impulse response h and the frequency domain response H to obtain a cyclically shifted time domain impulse response and frequency domain response Fast Fourier Transform form;

[0086] S103, according to Perform point operation on the Fourier transform FFT result to obtain the power delay spectrum PDP;

[0087] S104: Obtain a time domain impulse response according to the power delay profile PDP. The nth element of the corresponding PDP;

[0088] S105: according to the time domain impulse response The nth element of the corresponding PDP obtains the relationship between the power delay profile PDP before and after the timing adjustment and the timing adjustment amount Δ;

[0089] S106: Obtain a resampling coefficient coef according to the relationship between the power delay profile PDP before and after the timing adjustment and the timing adjustment amount Δ nl :

[0090] coef nl =|a nl | 2 =|sinc(nl-Δ)| 2

[0091] Where n and l are the serial indexes of the path of the time domain impulse response h, 0≤n,l <N。

[0092] The H of the k-th subcarrier in S100 can be expressed as H k :

[0093]

[0094] Among them, h n is the time domain impulse response of the nth path, N is the length of the IFFT transform, 0≤n <N,0≤k<N。

[0095] The k-th subcarrier in S101 It can be expressed as

[0096]

[0097] The time domain impulse response of the nth path is for:

[0098]

[0099] By further calculation:

[0100]

[0101] Where m is the number of the path of the time domain impulse response h, 0≤m <N。

[0102] In order to facilitate the subsequent formula derivation, the time domain impulse response CIR and the frequency domain response CFR are converted into a fast Fourier transform FFT form. The time domain impulse response CIR and the frequency domain response CFR can be written in a matrix form:

[0103] H=Fh

[0104] Wherein, F is the fast Fourier transform (FFT) matrix.

[0105] According to the matrix form of the time domain impulse response CIR and the frequency domain response CFR, the frequency domain response It can be expressed as:

[0106]

[0107] Where E is represented by:

[0108]

[0109] The corresponding time domain impulse response It can be expressed as:

[0110]

[0111] Among them, F -1 is the inverse of the matrix F.

[0112] Then, the power delay spectrum PDP is The diagonal elements of :

[0113]

[0114] Among them, X H represents the conjugate transposed matrix of matrix X, Λ=hh H is the unit matrix, A=F -1 ·E·F.

[0115] The step of obtaining the power delay profile PDP in S103 further includes: making the multipaths of the time domain impulse response uncorrelated, then:

[0116]

[0117] Where Λ=hh H is the identity matrix, h H represents the conjugate transposed matrix of matrix h, |h0|2 represents the power gain of the channel on the 0th path to the signal, |h N-1 | 2 It represents the power gain of the channel to the signal on the N-1th path.

[0118] The time domain impulse response is obtained as described in S104 The corresponding nth element of PDP is expressed as:

[0119]

[0120] Among them, A H It represents the conjugate transposed matrix of matrix A, n and l are the serial index of the path of the time domain impulse response h, 0≤n,l <N。

[0121]

[0122] Further derivation of the above formula gives:

[0123]

[0124]

[0125] The relationship between the power delay profile PDP before and after the timing adjustment in S105 and the timing adjustment amount Δ is expressed as:

[0126]

[0127] During the PDP estimation process, the PDPs between different subframes need to be merged to improve the accuracy of PDP estimation. However, if the time offset is fixed, after the time offset is adjusted at the front end, the timing positions of the PDP of the current subframe and the historical PDP may not be aligned, resulting in the inability to directly merge the two. If the time offset adjusted by the front end is converted to an integer sampling point at the current resolution, the timing positions of the two can be aligned by cyclically shifting the historical PDP, but if it is not an integer sampling point, the two cannot be aligned by cyclic shifting.

[0128] The historical PDP is aligned with the current instantaneous PDP by setting a timing adjustment Δ, which may be an integer or not. The cyclic shift Δ in S101 is determined according to the timing adjustment of the timing position sampling points of the first and second subframes in the channel.

[0129] Therefore, the PDP resampling function is added before PDP filtering, and the resampling coefficient is:

[0130] coef nl =|a nl | 2 =|sinc(nl-Δ)|2

[0131] Example 2: Reference Figure 2 and Figure 3

[0132] Compared with Example 1, the difference is that Figure 2 and Figure 3 As shown in the step diagram of the PDP filtering method, the present invention also provides a PDP filtering method, comprising:

[0133] The present invention also provides a PDP filtering method, comprising the above-mentioned PDP resampling shift algorithm, and further comprising:

[0134] S1. receiving subframe signal data from a wireless communication system;

[0135] S2. Extracting a pilot signal from the received subframe signal data, performing least squares channel estimation LS and inverse Fourier transform FFT, and obtaining a channel frequency domain response CFR and a time domain impulse response CIR;

[0136] S3, calculating the power of the time domain impulse response CIR to obtain a historical PDP;

[0137] S4, calculate the obtained channel frequency domain response CFR and time domain impulse response CIR according to steps S100-S106 to obtain instantaneous PDP and resampling coefficient coef nl ;

[0138] S5, filtering the historical PDP;

[0139] S6, resampling the historical PDP so that the historical PDP is aligned with the current PDP timing position and then filtering the two together to obtain a filtered PDP;

[0140] S7. De-noise the filtered PDP to obtain a de-noised PDP.

[0141] Since the historical PDP has no sampling deviation relative to the current instantaneous PDP, it only shows one main path, and this main path corresponds to the strongest path in signal propagation. When the historical PDP and the current instantaneous PDP are directly filtered, the filtered PDP may only reflect one of the main paths, usually the stronger one, due to the two main paths of the current subframe PDP, and thus cannot accurately reflect the actual channel state of the current subframe because it ignores another important path that appears due to the timing offset. Therefore, it is necessary to resample the historical PDP according to the timing adjustment amount Δ to match the timing position of the previous and next subframes.

[0142] After resampling the historical PDP, the historical PDP has two main paths, which are close to the instantaneous PDP and can match the timing positions of the previous and next subframes, so that the filtering process can more accurately reflect the channel state of the current subframe, including all important path information. The historical PDP filtered after resampling will be closer to the real PDP of the current subframe, thereby improving the accuracy of channel estimation and system performance.

[0143] Example 3: Reference Figure 4 and Figure 5

[0144] Compared with the above embodiment, the difference is that, assuming that the current channel is Gaussian white noise (AWGN), the timing positions of the two subframes before and after are offset by 0.5 sampling points. If the historical PDP is not resampled, the two subframes before and after and the filtered PDP are as follows: Figure 4 As shown in the figure, since the current subframe is 0.5 sampling points off from the previous subframe, there are two main paths on the PDP of the current subframe, while the historical PDP has no sampling deviation and only one main path. The two are directly filtered, and the filtered PDP has only one main path. As a result, the filtered PDP is quite different from the current subframe PDP, and cannot accurately reflect the channel status of the current subframe.

[0145] By adopting the PDP resampling shift algorithm and PDP filtering method of the present invention, the historical PDP is resampled according to the timing adjustment amount Δ, and the PDP before and after filtering is as follows: Figure 5 As shown in Figure 2, the historical PDP after resampling has two main paths, which are similar to the current instantaneous PDP waveform. The filtered PDP can reflect the real state of the current channel.

Claims

1. A PDP resampling shift algorithm, characterized in that: S100, let the channel time domain impulse response CIR be h, the frequency domain response CFR be H, and find the frequency domain response H of the kth subcarrier k ; S101, cyclically shift the channel time domain impulse response h and the channel frequency domain response H, the timing adjustment amount of the cyclic shift is Δ, and the results are respectively expressed as and Indicates that the frequency domain response of the kth subcarrier is obtained Thus, the time domain impulse response of the nth path is obtained S102, performing fast Fourier transform FFT on the time domain impulse response h and the frequency domain response H to obtain a cyclically shifted time domain impulse response and frequency domain response Fast Fourier Transform form; S103, according to Perform point operation on the Fourier transform FFT result to obtain the power delay spectrum PDP; S104: Obtain a time domain impulse response according to the power delay profile PDP. The nth element of the corresponding PDP; S105: according to the time domain impulse response The nth element of the corresponding PDP obtains the relationship between the power delay profile PDP before and after the timing adjustment and the timing adjustment amount Δ; S106: Obtain a resampling coefficient coef according to the relationship between the power delay profile PDP before and after the timing adjustment and the timing adjustment amount Δ nl : coefficient nl =|a nl | 2 =|sinc(nl-Δ)| 2 Where n and l are the serial indexes of the path of the time domain impulse response h, 0≤n,l <N。 2. A PDP resampling shift algorithm according to claim 1, characterized in that: Obtaining the power delay profile PDP in S103 includes: Assuming that the multipaths of the time domain impulse response are uncorrelated, we have: Where Λ=hh H is the identity matrix, h H represents the conjugate transposed matrix of matrix h, |h0| 2 represents the power gain of the channel on the 0th path to the signal, |h N-1 | 2 It represents the power gain of the channel to the signal on the N-1th path.

3. A PDP resampling shift algorithm according to claim 1, characterized in that: The time domain impulse response is obtained as described in S104 The corresponding nth element of PDP is expressed as: Among them, A H It represents the conjugate transposed matrix of matrix A, n and l are the serial index of the path of the time domain impulse response h, 0≤n,l <N。 4. A PDP resampling shift algorithm according to claim 1, characterized in that: The relationship between the power delay profile PDP before and after the timing adjustment in S105 and the timing adjustment amount Δ is expressed as:

5. A PDP resampling shift algorithm according to claim 1, characterized in that: The cyclic shift Δ in S101 is determined according to the timing adjustment amount of the timing position sampling points of the first and second subframes in the channel.

6. A method for PDP filtering, characterized in that: The method comprises a PDP resampling shift algorithm as claimed in claims 1 to 5, further comprising: S1. receiving subframe signal data from a wireless communication system; S2. Extracting a pilot signal from the received subframe signal data, performing least squares channel estimation LS and inverse Fourier transform FFT, and obtaining a channel frequency domain response CFR and a time domain impulse response CIR; S3, calculating the power of the time domain impulse response CIR to obtain a historical PDP; S4, calculate the obtained channel frequency domain response CFR and time domain impulse response CIR according to steps S100-S106 to obtain instantaneous PDP and resampling coefficient coef nl ; S5, filtering the historical PDP; S6, resampling the historical PDP so that the historical PDP is aligned with the current PDP timing position and then filtering the two together to obtain a filtered PDP; S7. De-noise the filtered PDP to obtain a de-noised PDP.

7. A PDP method according to claim 6, characterized in that: The resampling of the historical PDP in S6 further includes: The historical PDP is resampled according to the timing adjustment amount Δ to match the timing positions of the preceding and succeeding subframes.

8. A PDP method according to claim 7, characterized in that: After the historical PDP is resampled, the historical PDP has two main paths, which are close to the instantaneous PDP and are used to accurately reflect the actual channel state of the current subframe.

9. A PDP filter device, characterized in that: include: Acquisition module: used for receiving signal data from a wireless communication system; Conversion module: obtains the channel frequency domain response CFR and time domain impulse response CIR through least squares channel estimation LS and inverse Fourier transform FFT; Calculation module: Calculate the power according to the time domain impulse response CIR, and obtain the historical PDP, instantaneous PDP and resampling coefficient coef nl ; Filtering module: filtering the historical PDP; Resampling module: resampling the historical PDP so that the historical PDP is aligned with the instantaneous PDP timing position; Processing module: filtering and denoising the resampled historical PDP and instantaneous PDP.

10. A filtering device, characterized in that: include: at least one processor; and a memory communicatively coupled to the at least one processor; The memory stores instructions executable by the at least one processor, and the instructions are configured to execute the PDP resampling shift algorithm and PDP filtering method described in any one of claims 1 to 8.

11. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is processed and executed, the steps of the PDP resampling shift algorithm and the PDP filtering method as described in any one of claims 1 to 8 are implemented.

Images