Data processing method and device, equipment and medium

By calculating the mutual correlation coefficient matrix and its inverse matrix in the communication system, disassembly and merge the Cholesky decomposition algorithm, the problem of high complexity of inverse calculation of the noise correlation matrix is ​​solved, and the rapid computing and memory optimization of the system are realized.

CN119094283APending Publication Date: 2024-12-06SHANGHAI SMARTLOGIC TECHNOLOGY LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411232031.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-03
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

When the prior art performs multi-antenna equalization in communication systems, the inverse calculation of the noise correlation matrix is ​​complex, resulting in a long time-consuming system and a large memory footprint, which makes it difficult to effectively reduce.

Method used

By obtaining the interference and covariance matrix, compute the mutual correlation coefficient matrix and its related intermediate variables, and performing division inverse calculation and dynamic calibration based on these variables, disassembly and merge the inverse matrix calculation process in the Cholesky decomposition algorithm, using high instructions The parallelism degree of DSP processor performs rapid calculations.

Benefits of technology

The rapid calculation of Cholesky decomposition is realized, which reduces the system time, reduces the memory burden, and improves the system's processing efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119094283A_ABST
    Figure CN119094283A_ABST
Patent Text Reader

Abstract

The invention discloses a data processing method and device, equipment and a medium, and relates to the technical field of communication, and the method comprises the steps: obtaining interference and a covariance matrix; calculating a cross correlation coefficient matrix and related intermediate variables corresponding to the cross correlation coefficient matrix according to the interference and covariance matrix; performing segmented inversion operation and dynamic calibration on the cross correlation coefficient matrix based on the related intermediate variables corresponding to the cross correlation coefficient matrix to obtain an inverse matrix corresponding to the cross correlation coefficient matrix, and calculating the cross correlation coefficient matrix and the related intermediate variables corresponding to the cross correlation coefficient matrix to obtain the cross correlation coefficient matrix. According to the method, the calculation process of the inverse matrix corresponding to the cross correlation coefficient matrix in the cholesky decomposition algorithm is disassembled and merged, so that the rapid calculation of the cholesky decomposition is realized, the time consumption of the system is reduced, and the burden of the memory is relieved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of communication technology, and in particular to data processing methods, devices, equipment and media. Background Art

[0002] In the communication system, in order to merge the multi-antenna equalization IRC (Inter-Symbol Interference Compensation), the concept of multi-antenna receiving noise correlation matrix is ​​introduced into the system, and the antenna correlation noise and the interference of the adjacent cells are eliminated by inverting the noise correlation matrix to achieve a better suppression effect. In the whole operation process, due to the high complexity of the inversion operation of the noise correlation matrix, this part of the calculation often becomes the bottleneck of the whole system. At present, the industry generally uses the Cholesky decomposition method or its variant to decompose and invert the matrix.

[0003] The existing technology simplifies matrix decomposition and inversion in theory, which reduces the burden on system resources to a certain extent. However, in actual implementation, it is still time-consuming and occupies a lot of memory.

[0004] The above contents are only used to assist in understanding the technical solution of the present application and do not constitute an admission that the above contents are prior art. Summary of the invention

[0005] The main purpose of this application is to provide a data processing method, device, equipment and medium, aiming to accelerate Cholesky decomposition to reduce system time consumption and reduce memory burden.

[0006] To achieve the above objectives, the present application provides a data processing method, the method comprising:

[0007] Get interference and covariance matrices;

[0008] Calculating the mutual correlation coefficient matrix and its corresponding related intermediate variables according to the interference and covariance matrix;

[0009] The mutual correlation coefficient matrix is ​​subjected to partial inversion operation and dynamic calibration based on relevant intermediate variables corresponding to the mutual correlation coefficient matrix to obtain an inverse matrix corresponding to the mutual correlation coefficient matrix.

[0010] In an optional embodiment of the present application, the step of calculating the mutual correlation coefficient matrix and its corresponding related intermediate variables according to the interference and covariance matrix includes:

[0011] Reading data at the same position of different matrices in the interference and covariance matrix to obtain parallel read data;

[0012] Calculate the first column of data of the mutual correlation coefficient matrix and its corresponding subsequent calculation related intermediate variables according to the parallel read data;

[0013] The next column of data and its corresponding subsequent calculation related intermediate variables are calculated based on the first column of data and its corresponding subsequent calculation related intermediate variables, until the mutual correlation coefficient matrix and its corresponding related intermediate variables are obtained.

[0014] In an optional embodiment of the present application, the step of reading data at the same position of different matrices in the interference and covariance matrix to obtain parallel reading data includes:

[0015] Retrieving preset position data for the interference and covariance matrix;

[0016] The preset position data are arranged in a target data format, and the preset position data arranged in the target data format are read through a preset row and column mode to obtain the parallel read data.

[0017] In an optional embodiment of the present application, the subsequent calculation-related intermediate variables corresponding to the first column of data include a reciprocal value and a square root reciprocal value, and the step of calculating the first column of data of the mutual correlation coefficient matrix and the subsequent calculation-related intermediate variables corresponding thereto according to the parallel reading data includes:

[0018] Reading the data of the first row and the first column of the interference and covariance matrix in the parallel read data, and dynamically calibrating the data of the first row and the first column to obtain a temporary variable and a first shift value;

[0019] Performing a reciprocal table lookup operation on the temporary variable to obtain the reciprocal value;

[0020] Performing a modulo-2 and integer operation on the first shift value to obtain a second shift value;

[0021] A square root reciprocal table lookup operation is performed on the second shift value and the data in the first row and first column to obtain the square root reciprocal value.

[0022] In an optional embodiment of the present application, the step of calculating the next column of data and its corresponding subsequent calculation-related intermediate variables based on the first column of data and its corresponding subsequent calculation-related intermediate variables includes:

[0023] According to the first column of data and its corresponding subsequent calculation-related intermediate variables, combined with the partial position data of the interference and covariance matrix, at least one of complex multiplication, subtraction, reciprocal table lookup and square root reciprocal table lookup is performed to obtain the next column of data and its corresponding subsequent calculation-related intermediate variables in turn.

[0024] In an optional embodiment of the present application, the step of performing partial inversion operation and dynamic calibration on the mutual correlation coefficient matrix based on the relevant intermediate variables corresponding to the mutual correlation coefficient matrix to obtain the inverse matrix corresponding to the mutual correlation coefficient matrix includes:

[0025] Dividing the mutual correlation coefficient matrix into parts according to the computational complexity to obtain each part matrix;

[0026] Based on the relevant intermediate variables corresponding to the mutual correlation coefficient matrix, the inverse operation is performed on each of the partial matrices to obtain a preliminary inverse matrix;

[0027] The preliminary inverse matrix is ​​dynamically calibrated to obtain an inverse matrix corresponding to the mutual correlation coefficient matrix.

[0028] In an optional embodiment of the present application, the partial matrices include an upper left partial matrix, a lower right partial matrix, and a lower left partial matrix, and the step of performing inverse operations on the partial matrices based on the relevant intermediate variables corresponding to the mutual correlation coefficient matrix to obtain a preliminary inverse matrix includes:

[0029] According to the relevant intermediate variables corresponding to the mutual correlation coefficient matrix, complex multiplication and / or shift operation is performed on the upper left corner partial matrix to obtain an upper left corner preliminary inverse matrix;

[0030] According to the relevant intermediate variables corresponding to the mutual correlation coefficient matrix, complex multiplication and / or shift operation is performed on the lower right corner partial matrix to obtain a lower right corner preliminary inverse matrix;

[0031] According to the relevant intermediate variables corresponding to the mutual correlation coefficient matrix, at least one of complex multiplication, shift operation and accumulation operation is performed on the lower left corner partial matrix to obtain a lower left corner preliminary inverse matrix;

[0032] The preliminary inverse matrix is ​​composed of the upper left preliminary inverse matrix, the lower right preliminary inverse matrix and the lower left preliminary inverse matrix.

[0033] In addition, to achieve the above-mentioned purpose, the present application also proposes a data processing device, the data processing device comprising:

[0034] An acquisition module, used for acquiring interference and covariance matrices;

[0035] A calculation module, used for calculating the mutual correlation coefficient matrix and its corresponding related intermediate variables according to the interference and covariance matrix;

[0036] The inversion module is used to perform partial inversion operation and dynamic calibration on the mutual correlation coefficient matrix based on relevant intermediate variables corresponding to the mutual correlation coefficient matrix to obtain an inverse matrix corresponding to the mutual correlation coefficient matrix.

[0037] In addition, to achieve the above objectives, the present application also proposes a data processing device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the data processing method described above.

[0038] In addition, to achieve the above objectives, the present application also proposes a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the data processing method described above are implemented.

[0039] One or more technical solutions proposed in this application have at least the following technical effects:

[0040] By obtaining interference and covariance matrices; calculating the mutual correlation coefficient matrix and its corresponding related intermediate variables according to the interference and covariance matrix; performing partial inversion operation and dynamic calibration on the mutual correlation coefficient matrix based on the related intermediate variables corresponding to the mutual correlation coefficient matrix to obtain the inverse matrix corresponding to the mutual correlation coefficient matrix; by calculating the mutual correlation coefficient matrix and its corresponding related intermediate variables, the calculation process of the inverse matrix corresponding to the mutual correlation coefficient matrix in the Cholesky decomposition algorithm is decomposed and merged, thereby realizing the fast calculation of Cholesky decomposition, reducing system time consumption and reducing memory burden. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0042] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0043] Figure 1 A flowchart of the first embodiment of the data processing method of the present application is provided;

[0044] Figure 2 It is a schematic diagram of the overall process provided according to the first embodiment of the present application;

[0045] Figure 3 A flowchart of the second embodiment of the data processing method of this application is provided;

[0046] Figure 4 is a schematic diagram of a calculation process according to the second embodiment of the present application;

[0047] Figure 5 Schematic diagram of the covariance matrix arrangement format according to the second embodiment of the present application;

[0048] Figure 6 A flowchart of the third embodiment of the data processing method of this application is provided;

[0049] Figure 7 Schematic diagram of the inversion process according to the third embodiment of the present application;

[0050] Figure 8 is a schematic diagram of matrix splitting according to the third embodiment of the present application;

[0051] Fig. 9 This is a schematic diagram of the module structure of the data processing device according to an embodiment of the present application;

[0052] Fig.10 Schematic diagram of the device structure of the hardware operating environment involved in the data processing method in the embodiment of the present application.

[0053] The purpose, features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0054] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0055] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0056] The main solution of the embodiment of the present application is: by obtaining the interference and covariance matrix; according to the interference and covariance matrix, calculating the mutual correlation coefficient matrix and its corresponding related intermediate variables; based on the related intermediate variables corresponding to the mutual correlation coefficient matrix, performing partial inversion operation and dynamic calibration on the mutual correlation coefficient matrix to obtain the inverse matrix corresponding to the mutual correlation coefficient matrix, by calculating the mutual correlation coefficient matrix and its corresponding related intermediate variables, the calculation process of the inverse matrix corresponding to the mutual correlation coefficient matrix in the Cholesky decomposition algorithm is disassembled and merged, thereby realizing the fast calculation of Cholesky decomposition, so as to reduce system time consumption and reduce memory burden.

[0057] In this embodiment, for the convenience of description, the following description is made with the data processing device as the execution subject.

[0058] At present, in the communication system, in order to merge the multi-antenna equalization IRC, the concept of multi-antenna receiving noise correlation matrix is ​​introduced into the system, and the antenna correlation noise and the interference of the adjacent cells are eliminated by inverting the noise correlation matrix, so as to achieve a better suppression effect. In the whole operation process, due to the high complexity of the inversion operation of the noise correlation matrix, this part of the calculation often becomes the bottleneck of the whole system. At present, the industry generally uses the Cholesky decomposition method or its variant to decompose and invert the matrix.

[0059] The existing technology simplifies matrix decomposition and inversion in theory, which reduces the burden on system resources to a certain extent. However, in actual implementation, it is still time-consuming and occupies a lot of memory. To address this problem, this patent designs a new engineering implementation solution for Cholesky decomposition based on a DSP processor with a SIMD-VLIW architecture with high instruction parallelism, accelerates the Cholesky decomposition, reduces system time consumption, and reduces the memory burden.

[0060] The equalization adopts the MMSE minimum mean square error detection algorithm, and the operator expression is:

[0061] W=(H H H+R e ) -1 H H

[0062] Among them, Re is the noise and interference covariance matrix of the current RE, H is the channel estimation result of multiple antennas and multiple layers on each subcarrier, and H H is the conjugate of H. The detection value of the transmitted signal vector obtained by the MMSE detection algorithm is:

[0063] d mmse =W*r=(H H H+R e ) -1 H H *r

[0064] r is the received data on each subcarrier, according to the Cholesky decomposition formula

[0065] A=LL H

[0066] Perform whitening noise filtering on the channel estimation results and received signals

[0067] H filter =L -1 H

[0068] r filter =L -1 r

[0069] Rewrite the above formula as:

[0070] d mmse =(H H H+R e ) -1 H H *r=(H H H+σ 2 I) -1 (H H *r)

[0071] =[(L -1 H) H L -1 H+I] -1 [(L -1 H) H (L -1 r)

[0072] This transforms the original calculation into finding L -1 , and the operations of matrix multiplication. In this design, the final matrix L is no longer calculated, but the intermediate variables that need to participate in matrix decomposition and inversion are calculated. -1 The calculation process is disassembled and merged, and then the parallel processing advantage of DSP on data and instructions is used to realize the fast calculation of Cholesky decomposition.

[0073] It should be noted that the execution subject of this embodiment may be a computing service device with data processing, network communication and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc., or an electronic device capable of realizing the above functions, a data processing device, etc. The following takes a data processing device as an example to illustrate this embodiment and the following embodiments.

[0074] Based on this, the present application embodiment provides a data processing method, referring to Figure 1 , Figure 1 This is a flowchart of the first embodiment of the data processing method of the present application.

[0075] In this embodiment, the data processing method includes steps S10 to S30:

[0076] Step S10, obtaining interference and covariance matrix;

[0077] Reference Figure 2 , Figure 2 This is a schematic diagram of the overall process provided according to the first embodiment of the present application, such as Figure 2As shown, due to the large number of data variables and complicated calculations, the embodiment of the present application first divides the entire calculation process into two parts based on the variable types involved in the calculation process, the coupling relationship between the variables, combined with the processor architecture and the memory allocation of data, and each part is divided into multiple small calculation modules to achieve the purpose of accelerating the calculation.

[0078] Optionally, in the embodiment of the present application, the input interference and covariance matrix (also called interference and noise variance matrix) is of RBG (Resource Block Group) dimension, and one RBG contains 4 RBs, so the amount of processed data is 1 / 48 of the original algorithm.

[0079] Step S20, calculating the mutual correlation coefficient matrix and its corresponding related intermediate variables according to the interference and covariance matrix;

[0080] Optionally, in an embodiment of the present application, the first part implements the Cholesky decomposition calculation of the intermediate variables involved in subsequent matrix decomposition and inversion. In the embodiment of the present application, the matrix calculation method of [4,4] is used as an example to illustrate, and the L matrix is ​​divided into 4 groups by column for calculation.

[0081] Optionally, to facilitate computational acceleration and parallel computing, for the input interference and noise variance matrix Re of dimension [4,4], the lower triangular data is taken, and the data is arranged into the target data format in advance. Then, relying on the processor's unique row and column mode reading method, the data at the same position of different matrices is read. Multiple groups of data can be read at one time to achieve the purpose of parallel processing.

[0082] Optionally, according to the bit width of the calculated data or input data, and the data placement format, combined with the bit width configuration of the DSP processor, the parallelism of this part is designed to be n in the embodiment of the present application, that is, n groups of data are calculated in parallel in one cycle, which can reduce the time consumption to at least 1 / n of the original time consumption; in order to maximize the use of the computing resources of the processor, in addition to using the method of parallel calculation of multiple groups of data, when writing the code, the calculation process is used as a framework to allocate the calculation of some data to the spare resource components in advance, and the idle resource bubbles in the processing process are used to perform partial or full processing in advance for some data that need to be processed or processed later, so as to reduce unnecessary time-consuming calculations when used later. For example, when calculating the intermediate quantity R11_inv_fixed of the first row and first column of data, considering that there are spare processor components and the data required for the calculation is only used here, this part of the calculation is advanced to the first calculation module, which can save resources and time for subsequent processing. On the other hand, when performing the table lookup calculation involved in the inversion, separating the table lookup calculation from the main process and calculating it separately can simplify the entire processing flow, which not only improves the readability of the code, but also makes the writing of the entire code smoother and more fluent. In addition, for the multiplier unit with the heaviest computational burden, when implementing this application, under the premise of not affecting the overall calculation, by adjusting the order of operations, the multiplication operation is distributed as evenly as possible to multiple multiplier units to reduce the short board effect, make the operation of the entire system module more balanced, and also achieve the maximum utilization of the components.

[0083] Step S30 , performing partial inversion operation and dynamic calibration on the mutual correlation coefficient matrix based on relevant intermediate variables corresponding to the mutual correlation coefficient matrix to obtain an inverse matrix corresponding to the mutual correlation coefficient matrix.

[0084] Optionally, in the embodiment of the present application, the second part implements Cholesky's method for finding L -1 In the first part, various intermediate variables calculated in the first part are used to calculate the final inversion result U. The inversion part is divided into four parts according to the computational complexity of the output lower triangle. The first three parts realize the preliminary calculation of the U matrix, and the fourth part dynamically calibrates the preliminary U matrix to 16 bits.

[0085] Optionally, the U matrix of [4,4] is divided into three parts, namely, the LA part on the upper left, the LB part on the lower right, and the LD part on the lower left, for calculation. In the calculation of this part, in addition to the acceleration ideas used in the first part such as parallel processing of n groups of data, the strength of data processing is also used to optimize the idleness of the operation unit to improve the processing efficiency. For example, in the dynamic calibration of the initially calculated U, the original cycle processes n groups of data in 10 cycles, but the data processing strength is doubled, and 2n groups of data are input at a time. The planning configuration of the previous calculation multiplier unit is optimized, and 2n groups of data can be processed in 14 cycles, and the efficiency is improved by 30%. In addition, by adjusting the data output order and the memory placement order, the time consumption of some calculations can also be optimized. For example, in the calculation of the LD part, when outputting data, the second column of data is output first, and then the first column of data is output. The first column of data is placed behind the second column of data in the memory, so that the output storage unit can be arranged smoothly, which also improves the efficiency of the entire cycle.

[0086] Optionally, the above optimization is mainly aimed at time consumption, with the goal of improving calculation speed and efficiency. In the implementation of Cholesky decomposition, in addition to the huge time consumption, the pressure on the memory is also relatively large due to the large number of intermediate variables that need to be stored. In the embodiment of the present application, under the premise of avoiding memory read and write conflicts, the memory consumption is also optimized by rearranging the order of operations of each variable and reusing the memory. For example, the calculation of R11_inv_fixed mentioned above, by advancing it from the second column calculation module to the first column calculation module, can save a part of the memory for storing the pre-data d1_fixed_tmp; for another example, the intermediate variables L2f_fixed of the second calculation module in the first part and L3f_fixed of the third calculation module have a short existence period, and the data does not conflict, so the two reuse memory, thereby reducing memory consumption.

[0087] This embodiment adopts the above scheme, specifically by obtaining interference and covariance matrices; calculating the mutual correlation coefficient matrix and its corresponding related intermediate variables according to the interference and covariance matrix; performing partial inversion operation and dynamic calibration on the mutual correlation coefficient matrix based on the related intermediate variables corresponding to the mutual correlation coefficient matrix to obtain the inverse matrix corresponding to the mutual correlation coefficient matrix, and by calculating the mutual correlation coefficient matrix and its corresponding related intermediate variables, the calculation process of the inverse matrix corresponding to the mutual correlation coefficient matrix in the Cholesky decomposition algorithm is decomposed and merged, thereby realizing fast calculation of Cholesky decomposition, reducing system time consumption, and reducing memory burden.

[0088] Based on the first embodiment of the present application, the second embodiment of the present application is proposed. In the second embodiment of the present application, the same or similar contents as those of the above-mentioned embodiment 1 can be referred to the above introduction and will not be repeated later. Figure 3 , step S20 also includes steps S201 to S203:

[0089] Step S201, reading data at the same position of different matrices in the interference and covariance matrix to obtain parallel read data;

[0090] Optionally, the step of reading data at the same position of different matrices in the interference and covariance matrix to obtain parallel reading of data includes:

[0091] Retrieving preset position data for the interference and covariance matrix;

[0092] The preset position data are arranged in a target data format, and the preset position data arranged in the target data format are read through a preset row and column mode to obtain the parallel read data.

[0093] Reference Figure 4 , Figure 4 is a schematic diagram of a calculation process according to the second embodiment of the present application, as shown in Figure 4 As shown, the first part implements the Cholesky decomposition calculation of the intermediate variables involved in the subsequent matrix decomposition and inversion. According to the matrix calculation method [4,4], the L matrix is ​​divided into 4 groups by column for calculation.

[0094] Reference Figure 5 , Figure 5 This is a schematic diagram of the covariance matrix arrangement format according to the second embodiment of the present application. In order to facilitate the acceleration and parallel calculation of the calculation, for the input interference and noise variance matrix Re of dimension [4,4], the lower triangular data is taken and the data is arranged in advance as follows: Figure 5 The format shown is then read by relying on the row-column mode reading method unique to the processor in the embodiment of the present application to read the data at the same position of different matrices. Multiple groups of data can be read at one time to achieve the purpose of parallel processing.

[0095] Optionally, according to the bit width of the calculated data or input data, and the data placement format, combined with the bit width configuration of the DSP processor, this solution designs the parallelism of this part to be n, that is, n groups of data are calculated in parallel in one cycle, which can reduce the time consumption to at least 1 / n of the original time consumption; in order to maximize the use of the processor's computing resources, in addition to using the method of parallel calculation of multiple groups of data, when writing the code, the calculation process is used as a framework to allocate the calculation of some data to the spare resource components in advance, and the idle resource bubbles in the processing process are used to perform partial or full processing in advance for some data that need to be processed or processed later, so as to reduce unnecessary time-consuming calculations when used later. For example, when calculating the intermediate quantity R11_inv_fixed of the first row and first column of data, considering that there are spare processor components and the data required for the calculation is only used here, this part of the calculation is advanced to the first calculation module, which can save resources and time for subsequent processing. On the other hand, when performing the table lookup calculation involved in the inversion, the table lookup calculation is separated from the main process and calculated separately, which can simplify the entire processing flow, not only improving the readability of the code, but also making the writing of the entire code smoother and more fluent. In addition, for the multiplier unit with the heaviest computational burden, in the embodiment of the present application, when implementing, by adjusting the order of operations without affecting the overall calculation, the multiplication operation is distributed as evenly as possible to multiple multiplier units to reduce the short board effect, so that the operation of the entire system module is more balanced, and the maximum utilization of the components can be achieved.

[0096] Step S202, calculating the first column of data of the mutual correlation coefficient matrix and its corresponding subsequent calculation-related intermediate variables according to the parallel read data;

[0097] Optionally, the subsequent calculation-related intermediate variables corresponding to the first column of data include a reciprocal value and a square root reciprocal value.

[0098] Optionally, the step of calculating the first column of data of the mutual correlation coefficient matrix and the corresponding subsequent calculation of related intermediate variables according to the parallel read data includes:

[0099] Reading the data of the first row and the first column of the interference and covariance matrix in the parallel read data, and dynamically calibrating the data of the first row and the first column to obtain a temporary variable and a first shift value;

[0100] Performing a reciprocal table lookup operation on the temporary variable to obtain the reciprocal value;

[0101] Performing a modulo-2 and integer operation on the first shift value to obtain a second shift value;

[0102] A square root reciprocal table lookup operation is performed on the second shift value and the data in the first row and first column to obtain the square root reciprocal value.

[0103] Optionally, for the first column, calculate the intermediate variables Q_d1_tmp, Q_d1, L11f_inv_fixed, and R11_inv_fixed. Read the data Re(1,1) of the first row and first column of the n Re matrices, perform 32-bit dynamic calibration on them to obtain the intermediate result d1_fixed_tmp, record the corresponding standard value Q_d1_tmp, perform a reciprocal table lookup on d1_fixed_tmp, and obtain the result R11_inv_fixed; round up Q_d1_tmp / 2 and multiply it by 2 to obtain Q_d1, record the standard value and perform dynamic shift on Re(1,1) to obtain d1_fixed, perform a square root reciprocal table lookup operation on it to obtain L11f_inv_fixed.

[0104] Step S203, calculating the next column of data and its corresponding subsequent calculation related intermediate variables based on the first column of data and its corresponding subsequent calculation related intermediate variables, until the mutual correlation coefficient matrix and its corresponding related intermediate variables are obtained.

[0105] Optionally, the step of calculating the next column of data and its corresponding subsequent calculation-related intermediate variables based on the first column of data and its corresponding subsequent calculation-related intermediate variables includes:

[0106] According to the first column of data and its corresponding subsequent calculation-related intermediate variables, combined with the partial position data of the interference and covariance matrix, at least one of complex multiplication, subtraction, reciprocal table lookup and square root reciprocal table lookup is performed to obtain the next column of data and its corresponding subsequent calculation-related intermediate variables in turn.

[0107] Optionally, since the table lookup operation is relatively complex and the process is relatively long, the second and third columns of calculations separate the table lookup part from other calculations and further divide them into two parts.

[0108] Optionally, for the second column, the first part outputs the intermediate quantities R2f_fixed, L32f_fixed, L42f_fixed, and L2f_fixed. Read n Re(2,1) at a time, take its conjugate, perform 32-bit complex multiplication with R11_inv_fixed, right shift Q_d1_tmp+4bit to get R2f_fixed; perform 32it complex multiplication of R2f_fixed and Re(2,1), expand to 64bit, do not shift, the result is recorded as L2f_fixed_tmp1, subtract L2f_fixed_tmp1 from Re(2,2)*2^26, right shift 26bit, get 32-bit L2f_fixed, remove the imaginary part output; take Re(3,1 ) performs 32-bit complex multiplication with R2f_fixed, expands to 64 bits, does not shift, and the result is recorded as L32f_fixed_tmp1. Subtract L32f_fixed_tmp1 from Re(3,2)*2^26 and shift right 26 bits to get L32f_fixed. The same calculation is used to calculate Re(4,1), Re(4,2), and R2f_fixed to get L42f_fixed. The real numbers involved are copied to two consecutive 32-bit positions through the processor's unique interleaving component to correspond to the 32-bit complex numbers. The second part calculates and outputs L2f_inv_fixed, L22f_inv_fixed, Q_d2, and Q_d2_tmp; performs 32-bit dynamic calibration on L2f_fixed to obtain d2_fixed_tmp, and records the calibration value Q_d2_tmp; performs a reciprocal table lookup on d2_fixed_tmp to obtain L2f_inv_fixed; rounds Q_d2_tmp / 2 up and multiplies it by 2 to obtain Q_d2, performs a 32-bit dynamic shift on L2f_fixed to obtain d2_fixed, and then looks up the first-order square root reciprocal table to obtain L22f_inv_fixed.

[0109] Optionally, for the third column, basically the same as the second column, the first part outputs R3f_fixed, L32XiL2f_fixed, L43f_fixed and L3f_fixed for the second part; the second part outputs L3f_inv_fixed, Q_d3, L33f_inv_fixed, Q_d3_tmp.

[0110] Optionally, the fourth column calculates and outputs R4f_fixed, L42fXiL2f_fixed, L43fXiL3f_fixed, Q_d4, L44_inv_fixed: take the conjugate of Re(4,1), perform 32-bit complex multiplication with R11_inv_fixed, right shift Q_d1_tmp+4bit, and get R4f_fixed; take Re(4,1) and perform 32-bit complex multiplication with R4f_fixed, expand it to 64bit, and the result is Recorded as R41XR4f_fixed; take the conjugate of L42f_fixed, do 32-bit complex multiplication with L2f_inv_fixed, right shift Q_d2_tmp+2bit to get L42fXiL2f_fixed, then use L42f_fixed and L42fXiL2f_fixed to do 32-bit complex multiplication, expand to 64bit without shifting, and the result is L42XL42fXiL2f_fixed; take the conjugate of L43f_fixed Conjugate, do 32-bit complex multiplication with L3f_inv_fixed, right shift Q_d3_tmp+2bit, get L43fXiL3f_fixed, then do 32-bit complex multiplication with L43f_fixed and L43fXiL3f_fixed, expand to 64bit without shifting, and the result is L43XL43fXiL3f_fixed; subtract R41XR4f_fixed from Re(4,4)*2^26, and subtract L42XL42f XiL2f_fixed, subtract L43XL43fXiL3f_fixed, right shift 26 bits, remove the imaginary part of the result, and get L4f_fixed; perform 32-bit dynamic calibration on L4f_fixed, record the calibration value Q_d4_tmp, round up Q_d4_tmp / 2 and multiply by 2 to get Q_d4, then shift L4f_fixed to get d4_fixed; take the square root and inverse of d4_fixed and look up the table to get L44f_inv_fixed.

[0111] This embodiment adopts the above scheme, specifically by reading the data at the same position of different matrices in the interference and covariance matrix, to obtain parallel reading data; calculating the first column data of the mutual correlation coefficient matrix and its corresponding subsequent calculation related intermediate variables according to the parallel reading data; calculating the next column data and its corresponding subsequent calculation related intermediate variables based on the first column data and its corresponding subsequent calculation related intermediate variables, until the mutual correlation coefficient matrix and its corresponding related intermediate variables are obtained, by calculating the intermediate variables that need to participate in matrix decomposition and inversion. In the implementation, by L in the cholesky decomposition algorithm -1The calculation process is disassembled and merged, and then the parallel processing advantage of DSP on data and instructions is used to realize the fast calculation of Cholesky decomposition.

[0112] Based on any of the above embodiments of the present application, a third embodiment of the present application is proposed. In the third embodiment of the present application, the same or similar contents as any of the above embodiments can be referred to the above introduction and will not be repeated in the following. Figure 6 , step S30 also includes steps S301 to S303:

[0113] Step S301, dividing the mutual correlation coefficient matrix into parts according to the computational complexity to obtain matrices of each part;

[0114] Step S302, performing inverse operations on each of the partial matrices based on the relevant intermediate variables corresponding to the mutual correlation coefficient matrix to obtain a preliminary inverse matrix;

[0115] Step S303: dynamically calibrate the preliminary inverse matrix to obtain an inverse matrix corresponding to the mutual correlation coefficient matrix.

[0116] Optionally, the partial matrices include an upper left corner partial matrix (LA part), a lower right corner partial matrix (LB part) and a lower left corner partial matrix (LD part).

[0117] Reference Figure 7 , Figure 7 FIG. 1 is a schematic diagram of an inversion process according to the third embodiment of the present application. Figure 7 As shown, the second part implements Cholesky's method for finding L -1 In the first part, various intermediate variables calculated in the first part are used to calculate the final inversion result U. The inversion part is divided into four parts according to the computational complexity of the output lower triangle. The first three parts realize the preliminary calculation of the U matrix, and the fourth part dynamically calibrates the preliminary U matrix to 16 bits.

[0118] Reference Figure 8 , Figure 8 FIG. 1 is a schematic diagram of matrix splitting according to the third embodiment of the present application. Figure 8 As shown, the [4,4] U matrix is ​​divided into Figure 8As shown, the LA part on the upper left, the LB part on the lower right, and the LD part on the lower left are calculated separately. In the calculation of this part, in addition to the acceleration ideas used in the first part such as parallel processing of n groups of data, the strength of data processing is also used to optimize the idleness of the operation unit to improve the processing efficiency. For example, in the dynamic calibration of the initially calculated U, the original cycle processes n groups of data in 10 cycles, but the data processing strength is doubled, and 2n groups of data are input at a time. The planning and configuration of the calculation multiplier unit before optimization can achieve 14 cycles to process 2n groups of data, and the efficiency is improved by 30%; in addition, by adjusting the data output order and the memory placement order, the time consumption of some calculations can also be optimized. For example, in the calculation of the LD part, when outputting data, the second column of data is output first, and then the first column of data is output. The first column of data is placed behind the second column of data in the memory, so that the output storage unit can be arranged smoothly, which also improves the efficiency of the entire cycle.

[0119] Optionally, the step of performing inverse operations on the partial matrices based on the relevant intermediate variables corresponding to the mutual correlation coefficient matrix to obtain a preliminary inverse matrix comprises:

[0120] According to the relevant intermediate variables corresponding to the mutual correlation coefficient matrix, complex multiplication and / or shift operation is performed on the upper left corner partial matrix to obtain an upper left corner preliminary inverse matrix;

[0121] According to the relevant intermediate variables corresponding to the mutual correlation coefficient matrix, complex multiplication and / or shift operation is performed on the lower right corner partial matrix to obtain a lower right corner preliminary inverse matrix;

[0122] According to the relevant intermediate variables corresponding to the mutual correlation coefficient matrix, at least one of complex multiplication, shift operation and accumulation operation is performed on the lower left corner partial matrix to obtain a lower left corner preliminary inverse matrix;

[0123] The preliminary inverse matrix is ​​composed of the upper left preliminary inverse matrix, the lower right preliminary inverse matrix and the lower left preliminary inverse matrix.

[0124] Optionally, in the LA part, since R2f_fixed is a 32-bit complex number, two sets of data need to be loaded for calculation to correspond to the input of other variables; Q_d1 and Q_d2 are loaded to calculate the shift value; the opposite of the complex conjugate of R2f_fixed is taken, and a 32-bit complex multiplication is performed with L22f_inv_fixed, and the specified number of bits are shifted right to obtain U21; L11f_inv_fixed and L22f_inv_fixed are shifted left by the specified number of bits to obtain U11 and U22.

[0125] Optionally, the calculation of the LB part is basically the same as that of the LA, and U43, U33, and U44 are calculated from L43XiL3f_fixed, L33f_inv_fixed, L44f_inv_fixed, Q_d3, and Q_d4.

[0126] Optionally, the LD part takes the inverse of the conjugate of the complex value of L32fXiL2f_fixed, performs 32-bit complex multiplication with it, and right shifts Q_d3 / 2+19bit to obtain U32; takes the inverse of the conjugate of the complex value of R3f_fixed, performs 32-bit complex multiplication with L33f_inv_fixed, and right shifts Q_d3 / 2+19bit to obtain U31_1, and U32 is the complex conjugate of R2f_fixed. The opposite number is multiplied by 32-bit complex number, shifted right by 26 bits, and U31_2 is obtained. U31_1 is accumulated with U31_2 without shifting, and U31 is obtained. The conjugate of L32fXiL2f_fixed and the conjugate of L43XiL3f_fixed are multiplied by 32-bit complex number, expanded to 64 bits without shifting, and U42_tmp1 is obtained. 2*2^26 is multiplied by the conjugate of the opposite number of L42XiL2f_fixed, and then the conjugate is added. Add U42_tmp1 and right shift 26 bits to get U42_tmp. Perform 32-bit complex multiplication of U42_tmp and L44f_inv_fixed, right shift Q_d4 / 2+19 bits to get U42. Perform 32-bit complex conjugate multiplication of U42_tmp and the conjugate of R2f_fixed, expand to 64 bits without shifting, and get U41_tmp1. The conjugate of L43XiL3f_fixed and The conjugate of R3f_fixed is multiplied by complex numbers, expanded to 64 bits without shifting, and U41_tmp2 is obtained. 2*2^26 is multiplied by the conjugate opposite of R4f_fixed, U41_tmp1 is subtracted, U41_tmp2 is added, and the result is shifted right by 26 bits. The result is recorded as U41_tmp. U41_tmp is multiplied by L44f_inv_fixed by complex numbers, and shifted right by Q_d4 / 2+19 bits to obtain U41. At this point, the inverse matrix U of L is obtained.

[0127] This embodiment adopts the above scheme, specifically by dividing the mutual correlation coefficient matrix according to the computational complexity to obtain each partial matrix; performing inverse operation on each partial matrix based on the relevant intermediate variables corresponding to the mutual correlation coefficient matrix to obtain a preliminary inverse matrix; dynamically calibrating the preliminary inverse matrix to obtain the inverse matrix corresponding to the mutual correlation coefficient matrix. In addition to the acceleration ideas used in the first part such as parallel processing of n groups of data, the strength of data processing is also used to optimize the idleness of the operation unit, thereby improving the processing efficiency, improving the efficiency of the entire cycle, and reducing memory consumption.

[0128] While improving the efficiency through the above data processing method, an optional embodiment of the present application also provides a method for reducing the complexity and cost of the above hardware, which is achieved by lightweight processing of Turbo coding in the running terminal of the data processing method. The data processing method also includes:

[0129] Obtaining a data block to be encoded through a sub-encoding unit in a Turbo encoder;

[0130] Determine at least one original bit sequence of a preset length from the data block to be encoded;

[0131] Acquire a bit coding table and a state value update table according to the preset length, wherein the bit coding table stores mapping relationship data between the original bit sequence, the current state value of the sub-coding unit and the coding bit, and the state value update table stores mapping relationship data between the original bit sequence, the current state value of the sub-coding unit and the updated state value of the sub-coding unit;

[0132] Based on the original bit sequence, the current state value of the sub-coding unit, the bit coding table and the state value update table, each of the original bit sequences in at least one of the original bit sequences is encoded in turn to obtain a coded bit sequence corresponding to each of the original bit sequences, so as to form a coded data block containing each of the coded bit sequences.

[0133] In an optional embodiment of the present application, based on the original bit sequence, the current state value of the sub-coding unit, the bit coding table, and the state value update table, encoding each of the original bit sequences in at least one of the original bit sequences in turn to obtain a coded bit sequence corresponding to each of the original bit sequences to form a coded data block containing each of the coded bit sequences, including:

[0134] Traversing each of the original bit sequences in turn;

[0135] For the traversed target original bit sequence, traverse each original bit in the target original bit sequence in turn;

[0136] For the target original bit in the traversed target original bit sequence, query the bit encoding table based on the target original bit sequence and the current state value of the sub-coding unit to obtain the target encoding bit corresponding to the target original bit; and query the state value update table based on the target original bit sequence and the current state value of the sub-coding unit to obtain the updated state value of the sub-coding unit, and determine the updated state value of the sub-coding unit as the current state value of the sub-coding unit;

[0137] If the coded bits corresponding to the original bits in the target original bit sequence are obtained, then the traversal of the original bits in the target original bit sequence is terminated to obtain the target coded bit sequence corresponding to the target original bit sequence;

[0138] If the coded bit sequences corresponding to the original bit sequences are obtained, the traversal of the original bit sequences is terminated to form the coded data blocks containing the coded bit sequences.

[0139] In an optional embodiment of the present application, the preset length includes a preset first length and a preset second length, the preset first length is greater than the preset second length and is an integer multiple of the preset second length, and the total length of the data block to be encoded is greater than or equal to the preset first length and is an integer multiple of the preset second length; determining at least one original bit sequence of a preset length from the data block to be encoded includes:

[0140] If the total length of the data block to be encoded is not an integer multiple of the preset first length, at least one original bit sequence of the preset first length and one original bit sequence of the preset second length are determined from the data block to be encoded, wherein the original bit sequence of the preset second length is after at least one original bit sequence of the preset first length.

[0141] In an optional embodiment of the present application, the step of obtaining a bit encoding table and a state value update table according to the preset length includes:

[0142] Acquire a first bit encoding table and a first state value update table according to the preset first length, wherein the first bit encoding table stores the mapping relationship data between the original bit sequence of the preset first length, the current state value of the sub-coding unit and the encoding bit, and the first state value update table stores the mapping relationship data between the original bit sequence of the preset first length, the current state value of the sub-coding unit and the updated state value of the sub-coding unit;

[0143] A second bit coding table and a second state value update table are obtained according to the preset second length, wherein the second bit coding table stores the original bit sequence of the preset second length, the mapping relationship data between the current state value of the sub-coding unit and the coding bit, and the second state value update table stores the mapping relationship data between the original bit sequence of the preset second length, the current state value of the sub-coding unit and the updated state value of the sub-coding unit.

[0144] In an optional embodiment of the present application, based on the original bit sequence, the current state value of the sub-coding unit, the bit coding table, and the state value update table, encoding each of the original bit sequences in at least one of the original bit sequences in turn to obtain a coded bit sequence corresponding to each of the original bit sequences to form a coded data block containing each of the coded bit sequences, including:

[0145] Based on the original bit sequence of the preset first length, the current state value of the sub-coding unit, the first bit coding table, and the first state value update table, encoding processing is performed on each original bit sequence of the preset first length in at least one original bit sequence of the preset first length in turn to obtain a coded bit sequence corresponding to each original bit sequence of the preset first length;

[0146] Based on the original bit sequence of the preset second length, the current state value of the sub-coding unit, the second bit coding table and the second state value update table, the original bit sequence of the preset second length is encoded to obtain a coded bit sequence corresponding to the original bit sequence of the preset second length, so as to form the coded data block including the coded bit sequences of the preset first length and the coded bit sequences of the preset second length.

[0147] In an optional embodiment of the present application, the to-be-encoded data block is obtained from an original bit stream; and after the encoding process is performed on each of the original bit sequences in at least one of the original bit sequences in sequence based on the original bit sequence, the current state value of the sub-coding unit, the bit encoding table, and the state value update table to obtain a coded bit sequence corresponding to each of the original bit sequences to form a coded data block containing each of the coded bit sequences, the method further includes:

[0148] Acquire the original bit stream and a check bit stream consisting of the coded data blocks;

[0149] A Turbo hybrid coded bit stream is outputted according to the original bit stream and the check bit stream.

[0150] In an optional embodiment of the present application, the Turbo encoding method is applied to a digital signal processor DSP.

[0151] This application also provides a data processing device, please refer to Fig. 9 , the data processing device comprises:

[0152] An acquisition module 10, used to acquire interference and covariance matrices;

[0153] A calculation module 20, used for calculating the mutual correlation coefficient matrix and its corresponding related intermediate variables according to the interference and covariance matrix;

[0154] The inversion module 30 is used to perform partial inversion operation and dynamic calibration on the mutual correlation coefficient matrix based on relevant intermediate variables corresponding to the mutual correlation coefficient matrix to obtain an inverse matrix corresponding to the mutual correlation coefficient matrix.

[0155] The data processing device provided by the present application adopts the data processing method in the above embodiment to solve the technical problem of data processing. Compared with the prior art, the beneficial effects of the data processing device provided by the present application are the same as the beneficial effects of the data processing method provided by the above embodiment, and other technical features in the data processing device are the same as the features disclosed in the above embodiment method, which will not be described in detail here.

[0156] The present application provides a data processing device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the data processing method in the above-mentioned embodiment 1.

[0157] Reference below Fig.10 , which shows a schematic diagram of the structure of a data processing device suitable for implementing an embodiment of the present application. The data processing device in the embodiment of the present application may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Fig.10 The data processing device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0158] like Fig.10As shown, the data processing device may include a processing device 1001 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. In RAM1004, various programs and data required for the operation of the data processing device are also stored. The processing device 1001, ROM1002, and RAM1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems can be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touch pad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD: Liquid Crystal Display), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the data processing device to communicate with other devices wirelessly or by wire to exchange data. Although the figure shows a data processing device with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented or have alternatively.

[0159] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0160] The data processing device provided by the present application adopts the data processing method in the above embodiment to solve the technical problem of data processing. Compared with the prior art, the beneficial effects of the data processing device provided by the present application are the same as the beneficial effects of the data processing method provided by the above embodiment, and other technical features in the data processing device are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0161] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0162] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0163] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, computer programs) stored thereon, wherein the computer-readable program instructions are used to execute the data processing method in the above-mentioned embodiment.

[0164] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination of the above.

[0165] The computer-readable storage medium may be included in a data processing device, or may exist independently without being incorporated into a data processing device.

[0166] The computer-readable storage medium carries one or more programs. When the one or more programs are executed by a data processing device, the data processing device: obtains interference and covariance matrices; calculates a mutual correlation coefficient matrix and its corresponding related intermediate variables according to the interference and covariance matrix; performs partial inversion operations and dynamic calibration on the mutual correlation coefficient matrix based on the related intermediate variables corresponding to the mutual correlation coefficient matrix to obtain an inverse matrix corresponding to the mutual correlation coefficient matrix; by calculating the mutual correlation coefficient matrix and its corresponding related intermediate variables, the calculation process of the inverse matrix corresponding to the mutual correlation coefficient matrix in the Cholesky decomposition algorithm is decomposed and merged, thereby realizing fast calculation of the Cholesky decomposition, reducing system time consumption, and reducing memory burden.

[0167] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0168] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0169] The modules involved in the embodiments described in this application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0170] The readable storage medium provided in the present application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned data processing method, and can solve the technical problems of data processing. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in the present application are the same as the beneficial effects of the data processing method provided in the above-mentioned embodiment, and will not be repeated here.

[0171] The present application also provides a computer program product, including a computer program, which implements the steps of the above-mentioned data processing method when executed by a processor.

[0172] The computer program product provided by this application can solve the technical problem of data processing. Compared with the prior art, the beneficial effects of the computer program product provided by this application are the same as the beneficial effects of the data processing method provided by the above embodiment, which will not be repeated here.

[0173] The above descriptions are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect applications in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A data processing method, characterized in that: The method comprises: Get interference and covariance matrices; Calculating the mutual correlation coefficient matrix and its corresponding related intermediate variables according to the interference and covariance matrix; The mutual correlation coefficient matrix is ​​subjected to partial inversion operation and dynamic calibration based on relevant intermediate variables corresponding to the mutual correlation coefficient matrix to obtain an inverse matrix corresponding to the mutual correlation coefficient matrix.

2. The method according to claim 1, characterized in that The step of calculating the mutual correlation coefficient matrix and its corresponding related intermediate variables according to the interference and covariance matrix comprises: Reading data at the same position of different matrices in the interference and covariance matrix to obtain parallel read data; Calculate the first column of data of the mutual correlation coefficient matrix and its corresponding subsequent calculation related intermediate variables according to the parallel read data; The next column of data and its corresponding subsequent calculation related intermediate variables are calculated based on the first column of data and its corresponding subsequent calculation related intermediate variables, until the mutual correlation coefficient matrix and its corresponding related intermediate variables are obtained.

3. The method according to claim 2, characterized in that The step of reading data at the same position of different matrices in the interference and covariance matrix to obtain parallel reading of data comprises: Retrieving preset position data for the interference and covariance matrix; The preset position data are arranged in a target data format, and the preset position data arranged in the target data format are read through a preset row and column mode to obtain the parallel read data.

4. The method according to claim 2, characterized in that The subsequent calculation-related intermediate variables corresponding to the first column of data include a reciprocal value and a square root reciprocal value, and the step of calculating the first column of data of the mutual correlation coefficient matrix and the subsequent calculation-related intermediate variables corresponding thereto according to the parallel reading data includes: Reading the data of the first row and the first column of the interference and covariance matrix in the parallel read data, and dynamically calibrating the data of the first row and the first column to obtain a temporary variable and a first shift value; Performing a reciprocal table lookup operation on the temporary variable to obtain the reciprocal value; Performing a modulo-2 and integer operation on the first shift value to obtain a second shift value; A square root reciprocal table lookup operation is performed on the second shift value and the data in the first row and first column to obtain the square root reciprocal value.

5. The method according to claim 2, characterized in that The step of calculating the next column of data and its corresponding subsequent calculation-related intermediate variables based on the first column of data and its corresponding subsequent calculation-related intermediate variables comprises: According to the first column of data and its corresponding subsequent calculation-related intermediate variables, combined with the partial position data of the interference and covariance matrix, at least one of complex multiplication, subtraction, reciprocal table lookup and square root reciprocal table lookup is performed to obtain the next column of data and its corresponding subsequent calculation-related intermediate variables in turn.

6. The method according to claim 1, characterized in that The step of performing partial inversion operation and dynamic calibration on the mutual correlation coefficient matrix based on the relevant intermediate variables corresponding to the mutual correlation coefficient matrix to obtain the inverse matrix corresponding to the mutual correlation coefficient matrix comprises: Dividing the mutual correlation coefficient matrix into parts according to the computational complexity to obtain each part matrix; Based on the relevant intermediate variables corresponding to the mutual correlation coefficient matrix, the inverse operation is performed on each of the partial matrices to obtain a preliminary inverse matrix; The preliminary inverse matrix is ​​dynamically calibrated to obtain an inverse matrix corresponding to the mutual correlation coefficient matrix.

7. The method according to claim 6, characterized in that The partial matrices include an upper left corner partial matrix, a lower right corner partial matrix and a lower left corner partial matrix, and the step of performing inverse operations on the partial matrices based on the relevant intermediate variables corresponding to the mutual correlation coefficient matrix to obtain a preliminary inverse matrix includes: According to the relevant intermediate variables corresponding to the mutual correlation coefficient matrix, complex multiplication and / or shift operation is performed on the upper left corner partial matrix to obtain an upper left corner preliminary inverse matrix; According to the relevant intermediate variables corresponding to the mutual correlation coefficient matrix, complex multiplication and / or shift operation is performed on the lower right corner partial matrix to obtain a lower right corner preliminary inverse matrix; According to the relevant intermediate variables corresponding to the mutual correlation coefficient matrix, at least one of complex multiplication, shift operation and accumulation operation is performed on the lower left corner partial matrix to obtain a lower left corner preliminary inverse matrix; The preliminary inverse matrix is ​​composed of the upper left preliminary inverse matrix, the lower right preliminary inverse matrix and the lower left preliminary inverse matrix.

8. A data processing device, characterized in that: The device comprises: An acquisition module, used for acquiring interference and covariance matrices; A calculation module, used for calculating the mutual correlation coefficient matrix and its corresponding related intermediate variables according to the interference and covariance matrix; The inversion module is used to perform partial inversion operation and dynamic calibration on the mutual correlation coefficient matrix based on relevant intermediate variables corresponding to the mutual correlation coefficient matrix to obtain an inverse matrix corresponding to the mutual correlation coefficient matrix.

9. A data processing device, characterized in that: The device comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the data processing method according to any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the data processing method according to any one of claims 1 to 7 are implemented.