WO2024061266A1 - Mlse均衡器的实现方法和芯片、电子设备、计算机可读介质 - Google Patents

Mlse均衡器的实现方法和芯片、电子设备、计算机可读介质 Download PDF

Info

Publication number
WO2024061266A1
WO2024061266A1 PCT/CN2023/120012 CN2023120012W WO2024061266A1 WO 2024061266 A1 WO2024061266 A1 WO 2024061266A1 CN 2023120012 W CN2023120012 W CN 2023120012W WO 2024061266 A1 WO2024061266 A1 WO 2024061266A1
Authority
WO
WIPO (PCT)
Prior art keywords
path
symbol
optimal
group
under
Prior art date
Application number
PCT/CN2023/120012
Other languages
English (en)
French (fr)
Inventor
钟一鸣
李运鹏
Original Assignee
中兴通讯股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 中兴通讯股份有限公司 filed Critical 中兴通讯股份有限公司
Publication of WO2024061266A1 publication Critical patent/WO2024061266A1/zh

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L25/00Baseband systems
    • H04L25/02Details ; arrangements for supplying electrical power along data transmission lines
    • H04L25/03Shaping networks in transmitter or receiver, e.g. adaptive shaping networks

Definitions

  • This application relates to but is not limited to the field of information processing technology.
  • ISI inter-symbol crosstalk
  • noise interference noise interference
  • attenuation of signal power etc. Therefore, in the presence of so many multiple distortion sources, signal compensation equalization technology has become an indispensable component of almost all communication systems.
  • Maximum Likelihood Sequence Estimation (MLSE) equalizer is one of the equalizers with excellent performance, which can effectively compensate for the nonlinearity and signal power of the receiving side signal.
  • DFE Decision Feedback Equalizer
  • the performance of the MLSE equalizer is better than the decision feedback equalizer (DFE, Decision Feedback Equalizer) in compensating the nonlinearity and mid-to-high frequency components of the signal.
  • DFE Decision Feedback Equalizer
  • the MLSE algorithm is not suitable for the chip.
  • the memory, computing and other resource consumption is much greater than that of the DFE equalizer.
  • the number of tap coefficients and calculation depth are directly proportional to the complexity of the algorithm implementation and the consumption of chip memory, computing and other resources.
  • the MLSE equalizer For example, every time the number of tap coefficients of the MLSE equalizer increases by one, the number of required branch measurement sub-modules, plus ratio selection sub-modules and path management sub-modules needs to be doubled. Therefore, when using an MLSE algorithm with better performance and a larger number of tap coefficients, the MLSE equalizer's demand for data storage, computing and other resources also increases exponentially, which is proposed for chips with limited area, power consumption and computing resources. faced a severe test.
  • This application provides an implementation method of an MLSE equalizer, a chip, an electronic device, and a computer-readable medium.
  • this application provides a method for implementing an MLSE equalizer, which includes: determining the first path corresponding to the k-th symbol that requires calculation of the branch path metric value according to the optimal path of the k-1th symbol; where, k is an integer greater than or equal to 2; calculate the branch path metric value of the k-th symbol under the first path; according to the branch path metric value of the k-th symbol under the first path and the The optimal cumulative path distance of the k-1th symbol under the optimal path of the k-1th symbol determines the optimal path of the k-th symbol and the k-th symbol under the optimal path of the k-th symbol. The optimal accumulated path distance under the optimal path; storing the optimal path of the k-th symbol and the optimal accumulated path distance of the k-th symbol under the optimal path of the k-th symbol.
  • the present application provides an electronic device, including: at least one processor; and a memory, at least one program is stored on the memory, and when the at least one program is executed by the at least one processor, the method described herein is implemented. Any implementation method of MLSE equalizer.
  • the present application provides a computer-readable medium.
  • a computer program is stored on the computer-readable medium.
  • the computer program is executed by a processor, the method for implementing any of the MLSE equalizers described herein is implemented.
  • this application provides a maximum likelihood estimation equalizer implementation chip, including: a data interface and a processing circuit; the processing circuit is configured to read instructions stored on a peripheral control circuit through the data interface, and execute The implementation method of any maximum likelihood estimation equalizer described in this article.
  • Figure 1 is a schematic diagram of the implementation of an MLSE equalizer in related technologies
  • Figure 2 is a flow chart of the implementation method of the MLSE equalizer provided by this application;
  • Figure 3 is a block diagram of the implementation chip of the MLSE equalizer provided by this application.
  • FIG. 4 is a block diagram of the processing circuit of the present application.
  • Figure 1 is a schematic diagram of the implementation of an MLSE equalizer in related technologies. As shown in Figure 1, the MLSE equalizer includes four sub-modules: branch metric sub-module 101, plus ratio selection sub-module 102, path management sub-module 103 and output decision sub-module 104.
  • X(k) is the signal to be equalized input to the MLSE equalizer
  • Y(k) is the level signal after decoding and output decision by the MLSE equalizer.
  • the branch metric submodule 101 is configured to calculate the k-th symbol in different paths
  • the branch path metric below.
  • the distance between sequence sampling points is used to represent the branch path metric value.
  • the distance can be any one of Euclidean distance, Manhattan distance, Chebyshev distance, etc.
  • the branch path metric value calculation formula is as formula (1).
  • x ki is the possible output level of the standard constellation point corresponding to the ki-th symbol.
  • each symbol has two possible outputs Levels, +1 and -1 respectively, x(k) is the k-th symbol of the currently input signal to be equalized, n is the number of tap coefficients of MLSE, ⁇ i is the symbol crosstalk coefficient, that is, the tap coefficient.
  • each symbol has m possible output levels, and there are m n branch metric values in formula (1). If each branch metric submodule 101 calculates one branch metric value, then m n branch metric submodules 101 are required. Taking NRZ modulation as an example, each symbol has 2 possible output levels.
  • the plus selection sub-module 102 is configured to accumulate the optimal value of the previous symbol under the path corresponding to the branch path metric value according to the branch path metric value of the k-th symbol output by the branch metric sub-module 101. Accumulate the path distance, and then select the optimal path for the current symbol (that is, the surviving path).
  • D(x k-(n-2) , x k-(n-3) ,..., x k ) is the k-th symbol in the path x k-(n-2) x k-(n-3) ...the optimal cumulative path distance under x k
  • D(x k-(n-1) , x k-(n-2) ,..., x k-1 ) is the optimal cumulative path distance of the k-1th symbol under the path x k-(n-1) x k-(n-2) ...x k-1
  • P(x k-(n- 1) , x k-(n-2) ,...,x k ) is the branch path metric value of the k-th symbol under the path x k-(n-1) x k-(n-2) ...x k .
  • each addition selection sub-module 102 calculates a possible value of D(x k-(n-2) , x k-(n-3) ,..., x k ), then m n-1 addition selection sub-modules 102.
  • the number of ratio selection sub-modules 102 required is 2 n-1 .
  • the path management sub-module 103 is configured to store the surviving path nodes selected by the comparison selection sub-module 102 and the optimal path values under different paths, and subsequently perform backtracking decoding related operations and output the decision decoding value.
  • the number of optimal path values stored in the path management submodule 103 is m n-1 .
  • the judgment output sub-module 104 is configured to judge and output the judgment decoding value in the survivor path to obtain a judged level signal.
  • Figure 2 is a flow chart of the implementation method of the MLSE equalizer provided by this application.
  • a method for implementing an MLSE equalizer comprising steps 200 to 203 .
  • step 200 the first path corresponding to the k-th symbol that requires calculation of the branch path metric is determined based on the optimal path of the k-1th symbol; where k is an integer greater than or equal to 2.
  • the number of first paths is less than the number of second paths, and the values of all symbols in the second paths include m possible output levels; m is an integer greater than or equal to 1.
  • the number of the first paths is less than the number of the second paths.
  • the second paths are the paths for which branch path metric values need to be calculated in related technologies, which reduces the number of paths for which branch path metric values need to be calculated, thereby further reducing the need to determine and store The optimal number of paths thus reduces the required computing and storage resources.
  • the value of the first target symbol in the first path is the same as the value of the corresponding symbol in the optimal path of the k-1th symbol; wherein the first target symbol includes the k-( n-1) symbols to the kt-th symbol; n is the number of tap coefficients, n is less than or equal to k, and t is an integer greater than or equal to 1 and less than or equal to n-2; the value of the second target symbol in the first path includes m possible output levels of the standard constellation point; where, the The second target symbol includes the k-t+1-th symbol to the k-th symbol.
  • the first target symbol in the first path has only one value
  • the second target symbol has m values
  • the first target symbol includes: k-(n-1)th symbol to k-1th symbol
  • the second target symbol includes: k-th symbol. That is the case where t takes a value of 1.
  • m is the number of possible output levels of each symbol corresponding to the modulation mode of the signal to be equalized. For example, for the NRZ modulation mode, the value of m is 2.
  • the embodiment of the present application does not limit the modulation method.
  • it can be any of NRZ, Quadrature Amplitude Modulation (QAM, Quadrature Amplitude Modulation), PSK (Phase Shift Keying), N-level pulse amplitude modulation (PAM-N, N Level Pulse Amplitude Modulation), etc. A sort of.
  • n 3 x k-2 x k-1 for the k-1th symbol. Assuming that these two optimal paths are (-1, -1, -1) and (-1, +1, +1), then based on these two optimal paths, it can be determined that the first path includes four paths: (-1, -1, -1), (-1, -1, +1), (+1, +1, -1) and (+1, +1, +1), which is the case where t is 1 above.
  • n 4, then there are four optimal paths x k-4 x k-3 x k-2 x k-1 for the k-1th symbol. Assume that these four optimal paths are (-1, -1, -1, -1), (+1, -1, -1, +1), (-1, +1, +1, +1) and (+1, +1, +1, -1), then According to these four optimal paths, it can be determined that the first path includes (-1, -1, -1, -1, +1), (-1, -1, +1, +1), (-1, -1, +1, -1), (+1, +1, +1, +1), (+1, +1, +1, -1), (+1, There are eight paths: +1, -1, +1) and (+1, +1, -1, -1), which is the case where the value of t is 2 above.
  • step 201 the branch path metric value of the k-th symbol under the first path is calculated.
  • the branch path metric value of the k-th symbol under the first path is calculated according to formula (1).
  • step 202 determine the k-th symbol based on the branch path metric value of the k-th symbol under the first path and the optimal cumulative path distance of the k-1-th symbol under the optimal path of the k-1-th symbol.
  • the optimal path and the optimal cumulative path distance of the k-th symbol under the optimal path of the k-th symbol are the optimal cumulative path distance of the k-th symbol under the optimal path of the k-th symbol.
  • the optimal path of the k-th symbol and the optimal cumulative path distance of the k-th symbol under the optimal path of the k-th symbol include: grouping the first path, and the paths in the same group include the first path in the first path.
  • the selected path includes: the value from the k-t+1st symbol to the k-1th symbol in the optimal path of the k-1st symbol and the k-t+1th symbol in the path within the group
  • the first path is x k-(n-1) x k-(n-2) .........x kt x k-t+1 .........x k-1 x k
  • the optimal path of the k-1th symbol is x k-1-(n-1) x k-1- (n-2) .........x k-1-t x kt .........x k-1
  • the portion of x k-(n-1) x k-(n-2) .........x kt in the path belonging to the same group in the first path is the same as the portion of x k- (n-1) x k-(n-2) .
  • x kt in the optimal path of the k-1th symbol and the portion of x k-t+1 .........x k -1 x k in the path belonging to the same group in the first path adopts the possible output levels of the standard constellation points.
  • the branch path metric value of the k-th symbol under the path within the group and the k-1th symbol are selected from the optimal path of the k-1th symbol.
  • the optimal cumulative path distance under the selected path. Determining the optimal path and the optimal cumulative path distance corresponding to the group includes: for each path in the group, calculate the branch path metric value of the k-th symbol under the path and the k-th The sum of the optimal cumulative path distances of -1 symbol under the path selected from the optimal path of the k-1th symbol, the calculated sum value is used as the cumulative path distance of the k-th symbol under the path ; Use the path corresponding to the smallest accumulated path distance among the paths in the group as the optimal path of the corresponding group; use the minimum value of the accumulated path distance among the paths in the group as the optimal accumulated path distance of the corresponding group.
  • n no matter how large the value of n is, for the calculation of each symbol, only m t+1 branch path metric values need to be calculated, that is, m t+1 first paths, and m t optimal paths are saved. and the optimal cumulative path distance, which reduces the number of branch path metric values that need to be calculated, as well as the number of saved optimal paths and optimal cumulative path distances, that is, reducing computing resources and storage resources.
  • the optimal path of the k-th symbol and the optimal cumulative path distance of the k-th symbol under the optimal path of the k-th symbol include: grouping the first path, and the paths in the same group include selection from the first path Paths with the same value for the k-th symbol; for paths within the same group, the branch path metric value of the k-th symbol under the path within the group and the k-1th symbol under the path from the k-1th symbol
  • the optimal cumulative path distance under the path selected from the optimal path determines the optimal path and the optimal cumulative path distance corresponding to the group; where, the path selected from the optimal path of the k-1th symbol includes: The value of the k-1th symbol in the optimal path of k-1 symbols is the same as the value of the k-1th symbol in the path within the group; the optimal
  • x k-(n-1) x k-(n- 2) ...the part of x k-1 is the same as the part of x k-(n-1) x k-(n-2) ...x k-1 in the optimal path of the k-1th symbol
  • the part of x k in the fifth path belonging to the same group adopts the possible output of the standard constellation point
  • the optimal accumulated path distance determines the corresponding optimal path and the optimal accumulated path distance within the group including: for each path within the group, calculate the branch path metric value and k-1th symbol under the path within the group.
  • the sum of the optimal cumulative path distances of symbols under the path selected from the optimal path of the k-1th symbol, and the calculated sum value is used as the cumulative path distance of the k-th symbol under the path;
  • the path corresponding to the smallest accumulated path distance among the paths is regarded as the optimal path of the corresponding group;
  • the minimum value of the accumulated path distance among the paths within the group is regarded as the optimal accumulated path distance of the corresponding group.
  • n no matter how large the value of n is, for the calculation of each symbol, only m 2 branch path metric values need to be calculated, that is, m 2 first paths, and m optimal paths and optimal cumulative paths are saved. distance, reducing the number of branch path metrics that need to be calculated, as well as the number of optimal path and optimal cumulative path distances that are saved.
  • step 203 the optimal path of the k-th symbol and the optimal accumulated path distance of the k-th symbol under the optimal path of the k-th symbol are stored.
  • the value of each symbol in the optimal path may be cached.
  • the method further includes: according to the k-th symbol The optimal accumulated path distance under the optimal path of the k-th symbol obtains the determined level signal.
  • obtaining the determined level signal based on the optimal cumulative path distance of the k-th symbol under the k-th symbol's optimal path includes: The optimal accumulated path distance under the optimal path is subjected to relevant operations of backtracking decoding to obtain the decision decoding value, and the decision decoding value is output to obtain the post-decision level signal.
  • the implementation method of the MLSE equalizer determines the branch path metric value that needs to be calculated corresponding to the kth symbol based on the optimal path of the k-1th symbol
  • the first path of the branch path metric is reduced, thereby reducing the number of paths that need to calculate the branch path metric value, thereby further reducing the number of optimal paths that need to be determined and stored, thereby reducing the required computing resources and storage resources.
  • Each symbol has two values: +1 and -1.
  • the standard constellation point corresponding to the first symbol of the signal to be equalized has two values, which are +1 and -1.
  • the first symbol does not have the optimal path and optimal accumulation path distance of the previous symbol. Initialize the optimal accumulation.
  • the path distance is 0.
  • the branch path metric value corresponding to each first path is calculated according to the value of the second symbol of the signal to be equalized.
  • the optimal path of group 1 and the optimal path of group 2 form the optimal path of the second symbol, and the optimal accumulated path distance of group 1 and the optimal accumulated path distance of group 2 form the optimal accumulated path of the second symbol. distance.
  • the path from the first symbol to the second symbol in the first path for which the branch path metric needs to be calculated for the third symbol adopts the optimal path of the second symbol.
  • the branch path metric value corresponding to each first path is calculated according to the value of the third symbol of the signal to be equalized.
  • the optimal path of group 1 and the optimal path of group 2 form the optimal path of the third symbol
  • the optimal accumulated path distance of group 1 and the optimal accumulated path distance of group 2 form the optimal accumulated path of the third symbol. distance.
  • the branch path metric value corresponding to each first path is calculated according to the value of the fourth symbol of the signal to be equalized.
  • the optimal path of group 1 and the optimal path of group 2 form the optimal path of the fourth symbol
  • the optimal accumulated path distance of group 1 and the optimal accumulated path distance of group 2 form the optimal accumulated path of the fourth symbol. distance.
  • the standard constellation point corresponding to the first symbol of the signal to be equalized has two values, which are +1 and -1.
  • the first symbol does not have the optimal path and optimal accumulation path distance of the previous symbol. Initialize the optimal accumulation.
  • the path distance is 0.
  • the branch path metric value corresponding to each first path is calculated according to the value of the second symbol of the signal to be equalized.
  • the optimal path of group 1 and the optimal path of group 2 form the optimal path of the second symbol, and the optimal accumulated path distance of group 1 and the optimal accumulated path distance of group 2 form the optimal accumulated path of the second symbol. distance.
  • the path from the first symbol to the second symbol in the first path for which the branch path metric needs to be calculated for the third symbol adopts the optimal path of the second symbol.
  • the branch path metric value corresponding to each first path is calculated according to the value of the third symbol of the signal to be equalized.
  • the optimal path of group 1 and the optimal path of group 2 form the optimal path of the third symbol
  • the optimal accumulated path distance of group 1 and the optimal accumulated path distance of group 2 form the optimal accumulated path of the third symbol. distance.
  • the branch path metric value corresponding to each first path is calculated according to the value of the fourth symbol of the signal to be equalized.
  • the optimal path of group 1, the optimal path of group 2, the optimal path of group 3 and the optimal path of group 4 form the optimal path of the fourth symbol.
  • the optimal cumulative path distance of group 1, the optimal path of group 2 The optimal accumulated path distance, the optimal accumulated path distance of group 3 and the optimal accumulated path distance of group 4 constitute the optimal accumulated path distance of the fourth symbol.
  • another embodiment of the present application provides an electronic device, comprising: at least one processor; a memory, wherein at least one program is stored in the memory, and when the at least one program is executed by the at least one processor, any one of the above-mentioned methods for implementing the MLSE equalizer is implemented.
  • the processor is a device with data processing capabilities, including but not limited to a central processing unit (CPU), etc.
  • the memory is a device with data storage capabilities, including but not limited to random access memory (RAM, more specifically such as SDRAM). , DDR, etc.), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory (FLASH).
  • RAM random access memory
  • ROM read-only memory
  • EEPROM electrically erasable programmable read-only memory
  • FLASH flash memory
  • the processor and the memory are connected to each other through a bus and are further connected to other components of the computing device.
  • another embodiment of the present application provides a computer-readable medium.
  • a computer program is stored on the computer-readable medium.
  • the computer program is executed by a processor, the method for implementing any of the above MLSE equalizers is implemented.
  • Figure 3 is a block diagram of the implementation chip of the MLSE equalizer provided by this application.
  • another embodiment of the present application provides an implementation chip of an MLSE equalizer, including: a data interface 301 and a processing circuit 302; the processing circuit is configured to read instructions stored on the peripheral control circuit through the data interface, and execute the above Any implementation method of MLSE equalizer.
  • the processing circuit 302 includes: a forward path indication subcircuit 401 configured to determine the required calculation branch corresponding to the k-th symbol according to the optimal path of the k-1th symbol.
  • the branch metric subcircuit 402 is configured to calculate the branch path metric value of the k-th symbol under the first path; plus ratio
  • the sub-selection circuit 403 is configured to calculate the branch path metric value of the k-th symbol under the first path and the k-1th symbol under the optimal path of the k-1th symbol.
  • the optimal accumulated path distance determines the optimal path of the k-th symbol and the optimal accumulated path distance of the k-th symbol under the optimal path of the k-th symbol; the path management subcircuit 404 is configured to store The optimal path of the k-th symbol and the optimal accumulated path distance of the k-th symbol under the optimal path of the k-th symbol.
  • the MLSE equalizer is implemented on a chip, such as an Application Specific Integrated Circuit (ASIC) chip or a Field-Programmable Gate Array (FPGA) chip.
  • ASIC Application Specific Integrated Circuit
  • FPGA Field-Programmable Gate Array
  • an output decision subcircuit 405 configured to The determined level signal is obtained based on the optimal accumulated path distance of the k-th symbol under the optimal path of the k-th symbol.
  • the number of first paths is less than the number of second paths, and the values of all symbols in the second paths include m possible output levels; m is an integer greater than or equal to 1.
  • the value of the first target symbol in the first path is the same as the value of the corresponding symbol in the optimal path of the k-1th symbol; wherein the first target symbol includes The k-(n-1)th symbol to the k-tth symbol; n is the number of tap coefficients, n is less than or equal to k, t is an integer greater than or equal to 1, and less than or equal to n-2; the first The values of the second target symbol in the path include m possible output levels of the standard constellation point; wherein the second target symbol includes the k-t+1-th symbol to the k-th symbol.
  • the gab selection sub-circuit 403 is configured to adopt the following manner to implement the branch path metric value of the k-th symbol under the first path and the k-1th symbol under the first path.
  • the optimal cumulative path distance under the optimal path of the k-1th symbol is determined to determine the optimal path of the k-th symbol and the optimal path of the k-th symbol under the optimal path of the k-th symbol.
  • the first paths, and the paths in the same group include paths with the same value from the k-t+1th symbol to the k-th symbol in the first path; for the paths in the same group Path, based on the branch path metric value of the k-th symbol under the path within the group and the optimal cumulative path of the k-1th symbol under the path selected from the optimal path of the k-1th symbol
  • the distance determines the optimal path corresponding to the group and the optimal accumulated path distance; wherein, the path selected from the optimal path of the k-1th symbol includes: the optimal path of the k-1th symbol Paths whose values from the k-t+1th symbol to the k-1th symbol are the same as the values from the k-t+1th symbol to the k-1th symbol in the path within the group; all groups
  • the corresponding optimal path constitutes the optimal path of the k-th symbol; the optimal accumulated path distance corresponding to all groups constitutes the optimal accumulated path of the k-th symbol under the optimal path of the group
  • the gab selection sub-circuit 403 is configured to adopt the following manner to implement the branch path metric value of the k-th symbol under the path within the group and the k-1th symbol under the path from the group.
  • the path selected from the optimal path of the k-1th symbol The optimal accumulated path distance under the path determines the optimal path and the optimal accumulated path distance corresponding to the group: for each path in the group, calculate the branch path metric value of the k-th symbol under the path and the The sum of the optimal accumulated path distances of the k-1th symbol under the path selected from the optimal path of the k-1th symbol, the calculated sum value is used as the calculated sum of the k-th symbol in the
  • the accumulated path distance under the path; the path corresponding to the minimum accumulated path distance among the paths within the group is regarded as the optimal path of the corresponding group; the minimum value of the accumulated path distance among the paths within the group is regarded as the optimal path for the corresponding group Accumulate path distance.
  • the first target symbol when t is equal to 1, includes: the k-(n-1)th symbol to the k-1th symbol, and the second The target symbols include: the k-th symbol.
  • the plus bit selection sub-circuit 403 is configured to adopt the following manner to implement the branch path metric value of the k-th symbol under the first path and the k-1th symbol under the first path.
  • the optimal cumulative path distance under the optimal path of the k-1th symbol is determined to determine the optimal path of the k-th symbol and the optimal path of the k-th symbol under the optimal path of the k-th symbol.
  • Accumulated path distance Group the first path.
  • Paths in the same group include paths with the same value of the k-th symbol in the first path; for paths in the same group, according to the k-th symbol in the group, The branch path metric value under the path and the optimal cumulative path distance of the k-1th symbol under the path selected from the optimal path of the k-1th symbol determine the optimal path corresponding to the group and the optimal accumulated path distance; wherein, the path selected from the optimal path of the k-1th symbol includes: the selection of the k-1th symbol in the optimal path of the k-1th symbol.
  • Such software may be distributed on computer-readable media, which may include computer storage media (or non-transitory media) and communication media (or transitory media).
  • computer storage media includes volatile and nonvolatile media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. removable, removable and non-removable media.
  • Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, Digital Versatile Disk (DVD) or other optical disk storage, magnetic cassettes, tapes, disk storage or other magnetic storage, or may be used Any other medium that stores the desired information and can be accessed by a computer.
  • communication media typically embodies computer readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media .
  • Example embodiments have been disclosed herein, and although specific terms are employed, they are used and should be interpreted in a general illustrative sense only and not for purpose of limitation. In some instances, it will be apparent to those skilled in the art that features, characteristics and/or elements described in connection with a particular embodiment may be used alone, or may be used in conjunction with other embodiments, unless expressly stated otherwise. Features and/or components are used in combination. Accordingly, it will be understood by those skilled in the art that various changes in form and details may be made without departing from the scope of the present application as set forth in the appended claims.

Landscapes

  • Engineering & Computer Science (AREA)
  • Power Engineering (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Digital Transmission Methods That Use Modulated Carrier Waves (AREA)

Abstract

本申请提供了一种最大似然估计MLSE均衡器的实现方法和芯片、电子设备、计算机可读介质,MLSE均衡器的实现方法包括:根据第k-1个符号的最优路径确定第k个符号对应的需要计算分支路径度量值的第一路径;其中,k为大于或等于2的整数;计算第k个符号在第一路径下的分支路径度量值;根据第k个符号在第一路径下的分支路径度量值和第k-1个符号在第k-1个符号的最优路径下的最优累加路径距离确定第k个符号的最优路径和第k个符号在第k个符号的最优路径下的最优累加路径距离;存储第k个符号的最优路径和第k个符号在第k个符号的最优路径下的最优累加路径距离。

Description

MLSE均衡器的实现方法和芯片、电子设备、计算机可读介质
相关申请的交叉引用
本申请要求2022年9月22日提交给中国专利局的第202211161260.8号专利申请的优先权,其全部内容通过引用合并于此。
技术领域
本申请涉及但不限于信息处理技术领域。
背景技术
随着人们对通信性能和通信速率要求的不断提高,通信设备及器件的带宽、通信信道对于数字通信的影响也越来越重要,这些影响因素中就包括越来越严重的码间串扰(ISI,Intersymbol Interference),噪声干扰以及信号功率的衰减等。所以,在这么多重失真源的存在下,信号补偿的均衡技术几乎已经成为所有通信***中不可缺少的组成部分。
最大似然估计(MLSE,Maximum Likelihood Sequence Estimation)均衡器就是其中性能优良的均衡器之一,能有效地补偿接收侧信号的非线性和信号功率。已有相关的仿真实验及论文论证MLSE均衡器性能对于信号的非线性以及中高频分量补偿效果要好于判决反馈(DFE,Decision Feedback Equalizer)均衡器,但由于算法原理及架构不同,MLSE算法对于芯片内存、计算等资源消耗要远远大于DFE均衡器。特别是针对多抽头系数的MLSE均衡器,抽头系数的个数、计算深度与算法实现的复杂度和对芯片内存、计算等资源的消耗成正比。例如,MLSE均衡器的抽头系数个数每增加1个,所需要的支路度量子模块、加比选子模块和路径管理子模块的数量就需要增加一倍。故当使用性能更优,抽头系数数量更多的MLSE算法时,MLSE均衡器对数据的存储、计算等资源的需求也成指数级增加,这些对于面积、功耗和计算等资源有限的芯片提出了严峻的考验。
发明内容
本申请提供一种MLSE均衡器的实现方法和芯片、电子设备、计算机可读介质。
第一方面,本申请提供一种MLSE均衡器的实现方法,包括:根据第k-1个符号的最优路径确定第k个符号对应的需要计算分支路径度量值的第一路径;其中,k为大于或等于2的整数;计算所述第k个符号在所述第一路径下的分支路径度量值;根据所述第k个符号在所述第一路径下的分支路径度量值和所述第k-1个符号在所述第k-1个符号的最优路径下的最优累加路径距离确定第k个符号的最优路径和所述第k个符号在所述第k个符号的最优路径下的最优累加路径距离;存储所述第k个符号的最优路径和所述第k个符号在所述第k个符号的最优路径下的最优累加路径距离。
第二方面,本申请提供一种电子设备,包括:至少一个处理器;存储器,存储器上存储有至少一个程序,当所述至少一个程序被所述至少一个处理器执行时,实现本文所述的任意一种MLSE均衡器的实现方法。
第三方面,本申请提供一种计算机可读介质,计算机可读介质上存储有计算机程序,所述计算机程序被处理器执行时实现本文所述的任意一种MLSE均衡器的实现方法。
第四方面,本申请提供一种最大似然估计均衡器的实现芯片,包括:数据接口和处理电路;所述处理电路,配置为通过所述数据接口读取***控制电路上存储的指令,执行本文所述的任意一种最大似然估计均衡器的实现方法。
附图说明
图1为相关技术中MLSE均衡器的实现示意图;
图2为本申请提供的MLSE均衡器的实现方法的流程图;
图3为本申请提供的MLSE均衡器的实现芯片的组成框图;
图4为本申请的处理电路的组成框图。
具体实施方式
为使本领域的技术人员更好地理解本申请的技术方案,下面结合附图对本申请提供的MLSE均衡器的实现方法和芯片、电子设备、计算机可读介质进行详细描述。
在下文中将参考附图更充分地描述示例实施方式,但是所述示例实施方式可以以不同形式来体现且不应当被解释为限于本文阐述的实施方式。反之,提供这些实施方式的目的在于使本申请透彻和完整,并将使本领域技术人员充分理解本申请的范围。
在不冲突的情况下,本申请各实施方式及实施方式中的各特征可相互组合。
如本文所使用的,术语“和/或”包括至少一个相关列举条目的任何和所有组合。
本文所使用的术语仅用于描述特定实施方式,且不意欲限制本申请。如本文所使用的,单数形式“一个”和“该”也意欲包括复数形式,除非上下文另外清楚指出。还将理解的是,当本说明书中使用术语“包括”和/或“由……制成”时,指定存在所述特征、整体、步骤、操作、元件和/或组件,但不排除存在或添加至少一个其它特征、整体、步骤、操作、元件、组件和/或其群组。
除非另外限定,否则本文所用的所有术语(包括技术和科学术语)的含义与本领域普通技术人员通常理解的含义相同。还将理解,诸如那些在常用字典中限定的那些术语应当被解释为具有与其在相关技术以及本申请的背景下的含义一致的含义,且将不解释为具有理想化或过度形式上的含义,除非本文明确如此限定。
图1为相关技术中MLSE均衡器的实现示意图。如图1所示,MLSE均衡器包括支路度量子模块101,加比选子模块102,路径管理子模块103和输出判决子模块104四个子模块。
图1中,X(k)为输入到MLSE均衡器中的待均衡信号,Y(k)为经过MLSE均衡器译码输出判决后的电平信号。
其中,支路度量子模块101配置为计算第k个符号在不同路径 下的分支路径度量值。通常情况下采用序列采样点间的距离来表示分支路径度量值。其中,距离可以是欧式距离、曼哈顿距离、切比雪夫距离等中的任意一个。例如,采用欧式距离时,分支路径度量值计算公式如公式(1)。
其中,xk-i为第k-i个符号对应的标准星座点的可能输出电平,以非归零码(NRZ,Non-Return-to-Zero)调制方式为例,每个符号均有两种可能输出电平,分别+1和-1,x(k)为当前输入的待均衡信号的第k个符号,n为MLSE的抽头系数的数量,αi为符号串扰系数,也即抽头系数。
在相关技术中,每个符号有m种可能输出电平,公式(1)中共有mn种分支度量值,如果每个支路度量子模块101计算一种分支度量值,则需要mn个支路度量子模块101。以NRZ调制方式为例,每个符号有2种可能输出电平,当n=3时,公式(1)中共有8种分支度量值,分别对应(xk-2,xk-1,xk)的可能输出电平为:(-1,-1,-1)、(+1,-1,-1)、(-1,+1,-1)、(+1,+1,-1)、(-1,-1,+1)、(+1,-1,+1)、(-1,+1,+1)、(+1,+1,+1),也就是需要8个支路度量子模块101。
其中,加比选子模块102配置为根据支路度量子模块101输出的第k个符号在不同路径下的分支路径度量值,累加前一符号在与分支路径度量值对应的路径下的最优累加路径距离,然后选择出当前符号的最优路径(也即幸存路径)。
以欧式距离为例,第k个符号在路径xk-(n-2)xk-(n-3)…xk下的最优累加路径距离的计算公式如公式(2)。
其中,D(xk-(n-2),xk-(n-3),…,xk)为第k个符号在路径 xk-(n-2)xk-(n-3)…xk下的最优累加路径距离,D(xk-(n-1),xk-(n-2),…,xk-1)为第k-1个符号在路径xk-(n-1)xk-(n-2)…xk-1下的最优累加路径距离,P(xk-(n-1),xk-(n-2),…,xk)为第k个符号在路径xk-(n-1)xk-(n-2)…xk下的分支路径度量值。
在相关技术中,如果每个加比选子模块102计算D(xk-(n-2),xk-(n-3),…,xk)的一种可能取值,则需要mn-1个加比选子模块102。以NRZ调整方式为例,需要加比选子模块102的数量为2n-1
其中,路径管理子模块103配置为存储加比选子模块102选择出来的幸存路径节点和不同路径下的最优路径值,并在后续进行回溯译码的相关操作,输出判决译码值。
在相关技术中,路径管理子模块103中存储的最优路径值的数量为mn-1
其中,判断输出子模块104配置为将幸存路径下的判决译码值判决输出得到判决后的电平信号。
图2为本申请提供的MLSE均衡器的实现方法的流程图。
第一方面,参照图2,在本申请一个实施方式中,提供一种MLSE均衡器的实现方法,包括步骤200至203。
在步骤200,根据第k-1个符号的最优路径确定第k个符号对应的需要计算分支路径度量值的第一路径;其中,k为大于或等于2的整数。
在一些示例性实施方式中,第一路径的数量小于第二路径的数量,第二路径中所有符号的取值均包括m种可能输出电平;m为大于或等于1的整数。该第一路径的数量小于第二路径的数量,第二路径即为相关技术中需要计算分支路径度量值的路径,减少了需要计算分支路径度量值的路径数,从而进一步减少了需要确定以及存储的最优路径数量,因此减少了所需要的计算资源和存储资源。
在一些示例性实施方式中,第一路径中的第一目标符号的取值与第k-1个符号的最优路径中对应符号的取值相同;其中,第一目标符号包括第k-(n-1)个符号到第k-t个符号;n为抽头系数的数量, n小于或等于k,t为大于或等于1,且小于或等于n-2的整数;第一路径中的第二目标符号的取值包括标准星座点的m种可能输出电平;其中,第二目标符号包括第k-t+1个符号到第k个符号。
也就是说,第一路径中的第一目标符号的取值只有一个,第二目标符号的取值有m个。
在一些示例性实施方式中,第一目标符号包括:第k-(n-1)个符号到第k-1个符号,第二目标符号包括:第k个符号。也就是t取值为1的情况。
在一些示例性实施方式中,m为待均衡信号的调制方式对应的每一个符号的可能输出电平的数量。例如针对NRZ调制方式,m的取值为2。
本申请实施方式对调制方式不作限定。例如,可以为NRZ、正交调幅(QAM,Quadrature Amplitude Modulation)、移项键控(PSK,Phase Shift Keying)、N电平脉冲幅度调制(PAM-N,N Level Pulse Amplitude Modulation)等中的任意一种。
以NRZ调制方式为例,每一个符号的可能输出电平有两种,分别为-1和+1。
假设n取3,那么第k-1个符号的最优路径xk-3xk-2xk-1有两条,假设这两条最优路径分别为(-1,-1,-1)和(-1,+1,+1),那么根据这两条最优路径可以确定第一路径包括(-1,-1,-1)、(-1,-1,+1)、(+1,+1,-1)和(+1,+1,+1)四条路径,也就是上面t取值为1的情况。
假设n取4,那么第k-1个符号的最优路径xk-4xk-3xk-2xk-1有四条,假设这四条最优路径分别为(-1,-1,-1,-1)、(+1,-1,-1,+1)、(-1,+1,+1,+1)和(+1,+1,+1,-1),那么根据这四条最优路径可以确定第一路径包括(-1,-1,-1,-1)、(-1,-1,-1,+1)、(-1,-1,+1,+1)、(-1,-1,+1,-1)、(+1,+1,+1,+1)、(+1,+1,+1,-1)、(+1,+1,-1,+1)和(+1,+1,-1,-1)八条路径,也就是上面t取值为2的情况。
在步骤201,计算第k个符号在第一路径下的分支路径度量值。
在一些示例性实施方式中,按照公式(1)计算第k个符号在第一路径下的分支路径度量值。
在步骤202,根据第k个符号在第一路径下的分支路径度量值和第k-1个符号在第k-1个符号的最优路径下的最优累加路径距离,确定第k个符号的最优路径和第k个符号在第k个符号的最优路径下的最优累加路径距离。
在一些示例性实施方式中,根据第k个符号在第一路径下的分支路径度量值和第k-1个符号在第k-1个符号的最优路径下的最优累加路径距离,确定第k个符号的最优路径和第k个符号在第k个符号的最优路径下的最优累加路径距离包括:对第一路径进行分组,同一个分组内的路径包括第一路径中第k-t+1个符号到第k个符号的取值相同的路径;对于同一分组内的路径,根据第k个符号在分组内的路径下的分支路径度量值和第k-1个符号在从第k-1个符号的最优路径中选择出的路径下的最优累加路径距离确定分组对应的最优路径和最优累加路径距离;其中,从第k-1个符号的最优路径中选择出的路径包括:第k-1个符号的最优路径中第k-t+1个符号到第k-1个符号的取值与分组内的路径中第k-t+1个符号到第k-1个符号的取值相同的路径;所有分组对应的最优路径组成第k个符号的最优路径;所有分组对应的最优累加路径距离组成第k个符号在第k个符号的最优路径下的最优累加路径距离。
在一些示例性实施方式中,假设第一路径为xk-(n-1)xk-(n-2)……xk-txk-t+1……xk-1xk,第k-1个符号的最优路径为xk-1-(n-1)xk-1- (n-2)……xk-1-txk-t……xk-1,那么第一路径中属于同一个分组的路径中xk-(n-1)xk-(n-2)……xk-t的部分是与第k-1个符号的最优路径中xk- (n-1)xk-(n-2)……xk-t的部分相同的,第一路径中属于同一个分组的路径中xk-t+1……xk-1xk的部分采用标准星座点的可能输出电平,具体的分组过程以及最优路径和最优累加路径距离的确定过程可以参考示例2中的描述。
在一些示例性实施方式中,根据第k个符号在分组内的路径下的分支路径度量值和第k-1个符号在从第k-1个符号的最优路径中选 择出的路径下的最优累加路径距离确定分组对应的最优路径和最优累加路径距离包括:针对分组内的每一条路径,计算第k个符号在路径下的分支路径度量值和第k-1个符号在从第k-1个符号的最优路径中选择出的路径下的最优累加路径距离之和,计算得到的和值作为所述第k个符号在路径下的累加路径距离;将分组内的路径中累加路径距离最小对应的路径作为对应分组的最优路径;将分组内的路径中所述累加路径距离的最小值作为对应分组的最优累加路径距离。
本申请实施方式中,不管n取多大值,针对每一个符号的计算,均只需要计算mt+1个分支路径度量值,即mt+1条第一路径,保存mt条最优路径和最优累加路径距离,减少了需要计算的分支路径度量值的数量,以及保存的最优路径和最优累加路径距离的数量,也就是减少了计算资源和存储资源。
在一些示例性实施方式中,根据第k个符号在第一路径下的分支路径度量值和第k-1个符号在第k-1个符号的最优路径下的最优累加路径距离,确定第k个符号的最优路径和第k个符号在第k个符号的最优路径下的最优累加路径距离包括:对第一路径进行分组,同一个分组内的路径包括第一路径中选择第k个符号的取值相同的路径;对于同一分组内的路径,根据第k个符号在分组内的路径下的分支路径度量值和第k-1个符号在从第k-1个符号的最优路径中选择出的路径下的最优累加路径距离确定分组对应的最优路径和最优累加路径距离;其中,从第k-1个符号的最优路径中选择出的路径包括:第k-1个符号的最优路径中第k-1个符号的取值与分组内的路径中第k-1个符号的取值相同的路径;所有分组对应的最优路径组成第k个符号的最优路径;所有分组对应的最优累加路径距离组成第k个符号在第k个符号的最优路径下的最优累加路径距离。
在一些示例性实施方式中,假设第一路径为xk-(n-1)xk-(n-2)……xk-1xk,第k-1个符号的最优路径为xk-1-(n-1)xk-1-(n-2)……xk-1,那么属于同一个分组的第五路径中xk-(n-1)xk-(n-2)……xk-1的部分是与第k-1个符号的最优路径中xk-(n-1)xk-(n-2)……xk-1的部分相同的,属于同一个分组的第五路径中xk的部分采用标准星座点的可能输出 电平,具体的分组过程以及最优路径和最优累加路径距离的确定过程可以参考示例1中的描述。
在一些示例性实施方式中,根据第k个符号在分组内的路径下的分支路径度量值和第k-1个符号在从第k-1个符号的最优路径中选择出的路径下的最优累加路径距离确定分组内对应的最优路径和最优累加路径距离包括:针对分组内的每一条路径,计算第k个符号在分组内的路径下的分支路径度量值和第k-1个符号在从第k-1个符号的最优路径中选择出的路径下的最优累加路径距离之和,计算得到的和值作为第k个符号在路径下的累加路径距离;将分组内的路径中累加路径距离最小对应的路径作为对应分组的最优路径;将分组内的路径中所述累加路径距离的最小值作为对应分组的最优累加路径距离。
本申请实施方式中,不管n取多大值,针对每一个符号的计算,均只需要计算m2个分支路径度量值,即m2条第一路径,保存m条最优路径和最优累加路径距离,减少了需要计算的分支路径度量值的数量,以及保存的最优路径和最优累加路径距离的数量。
在步骤203,存储第k个符号的最优路径和第k个符号在第k个符号的最优路径下的最优累加路径距离。
在一些示例性实施方式中,在缓存第k个符号的最优路径时,可以缓存最优路径中每一个符号的取值。
在一些示例性实施方式中,存储第k个符号的最优路径和第k个符号在第k个符号的最优路径下的最优累加路径距离后,该方法还包括:根据第k个符号在第k个符号的最优路径下的最优累加路径距离得到判决后的电平信号。
在一些示例性实施方式中,根据第k个符号在第k个符号的最优路径下的最优累加路径距离得到判决后的电平信号包括:对第k个符号在第k个符号的最优路径下的最优累加路径距离进行回溯译码的相关操作,得到判决译码值,对判决译码值判决输出得到判决后的电平信号。
本申请实施方式提供的MLSE均衡器的实现方法,基于第k-1个符号的最优路径确定第k个符号对应的需要计算分支路径度量值 的第一路径,减少了需要计算分支路径度量值的路径数,从而进一步减少了需要确定以及存储的最优路径数量,因此减少了所需要的计算资源和存储资源。
为了使本申请实施方式的MLSE均衡器的实现方法更加易于理解,下面列举两个示例进行说明,所列举的示例不用于限定本申请实施方式的MLSE均衡器的实现方法的保护范围。
示例1
本示例中,n=3,t=1,采用NRZ调制方式,每一个符号都有+1和-1两种取值。
本示例的MLSE均衡器的实现过程如下:
获取待均衡信号。
待均衡信号的第1个符号对应的标准星座点有两种取值,分别为+1和-1,第1个符号没有上一个符号的最优路径和最优累加路径距离,初始化最优累加路径距离为0。
第2个符号需要计算分支路径度量值的第一路径包括四条,分别为(x1=-1,x2=-1)、(x1=-1,x2=+1)、(x1=+1,x2=-1)、(x1=+1,x2=+1)。
根据待均衡信号的第2个符号的值计算每一条第一路径对应的分支路径度量值。
上述四条路径中,将路径(x1=-1,x2=-1)和路径(x1=+1,x2=-1)划分为分组1,将路径(x1=-1,x2=+1)和路径(x1=+1,x2=+1)划分为分组2。
针对分组1,计算第2个符号在路径(x1=-1,x2=-1)下的累加路径距离D(x1=-1,x2=-1)=D(x1=-1)+P(x1=-1,x2=-1);计算第2个符号在路径(x1=+1,x2=-1)下的累加路径距离D(x1=+1,x2=-1)=D(x1=+1)+P(x1=+1,x2=-1)=P(x1=+1,x2=-1)。假设D(x1=-1,x2=-1)小于D(x1=+1,x2=-1),则分组1对应的最优路径为路径(x1=-1,x2=-1),分组1的最优累加路径距离为D(x1=-1,x2=-1)。
针对分组2,计算第2个符号在路径(x1=-1,x2=+1)下的累加 路径距离D(x1=-1,x2=+1)=D(x1=-1)+P(x1=-1,x2=+1);计算第2个符号在路径(x1=+1,x2=+1)下的累加路径距离D(x1=+1,x2=+1)=D(x1=+1)+P(x1=+1,x2=+1)。假设D(x1=-1,x2=+1)小于D(x1=+1,x2=+1),则分组2对应的最优路径为路径(x1=-1,x2=+1),分组2的最优累加路径距离为D(x1=-1,x2=+1)。
分组1的最优路径和分组2的最优路径组成第2个符号的最优路径,分组1的最优累加路径距离和分组2的最优累加路径距离组成第2个符号的最优累加路径距离。
那么第2个符号的最优路径包括路径(x1=-1,x2=-1)和路径(x1=-1,x2=+1),第2个符号的最优累加路径距离包括D(x1=-1,x2=-1)和D(x1=-1,x2=+1)。
根据第2个符号的最优路径确定第3个符号需要计算分支路径度量值的第一路径包括四条,分别为(x1=-1,x2=-1,x3=-1)、(x1=-1,x2=-1,x3=+1)、(x1=-1,x2=+1,x3=-1)和(x1=-1,x2=+1,x3=+1)。也就是说,第3个符号需要计算分支路径度量值的第一路径中第1个符号到第2个符号的路径采用第2个符号的最优路径。
根据待均衡信号的第3个符号的值计算每一条第一路径对应的分支路径度量值。
上述四条路径中,将路径(x1=-1,x2=-1,x3=-1)和路径(x1=-1,x2=+1,x3=-1)划分为分组1,将路径(x1=-1,x2=-1,x3=+1)和路径(x1=-1,x2=+1,x3=+1)划分为分组2。
针对分组1,计算第3个符号在路径(x1=-1,x2=-1,x3=-1)下的累加路径距离D(x1=-1,x2=-1,x3=-1)=D(x1=-1,x2=-1)+P(x1=-1,x2=-1,x3=-1);计算第3个符号在路径(x1=-1,x2=+1,x3=-1)下的累加路径距离D(x1=-1,x2=+1,x3=-1)=D(x1=-1,x2=+1)+P(x1=-1,x2=+1,x3=-1)。假设D(x1=-1,x2=-1,x3=-1)大于D(x1=-1,x2=+1,x3=-1),则分组1对应的最优路径为路径D(x1=-1,x2=+1,x3=-1),分组1的最优累加路径距离为D(x1=-1,x2=+1,x3=-1)。
针对分组2,计算第3个符号在路径(x1=-1,x2=-1,x3=+1)下的累加路径距离D(x1=-1,x2=-1,x3=+1)=D(x1=-1,x2=-1)+P(x1=-1, x2=-1,x3=+1);计算第3个符号在路径(x1=-1,x2=+1,x3=+1)下的累加路径距离D(x1=-1,x2=+1,x3=+1)=D(x1=-1,x2=+1)+P(x1=-1,x2=+1,x3=+1)。假设D(x1=-1,x2=-1,x3=+1)小于D(x1=-1,x2=+1,x3=+1),则分组2对应的最优路径为路径(x1=-1,x2=-1,x3=+1),分组2的最优累加路径距离为D(x1=-1,x2=-1,x3=+1)。
分组1的最优路径和分组2的最优路径组成第3个符号的最优路径,分组1的最优累加路径距离和分组2的最优累加路径距离组成第3个符号的最优累加路径距离。
那么第3个符号的最优路径包括路径(x1=-1,x2=+1,x3=-1)和路径(x1=-1,x2=-1,x3=+1),第3个符号的最优累加路径距离包括D(x1=-1,x2=+1,x3=-1)和D(x1=-1,x2=-1,x3=+1)。
根据第3个符号的最优路径确定第4个符号需要计算分支路径度量值的第一路径包括四条,分别为(x1=-1,x2=+1,x3=-1,x4=-1)、(x1=-1,x2=+1,x3=-1,x4=+1)、(x1=-1,x2=-1,x3=+1,x4=-1)和(x1=-1,x2=-1,x3=+1,x4=+1)。也就是说,第4个符号需要计算分支路径度量值的第一路径中第1个符号到第3个符号的路径采用第3个符号的最优路径。
根据待均衡信号的第4个符号的值计算每一条第一路径对应的分支路径度量值。
上述四条路径中,将路径(x1=-1,x2=+1,x3=-1,x4=-1)和路径(x1=-1,x2=-1,x3=+1,x4=-1)划分为分组1,将路径(x1=-1,x2=+1,x3=-1,x4=+1)和路径(x1=-1,x2=-1,x3=+1,x4=+1)划分为分组2。
针对分组1,计算第4个符号在路径(x1=-1,x2=+1,x3=-1,x4=-1)下的累加路径距离D(x1=-1,x2=+1,x3=-1,x4=-1)=D(x1=-1,x2=+1,x3=-1)+P(x1=-1,x2=+1,x3=-1,x4=-1);计算第4个符号在路径(x1=-1,x2=-1,x3=+1,x4=-1)下的累加路径距离D(x1=-1,x2=-1,x3=+1,x4=-1)=D(x1=-1,x2=-1,x3=+1)+P(x1=-1,x2=-1,x3=+1,x4=-1)。假设D(x1=-1,x2=+1,x3=-1,x4=-1)小于D(x1=-1,x2=-1, x3=+1,x4=-1),则分组1对应的最优路径为路径(x1=-1,x2=+1,x3=-1,x4=-1),分组1的最优累加路径距离为D(x1=-1,x2=+-1,x3=-1,x4=-1)。
针对分组2,计算第4个符号在路径(x1=-1,x2=+1,x3=-1,x4=+1)下的累加路径距离D(x1=-1,x2=+1,x3=-1,x4=+1)=D(x1=-1,x2=+1,x3=-1)+P(x1=-1,x2=+1,x3=-1,x4=+1);计算第4个符号在路径(x1=-1,x2=-1,x3=+1,x4=+1)下的累加路径距离D(x1=-1,x2=-1,x3=+1,x4=+1)=D(x1=-1,x2=-1,x3=+1)+P(x1=-1,x2=-1,x3=+1,x4=+1)。假设D(x1=-1,x2=+1,x3=-1,x4=+1)大于D(x1=-1,x2=-1,x3=+1,x4=+1),则分组2对应的最优路径为路径(x1=-1,x2=-1,x3=+1,x4=+1),分组2的最优累加路径距离为D(x1=-1,x2=-1,x3=+1,x4=+1)。
分组1的最优路径和分组2的最优路径组成第4个符号的最优路径,分组1的最优累加路径距离和分组2的最优累加路径距离组成第4个符号的最优累加路径距离。
那么第4个符号的最优路径包括路径(x1=-1,x2=+1,x3=-1,x4=-1)和路径(x1=-1,x2=-1,x3=+1,x4=+1),第4个符号的最优累加路径距离包括D(x1=-1,x2=+1,x3=-1,x4=-1)和D(x1=-1,x2=-1,x3=+1,x4=+1)。
示例2
本示例中,n=4,t=2,采用NRZ调制方式,每一个符号都有+1和-1两种取值。
示例的MLSE均衡器的实现过程如下:
获取待均衡信号。
待均衡信号的第1个符号对应的标准星座点有两种取值,分别为+1和-1,第1个符号没有上一个符号的最优路径和最优累加路径距离,初始化最优累加路径距离为0。
第2个符号需要计算分支路径度量值的第一路径包括四条,分别为(x1=-1,x2=-1)、(x1=-1,x2=+1)、(x1=+1,x2=-1)、(x1=+1,x2=+1)。
根据待均衡信号的第2个符号的值计算每一条第一路径对应的分支路径度量值。
上述四条路径中,将路径(x1=-1,x2=-1)和路径(x1=+1,x2=-1)划分为分组1,将路径(x1=-1,x2=+1)和路径(x1=+1,x2=+1)划分为分组2。
针对分组1,计算第2个符号在路径(x1=-1,x2=-1)下的累加路径距离D(x1=-1,x2=-1)=D(x1=-1)+P(x1=-1,x2=-1);计算第2个符号在路径(x1=+1,x2=-1)下的累加路径距离D(x1=+1,x2=-1)=D(x1=+1)+P(x1=+1,x2=-1)。假设D(x1=-1,x2=-1)小于D(x1=+1,x2=-1),则分组1对应的最优路径为路径(x1=-1,x2=-1),分组1的最优累加路径距离为D(x1=-1,x2=-1)。
针对分组2,计算第2个符号在路径(x1=-1,x2=+1)下的累加路径距离D(x1=-1,x2=+1)=D(x1=-1)+P(x1=-1,x2=+1);计算第2个符号在路径(x1=+1,x2=+1)下的累加路径距离D(x1=+1,x2=+1)=D(x1=+1)+P(x1=+1,x2=+1)。假设D(x1=-1,x2=+1)小于D(x1=+1,x2=+1),则分组2对应的最优路径为路径(x1=-1,x2=+1),分组2的最优累加路径距离为D(x1=-1,x2=+1)。
分组1的最优路径和分组2的最优路径组成第2个符号的最优路径,分组1的最优累加路径距离和分组2的最优累加路径距离组成第2个符号的最优累加路径距离。
那么第2个符号的最优路径包括路径(x1=-1,x2=-1)和路径(x1=-1,x2=+1),第2个符号的最优累加路径距离包括D(x1=-1,x2=-1)和D(x1=-1,x2=+1)。
根据第2个符号的最优路径确定第3个符号需要计算分支路径度量值的第一路径包括四条,分别为(x1=-1,x2=-1,x3=-1)、(x1=-1,x2=-1,x3=+1)、(x1=-1,x2=+1,x3=-1)和(x1=-1,x2=+1,x3=+1)。也就是说,第3个符号需要计算分支路径度量值的第一路径中第1个符号到第2个符号的路径采用第2个符号的最优路径。
根据待均衡信号的第3个符号的值计算每一条第一路径对应的分支路径度量值。
上述四条路径中,将路径(x1=-1,x2=-1,x3=-1)和路径(x1=-1,x2=+1,x3=-1)划分为分组1,将路径(x1=-1,x2=-1,x3=+1)和路径(x1=-1,x2=+1,x3=+1)划分为分组2。
针对分组1,计算第3个符号在路径(x1=-1,x2=-1,x3=-1)下的累加路径距离D(x1=-1,x2=-1,x3=-1)=D(x1=-1,x2=-1)+P(x1=-1,x2=-1,x3=-1);计算第3个符号在路径(x1=-1,x2=+1,x3=-1)下的累加路径距离D(x1=-1,x2=+1,x3=-1)=D(x1=-1,x2=+1)+P(x1=-1,x2=+1,x3=-1)。假设D(x1=-1,x2=-1,x3=-1)大于D(x1=-1,x2=+1,x3=-1),则分组1对应的最优路径为路径D(x1=-1,x2=+1,x3=-1),分组1的最优累加路径距离为D(x1=-1,x2=+1,x3=-1)。
针对分组2,计算第3个符号在路径(x1=-1,x2=-1,x3=+1)下的累加路径距离D(x1=-1,x2=-1,x3=+1)=D(x1=-1,x2=-1)+P(x1=-1,x2=-1,x3=+1);计算第3个符号在路径(x1=-1,x2=+1,x3=+1)下的累加路径距离D(x1=-1,x2=+1,x3=+1)=D(x1=-1,x2=+1)+P(x1=-1,x2=+1,x3=+1)。假设D(x1=-1,x2=-1,x3=+1)小于D(x1=-1,x2=+1,x3=+1),则分组2对应的最优路径为路径(x1=-1,x2=-1,x3=+1),分组2的最优累加路径距离为D(x1=-1,x2=-1,x3=+1)。
分组1的最优路径和分组2的最优路径组成第3个符号的最优路径,分组1的最优累加路径距离和分组2的最优累加路径距离组成第3个符号的最优累加路径距离。
那么第3个符号的最优路径包括路径(x1=-1,x2=+1,x3=-1)和路径(x1=-1,x2=-1,x3=+1),第3个符号的最优累加路径距离包括D(x1=-1,x2=+1,x3=-1)和D(x1=-1,x2=-1,x3=+1)。
根据第3个符号的最优路径确定第4个符号需要计算分支路径度量值的第一路径包括八条,分别为(x1=-1,x2=+1,x3=-1,x4=-1)、(x1=-1,x2=+1,x3=-1,x4=+1)、(x1=-1,x2=+1,x3=+1,x4=-1)、(x1=-1,x2=+1,x3=+1,x4=+1)、(x1=-1,x2=-1,x3=+1,x4=-1)、(x1=-1,x2=-1,x3=+1,x4=+1)、(x1=-1,x2=-1,x3=-1,x4=-1)和(x1=-1,x2=-1,x3=-1,x4=+1)。也就是说,第4个符号需要计 算分支路径度量值的第一路径中第1个符号到第2个符号的路径采用第3个符号的最优路径。
根据待均衡信号的第4个符号的值计算每一条第一路径对应的分支路径度量值。
上述八条路径中,将路径(x1=-1,x2=+1,x3=-1,x4=-1)和路径(x1=-1,x2=-1,x3=-1,x4=-1)划分为分组1,将路径(x1=-1,x2=+1,x3=-1,x4=+1)和路径(x1=-1,x2=-1,x3=-1,x4=+1)划分为分组2,将路径(x1=-1,x2=+1,x3=+1,x4=-1)和路径(x1=-1,x2=-1,x3=+1,x4=-1)划分为分组3,将路径(x1=-1,x2=+1,x3=+1,x4=+1)和路径(x1=-1,x2=-1,x3=+1,x4=+1)划分为分组4。
针对分组1,计算第4个符号在路径(x1=-1,x2=+1,x3=-1,x4=-1)下的累加路径距离D(x1=-1,x2=+1,x3=-1,x4=-1)=D(x1=-1,x2=+1,x3=-1)+P(x1=-1,x2=+1,x3=-1,x4=-1);计算第4个符号在路径(x1=-1,x2=-1,x3=-1,x4=-1)下的累加路径距离D(x1=-1,x2=-1,x3=-1,x4=-1)=D(x1=-1,x2=-1,x3=-1)+P(x1=-1,x2=-1,x3=-1,x4=-1)。假设D(x1=-1,x2=+1,x3=-1,x4=-1)小于D(x1=-1,x2=-1,x3=-1,x4=-1),则分组1对应的最优路径为路径(x1=-1,x2=+1,x3=-1,x4=-1),分组1的最优累加路径距离为D(x1=-1,x2=+-1,x3=-1,x4=-1)。
针对分组2,计算第4个符号在路径(x1=-1,x2=+1,x3=-1,x4=+1)下的累加路径距离D(x1=-1,x2=+1,x3=-1,x4=+1)=D(x1=-1,x2=+1,x3=-1)+P(x1=-1,x2=+1,x3=-1,x4=+1);计算第4个符号在路径(x1=-1,x2=-1,x3=-1,x4=+1)下的累加路径距离D(x1=-1,x2=-1,x3=-1,x4=+1)=D(x1=-1,x2=-1,x3=-1)+P(x1=-1,x2=-1,x3=-1,x4=+1)。假设D(x1=-1,x2=+1,x3=-1,x4=+1)大于D(x1=-1,x2=-1,x3=-1,x4=+1),则分组2对应的最优路径为路径(x1=-1,x2=-1,x3=-1,x4=+1),分组2的最优累加路径距离为D(x1=-1,x2=-1,x3=-1,x4=+1)。
针对分组3,计算第4个符号在路径(x1=-1,x2=+1,x3=+1,x4=-1)下的累加路径距离D(x1=-1,x2=+1,x3=+1,x4=-1)=D(x1=-1, x2=+1,x3=+1)+P(x1=-1,x2=+1,x3=+1,x4=-1);计算第4个符号在路径(x1=-1,x2=-1,x3=+1,x4=-1)下的累加路径距离D(x1=-1,x2=-1,x3=+1,x4=-1)=D(x1=-1,x2=-1,x3=+1)+P(x1=-1,x2=-1,x3=+1,x4=-1)。假设D(x1=-1,x2=+1,x3=+1,x4=-1)小于D(x1=-1,x2=-1,x3=+1,x4=-1),则分组3对应的最优路径为路径(x1=-1,x2=+1,x3=+1,x4=-1),分组3的最优累加路径距离为D(x1=-1,x2=+1,x3=+1,x4=-1)。
针对分组4,计算第4个符号在路径(x1=-1,x2=+1,x3=+1,x4=+1)下的累加路径距离D(x1=-1,x2=+1,x3=+1,x4=+1)=D(x1=-1,x2=+1,x3=+1)+P(x1=-1,x2=+1,x3=+1,x4=+1);计算第4个符号在路径(x1=-1,x2=-1,x3=+1,x4=+1)下的累加路径距离D(x1=-1,x2=-1,x3=+1,x4=+1)=D(x1=-1,x2=-1,x3=+1)+P(x1=-1,x2=-1,x3=+1,x4=+1)。假设D(x1=-1,x2=+1,x3=+1,x4=+1)大于D(x1=-1,x2=-1,x3=+1,x4=+1),则分组4对应的最优路径为路径(x1=-1,x2=-1,x3=+1,x4=+1),分组4的最优累加路径距离为D(x1=-1,x2=-1,x3=+1,x4=+1)。
分组1的最优路径、分组2的最优路径、分组3的最优路径和分组4的最优路径组成第4个符号的最优路径,分组1的最优累加路径距离、分组2的最优累加路径距离、分组3的最优累加路径距离和分组4的最优累加路径距离组成第4个符号的最优累加路径距离。
那么第4个符号的最优路径包括路径(x1=-1,x2=+1,x3=-1,x4=-1)、路径(x1=-1,x2=-1,x3=-1,x4=+1)、路径(x1=-1,x2=+1,x3=+1,x4=-1)和路径(x1=-1,x2=-1,x3=+1,x4=+1),第4个符号的最优累加路径距离包括D(x1=-1,x2=+1,x3=-1,x4=-1)、D(x1=-1,x2=-1,x3=-1,x4=+1)、D(x1=-1,x2=+1,x3=+1,x4=-1)和D(x1=-1,x2=-1,x3=+1,x4=+1)。
第二方面,本申请另一个实施方式提供一种电子设备,包括:至少一个处理器;存储器,存储器上存储有至少一个程序,当至少一个程序被至少一个处理器执行时,实现上述任意一种MLSE均衡器的实现方法。
其中,处理器为具有数据处理能力的器件,其包括但不限于中央处理器(CPU)等;存储器为具有数据存储能力的器件,其包括但不限于随机存取存储器(RAM,更具体如SDRAM、DDR等)、只读存储器(ROM)、带电可擦可编程只读存储器(EEPROM)、闪存(FLASH)。
在一些实施方式中,处理器、存储器通过总线相互连接,进而与计算设备的其它组件连接。
第三方面,本申请另一个实施方式提供一种计算机可读介质,计算机可读介质上存储有计算机程序,所述计算机程序被处理器执行时实现上述任意一种MLSE均衡器的实现方法。
图3为本申请提供的MLSE均衡器的实现芯片的组成框图。
第四方面,本申请另一个实施方式提供一种MLSE均衡器的实现芯片,包括:数据接口301和处理电路302;处理电路,配置为通过数据接口读取***控制电路上存储的指令,执行上述任意一种MLSE均衡器的实现方法。
在一些示例性实施方式中,如图4所示,处理电路302包括:前向路径指示子电路401,配置为根据第k-1个符号的最优路径确定第k个符号对应的需要计算分支路径度量值的第一路径;其中,k为大于或等于2的整数;支路度量子电路402,配置为计算所述第k个符号在所述第一路径下的分支路径度量值;加比选子电路403,配置为根据所述第k个符号在所述第一路径下的分支路径度量值和所述第k-1个符号在所述第k-1个符号的最优路径下的最优累加路径距离,确定第k个符号的最优路径和所述第k个符号在所述第k个符号的最优路径下的最优累加路径距离;路径管理子电路404,配置为存储所述第k个符号的最优路径和所述第k个符号在所述第k个符号的最优路径下的最优累加路径距离。
在一些示例性实施方式中,MLSE均衡器的实现芯片,如专用集成电路(ASIC,Application Specific Integrated Circuit)芯片或现场可编辑门阵列(FPGA,Field-Programmable Gate Array)芯片。
在一些示例性实施方式中,还包括:输出判决子电路405,配置 为根据所述第k个符号在所述第k个符号的最优路径下的最优累加路径距离得到判决后的电平信号。
在一些示例性实施方式中,第一路径的数量小于第二路径的数量,第二路径中所有符号的取值均包括m种可能输出电平;m为大于或等于1的整数。
在一些示例性实施方式中,第一路径中的第一目标符号的取值与所述第k-1个符号的最优路径中对应符号的取值相同;其中,所述第一目标符号包括第k-(n-1)个符号到第k-t个符号;n为抽头系数的数量,n小于或等于k,t为大于或等于1,且小于或等于n-2的整数;所述第一路径中的第二目标符号的取值包括标准星座点的m种可能输出电平;其中,所述第二目标符号包括第k-t+1个符号到第k个符号。
在一些示例性实施方式中,加比选子电路403配置为采用以下方式实现根据所述第k个符号在所述第一路径下的分支路径度量值和所述第k-1个符号在所述第k-1个符号的最优路径下的最优累加路径距离,确定第k个符号的最优路径和所述第k个符号在所述第k个符号的最优路径下的最优累加路径距离:对所述第一路径进行分组,同一个分组内的路径包括第一路径中选择第k-t+1个符号到第k个符号的取值相同的路径;对于同一分组内的路径,根据第k个符号在分组内的路径下的分支路径度量值和所述第k-1个符号在从第k-1个符号的最优路径中选择出的路径下的最优累加路径距离确定分组对应的最优路径和最优累加路径距离;其中,从所述第k-1个符号的最优路径中选择出的路径包括:所述第k-1个符号的最优路径中第k-t+1个符号到第k-1个符号的取值与所述分组内的路径中第k-t+1个符号到第k-1个符号的取值相同的路径;所有分组对应的最优路径组成所述第k个符号的最优路径;所有分组对应的最优累加路径距离组成所述第k个符号在所述第k个符号的最优路径下的最优累加路径距离。
在一些示例性实施方式中,加比选子电路403配置为采用以下方式实现根据所述第k个符号在分组内的路径下的分支路径度量值和所述第k-1个符号在从所述第k-1个符号的最优路径中选择出的路 径下的最优累加路径距离确定分组对应的最优路径和最优累加路径距离:针对分组内的每一条路径,计算所述第k个符号在所述路径下的分支路径度量值和所述第k-1个符号在从所述第k-1个符号的最优路径中选择出的路径下的最优累加路径距离之和,计算得到的和值作为所述第k个符号在所述路径下的累加路径距离;将分组内的路径中所述累加路径距离最小对应的路径作为对应分组的最优路径;将分组内的路径中所述累加路径距离的最小值作为对应分组的最优累加路径距离。
在一些示例性实施方式中,在t等于1的情况下,所述第一目标符号包括:所述第k-(n-1)个符号到所述第k-1个符号,所述第二目标符号包括:所述第k个符号。
在一些示例性实施方式中,加比选子电路403配置为采用以下方式实现根据所述第k个符号在所述第一路径下的分支路径度量值和所述第k-1个符号在所述第k-1个符号的最优路径下的最优累加路径距离,确定第k个符号的最优路径和所述第k个符号在所述第k个符号的最优路径下的最优累加路径距离:对第一路径进行分组,同一个分组内的路径包括第一路径中第k个符号的取值相同的路径;对于同一分组内的路径,根据所述第k个符号在分组内的路径下的分支路径度量值和所述第k-1个符号在从所述第k-1个符号的最优路径中选择出的路径下的最优累加路径距离确定分组对应的最优路径和最优累加路径距离;其中,从所述第k-1个符号的最优路径中选择出的路径包括:所述第k-1个符号的最优路径中第k-1个符号的取值与所述分组内的路径中第k-1个符号的取值相同的路径;所有分组对应的最优路径组成所述第k个符号的最优路径;所有分组对应的最优累加路径距离组成所述第k个符号在所述第k个符号的最优路径下的最优累加路径距离。
上述MLSE均衡器的实现芯片的具体实现过程与前述实施方式的MLSE均衡器的实现方法的具体实现过程相同,这里不再赘述。
本领域普通技术人员可以理解,上文中所公开方法中的全部或某些步骤、***、装置中的功能模块/单元可以被实施为软件、固件、 硬件及其适当的组合。在硬件实施方式中,在以上描述中提及的功能模块/单元之间的划分不一定对应于物理组件的划分;例如,一个物理组件可以具有多个功能,或者一个功能或步骤可以由若干物理组件合作执行。某些物理组件或所有物理组件可以被实施为由处理器,如中央处理器、数字信号处理器或微处理器执行的软件,或者被实施为硬件,或者被实施为集成电路,如专用集成电路。这样的软件可以分布在计算机可读介质上,计算机可读介质可以包括计算机存储介质(或非暂时性介质)和通信介质(或暂时性介质)。如本领域普通技术人员公知的,术语计算机存储介质包括在用于存储信息(诸如计算机可读指令、数据结构、程序模块或其它数据)的任何方法或技术中实施的易失性和非易失性、可移除和不可移除介质。计算机存储介质包括但不限于RAM、ROM、EEPROM、闪存或其它存储器技术、CD-ROM、数字多功能盘(DVD)或其它光盘存储、磁盒、磁带、磁盘存储或其它磁存储器、或者可以用于存储期望的信息并且可以被计算机访问的任何其它的介质。此外,本领域普通技术人员公知的是,通信介质通常包含计算机可读指令、数据结构、程序模块或者诸如载波或其它传输机制之类的调制数据信号中的其它数据,并且可包括任何信息递送介质。
本文已经公开了示例实施方式,并且虽然采用了具体术语,但它们仅用于并仅应当被解释为一般说明性含义,并且不用于限制的目的。在一些实例中,对本领域技术人员显而易见的是,除非另外明确指出,否则可单独使用与特定实施方式相结合描述的特征、特性和/或元素,或可与其它实施方式相结合描述的特征、特性和/或元件组合使用。因此,本领域技术人员将理解,在不脱离由所附的权利要求阐明的本申请的范围的情况下,可进行各种形式和细节上的改变。

Claims (10)

  1. 一种最大似然估计均衡器的实现方法,包括:
    根据第k-1个符号的最优路径确定第k个符号对应的需要计算分支路径度量值的第一路径;其中,k为大于或等于2的整数;
    计算所述第k个符号在所述第一路径下的分支路径度量值;
    根据所述第k个符号在所述第一路径下的分支路径度量值和所述第k-1个符号在所述第k-1个符号的最优路径下的最优累加路径距离,确定第k个符号的最优路径和所述第k个符号在所述第k个符号的最优路径下的最优累加路径距离;
    存储所述第k个符号的最优路径和所述第k个符号在所述第k个符号的最优路径下的最优累加路径距离。
  2. 根据权利要求1所述的最大似然估计均衡器的实现方法,其中,所述第一路径的数量小于第二路径的数量,所述第二路径中所有符号的取值均包括m种可能输出电平;m为大于或等于1的整数。
  3. 根据权利要求1所述的最大似然估计均衡器的实现方法,其中,所述第一路径中的第一目标符号的取值与所述第k-1个符号的最优路径中对应符号的取值相同;其中,所述第一目标符号包括第k-(n-1)个符号到第k-t个符号;n为抽头系数的数量,n小于或等于k,t为大于或等于1,且小于或等于n-2的整数;
    所述第一路径中的第二目标符号的取值包括标准星座点的m种可能输出电平;其中,所述第二目标符号包括第k-t+1个符号到第k个符号。
  4. 根据权利要求3所述的最大似然估计均衡器的实现方法,其中,根据所述第k个符号在所述第一路径下的分支路径度量值和所述第k-1个符号在所述第k-1个符号的最优路径下的最优累加路径距离,确定第k个符号的最优路径和所述第k个符号在所述第k个符号的最优路径下的最优累加路径距离包括:
    对所述第一路径进行分组,同一个分组内的路径包括所述第一路径中第k-t+1个符号到所述第k个符号的取值相同的路径;
    对于同一分组内的路径,根据所述第k个符号在所述分组内的路径下的分支路径度量值和所述第k-1个符号在从所述第k-1个符号的最优路径中选择出的路径下的最优累加路径距离确定所述分组对应的最优路径和最优累加路径距离;其中,从所述第k-1个符号的最优路径中选择出的路径包括:所述第k-1个符号的最优路径中第k-t+1个符号到第k-1个符号的取值与所述分组内的路径中第k-t+1个符号到第k-1个符号的取值相同的路径;
    所有分组对应的最优路径组成所述第k个符号的最优路径;
    所有分组对应的最优累加路径距离组成所述第k个符号在所述第k个符号的最优路径下的最优累加路径距离。
  5. 根据权利要求4所述的最大似然估计均衡器的实现方法,其中,所述根据所述第k个符号在所述分组内的路径下的分支路径度量值和所述第k-1个符号在从所述第k-1个符号的最优路径中选择出的路径下的最优累加路径距离确定所述分组对应的最优路径和最优累加路径距离包括:
    针对所述分组内的每一条路径,计算所述第k个符号在所述路径下的分支路径度量值和所述第k-1个符号在从所述第k-1个符号的最优路径中选择出的路径下的最优累加路径距离之和,计算得到的和值作为所述第k个符号在所述路径下的累加路径距离;
    将所述分组内的路径中所述累加路径距离最小对应的路径作为对应分组的最优路径;
    将所述分组内的路径中所述累加路径距离的最小值作为对应分组的最优累加路径距离。
  6. 根据权利要求3所述的最大似然估计均衡器的实现方法,其中,在t等于1的情况下,所述第一目标符号包括:所述第k-(n-1)个符号到所述第k-1个符号,所述第二目标符号包括:所述第k个符号。
  7. 根据权利要求6所述的最大似然估计均衡器的实现方法,其中,根据所述第k个符号在所述第一路径下的分支路径度量值和所述第k-1个符号在所述第k-1个符号的最优路径下的最优累加路径距离,确定第k个符号的最优路径和所述第k个符号在所述第k个符号的最 优路径下的最优累加路径距离包括:
    对所述第一路径进行分组,同一个分组内的路径包括所述第一路径中所述第k个符号的取值相同的路径;
    对于同一分组内的路径,根据所述第k个符号在所述分组内的路径下的分支路径度量值和所述第k-1个符号在从所述第k-1个符号的最优路径中选择出的路径下的最优累加路径距离确定所述分组对应的最优路径和最优累加路径距离;其中,从所述第k-1个符号的最优路径中选择出的路径包括:所述第k-1个符号的最优路径中第k-1个符号的取值与所述分组内的路径中第k-1个符号的取值相同的路径;
    所有分组对应的最优路径组成所述第k个符号的最优路径;
    所有分组对应的最优累加路径距离组成所述第k个符号在所述第k个符号的最优路径下的最优累加路径距离。
  8. 一种电子设备,包括:
    至少一个处理器;
    存储器,所述存储器上存储有至少一个程序,当所述至少一个程序被所述至少一个处理器执行时,实现权利要求1-7任意一项所述的最大似然估计均衡器的实现方法。
  9. 一种计算机可读介质,所述计算机可读介质上存储有计算机程序,所述计算机程序被处理器执行时实现权利要求1-7任意一项所述的最大似然估计均衡器的实现方法。
  10. 一种最大似然估计均衡器的实现芯片,包括:数据接口和处理电路;
    所述处理电路,配置为通过所述数据接口读取***控制电路上存储的指令,执行如权利要求1-7任意一项所述的最大似然估计均衡器的实现方法。
PCT/CN2023/120012 2022-09-22 2023-09-20 Mlse均衡器的实现方法和芯片、电子设备、计算机可读介质 WO2024061266A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202211161260.8A CN117792837A (zh) 2022-09-22 2022-09-22 Mlse均衡器的实现方法和芯片、电子设备、计算机可读介质
CN202211161260.8 2022-09-22

Publications (1)

Publication Number Publication Date
WO2024061266A1 true WO2024061266A1 (zh) 2024-03-28

Family

ID=90389722

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/120012 WO2024061266A1 (zh) 2022-09-22 2023-09-20 Mlse均衡器的实现方法和芯片、电子设备、计算机可读介质

Country Status (2)

Country Link
CN (1) CN117792837A (zh)
WO (1) WO2024061266A1 (zh)

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1787386A (zh) * 2004-12-08 2006-06-14 中兴通讯股份有限公司 一种维特比译码器路径度量存储的方法
CN1841546A (zh) * 2005-03-31 2006-10-04 索尼株式会社 最大似然解码装置、信号估计方法和再现装置
CN1994003A (zh) * 2004-10-29 2007-07-04 中兴通讯股份有限公司 一种适用于edge***的8psk均衡解调的方法及装置
CN102948097A (zh) * 2010-06-18 2013-02-27 思科技术公司 用于ber估计的方法和电路
CN103685105A (zh) * 2013-12-25 2014-03-26 北京华力创通科技股份有限公司 一种最大似然均衡中输出软信息的方法及装置
CN110061761A (zh) * 2018-01-19 2019-07-26 华为技术有限公司 信号均衡方法及装置、光接收机
US20220141057A1 (en) * 2020-10-29 2022-05-05 Samsung Electronics Co., Ltd. Data receiving device and method

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1994003A (zh) * 2004-10-29 2007-07-04 中兴通讯股份有限公司 一种适用于edge***的8psk均衡解调的方法及装置
CN1787386A (zh) * 2004-12-08 2006-06-14 中兴通讯股份有限公司 一种维特比译码器路径度量存储的方法
CN1841546A (zh) * 2005-03-31 2006-10-04 索尼株式会社 最大似然解码装置、信号估计方法和再现装置
CN102948097A (zh) * 2010-06-18 2013-02-27 思科技术公司 用于ber估计的方法和电路
CN103685105A (zh) * 2013-12-25 2014-03-26 北京华力创通科技股份有限公司 一种最大似然均衡中输出软信息的方法及装置
CN110061761A (zh) * 2018-01-19 2019-07-26 华为技术有限公司 信号均衡方法及装置、光接收机
US20220141057A1 (en) * 2020-10-29 2022-05-05 Samsung Electronics Co., Ltd. Data receiving device and method

Also Published As

Publication number Publication date
CN117792837A (zh) 2024-03-29

Similar Documents

Publication Publication Date Title
JP3459879B2 (ja) デジタル通信レシーバー用改良型検知
CN109873777B (zh) 一种纠错方法和纠错装置
US11038538B2 (en) Maximum likelihood error detection for decision feedback equalizers with PAM modulation
EP0050930B1 (en) Improvements in or relating to data transmission systems
JP7200363B2 (ja) 狭帯域フィルタ済み信号のためのノイズホワイトニング後補償の効率的実施
US20230370192A1 (en) Receiver filtering
JP2009225005A (ja) データ処理装置、データ処理方法、及び、プログラム
WO2024061266A1 (zh) Mlse均衡器的实现方法和芯片、电子设备、计算机可读介质
US11804991B2 (en) Sequence detection device using path-selective sequence detection and associated sequence detection method
WO2023273589A1 (zh) 一种信号判决均衡方法以及装置
CN113796016A (zh) 符号判定装置和符号判定方法
JP2000315968A (ja) 適応型信号推定器
JP2014033347A (ja) アダプティブイコライザ、イコライザ調整方法、それを用いた半導体装置および情報ネットワーク装置
US20230308322A1 (en) Error detection and correction device capable of detecting head position of suspicious error and performing forward error propagation path tracking for providing information needed by follow-up error correction and associated method
US11831475B1 (en) Receiver using pseudo partial response maximum likelihood sequence detection
US20230308315A1 (en) Reduced-complexity maximum likelihood sequence detector suitable for m-ary signaling
US11936505B2 (en) Decision feedback equalization with efficient burst error correction
US11811566B2 (en) Methods and systems for performing adaptive equalization of data
WO2024109014A1 (zh) 误码检测方法以及相关设备
CN115473777A (zh) 一种自适应soqpsk调制解调方法、终端和接收机
WO2023179850A1 (en) Equalisation module for a digital receiver
KR100442813B1 (ko) 비선형 통신 채널 신호의 검출 방법 및 그 장치
CN116805894A (zh) 检错纠错装置及相关方法
CN117792836A (zh) 最大似然序列的检测电路、检测方法、装置和电子设备
CN114938321A (zh) 一种利用数据符号用作参考信号进行信道估计的方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23867553

Country of ref document: EP

Kind code of ref document: A1