Sphere decoding detection method and device, electronic device, and storage medium

Abstract

The disclosed embodiments provide a spherical decoding detection method and device, an electronic device, and a storage medium. The method includes: performing QR decomposition on a channel matrix to obtain a received signal of each layer, and constructing an index of a hard-decision constellation point lookup table according to the received signal; calculating the Euclidean distance of the constellation points of the corresponding layer; performing a hard-decision constellation point search on all preset hard-decision search paths, summing the Euclidean distances of the constellation points of all layers, and obtaining a hard-decision global optimal path; performing a soft-decision constellation point search according to the constellation points corresponding to the hard-decision global optimal path, obtaining a soft-decision global optimal path, a soft-decision shortest Euclidean distance, an updated inverse polarity bit matrix, and a shortest Euclidean distance, and calculating the LLR soft information corresponding to each bit of each layer according to the updated soft-decision global optimal path, the soft-decision shortest Euclidean distance, the inverse polarity bit matrix, and the shortest Euclidean distance corresponding to the inverse polarity bit matrix; the disclosed embodiments make full use of channel information to improve decoding performance.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technology, and in particular to a sphere decoding detection method and device, electronic equipment, and a storage medium. Background Art

[0002] Among the commonly used multiple input multiple output (MIMO) detection algorithms, the maximum likelihood (ML) detection algorithm has the best performance, but its computational complexity is exponentially related to the modulation order and the number of transmitting antennas, making it difficult to implement on a hardware system in an actual system. Common detection algorithms with lower computational complexity include the zero forcing algorithm and the minimum mean square error algorithm. These two algorithms are simple to implement, but have a large performance loss. Therefore, the current research focus of MIMO detection algorithms is on how to reduce the complexity of algorithm implementation and achieve or approach the detection performance of the ML algorithm.

[0003] The computational complexity of the sphere decoding (SD) algorithm is much smaller than that of the ML detection algorithm, but its performance is closest to that of the ML detection algorithm. It is an ideal signal detection algorithm. The detection principle of the sphere decoding algorithm is to find the optimal signal point by searching all constellation point grids in a hypersphere of a certain radius. The standard sphere decoding detection algorithm has a relatively high implementation complexity, which is not conducive to hardware implementation.

[0004] The relevant patent embodiment discloses a fixed complexity MIMO receiver signal search spherical decoding algorithm, which uses the results of Cholesky decomposition and zero-forcing equalization to iteratively search all constellation points under different modulation modes, calculate the cumulative Euclidean distance, and then calculate the corresponding soft bit information. This patent requires traversing and searching all constellation points to obtain the optimal path, and the hardware implementation has a high degree of recurrence.

[0005] The relevant patent embodiment discloses a spherical decoding detection method based on a very large scale integrated circuit, which uses the configured search nodes to start from the first layer to perform table lookup sorting, and retain the node with the smallest Euclidean distance, and then sequentially perform table lookup sorting on the remaining layers, and retain the search node with the smallest weight in the current layer, and when the last layer is searched, the decoding result is output. The result of this patent embodiment only contains the symbol information of the modulation result, and the utilization rate of the channel information is low.

[0006] The relevant patent embodiment also discloses a spherical decoding detection method based on depth-first search. During the search process, it is necessary to configure the upper limit of the total search nodes and the upper limit of the search nodes at different layers. Then, the search is performed layer by layer according to the depth-first principle, and the current search radius is updated to the weight of the bottom-level node searched. When the total number of search points is equal to the upper limit or the search layer cannot continue to perform the search, the decoding result is output. The bottom-level search weight of the patent embodiment is constantly updated, and the subsequent search range is also updated accordingly. The amount of calculation during its hardware implementation cannot be fixed. Summary of the invention

[0007] The main purpose of the embodiments of the present disclosure is to provide a sphere decoding detection method and device, an electronic device, and a storage medium to improve decoding performance.

[0008] To achieve the above objective, a first aspect of an embodiment of the present disclosure proposes a sphere decoding detection method, comprising:

[0009] Perform QR decomposition on the channel matrix to obtain a received signal of each layer, and construct an index of a hard-decision constellation point lookup table according to the received signal;

[0010] Calculate the Euclidean distance of the constellation points of the corresponding layer according to the constellation points corresponding to the search nodes of each layer in the index of the hard-determined constellation point lookup table;

[0011] Perform hard decision constellation point search on all preset hard decision search paths, sum the Euclidean distances of constellation points of all layers under each hard decision search path, and obtain a hard decision global optimal path, a hard decision shortest Euclidean distance, an inverse polarity bit matrix, and a shortest Euclidean distance corresponding to the inverse polarity bit matrix;

[0012] A soft-decision constellation point search is performed according to the constellation points corresponding to the hard-decision global optimal path, the hard-decision global optimal path is updated to obtain a soft-decision global optimal path, the hard-decision shortest Euclidean distance is updated to obtain a soft-decision shortest Euclidean distance, the inverse polarity bit matrix and the shortest Euclidean distance corresponding to the inverse polarity bit matrix are updated, and the LLR soft information corresponding to each bit of each layer is calculated according to the updated soft-decision global optimal path, the soft-decision shortest Euclidean distance, the inverse polarity bit matrix, and the shortest Euclidean distance corresponding to the inverse polarity bit matrix.

[0013] To achieve the above objective, a second aspect of an embodiment of the present disclosure provides a sphere decoding detection device, comprising:

[0014] A lookup table construction module is used to perform QR decomposition on the channel matrix to obtain a received signal of each layer, and construct an index of a hard-determined constellation point lookup table according to the received signal;

[0015] A calculation module, configured to calculate the Euclidean distance of the constellation points of the corresponding layer according to the constellation points corresponding to the search nodes of each layer in the index of the hard-determined constellation point lookup table;

[0016] A hard decision search module is used to perform hard decision constellation point search on all preset hard decision search paths, sum the Euclidean distances of the constellation points of all layers under each hard decision search path, and obtain a hard decision global optimal path, a hard decision shortest Euclidean distance, an inverse polarity bit matrix, and the shortest Euclidean distance corresponding to the inverse polarity bit matrix;

[0017] A soft decision search module is used to perform soft decision constellation point search according to the constellation points corresponding to the hard decision global optimal path, update the hard decision global optimal path to obtain the soft decision global optimal path, update the hard decision shortest Euclidean distance to obtain the soft decision shortest Euclidean distance, update the inverse polarity bit matrix and the shortest Euclidean distance corresponding to the inverse polarity bit matrix, and calculate the LLR soft information corresponding to each bit of each layer according to the updated soft decision global optimal path, the soft decision shortest Euclidean distance, the inverse polarity bit matrix, and the shortest Euclidean distance corresponding to the inverse polarity bit matrix.

[0018] To achieve the above objective, a third aspect of an embodiment of the present disclosure provides an electronic device, including:

[0019] Field Programmable Gate Array FPGA or Integrated Circuit ASIC;

[0020] A memory in communication with the FPGA or ASIC;

[0021] The memory stores instructions that can be executed by the FPGA or ASIC, and the instructions are executed by the FPGA or ASIC so that the FPGA or ASIC can execute the method described in the first aspect above.

[0022] To achieve the above objectives, the fourth aspect of an embodiment of the present disclosure proposes a storage medium, which stores instructions that can be executed by an FPGA or an ASIC, and the instructions are executed by the FPGA or the ASIC to implement the method described in the first aspect above.

[0023] The spherical decoding detection method and device, electronic device, and storage medium proposed in the embodiments of the present disclosure construct an index of a hard-decision constellation point lookup table according to a received signal, and calculate the Euclidean distance of the constellation point of the corresponding layer according to the constellation point corresponding to the search node of each layer in the index of the hard-decision constellation point lookup table, perform a hard-decision constellation point search on all preset hard-decision search paths, sum the Euclidean distances of the constellation points of all layers, thereby obtaining a hard-decision global optimal path, a hard-decision shortest Euclidean distance, an inverse polarity bit matrix, and the shortest Euclidean distance corresponding to the inverse polarity bit matrix, and then perform a soft-decision constellation point search according to the constellation point corresponding to the hard-decision global optimal path, update the hard-decision global optimal path to obtain a soft-decision constellation point. A global optimal path is determined, the hard-decision shortest Euclidean distance is updated to obtain a soft-decision shortest Euclidean distance, the inverse polarity bit matrix and the shortest Euclidean distance corresponding to the inverse polarity bit matrix are updated, and the LLR soft information corresponding to each bit of each layer is calculated according to the updated soft-decision global optimal path, the soft-decision shortest Euclidean distance, the inverse polarity bit matrix, and the shortest Euclidean distance corresponding to the inverse polarity bit matrix. The hard-decision global optimal path is used to further update the hard-decision shortest Euclidean distance in the soft-decision constellation point to obtain the soft-decision shortest Euclidean distance, update the inverse polarity bit matrix, and update the shortest Euclidean distance of the inverse polarity bit matrix, thereby making full use of channel information and improving decoding performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a schematic diagram of a search tree of the sphere decoding detection method provided in an embodiment of the present disclosure.

[0025] Figure 2 It is a flow chart of the sphere decoding detection method provided by the embodiment of the present disclosure.

[0026] Figure 3 yes Figure 2 Flow chart of step 201 in FIG.

[0027] Figure 4 yes Figure 2 Flow chart of step 202 in FIG.

[0028] Figure 5 yes Figure 2 Flow chart of step 203 in FIG. DETAILED DESCRIPTION

[0029] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0030] It should be noted that, although the functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first", "second", etc. in the specification, claims and the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0031] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which the present invention belongs. The terms used herein are only for the purpose of describing the embodiments of the present invention and are not intended to limit the present invention.

[0032] First, some nouns involved in this application are analyzed:

[0033] Multiple input multiple output (MIMO) is an antenna diversity technology. It is an antenna system that uses multiple antennas at both the transmitting and receiving ends to form multiple channels between the transmitting and receiving ends. The core principle of the MIMO system is to use the spatial freedom provided by multiple transmitting antennas and multiple receiving antennas to improve the spectrum efficiency in wireless communications, thereby increasing the transmission rate and improving the communication quality.

[0034] Maximum Likelihood (ML): The ML detection algorithm takes into account the effect of time dispersion on the received signal and uses the entire received signal to determine the sequence that is most likely to be sent.

[0035] Log likelihood ratio (LLR): commonly used in soft decoding.

[0036] Sphere Decoding (SD): The SD algorithm is closer to the ML algorithm, and its computational complexity is relatively smaller than that of the ML algorithm. The complexity of the SD algorithm depends on the size of the search radius. The general principle of SD is: first preset a sphere with the received signal as the center in the received signal space, and then map the sphere to an ellipsoid in the transmitted signal space, and search for possible transmitted signal points in the ellipsoid. Once a transmitted signal is found, the distance between the mapping point of the signal point and the received signal is the radius of the preset sphere, so that subsequent searches can be carried out in a smaller atmosphere.

[0037] Field Programmable Gate Array (FPGA): A semi-custom circuit in the field of Application Specific Integrated Circuit (ASIC).

[0038] Application Specific Integrated Circuit (ASIC) is an integrated circuit designed and manufactured to meet specific user requirements and the needs of specific electronic systems. ASIC design can be carried out using CPLD (Complex Programming Logic Device) and FPGA (Field Programmable Logic Array).

[0039] Quadrature Amplitude Modulation (QAM): It is a vector modulation that maps the input bits (generally using Gray code) to a complex plane (constellation) to form complex modulation symbols, and then uses amplitude modulation on the I and Q components of the symbols (corresponding to the real and imaginary parts of the complex plane, that is, the horizontal and vertical directions), respectively modulating them on two mutually orthogonal (orthogonal in the time domain) carriers (cos wt and sin wt). Compared with amplitude modulation (AM), its spectrum utilization will be doubled. QAM is a technology of amplitude and phase joint modulation, which uses the amplitude and phase of the carrier to transmit information bits at the same time. Therefore, higher bandwidth utilization can be achieved under the same minimum distance. At present, the highest QAM has reached 1024-QAM (i.e. 1024 samples). The more samples there are, the higher the transmission efficiency is. For example, a 16-QAM signal has 16 samples. Each sample represents a vector state. 16-QAM has 16 states. Each 4-bit binary number specifies one of the 16 states. 16-QAM specifies 16 combinations of carrier and phase. Each symbol and period of 16-QAM transmits 4 bits.

[0040] By calculating the accumulated Euclidean distance sum of each path and the corresponding constellation point in the hard decision search step, the constellation point corresponding to the minimum value of all accumulated Euclidean distances is the global shortest path of the hard decision algorithm, that is, the hard decision global optimal path, and then using the hard decision global optimal path to perform step 204, that is, using the hard decision global optimal path to further update the hard decision shortest Euclidean distance in the soft decision constellation point to obtain the soft decision shortest Euclidean distance, update the inverse polarity bit matrix, and update the shortest Euclidean distance of the inverse polarity bit matrix, thereby fully utilizing the channel information and improving the decoding performance.

[0041] The application scenario of the embodiment of the present disclosure is applied to the MIMO system, which performs hard and soft judgment by constructing a search tree to control the search path of the constellation points of the entire process of hard and soft judgment through the search tree, and calculates the accumulated Euclidean distance sum of each path and the corresponding constellation point in the hard judgment search step. The constellation point corresponding to the minimum value of all accumulated Euclidean distances is the global optimal path for hard judgment, and then uses the global optimal path for hard judgment to further update the shortest Euclidean distance for hard judgment in the soft judgment constellation point to obtain the shortest Euclidean distance for soft judgment, update the reverse polarity bit matrix, and update the shortest Euclidean distance of the reverse polarity bit matrix, so as to make full use of the channel environment and improve the decoding performance. Please refer to Figure 1 Schematically, the search tree provided by the embodiment of the present disclosure, wherein the highest layer is the Nth layer, the next highest layer is the N-1th layer, the N-1th layer includes two nodes, namely node 1 and node 2, and node 1 and node 2 of the N-1th layer are child nodes of the Nth layer; if N=4, then it is defined that the nodes under the highest layer 4 are all child nodes of the 4th layer, for example, the child nodes of the 4th layer include: node 1 and node 2 of the 3rd layer, node 3, node 4 and node 5 of the 2nd layer, and all nodes of the 1st layer (not shown).

[0042] The solution provided by the embodiments of the present disclosure relates to a sphere decoding detection method and device, an electronic device, and a storage medium, which are specifically described through the following embodiments. First, the sphere decoding detection method in the embodiments of the present disclosure is described.

[0043] The spherical decoding detection method provided in the embodiments of the present disclosure is applied to a MIMO system, which can be applied to a terminal or a server, or can be software running in a terminal or a server. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, or a smart watch, etc.; the server can be configured as an independent physical server, or a server cluster or a distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application that implements the spherical decoding detection method, etc., but is not limited to the above forms.

[0044] In some embodiments, the sphere decoding detection method may be implemented by FPGA; in other embodiments, the sphere decoding detection method may be implemented by ASIC.

[0045] Figure 2 is an optional flow chart of the sphere decoding detection method provided by the embodiment of the present disclosure, Figure 2 The method may include but is not limited to steps 201 to 204.

[0046] Step 201: Perform QR decomposition on the channel matrix to obtain the received signal of each layer, and construct the index of the hard decision constellation point lookup table according to the received signal;

[0047] Step 202, calculating the Euclidean distance of the constellation points of the corresponding layer according to the constellation points corresponding to the search nodes of each layer in the index of the hard-determined constellation point lookup table;

[0048] Step 203: perform hard decision constellation point search on all preset hard decision search paths, sum the Euclidean distances of constellation points of all layers under each hard decision search path, and obtain the hard decision global optimal path, the hard decision shortest Euclidean distance, the inverse polarity bit matrix, and the shortest Euclidean distance corresponding to the inverse polarity bit matrix;

[0049] Step 204: perform a soft-judgment constellation point search based on the constellation points corresponding to the hard-judgment global optimal path, update the hard-judgment global optimal path to obtain the soft-judgment global optimal path, update the hard-judgment shortest Euclidean distance to obtain the soft-judgment shortest Euclidean distance, update the inverse polarity bit matrix and the shortest Euclidean distance corresponding to the inverse polarity bit matrix, and calculate the LLR soft information corresponding to each bit of each layer based on the updated soft-judgment global optimal path, the soft-judgment shortest Euclidean distance, the inverse polarity bit matrix, and the shortest Euclidean distance corresponding to the inverse polarity bit matrix.

[0050] In step 201 of some embodiments, by performing QR decomposition on the channel matrix corresponding to the original signal, the correlation between the original signals can be removed to obtain the received signal of each layer. Step 201 also includes:

[0051] In some embodiments, see Figure 3 ,Step 201 may include but is not limited to steps 301 to 303.

[0052] Step 301: Perform QR decomposition on the channel matrix to obtain an upper triangular matrix and a received signal at each layer;

[0053] Step 302: performing equalization processing on the received signal;

[0054] Step 303: construct an index of a hard-decision constellation point lookup table according to the signal component of the received signal after equalization processing, and obtain the constellation point of each layer; specifically, in one embodiment, the signal component of the received signal is used as the index of the hard-decision constellation point lookup table, and the obtained constellation point of each layer is used as a candidate constellation point, and the candidate constellation point is used as the constellation point for calculating the Euclidean distance of the constellation point in step 202.

[0055] By performing equalization processing on the received signals in step 302, interference between received signals can be reduced.

[0056] In step 301, the channel matrix H is subjected to QR decomposition, H = Q·R, where Q is the chief matrix, QQ H =I, R is an upper triangular matrix, and R is represented by the following formula:

[0057]

[0058] Received signal Z = Q H Y is expressed as follows:

[0059]

[0060] Among them, Z N Only the signal components of the Nth layer are included, Z N-1 includes the signal components of the Nth layer and the signal components of the N-1th layer, ..., and so on, Z1 includes all the signal components from the Nth layer to the 1st layer; S N Indicates the constellation point of the Nth layer

[0061] When step 201 of the embodiment of the present disclosure starts to be executed, a preset hard judgment start flag signal takes effect, and the hardware module of the FPGA or ASIC prepares for the hard judgment process.

[0062] In some embodiments, see Figure 4 ,Step 202 may include but is not limited to steps 401 to 402.

[0063] Step 401, obtaining candidate constellation points of each layer according to the index of the hard-decision constellation point lookup table;

[0064] Step 402: Calculate the Euclidean distance of the constellation points of the corresponding layer according to the candidate constellation points of each layer.

[0065] In some embodiments, see Figure 5 ,Step 203 may include but is not limited to steps 501 to 502.

[0066] Step 501: sum the Euclidean distances of the constellation points of all layers to obtain the total accumulated Euclidean distances of the constellation points;

[0067] Step 502: Obtain all constellation points corresponding to the minimum values ​​of the accumulated Euclidean distances, and use the obtained constellation points as hard-determined global optimal paths.

[0068] In this embodiment, by calculating the accumulated Euclidean distance sum of each path and the corresponding constellation point in the hard decision search step, the constellation point corresponding to the minimum value of all accumulated Euclidean distances is the global shortest path of the hard decision process, that is, the hard decision global optimal path, that is, the search path with the smallest accumulated Euclidean distance is selected as the hard decision global optimal path.

[0069] In some embodiments, step 202 may further include:

[0070] Search the sub-nodes corresponding to different layers in turn according to the preset search tree; thus, the constellation points corresponding to the high-level layers can be continuously updated iteratively;

[0071] Calculate and retain the Euclidean distance of the constellation point corresponding to the current child node.

[0072] Specifically, the sub-nodes corresponding to different layers are searched in sequence according to the preset search tree, including:

[0073] Keep the path from the highest layer to the i-th layer unchanged, and search all child nodes of the i-1th layer under the path;

[0074] Jump back to the i-th layer;

[0075] Among them, the highest layer is the N layer, N is a positive integer, and i is a positive integer greater than 2.

[0076] For example, the highest layer is taken as the Nth layer for explanation:

[0077] Keep the first path from layer N to layer 2 unchanged, search all child nodes of layer 1 under the current first path until the search is complete, and then jump back to layer 2;

[0078] Keep the second path from layer N to layer 3 fixed, search all the child nodes of layer 2 under the current second path until the search is completed, and jump back to layer 3;

[0079] And so on, until all hard judgment search paths are traversed;

[0080] The hard decision end flag signal is generated.

[0081] Through each sub-step of the above step 202, until all hard judgment search paths specified by the search tree are traversed, the search table will give the hardware system a hard judgment end flag signal, and at this time, the constellation point corresponding to the hard judgment global optimal path, the shortest Euclidean distance corresponding to the hard judgment global optimal path, and the shortest Euclidean distance corresponding to the reverse polarity bit matrix will be sent to the back-end soft judgment module. Among them, the hard judgment process can be implemented by the hard judgment module inside the FPGA or ASIC, and the soft judgment process can be implemented by the soft judgment module inside the FPGA or ASIC.

[0082] In step 401 of some embodiments, the constellation point lookup table corresponding to 256QAM (256-order quadrature amplitude modulation) is taken as an example for explanation. 256QAM contains a total of 256 constellation points, that is, 256 constellation points to be detected. For each constellation point to be detected, a constellation point whose Euclidean distance is within a fixed range is selected as a candidate constellation point for its constellation point search. The candidate constellation points are arranged in the constellation point lookup table in the order of the Euclidean distance to the constellation point to be detected from near to far; as shown in Table 1 below, the constellation point lookup table, further, in order to reduce the table specifications of the constellation point lookup table, the 256 constellation points to be detected of 256QAM can be transformed to one of the quadrants through quadrant transformation, that is, the 256 constellation points to be detected are reduced by three quarters, and 64 constellation points to be detected are sorted, and the specifications of the constellation point lookup table are also reduced by three quarters, that is, 64 constellation points to be detected (respectively 1, 2, 3, ... 64), recorded as B i,n , where i represents the first point, n represents the constellation point, B i,1 Indicates that the constellation point to be detected corresponding to the i-th layer is 1, B i,2 Indicates that the constellation point to be detected corresponding to the i-th layer is 2, B i,3 Indicates that the constellation point to be detected corresponding to the i-th layer is 3, B i,64 It means that the number of to-be-detected constellation points corresponding to the i-th layer is 64, and the number of hard-decision constellation points corresponding to each to-be-detected constellation point is N, and the number of soft-decision constellation points corresponding to each to-be-detected constellation point is M. That is, each to-be-detected constellation point will have N hard-decision constellation points with a close Euclidean distance to it, and M soft-decision constellation points with a close Euclidean distance to it.

[0083] Constellation points to be detected Hard determination of the number of constellation points Number of soft-judged constellation points <![CDATA[B i,1 Corresponding constellation point 1]]> N M <![CDATA[B i,2 Corresponding constellation point 2]]> N M <![CDATA[B i,3 Corresponding constellation point 3]]> N M …… N M <![CDATA[B i,n Corresponding constellation point 64]]> N M

[0084] Table 1

[0085] In some embodiments, the highest layer is layer N, and step 402 includes:

[0086] Calculate the Euclidean distance of the constellation points in the Nth layer according to the constellation points corresponding to the first search node in the Nth layer;

[0087] Remove the signal component of the first search node of the Nth layer from the correlation signal of the N-1th layer, and use it as the index of the hard decision constellation point lookup table of the N-1th layer to obtain the constellation point corresponding to the first search node of the N-1th layer;

[0088] The Euclidean distance of the constellation points at the N-1th layer is calculated according to the constellation points corresponding to the first search node at the N-1th layer.

[0089] This process is repeated in this way until the constellation point corresponding to the first search node of the first layer is obtained according to all constellation point information of the N-1 layer, thus completing the hard-decision constellation point search of a complete path.

[0090] For example, taking the search node 1 at the highest level as an example, step 402 includes:

[0091] According to the constellation point B corresponding to the search node 1 of the highest layer N N,1 , calculate the Euclidean distance of the Nth layer constellation points;

[0092] Remove the signal component of the search node 1 of the highest layer N from the relevant signal of the next highest layer N-1, and use it as the index of the hard decision constellation point lookup table of the N-1 layer to obtain the constellation point B corresponding to the search node 1 of the N-1 layer. N-1,1 ;

[0093] The constellation point B corresponding to the search node 1 at the N-1th layer N-1,1 Calculate the Euclidean distance of the N-1th layer constellation points;

[0094] And so on, until the constellation point B corresponding to the search node 1 of the first layer is obtained according to all the constellation point information of the N-1 layer. 1,1 , then a hard-decision constellation point search for a complete path is completed.

[0095] In step 203, all preset hard decision search paths refer to all hard decision search paths specified by a preset search tree; the embodiment of the present disclosure constructs a search tree, and searches all specified hard decision search paths to perform hard decision constellation point search, sums the Euclidean distances of the constellation points of all layers, calculates the accumulated Euclidean distance sum of all paths and the corresponding constellation points, and uses the constellation point corresponding to the minimum value of all accumulated Euclidean distances as the global optimal path for hard decision.

[0096] In this embodiment, by calculating the accumulated Euclidean distance sum of each path and the corresponding constellation point in the hard decision search step, the constellation point corresponding to the minimum value of all accumulated Euclidean distances is the global shortest path of the hard decision algorithm, that is, the global optimal path of the hard decision.

[0097] In step 203 of some embodiments, the constellation point and the shortest Euclidean distance corresponding to the global optimal path may also be obtained. In step 203, the hard-judged constellation point search is performed from the highest layer according to the depth-first principle. For example, starting from the Nth layer of the highest layer, the hard-judged constellation point search is performed from the N-1th layer to the 1st layer in sequence until the constellation point search of a complete path is completed.

[0098] By calculating the accumulated Euclidean distance sum of each path and the corresponding constellation point in the hard decision search step, the constellation point corresponding to the minimum value of all accumulated Euclidean distances is the global shortest path of the hard decision algorithm, that is, the hard decision global optimal path, and then using the hard decision global optimal path to perform step 204, that is, using the hard decision global optimal path to further update the hard decision shortest Euclidean distance in the soft decision constellation point to obtain the soft decision shortest Euclidean distance, update the inverse polarity bit matrix, and update the shortest Euclidean distance of the inverse polarity bit matrix, thereby fully utilizing the channel information and improving the decoding performance.

[0099] In step 204, the highest layer N does not participate in the soft decision process; the shortest Euclidean distance corresponding to each maximum likelihood solution and its corresponding inverse bit matrix is ​​saved, and the shortest Euclidean distance is searched for soft decision constellation points to calculate the LLR soft information corresponding to each bit of each layer.

[0100] In some embodiments, the next highest layer is layer N-1, and step 204 includes:

[0101] The constellation point corresponding to the global optimal path of each layer of hard judgment is used as the base address of the soft judgment search table;

[0102] Find the constellation point corresponding to the first search node of each layer from the soft judgment search table according to the base address;

[0103] According to the constellation points, the global optimal path of hard judgment is updated to obtain the global optimal path of soft judgment, the shortest Euclidean distance of hard judgment is updated to obtain the shortest Euclidean distance of soft judgment, and the inverse polarity bit matrix and the shortest Euclidean distance corresponding to the inverse polarity bit matrix are updated;

[0104] The LLR soft information corresponding to each bit of each layer is calculated according to the updated soft-judgment global optimal path, the soft-judgment shortest Euclidean distance, the inverse polarity bit matrix, and the shortest Euclidean distance corresponding to the inverse polarity bit matrix.

[0105] For example, the next highest layer is the N-1th layer, and step 204 includes:

[0106] The constellation point corresponding to the global optimal path of the N-1th layer is hard judged As the base address of the soft judgment search table;

[0107] Based on the base address Searching the constellation point corresponding to the first search node 1 of the N-1th layer from the soft judgment search table; wherein all constellation points from the N-2th layer to the 1st layer are processed according to the aforementioned hard judgment process;

[0108] Jump back to layer N-1;

[0109] Based on the base address Searching the constellation point corresponding to the second search node 2 of the N-1th layer from the soft judgment search table; wherein all constellation points from the N-2th layer to the 1st layer are processed according to the aforementioned hard judgment process;

[0110] Jump back to layer N-1;

[0111] And so on, until the soft judgment constellation point search of all layers is completed;

[0112] The shortest Euclidean distance is calculated according to the reverse polarity bit matrix corresponding to the soft-judged constellation point;

[0113] The LLR soft information corresponding to each bit of each layer is calculated according to the shortest Euclidean distance corresponding to the inverse polarity bit matrix.

[0114] In the disclosed embodiment, the constellation point corresponding to the global optimal path is hard-determined at the N-1th layer. As the base address of the soft judgment lookup table, by looking up the soft judgment constellation point lookup table, the soft judgment constellation point corresponding to the soft judgment search node 1 of layer N-1 is obtained, and the constellation points of the remaining N-2 layers are processed by normal hard judgment. After the hard judgment search is completed, jump back to the N-1 layer and use Find the soft decision constellation point corresponding to the second search node of the N-1th layer, and so on until the search of all soft decision search nodes of the N-1th layer is completed. Through the soft decision search of the corresponding layer, the optimal path of the constellation point of this layer can be updated, and the shortest path of the N-1th layer, ..., 1st layer can also be updated (or the result of hard decision processing is still retained).

[0115] In step 204, the highest layer N does not participate in the soft decision processing, and directly uses the constellation point corresponding to the hard decision global optimal path As the shortest path for soft judgment, soft judgment starts from the N-1th layer and converts the constellation point corresponding to the global optimal path of hard judgment in the N-1th layer into As the base address of the soft judgment table lookup, by looking up the soft judgment constellation point table, the soft judgment constellation point corresponding to the first soft judgment search node 1 of the N-1 layer is obtained, and the constellation points of the remaining N-2 layers from the N-2 layer to the 1st layer are processed by the aforementioned hard judgment process. After the hard judgment process is completed, jump back to the N-1 layer and use the base address Find the soft decision constellation point corresponding to the second soft decision search node 2 of the N-1th layer, and so on until the search of all soft decision search nodes of the N-1th layer is completed. Through the soft decision search of the corresponding layer, the optimal path of the constellation point of this layer can be updated, and the shortest path from the N-1th layer, the N-2th layer, ..., to the 1st layer can also be updated, and the result of the hard decision processing can be saved. Then, the above process is repeated for the remaining N-2 layers until the soft decision search process from the N-2th layer to the 1st layer is completed. By repeating the above process until the soft decision search path specified by the search tree is completed, the soft decision search table will give a soft decision search end flag signal, and the soft information LLR of each bit of each layer is output at this time. The soft information LLR can be expressed as follows:

[0116]

[0117] Among them, L j,b represents the soft information of the bth bit of the jth layer, λ ML represents the shortest Euclidean distance, represents the bit at the (j,b)th position corresponding to the shortest Euclidean distance, represents the bit at position (j,b) in the reverse polarity bit matrix, Represents the minimum partial Euclidean distance corresponding to the bit at the (j, b)th position in the inverse polarity bit matrix.

[0118] In an actual application scenario, the search process of hard decision and soft decision in the sphere decoding detection method includes the following steps:

[0119] The hard judgment flag signal takes effect; specifically, when the process starts, the hard judgment flag signal takes effect, and the hard judgment module prepares for the hard judgment search process;

[0120] According to the depth-first principle, the hard-decision constellation point search is performed from the highest layer to perform a constellation point search for a complete hard-decision path; specifically, the hard-decision constellation point search is started from the Nth layer of the highest layer, and the hard-decision constellation points from the N-1th layer to the 1st layer are searched in sequence to complete the constellation point search for a complete hard-decision path;

[0121] Keep the path from layer N to layer 2 unchanged, search all the child nodes of layer 1 under the current path until the search is finished, and jump back to layer 2; keep the path from layer N to layer 3 unchanged, search all the child nodes of layer 2 under the current path until the search is finished; and so on, until the traversal of all hard-judged paths is completed;

[0122] Hard judgment search end flag signal generation;

[0123] The soft judgment flag signal takes effect; specifically, the soft judgment flag signal takes effect, and the software module prepares for the soft judgment search process;

[0124] According to the depth-first principle, the soft judgment search point search is started from the first soft judgment search point of the second highest layer, and the soft judgment point search of the first complete soft judgment path is performed; specifically, starting from the first soft judgment search point of the N-1th layer of the second highest layer, the N-2th to the 1st layer are searched in sequence to complete the search of the first complete soft judgment path; wherein, the highest layer N does not participate in the soft judgment, and the soft judgment is started from the N-1th layer of the second highest layer;

[0125] Jump back to the second soft judgment search node of the N-1th layer of the next higher layer to start the soft judgment search point search, and perform the soft judgment point search of the second complete soft judgment path; specifically, starting from the second soft judgment search point of the N-1th layer of the next higher layer, search the N-2th to the 1st layer in sequence, and complete the search of the second complete soft judgment path, until the search of all the soft judgment search nodes of the N-1th layer is completed;

[0126] Repeat the previous two steps to complete the soft judgment search point search of the N-2th layer, the N-3th layer, ..., until the 1st layer;

[0127] A soft judgment end flag signal is generated; that is, the soft judgment search process is ended, and a complete constellation point search process, namely, a hard judgment search point search process and a soft judgment search point search process, is completed.

[0128] In the spherical decoding detection method provided by the embodiment of the present disclosure, an index of a hard-decision constellation point lookup table is constructed according to a received signal, and the Euclidean distance of the constellation point of the corresponding layer is calculated according to the constellation point corresponding to the search node of each layer in the index of the hard-decision constellation point lookup table, and a hard-decision constellation point search is performed on all preset hard-decision search paths, and the Euclidean distances of the constellation points of all layers are summed up to obtain a hard-decision global optimal path, a hard-decision shortest Euclidean distance, an inverse polarity bit matrix, and the shortest Euclidean distance corresponding to the inverse polarity bit matrix, and then a soft-decision constellation point search is performed according to the constellation point corresponding to the hard-decision global optimal path, and the hard-decision global optimal path is updated to obtain the soft-decision global optimal path. Path, update the shortest Euclidean distance of hard judgment to obtain the shortest Euclidean distance of soft judgment, update the inverse polarity bit matrix and the shortest Euclidean distance corresponding to the inverse polarity bit matrix, and calculate the LLR soft information corresponding to each bit of each layer according to the updated global optimal path of soft judgment, the shortest Euclidean distance of soft judgment, the inverse polarity bit matrix, and the shortest Euclidean distance corresponding to the inverse polarity bit matrix; select a fixed number of constellation points in the search of each layer (the number of each layer can be different), traverse all paths in turn, and obtain the accumulated Euclidean distance and corresponding constellation point of each path at the same time. The constellation point corresponding to the minimum value of all accumulated Euclidean distances is the global optimal path of hard judgment. Using the optimal path obtained by hard judgment, further update the shortest Euclidean distance of hard judgment in the candidate soft judgment constellation points to obtain the shortest Euclidean distance of soft judgment, update the inverse polarity bit matrix, and update the shortest Euclidean distance of the inverse polarity bit matrix, so that the channel information can be fully utilized and the decoding performance can be improved.

[0129] The disclosed embodiment can switch the corresponding constellation point lookup table and constellation point search table according to different modulation modes and antenna layers, so as to implement the sphere decoding algorithm through hardware with low and fixed computational complexity. The disclosed embodiment can better improve the decoding performance of sphere decoding, and through the sphere decoding method combining hard judgment and soft judgment, it can effectively utilize channel information (such as channel environment, etc.) to improve the decoding performance of the algorithm.

[0130] The disclosed embodiment constructs search trees and constellation point lookup tables of different constellation points according to different channel qualities (channel environments). The receiver can realize online switching according to the actual channel environment during operation, that is, the path of the constellation point search tree and the constellation point lookup table can be switched online, and all constellation point search paths can be padded to the same number of search nodes as the longest path using invalid paths, which can better ensure the decoding performance of the algorithm, reduce the complexity of the algorithm in the hardware implementation process, and fix the computational complexity of the algorithm, making it easier to implement in hardware. Compared with the MMSE signal detection algorithm that is widely used in physical receivers, the disclosed embodiment has better decoding performance. In addition, the disclosed embodiment uses FPGA or ASIC hardware implementation, which can realize the quantitative production of products and improve communication quality.

[0131] The present disclosure also provides a sphere decoding detection device, which can implement the above sphere decoding detection method, and the device includes:

[0132] A lookup table construction module is used to perform QR decomposition on the channel matrix to obtain the received signal of each layer, and construct the index of the hard decision constellation point lookup table according to the received signal;

[0133] A calculation module, used to calculate the Euclidean distance of the constellation points of the corresponding layer according to the constellation points corresponding to the search nodes of each layer in the index of the hard-determined constellation point lookup table;

[0134] A hard decision search module is used to search for hard decision constellation points on all preset hard decision search paths, sum the Euclidean distances of constellation points of all layers, and obtain the global optimal hard decision path, the shortest Euclidean distance of hard decision, the inverse polarity bit matrix, and the shortest Euclidean distance corresponding to the inverse polarity bit matrix;

[0135] The soft judgment search module is used to search the soft judgment constellation points according to the constellation points corresponding to the hard judgment global optimal path, update the hard judgment global optimal path to obtain the soft judgment global optimal path, update the hard judgment shortest Euclidean distance to obtain the soft judgment shortest Euclidean distance, update the inverse polarity bit matrix and the shortest Euclidean distance corresponding to the inverse polarity bit matrix, and calculate the LLR soft information corresponding to each bit of each layer according to the updated soft judgment global optimal path, the soft judgment shortest Euclidean distance, the inverse polarity bit matrix, and the shortest Euclidean distance corresponding to the inverse polarity bit matrix.

[0136] The present disclosure also provides an electronic device, including:

[0137] Field Programmable Gate Array FPGA or Integrated Circuit ASIC;

[0138] Memory that communicates with the FPGA or ASIC;

[0139] The memory stores instructions that can be executed by the FPGA or ASIC, and the instructions are executed by the FPGA or ASIC so that the FPGA or ASIC can perform the aforementioned sphere decoding detection method.

[0140] The FPGA or ASIC may include but is not limited to a hard decision module and a soft decision module. The hard decision module includes the above-mentioned lookup table construction module, calculation module, and hard decision search module. The soft decision module includes the above-mentioned soft decision search module.

[0141] The embodiment of the present disclosure further provides a storage medium, which stores instructions that can be executed by FPGA or ASIC, and the instructions are executed by FPGA or ASIC to implement the sphere decoding detection method as described above.

[0142] The spherical decoding detection method, spherical decoding detection device, electronic device and storage medium proposed in the embodiments of the present disclosure construct an index of a hard-decision constellation point lookup table according to a received signal, and calculate the Euclidean distance of the constellation point of the corresponding layer according to the constellation point corresponding to the search node of each layer in the index of the hard-decision constellation point lookup table, perform a hard-decision constellation point search on all preset hard-decision search paths, sum the Euclidean distances of the constellation points of all layers, thereby obtaining the shortest Euclidean distance corresponding to the hard-decision global optimal path and the inverse polarity bit matrix, and then perform a soft-decision constellation point search according to the constellation point corresponding to the hard-decision global optimal path, and update the hard-decision global optimal path to obtain The global optimal path of soft judgment is used, the shortest Euclidean distance of hard judgment is updated to obtain the shortest Euclidean distance of soft judgment, the inverse polarity bit matrix and the shortest Euclidean distance corresponding to the inverse polarity bit matrix are updated, and the LLR soft information corresponding to each bit of each layer is calculated according to the updated global optimal path of soft judgment, the shortest Euclidean distance of soft judgment, the inverse polarity bit matrix, and the shortest Euclidean distance corresponding to the inverse polarity bit matrix. The global optimal path of hard judgment is used to further update the shortest Euclidean distance of hard judgment in the soft judgment constellation point to obtain the shortest Euclidean distance of soft judgment, update the inverse polarity bit matrix, and update the shortest Euclidean distance of the inverse polarity bit matrix, thereby making full use of channel information and improving decoding performance.

[0143] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0144] The embodiments described in the embodiments of the present disclosure are intended to more clearly illustrate the technical solutions of the embodiments of the present disclosure and do not constitute a limitation on the technical solutions provided by the embodiments of the present disclosure. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of the present disclosure are also applicable to similar technical problems.

[0145] It can be understood by those skilled in the art that Figure 2-5 The sphere decoding detection method shown in the figure does not constitute a limitation on the embodiments of the present disclosure, and may include more or fewer steps than those shown in the figure, or a combination of certain steps, or different steps.

[0146] The device embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0147] Those skilled in the art will appreciate that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices may be implemented as software, firmware, hardware, or a suitable combination thereof.

[0148] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0149] It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0150] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0151] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0152] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0153] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including multiple instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store programs.

[0154] The preferred embodiments of the present disclosure are described above with reference to the accompanying drawings, but the scope of the rights of the present disclosure is not limited thereto. Any modification, equivalent substitution and improvement made by those skilled in the art without departing from the scope and essence of the present disclosure should be within the scope of the rights of the present disclosure.

Claims

1. A sphere decoding detection method, characterized in that: include: Perform QR decomposition on the channel matrix to obtain a received signal of each layer, and construct an index of a hard-decision constellation point lookup table according to the received signal; Calculate the Euclidean distance of the constellation points of the corresponding layer according to the constellation points corresponding to the search nodes of each layer in the index of the hard-determined constellation point lookup table; Perform hard decision constellation point search on all preset hard decision search paths, sum the Euclidean distances of constellation points of all layers under each hard decision search path, and obtain a hard decision global optimal path, a hard decision shortest Euclidean distance, an inverse polarity bit matrix, and a shortest Euclidean distance corresponding to the inverse polarity bit matrix; A soft-decision constellation point search is performed according to the constellation points corresponding to the hard-decision global optimal path, the hard-decision global optimal path is updated to obtain a soft-decision global optimal path, the hard-decision shortest Euclidean distance is updated to obtain a soft-decision shortest Euclidean distance, the inverse polarity bit matrix and the shortest Euclidean distance corresponding to the inverse polarity bit matrix are updated, and the LLR soft information corresponding to each bit of each layer is calculated according to the updated soft-decision global optimal path, the soft-decision shortest Euclidean distance, the inverse polarity bit matrix, and the shortest Euclidean distance corresponding to the inverse polarity bit matrix.

2. The method according to claim 1, characterized in that The hard decision constellation point search is performed on all preset hard decision search paths, and the Euclidean distances of the constellation points of all layers are summed to obtain the shortest Euclidean distance corresponding to the hard decision global optimal path and the reverse polarity bit matrix, including: The Euclidean distances of the constellation points of all layers are summed to obtain the total accumulated Euclidean distances of the constellation points; Obtain all constellation points corresponding to the minimum values ​​of the accumulated Euclidean distances, and use the obtained constellation points as the hard-determined global optimal path.

3. The method according to claim 1, characterized in that: The calculating of the Euclidean distance of the constellation points of the corresponding layer according to the constellation points corresponding to the search nodes of each layer in the index of the hard-determined constellation point lookup table also includes: Search the child nodes corresponding to different layers in sequence according to the preset search tree; Calculate and retain the Euclidean distance of the constellation point corresponding to the current child node.

4. The method according to claim 3, characterized in that The searching of sub-nodes corresponding to different layers in sequence according to the preset search tree includes: Keep the path from the highest layer to the i-th layer unchanged, and search all child nodes of the i-1th layer under the path; Jump back to the i-th layer; Among them, the highest layer is the Nth layer, N is a positive integer, and i is a positive integer greater than 2.

5. The method according to claim 1, characterized in that The highest layer is the Nth layer, and the calculation of the Euclidean distance of the constellation points of the corresponding layer according to the candidate constellation points of each layer includes: Calculate the Euclidean distance of the constellation points in the Nth layer according to the constellation points corresponding to the first search node in the Nth layer; Remove the signal component of the first search node of the Nth layer from the correlation signal of the N-1th layer, and use it as the index of the hard decision constellation point lookup table of the N-1th layer to obtain the constellation point corresponding to the first search node of the N-1th layer; The Euclidean distance of the constellation points at the N-1th layer is calculated according to the constellation points corresponding to the first search node at the N-1th layer.

6. The method according to any one of claims 1 to 5, characterized in that: The QR decomposition of the channel matrix is ​​performed to obtain a received signal of each layer, and an index of a hard-determined constellation point lookup table is constructed according to the received signal, including: Performing QR decomposition on the channel matrix to obtain an upper triangular matrix and the received signal of each layer; performing equalization processing on the received signal; The index of the hard-decision constellation point lookup table is constructed according to the signal components of the received signal after equalization processing to obtain the constellation points of each layer.

7. The method according to any one of claims 1 to 5, characterized in that: The next higher layer is the N-1th layer, and the soft-decision constellation point search is performed according to the hard-decision global optimal path and the shortest Euclidean distance corresponding to the reverse polarity bit matrix to obtain LLR soft information corresponding to each bit of each layer, including: The constellation point corresponding to the global optimal path of each layer of hard judgment is used as the base address of the soft judgment search table; Searching the constellation point corresponding to the first search node of each layer from the soft judgment search table according to the base address; updating the hard-decision global optimal path according to the constellation point to obtain the soft-decision global optimal path, updating the hard-decision shortest Euclidean distance to obtain the soft-decision shortest Euclidean distance, and updating the inverse polarity bit matrix and the shortest Euclidean distance corresponding to the inverse polarity bit matrix; The LLR soft information corresponding to each bit of each layer is calculated according to the updated soft-decision global optimal path, the soft-decision shortest Euclidean distance, the inverse polarity bit matrix, and the shortest Euclidean distance corresponding to the inverse polarity bit matrix.

8. A sphere decoding detection device, characterized in that: include: A lookup table construction module is used to perform QR decomposition on the channel matrix to obtain a received signal of each layer, and construct an index of a hard-determined constellation point lookup table according to the received signal; A calculation module, configured to calculate the Euclidean distance of the constellation points of the corresponding layer according to the constellation points corresponding to the search nodes of each layer in the index of the hard-determined constellation point lookup table; A hard decision search module is used to perform hard decision constellation point search on all preset hard decision search paths, sum the Euclidean distances of the constellation points of all layers under each hard decision search path, and obtain a hard decision global optimal path, a hard decision shortest Euclidean distance, an inverse polarity bit matrix, and the shortest Euclidean distance corresponding to the inverse polarity bit matrix; A soft decision search module is used to perform soft decision constellation point search according to the constellation points corresponding to the hard decision global optimal path, update the hard decision global optimal path to obtain the soft decision global optimal path, update the hard decision shortest Euclidean distance to obtain the soft decision shortest Euclidean distance, update the inverse polarity bit matrix and the shortest Euclidean distance corresponding to the inverse polarity bit matrix, and calculate the LLR soft information corresponding to each bit of each layer according to the updated soft decision global optimal path, the soft decision shortest Euclidean distance, the inverse polarity bit matrix, and the shortest Euclidean distance corresponding to the inverse polarity bit matrix.

9. An electronic device, characterized in that: include: Field Programmable Gate Array FPGA or Integrated Circuit ASIC; A memory in communication with the FPGA or ASIC; The memory stores instructions that can be executed by the FPGA or ASIC, and the instructions are executed by the FPGA or ASIC so that the FPGA or ASIC can execute the method according to any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium stores instructions that can be executed by the FPGA or ASIC, and the instructions are executed by the FPGA or ASIC to implement the method according to any one of claims 1 to 7.

Images