Verify data integrity in the receiver
By calculating the product of the parity check matrix and data messages in parallel to verify CRC, the problem of low processing efficiency of CRC is solved and more efficient data integrity verification is achieved.
Patent Information
- Application Number
- CN202180023737.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Priority Date
- 2020-03-23
- Filing Date
- 2021-03-05
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2041-03-05
AI Technical Summary
In the prior art, when cyclic redundancy verification (CRC) is used in wireless communication networks to verify data integrity, the processing volume increases with the message length, resulting in low processing efficiency. Especially in software implementation, it is difficult to efficiently calculate CRC.
By dividing the parity check matrix into column groups or non-overlapping matrix element groups, multiplying it with the corresponding part of the data message, performing parallel calculation of partial check sub vectors, and summing them to obtain the complete check sub vector, verifying the data The integrity of the message.
The processing efficiency and resource utilization are improved, so that parallel operations can be processed independently when verifying data integrity in the receiver, reducing the overhead of processing time.
Smart Images

Figure CN115280696B_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates generally to the field of data integrity. More particularly, it relates to verifying data integrity in a receiver in a wireless communication network. Background Art
[0002] To verify data integrity in a receiver, a cyclic redundancy check, CRC, is often used to detect errors in messages transmitted over a potentially unreliable channel, such as a radio channel, in a wireless communication network.
[0003] CRC is often implemented in dedicated hardware, but can also be implemented in software on a processor.
[0004] A drawback of hardware and / or software implemented CRCs is that the amount of processing increases with the length of the message and renders the processing inefficient in terms of processing time and overhead incurred.
[0005] Therefore, an alternative method for verifying data integrity in the receiver is needed. Summary of the invention
[0006] It should be emphasized that the term "include / comprises" when used in this specification is used to specify the presence of stated features, integers, steps or components, but does not exclude the presence or addition of one or more other features, integers, steps, components or groups thereof. As used herein, the singular forms "a", "a", and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.
[0007] Generally speaking, when referring to an arrangement in this document, the arrangement is to be understood as a physical product; for example, a device. The physical product may include one or more parts, such as control circuitry in the form of one or more controllers, one or more processors, and the like.
[0008] It is an object of some embodiments to address or mitigate, alleviate, or eliminate at least some of the above or other disadvantages.
[0009] According to a first aspect, this is achieved by a method for verifying data integrity in a receiver in a wireless communication network.
[0010] The method comprises: receiving a data message, wherein the data message comprises a group of data elements and a checksum; and calculating a complete syndrome vector based on a partial syndrome vector, wherein the partial syndrome vector is calculated in parallel by multiplying portions of a parity check matrix with corresponding portions of the received data message.
[0011] The method further includes: determining that all vector elements of the complete syndrome vector are zero; and verifying that the received data message is correct when all vector elements of the complete syndrome vector are zero, and incorrect otherwise.
[0012] In some embodiments, the method further includes: dividing the parity check matrix into column groups and associating each column group with a corresponding portion of the data message; calculating partial syndrome vectors in parallel by multiplying each column group with its corresponding portion of the data message; and summing the calculated partial syndrome vectors to obtain a complete syndrome vector.
[0013] In some embodiments, the method further includes: dividing the parity check matrix into non-overlapping matrix element groups, and associating each non-overlapping matrix element group with a corresponding portion of the data message; calculating partial syndrome vectors in parallel by multiplying each non-overlapping matrix element group with its corresponding portion of the data message; and summing the calculated partial syndrome vectors to obtain a complete syndrome vector.
[0014] In some embodiments, the method further includes: partitioning the parity check matrix into non-overlapping groups of matrix elements from the column groups, and associating each non-overlapping group of matrix elements with a corresponding portion of the data message; calculating partial syndrome vectors in parallel by multiplying each non-overlapping group of matrix elements from the column groups with its corresponding portion of the data message; and summing the calculated partial syndrome vectors to obtain a complete syndrome vector.
[0015] In some embodiments, the parity check matrix is associated with a checksum polynomial from which the parity check matrix is calculated.
[0016] In some embodiments, the checksum is based on groups of data elements and is appended to the groups of data elements in the data message to form a codeword.
[0017] In some embodiments, the received data message has been calculated by the transmitter and transmitted to the receiver.
[0018] In some embodiments, the checksum is one or more of: a cyclic redundancy check CRC and a block code checksum.
[0019] In some embodiments, a complete syndrome vector including all zero vector elements indicates a matching checksum and / or an error-free group of data elements.
[0020] In some embodiments, the parity check matrix is pre-computed and stored in a memory of the receiver and / or in a remote memory accessible to the receiver.
[0021] In some embodiments, the parity-check matrix is a binary matrix.
[0022] In some embodiments, the parity check matrix includes rows that are shifted versions of each other.
[0023] In some embodiments, given a portion of a data message, a corresponding partial syndrome is calculated by first reading a pre-computed portion of a parity check matrix from memory and initializing a partial syndrome vector.
[0024] In some embodiments, the bits of the portion of the data message are accessed one by one, and when a one is encountered, a first set of bits of the parity check matrix is added to the partial syndrome vector.
[0025] In some embodiments, the bits of the portion of the data message are accessed one by one, and when a zero is encountered, the portion of the syndrome vector is left unchanged.
[0026] In some embodiments, where each accessed column in the parity check matrix is a shifted version of the previous column, only one new bit is shifted in when every bit of the portion of the data message and the portion of the parity check matrix has been accessed.
[0027] In some embodiments, one or more software threads and / or processing elements for execution are assigned to one or more portions of the parity check matrix and corresponding portions of the data message to calculate partial syndrome vectors in parallel.
[0028] A second aspect is a computer program product comprising a non-transitory computer readable medium having thereon a computer program comprising program instructions. The computer program is loadable into a data processing unit and is configured to: when the computer program is run by the data processing unit, cause the execution of the method according to the first aspect.
[0029] A third aspect is an apparatus for verifying data integrity in a receiver in a wireless communication network.
[0030] The apparatus includes a controller configured to cause: receiving a data message, wherein the data message includes a group of data elements and a checksum; and calculating a complete syndrome vector based on a partial syndrome vector, wherein the partial syndrome vector is calculated in parallel by multiplying portions of a parity check matrix with corresponding portions of the received data message.
[0031] The controller is further configured to cause: determining that all vector elements of the complete syndrome vector are zero; and verifying that the received data message is correct when all vector elements of the complete syndrome vector are zero, and incorrect otherwise.
[0032] In some embodiments, the controller is further configured to cause: partitioning the parity check matrix into column groups and associating each column group with a corresponding portion of the data message; calculating partial syndrome vectors in parallel by multiplying each column group with its corresponding portion of the data message; and summing the calculated partial syndrome vectors to obtain a complete syndrome vector.
[0033] In some embodiments, the controller is further configured to cause: partitioning the parity check matrix into non-overlapping groups of matrix elements and associating each non-overlapping group of matrix elements with a corresponding portion of a data message; calculating partial syndrome vectors in parallel by multiplying each non-overlapping group of matrix elements with its corresponding portion of the data message; and summing the calculated partial syndrome vectors to obtain a complete syndrome vector.
[0034] In some embodiments, the controller is further configured to cause: partitioning the parity check matrix into non-overlapping groups of matrix elements from the column groups, and associating each non-overlapping group of matrix elements with a corresponding portion of the data message; calculating partial syndrome vectors in parallel by multiplying each non-overlapping group of matrix elements from the column groups with its corresponding portion of the data message; and summing the calculated partial syndrome vectors to obtain a complete syndrome vector.
[0035] In some embodiments, the parity check matrix is associated with a checksum polynomial from which the parity check matrix is calculated.
[0036] In some embodiments, the checksum is based on groups of data elements and is appended to the groups of data elements in the data message to form a codeword.
[0037] In some embodiments, the received data message has been calculated by the transmitter and transmitted to the receiver.
[0038] In some embodiments, the checksum is one or more of: a cyclic redundancy check CRC and a block code checksum.
[0039] In some embodiments, a complete syndrome vector including all zero vector elements indicates a matching checksum and / or an error-free group of data elements.
[0040] In some embodiments, the parity check matrix is pre-computed and stored in a memory of the receiver and / or in a remote memory accessible to the receiver.
[0041] In some embodiments, the parity-check matrix is a binary matrix.
[0042] In some embodiments, the parity check matrix includes rows that are shifted versions of each other.
[0043] In some embodiments, the controller is further configured to cause: given a portion of the data message, calculating a corresponding partial syndrome by reading a pre-computed portion of the parity check matrix from the memory and initializing the partial syndrome vector.
[0044] In some embodiments, the controller is further configured to cause: accessing bits of the portion of the data message one by one, and when a one is encountered, adding a first set of bits of the parity check matrix to the partial syndrome vector.
[0045] In some embodiments, the controller is further configured to cause: accessing bits of the portion of the data message one by one, and when a zero is encountered, leaving the portion of the syndrome vector unchanged.
[0046] In some embodiments, the controller is further configured to cause: shifting a row of the parity check matrix when each bit of the portion of the data message has been accessed.
[0047] In some embodiments, one or more software threads and / or processing elements for execution are assigned to one or more portions of the parity check matrix and corresponding portions of the data message to calculate partial syndrome vectors in parallel.
[0048] In some embodiments, the device is operatively connected to general purpose hardware having one or more parallel processing elements configured to process parallelized computations.
[0049] In some embodiments, each of the one or more parallel processing elements is configured to compute in parallel a respective one or more of the partial syndrome vectors by multiplying a portion of the parity check matrix with a corresponding portion of the data message.
[0050] In some embodiments, general purpose hardware having one or more parallel processing elements is included in the receiver and / or in a cloud environment.
[0051] In some embodiments, the general purpose hardware includes a graphics processing unit (GPU).
[0052] A fourth aspect is a receiver in a wireless communication network, the receiver comprising the device according to the third aspect.
[0053] In some embodiments, the receiver is configured to receive the data and verify the integrity of the data.
[0054] Any of the above aspects may additionally have features that are the same as or correspond to any of the various features as described above for any of the other aspects.
[0055] An advantage of some embodiments is that an alternative method for verifying data integrity in a receiver is provided.
[0056] Another advantage of some embodiments is that parallelization of operations improves processing efficiency in terms of processing time and incurred overhead than was possible according to prior art approaches.
[0057] Yet another advantage of some embodiments is that parallelization of operations enables processing elements in hardware to process operations independently and at different timings from each other.
[0058] Yet another advantage of some embodiments is that parallelization of operations enables a more efficient implementation in terms of utilizing available resources in hardware than is possible according to prior art approaches.
[0059] It should be noted that even though embodiments are described herein in the context of verifying data integrity in a receiver, some embodiments may be equally applicable and / or beneficial in other contexts. BRIEF DESCRIPTION OF THE DRAWINGS
[0060] Further objects, features and advantages will appear from the following detailed description of the embodiments with reference to the accompanying drawings.The drawings are not necessarily drawn to scale, emphasis instead being placed upon illustrating example embodiments.
[0061] Figure 1 is a flow chart illustrating example method steps according to some embodiments;
[0062] Figure 2a is a schematic block diagram illustrating example operations according to some embodiments;
[0063] Figure 2b is a schematic block diagram illustrating example operations according to some embodiments;
[0064] Figure 2c is a schematic block diagram illustrating example operations according to some embodiments;
[0065] Figure 3a is a schematic block diagram illustrating example operations according to some embodiments;
[0066] Figure 3b is a schematic diagram illustrating an example portion of a parity check matrix according to some embodiments;
[0067] Figure 3c is a schematic diagram illustrating an example portion of a data message and an example portion of a parity check matrix according to some embodiments;
[0068] Figure 4 is a schematic block diagram illustrating an example apparatus according to some embodiments; and
[0069] Figure 5 is a schematic diagram illustrating an example computer-readable medium in accordance with some embodiments. DETAILED DESCRIPTION
[0070] As mentioned above, it should be emphasized that the term "includes / comprising" when used in this specification is used to specify the presence of stated features, integers, steps or components, but does not exclude the presence or addition of one or more other features, integers, steps, components or groups thereof. As used herein, the singular forms "a", "a", and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.
[0071] Embodiments of the present disclosure will be described and illustrated more fully below with reference to the accompanying drawings. However, the solutions disclosed herein can be implemented in many different forms and should not be construed as being limited to only the embodiments set forth herein.
[0072] As mentioned above, a drawback of hardware and / or software implemented CRCs is that the amount of processing increases with message length, and renders the processing inefficient in terms of processing time and incurred overhead.
[0073] In the following, embodiments will be presented in which alternative approaches for verifying data integrity in a receiver are described.
[0074] As described herein, general purpose hardware generally includes hardware including one or more parallel processing elements configured to process operations in parallel, wherein the one or more processing elements further include local data memory for processing the operations.
[0075] As described herein, operations generally include matrix and vector operations, where operations may include calculations of matrix multiplications and matrix products, and vector operations may include summations of matrix products.
[0076] As described herein, the checksum typically includes a cyclic redundancy check CRC and / or a block code checksum.
[0077] The CRC is usually implemented using some kind of shift register technique, usually in a serial manner. The data message is processed piecewise while the shift register is updated. Finally, the shift register state forms the checksum to be appended. In its most basic form, the data message is processed one bit at a time, but several consecutive bits, such as 8 or 32 bits, may also be processed together.
[0078] The CRC calculation can also be parallelized by splitting the data message into multiple subsequences and calculating a temporary CRC for each subsequence. A relatively complex linear function is then applied to each temporary CRC, and the final CRC is obtained by adding the results using XOR addition.
[0079] Because the amount of processing increases with the length of the message, the processing time will be long for long messages. This is especially true for software implementations. Therefore, this makes it problematic to implement CRC on general-purpose processors.
[0080] Since the parallelization is based on serial calculation for each subsequence, the complex functions that need to be performed for all subsequences in the end add extra computational burden, and thus it is difficult to efficiently calculate CRC on modern hardware including one or more processing elements.
[0081] It should be noted that even though embodiments are described herein in the context of verifying data integrity in a receiver in a wireless communication network, where the receiver may be operably connected to general-purpose hardware having one or more parallel processing elements configured to handle parallelized computations, some embodiments may be equally applicable and / or beneficial in other contexts where checksums are used to verify data integrity of received messages.
[0082] It should also be noted that even though embodiments are described herein in the context of general purpose hardware having one or more parallel processing elements configured to process parallelized computations, some embodiments may be equally applicable and / or beneficial in other contexts where each of the one or more processing elements is configured to process a set of computations serially.
[0083] Furthermore, it should be noted that even though embodiments are described herein in the context of summing calculated partial syndrome vectors to obtain a complete syndrome vector, some embodiments may also be equally applicable and / or beneficial in other contexts where two or more of the calculated partial syndrome vectors calculated in one of one or more parallel processing elements are also summed in the one parallel processing element, and where the summed two or more calculated partial syndrome vectors form a subset of those calculated partial syndrome vectors that are summed to obtain the complete syndrome vector.
[0084] Figure 1 is a flow chart showing method steps of an example method 100 according to some embodiments. The method 100 is used to verify data integrity in a receiver in a wireless communication network. Thus, the method 100 may be performed, for example, by Figure 4 The device 400 and / or controller 410 are executed; all of which will be described later in this document.
[0085] The method 100 includes the following steps.
[0086] In step 101, a data message is received, wherein the data message includes a data element group and a checksum (refer to Figures 2a to 2c ).
[0087] Alternatively or additionally, a data element group may comprise a sequence of data elements.
[0088] In some embodiments, the received data message has been calculated by the transmitter and transmitted to the receiver.
[0089] For example, the receiver receiving the message may be included in a mobile communication device and operate in a wireless communication network.
[0090] In some embodiments, the checksum is based on a group of data elements and is appended to the group of data elements in the data message to form a codeword (see Figure 2a ).
[0091] In some embodiments, the checksum is one or more of: a cyclic redundancy check CRC and a block code checksum.
[0092] As mentioned above, checksums (e.g., CRCs) are often used to detect errors in messages transmitted over potentially unreliable channels (such as radio channels). For example, the 5G NR standard uses several different CRCs in the air interface; these are described in 3GPP TS 38.212 Section 5.1.
[0093] As mentioned above, CRC is a checksum, wherein the principle of the checksum is that for the transmitter, a sequence, i.e. a checksum, is calculated based on a group of data elements and appended to the data message including the group of data elements (refer to Figure 2a ).
[0094] The receiver then performs the same calculation and checks (ie verifies) that the checksum matches, in which case the data message is declared error-free. If the checksum does not match, the non-matching checksum indicates that the data message contains errors and is typically discarded.
[0095] The receiver can emulate the transmitter by calculating a CRC based on the received data message, and then comparing that CRC to the appended CRC sequence. Another way to check the CRC is to run the same CRC calculation algorithm on the entire data message (including the CRC). In this case, the calculated value should be zero.
[0096] CRC is a linear block code in which the transmitted data message including the checksum is considered as a codeword. In addition, CRC is usually applied to binary sequences and is therefore a binary linear block code. CRC can also be used on data messages based on larger alphabets.
[0097] CRC is defined mathematically using long polynomial division. The divisor polynomial is based on the data message, and the dividend is a fixed polynomial (the CRC polynomial). The quotient is discarded, and the resulting CRC sequence is derived from the remainder.
[0098] In step 102, a complete syndrome vector is calculated based on the partial syndrome vectors, wherein the partial syndrome vectors are calculated in parallel by multiplying portions of the parity check matrix with corresponding portions of the received data message (see Figures 3a to 3c ).
[0099] In some embodiments, the parity check matrix is associated with a checksum polynomial (eg, a CRC polynomial) from which the parity check matrix is calculated.
[0100] In some embodiments, the parity-check matrix is a binary matrix.
[0101] In some embodiments, the parity check matrix includes rows that are shifted versions of each other.
[0102] Alternatively or additionally, a parity check matrix of this form (ie, rows are shifted versions of each other) requires only one row to be stored in memory, thus reducing the required memory size and the number of read operations.
[0103] In some embodiments, where each accessed column in the parity check matrix is a shifted version of the previous column, where when every bit of the portion of the data message and the portion of the parity check matrix has been accessed, only one new bit is shifted in.
[0104] Alternatively or additionally, the rows of the parity check matrix may be generated from a linear feedback shift register whose feedback polynomial is the same as (or the reverse of) the polynomial defining the CRC.
[0105] In some embodiments, the bits of the portion of the data message are accessed one by one, and when a one is encountered, a first set of bits of the parity check matrix is added to the partial syndrome vector.
[0106] In some embodiments, the bits of the portion of the data message are accessed one by one, and when a zero is encountered, the portion of the syndrome vector is left unchanged.
[0107] Alternatively or additionally, the complete syndrome vector is calculated as the sum of the calculated partial syndrome vectors (see Figures 3a to 3c ).
[0108] Alternatively or additionally, the integrity syndrome vector indicates data integrity of the received data message.
[0109] In some embodiments, a complete syndrome vector including all zero vector elements indicates a matching checksum and / or an error-free group of data elements.
[0110] In some embodiments, the parity check matrix is pre-computed and stored in a memory of the receiver and / or in a remote memory accessible to the receiver.
[0111] In some embodiments, given a portion of a data message, a corresponding partial syndrome is calculated by first reading a pre-computed portion of a parity check matrix from memory and initializing a partial syndrome vector.
[0112] In some embodiments, one or more software threads and / or processing elements for execution are assigned to one or more portions of the parity check matrix and corresponding portions of the data message to calculate partial syndrome vectors in parallel.
[0113] In optional step 102a-1, in some embodiments, the parity check matrix is divided into column groups, and each column group is associated with a corresponding portion of the data message (see Figures 3a to 3c ).
[0114] In optional step 102b-1, in some embodiments, partial syndrome vectors are calculated in parallel by multiplying each column group by its corresponding portion of the data message (see Figures 3a to 3c ).
[0115] In optional step 102a-2, in some embodiments, the parity check matrix is divided into non-overlapping groups of matrix elements, and each non-overlapping group of matrix elements is associated with a corresponding portion of the data message (see Figures 3a to 3c ).
[0116] Alternatively or additionally, groups of all-zero matrix elements may be discarded, since their partial syndrome vectors will be zero and will not contribute to the sum (cf. step 102c).
[0117] In optional step 102b-2, in some embodiments, partial syndrome vectors are calculated in parallel by multiplying each non-overlapping group of matrix elements by its corresponding portion of the data message (see Figures 3a to 3c ).
[0118] In optional step 102a-3, in some embodiments, the parity check matrix is divided into non-overlapping groups of matrix elements from the column groups, and each non-overlapping group of matrix elements is associated with a corresponding portion of the data message (see Figures 3a to 3c ).
[0119] Alternatively or additionally, groups of all-zero matrix elements may be discarded, since their partial syndrome vectors will be zero and will not contribute to the sum (cf. step 102c).
[0120] In optional step 102b-3, in some embodiments, the partial syndrome vectors are calculated in parallel by multiplying each non-overlapping group of matrix elements from the column group with its corresponding portion of the data message (see Figures 3a to 3c ).
[0121] In optional step 102c, in some embodiments, the calculated partial syndrome vectors are summed to obtain a complete syndrome vector (see Figure 3a ).
[0122] For example, by having each parallel processing element calculate 10 partial syndrome vectors in succession, an implementation can calculate 1000 partial syndrome vectors using only 100 parallel processing elements. Each parallel processing element can then calculate the sum of its 10 partial syndrome vectors. Thus, it can be considered that 100 partial syndrome vectors are calculated in parallel by 100 parallel processing elements, each parallel processing element calculating its partial syndrome vector by iterating over some portions.
[0123] Alternatively or additionally, the sum of the partial syndrome vectors may be calculated, either by calculating the sum of subsets of the partial syndrome vectors in one or more parallel processing elements and then summing the subset sums to obtain the complete syndrome vector, or by calculating the sum of the partial syndrome vectors calculated by one or more parallel processing elements to obtain the complete syndrome vector.
[0124] In step 103 , a first possibility is that all vector elements of the complete syndrome vector are determined to be zero, ie, the complete syndrome vector is a zero vector (“yes” path out of step 103 ).
[0125] In step 103 , a second possibility is that one or more of the vector elements of the complete syndrome vector are determined to be non-zero, ie, the complete syndrome vector is a non-zero vector (“No” path out of step 103 ).
[0126] In step 104 (not shown), it is verified whether the received data message is correct or incorrect.
[0127] In step 104a, when all vector elements of the complete syndrome vector are zero (ie, a zero vector), it is verified that the received data message is correct ("yes" path out of step 103).
[0128] In step 104b, when one or more of the vector elements of the complete syndrome vector are non-zero (ie, a non-zero vector), it is verified that the received data message is incorrect ("No" path out of step 103).
[0129] Figure 1 Any of the above steps may additionally have features that are the same as or correspond to any of the various features described below with respect to FIGS. 2 to 5 .
[0130] Figure 2ais a schematic block diagram illustrating example operations according to some embodiments. The operations of FIG. 200a are performed in a transmitter, wherein these operations enable verification of data integrity in a receiver in a wireless communication network. Thus, the operations of FIG. 200a may be configured, for example, to perform Figure 1 One or more of the method steps of .
[0131] Figure 2a A data message is shown including a checksum sequence (CRC (L bits)), which is based on a data element group, ie, a data sequence (message (K bits)), appended to a data message further including the data element group.
[0132] Figure 2a Further illustrated is a CRC calculation wherein a block-by-block or bit-by-bit input of a group of data elements is provided and a calculated CRC is output and appended to the data message.
[0133] Figure 2b is a schematic block diagram illustrating example operations according to some embodiments. The operations of FIG. 200b are used to verify data integrity in a receiver in a wireless communication network. Thus, the operations of FIG. 200b may be configured, for example, to perform Figure 1 One or more of the method steps of .
[0134] Figure 2b A received data message (N=K+L bits) including a CRC is shown, wherein the CRC is calculated for verifying data integrity, and wherein a block-by-block or bit-by-bit input of the received data message including the CRC is provided as input to the calculation, and a vector (L bits) is output to verify that the vector includes all-zero vector elements.
[0135] Figure 2c is a schematic block diagram illustrating example operations according to some embodiments. The operations of FIG. 200c are used to verify data integrity in a receiver in a wireless communication network. Thus, the operations of FIG. 200c may be configured, for example, to perform Figure 1 One or more of the method steps of .
[0136] Figure 2c A received data message (N bits) and a checksum associated parity check matrix (L×N) including a CRC are shown, wherein the received data message and the checksum associated parity check matrix are multiplied in a matrix-vector multiplication and an output vector (L bits) is output to verify that the vector includes all zero vector elements.
[0137] Figure 3a is a schematic block diagram illustrating example operations according to some embodiments. The operations of FIG. 300a are used to verify data integrity in a receiver in a wireless communication network. Thus, the operations of FIG. 300a may be configured, for example, to perform Figure 1 One or more of the method steps of .
[0138] Figure 3a Calculation of a complete syndrome vector based on partial syndrome vectors is shown, wherein the partial syndrome vectors are calculated in parallel by multiplying (MULT) portions of a parity check matrix with corresponding portions of a received data message (i.e., portions of a received sequence), and the calculated partial syndrome vectors are summed (SUM) to obtain a complete syndrome vector.
[0139] Figure 3a It is further shown that: determining that all vector elements of the complete syndrome vector are zero (checking whether they are zero), and verifying that the received data message is correct when all vector elements of the complete syndrome vector are zero, and incorrect otherwise.
[0140] As described herein, matrix operations can be applied to matrices of arbitrary size, limited only by the size of the local data memory associated with one or more processing elements. In order to make processing more efficient and utilize one or more processing elements as much as possible, the matrix is split into several parts for processing.
[0141] Any binary linear block code can be defined by a parity check matrix (usually denoted H). H is an L×N binary matrix, where L is the number of redundant bits added, N is the length of the codeword, and K=NL is the length of the data message. A binary sequence of length N If and only if it satisfies the matrix equation is the codeword (indicating that the CRC is correct). The CRC is a special case of a binary linear block code in which the redundant bits are the calculated CRC. Therefore, L is the length of the CRC.
[0142] Parity check matrices are not unique: If A is a non-singular L×L matrix, then AH and H are parity check matrices for the same code.
[0143] Therefore, we can calculate the syndrome vector of length L And check that it is all zero to perform a check CRC. Because all elements are binary, matrix multiplication can be implemented by logical AND (i.e., binary multiplication) and logical XOR (i.e., binary addition). This makes both software implementation and hardware implementation possible.
[0144] There are different ways to parallelize the computation. Items 1 to 3 below describe different computations:
[0145] 1. Parallel Computing different parts of , checking that each part is zero. This provides at most L-way parallelization.
[0146] 2. Change the sequence Divide into M parts (M≤N) and calculate the sum of the corresponding parts, then sum the sums of these parts together to get This provides up to N-way parallelization. The summation can be done partially in parallel in the tree, Before, logM final steps are required. Each part can be or it can be a subsequence of A collection of elements of any segment.
[0147] 3. The combination of items 1 and 2 provides at most L×N-way parallelization.
[0148] Because L is usually a fixed small number (usually L ≤ 32), and N can be on the order of thousands, the second item above can often provide sufficient parallelization by itself.
[0149] Then suppose the elements of H are , so that each row is the row above shifted left by one step. This means that for a sequence of length N, it is necessary to store only N+L-1 precomputed parity bits (or, by not storing the L-1 starting zeros in the first row and only storing N of the parity bits).
[0150] Syndrome Vector can be written as the (modulo-2) sum ,in is the jth column of H. Given the assumptions here, we get Note that each term in this sum is either an all-zero vector (when c j = 0), or a subsequence of the q sequence (when c j =1).
[0151] Given a subsequence , by first reading the precomputed parity sequence from memory And will Initialize to all zero vector to calculate the corresponding partial syndrome Then, access bit c one by one b 、c b+1 etc.: When a "one" is encountered, add the first L bits of the parity sequence to ; When encountering "zero", After each bit is accessed, the parity sequence is shifted left by one step.
[0152] Given some processing elements, i.e. one or more cores, each core can be assigned to calculate the partial syndrome sum. For each such subsequence, the above method can be used. If a core is assigned multiple subsequences, it can simply iterate over those subsequences and add up the partial syndromes. When the cores have completed their calculations, their results can be added together to get the final sum, which is the overall syndrome.
[0153] Figure 3b is a diagram illustrating an example portion of a parity check matrix according to some embodiments.
[0154] Figure 3b A partitioned parity check matrix is shown, the matrix including elements corresponding to portions thereof, which portions of the matrix may be multiplied with corresponding portions of a data message (eg, subsets of data sequence elements).
[0155] Figure 3b Further shown are portions of the partitioned parity-check matrix, including one or more of: non-contiguous portions, portions spanning a subset of rows and columns, non-rectangular portions, and portions having only zero elements (which may be skipped).
[0156] Figure 3c is a schematic diagram illustrating an example portion of a data message and an example portion of a parity check matrix according to some embodiments.
[0157] Figure 3c A parity check matrix is shown, wherein the parity check matrix is partitioned, and wherein the partitioned matrix portion corresponds to a portion of a data message, ie, a subset of a data sequence including elements associated with the partitioned matrix portion.
[0158] Figure 3c Further shown is a partial syndrome vector, ie a partial syndrome, which is calculated by multiplying the partitioned matrix portion with a corresponding portion of the data message, ie a subset of the data sequence including elements associated with the partitioned matrix portion.
[0159] It should be noted that the calculation of the partial syndrome vectors is performed in parallel and may be performed by hardware (eg, a GPU) including one or more processing elements, wherein the one or more processing elements are configured to perform the calculations in parallel.
[0160] The hardware including one or more processing elements is thus configured to minimize processing time, and by using one or more processing elements for parallel computations, these computations will be time-series independent.
[0161] Figure 44 is a schematic block diagram illustrating an example device 400 according to some embodiments. The device 400 is used to verify data integrity in a receiver in a wireless communication network. Thus, the device 400 and / or the controller 410 may, for example, be configured to perform Figure 1 One or more of the method steps of , and / or one or more of any steps otherwise described herein.
[0162] The apparatus 400 includes a controller 410 (e.g., device control circuitry) configured to cause: receiving a data message, wherein the data message includes a group of data elements and a checksum; and calculating a complete syndrome vector based on a partial syndrome vector, wherein the partial syndrome vector is calculated in parallel by multiplying portions of a parity-check matrix with corresponding portions of the received data message.
[0163] The controller 410 is further configured to cause: determining that all vector elements of the complete syndrome vector are zero; and verifying that the received data message is correct when all vector elements of the complete syndrome vector are zero, and incorrect otherwise.
[0164] As mentioned above, the device 400 includes a controller (CNTR; for example, a control circuit or a control module) 410, which may include (or be otherwise associated with; for example, connected or connectable to) a receiver 401, such as a receiving circuit or a receiving module, and the receiver 401 is configured to receive a data message, wherein the data message includes a data element group and a checksum (with Figure 1 compared to step 101).
[0165] The controller 410 further includes (or is otherwise associated with; e.g., connected or connectable to) a computer 402, such as a computing circuit or computing module, configured to calculate a complete syndrome vector based on the partial syndrome vector, wherein the partial syndrome vector is calculated in parallel (with the partial syndrome vector being multiplied by a portion of the parity check matrix with a corresponding portion of the received data message). Figure 1 compared with step 102).
[0166] The controller 410 further includes (or is otherwise associated with; for example, connected or connectable to) a determiner 403, such as a determination circuit or a determination module, the determiner 403 being configured to determine that all vector elements of the complete syndrome vector are zero (with Figure 1 compared with step 103).
[0167] The controller 410 further includes (or is otherwise associated with; e.g., connected or connectable to) validators 404a and 404b, such as validation circuits or validation modules, configured to validate that a received data message is correct when all vector elements of a complete syndrome vector are zero (as opposed to Figure 1 104a), otherwise it is incorrect (compared with step 104a of Figure 1 compared to step 104b).
[0168] In some embodiments, the controller 410 further includes (or is otherwise associated with; for example, connected or connectable to): a divider 402a-1, such as a dividing circuit or a dividing module, the divider 402a-1 is configured to divide the parity check matrix into column groups and associate each column group with a corresponding portion of the data message (with Figure 1 and a computer 402b-1, such as a computing circuit or a computing module, configured to calculate in parallel the partial syndrome vectors (with Figure 1 compared to step 102b-1).
[0169] In some embodiments, the controller 410 further includes (or is otherwise associated with; for example, connected or connectable to): a divider 402a-2, such as a dividing circuit or a dividing module, the divider 402a-2 is configured to divide the parity check matrix into non-overlapping matrix element groups, and associate each non-overlapping matrix element group with a corresponding portion of the data message (with Figure 1 and a computer 402b-2, such as a computing circuit or a computing module, configured to calculate the partial syndrome vectors (with Figure 1 compared to step 102b-2).
[0170] In some embodiments, the controller 410 further includes (or is otherwise associated with; e.g., connected or connectable to): a divider 402a-3, e.g., a divider circuit or a divider module, the divider 402a-3 being configured to divide the parity check matrix into non-overlapping groups of matrix elements from the column groups, and to associate each non-overlapping group of matrix elements with a corresponding portion of the data message (with Figure 1 and a computer 402b-3, such as a computing circuit or computing module, configured to calculate in parallel the partial syndrome vectors (with Figure 1 compared to step 102b-3).
[0171] In some embodiments, the controller 410 further includes (or is otherwise associated with; for example, connected to or connectable to) a summer 402c, such as a summing circuit or a summing module, the summer 402c being configured to sum the calculated partial syndrome vectors to obtain a complete syndrome vector (with Figure 1 compared to step 102c).
[0172] In some embodiments, the device is operatively connected to general purpose hardware having one or more parallel processing elements configured to process parallelized computations.
[0173] In some embodiments, each of the one or more parallel processing elements is configured to compute in parallel a respective one or more of the partial syndrome vectors by multiplying a portion of the parity check matrix with a corresponding portion of the data message.
[0174] In some embodiments, general purpose hardware having one or more parallel processing elements is included in the receiver and / or in a cloud environment.
[0175] In some embodiments, the general purpose hardware includes a graphics processing unit (GPU).
[0176] In some embodiments, the device 400 may further optionally include (or be otherwise associated; e.g., connected or connectable to) a transceiver TX / RX 420, such as a transceiver circuit or a transceiver module, configured to send and receive radio signals, e.g., based on verification of data integrity in a receiver in a wireless communication network.
[0177] Generally speaking, when referring to an arrangement in this document, the arrangement is to be understood as a physical product; for example, a device. The physical product may include one or more parts, such as control circuitry in the form of one or more controllers, one or more processors, and the like.
[0178] The described embodiments and their equivalents may be implemented in software or hardware or a combination thereof. These embodiments may be performed by general purpose circuits. Examples of general purpose circuits include: digital signal processors (DSPs), central processing units (CPUs), graphics processing units (GPUs), coprocessor units, field programmable gate arrays (FPGAs), and other programmable hardware. Alternatively or additionally, the embodiments may be performed by dedicated circuits such as application specific integrated circuits (ASICs). General purpose circuits and / or dedicated circuits may be associated with or included in a device such as a wireless communication device, for example.
[0179] Embodiments may occur in an electronic device (such as a wireless communication device), which includes an arrangement, circuitry and / or logic according to any of the embodiments described herein. Alternatively or additionally, an electronic device (such as a wireless communication device) may be configured to perform a method according to any of the embodiments described herein.
[0180] According to some embodiments, the computer program product comprises a computer readable medium, such as for example a Universal Serial Bus (USB) memory, a plug-in card, an embedded drive or a Read Only Memory (ROM).
[0181] Figure 5 An example computer readable medium is shown in the form of a compact disk (CD) ROM 500. The computer readable medium has a computer program stored thereon, the computer program comprising program instructions. The computer program may be loaded into a data processor (PROC) 520, which may be included in a wireless communication device 510, for example. When loaded into the data processor, the computer program may be stored in a memory (MEM) 530 associated with or included in the data processor.
[0182] In some embodiments, the computer program may, when loaded into and executed by a data processing unit, cause the computer to be executed according to, for example, Figure 1 The performance of one or more of the method steps and / or any steps otherwise described herein.
[0183] In some embodiments, the computer program may, when loaded into and executed by a data processing unit, cause the computer to be executed according to, for example, Figures 2a to 2c and / or Figure 3a The performance of one or more of the operations and / or any operations otherwise described herein.
[0184] Generally, all terms used herein are to be interpreted according to their ordinary meaning in the relevant technical field, unless a different meaning is clearly given and / or implied from the context in which the term is used.
[0185] Various embodiments have been referenced herein. However, those skilled in the art will recognize many variations to the described embodiments that would still fall within the scope of the claims.
[0186] For example, the method embodiments described herein disclose example methods by steps performed in a certain order. However, it is to be appreciated that these sequences of events may occur in another order without departing from the scope of the claims. In addition, even if some method steps have been described as being performed in sequence, they may also be performed in parallel. Therefore, unless a step is explicitly described as being after or before another step and / or in a situation where it is implied that a step must be after or before another step, the steps of any method disclosed herein do not necessarily have to be performed in the exact order disclosed.
[0187] Likewise, it should be noted that in the description of the embodiments, the division of the functional block into specific units is by no means intended to be limiting. On the contrary, these divisions are merely examples. The functional block described as a unit herein can be split into two or more units. In addition, the functional block described as being implemented as two or more units herein can be merged into fewer (e.g., single) units.
[0188] Where appropriate, any features of any embodiment disclosed herein may be applied to any other embodiment. Likewise, any advantages of any embodiment in these embodiments may be applied to any other embodiment, and vice versa.
[0189] Therefore, it should be understood that the details of the described embodiments are merely examples set forth for illustrative purposes and that all modifications coming within the scope of the claims are intended to be embraced therein.
Claims
1. A method for verifying data integrity in a receiver in a wireless communication network, the method comprising the following steps: receiving (101) a data message, wherein the data message comprises a group of data elements and a checksum, wherein the checksum is a cyclic redundancy check CRC, A complete syndrome vector is calculated (102) based on the partial syndrome vector, wherein the partial syndrome vector is calculated in parallel by multiplying portions of a parity check matrix with corresponding portions of the received data message, wherein the parity check matrix is associated with a check sum polynomial from which the parity check matrix is calculated, and wherein each accessed column in the parity check matrix is a shifted version of a previous column, wherein when each bit of the portion of the data message and the portion of the parity check matrix has been accessed, only a new bit is shifted in, Determine (103) that all vector elements of the complete syndrome vector are zero, and It is verified (104) that the received data message is correct (104a) when all the vector elements of the complete syndrome vector are zero, and incorrect (104b) otherwise.
2. The method according to claim 1, further comprising the steps of: dividing (102a-1) the parity check matrix into column groups and associating each column group with a corresponding portion of the data message, computing (102b-1) partial syndrome vectors in parallel by multiplying each column group by its corresponding portion of said data message, and The calculated partial syndrome vectors are summed (102c) to obtain a complete syndrome vector.
3. The method according to any one of claims 1 to 2, further comprising the following steps: dividing (102a-2) the parity check matrix into non-overlapping groups of matrix elements and associating each non-overlapping group of matrix elements with a corresponding portion of the data message, computing (102b-2) partial syndrome vectors in parallel by multiplying each non-overlapping group of matrix elements by its corresponding portion of said data message, and The calculated partial syndrome vectors are summed (102c) to obtain a complete syndrome vector.
4. The method according to any one of claims 1 to 2, further comprising the following steps: partitioning (102a-3) the parity check matrix into non-overlapping groups of matrix elements from the column groups and associating each non-overlapping group of matrix elements with a corresponding portion of the data message, computing (102b-3) partial syndrome vectors in parallel by multiplying each non-overlapping group of matrix elements from said column groups with its corresponding portion of said data message, and The calculated partial syndrome vectors are summed (102c) to obtain a complete syndrome vector.
5. The method according to any one of claims 1 to 2, wherein: The checksum is one or more of the following: a cyclic redundancy check CRC and a block code checksum.
6. The method according to any one of claims 1 to 2, wherein: The complete syndrome vector including all zero vector elements indicates a matching checksum and / or an error-free group of data elements.
7. The method according to any one of claims 1 to 2, wherein: The parity check matrix is pre-computed and stored in a memory of the receiver and / or in a remote memory accessible to the receiver.
8. The method according to any one of claims 1 to 2, wherein: The parity check matrix is a binary matrix.
9. The method according to any one of claims 1 to 2, wherein: The parity check matrix includes rows that are shifted versions of each other.
10. The method according to any one of claims 1 to 2, wherein: Given the portion of the data message, a corresponding partial syndrome is calculated by first reading a pre-computed portion of the parity check matrix from memory and initializing the partial syndrome vector.
11. The method according to any one of claims 1 to 2, wherein: The bits of the portion of the data message are accessed one by one, and when a one is encountered, a first set of bits of the parity check matrix is added to the partial syndrome vector.
12. The method according to any one of claims 1 to 2, wherein: The bits of the portion of the data message are accessed one by one, and when a zero is encountered, the portion of the syndrome vector is left unchanged.
13. The method according to any one of claims 1 to 2, wherein: One or more software threads and / or processing elements for execution are assigned to one or more portions of the parity check matrix and corresponding portions of the data message to calculate the partial syndrome vectors in parallel.
14. A computer program product comprising a non-transitory computer-readable medium having thereon a computer program comprising program instructions, the computer program being loadable into a data processing unit and being configured to cause the execution of the method according to any one of claims 1 to 13 when the computer program is run by the data processing unit.
15. A receiver in a wireless communication network, comprising an apparatus for verifying data integrity in a receiver in a wireless communication network, the apparatus comprising a controller (410), the controller being configured to cause: Receive a data message where The data message comprises a data element group and a checksum, wherein the checksum is a cyclic redundancy check CRC, a complete syndrome vector is calculated based on the partial syndrome vector, wherein the partial syndrome vector is calculated in parallel by multiplying portions of a parity check matrix with corresponding portions of a received data message, wherein the parity check matrix is associated with a check sum polynomial from which the parity check matrix is calculated, and wherein each accessed column in the parity check matrix is a shifted version of a previous column, wherein when each bit of the portion of the data message and the portion of the parity check matrix has been accessed, only a new bit is shifted in, determining that all vector elements of the complete syndrome vector are zero, and Verifying that the received data message is correct when all of the vector elements of the complete syndrome vector are zero, and incorrect otherwise.
