A preamble sequence design method for future multiple access

Abstract

The present invention belongs to the field of wireless communication technology, and specifically is a method for designing a preamble sequence for future multiple access. The present invention specifically includes: determining the number of bits required for source coding based on the scale of potential users; using cyclic redundancy check and FEC coding scheme for channel coding, and outputting FEC-encoded codewords; passing the FEC-encoded codewords through an interleaved memory to combat deep channel fading; establishing a mapping relationship between codewords and sparse modulation symbols, and then mapping the codewords to preamble sequences through sparse modulation, which satisfy the Bernoulli characteristic; the preamble sequence designed by the present invention can effectively reduce the missed detection probability and false alarm probability of active user detection by the base station, and the potential user scale that can be served can reach 10 7 Orders of magnitude or more.

Description

Technical Field

[0001] The present invention belongs to the technical field of wireless communications, and in particular relates to a preamble sequence design method for future multiple access scenarios. Background Art

[0002] To address the large-scale access issues brought about by the growing number of mobile broadband users, the 3rd Generation Partnership Project (3GPP) has listed massive machine type communications (mMTC) and user connectivity as important applications for the next generation of mobile communication systems.

[0003] In the context of the Internet of Things, the potential user scale of the next generation communication system may be very large, reaching 10 6 or 10 7 The number of active users at a given time point is of the order of magnitude, while the size of the active user group at the same time point is relatively small. In this scenario, assigning orthogonal preamble sequences to users is unrealistic, which will introduce additional multi-access interference. Traditional pseudo-random sequences such as m-sequences and Gold sequences, as well as the widely used ZC (Zadoff-Chu) sequence, have limited sequence spaces and cannot be extended to ultra-large-scale multiple access. While the recently proposed random Gaussian codebook has a huge sequence space, its application is limited by the high complexity of receiver processing, which is proportional to the total number of users.

[0004] Therefore, how to design a new preamble sequence technology to effectively resist multiple access interference, accommodate a large number of users and facilitate receiver processing is of great significance to the further development of the next generation mobile communication system. Summary of the Invention

[0005] The purpose of the present invention is to provide a novel preamble sequence design method to reduce multi-access interference in ultra-large-scale multiple access scenarios, so as to serve future multiple access scenarios.

[0006] The novel preamble sequence design method for future multiple access provided by the present invention is based on sparse modulation and FEC coding technology, and the specific steps are as follows:

[0007] Step (1) determining the number of bits required for source coding based on the potential user scale;

[0008] Assume that the number of potential users is N and the number of bits required to identify users is B i , satisfying the relationship So given N,B i Determined by the following formula:

[0009]

[0010] in, Indicates rounding up; in actual scenarios, the user identity (Identity, ID) can be assigned a value between 0 and N numbers in, and then convert ID into a number containing B i The binary form of a bit number;

[0011] Step (2): Channel coding using cyclic redundancy check and FEC coding scheme

[0012] That is, based on the number of bits required for source coding in step (1), cyclic redundancy check and FEC coding scheme are used for channel coding, and the FEC-coded codeword is output;

[0013] In order to reduce the false alarm rate at the base station and facilitate receiver processing, it is decided whether to add an error detection code after source coding. The commonly used error detection method is cyclic redundancy check (CRC). Let the number of bits contained in the error detection code be B c (B c ≥0), then the code length after cyclic redundancy check is (B i +B c ), in order to reduce the bit error rate, an FEC coding scheme is adopted to output the FEC-encoded codeword.

[0014] Step (3), interweaving;

[0015] The codewords after FEC encoding are passed through an interleaving memory to combat deep channel fading.

[0016] Step (4), sparse modulation;

[0017] A mapping relationship between codewords and sparse modulation symbols is established, and then the codewords obtained in step (3) are mapped to a leading sequence through sparse modulation; so that the sequence satisfies the Bernoulli characteristic and has certain sparse characteristics.

[0018] The preamble sequence designed by the present invention can effectively reduce the probability of missed detection (MD) and false alarm (FA) of active user detection by the base station, and the potential user scale that can be served can reach 10 7 Orders of magnitude or more. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 Flow chart of the method of the present invention.

[0020] Figure 2The figure is a simulation comparison diagram of the present invention and the existing design, as well as the performance of the present invention under large-scale address access. DETAILED DESCRIPTION

[0021] The present invention will be further described below in conjunction with the accompanying drawings:

[0022] like Figure 1 As shown, the present invention provides a preamble sequence design method for future multiple access, comprising the following steps:

[0023] Step (1) Determine the number of bits required for source coding based on the potential user scale

[0024] Potential user size N and the number of bits required to identify users B i Needs Satisfaction So given N,B i Determined by the following formula:

[0025]

[0026] in Indicates rounding up; in actual scenarios, the user identity (Identity, ID) can be assigned a value between 0 and N numbers in, and then convert ID into a number containing B i The binary form of a bit number.

[0027] Step (2) Use cyclic redundancy check and FEC coding scheme for channel coding

[0028] In order to reduce the false alarm rate at the base station and facilitate the processing of the receiver, it is decided whether to add an error detection code after the source coding. The error detection method generally used is the Cyclic Redundancy Check (CRC). Assume that the number of bits contained in the error detection code is B c (B c ≥0), then the code length after cyclic redundancy check is (B i +B c ); Because the symbol duration of the preamble sequence is limited, the codeword length is limited. When selecting the FEC coding scheme, it is considered suitable to be applied to the short code BCH (Bose-Chaudhuri-Hocquenghem, BCH) code; considering the original BCH code, the specific steps of its encoding method are as follows:

[0029] Step (2.1) determines the generating polynomial

[0030] Step (2.1.1) Let the total code length after FEC encoding be n, based on which we can select a polynomial of degree m over the finite field GF(2) and construct the extended field GF(2) of GF(2) m), m and n satisfy the relationship n=2 m -1;

[0031] Step (2.1.2) Let α be GF(2 m ) primitive element, if the number of error bits to be corrected is t, then the generating polynomial g(x) of the BCH code is α 1 ,α 2 ,…,α 2t The lowest degree polynomial on GF(2) with roots; let m i (x) is α i The minimum polynomial of

[0032] g(x)=LCM[m1(x),m2(x),…,m 2t (x)]

[0033] Where LCM[·] means taking the least common multiple; since α 2i and α i The minimal polynomial of is the same, so

[0034] g(x)=LCM[m1(x),m3(x),…,m 2t-1 (x)]

[0035] Step (2.2) Determine the parameters of the BCH code

[0036] The code length after FEC encoding in step (2.2.1) is n=2 m -1;

[0037] Step (2.2.2) Let k = (B i +B c ), the number of parity bits required is r=nk≤mt;

[0038] Step (2.2.3) Minimum distance d between codewords min ≥2t+1;

[0039] The number of bits that can be corrected in step (2.2.4) is t;

[0040] Step (2.3) obtains the FEC encoding codeword;

[0041] Step (2.3.1) Let the information code polynomial after source coding and cyclic redundancy check be q(x), and divide x by g(x) (n-k) q(x), we get the quotient S(x) and the remainder r(x), that is

[0042]

[0043] Then the code group c(x)=x can be output (n-k) q(x)+r(x);

[0044] Step (2.3.2) outputs the FEC-encoded codeword based on the encoding group obtained in step 2.3.1).

[0045] Step (3) passes the FEC-encoded codeword through an interleaving memory to combat deep channel fading.

[0046] Step (4) Mapping the codeword to the preamble sequence through sparse modulation

[0047] The sparse modulation method specifically adopted in this embodiment is sparse index modulation; the detailed description and steps are as follows:

[0048] Step (4.1) determines the pattern of the sparse index modulation signal

[0049] Let the preamble sequence after sparse index modulation be Where L is the symbol duration of the sequence, x i Denotes the i-th element of x (1≤i≤L), and defines the signal x that satisfies the following three conditions as a sparse index modulation signal:

[0050] Step (4.1.1) divides x into Q equal segments, and the length of each segment is Segments are independent and identically distributed;

[0051] Step (4.1.2) specifies that each segment has only one non-zero entry, i.e.

[0052]

[0053] in represents the indicator function, the positions of non-zero entries are random;

[0054] In step (4.1.3), for the qth segment of x, without loss of generality, let the lth element be a non-zero item, then

[0055]

[0056] in represents the set of constellation points, so

[0057] Step (4.2) determines the number of bits that the preamble sequence can carry

[0058] It can be verified that if the leading sequence x meets the above requirements, then for the qth segment of x, the number of bits that the segment can carry is Taking into account the actual needs, in order to eliminate phase ambiguity in real communication scenarios, it is necessary to set the first non-zero item of the preamble sequence signal x as the reference symbol, so the number of bits that the preamble sequence can actually carry is recorded as B p, expressed as

[0059]

[0060] Obviously, B p Must meet B p ≥n, that is, the length of the FEC code cannot be greater than the number of bits that the preamble sequence can carry; however, in order to save system resources, the FEC coding parameters and the preamble sequence construction conditions must be coordinated in advance so that B p =n.

[0061] Step (4.3) simplifies the mapping relationship

[0062] Due to the special properties of the preamble sequence x, in practice, only the mapping between the signal segments and specific bit vectors needs to be constructed, which greatly compresses the space before and after the mapping. However, since the non-zero entries in the first segment of x are reference symbols, the only information carried is the position of the non-zero entries. Therefore, the mapping relationship between the first segment and the other segments is considered separately.

[0063] Step (4.3.1) determines the first segment mapping relationship

[0064] Because there are L non-zero positions in this segment s So the number of bits that can be carried is log2 L s , construct log2 L s Gray code with bits (the distance between adjacent code words is 1), corresponding to L s This situation requires that the Hamming distance of the codes corresponding to adjacent non-zero items is the smallest, which is only 1.

[0065] Step (4.3.2) determines the mapping relationship of other segments

[0066] consider That is QPSK, let x i (1≤i≤Q) represents the i-th segment of x; x q The number of bits that can be carried B s , so it constitutes x q The set of all cases has vector, denoted as It can be proved that i≠j, there is Where ||·||2 represents the Euclidean distance, and the two vectors with the longest distance have the same non-zero entries but opposite polarities. Therefore, when constructing the mapping relationship, the two segments with the longest distance must also have the largest Hamming distance corresponding to the codeword vectors. This is constructed as follows:

[0067] The number of bits to be constructed is B s =log2 Ls +2 standard Gray code, and divide it into 4 equal parts in sequence; the code words of the first Gray code are matched one-to-one with all possible segments with non-zero items of 1, and a mapping relationship is established; the code words of the first Gray code are inverted, and the code words after the operation are matched one-to-one with all possible segments with non-zero items of -1, and a mapping relationship is established. In fact, the code word set after the inversion is equal to the code word set composed of the third Gray code; the code words of the fourth Gray code are matched one-to-one with all possible segments with non-zero items of -i; the code words of the fourth Gray code are inverted, and the code words after the operation are matched one-to-one with all possible segments with non-zero items of i, and a mapping relationship is established. In fact, the code word set after the inversion is equal to the code word set composed of the second Gray code; note that the above mappings all require that the Hamming distance of the codes corresponding to adjacent non-zero items is the minimum, which is only 1.

[0068] Step (4.4) completes the codeword vector c∈{0,1} n To the leader sequence Mapping

[0069] The codeword output after the interleaving memory is specified as c∈{0,1+ n , the 1st to the log2 L of c s bits are mapped according to the first mapping relationship in step 4.3.1), and the remaining n-log2 L s After the bits are divided into Q-1 equal parts, they are mapped according to the mapping relationship of other segments in step 4.3.3) to complete the mapping of codeword c to sequence x.

[0070] Experiment: Considering a massive multiple access scenario, the user scale is huge, but the size of the active user group at a given time node is relatively small. The number of active users is set to K = 40, the base station has M = 200 antennas, the user end is a single antenna, the sequence design length is L = 104, the channel model is a block fading channel, the channel coefficient remains unchanged within the coherence time, and considering the power allocation, the channel is equivalent to a Rayleigh fading channel. The new preamble sequence of this embodiment is divided into Q = 13 segments, each segment is L S =8, each segment has only one non-zero item and the non-zero item is a QPSK symbol, then the number of bits that the preamble sequence of this embodiment can carry is Using BCH encoding, set the source encoding result after cyclic redundancy check, that is, the code length before encoding is k = (B i +B c ) = 24, and then BCH encoded into n = 63 bits. After passing through the interleaving memory, the codeword is mapped into a sparse index modulation signal. After a given preamble sequence, two scenarios are considered: the traditional multiple access scenario and the future-oriented ultra-large-scale multiple access scenario. The detailed description is as follows:

[0071] Scenario (1), for traditional multiple access, the number of users is 10 5 Below, the total number of users N is set to 2 16 For Gaussian sequences, the Multiple-Measurement Vector Approximate Message Passing (MMV-AMP) algorithm is used for detection. This detection algorithm requires the base station to store the leading sequences of all users. The present invention uses the Bilinear Generalized Approximate Message Passing (BiG-AMP) algorithm for detection. The sent leading sequence is obtained through sparse matrix decomposition to further determine which users are active. The simulation results are shown in Figure 2. Figure 2 As shown, the performance of the present invention is better than that of the Gaussian sequence, measured by the probability of missed detection.

[0072] Scenario (2): For future ultra-large-scale multiple access, the number of users may exceed 10 5 Even more, set N=2 20 and N = 2 24 Due to the extremely large number of users, the base station is not required to store the user's preamble sequence. The Bilinear Generalized Approximate Message Passing (BiG-AMP) algorithm is used for detection. The preamble sequence sent is obtained through sparse matrix decomposition. The simulation results are as follows: Figure 2 As shown, with the missed detection probability as the measure, in the large-scale site scenario, the present invention supports an extremely large number of users, reaching 10 6 , 10 7 Even more.

[0073] The above descriptions are only partial embodiments of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the present invention, and these improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A preamble sequence design method for future multiple access, characterized in that: The specific steps are as follows: Step (1) determining the number of bits required for source coding based on the potential user scale; Assume that the number of potential users is N and the number of bits required to identify users is B i , satisfying the relationship Given N,B i Determined by the following formula: in, Indicates rounding up; assign user identity (ID) to 0 to N numbers in, and then convert ID into a number containing B i The binary form of a bit number; Step (2), based on the number of bits required for source coding determined in step (1), cyclic redundancy check and FEC coding scheme are used for channel coding, and the FEC-coded codeword is output; Step (3), the codeword after FEC encoding is passed through the interleaving memory to combat deep channel fading; Step (4), sparse modulation; establishing a mapping relationship from codewords to sparse modulation symbols, and then mapping the codewords obtained in step (3) to a leading sequence through sparse modulation; making the sequence satisfy the Bernoulli characteristic and have sparse features; In step (2), after the source code is encoded, it is first selected whether to add an error detection code; the error detection method used is a cyclic redundancy check (CRC), and the number of bits contained in the error detection code is B c (B c ≥0), then the code length after cyclic redundancy check is (B i +B c ), in order to reduce the bit error rate, an FEC coding scheme is adopted and an FEC-encoded codeword is output; In step (2), the specific steps of FEC encoding are as follows: Step (2.1) determines the generating polynomial; Step (2.1.1) Let the total code length after FEC encoding be n, and based on this, select a reduced polynomial of degree m over the finite field GF(2), and construct the extended field GF(2) of GF(2) m ), m and n satisfy the relationship n=2 m -1; Step (2.1.2) Let α be GF(2 m ) primitive element, if the number of error bits to be corrected is t, then the generating polynomial g(x) of the BCH code is α 1 ,α 2 ,…,α 2t The lowest degree polynomial on GF(2) with roots; let m i (x) is α i The minimum polynomial of , then: g(x)=LCM[m1(x),m2(x),…,m 2t (x)] Among them, LCM[·] means taking the least common multiple; since α 2i and α i The minimal polynomial is the same, so we have: g(x)=LCM[m1(x),m3(x),…,m 2t-1 (x)] Step (2.2) determines the parameters of the BCH code: The code length after FEC encoding in step (2.2.1) is n=2 m -1; Step (2.2.2) Let k = (B i +B c ), the number of parity bits required is r = nk ≤ mt; Step (2.2.3) Minimum distance d between codewords min ≥2t+1; The number of bits corrected in step (2.2.4) is t; Step (2.3) obtains the FEC codeword: Step (2.3.1) Let the information code polynomial after source coding and cyclic redundancy check be q(x), and divide x by g(x) (n-k) q(x), we get the quotient S(x) and the remainder r(x), namely: Output code group c(x) = x (n-k) q(x)+r(x); Step (2.3.2) outputs the FEC-encoded codeword according to the code group obtained in step (2.3.1); The mapping relationship between codewords and sparse modulation symbols is established in step (4), and the codewords obtained in step (3) are mapped to the preamble sequence through sparse modulation. The specific steps are: Step (4.1) determines the pattern of the sparse index modulation signal Let the preamble sequence after sparse index modulation be Where L is the symbol duration of the sequence, x i Represents the i-th element of x, 1≤i≤L, and defines the signal x that satisfies the following three conditions as a sparse index modulation signal: Step (4.1.1) divides x into Q equal segments, and the length of each segment is Segments are independent and identically distributed; Step (4.1.2) specifies that each segment has exactly one non-zero entry, namely: in, represents the indicator function, the positions of non-zero entries are random; In step (4.1.3), for the qth segment of x, let the lth element be a non-zero item, then: in, Represents the set of constellation points, Step (4.2) determines the number of bits that the preamble sequence can carry: The number of bits that the preamble sequence can carry is denoted as B p , expressed as: Step (4.3) simplifies the mapping relationship: Based on the preamble sequence x, we only need to construct a mapping between signal segments and specific bit vectors, which greatly compresses the space before and after the mapping. Since the non-zero items in the first segment of x are reference symbols, the only information that actually carries is the position of the non-zero items. Therefore, the mapping relationship between the first segment and other segments is considered separately: Step (4.3.1) determines the first segment mapping relationship: There are L non-zero positions in this segment. s In this case, the number of bits carried is log2 L s , construct log2 L s bits of Gray code, corresponding to L s This situation requires that the Hamming distance of the codes corresponding to adjacent non-zero items is the smallest, which is only 1. Step (4.3.2) determines the mapping relationship of other segments: make That is QPSK, let represents the i-th segment of x, x q The number of bits that can be carried B s , so it constitutes x q The set of all cases has vector, denoted as It can be proved that have Where ‖·‖2 represents the Euclidean distance, and the non-zero entries of the two vectors with the longest distance have the same position but opposite polarity. Therefore, when constructing the mapping relationship, the Hamming distance of the codeword vectors corresponding to the two segments with the longest distance must also be the largest. The specific construction is as follows: The number of bits to be constructed is B s =log2 L s +2 standard Gray code, and divided into 4 equal parts in sequence; the code words of the first Gray code are matched one-to-one with all possible segments with non-zero items of 1, and a mapping relationship is established; the code words of the first Gray code are inverted, and the code words after the operation are matched one-to-one with all possible segments with non-zero items of -1, and a mapping relationship is established; the code word set after the inversion is equal to the code word set composed of the third Gray code; the code words of the fourth Gray code are matched one-to-one with all possible segments with non-zero items of -i; the code words of the fourth Gray code are inverted, and the code words after the operation are matched one-to-one with all possible segments with non-zero items of i, and a mapping relationship is established; the code word set after the inversion is equal to the code word set composed of the second Gray code; note that the mapping requires that the Hamming distance of the codes corresponding to adjacent non-zero items is the minimum, which is only 1; Step (4.4) completes the codeword vector c∈{0,1} n To the leader sequence Mapping; Specify the codeword output after the interleaving memory as c∈{0,1} n , the 1st to the log2 L of c s bits are mapped according to the first mapping relationship in step (4.3.1), and the remaining n-log2 L s After the bits are divided into Q-1 equal parts, they are mapped according to the mapping relationship of other segments in step (4.3.3) to complete the mapping of codeword c to sequence x.

2. The leader sequence design method according to claim 1, wherein: In step (4), coordinate the FEC encoding parameters and the preamble sequence construction conditions so that B p =n.

Images