Methods and systems for constructing low-correlation pilot sequences for satellite internet

Abstract

This invention discloses a method, system, and apparatus for constructing low-correlation pilot sequences for satellite internet. The method includes: pilot detection using a zero-correlation domain shifted superposition pilot sequence set, and channel estimation; decoding using the SCJD algorithm to recover the remaining k-rb bits corresponding to no more than T conflicting terminals in each frame, and recovering the first rb bits of each UE using the pilot sequence obtained from pilot detection; and the satellite broadcasting a decoding status feedback message to all active users. This invention reduces the probability of pilot collisions, thereby achieving a lower access failure probability. Furthermore, this invention improves communication spectrum efficiency through the design of a data transmission protocol.

Description

Technical Field

[0001] This invention relates to the field of massive machine communication technology for satellite internet, and in particular to a method, system, and apparatus for constructing low-correlation pilot sequences for massive access services in satellite internet. Background Technology

[0002] With the rollout of fifth-generation (5G) communication networks and the upcoming development of sixth-generation (6G) communication networks, satellite internet, capable of supporting ubiquitous access for massive numbers of user equipment (UEs), has become one of the major new infrastructure construction projects with national strategic needs. According to the International Mobile Telecommunications 2030 (IMT-2030) plan, 6G networks require satellites to support ultra-high-density mMTC-s service scenarios with 1 million UEs per square kilometer. Therefore, the key challenges that need to be addressed in mMTC-s scenarios include: how to effectively resolve pilot conflicts between UEs, reduce AFP (Advanced Pilot Frequency), and how to achieve high-spectrum-efficiency transmission of short packet communication under limited spectrum resources.

[0003] Faced with the challenges of pilot collisions and communication spectrum efficiency in massive connections, existing non-orthogonal pilot schemes, while improving pilot collision performance, suffer from low system reliability due to limited pilot resource expansion or high average cross-correlation values ​​of pilot sequences. Furthermore, they are insufficient to handle the explosive growth in the number of terminals because they cannot recover data from terminals that have already experienced collisions. Therefore, research is urgently needed to develop system solutions that address the reliability limitations of existing traditional communication solutions and meet the needs of large-scale access service scenarios. Summary of the Invention

[0004] The main objective of this invention is to propose a method, system, and apparatus for constructing low-correlation pilot sequences for satellite internet. This invention addresses the core challenges of massive device access scenarios in massive machine-type communications (mMTC-s) satellite internet, including pilot collisions, low spectrum efficiency of short packet communication, and high reliability requirements. It aims to provide large-scale access services for a massive number of users in mMTC-s scenarios and reduces the probability of pilot collisions (PCP) by expanding the size of the pilot sequence set, thereby achieving a lower access failure probability (AFP). Furthermore, it improves communication spectrum efficiency through the design of data transmission protocols.

[0005] To achieve the above objectives, the present invention provides a method for constructing low-correlation pilot sequences for the Internet, the method comprising the following steps:

[0006] Step S100: After the receiving satellite receives the concatenated information sent by the transmitting end, it performs pilot detection and channel estimation using a set of zero-correlation domain shift-superimposed pilot sequences, wherein the least squares algorithm is applied to obtain the channel response; wherein the concatenated information is obtained by each active UE at the transmitting end dividing a k-bit information into r+1 blocks, wherein the first r blocks are all b bits long, each block is mapped to a zero-correlation domain sequence according to b bits, and these r sequences are superimposed to generate a zero-correlation domain shift-superimposed pilot sequence. Subsequently, the remaining k-rb bits are encoded using a T-fold codebook of order T and concatenated with the zero-correlation domain shift-superimposed pilot sequence to obtain the result.

[0007] Step S200: Decode the remaining k-rb bits of no more than T conflicting terminals in each frame using the SCJD algorithm, and recover the first rb bits of each UE using the pilot sequence obtained from pilot detection;

[0008] In step S300, the satellite broadcasts a decoding status feedback message to all active users.

[0009] A further technical solution of the present invention is that, prior to step S100, the following is included:

[0010] Step S000: Construct a system model for a code domain unlicensed large-scale access protocol based on satellite internet.

[0011] A further technical solution of the present invention is that step S000 includes:

[0012] Step S001: In a communication system covered by a single sub-beam, satellite Sa provides mMTC-s service to terminals within its beam coverage area. It is assumed that these terminals are quasi-stationary and uniformly distributed within the coverage area. It is approximately assumed that all terminals are equidistant from Sa, thus achieving almost simultaneous signal arrival times. It is assumed that the duration of each frame matches the duration of a time slot, denoted as n, where propagation delay and transceiver processing delay are both included within a random access time slot. Considering the maximum bidirectional propagation delay is approximately 26 milliseconds, the duration of the random access time slot is set to 50 milliseconds. The mMTC-s scenario allows for a potentially unlimited number of terminal accesses, but in each random access time slot, only a small subset of terminals are activated with probability pa = 0.01 and communicate with the satellite; this subset of activated terminals is denoted as K.

[0013] A shadowed Ricean fading channel model is used to simulate a real wireless propagation environment. The probability density function of the shadowed Ricean fading channel model is:

[0014] f(h 2 )=(2dn / (2dn+Ω) n (1 / 2d)exp(-h 2 / 2d)1F1(n,1,Ωh 2 / 2d(2dn+Ω));

[0015] Where d represents the multipath component, Ω represents the average power at line of sight, n represents the Nakagami-n parameter, and 1F1(i,j,k) represents the confluence hypergeometry function. By setting the parameters d = 0.063, n = 1, and Ω = 0.000897, the above equation can be simplified to:

[0016] f(r) = we -ηr r>o;

[0017] In the formula

[0018] Where w and η are the coefficients of the exponential distribution, and r = h 2 As the fading coefficient of SR, P T It is the user power, σ 2 It is the noise variance.

[0019] A further technical solution of the present invention is that the specific description of the uplink code field unlicensed massive access protocol in step S001 is as follows:

[0020] First, set the zero-correlation domain shift superposition pilot set as Where L is the pilot length of the zero-correlation domain displacement superimposed pilot, and S represents the size of the pilot set of the zero-correlation domain displacement superimposed pilot; the terminal activated at the UE uses its first r·b bits of information to select the pilot, and then performs multi-level encoding operations on the remaining information. The encoded signal is then modulated and spread spectrum processed before being sent to the satellite.

[0021] Let X u Y represents the codeword after the UE information has been encoded and modulated using a T-order codebook. p and Y d S represents a The received pilot signals and data information are specifically represented as follows:

[0022]

[0023] as well as

[0024]

[0025] Where j and u represent different user indices, P t It is the user power, P Uj It is an LPS pilot, Y d This represents the received coded signal, S represents the pilot set size, and φ represents the received coded signal. j This represents the set of terminals whose first r parts are all identical; that is, these terminals select the same zero-correlation domain periodic sequence for superposition in their first r parts, and correspond to the same zero-correlation domain shifted superposition pilot sequence P in the received frame. Uj , Representing the Kronecker product, W is additive white Gaussian noise, and is CN(0, σ). 2 ), h u Represents the SR coefficient of UEu;

[0026] Satellite S a Based on the received Y p The data is subjected to pilot detection, and the first r·b bits of the received frame are decoded. Then, the selected P... Uj The terminal performs channel estimation;

[0027] The channel weighted sum estimate of the terminal is expressed as follows:

[0028]

[0029] This allows us to obtain the selected pilot P. Uj The remaining portion of the received data Y in the transmission frame d The encoded message in the code is represented as follows:

[0030]

[0031] Where u represents different users, h u X represents the channel fading coefficient for different users. u P represents the transmitted user-coded data. t It is the user's power. This represents the received coded signal, and S represents the pilot set size;

[0032] Subsequently, S a The remaining portion of the transmitted frame is decoded using SCJD. SCJD first iteratively performs serial cancellation decoding, where the residual signal obtained after the i-th step of serial cancellation decoding is as follows:

[0033]

[0034] Where q represents different users, h q X represents the channel fading coefficient for different users. q P represents the transmitted user-coded data. t It is the user's power. This represents the received coded signal, and S represents the pilot set size;

[0035] However, if the serial decoding in step i fails, joint decoding is initiated. Joint decoding can recover at most Ti conflicting terminals. The input signals for joint decoding are as follows:

[0036]

[0037] Where δ is the residual interference coefficient after decoding; T is the multi-order coding order; v represents different active users; h v P represents the channel fading coefficient for different users. t It is the user's power;

[0038] set up Exhaustive search of residual signals All possible combinations, and compare each estimate with the actual residual signal. Compare to obtain the difference g e Then it is compared with the threshold θ as follows:

[0039]

[0040] Where I represents the error between the actual received signal and the estimated signal, h l This represents the channel fading coefficient for different users. V represents an estimate of I. T-i This indicates that it belongs to the residual signal. The code X u Set, threshold θ = (N) c -L)σ 2 , where Nc The length of the T-order codeword is represented by continuous iteration until g. e If the value drops below θ, the estimation is considered successful.

[0041] Next, iterative joint decoding is performed, including decoding of both internal and external codes, to recover all remaining messages from up to Ti terminals; however, if after exhausting all Ti attempts, g e >θ, meaning that a valid set V cannot be found in the set. T-i S a The declaration indicates that the joint decoding failed and is discarded.

[0042] A further technical solution of the present invention is that the method further includes: designing zero-correlation domain shift-superimposed pilots: a random combination superposition method is introduced, that is, r sequences are randomly selected from the set of zero-correlation domain periodic sequences, and then superimposed, with the pilot length remaining unchanged, and the corresponding elements are summed to obtain the zero-correlation domain shift-superimposed pilot sequence P. U :

[0043]

[0044] Where p v It is an LPS sequence, where i, j, ..., v represent different sequence indices, and Mτ z Indicates the size of the set of zero-correlation domain sequences;

[0045] By iterating through all r random combinations of sequences, a set of zero-correlation domain shift-superimposed pilot sequences is obtained; based on the construction process of the zero-correlation domain shift-superimposed pilot sequence set, the size of the zero-correlation domain shift-superimposed pilot sequence set is... Where Z = M·τ z The value represents the size of the set of zero-correlation domain sequences, M represents the number of root LPS sequences, and τ represents the value of the set of zero-correlation domain sequences. z This represents the number of bits that the zero-correlation field sequence can be shifted.

[0046] A further technical solution of the present invention is that the method further includes: correlation analysis of zero-correlation domain shift superimposed pilots:

[0047] set up Represents the pilot sequence of the zero correlation domain and The cross-correlation function between them, where τ u , τ v This represents the different shift values ​​corresponding to different sequences, where the shift difference between two sequences is Δτ. u,v =|τ u -τ v | bits, where Δτ u，v Let the shift difference between the two sequences be represented by... The zero-correlation domain displacement superposition pilot P Uk and P Uq The cross-correlation function between them, where P Uk P Uq This represents different user indexes, and their index sets are: and in To represent different LPS sequences, τ a , τ b , τ c , τ d Let k1, k2, q1, and q2 represent different shift values, and r = 2. Therefore, the cross-correlation of any two zero-correlation domain shifts superimposed with pilots is as follows:

[0048]

[0049] in Let Δτ represent the cross-correlation value between any two ZSPs. a，c ,Δτ a，d ,Δτ b，c ,Δτ b，d Indicates different shift values, To represent different LPS sequences, τ a , τ b , τ c , τ d These represent different shift values, and k1, k2, q1, q2 represent different root sequences;

[0050] If ZSP P is constituted Uk and P Uq If the zero-correlation domain sequences are different, then the cross-correlation value between two ZSPs is zero; probability statistics show that the number of ZSPs composed of completely different sequences is... Where Z = M·τ z The value represents the size of the set of zero-correlation domain sequences, M represents the number of root LPS sequences, and τ represents the value of the set of zero-correlation domain sequences. z Let r represent the number of shiftable bits of the zero-correlation domain sequence, and r represent the superposition coefficient; therefore, the shifted superposition pilots of other zero-correlation domains in this set remain consistent with P. Uk The probabilities of orthogonality are as follows:

[0051]

[0052] Where Z = M·τ z The value represents the size of the set of zero-correlation domain sequences, M represents the number of root LPS sequences, and τ represents the value of the set of zero-correlation domain sequences. z The number of bits that can be shifted in the zero-correlation domain sequence is represented by r, which represents the superposition coefficient.

[0053] For r << (τ) z ·M), PO →1, meaning that most of the zero-correlation domain shift superposition pilots remain orthogonal, with only a small portion exhibiting non-orthogonal characteristics; let... Represents the pilot sequence of the zero correlation domain The autocorrelation function, then The average autocorrelation is shown below:

[0054]

[0055] Where L represents the sequence length, M represents the number of root LPS sequences, and N represents the perfect sequence length;

[0056] Analysis of the superposition of two zero-correlation domain displacement pilots P Uk and P Uq The cross-correlation value needs to consider three different cases:

[0057] Case 1: The pilot signals of the two zero-correlation domain shift superpositions have the same combination of zero-correlation domain sequences;

[0058] Case 2: Two zero-correlation domain shifts are superimposed, and pilots share d zero-correlation domain sequences;

[0059] Case 3: The zero-correlation domain sequences of the two zero-correlation domain shift-superimposed pilots do not overlap at all; since the two zero-correlation domain shift-superimposed pilots are orthogonal in Case 3, we mainly focus on the first two cases; therefore, the probability that the two zero-correlation domain shift-superimposed pilots share d zero-correlation domain sequences is given by the following formula:

[0060]

[0061] Where Z represents the size of the set of zero-correlation domain sequences, r represents the superposition coefficient, and d represents that the two ZSPs share d zero-correlation domain sequences;

[0062] Therefore P Uk and P Uq The average cross-correlation value is derived as follows:

[0063]

[0064] in denoted by , where L represents the sequence length, N represents the perfect sequence length, V′ represents the number of columns in the Hadamard matrix recombination, and r represents the superposition coefficient.

[0065] Therefore, when r = 2, the average cross-correlation value between any two zero-correlation domain shift superimposed pilots is O(1 / L); furthermore, if r increases, the average cross-correlation value will also increase.

[0066] To achieve the above objectives, the present invention also proposes a low-correlation pilot sequence construction system for the Internet, the system comprising a memory, a processor, and a low-correlation pilot sequence construction program for the Internet stored on the processor, the low-correlation pilot sequence construction program for the Internet being executed by the processor to perform the steps of the method described above.

[0067] To achieve the above objectives, the present invention also proposes a computer-readable storage device storing a low-correlation pilot sequence construction program for the Internet, wherein the low-correlation pilot sequence construction program for the Internet is executed by a processor to perform the steps of the method described above.

[0068] The beneficial effects of the low-correlation pilot sequence construction method, system, and device for satellite internet of the present invention are:

[0069] 1. This invention addresses the problem of massive device access in mMTC-s communication scenarios under satellite internet by proposing a ZT-Collision Resolution Grant-Free Random Access (ZT-GFRA) scheme; ZT: pilot sequence with low cross-correlation value in the zero-correlation domain + T-order high-dimensional universal codebook. The scheme improves the system's spectral efficiency by designing an efficient and reliable pilot allocation strategy. This scheme not only adapts to the needs of massive device access but also achieves lower PCP, decoding failure probability (DFP), and AFP in the short-packet communication mMTC-s scenario, providing an important reference for addressing the high-density access challenges of future satellite internet.

[0070] 2. This invention designs a ZSP sequence set as the pilot set for the ZT-GFRA scheme. Utilizing the idea of ​​random combination and superposition, based on the zero correlation domain sequence, a ZSP sequence with a larger pilot set is generated. The scale of its pilot set far exceeds the pilot length, which can effectively reduce PCP and make it more suitable for future high-density massive device access scenarios.

[0071] 3. This invention analyzes the average cross-correlation characteristics of the ZSP sequence, derives an expression for the average cross-correlation value with respect to the superposition coefficient r, and thus provides the optimal value of the superposition coefficient r for the ZSP sequence. It also derives that this pilot sequence has an average cross-correlation value close to 0, which can significantly reduce interference between terminals. Monte Carlo simulations verify that the ZSP exhibits superior performance in shadow-Rice fading channels. Attached Figure Description

[0072] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0073] Figure 1 This is a flowchart illustrating a preferred embodiment of the method for constructing low-correlation pilot sequences for satellite internet according to the present invention.

[0074] Figure 2 This is a simulation diagram of the AFP performance of the ZT-GFRA scheme based on ZSP with respect to r and SNR, where K = 1000 and T = 2;

[0075] Figure 3 The simulation of the PCP of ZSP and other pilots with respect to the number of active UEs K and pilot length, where r = 2 and pilot length L = 168 / 167 or L = 336 / 337;

[0076] Figure 4 The simulation is of the DFP of ZSP and other pilots with respect to the number of active UEs K and pilot length, where r = 2 and pilot length L = 168 / 167 or L = 336 / 337;

[0077] Figure 5 The simulation of the ZT-GFRA scheme with respect to the number of active UEs K and pilot length in AFP is compared with other schemes, where r = 2 and pilot length L = 168 or L = 336.

[0078] Figure 6 The simulation of the ZT-GFRA scheme with respect to SNR in AFP compared to other schemes, where r = 2 and pilot length L = 168 or L = 336;

[0079] Figure 7 The simulation of the ZT-GFRA scheme with respect to the number of active UEs K and pilot length in AFP is compared with other schemes, where r = 2 and pilot length L = 28 / 29;

[0080] Figure 8 The simulation of the ZT-GFRA scheme with respect to SNR in AFP compared to other schemes is shown, where r = 2 and pilot length L = 28 / 29.

[0081] Figure 9 This is a performance simulation analysis of different pilot combined with high-order codebook coding schemes, where r = 2 and pilot length L = 168 or L = 336;

[0082] Figure 10This is a schematic diagram of a satellite internet communication scenario;

[0083] Figure 11 This is a schematic diagram of a method for constructing low-correlation sequences based on block shifting of zero-correlation domain sequences;

[0084] Figure 12 This is a hardware architecture diagram of the low-correlation pilot sequence construction system for satellite internet of the present invention.

[0085] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0086] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0087] Pilot collisions are a critical issue that urgently needs to be addressed in Grant-Free Random Access (GFRA) schemes. Due to the significant path loss in satellite-to-ground links, the satellite end struggles to effectively distinguish terminals that have selected the same pilot signal through power control. Therefore, an intuitive solution is to increase pilot resources and expand the pilot set to minimize the probability of pilot collisions, or even completely avoid them. Furthermore, in mMTC-s scenarios, terminal message lengths are typically short. As terminal density increases, simply expanding the pilot set by increasing pilot length will lead to a significant decrease in spectral efficiency. Therefore, it is necessary to expand the size of the pilot set without significantly increasing pilot length.

[0088] Traditional pilot sequences are typically designed with mutual orthogonality as the principle, such as Kasami codes and Zadoff-Chu sequences (ZCSs). However, the size of these orthogonal pilot sets is limited by the pilot length, leading to severe pilot collisions in large-scale access scenarios. To address this issue, researchers have begun exploring the design of non-orthogonal pilot sequences to expand the number of pilot resources and meet the challenges of large-scale access scenarios. Related literature proposes a non-orthogonal sequence design that can significantly reduce the probability of pilot collisions. This scheme fully utilizes the potential of ZCSs, generating pilot sequences using all available root ZCSs and integrating them into a single pilot set. Since a ZCS of length N can have at most N-1 root sequences, the multi-root ZCS (mZCS) pilot set constructed by this method can reach a size of N(N-1), which is N-1 times larger than orthogonal pilot sets, while retaining the excellent characteristics of ZCSs. Building upon this, some literature has further introduced the structure of m-sequences into the ZCS, increasing the maximum cross-correlation value of the sequence set and adding more non-orthogonal sequences, thus expanding the number of pilots to N(N+1). However, since the cross-correlation value within the same root sequence in the mZCS pilot set is always zero, while the cross-correlation value between different root sequences is high, this design inevitably leads to interference between most terminals, maintaining orthogonality only in a very few cases. This makes the mZCS relatively vulnerable to channel noise in GFRA systems, ultimately resulting in a high system AFP.

[0089] To address user interference caused by non-orthogonal pilots, recent literature has proposed using the average cross-correlation value instead of the maximum cross-correlation value as a key indicator for measuring the correlation of non-orthogonal pilot sets in large-scale access systems. Furthermore, low-correlation zone periodic sequences (LPS), distinct from traditional Gaussian, Gold, and mZCS sequences, have been designed. LPS maintains low cross-correlation even with random non-orthogonal shifts, helping to reduce terminal conflicts. However, compared to orthogonal pilots of the same length, LPS only expands the pilot set by approximately M times (where M represents the order of the Hadamard matrix used to construct the LPS), which is insufficient to cope with the explosive growth in the number of terminals. Therefore, it is necessary to design pilot sequences more suitable for massive communication service scenarios, capable of further expanding the available pilot set while maintaining low average cross-correlation characteristics, thereby reducing PCP in mMTC-s.

[0090] On the other hand, in satellite internet scenarios, code domain GFRA is becoming an important research direction for addressing the access needs of massive user devices. Due to the wide-area coverage and unique physical characteristics of satellite systems, random access in mMTC-s in satellite internet faces severe challenges, including severe pilot collisions, low spectrum resource utilization efficiency, and high AFP (Area of ​​Presence). Traditional ALOHA-like protocols based on time slots and frequency domains are difficult to apply directly in satellite scenarios because these protocols are prone to severe resource conflicts and system performance degradation when faced with extremely high user density. Code domain GFRA, by utilizing codeword resources shared by multiple users, achieves efficient resource allocation and conflict management during the access process, becoming one of the current research hotspots. For code domain GFRA, Polyanskiy et al. proposed a T-fold random access scheme based on multi-order codebooks in their literature. This scheme addresses the shortcomings of ALOHA-like protocols in collision management by optimizing codeword design. However, this method still has bottlenecks in large-scale access systems. To better suit mMTC-s services, some literature has proposed a pilot-based architecture called the LT-order Collision Resolution Code Domain GFRA (LT-GFRA) scheme. This scheme combines a ZCS sequence set of length L with a T-order codebook, theoretically resolving collisions of up to LT terminals. However, the LT-GFRA scheme is limited by the ratio between the pilot sequence length and the number of terminals, and the codebook complexity of T-fold related schemes increases significantly with the increase of T. This indicates that the above two schemes have inherent limitations in practical applications.

[0091] Therefore, this invention designs a data transmission protocol based on non-orthogonal pilot sequences and uplink code domain framework for the mMTC-s service scenario under wide-area coverage of satellite Internet, which meets the requirements of high reliability and high data transmission efficiency under massive device access.

[0092] Specifically, this invention proposes a method for constructing low-correlation pilot sequences for the Internet. The scheme involves each active UE dividing a k-bit information block into r+1 blocks, where the first r blocks are each b bits long. Each block is mapped to a zero-correlation zone sequence based on its b bits, and these r sequences are superimposed to generate a zero-correlation-zone shift-and-superposition pilot (ZSP) sequence. Subsequently, the remaining k-rb bits are encoded using a T-order T-fold codebook, concatenated with the ZSP, and transmitted to the satellite.

[0093] At the receiving end, the satellite uses the ZSP set for pilot detection and channel estimation, employing the Least-Squares (LS) algorithm to obtain the channel response. The SCJD algorithm is used to decode and recover the remaining k-rb bits corresponding to no more than T conflicting terminals in each frame. Furthermore, the first rb bits of each UE are recovered from the pilot sequence obtained from pilot detection. Finally, the satellite broadcasts a decoding status feedback message to all active users.

[0094] Please refer to Figure 1 A preferred embodiment of the method for constructing low-correlation pilot sequences for the Internet according to the present invention includes the following steps:

[0095] Step S100: After the receiving satellite receives the concatenated information sent by the transmitting end, it performs pilot detection and channel estimation using a set of zero-correlation domain shifted superimposed pilot sequences, wherein the least squares algorithm is applied to obtain the channel response. The concatenated information is obtained by each active UE at the transmitting end dividing a k-bit information into r+1 blocks, wherein the first r blocks are all b bits long. Each block is mapped to a zero-correlation domain sequence based on b bits, and these r sequences are superimposed to generate a zero-correlation domain shifted superimposed pilot sequence. Subsequently, the remaining k-rb bits are encoded using a T-fold codebook of order T and concatenated with the zero-correlation domain shifted superimposed pilot sequence.

[0096] Step S200: The remaining k-rb bits of each frame corresponding to no more than T conflicting terminals are recovered by decoding using the SCJD algorithm, and the first rb bits of each UE are recovered by using the pilot sequence obtained by pilot detection.

[0097] In step S300, the satellite broadcasts a decoding status feedback message to all active users.

[0098] The ZSP sequence set designed in this invention can significantly expand the size of the usable pilot set without increasing the pilot length, effectively reducing PCP. Furthermore, ZSP possesses excellent low cross-correlation characteristics, giving it superior noise immunity performance compared to existing non-orthogonal pilots in practical noisy channel transmission. See the detailed flowchart below. Figure 11 Meanwhile, by resolving UE conflicts of order no higher than T from a code domain perspective, the number of UEs the system can accommodate is greatly increased. Furthermore, by utilizing pilot sequences for multiple purposes, enabling them to both estimate channel state information and carry user data, spectral efficiency is further improved, meeting the needs of short packet communication services.

[0099] In this embodiment, the steps preceding step S100 include:

[0100] Step S000: Construct a system model for a code domain unlicensed large-scale access protocol based on satellite internet.

[0101] Step S000 specifically includes:

[0102] Step S001: In a communication system covered by a single sub-beam, satellite Sa provides mMTC-s service to terminals within its beam coverage area. It is assumed that these terminals are quasi-stationary and uniformly distributed within the coverage area. It is approximately assumed that all terminals are equidistant from Sa, thus achieving almost simultaneous signal arrival times. It is assumed that the duration of each frame matches the duration of a time slot, denoted as n, where propagation delay and transceiver processing delay are both included within a random access time slot. Considering the maximum bidirectional propagation delay is approximately 26 milliseconds, the duration of the random access time slot is set to 50 milliseconds. The mMTC-s scenario allows for a potentially unlimited number of terminal accesses, but in each random access time slot, only a small subset of terminals are activated with probability pa = 0.01 and communicate with the satellite; this subset of activated terminals is denoted as K.

[0103] A shadowed Ricean fading channel model is used to simulate a real wireless propagation environment. The probability density function of the shadowed Ricean fading channel model is:

[0104] f(h 2 )=(2dn / (2dn+Ω) n (1 / 2d)exp(-h 2 / 2d)1F1(n,1,Ωh 2 / 2d(2dn+Ω));

[0105] Where d represents the multipath component, Ω represents the average power at line of sight, n represents the Nakagami-n parameter, and 1F1(i,j,k) represents the confluence hypergeometry function. By setting the parameters d = 0.063, n = 1, and Ω = 0.000897, the above equation can be simplified to:

[0106] f(r) = we -ηr r>o;

[0107] In the formula

[0108] Where w and η are the coefficients of the exponential distribution, and r = h 2 As the fading coefficient of SR, P T It is the user power, σ 2 It is the noise variance.

[0109] In this embodiment, the specific description of the uplink code field unlicensed large-scale access protocol in step S001 is as follows:

[0110] First, set the zero-correlation domain shift superposition pilot set as Where L is the pilot length of the zero-correlation domain displacement superimposed pilot, and S represents the size of the pilot set of the zero-correlation domain displacement superimposed pilot; the terminal activated at the UE uses its first r·b bits of information to select the pilot, and then performs multi-level encoding operations on the remaining information. The encoded signal is then modulated and spread spectrum processed before being sent to the satellite.

[0111] Let X u Y represents the codeword after the UE information has been encoded and modulated using a T-order codebook. p and Y d S represents a The received pilot signals and data information are specifically represented as follows:

[0112]

[0113] as well as

[0114]

[0115] Where j and u represent different user indexes, P t It is the user power, P Uj It is an LPS pilot, Y d This represents the received coded signal, S represents the pilot set size, and φ represents the received coded signal. j This represents the set of terminals whose first r parts are all identical; that is, these terminals select the same zero-correlation domain periodic sequence for superposition in their first r parts, and correspond to the same zero-correlation domain shifted superposition pilot sequence P in the received frame. Uj , Representing the Kronecker product, W is additive white Gaussian noise, and is CN(0, σ). 2 ), h u Represents the SR coefficient of UEu;

[0116] Satellite S a Based on the received Y p The data is subjected to pilot detection, and the first r·b bits of the received frame are decoded. Then, the selected P... Uj The terminal performs channel estimation;

[0117] The channel weighted sum estimate of the terminal is expressed as follows:

[0118]

[0119] This allows us to obtain the selected pilot P. Uj The remaining portion of the received data Y in the transmission frame d The encoded message in the code is represented as follows:

[0120]

[0121] Where u represents different users, h u X represents the channel fading coefficient for different users. u P represents the transmitted user-coded data. t It is the user's power. This represents the received coded signal, and S represents the pilot set size;

[0122] Subsequently, S a The remaining portion of the transmitted frame is decoded using SCJD. SCJD first iteratively performs serial cancellation decoding, where the residual signal obtained after the i-th step of serial cancellation decoding is as follows:

[0123]

[0124] Where q represents different users, h q X represents the channel fading coefficient for different users. q P represents the transmitted user-coded data. t It is the user's power. This represents the received coded signal, and S represents the pilot set size;

[0125] However, if the serial decoding in step i fails, joint decoding is initiated. Joint decoding can recover at most Ti conflicting terminals. The input signals for joint decoding are as follows:

[0126]

[0127] Where δ is the residual interference coefficient after decoding; T is the multi-order coding order; v represents different active users; h v P represents the channel fading coefficient for different users. t It is the user's power;

[0128] set up Exhaustive search of residual signals All possible combinations, and compare each estimate with the actual residual signal. Compare to obtain the difference g e Then it is compared with the threshold θ as follows:

[0129]

[0130] Where I represents the error between the actual received signal and the estimated signal, h l This represents the channel fading coefficient for different users. V represents an estimate of I. T-i This indicates that it belongs to the residual signal. The code X u Set, threshold θ = (N) c -L)σ 2 , where Nc The length of the T-order codeword is represented by continuous iteration until g. e If the value drops below θ, the estimation is considered successful.

[0131] Next, iterative joint decoding is performed, including decoding of both internal and external codes, to recover all remaining messages from up to Ti terminals; however, if after exhausting all Ti attempts, g e >θ, meaning that a valid set V cannot be found in the set. T-i S a The declaration indicates that the joint decoding failed and is discarded.

[0132] In this embodiment, the method further includes: designing zero-correlation domain displacement superimposed pilots: a random combination superposition method is introduced, that is, r sequences are randomly selected from the set of zero-correlation domain periodic sequences, and then superimposed. The pilot length remains unchanged, and the corresponding elements are summed to obtain the zero-correlation domain displacement superimposed pilot sequence P. U :

[0133]

[0134] Where p v It is an LPS sequence, where i, j, ..., v represent different sequence indices, and Mτ z Indicates the size of the set of zero-correlation domain sequences;

[0135] By iterating through all r random combinations of sequences, a set of zero-correlation domain shift-superimposed pilot sequences is obtained; based on the construction process of the zero-correlation domain shift-superimposed pilot sequence set, the size of the zero-correlation domain shift-superimposed pilot sequence set is... Where Z = M·τ z The value represents the size of the set of zero-correlation domain sequences, M represents the number of root LPS sequences, and τ represents the value of the set of zero-correlation domain sequences. z This represents the number of bits that the zero-correlation field sequence can be shifted.

[0136] In this embodiment, the method further includes: correlation analysis of zero-correlation domain displacement superimposed pilots:

[0137] set up Represents the pilot sequence of the zero correlation domain and The cross-correlation function between them, where τ u , τ v This represents the different shift values ​​corresponding to different sequences, where the shift difference between two sequences is Δτ. u，v =|τ u -τ v |bits, where Δτ u，v Let the shift difference between the two sequences be represented by... The zero-correlation domain displacement superposition pilot P Ukand P Uq The cross-correlation function between them, where P Uk P Uq This represents different user indexes, and their index sets are: and in To represent different LPS sequences, τ a , τ b , τ c , τ d Let k1, k2, q1, and q2 represent different shift values, and r = 2. Therefore, the cross-correlation of any two zero-correlation domain shifts superimposed with pilots is as follows:

[0138]

[0139] in Let Δτ represent the cross-correlation value between any two ZSPs. a，c ,Δτ a，d ,Δτ b，c ,Δτ b，d Indicates different shift values, To represent different LPS sequences, τ a , τ b , τ c , τ d These represent different shift values, and k1, k2, q1, q2 represent different root sequences;

[0140] If ZSP P is constituted Uk and P Uq If the zero-correlation domain sequences are different, then the cross-correlation value between two ZSPs is zero; probability statistics show that the number of ZSPs composed of completely different sequences is... Where Z = M·τ z The value represents the size of the set of zero-correlation domain sequences, M represents the number of root LPS sequences, and τ represents the value of the set of zero-correlation domain sequences. z Let r represent the number of shiftable bits of the zero-correlation domain sequence, and r represent the superposition coefficient; therefore, the shifted superposition pilots of other zero-correlation domains in this set remain consistent with P. Uk The probabilities of orthogonality are as follows:

[0141]

[0142] Where Z = M·τ z The value represents the size of the set of zero-correlation domain sequences, M represents the number of root LPS sequences, and τ represents the value of the set of zero-correlation domain sequences. z The number of bits that can be shifted in the zero-correlation domain sequence is represented by r, which represents the superposition coefficient.

[0143] For r << (τ) z ·M), P O→1, meaning that most of the zero-correlation domain shift superposition pilots remain orthogonal, with only a small portion exhibiting non-orthogonal characteristics; let... Represents the pilot sequence of the zero correlation domain The autocorrelation function, then The average autocorrelation is shown below:

[0144]

[0145] Where L represents the sequence length, M represents the number of root LPS sequences, and N represents the perfect sequence length;

[0146] Analysis of the superposition of two zero-correlation domain displacement pilots P Uk and P Uq The cross-correlation value needs to consider three different cases:

[0147] Case 1: The pilot signals of the two zero-correlation domain shift superpositions have the same combination of zero-correlation domain sequences;

[0148] Case 2: Two zero-correlation domain shifts are superimposed, and pilots share d zero-correlation domain sequences;

[0149] Case 3: The zero-correlation domain sequences of the two zero-correlation domain shift-superimposed pilots do not overlap at all; since the two zero-correlation domain shift-superimposed pilots are orthogonal in Case 3, we mainly focus on the first two cases; therefore, the probability that the two zero-correlation domain shift-superimposed pilots share d zero-correlation domain sequences is given by the following formula:

[0150]

[0151] Where Z represents the size of the set of zero-correlation domain sequences, r represents the superposition coefficient, and d represents that the two ZSPs share d zero-correlation domain sequences;

[0152] Therefore P Uk and P Uq The average cross-correlation value is derived as follows:

[0153]

[0154] in denoted by , where L represents the sequence length, N represents the perfect sequence length, V′ represents the number of columns in the Hadamard matrix recombination, and r represents the superposition coefficient.

[0155] Therefore, when r = 2, the average cross-correlation value between any two zero-correlation domain shift superimposed pilots is O(1 / L); furthermore, if r increases, the average cross-correlation value will also increase.

[0156] The following combination Figures 1 to 11The simulation results and technical solutions of the Internet low-correlation pilot sequence construction method of the present invention are further described in detail.

[0157] I. The simulation results of the Internet low-correlation pilot sequence construction method of this invention are as follows:

[0158] Figure 2 This paper presents the AFP performance of the ZSP-based ZT-GFRA scheme under different superposition coefficients r and signal-to-noise ratios. We observe that the system AFP performance is optimal when r = 2, with a pilot length of L = 168, resulting in a ZSP pilot set size of 11476. When r = 1, the smaller available pilot set (672) leads to higher PCP, thus affecting AFP performance. When r > 2, even with the pilot set further expanded to 573800, the average cross-correlation value of ZSP increases with r. Therefore, when pilot resources are absolutely sufficient relative to the number of UEs, the system performance deteriorates when r > 2, indicating that ZSP performance is optimal when r = 2.

[0159] PCP performance is mainly related to the size of the pilot set and the number of active terminals K. Figure 3 The relationship between PCP and the number of activated terminals for different pilots with different pilot lengths is shown, and the PCP performance of the GFRA scheme under three different pilots (ZSP, LPS, mZCS) is compared. It can be observed that since mZCS is obtained by shifting N-1 root ZCSs, the number of pilots can reach N(N-1), thus exhibiting the lowest PCP for the same pilot length. ZSP, due to its... The sequence ZSP has the largest number of pilot sequences, where z represents the size of the null correlation domain sequence set and r represents the superposition coefficient. Its PCP performance is the second largest. LPS, due to its smallest pilot set, has the highest PCP. Furthermore, it can be observed that under the same parameters (pilot length set to 168 / 336), ZSP's PCP is approximately 10 times lower than LPS.

[0160] Figure 4 This paper presents a comparison of the DFP performance of the GFRA scheme under different pilots (ZSP, LPS, mZCS). DFP is not only affected by SNR and the number of active terminals, but also depends primarily on factors such as pilot length, encoding / decoding scheme, decoding algorithm, and the average cross-correlation of the pilot sequence set. Here, we keep other variables constant under different pilots to compare the impact of pilots on the system's DFP performance. The average cross-correlation values ​​for ZSP and LPS are both O(1 / L) (the superposition coefficient for ZSP is set to r = 2), which is higher than that for mZCS. Low. It can be observed that as the number of UEs increases, the DFP of ZSP significantly outperforms LPS and mZCS. Furthermore, when K is small, the DFP of ZSPs with different pilot lengths shows only minimal difference, due to the large pilot set of ZSPs (far exceeding the number of terminals) and low cross-correlation values. In this case, DFP performance is primarily limited by the decoder.

[0161] Figure 5 The simulation results of the ZT-GFRA scheme versus the LSP-GFRA scheme with respect to the number of active UEs K and the pilot length in the AFP are presented. Figure 6 The simulation of SNR using AFP is shown. It can be observed that when the SNR is below 5dB, the ZT-GFRA scheme performs worse than LPS-GFRA. This is due to the decreased decoding performance of the SCJD decoder at low SNR. As the SNR increases, the ZT-GFRA scheme gradually outperforms the LPS-GFRA scheme. This is because the ZSP sequence upon which the ZT-GFRA scheme is based has a sufficiently large pilot set, thus possessing a stronger ability to resolve pilot collisions. For users experiencing collisions, it can also successfully recover some UE codewords.

[0162] Figure 7 The simulation of the AFP with respect to the number of active UEs K and pilot length is presented, comparing the ZT-GFRA scheme with the LT-GFRA scheme. Figure 8 The simulation of SNR using AFP is presented. It can be observed that when the number of users is small, the difference between the two schemes is minimal because the number of active UEs is much smaller than the pilot set size, resulting in minimal UE conflicts. However, as the number of users increases, the drawbacks of the LT-GFRA scheme gradually become apparent. Due to limited orthogonal pilot resources, conflicts increase with the increase of K, leading to a gradual deterioration in system access performance. In summary, compared with the two large-scale access schemes, the ZT-GFRA scheme demonstrates significant performance advantages. It reduces the probability of conflicts through more abundant pilot resources and employs a multi-order coding scheme to further resolve some conflicting terminals, making it more suitable for scenarios with massive device access.

[0163] Figure 9 By combining ZSP, LPS pilot, and T-fold coding schemes for simulation comparison, it can be observed that the code domain GFRA system under ZSP has significantly better performance than the latter. This is because ZSP can significantly increase the number of pilots while ensuring low correlation characteristics, allowing the system to accommodate more terminals, thus resulting in better system access performance.

[0164] II. The system model of the code-domain unlicensed large-scale access protocol based on satellite internet in the low-correlation pilot sequence construction method of this invention is as follows:

[0165] In real-world scenarios, it is often a multi-beam system, while we study it separately. Figure 10 This paper presents a communication system under the coverage of a single sub-beam. In this scenario, satellite Sa provides mMTC-s service to terminals within its beam coverage area. These terminals are assumed to be quasi-stationary and uniformly distributed within the coverage area. Given that the link length between the satellite and the ground can reach hundreds of kilometers, we approximate that all terminals are equidistant from Sa, thus achieving almost simultaneous signal arrival times. We assume that the duration of each frame matches the length of a time slot, denoted as n, where propagation delay and transceiver processing delay are both included within a random access time slot. Considering a maximum two-way propagation delay of approximately 26 milliseconds, we set the duration of the random access time slot to 50 milliseconds. The mMTC-s scenario allows for a potentially unlimited number of terminal accesses, but in each random access time slot, only a small subset of terminals are activated with probability pa = 0.01 and communicate with the satellite; this subset of activated terminals is denoted as K.

[0166] In wireless communication systems, the channel characteristics of satellite-to-ground links are affected by numerous factors, such as the presence of obstacles like buildings, vegetation, and terrain, due to multipath effects and shadowing fading in the environment. These factors complicate the signal propagation path, leading to signal instability. To accurately simulate this situation, the Shadowed Rician Fading Channel (SR) model is commonly used. This model fully considers shadowing and fading effects, thus more realistically reflecting the actual wireless propagation environment. The probability density function (PDF) of SR is as follows:

[0167] f(h 2 )=(2dn / (2dn+Ω) n (1 / 2d)exp(-h 2 / 2d)1F1(n,1,Ωh 2 / 2d(2dn+Ω));

[0168] Where d represents the multipath component, Ω represents the average power at line of sight (LoS), n represents the Nakagami-n parameter, and 1F1(i,j,k) represents the confluence hypergeometry function. By setting the parameters d = 0.063, n = 1, and Ω = 0.000897, the above equation can be simplified to:

[0169] f(r) = we -ηr , r>o;

[0170] In the formula Where w and η are the coefficients of the exponential distribution, and r = h2 As the fading coefficient of SR, P T It is the user power, σ 2 It is the noise variance.

[0171] III. The specific description of the uplink code field unlicensed large-scale access protocol in the low-correlation pilot sequence construction method of the satellite internet of the present invention is as follows:

[0172] First, set the ZSP pilot set as Where L is the pilot length of the ZSP, and S represents the pilot set size of the ZSP. The terminal activated at the UE selects the pilot using its first r·b bits of information, and then performs multi-level coding operations on the remaining information. The encoded signal is then modulated and spread-spectrum processed before being transmitted to the satellite.

[0173] Let X u Y represents the codeword after the UE information has been encoded and modulated using a T-order codebook. p and Y d S represents a The received pilot signals and data information are specifically represented as follows:

[0174]

[0175] as well as

[0176]

[0177] Where j and u represent different user indexes, P t It is the user power, P Uj It is an LPS pilot, Y d This represents the received coded signal, S represents the pilot set size, and φ represents the received coded signal. j This represents the set of terminals whose first r parts are all identical; that is, these terminals select the same zero-correlation domain periodic sequence for their first r parts, superimpose them, and correspond to the same ZSP sequence P in the received frame. Uj , Representing the Kronecker product, W is additive white Gaussian noise (AWGN) with a form CN(0, σ). 2 In addition, h u This represents the SR coefficient of UEu.

[0178] Satellite S a Based on the received Y p The data is subjected to pilot detection, and the first r·b bits of the received frame are decoded. Then, the selected P... Uj The terminal performs channel estimation.

[0179] The channel weighted sum estimate of the terminal is expressed as follows:

[0180]

[0181] This allows us to obtain the selected pilot P. Uj The remaining portion of the received data Y in the transmission frame d The encoded message in the code is represented as follows:

[0182]

[0183] Where u represents different users, h u X represents the channel fading coefficient for different users. u P represents the transmitted user-coded data. t It is the user's power. This represents the received coded signal, and S represents the pilot set size. Subsequently, S... a The remaining portion of the transmitted frame is decoded using SCJD. SCJD first iteratively performs serial cancellation decoding, where the residual signal obtained after the i-th step of serial cancellation decoding is as follows:

[0184]

[0185] Where q represents different users, h q X represents the channel fading coefficient for different users. q P represents the transmitted user-coded data. t It is the user's power. Let S represent the received coded signal and S represent the pilot set size. However, if the serial decoding in step i fails, joint decoding is initiated. Joint decoding can recover at most Ti conflicting terminals. The input signals for joint decoding are as follows:

[0186]

[0187] Where δ is the residual interference coefficient after decoding, T is the multi-order coding order, v represents different active users, and h v P represents the channel fading coefficient for different users. t This is the user's power. Let... We need to exhaustively enumerate the residual signals All possible combinations, and compare each estimate with the actual residual signal. Compare to obtain the difference g e Then it is compared with the threshold θ as follows:

[0188]

[0189] Where I represents the error between the actual received signal and the estimated signal, h l This represents the channel fading coefficient for different users. V represents an estimate of I. T-i This indicates that it belongs to the residual signal. The code X u Set, threshold θ = (N) c -L)σ 2 , where N c This represents the length of a T-order codeword. Iteration continues until g is reached. e If the value drops below θ, the estimation is considered successful.

[0190] Next, we iteratively perform joint decoding, including decoding of both internal and external codes, to recover all remaining messages from at most Ti terminals. However, if after exhausting all Ti attempts, g... e >θ, meaning that a valid set V cannot be found in the set. T-i S a The declaration indicates that the joint decoding failed and is discarded.

[0191] IV. The ZSP design scheme in the low-correlation pilot sequence construction method of satellite internet in this invention is as follows:

[0192] Zero-correlation domain sequences possess perfect autocorrelation and cross-correlation properties, ensuring that sequences periodically shifted within the zero-correlation domain from the same root sequence are mutually orthogonal, and that shifted sequences from different root sequences are also mutually orthogonal. Furthermore, compared to traditional orthogonal pilots (ZCS pilots), zero-correlation domain sequences have a larger sequence set. Therefore, we consider fully utilizing the properties of zero-correlation domain sequences to construct new sequences to further expand pilot resources.

[0193] To expand the sequence set, we introduce a random combination superposition method. This involves randomly selecting r sequences from the zero-correlation domain periodic sequence set, superimposing them while keeping the pilot length unchanged, and summing the corresponding elements to obtain the ZSP sequence P. U .

[0194]

[0195] Where p v It is an LPS sequence, where i, j, ..., v represent different sequence indices, and Mτ z This represents the size of the zero-correlation domain sequence set. By iterating through all r random combinations of sequences, we obtain the ZSP sequence set. Based on the construction process of the ZSP sequence set, we can determine its size as follows: Where Z = M·τ z The value represents the size of the set of zero-correlation domain sequences, M represents the number of root LPS sequences, and τ represents the value of the set of zero-correlation domain sequences. z This indicates the number of bits that the zero-correlation field sequence can be shifted. For example... Figure 11To better understand ZSP sets, we provide a specific example as follows:

[0196] First, a set of zero-correlation domain sequences is generated, where the order of the Hadamard matrix is ​​set to M = 2 (i.e., the number of root sequences), the length of the perfect sequence e = {1, 1, 0, 1, 0, 0, -1} is N = 7, and the number of columns after reassembling the Hadamard matrix is ​​V′ = 2. Therefore, the pilot length of the ZSP is equal to the length of the zero-correlation domain sequence, which is L = M·N = 14, and the number of shiftable bits within the zero-correlation domain is... Therefore, we can obtain the size of the null correlation domain as Z = M·τ z =10, the zero-correlation domain sequence is as follows:

[0197] p 1，1 ={+1,+1,0,0,0,+1,-1,0,+1,0,+1,-1,0,+1},

[0198] p 1，2 ={+1, 0, 0, 0, +1, -1, 0, +1, 0, +1, -1, 0, +1, +1},

[0199] P 1,3 ={0, 0, 0, +1, -1, 0, +1, 0, +1, -1, 0, +1, +1, +1},

[0200] p 1，4 ={0, 0, +1, -1, 0, +1, 0, +1, -1, 0, +1, +1, +1, 0}

[0201] P i,5 ={0, +1, -1, 0, +1, 0, +1, -1, 0, +1, +1, +1, 0, 0}

[0202] P 2,1 ={-1, +1, 0, 0, 0, +1, +1, 0, -1, 0, -1, -1, 0, +1},

[0203] p 2，2 ={+1, 0, 0, 0, +1, +1, 0, -1, 0, -1, -1, 0, +1, -1},

[0204] P 2,3 ={0, 0, 0, +1, +1, 0, -1, 0, -1, -1, 0, +1, -1, +1},

[0205] P 2,4 ={0, 0, +1, +1, 0, -1, 0, -1, -1, 0, +1, -1, +1, 0}

[0206] P 2,5= {0, +1, +1, 0, -1, 0, -1, -1, 0, +1, -1, +1, 0, 0}.

[0207] Let A q ={a1, a2, ..., a r} represents the index set of the q-th selection, where q = 1, 2, ..., S. Therefore, ZSP can generate a total of r = 2 sequences randomly selected from the zero-correlation domain periodic sequences and superimposed on them. A ZSP sequence, namely A q ={p i，j p u，v ), 1≤i, u≤2, 1≤j, v≤5. Then, the q-th ZSP can be defined as:

[0208]

[0209] Where a l P represents the sequence index. Uq Representing different ZSP sequences, A q Let represent the index set of the q-th selection, l represent different sequence indices, and r represent the superposition coefficient. Clearly, for ZSP, the probability that an active terminal does not collide with other active terminals is given by the following formula:

[0210]

[0211] Where P cf Let S represent the probability that an activated terminal does not collide with other activated terminals, S represent the size of the ZSP sequence set, and K represent the number of activated users. Therefore, if the superposition coefficient r = 1 (i.e., the traditional method of directly using zero-correlation domain pilots), and the number of activated UEs is set to K = 100, then the probability that an activated terminal does not collide is... Where τ m Let M represent the maximum number of shiftable bits in the low-relevance domain sequence, M represent the number of root LPS sequences, and K represent the number of active users. However, when using ZSP, with the superposition coefficient r = 2, the probability becomes... Clearly, ZSP significantly reduces the probability of terminal conflicts.

[0212] V. The correlation analysis of ZSP in the low-correlation pilot sequence construction method of satellite internet of the present invention is as follows: Let... Represents the pilot sequence of the zero correlation domain and The cross-correlation function between them, where τ u , τ v This represents the different shift values ​​corresponding to different sequences. The shift difference between two sequences is Δτ. u，v =|τ u -τ v |bits, where Δτu，v This represents the shift difference between two sequences. Furthermore, let... ZSP P Uk and P Uq The cross-correlation function between them, where P Uk P Uq This represents different user indexes, and their index sets are: and in To represent different LPS sequences, τ a , τ b , τ c , τ d Let k1, k2, q1, q2 represent different shift values, and r = 2 represent different root sequences. Therefore, the cross-correlation between any two ZSPs is as follows:

[0213] in Let Δτ represent the cross-correlation value between any two ZSPs. a，c ,Δτ a，d ,Δτ b，c ,Δτ b,d Indicates different shift values, To represent different LPS sequences, τ a , τ b , τ c , τ d The numbers k1, k2, q1, and q2 represent different shift values, and k1, k2, q2 represent different root sequences.

[0214] If ZSP P is constituted Uk and P Uq If the zero-correlation domain sequences are different, then the cross-correlation value between two ZSPs is zero. Probability statistics show that the number of ZSPs composed of completely different sequences is... Where Z = M·τ z The value represents the size of the set of zero-correlation domain sequences, M represents the number of root LPS sequences, and τ represents the value of the set of zero-correlation domain sequences. z Let r represent the number of shiftable bits of the zero-correlation domain sequence, and r represent the superposition coefficient. Therefore, other ZSPs in this set remain consistent with P. Uk The probabilities of orthogonality are as follows:

[0215]

[0216] Where Z = M·τ z The value represents the size of the set of zero-correlation domain sequences, M represents the number of root LPS sequences, and τ represents the value of the set of zero-correlation domain sequences. z Let r represent the number of shiftable bits in the zero-correlation domain sequence, and r represent the superposition coefficient. Note that for r << (τ) z ·M), P O→1, meaning that most ZSPs maintain orthogonality, with only a small portion exhibiting non-orthogonal properties. Let... Represents the pilot sequence of the zero correlation domain The autocorrelation function, then The average autocorrelation is shown below:

[0217]

[0218] Where L represents the sequence length, M represents the number of root LPS sequences, and N represents the perfect sequence length. Analyze two ZSP P Uk and P Uq The cross-correlation value needs to consider three different cases: Case 1: The two ZSPs have the same combination of zero-correlation domain sequences; Case 2: The two ZSPs share d zero-correlation domain sequences; Case 3: The combinations of zero-correlation domain sequences of the two ZSPs do not overlap at all. Since the two ZSPs are orthogonal to each other in Case 3, we mainly focus on the first two cases. Therefore, the probability that two ZSPs share d zero-correlation domain sequences is given by the following formula:

[0219]

[0220] Where Z represents the size of the zero-correlation domain sequence set, r represents the superposition coefficient, and d represents that the two ZSPs share d zero-correlation domain sequences.

[0221] Therefore P Uk and P Uq The average cross-correlation value is derived as follows:

[0222]

[0223] in Let represent the average cross-correlation value, L represent the sequence length, N represent the perfect sequence length, V′ represent the number of columns in the Hadamard matrix recombination, and r represent the superposition coefficient. Therefore, when r = 2, the average cross-correlation value between any two ZSPs is O(1 / L). Furthermore, if r increases, the average cross-correlation value will also increase.

[0224] Table 1. Probability distribution of cross-correlation values

[0225]

[0226] To clearly and intuitively illustrate the cross-correlation characteristics of ZSP, the table above summarizes the probability distribution of cross-correlation values ​​between the pilots selected by UE1 (mZCS, LPS, and ZSP, respectively) and the pilots selected by other terminals, where r = 2 and the pilot length is L = 168 / 167. It can be observed that for ZSP and LPS, the cross-correlation values ​​between other terminals and UE1 are approximately 97% equal to 0. However, since the mZCSs constructed from shifts of different root ZCS sequences exhibit the same cross-correlation value, this also results in 99% of the pilots having non-zero cross-correlation. This indicates that most mZCSs are non-orthogonal, leading to significant mutual interference between terminals and making mZCS vulnerable to noisy channels in GFRA.

[0227] It should be noted that the reliability of the system can be further improved by refining the decoding method.

[0228] The beneficial effects of the low-correlation pilot sequence construction method for satellite internet of the present invention are:

[0229] 1. This invention addresses the problem of massive device access in mMTC-s communication scenarios under satellite internet by proposing a ZT-Collision Resolution Grant-Free Random Access (ZT-GFRA) scheme; ZT: pilot sequence with low cross-correlation value in the zero-correlation domain + T-order high-dimensional universal codebook. The scheme improves the system's spectral efficiency by designing an efficient and reliable pilot allocation strategy. This scheme not only adapts to the needs of massive device access but also achieves lower PCP, decoding failure probability (DFP), and AFP in the short-packet communication mMTC-s scenario, providing an important reference for addressing the high-density access challenges of future satellite internet.

[0230] 2. This invention designs a ZSP sequence set as the pilot set for the ZT-GFRA scheme. Utilizing the idea of ​​random combination and superposition, based on the zero correlation domain sequence, a ZSP sequence with a larger pilot set is generated. The scale of its pilot set far exceeds the pilot length, which can effectively reduce PCP and make it more suitable for future high-density massive device access scenarios.

[0231] 3. This invention analyzes the average cross-correlation characteristics of the ZSP sequence, derives an expression for the average cross-correlation value with respect to the superposition coefficient r, and thus provides the optimal value of the superposition coefficient r for the ZSP sequence. It also derives that this pilot sequence has an average cross-correlation value close to 0, which can significantly reduce interference between terminals. Monte Carlo simulations verify that the ZSP exhibits superior performance in shadow-Rice fading channels.

[0232] To achieve the above objectives, this invention also proposes a low-correlation pilot sequence construction system for satellite internet, such as... Figure 12 As shown, the system includes a processor 1001, a CPU, a network interface 1004, a user interface 1003, a memory 1005, a communication bus 1002, and a low-correlation pilot sequence construction program for satellite internet stored on the processor. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen and an input unit such as a keyboard. Optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface). The memory 1005 may be high-speed RAM or non-volatile memory, such as a disk drive. Optionally, the memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0233] Those skilled in the art will understand that Figure 12 The system structure shown does not constitute a limitation on the system and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0234] like Figure 12 As shown, the memory 1005, which is a computer storage device, may include an operating device, a network communication module, a user interface module, and a low-correlation pilot sequence construction program for satellite internet.

[0235] exist Figure 12 In the system shown, network interface 1004 is mainly used to connect to the network server and communicate with the network server; user interface 1003 is mainly used to interact with user terminals and receive user input commands; and processor 1001 can be used to call the satellite internet low-correlation pilot sequence construction program stored in memory 1005.

[0236] To achieve the above objectives, the present invention also proposes a computer-readable storage device storing a joint order likelihood decoding program for satellite internet. When the joint order likelihood decoding program for satellite internet is run by a processor, the steps of the method described above are executed, and will not be repeated here.

[0237] The above description is only a preferred embodiment of the present invention and does not limit the patent scope of the present invention. All equivalent structural transformations made under the concept of the present invention using the contents of the present invention specification and drawings, or direct / indirect applications in other related technical fields, are included within the patent protection scope of the present invention.

Claims

1. A method for constructing low-correlation pilot sequences for the Internet, characterized in that, The method includes the following steps: Step S100: After the receiving satellite receives the concatenated information sent by the transmitting end, it performs pilot detection and channel estimation using a set of zero-correlation domain shifted superimposed pilot sequences, wherein the least squares algorithm is applied to obtain the channel response; wherein the concatenated information is obtained by each active UE at the transmitting end dividing a k-bit information into r+1 blocks, wherein the first r blocks are all b bits long, each block is mapped to a zero-correlation domain sequence according to b bits, and these r sequences are superimposed to generate a zero-correlation domain shifted superimposed pilot sequence. Subsequently, the remaining k-rb bits are encoded using a T-order T-fol d codebook and concatenated with the zero-correlation domain shifted superimposed pilot sequence to obtain the result. Step S200: Decode the remaining k-rb bits of no more than T conflicting terminals in each frame using the SCJD algorithm, and recover the first rb bits of each UE using the pilot sequence obtained from pilot detection; In step S300, the satellite broadcasts a decoding status feedback message to all active users; Prior to step S100, the following is included: Step S000: Construct a system model for a code domain unlicensed large-scale access protocol based on satellite internet; The specific description of the uplink code domain unlicensed mass access protocol is as follows: First, set the zero-correlation domain shift superposition pilot set as Where L is the pilot length of the zero-correlation domain displacement superimposed pilot, and S represents the size of the pilot set of the zero-correlation domain displacement superimposed pilot; the terminal activated at the UE uses its first rb bits of information to select the pilot, and then performs multi-level coding operations on the remaining information. The encoded signal is then modulated and spread spectrum processed before being sent to the satellite. Let X u Y represents the codeword after the UE information has been encoded and modulated using a T-order codebook. p and Y d S represents a The received pilot signals and data information are specifically represented as follows: as well as Where j and u represent different user indexes, P t It is the user power, P Uj It is an LPS pilot, Y d This represents the received coded signal, S represents the pilot set size, and φ represents the received coded signal. j This represents the set of terminals whose first r parts are all identical; that is, these terminals select the same zero-correlation domain periodic sequence for superposition in their first r parts, and correspond to the same zero-correlation domain shifted superposition pilot sequence P in the received frame. Uj , Representing the Kronecker product, W is additive white Gaussian noise, and is CN(0, σ). 2 ), h u Represents the SR coefficient of UEu; Satellite S a Based on the received Y p The data is subjected to pilot detection, and the first r·b bits of the received frame are decoded. Then, the selected P... Uj The terminal performs channel estimation; The channel weighted sum estimate of the terminal is expressed as follows: This allows us to obtain the selected pilot P. Uj The remaining portion of the received data Y in the transmission frame d The encoded message in the code is represented as follows: Where u represents different users, h u X represents the channel fading coefficient for different users. u P represents the transmitted user-coded data. t It is the user's power. This represents the received coded signal, and S represents the pilot set size; Subsequently, S a The remaining portion of the transmitted frame is decoded using SCJD. SCJD first iteratively performs serial cancellation decoding, where the residual signal obtained after the i-th step of serial cancellation decoding is as follows: Where q represents different users, h q X represents the channel fading coefficient for different users. q This represents the transmitted user-coded data, where Pt is the user power. This represents the received coded signal, and S represents the pilot set size; However, if the serial decoding in step i fails, joint decoding is initiated. Joint decoding can recover at most Ti conflicting terminals. The input signals for joint decoding are as follows: Where δ is the residual interference coefficient after decoding; T is the multi-order coding order; v represents different active users; h v P represents the channel fading coefficient for different users. t It is the user's power; set up Exhaustive search of residual signals All possible combinations, and compare each estimate with the actual residual signal. Compare to obtain the difference g e Then it is compared with the threshold θ as follows: Where I represents the error between the actual received signal and the estimated signal, h l This represents the channel fading coefficient for different users. V represents an estimate of I. T-i This indicates that it belongs to the residual signal. The code X u Set, threshold θ = (N) c -L)σ 2 , where N c The length of the T-order codeword is represented by continuous iteration until g. e If the value drops below θ, the estimation is considered successful. Next, iterative joint decoding is performed, including decoding of both internal and external codes, to recover all remaining messages from up to Ti terminals; however, if after exhausting all Ti attempts, g e >θ, meaning no valid set V can be found in the set. T-i S a The declaration indicates that the joint decoding failed and is discarded. The method further includes: designing zero-correlation domain displacement superimposed pilots: a random combination superposition method is introduced, that is, r sequences are randomly selected from the set of zero-correlation domain periodic sequences, and then superimposed. The pilot length remains unchanged, and the corresponding elements are summed to obtain the zero-correlation domain displacement superimposed pilot sequence P. U : Where p v It is an LPS sequence, where i, j, ..., v represent different sequence indices, and Mτ z Indicates the size of the set of zero-correlation domain sequences; By iterating through all r random combinations of sequences, a set of zero-correlation domain shift-superimposed pilot sequences is obtained; based on the construction process of the zero-correlation domain shift-superimposed pilot sequence set, the size of the zero-correlation domain shift-superimposed pilot sequence set is... Where Z = M·τ z The value represents the size of the set of zero-correlation domain sequences, M represents the number of root LPS sequences, and τ represents the value of the set of zero-correlation domain sequences. z This represents the number of bits that the zero-correlation field sequence can be shifted.

2. The method for constructing low-correlation pilot sequences for the Internet according to claim 1, characterized in that, Step S000 includes: Step S001: In a communication system covered by a single sub-beam, satellite Sa provides mMTC-s service to terminals within its beam coverage area. It is assumed that these terminals are quasi-stationary and uniformly distributed within the coverage area. It is approximately assumed that all terminals are equidistant from Sa, thus achieving almost simultaneous signal arrival times. It is assumed that the duration of each frame matches the duration of a time slot, denoted as n, where propagation delay and transceiver processing delay are both included within a random access time slot. Considering the maximum bidirectional propagation delay is approximately 26 milliseconds, the duration of the random access time slot is set to 50 milliseconds. The mMTC-s scenario allows for a potentially unlimited number of terminal accesses, but in each random access time slot, only a small subset of terminals are activated with probability pa = 0.01 and communicate with the satellite; this subset of activated terminals is denoted as K. A shadowed Ricean fading channel model is used to simulate a real wireless propagation environment. The probability density function of the shadowed Ricean fading channel model is: f(h 2 )=(2dn / (2dn+Ω) n (1 / 2d)exp(-h 2 / 2d)1F1(n,1,Ωh 2 / 2d(2dn+Ω)); Where d represents the multipath component, Ω represents the average power at line of sight, n represents the Nakagami-n parameter, and 1F1(i,j,k) represents the confluence hypergeometry function. By setting the parameters d = 0.063, n = 1, and Ω = 0.000897, the above equation can be simplified to: f(r)=we -ηr ,r>o; In the formula Where w and η are the coefficients of the exponential distribution, and r = h 2 As the fading coefficient of SR, P T It is the user power, σ 2 It is the noise variance.

3. The method for constructing low-correlation pilot sequences for the Internet according to claim 2, characterized in that, The method further includes: correlation analysis of zero-correlation domain shift superimposed pilots: set up Represents the pilot sequence of the zero correlation domain and The cross-correlation function between them, where τ u , τ v This represents the different shift values ​​corresponding to different sequences, where the shift difference between two sequences is Δτ. u,v =|τ u -τ v | bits, where Δτ u,v Let the shift difference between the two sequences be represented by... The zero-correlation domain displacement superposition pilot P Uk and P Uq The cross-correlation function between them, where P Uk P Uq This represents different user indexes, and their index sets are: and in To represent different LPS sequences, τ a , τ b , τ c , τ d Let k1, k2, q1, and q2 represent different shift values, and r = 2. Therefore, the cross-correlation of any two zero-correlation domain shifts superimposed with pilots is as follows: in Let Δτ represent the cross-correlation value between any two ZSPs. a,c ,Δτ a,d ,Δτ b,c ,Δτ b,d Indicates different shift values, To represent different LPS sequences, τ a , τ b , τ c , τ d These represent different shift values, and k1, k2, q1, q2 represent different root sequences; If ZSP P is constituted Uk and P Uq If the zero-correlation domain sequences are different, then the cross-correlation value between two ZSPs is zero; probability statistics show that the number of ZSPs composed of completely different sequences is... Where Z = M·τ z The value represents the size of the set of zero-correlation domain sequences, M represents the number of root LPS sequences, and τ represents the value of the set of zero-correlation domain sequences. z Let r represent the number of shiftable bits of the zero-correlation domain sequence, and r represent the superposition coefficient; therefore, the shifted superposition pilots of other zero-correlation domains in this set remain consistent with P. Uk The probabilities of orthogonality are as follows: Where Z = M·τ z The value represents the size of the set of zero-correlation domain sequences, M represents the number of root LPS sequences, and τ represents the value of the set of zero-correlation domain sequences. z The number of bits that can be shifted in the zero-correlation domain sequence is represented by r, which represents the superposition coefficient. For r << (τ) z ·M), P O →1, meaning that most of the zero-correlation domain shift superposition pilots remain orthogonal, with only a small portion exhibiting non-orthogonal characteristics; let... Represents the pilot sequence of the zero correlation domain The autocorrelation function, then The average autocorrelation is shown below: Where L represents the sequence length, M represents the number of root LPS sequences, and N represents the perfect sequence length; Analysis of the superposition of two zero-correlation domain displacement pilots P Uk and P Uq The cross-correlation value needs to consider three different cases: Case 1: The pilot signals of the two zero-correlation domain shift superpositions have the same combination of zero-correlation domain sequences; Case 2: Two zero-correlation domain shifts are superimposed, and pilots share d zero-correlation domain sequences; Case 3: The zero-correlation domain sequences of the two zero-correlation domain shift-superimposed pilots do not overlap at all; since the two zero-correlation domain shift-superimposed pilots are orthogonal in Case 3, we mainly focus on the first two cases; therefore, the probability that the two zero-correlation domain shift-superimposed pilots share d zero-correlation domain sequences is given by the following formula: Where Z represents the size of the set of zero-correlation domain sequences, r represents the superposition coefficient, and d represents that the two ZSPs share d zero-correlation domain sequences; Therefore P Uk and P Uq The average cross-correlation value is derived as follows: in denoted by , where L represents the sequence length, N represents the perfect sequence length, V′ represents the number of columns in the Hadamard matrix recombination, and r represents the superposition coefficient. Therefore, when r = 2, the average cross-correlation value between any two zero-correlation domain shift superimposed pilots is O(1 / L); furthermore, if r increases, the average cross-correlation value will also increase.

4. A system for constructing low-correlation pilot sequences for the Internet, characterized in that, The system includes a memory, a processor, and a low-correlation pilot sequence construction program for the Internet stored on the processor, the low-correlation pilot sequence construction program for the Internet being executed by the processor to perform the steps of the method as described in any one of claims 1 to 3.

5. A computer-readable storage device, characterized in that, The computer-readable storage device stores a low-correlation pilot sequence construction program for the Internet, which, when executed by a processor, performs the steps of the method as described in any one of claims 1 to 3.

Images