A large-scale unlicensed random access method for low-orbit satellite Internet of Things

The proposed tensor-based random access method for low-orbit satellite IoT devices addresses high access delays and channel complexities by using tensor signal construction and Bayesian learning, ensuring efficient device detection and channel estimation.

CN114630446BActive Publication Date: 2024-12-06ZHEJIANG UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210149690.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-02-18
Publication Date
2024-12-06
Estimated Expiration
2042-02-18

AI Technical Summary

Technical Problem

Existing satellite communication random access schemes for low-orbit IoT devices face high access delays and complex ground-satellite link channels due to mobility and channel composition complexities, which are not adequately addressed by existing unlicensed random access methods designed for Rayleigh channels.

Method used

A large-scale unlicensed random access method for low-orbit satellite IoT devices using tensor signal construction and tensor-based Bayesian learning to process preamble sequences, enabling efficient detection and channel estimation for activated devices.

Benefits of technology

The method effectively reduces access delay and computational complexity, allowing for seamless large-scale access of IoT devices by accurately detecting and estimating channel states, suitable for low-orbit satellite IoT systems.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114630446B_ABST
    Figure CN114630446B_ABST
Patent Text Reader

Abstract

The present invention discloses a large-scale unauthorized random access method for a low-orbit satellite Internet of Things. An activated Internet of Things device constructs a corresponding tensor-formed preamble sequence based on an assigned preamble sequence, and sends the tensor-formed preamble sequence to a low-orbit satellite through an uplink channel at the beginning of each time slot. The low-orbit satellite recovers the device state matrix based on the received signal, and then detects the activated device using a threshold method. For devices detected as being in an activated state, the corresponding channel state information is estimated based on the device state matrix. The present invention provides a simple but effective unauthorized random access method for a low-orbit satellite Internet of Things with large-scale Internet of Things device access.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of wireless communications, and in particular to a large-scale authorization-free random access method for a low-orbit satellite Internet of Things. Background Art

[0002] With the rapid development of the Internet of Things, a large number of IoT devices need to access wireless networks. However, a large number of IoT devices are distributed in remote areas such as oceans, forests, mountains and deserts, which do not have good wireless coverage. Whether in terms of construction cost or maintenance difficulty, low-orbit satellites are an effective means to achieve global seamless coverage. Therefore, it is urgent to study the giant site access technology for low-orbit satellite IoT. In order to reduce the access delay of IoT devices, it is necessary to design an unlicensed random access technology based on the characteristics of low-orbit satellite communications to meet the giant site access needs of massive IoT devices.

[0003] However, the existing satellite communication random access scheme mainly relies on sufficient access resources and static ground-satellite links. However, in the Internet of Things supported by low-orbit satellites, due to the mobility of low-orbit satellites, the ground-satellite link will change rapidly as the satellite moves, so large-scale access from IoT devices may lead to a shortage of access resources. Therefore, the existing satellite communication random access scheme cannot match the Internet of Things supported by low-orbit satellites. Although the unlicensed random access scheme can improve the spectrum efficiency and reduce the transmission delay, the joint activation detection and channel estimation algorithm in the existing unlicensed random access scheme is designed for Rayleigh channels rather than ground-satellite link channels. In the Rayleigh channel model, the line of sight (LoS) component in wave propagation can be ignored, while the scattering component dominates. However, for the ground-satellite link channel, the channel composition is more complicated, and both the LoS component and the scattering component need to be considered at the same time. Summary of the invention

[0004] The purpose of the present invention is to solve the problems of high access delay of IoT devices and complex ground-satellite link channels in the above-mentioned scheme, and proposes a large-scale unauthorized random access method for low-orbit satellite IoT.

[0005] The specific technical solutions adopted by the present invention are as follows:

[0006] A large-scale unlicensed random access method for low-orbit satellite Internet of Things comprises the following steps:

[0007] S1. For each IoT device d among all K IoT devices k , the assigned leading sequence is s k , through a tensor signal construction method, the IoT device d k The leading sequence T in tensor form k, k=1,...K, where K is the total number of IoT devices; at the beginning of each time slot, all activated IoT devices simultaneously send a preamble sequence in the form of a tensor to the low-orbit satellite, and the signal received by the low-orbit satellite is Y;

[0008] S2, based on the received signal Y, the low-orbit satellite restores the signal Y into the device state matrix M according to a tensor-based Bayesian learning method X , where M X The kth column is the IoT device d k Status information of , k=1,...K;

[0009] S3, low-orbit satellite based on the restored device state matrix M X ,Using a threshold judgment method to detect the IoT devices in the activated state;

[0010] S4. For each IoT device detected as being in an activated state, obtain corresponding channel state estimation information using a channel estimation method.

[0011] Preferably, in step S1, the tensor signal construction method is:

[0012] For each IoT device’s preamble sequence s k , whose length is L, and it is evenly divided into d segments, where the i-th segment is a k,i , i = 1, 2...d; then construct the transmission signal according to the following formula: Among them, τ i Indicates a k,i The dimension of is i=1, 2, ..., d, and the symbol ° represents the vector outer product.

[0013] Preferably, d is 3.

[0014] Preferably, in step S2, the tensor-based Bayesian learning method comprises the following specific steps:

[0015] a) Input the leading sequence signal Y received by the satellite and the number of iterations T, and initialize the current iteration round t = 0;

[0016] b) According to the following formula, the median matrix of the channel corresponding to the current iteration round is and the covariance matrix Update the parameters:

[0017]

[0018]

[0019] Among them, F β represents the expectation of the noise factor β, Ai Indicates that all a k,i The tensor factor matrix composed of A i The kth column of k,i , k = 1, 2...K; y(d+1) represents the matrix of the tensor signal Y expanded in mode d+1; Represents parameter v k Expectation of, k = 1, 2, ... K; (·) -1 represents matrix inversion; represents the mean estimation parameter μ k The first moment of the mean estimate parameter μ k is a hyperparameter; M represents the number of antennas, diag() is a diagonal matrix, ⊙ represents the Hadmard product, ◇ represents the Khatri-rao product, the superscript H represents the conjugate transpose, and the superscript * represents the conjugate operation; and is the initial preset parameter;

[0020] c) Based on the median matrix of the channel obtained in the previous iteration and the covariance matrix The variance of the channel distribution is estimated according to the following formula: and Update the parameters:

[0021]

[0022]

[0023] Among them, δ is the initial value, represents the mean estimation parameter μ k The first moment of represents the mean estimation parameter μ k The second moment of Represents the median matrix The kth column of M X (m, k) represents the median matrix The element in the mth row and kth column of Represents the covariance matrix The element at row k and column k;

[0024] d) Estimate the mean parameter of the channel distribution according to the following formula and Update the parameters:

[0025]

[0026]

[0027] Re-order By and Get the hyperparameter μ k The mean and the second moment for:

[0028]

[0029]

[0030] Where Gamma[] represents the gamma function, H[] represents the confluent hypergeometric function 1F1;

[0031] e) The expected F of the noise factor β according to the following formula β Update the parameters:

[0032]

[0033]

[0034]

[0035] Among them, τ i Indicates A i The row dimension of , Tr() means finding the trace, Indicates finding the F norm of the matrix;

[0036] f) After adding 1 to the current iteration round t, repeat steps b) to e) for the next iteration. After a total of T iteration rounds, the median matrix of the channel is finally output. As the device state matrix M X .

[0037] As a preference, the initial value δ is set to 10 -6 .

[0038] Preferably, in step S3, the threshold determination method is:

[0039] Set threshold θ0 = M*(0.15*α) 2 , where α is M X The largest element in; traverse the median matrix M in turn X ∈C M×K For each column in the median matrix M X ∈C M×K Any k-th column M in X (:,k), if the norm IoT devices k is determined to be in an activated state. Otherwise, IoT devices k Determined to be inactive.

[0040] Preferably, in step S4, the channel estimation method is:

[0041] If the IoT device k If it is detected as an activated device, the low-orbit satellite will be based on the device state matrix M X , extract M X The kth column data is used as the IoT device d k Estimated information of the channel state

[0042] The invention has the following beneficial effects: the tensor signal construction method proposed in the invention effectively solves a series of problems caused by the long preamble sequence by dividing a long preamble sequence into several segments for transmission. The tensor-based Bayesian learning method proposed in the invention has the characteristics of low computational complexity and can handle a large number of potential devices, which is suitable for the business characteristics of low-orbit satellite Internet of Things. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a block diagram of a large-scale unlicensed random access method for low-orbit satellite Internet of Things;

[0044] Figure 2 It is the performance comparison of channel normalization error and number of iterations under different signal-to-noise ratios;

[0045] Figure 3 It is the performance comparison of the detection error probability and noise ratio of the activated device under different numbers of antennas; DETAILED DESCRIPTION

[0046] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0047] On the contrary, the present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention as defined by the claims. Further, in order to make the public have a better understanding of the present invention, some specific details are described in detail in the detailed description of the present invention below. Those skilled in the art can fully understand the present invention without the description of these details.

[0048] like Figure 1As shown, it is a block diagram of a large-scale unlicensed random access method for low-orbit satellite Internet of Things provided in this embodiment. The satellite receives the tensor-form leading sequence transmitted by the Internet of Things device, detects the channel state matrix of the Internet of Things device, and then determines the activated Internet of Things device according to the channel state matrix and the threshold, and obtains the channel state estimation information of the activated device according to the channel state matrix. The specific implementation scheme of the large-scale unlicensed random access method for low-orbit satellite Internet of Things is described in detail below, which includes the following steps:

[0049] S1. For each IoT device d among all K IoT devices k , the assigned leading sequence is s k , through a tensor signal construction method, the IoT device d k The leading sequence T in tensor form k , k=1,...K, where K is the total number of IoT devices; at the beginning of each time slot, all activated IoT devices simultaneously send a preamble sequence in tensor form to the low-orbit satellite, and the signal received by the low-orbit satellite is Y.

[0050] In this embodiment, the tensor signal construction method described in step S1 is:

[0051] For each IoT device’s preamble sequence s k , whose length is L, and it is evenly divided into d segments (d is generally 3), where the i-th segment is a k,i , i = 1, 2...d; then construct the transmission signal according to the following formula: Among them, τ i Indicates a k,i The dimension of is i=1, 2, ..., d, and the symbol ° represents the vector outer product.

[0052] S2, based on the received signal Y, the low-orbit satellite restores the signal Y into the device state matrix M according to a tensor-based Bayesian learning method X , where M X The kth column is the IoT device d k Status information of , k=1,...K;

[0053] In this embodiment, the tensor-based Bayesian learning method described in step S2 specifically comprises the following steps:

[0054] a) Input the leading sequence signal Y received by the satellite and the number of iterations T, and initialize the current iteration round t = 0;

[0055] b) According to the following formula, the median matrix of the channel corresponding to the current iteration round is and the covariance matrix Update the parameters:

[0056]

[0057]

[0058] Among them, F β represents the expectation of the noise factor β, A i Indicates that all a k,i The tensor factor matrix composed of A i The kth column of k,i , k = 1, 2...K; y(d+1) represents the matrix of the tensor signal Y expanded in mode d+1; Represents parameter v k Expectation of, k = 1, 2, ... K; (·) -1 represents matrix inversion; represents the mean estimation parameter μ k The first moment of the mean estimate parameter μ k is a hyperparameter; M represents the number of antennas of the low-orbit satellite, diag() is a diagonal matrix, ⊙ represents the Hadmard product, ◇ represents the Khatri-rao product, the superscript H represents the conjugate transpose, and the superscript * represents the conjugate operation; it should be noted that when t = 0, there is no previous iteration round, so the above update formula and The initial preset parameters need to be adopted;

[0059] c) Based on the median matrix of the channel obtained in the previous iteration and the covariance matrix The variance of the channel distribution is estimated according to the following formula: and Update the parameters:

[0060]

[0061]

[0062] Among them, δ is the initial value, which can generally be set to 10 -6 ; represents the mean estimation parameter μ k The first moment of represents the mean estimation parameter μ k The second moment of Represents the median matrix The kth column of M X (m, k) represents the median matrix The element in the mth row and kth column of Represents the covariance matrix The element at row k and column k;

[0063] d) Estimate the mean parameter of the channel distribution according to the following formula and Update the parameters:

[0064]

[0065]

[0066] Re-order By and Get the hyperparameter μ k The mean and the second moment for:

[0067]

[0068]

[0069] Where Gamma[] represents the gamma function, H[] represents the confluent hypergeometric function 1F1;

[0070] e) The expected F of the noise factor β according to the following formula β Update the parameters:

[0071]

[0072]

[0073]

[0074] Among them, τ i Indicates A i The row dimension of , Tr() means finding the trace, Indicates finding the F norm of the matrix;

[0075] f) After adding 1 to the current iteration round t, repeat steps b) to e) for the next iteration. After a total of T iteration rounds, that is, after completing t = T-1 iteration rounds, the median matrix of the channel is finally output. As the device state matrix M X .

[0076] S3, low-orbit satellite based on the restored device state matrix M X ,Using a threshold judgment method to detect the IoT devices in the activated state;

[0077] In this embodiment, the threshold determination method described in step S3 is:

[0078] Set threshold θ0 = M*(0.15*α) 2 , where α is M X The largest element in; traverse the median matrix M in turn X ∈C M×K For each column in the median matrix M X ∈C M×K Any k-th column M in X (:,k), if the norm IoT devices k is determined to be in an activated state. Otherwise, IoT devices k Determined to be inactive.

[0079] S4. For each IoT device detected as being in an activated state, obtain corresponding channel state estimation information using a channel estimation method.

[0080] In this embodiment, the channel estimation method described in step S4 is:

[0081] If the IoT device k If it is detected as an activated device, the low-orbit satellite will be based on the device state matrix M X , extract M X The kth column data is used as the IoT device d k Estimated information of the channel state

[0082] Computer simulations show that Figure 2 As shown, the large-scale unlicensed random access method for low-orbit satellite Internet of Things proposed in the above embodiment of the present invention, as the number of algorithm iterations increases, its channel normalized error NMSE gradually decreases and tends to be stable, so the channel state information of the activated device can be better obtained. In addition, Figure 3 It shows that as the signal-to-noise ratio increases, the error probability of activating device detection gradually decreases. Therefore, the large-scale unlicensed random access method proposed in the present invention can provide an efficient large-scale device access method for low-orbit satellite Internet of Things communication systems.

Claims

1. A large-scale unlicensed random access method for low-orbit satellite Internet of Things, characterized in that The steps include: S1. For each IoT device d among all K IoT devices k , the assigned leading sequence is s k , through a tensor signal construction method, the IoT device d k The leading sequence T in tensor form k , k = 1, ... K, where K is the total number of IoT devices; at the beginning of each time slot, all activated IoT devices simultaneously send a preamble sequence in the form of a tensor to the low-orbit satellite, and the signal received by the low-orbit satellite is Y; S2, based on the received signal Y, the low-orbit satellite restores the signal Y into the device state matrix M according to a tensor-based Bayesian learning method X , where M X The kth column is the IoT device d k Status information, k = 1, ... K; S3, low-orbit satellite based on the restored device state matrix M X ,Using a threshold judgment method to detect the IoT devices in the activated state; S4. For each IoT device detected as being in an activated state, obtain corresponding channel state estimation information using a channel estimation method; In step S1, the tensor signal construction method is: For each IoT device’s preamble sequence s k , whose length is L, and it is evenly divided into d segments, where the i-th segment is a k,i , i = 1, 2…d; then construct the transmission signal according to the following formula: Among them, τ i Indicates a k,i The dimension, i = 1, 2, ..., d, symbol represents the vector outer product; In step S2, the tensor-based Bayesian learning method specifically comprises the following steps: a) Input the leading sequence signal Y received by the satellite and the number of iterations T, and initialize the current iteration round t = 0; b) According to the following formula, the median matrix of the channel corresponding to the current iteration round is and the covariance matrix Update the parameters: Among them, E β represents the expectation of the noise factor β, A i Indicates that all a k,i The tensor factor matrix composed of A i The kth column of k,i , k = 1, 2…K; y(d+1) represents the matrix of the tensor signal Y expanded in mode d+1; Represents parameter v k Expectation of, k = 1, 2, ... K; (·) -1 represents matrix inversion; represents the mean estimation parameter μ k The first moment of the mean estimate parameter μ k is a hyperparameter; M represents the number of antennas, diag() is a diagonal matrix, ⊙ represents the Hadmard product, ◇ represents the Khatri-rao product, the superscript H represents the conjugate transpose, and the superscript * represents the conjugate operation; and is the initial preset parameter; c) Based on the median matrix of the channel obtained in the previous iteration round and the covariance matrix The variance of the channel distribution is estimated according to the following formula: and Update the parameters: Among them, δ is the initial value, represents the mean estimation parameter μ k The first moment of represents the mean estimation parameter μ k The second moment of Represents the median matrix The kth column of M X (m,k) represents the median matrix The element in the mth row and kth column of Represents the covariance matrix The element at row k and column k; d) Estimate the mean parameter of the channel distribution according to the following formula and Update the parameters: Re-order By and Get the hyperparameter μ k The mean and the second moment for: Among them, Gamma[] represents the gamma function, H[] represents the confluent hypergeometric function 1F1; e) The expected noise coefficient β according to the following formula E β Update the parameters: Among them, τ i Indicates A i The row dimension of , Tr() means finding the trace, Indicates finding the F norm of the matrix; f) After adding 1 to the current iteration round t, repeat steps b) to e) for the next iteration. After a total of T iteration rounds, the median matrix of the channel is finally output. As the device state matrix M X ; In step S3, the threshold determination method is: Set threshold θ0 = M*(0.15*α) 2 , where α is M X The largest element in; traverse the median matrix M in turn X ∈C M×K For each column in the median matrix M X ∈C M×K Any k-th column M in X (:,k), if the norm IoT devices k is determined to be in an activated state. Otherwise, IoT devices k Determined to be inactive; In step S4, the channel estimation method is: If the IoT device k If it is detected as an activated device, the low-orbit satellite will be based on the device state matrix M X , extract M X The kth column data is used as the IoT device d k Estimated information of the channel state 2. The large-scale unlicensed random access method for low-orbit satellite Internet of Things according to claim 1 is characterized in that: The d is set to 3.

3. The large-scale unlicensed random access method for low-orbit satellite Internet of Things according to claim 1 is characterized in that: The initial value δ is set to 10 -6 .