Machine type communication oriented grant-free random multiple access method and apparatus

Abstract

The application discloses a machine type communication oriented grant-free random multiple access method, and specific steps are as follows: step 1, in a transmitting end, scrambling transmission of a user k corresponding information vector with a length of D to obtain a receiving signal y; step 2, in a receiving end, performing channel estimation, adaptive multi-user detection and decoding processing on the receiving signal y obtained in step 1. The method significantly improves detection precision. The application further provides a machine type communication oriented grant-free random multiple access method device.

Description

Technical Field

[0001] This invention belongs to the field of communication technology, specifically relating to an unlicensed random multiple access method for machine-type communication, and also to an unlicensed random multiple access device for machine-type communication. Background Technology

[0002] Massive machine-type communications (mMTC), as one of the most important application scenarios in sixth-generation (6G) wireless communication, aims to realize the Internet of Things, such as smart cities, smart healthcare, and smart homes. In future 6G communication systems, the potential number of MTC devices will reach 10 per square kilometer. 7 This inevitably poses a significant challenge to random access. One solution for effectively utilizing spectrum resources is unlicensed random multiple access (GF-RA), which can reduce communication latency by simplifying the handshake process between MTC devices and base stations (BS). GF-RA allows MTC devices to transmit pilots and data on available time-frequency resource blocks. However, due to the lack of interaction between MTC devices and base stations, the base station inevitably cannot obtain specific information about active MTC devices. Therefore, the GF-RA scheme faces the following three problems: identification of active MTC devices, estimation of channel state information between active MTC devices and base stations, and recovery of data transmitted by active MTC devices.

[0003] In traditional GF-RA schemes, pilot sequences are typically inserted to estimate the channel and identify active devices. However, when the number of MTC devices is large, a large number of pilot sequences are needed to accurately distinguish users and estimate the channel in order to ensure the accuracy of active device identification. A large number of pilot sequences will result in significant communication resource consumption. To reduce the overhead of pilot sequences, researchers are actively seeking pilot-free transmission schemes. However, when pilot signals are lacking, the activity detection, channel estimation, and data recovery of MTC devices will face significant challenges. To address this issue, existing research relies on some dedicated signature information for blind channel estimation and decoding. In his paper "Blind multiple user detection for grant-free MUSA without reference signal," Z. Yuan designed a minimum mean square error based successive interference (MMSE-SIC) algorithm, transforming the multi-user channel into a single-user channel estimation problem. In his paper "Joint user identification channel estimation, and signal detection for grant-free NOMA," S. Jiang decomposes the problem of device activity detection, channel estimation, and data recovery into two parts: slot multi-user detection (SMD) and combined signal and channel estimation (CSCE). The first part is solved using an approximate message passing (AMP) algorithm, and the second part is solved using a message passing algorithm based on structured Gaussian mixtures. However, this method has high implementation complexity. For low-density signature orthogonal frequency division multiplexing (LDS-OFDM) systems, Y. Zhang, in his paper "Bayesian receiver design for grant-free NOMA with messagepassing based structured signal estimation," proposes a message-passing Bayesian receiver and performs channel estimation and decoding on a sparse factor graph.The aforementioned pilotless scheme relies on sparse spreading sequences as user-specific signatures for channel estimation and decoding. However, as the number of potential users increases, the length of the spreading sequence will correspondingly increase to ensure reliable communication, which also leads to low communication efficiency. Summary of the Invention

[0004] The first objective of this invention is to provide an unlicensed random multiple access method for machine-type communication, which significantly improves detection accuracy.

[0005] A second objective of this invention is to provide an apparatus for an unlicensed random multiple access method for machine-type communication.

[0006] The first technical solution adopted in this invention is an unlicensed random multiple access method for machine-type communication, the specific steps of which are as follows:

[0007] Step 1: At the transmitting end, the information vector of length D corresponding to user k is scrambled and transmitted to obtain the received signal y;

[0008] Step 2: At the receiving end, perform channel estimation, adaptive multi-user detection, and decoding on the received signal y obtained in Step 1.

[0009] The invention is further characterized in that,

[0010] Step 1 is implemented in the following steps:

[0011] Step 1.1: Extract the information vector u of length D corresponding to user k. k Input encoder b k Generate a codeword c of length N using encoding. k =(c k,1 ,c k,2 ,c k,i ,…,c k,N ),c k,i ∈{0,1};

[0012] Step 1.2, Code word c k After BPSK modulation, it becomes x k =(x k,1 ,x k,2 ,x k,i ,…x k,N ),x k,i ∈{+1,-1};

[0013] Step 1.3, for x k Phase scrambling generation in Phase scrambling vector θ k,i It is uniformly distributed on [0, 2π].

[0014] Step 1.4, x k After passing through the Rayleigh channel, the received signal y = (y1, y2, y3) is obtained. i …,y N ), where y i Specifically, it is expressed as follows:

[0015]

[0016] Among them, h k Let α be the fading coefficient of the Rayleigh channel. k This indicates the user's activity level. If the user is active, α... k =1, otherwise α k =0; n i It is complex Gaussian noise; β will be used. k =α k h k Channel state information representing the user.

[0017] Step 2 is implemented in the following steps:

[0018] Step 2.1: In the first iteration, based on the received signal y, derive the user-specific likelihood function of the user channel information (CSI) using the phase rotation vector. And maximize this likelihood function to obtain the user

[0019] Step 2.2: Based on the estimated user channel information The AMUD module first restores the active state of each user; it then finds a threshold pair (γ, γ′) if... and If the user is detected as inactive, then the user is detected as active.

[0020] When user k is detected to be active, based on the currently estimated user channel information AMUD calculates x k,i Output information L a (x k,i The specific formula is as follows:

[0021]

[0022] Step 2.3: Based on the output information L a (x k,i Decode and output the feedback information L. e (x k,i This serves as prior information for the next iteration, at which point the first iteration is complete;

[0023] Step 2.4, at the start of the second iteration, based on the received signal y iFeedback information from DEC L e (x k,i ), using equation (3) to update And the user channel information of user k is estimated using equation (4);

[0024] Step 2.5: The AMUD module, based on the estimated user channel information... Feedback information from DEC L e (x k,i First, the active state of each user is restored, and then the soft estimate L of the symbol of each active user is calculated according to equation (5). a (x k,i );

[0025] Step 2.6: Based on the output information L a (x k,i Decode, recover the user's sent data, and output the L e (x k,i This serves as prior information for the next iteration; after reaching the preset number of iterations, the L output by DEC will be used as prior information. e (x k,i Hard decision is performed to obtain the estimated message vector.

[0026] Step 2.1 is implemented according to the following steps:

[0027] Step 2.1.1: In the first iteration, the received signal y... i Rewritten as in It is the equivalent disturbance of user k; according to the central limit theorem, ζ k,i The mean is considered to be E[ζ] k,i ], with variance Var[ζ k,i The complex Gaussian variable; given β k and x k,i y i The probability density function p(y) i |β k ,x k,i )for:

[0028]

[0029] Assume Var[Re(ζ)] k,i )]=Var[Im(ζ k,i )];

[0030] Therefore, we get

[0031] Where, let Pr(x) k，i= 1 / 2; When in a memoryless channel, given y, β k log-likelihood function As shown in the following formula:

[0032]

[0033] Step 2.1.2: Estimate the CSI of user k using maximum a posteriori (MAP), as shown in the following formula:

[0034]

[0035] Use the standard gradient descent method to find The solution obtained from the maximum likelihood estimation yields the user channel information (CSI).

[0036] The second technical solution adopted in this invention is an unlicensed random multiple access device for machine-type communication, comprising a transmitter and a receiver;

[0037] The transmitter scrambles and transmits the information vector of length D corresponding to user k to obtain the received signal y.

[0038] The receiver performs channel estimation, adaptive multi-user detection, and decoding on the received signal y.

[0039] The invention is further characterized in that,

[0040] The receiver consists of a channel estimation module, an adaptive multi-user detection module, and a decoder module;

[0041] The channel estimation module performs channel estimation on the received signal y;

[0042] The adaptive multi-user detection module performs adaptive multi-user detection on the received signal y.

[0043] The decoder module decodes the received signal y.

[0044] The beneficial effects of this invention are:

[0045] This invention presents an unlicensed random multiple access method for machine-type communication that does not require additional transmission power or bandwidth overhead at the transmitting end. Signal scrambling acts as a user-specific signature. Compared to traditional schemes that require pilot signals and incur additional overhead, the signal scrambling scheme proposed in this invention does not rely on sparse extension and can be applied to more general GF-RA systems. Simultaneously, this invention proposes a joint user activity detection, channel estimation, and multi-user decoding method based on signal scrambling at the receiving end. The soft-information iterative algorithm further improves the reliability of channel estimation and multi-user decoding, and the two-dimensional threshold method detects user activity, thereby significantly improving detection accuracy. The blind channel estimation and multi-user detection scheme in this invention also works well without requiring a known number of active users. Attached Figure Description

[0046] Figure 1 This is a block diagram of signal scrambling transmission in a user K random multiple access system according to the method of the present invention;

[0047] Figure 2 This is a block diagram of the SS-JACD receiver in the user K random multiple access system according to the method of the present invention;

[0048] Figure 3 This invention employs the BF algorithm and GDA algorithm to measure MSE under different numbers of users;

[0049] Figure 4 The BER of the method and pilot-assisted scheme of the present invention under different numbers of users. Detailed Implementation

[0050] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0051] This invention provides an unlicensed random multiple access method for machine-type communication, such as... Figure 1 As shown, Figure 1 A block diagram of signal scrambling transmission in a K-user random multiple access system is given, and the specific steps are as follows:

[0052] Step 1: At the transmitting end, the information vector of length D corresponding to user k is scrambled and transmitted to obtain the received signal y; at the transmitting end, the probability of each user being in an active state is α, and the probability of being in an inactive state is 1-α.

[0053] Step 1 is implemented in the following steps:

[0054] Step 1.1: Extract the information vector u of length D corresponding to user k. k Input encoder b k Generate a codeword c of length N using encoding. k =(ck,1 ,c k,2 ,c k,i ,…,c k,N ),c k,i ∈{0,1};

[0055] Step 1.2, Code word c k After BPSK modulation, it becomes x k =(x k,1 ,x k,2 ,x k,i ,…x k,N ),x k,i ∈{+1,-1};

[0056] Step 1.3, for x k Phase scrambling generation in Phase scrambling vector θ k,i It is uniformly distributed on [0, 2π].

[0057] Step 1.4, x k After passing through the Rayleigh channel, the received signal y = (y1, y2, y3) is obtained. i …,y N ), where y i Specifically, it is expressed as follows:

[0058]

[0059] Among them, h k Let α be the fading coefficient of the Rayleigh channel. k This indicates the user's activity level. If the user is active, α... k =1, otherwise α k =0; n i It is complex Gaussian noise; β will be used. k =α k h k Channel State Information (CSI) representing the user.

[0060] Step 2: At the receiving end, perform channel estimation, adaptive multi-user detection, and decoding on the received signal y obtained in Step 1.

[0061] Step 2 is implemented in the following steps, such as... Figure 2 As shown:

[0062] Step 2.1: In the first iteration, based on the received signal y, derive the user-specific likelihood function of the user channel information using the phase rotation vector. And maximize this likelihood function to obtain the user

[0063] Step 2.1.1: In the first iteration, the received signal y... i Rewritten as in It is the equivalent disturbance of user k; according to the central limit theorem, ζ k,i The mean is considered to be E[ζ] k,i ], with variance Var[ζ k,i The complex Gaussian variable; given β k and x k,i y i The probability density function p(y) i |β k ,x k,i )for:

[0064]

[0065] Assume Var[Re(ζ)] k,i )]=Var[Im(ζ k,i )];

[0066] Therefore, we get

[0067] In this case, due to the lack of prior information about the codewords, let Pr(x) k，i = 1 / 2; When in a memoryless channel, given y, β k log-likelihood function As shown in the following formula:

[0068]

[0069] Step 2.1.2: Estimate the CSI of user k using maximum a posteriori (MAP), as shown in the following formula:

[0070]

[0071] Use standard gradient descent (GDA) to find The solution obtained from the maximum likelihood estimation yields the user channel information.

[0072] Figure 3 The results show the mean square error (MSE) of the channel estimates obtained using brute-force search and GDA, employing ARA codes with a rate of 1 / 2 and repeating codes with a rate of 1 / 10 as the channel coding for the user, with a code length N = 1000. For large-scale user systems, the difference in estimation accuracy between the two search algorithms is very small. Therefore, considering implementation efficiency, we adopt the GDA algorithm as our search algorithm.

[0073] Step 2.2: Based on the estimated user channel information First, restore the active status of each user;

[0074] Existing solutions use user channel information The user's activity level is determined by comparing it with γ, which serves as a one-dimensional detection threshold; that is, if... Users are detected as inactive otherwise. However, we found that the maximum likelihood value in equation (4) is also related to the user's activity level; active users typically have a larger maximum likelihood value than inactive users. Therefore, we propose a two-dimensional threshold detection method, which finds a threshold pair (γ, γ′) if... and If the user is detected as inactive, then the user will be detected as active; otherwise, the user will be detected as active.

[0075] When user k is detected to be active, based on the currently estimated user channel information AMUD calculates x k,i Output information L a (x k,i The specific formula is as follows:

[0076]

[0077] Step 2.3: Based on the output information L from step 2.2 a (x k,i Decode and output the feedback information L. e (x k,i The information is fed back to the channel estimation module and AMUD module as prior information for the next iteration, at which point the first iteration is complete;

[0078] Step 2.4, at the start of the second iteration, based on the received signal y i Feedback information from DEC L e (x k,i The channel estimation module updates using equation (3). And the user channel information of user k is estimated using equation (4);

[0079] Step 2.5: Based on the estimated user channel information Feedback information from DEC L e (x k,i First, the active state of each user is restored, and then the soft estimate L of the symbol of each active user is calculated according to equation (5). a (x k,i );

[0080] Step 2.6: Based on the output information L a (x k,i Decode, recover the user's sent data, and output the Le (x k,i This serves as prior information for the next iteration; after reaching the preset number of iterations, the output L will be used as prior information. e (x k,i Hard decision is performed to obtain the estimated message vector.

[0081] Figure 4 A comparison of the BER performance of the pilot-assisted scheme and the scheme proposed in this patent is presented, where α = 1 and codeword length N = 1000. In the pilot-assisted scheme, 200 bytes of codeword length are used for pilot transmission, and 800 bytes are used for data transmission. The decoding iterations for K = 1, 20, 35, and 40 are 5, 25, 35, and 40, respectively. As can be seen from the figure, the performance of our proposed scheme is superior to that of the pilot-assisted scheme in all cases.

[0082] Example 1:

[0083] Unlicensed random multiple access methods for machine-type communication, such as Figure 1 As shown, Figure 1 A block diagram of signal scrambling transmission in a K-user random multiple access system is given, and the specific steps are as follows:

[0084] Step 1: At the transmitting end, the information vector of length D corresponding to user k is scrambled and transmitted to obtain the received signal y; at the transmitting end, the probability of each user being in an active state is α, and the probability of being in an inactive state is 1-α.

[0085] Step 1 is implemented in the following steps:

[0086] Step 1.1: Extract the information vector u of length D corresponding to user k. k Input encoder b k Generate a codeword c of length N using encoding. k =(c k,1 ,c k,2 ,c k,i ,…,c k,N ),c k,i ∈{0,1};

[0087] Step 1.2, Code word c k After BPSK modulation, it becomes x k =(x k,1 ,x k,2 ,x k,i ,…x k,N ),x k,i ∈{+1,-1};

[0088] Step 1.3, for x k Phase scrambling generation in Phase scrambling vector θ k,i It is uniformly distributed on [0, 2π].

[0089] Step 1.4, x k After passing through the Rayleigh channel, the received signal y = (y1, y2, y3) is obtained. i …,y N ), where y i Specifically, it is expressed as follows:

[0090]

[0091] Among them, h k Let α be the fading coefficient of the Rayleigh channel. k This indicates the user's activity level. If the user is active, α... k =1, otherwise α k =0; n i It is complex Gaussian noise; β will be used. k =α k h k Channel State Information (CSI) representing the user.

[0092] Step 2: At the receiving end, perform channel estimation, adaptive multi-user detection, and decoding on the received signal y obtained in Step 1.

[0093] Step 2 is implemented in the following steps, such as... Figure 2 As shown:

[0094] Step 2.1: In the first iteration, based on the received signal y, derive the user-specific likelihood function of the user channel information using the phase rotation vector. And maximize this likelihood function to obtain the user

[0095] Step 2.1.1: In the first iteration, the received signal y... i Rewritten as in It is the equivalent disturbance of user k; according to the central limit theorem, ζ k,i The mean is considered to be E[ζ] k,i ], with variance Var[ζ k,i The complex Gaussian variable; given β k and x k,i y i The probability density function p(y) i |β k ,x k,i )for:

[0096]

[0097] Assume Var[Re(ζ)]k,i )]=Var[Im(ζ k,i )];

[0098] Therefore, we get

[0099] In this case, due to the lack of prior information about the codewords, let Pr(x) k，i = 1 / 2; When in a memoryless channel, given y, β k log-likelihood function As shown in the following formula:

[0100]

[0101] Step 2.1.2: Estimate the CSI of user k using maximum a posteriori (MAP), as shown in the following formula:

[0102]

[0103] Use standard gradient descent (GDA) to find The solution obtained from the maximum likelihood estimation yields the user channel information.

[0104] In the method of this invention, the decoding iteration number for K=1 is 5 times;

[0105] Example 2:

[0106] Unlicensed random multiple access methods for machine-type communication, such as Figure 1 As shown, Figure 1 A block diagram of signal scrambling transmission in a K-user random multiple access system is given, and the specific steps are as follows:

[0107] Step 1: At the transmitting end, the information vector of length D corresponding to user k is scrambled and transmitted to obtain the received signal y; at the transmitting end, the probability of each user being in an active state is α, and the probability of being in an inactive state is 1-α.

[0108] Step 1 is implemented in the following steps:

[0109] Step 1.1: Extract the information vector u of length D corresponding to user k. k Input encoder b k Generate a codeword c of length N using encoding. k =(c k,1 ,c k,2 ,c k,i ,…,c k,N ),c k,i ∈{0,1};

[0110] Step 1.2, Code word c k After BPSK modulation, it becomes xk =(x k,1 ,x k,2 ,x k,i ,…x k,N ),x k,i ∈{+1,-1};

[0111] Step 1.3, for x k Phase scrambling generation in Phase scrambling vector θ k,i It is uniformly distributed on [0, 2π].

[0112] Step 1.4, x k After passing through the Rayleigh channel, the received signal y = (y1, y2, y3) is obtained. i …,y N ), where y i Specifically, it is expressed as follows:

[0113]

[0114] Among them, h k Let α be the fading coefficient of the Rayleigh channel. k This indicates the user's activity level. If the user is active, α... k =1, otherwise α k =0; n i It is complex Gaussian noise; β will be used. k =α k h k Channel State Information (CSI) representing the user.

[0115] Step 2: At the receiving end, perform channel estimation, adaptive multi-user detection, and decoding on the received signal y obtained in Step 1.

[0116] Step 2 is implemented in the following steps, such as... Figure 2 As shown:

[0117] Step 2.1: In the first iteration, based on the received signal y, derive the user-specific likelihood function of the user channel information using the phase rotation vector. And maximize this likelihood function to obtain the user

[0118] Step 2.1.1: In the first iteration, the received signal y... i Rewritten as in It is the equivalent disturbance of user k; according to the central limit theorem, ζ k,i The mean is considered to be E[ζ] k,i ], with variance Var[ζ k,iThe complex Gaussian variable; given β k and x k,i y i The probability density function p(y) i |β k ,x k,i )for:

[0119]

[0120] Assume Var[Re(ζ)] k,i )]=Var[Im(ζ k,i )];

[0121] Therefore, we get

[0122] In this case, due to the lack of prior information about the codewords, let Pr(x) k，i = 1 / 2; When in a memoryless channel, given y, β k log-likelihood function As shown in the following formula:

[0123]

[0124] Step 2.1.2: Estimate the CSI of user k using maximum a posteriori (MAP), as shown in the following formula:

[0125]

[0126] Use standard gradient descent (GDA) to find The solution obtained from the maximum likelihood estimation yields the user channel information.

[0127] In the method of this invention, the decoding iteration number for K=20 is 25.

[0128] Example 3:

[0129] Unlicensed random multiple access methods for machine-type communication, such as Figure 1 As shown, Figure 1 A block diagram of signal scrambling transmission in a K-user random multiple access system is given, and the specific steps are as follows:

[0130] Step 1: At the transmitting end, the information vector of length D corresponding to user k is scrambled and transmitted to obtain the received signal y; at the transmitting end, the probability of each user being in an active state is α, and the probability of being in an inactive state is 1-α.

[0131] Step 1 is implemented in the following steps:

[0132] Step 1.1: Extract the information vector u of length D corresponding to user k. kInput encoder b k Generate a codeword c of length N using encoding. k =(c k,1 ,c k,2 ,c k,i ,…,c k,N ),c k,i ∈{0,1};

[0133] Step 1.2, Code word c k After BPSK modulation, it becomes x k =(x k,1 ,x k,2 ,x k,i ,…x k,N ),x k,i ∈{+1,-1};

[0134] Step 1.3, for x k Phase scrambling generation in Phase scrambling vector θ k,i It is uniformly distributed on [0, 2π].

[0135] Step 1.4, x k After passing through the Rayleigh channel, the received signal y = (y1, y2, y3) is obtained. i …,y N ), where y i Specifically, it is expressed as follows:

[0136]

[0137] Among them, h k Let α be the fading coefficient of the Rayleigh channel. k This indicates the user's activity level. If the user is active, α... k =1, otherwise α k =0; n i It is complex Gaussian noise; β will be used. k =α k h k Channel State Information (CSI) representing the user.

[0138] Step 2: At the receiving end, perform channel estimation, adaptive multi-user detection, and decoding on the received signal y obtained in Step 1.

[0139] Step 2 is implemented in the following steps, such as... Figure 2 As shown:

[0140] Step 2.1: In the first iteration, based on the received signal y, derive the user-specific likelihood function of the user channel information using the phase rotation vector. And maximize this likelihood function to obtain the user

[0141] Step 2.1.1: In the first iteration, the received signal y... i Rewritten as in It is the equivalent disturbance of user k; according to the central limit theorem, ζ k,i The mean is considered to be E[ζ] k,i ], with variance Var[ζ k,i The complex Gaussian variable; given β k and x k,i y i The probability density function p(y) i |β k ,x k,i )for:

[0142]

[0143] Assume Var[Re(ζ)] k,i )]=Var[Im(ζ k,i )];

[0144] Therefore, we get

[0145] In this case, due to the lack of prior information about the codewords, let Pr(x) k，i = 1 / 2; When in a memoryless channel, given y, β k log-likelihood function As shown in the following formula:

[0146]

[0147] Step 2.1.2: Estimate the CSI of user k using maximum a posteriori (MAP), as shown in the following formula:

[0148]

[0149] Use standard gradient descent (GDA) to find The solution obtained from the maximum likelihood estimation yields the user channel information.

[0150] In the method of this invention, the decoding iteration number for K=35 is 35 times;

[0151] Example 4:

[0152] Unlicensed random multiple access methods for machine-type communication, such as Figure 1 As shown, Figure 1 A block diagram of signal scrambling transmission in a K-user random multiple access system is given, and the specific steps are as follows:

[0153] Step 1: At the transmitting end, the information vector of length D corresponding to user k is scrambled and transmitted to obtain the received signal y; at the transmitting end, the probability of each user being in an active state is α, and the probability of being in an inactive state is 1-α.

[0154] Step 1 is implemented in the following steps:

[0155] Step 1.1: Extract the information vector u of length D corresponding to user k. k Input encoder b k Generate a codeword c of length N using encoding. k =(c k,1 ,c k,2 ,c k,i ,…,c k,N ),c k,i ∈{0,1};

[0156] Step 1.2, Code word c k After BPSK modulation, it becomes x k =(x k,1 ,x k,2 ,x k,i ,…x k,N ),x k,i ∈{+1,-1};

[0157] Step 1.3, for x k Phase scrambling generation in Phase scrambling vector θ k,i It is uniformly distributed on [0, 2π].

[0158] Step 1.4, x k After passing through the Rayleigh channel, the received signal y = (y1, y2, y3) is obtained. i …,y N ), where y i Specifically, it is expressed as follows:

[0159]

[0160]

[0161] Among them, h k Let α be the fading coefficient of the Rayleigh channel. k This indicates the user's activity level. If the user is active, α... k =1, otherwise α k =0; n i It is complex Gaussian noise; β will be used. k =α k h k Channel State Information (CSI) representing the user.

[0162] Step 2: At the receiving end, perform channel estimation, adaptive multi-user detection, and decoding on the received signal y obtained in Step 1.

[0163] Step 2 is implemented in the following steps, such as... Figure 2 As shown:

[0164] Step 2.1: In the first iteration, based on the received signal y, derive the user-specific likelihood function of the user channel information using the phase rotation vector. And maximize this likelihood function to obtain the user

[0165] Step 2.1.1: In the first iteration, the received signal y... i Rewritten as in It is the equivalent disturbance of user k; according to the central limit theorem, ζ k,i The mean is considered to be E[ζ] k,i ], with variance Var[ζ k,i The complex Gaussian variable; given β k and x k,i y i The probability density function p(y) i |β k ,x k,i )for:

[0166]

[0167] Assume Var[Re(ζ)] k,i )]=Var[Im(ζ k,i )];

[0168] Therefore, we get

[0169] In this case, due to the lack of prior information about the codewords, let Pr(x) k，i = 1 / 2; When in a memoryless channel, given y, β k log-likelihood function As shown in the following formula:

[0170]

[0171] Step 2.1.2: Estimate the CSI of user k using maximum a posteriori (MAP), as shown in the following formula:

[0172]

[0173] Use standard gradient descent (GDA) to find The solution obtained from the maximum likelihood estimation yields the user channel information.

[0174] In the method of this invention, the decoding iteration number for K=40 is 40 times;

[0175] The present invention also provides an unlicensed random multiple access device for machine-type communication, comprising a transmitter and a receiver;

[0176] The transmitter scrambles and transmits the information vector of length D corresponding to user k to obtain the received signal y.

[0177] The receiver performs channel estimation, adaptive multi-user detection, and decoding on the received signal y.

[0178] The receiver consists of a channel estimation module, an adaptive multi-user detection module (AMUD), and a decoder module (DEC).

[0179] The channel estimation module performs channel estimation on the received signal y;

[0180] The adaptive multi-user detection module performs adaptive multi-user detection on the received signal y.

[0181] Specifically, the AMUD module is based on estimated user channel information. Feedback information from DEC L e (x k,i First, the active state of each user is restored, and then the soft estimate L of the symbol of each active user is calculated according to equation (5). a (x k,i );

[0182] The decoder module decodes the received signal y.

[0183] Specifically, DEC uses AMUD's output information L a (x k,i Decode, recover the user's sent data, and output the L e (x k,i The data is fed back to the channel estimation module and the AMUD module as prior information for the next iteration; after reaching the preset number of iterations, the L output by DEC is... e (x k,i Hard decision is performed to obtain the estimated message vector.

Claims

1. A method for unlicensed random multiple access for machine-type communication, characterized in that, The specific steps are as follows: Step 1: At the transmitting end, the information vector of length D corresponding to user k is scrambled and transmitted to obtain the received signal y; Step 1 is implemented in the following steps: Step 1.1: Extract the information vector of length D corresponding to user k. Input encoder Generate codewords of length N using encoding. ; Step 1.2, typing After BPSK modulation ; Step 1.3, for Phase scrambling generation ,in Phase scrambling vector In [0,2 Evenly distributed on the surface; Step 1.4 After passing through the Rayleigh channel, the received signal y=( , , …, ),in Specifically, it is expressed as follows: = + = + in, Let be the fading coefficient of the Rayleigh channel. This indicates the user's activity level. If the user is active, =1, otherwise =0; It is complex Gaussian noise; it will be used Channel state information representing the user; Step 2: At the receiving end, perform channel estimation, adaptive multi-user detection, and decoding on the received signal y obtained in Step 1.

2. The unlicensed random multiple access method for machine-type communication according to claim 1, characterized in that, Step 2 is implemented in the following steps: Step 2.1: In the first iteration, based on the received signal y, derive the user-specific likelihood function of the user channel information using the phase rotation vector. ( And maximize this likelihood function to obtain the user's CSI. ; Step 2.2: Based on the estimated user channel information The AMUD module first restores the active state of each user; then it finds a threshold pair ( ),if and ( ) If the condition is met, the user is detected as inactive; otherwise, the user is detected as active. When user k is detected to be active, based on the currently estimated user channel information AMUD calculation of The specific formula is as follows: (5)； Step 2.3: Based on the output information Decode and output This serves as prior information for the next iteration, at which point the first iteration is complete. Step 2.4, at the start of the second iteration, based on the received signal Feedback information from DEC Update using equation (3) And the user channel information of user k is estimated using equation (4); Step 2.5: The AMUD module, based on the estimated user channel information... Feedback information from DEC First, the active state of each user is restored, and then the soft estimate of the symbol of each active user is calculated according to equation (5). ; Step 2.6: Based on the output information Decode, recover the user's sent data, and output the data. As prior information for the next iteration; After reaching the preset number of iterations, the DEC output will be... Hard decision is performed to obtain the estimated message vector. .

3. The unlicensed random multiple access method for machine-type communication according to claim 1, characterized in that, Step 2.1 is implemented according to the following steps: Step 2.1.1: During the first iteration, the received signal... Rewritten as ,in It is the equivalent disturbance of user k; according to the central limit theorem, The mean is considered to be The variance is Complex Gaussian variables; Given and , probability density function for: (1) Assumption = Therefore, we get (2) Among them, let When in a memoryless channel, given y, log-likelihood function ( As shown in the following formula: + (3) Step 2.1.2: Estimate the CSI of user k using maximum a posteriori (MAP), as shown in the following formula: =angry max ( ),k=1,2,…,K.(4) Use the standard gradient descent method to find The solution obtained from the maximum likelihood estimation yields the user channel information. .

4. An unlicensed random multiple access device for machine-type communication, characterized in that, It includes a transmitter and a receiver; The transmitter scrambles and transmits the information vector of length D corresponding to user k to obtain the received signal y. The transmitter works as follows: First, the information vector of length D corresponding to user k. Input encoder Generate codewords of length N using encoding. ; Then, typing. After BPSK modulation ; After that, to Phase scrambling generation ,in Phase scrambling vector In [0,2 Evenly distributed on the surface; at last, After passing through the Rayleigh channel, the received signal y=( , , …, ),in Specifically, it is expressed as follows: = + = + in, Let be the fading coefficient of the Rayleigh channel. This indicates the user's activity level. If the user is active, =1, otherwise =0; It is complex Gaussian noise; it will be used Channel state information representing the user; The receiver performs channel estimation, adaptive multi-user detection, and decoding on the received signal y.

5. The unlicensed random multiple access device for machine-type communication according to claim 4, characterized in that, The receiver consists of a channel estimation module, an adaptive multi-user detection module, and a decoder module; The channel estimation module performs channel estimation on the received signal y; The adaptive multi-user detection module performs adaptive multi-user detection on the received signal y. The decoder module decodes the received signal y.

Images