Large-scale authorization-free random access method based on CR-NOMA-PD

Abstract

The invention belongs to the technical field of wireless communication, and discloses a large-scale authorization-free random multiple access method based on cognitive radio (CR), non-orthogonal multiple access (NOMA) and preamble delay (PD). The method comprises the following steps: firstly, proposing a three-step grant-free random access protocol based on CR-NOMA-PD, and designing an access channel and power level optimization strategy based on spectrum sensing; then, an uplink transmission scheme is analyzed, and a user transmission signal model and a base station receiving signal model are established; then, a multi-user signal detection algorithm based on channel filtering, power level detection and preamble detection is designed, and joint detection and access conflict detection of multi-user uplink signals are completed; and finally, the random access performance of the user is analyzed, and a mathematical model of the user access rate is established. The method is suitable for a large-scale machine communication scene, has good robustness, can effectively reduce large-scale authorization-free random access conflicts, improves the access success rate of a user, and improves the overload access capability of a wireless network.

Description

Technical Field

[0001] The present invention belongs to the field of wireless communication, and particularly relates to a large-scale grant-free random access method based on CR-NOMA-PD. Background Art

[0002] With the rapid development of the mobile Internet and the Internet of Things, the demand for massive machine type communication (mMTC) is increasing day by day. Compared with human-to-human communication, mMTC presents new characteristics such as a large number of terminals, sporadic and bursty data, and short packet transmission in the uplink, which requires new capabilities of wireless access technologies. However, traditional licensed multiple access technologies such as time division multiple access and code division multiple access rely on the base station for access authorization and do not support random multiple access. Limited by the number of orthogonal time-frequency resources or orthogonal spreading codes, the access carrying capacity of licensed multiple access is limited. In addition, the access protocol signaling overhead of licensed multiple access is large. When applied to mMTC, a large amount of signaling occupies limited uplink transmission resources, exacerbating spectrum congestion, resulting in a low access success rate and large access delay of terminals. Therefore, it is difficult for licensed multiple access technologies to support the efficient access of a large number of terminals, and exploring large-scale grant-free random multiple access (MGFRMA) technology has become the research focus and hotspot in the current field of wireless communication.

[0003] At present, significant progress has been made in MGFRMA research. In terms of protocol optimization, the 5G system has proposed a 2-Step grant-free random access protocol. By eliminating the access authorization process, the access delay and power consumption have been significantly reduced. However, the access competition of this protocol may lead to preamble collisions. For this reason, some relevant literature has proposed various improvement measures such as access category barring (ACB) mechanism, random backoff mechanism, and random power backoff. In the research of MGFRMA based on multiple antennas, aiming at the problems of pilot contamination and access conflict, some literature has studied and proposed a single orthogonal preamble (SOP) multi-cell multiplexing strategy, a concatenated orthogonal preamble (COP) transmission scheme, and a massive preamble-data superposition random access (mPDSRA) scheme. In the research based on non-orthogonal multiple access (NOMA), MGFRMA schemes such as grant-free non-orthogonal multiple access (GF-NOMA), hierarchical GF-NOMA, and combined multi-preamble and NOMA schemes have been proposed.

[0004] The existing MGFRMA schemes have significantly improved the access performance in the mMTC scenario through the combination of protocol optimization, multi-antenna technology, and NOMA technology, providing a technical basis for realizing efficient massive terminal access. However, there are still some deficiencies. For example, the preamble-data superposition random access (mPDSRA) scheme proposed in the literature DOI: 10.1109 / LWC.2020.2974724 improves the spectral efficiency, but requires a strict timing synchronization mechanism and requires the user data packet lengths to be consistent, limiting its applicability in complex network environments. Although the patent CN202310315163.8 introduces a channel interference power awareness mechanism, it adopts a single preamble fixed time slot transmission scheme, and the number of preambles has a great restriction on the random access performance of users.

[0005] In summary, a lot of research has been carried out on large-scale grant-free random multiple access, but the performance of the existing schemes is still not ideal, and there is still a large room for improvement in the random access performance of users. Summary of the Invention

[0006] Combining the technical advantages of spectrum sensing, non-orthogonal multiple access, and preamble delay, the present invention proposes a large-scale grant-free random access scheme based on CR-NOMA and preamble delay (CR-NOMA and Preamble Delay Based Massive Grant Free Random Multiple Access, CR-NOMA-PDMGFRMA). It allows end-users to obtain the occupancy status of the current channel by sensing the interference power level of the channel, thereby preferentially selecting the uplink access channel and transmission power level. At the same time, the preamble is randomly delayed for transmission to form a delayed preamble, reducing the preamble collision of users, thereby improving the user random access rate and enhancing the access bearing capacity of the wireless network for a large number of terminals.

[0007] The cognitive-assisted large-scale terminal grant-free random access method of the present invention comprises the following steps:

[0008] Step 1: Active users sense the interference power of the channel within the sensing time slot to preferentially select the access channel and uplink transmission power. Active users independently sense the interference power levels of all uplink channels, select the channel with the lowest and minimum interference power level below the threshold as the uplink access channel, and at the same time select a power level from the power level set as the uplink transmission power level according to the power level selection criterion.

[0009] Step 2: Active users perform contention access and data transmission within the transmission time slot: First, randomly select a preamble sequence from the preamble set to spread-spectrum the service data to form spread-spectrum data; then, the active user randomly delays a number of preamble sub-slots to transmit the preamble sequence to the base station through the selected channel at the selected power level, thereby forming a delayed preamble; finally, the active user continues to transmit the spread-spectrum data to the base station in the data sub-slot.

[0010] Step 3: Active users receive the access result feedback by the base station within the feedback waiting time slot. The base station uses band filtering to obtain the received signals on each channel, and uses successive interference cancellation (SIC) technology to detect the signals at each power level; perform preamble detection, data detection, and collision detection on the signals at each power level, recover the uplink signals of the users and obtain the access contention result, and then feedback the result to the users through the access response.

[0011] The beneficial effects of the present invention are as follows: (1) The random access scheme based on CR-NOMA-PD (Cognitive Radio and Non-Orthogonal Multiple Access combined with Preamble Delay) proposed by the present invention significantly improves the user access probability and the overload capacity of the system; (2) By optimizing channel selection through spectrum sensing and combining random delay preambles to expand the access space, the probability of user conflict is reduced; (3) Adopting multi-power level transmission and power domain interference cancellation technology to achieve efficient multi-user signal detection and separation, and still having a high access success rate and signal detection robustness even in high-load or interference environments; (4) This method greatly reduces the access delay and signaling overhead through the license-free access process, improving the spectrum utilization efficiency and resource allocation flexibility. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 FIG. is the system model diagram in an embodiment of the present invention.

[0013] Figure 2 FIG. is the time frame structure diagram of a user in an embodiment of the present invention.

[0014] Figure 3 FIG. is the schematic diagram of the access transmission protocol in an embodiment of the present invention.

[0015] Figure 4 FIG. is the implementation flow block diagram of the method in an embodiment of the present invention.

[0016] Figure 5 FIG. is the relationship curve diagram between the access probability of the terminal user and the number of users under different preamble numbers in the simulation result of the method in an embodiment of the present invention.

[0017] Figure 6 FIG. is the schematic diagram of the influence of the channel and the number of users on the user access probability in the simulation result of the method in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings of the specification.

[0019] Set Figure 1 the system scenario shown:

[0020] Consider a single-cell mMTC system scenario: The base station is equipped with M antennas, and N single-antenna users are evenly distributed in the cell. The user label set is The length of the single burst service data of each user is v bits. There are a total of N c uplink channels (channels) and K orthogonal preamble sequences of length J to serve the uplink random access of users, where K ≤ J. Let the uplink channel set and the preamble sequence set be and s kDenote the k-th preamble sequence, Denote the set of preamble labels. Different preamble sequences are orthogonal to each other and have an energy of 1, that is s k′ ∈SC, for k≠k′, there is Define the preamble matrix S = [s 1 ,…,s K T , SS H = I K , where S H denotes the Hermitian matrix of matrix S, and I K denotes the K-order identity matrix, denotes the complex number field of K rows and J columns.

[0021] Consider that the terminal (user) has the ability of spectrum sensing. Each active user performs spectrum sensing before accessing the transmission to obtain the channel state information and designs the Figure 2 user access transmission time frame structure shown as follows:

[0022] The access transmission time frame includes three parts: a sensing time slot, a transmission time slot, and a feedback waiting time slot, where the sensing time slot is T s , the transmission time slot is T t , and the feedback waiting time slot is T w ; the transmission time slot is further divided into N s preamble sub-slots and 1 data sub-slot, denotes the set of preamble sub-slot sequence numbers. In the sensing time slot, the active user uses a broadband sensor to sense the interference power levels of all N c uplink channels, and selects the channel with the interference power not higher than the threshold and the smallest as the uplink access channel. In the transmission time slot, the active user selects a preamble sequence from K preambles according to the access protocol, spreads the service data with the selected preamble sequence, and sequentially transmits the preamble sequence and the service data to the base station in a randomly selected preamble sub-slot and the subsequent data sub-slot through the selected channel. In the feedback time slot, the active user waits for the access feedback message from the base station to confirm whether retransmission of data is required.

[0023] Combined with Figure 1 the system model, Figure 2 the time frame structure, Figure 3 the schematic of the transmission protocol, and Figure 4 the random access process, the implementation steps of this method are specifically described as follows:

[0024] ​Step 1. The active user senses the interference power of the channel within the sensing time slot to preferentially access the channel and the uplink transmission power. The active user independently senses the interference power levels of all uplink channels, selects the channel with the lowest interference power level below the threshold as the uplink access channel, and simultaneously selects a power level from the power level set as the uplink transmission power level according to the power level selection criterion.

[0025] The steps for the active user to select the channel to be accessed are as follows:

[0026] Step 1a. Define the uplink transmission power level set and divide the channel interference power level sections.

[0027] Let the uplink transmission power level set of the active user be Q = {q 1 , …, q L}, this set contains L power levels, where q l represents the l-th power level, I q represents the label set of the transmission power level; the channel interference power levels are also evenly divided into L sections. Let and represent the starting value and the ending value of the i-th section respectively, i = 1, 2, …, L. It is stipulated that where represents the interference power threshold of the accessible channel; j ∈ I q , i < j, the power levels q i and q j in the set Q satisfy q i < q j , and the ending value of the i-th interference power level section is also less than the ending value of the j-th interference power level section, that is

[0028] Step 1b. The active user selects the channel to be accessed according to the interference power level.

[0029] Let the active user be n, n ∈ I a senses the interference power levels of all N c uplink channels as where represents the interference power level sensed by user n on channel u; the active user n selects the channel with the currently lowest interference power level as the best access channel from the channels that satisfy the interference power threshold, that is . The uplink access channel of the active user n is calculated as:

[0030]

[0031] If an active user perceives that the interference power levels on all current channels are higher than the interference power threshold the active user delays for one frame time and then retries to access.

[0032] Step 1c: The active user selects an uplink transmission power level.

[0033] According to the interference power level of the uplink access channel selected in Step 1b the uplink optimal transmission power level q is determined based on the interference interval to which it belongs i,n and the optimal transmission power level serial number l * is calculated as:

[0034]

[0035] where l * ∈I q ; represents the i-th interference section, and represent the starting value and the ending value of the i-th interference section respectively, and are calculated as If it is stipulated that l * = 1.

[0036] Step 2: The active user performs contention access and data transmission within the transmission time slot: First, a preamble sequence is randomly selected from the preamble set to spread-spectrum the service data to form spread-spectrum data; then, the active user transmits the preamble sequence to the base station through the selected channel at the selected power level with a random delay of several preamble sub-slots to form a delayed preamble; finally, the active user continues to transmit the spread-spectrum data to the base station in the data sub-slot.

[0037] The specific steps of the delayed preamble and data transmission are as follows:

[0038] Step 2a-1: The active user randomly selects a preamble sequence from K preambles and simultaneously randomly selects a sub-slot from the preamble sub-slots to generate a delayed preamble signal.

[0039] Considering that user n selects channel as the uplink channel, selects the power level q l,n as the transmission power level, selects the c n -th preamble to participate in the access contention and transmits the preamble sequence in the preamble sub-slot t n , c n ∈L p , n∈I a , t n ∈L i . Thus, the preamble selection vector a of user nn and the preamble time vector t n are respectively expressed as and where 1 [x] is an indicator function, which is 1 when the logical expression in the square brackets [], i.e., the condition x, holds [x] = 1, otherwise 1 [x] = 0. Thus, the preamble sequence p n of user n and the delayed preamble signal c n can be expressed as

[0040]

[0041] where and ||p n || 2 = 1; P = JN s represents the length of the delayed preamble signal.

[0042] Step 2a - 2, the active user spreads its v - bit service data using the selected preamble sequence to generate a spread - spectrum sequence of length D = vJ.

[0043] Let the service data of active user n be d n = [d 1 , d 2 ,... d v , and the data x n after spreading it using p n is expressed as

[0044] x n = [x 1 , x 2 ,... x D = [d 1 p n , d 2 p n ,... d v p n (5)

[0045] where ||x n || 2 = 1, D = vJ;

[0046] Step 2a - 3, the active user sequentially transmits the delayed preamble and the spread - spectrum data to the base station through the selected uplink channel at the selected power level.

[0047] Combining Equation (2) and Equation (3), the signal z n transmitted by active user n in one frame in the time domain is:

[0048]

[0049] Let denote the channel selection vector of active user n, and a frame of signal F transmitted by active user n in the two-dimensional time-frequency space n is expressed as

[0050]

[0051] where denotes the transmission signal of active user n on channel u;

[0052] Step 3: The active user receives the access result feedback by the base station during the feedback waiting time slot. The base station uses band filtering to obtain the received signals on each channel, and uses successive interference cancellation (SIC) technology to detect the signals at each power level; performs preamble detection, data detection, and collision detection on the signals at each power level, recovers the uplink signals of the users and obtains the access competition result, and then feeds back the result to the users through the access response.

[0053] The specific detection steps are as follows:

[0054] 1) The base station performs power level detection on the received signals.

[0055] Step 3a-1, the base station receives the sum signal from N c uplink channels.

[0056] Considering that the uplink channels from the active users to the base station experience Rayleigh fading and additive white Gaussian noise, and the channels are quasi-static, that is, the channel attenuation coefficients and additive noise remain unchanged within one frame time; a frame of signal G received by the base station is the sum signal of N users, which is expressed as:

[0057]

[0058] where F n = z n u n denotes a frame of signal of active user n; denotes the channel attenuation coefficient matrix of N c uplink channels from user n to the base station; denotes the attenuation coefficient vector of N c uplink channels from user n to base station antenna m, where denotes the attenuation coefficient of the u-th uplink channel from user n to base station antenna m, which follows a Gaussian distribution with a mean of 0 and a variance of 1; N denotes the channel additive white Gaussian noise matrix, Since the channel is quasi-static, each column of N is the same and follows a Gaussian distribution with a mean of 0 and a variance of σ 2 ; the channel attenuation coefficient matrix H n remains constant within one frame time and has a quasi-orthogonal property, that is, for the uplink channel attenuation coefficient matrices H n and H n′ of any two users n and n′, there is

[0059]

[0060] where represents the conjugate transpose matrix of H n , and M represents the number of base station antennas.

[0061] Step 3a-2, the base station performs band-pass filtering on the received uplink signal to obtain the received signals of N c uplink channels.

[0062] The base station received signal G is a broadband signal containing N c frequency bands; performing N c channel band-pass filtering on the received signal G to obtain the received signals of N c uplink channels; let the received signal of the base station on the uplink channel u be G u , which can be expressed as

[0063]

[0064] where represents the access channel selected by user n; N u represents the Gaussian white noise matrix on channel u, which follows a Gaussian distribution with a mean of 0 and a variance of σ 2 I M ; represents the channel attenuation coefficient vector from user n to the base station on channel u, which is the u-th column vector of the matrix H n ; represents the set of users selecting channel u. Introducing to represent the attenuation coefficients from N users to the base station on channel u; to represent the transmitted signals of N users on channel u, Equation (10) is rewritten as

[0065]

[0066] Step 3a-3, using successive interference cancellation, i.e., SIC technology, to detect the power level signals of each received signal G u on each uplink channel in descending order of power level

[0067] The base station detects each power level q in descending order l of the signals carried thereon Calculate the power level signal for the signal-to-interference-plus-noise ratio That is

[0068]

[0069] wherein represents the average power of the signal and is calculated as

[0070]

[0071] Here, trace() represents the trace of the matrix; and respectively represent the sets of users carried by the power levels q l and q j on channel u, q j <q l ; N u,i and N u,j respectively represent the number of elements in the sets and i.e., the number of users carried by the power levels q l and q j on channel u,

[0072] Judge whether it is greater than the detection threshold of the base station receiver, i.e., the receiving sensitivity That is whether it holds.

[0073] If not, the signals carried by the power levels q l , q l-1 , …, q 1 on channel u cannot be detected, and the active users carried by these power levels fail to access.

[0074] If so, the power level signal output by the base station is the estimated value of and then continue to detect the next power level signal on channel u until the detection of the power level signals on channel u is completed, and then transfer to the remaining channels for power level detection.

[0075] When all channels have been traversed, the power level signals on all channels can be activated.

[0076] 2) The base station performs preamble detection on the detected power level signals;

[0077] Step 3b-1, the base station separates the preamble signal and data signal of the active users from the detected power level signal among them.

[0078] Decompose into wherein, represents that the power level carried by channel u detected by the base station is q l of the preamble signal, represents that the power level carried by channel u detected by the base station is q l of the data signal, u ∈ CH, l ∈ I q ; and are respectively expressed as

[0079]

[0080] In the formula, and respectively represent the residual white noise of the preamble sub-slot and data sub-slot on channel u, and have the same probability distribution; and respectively represent the delayed preamble signal and data signal of user n carried by the power level q on channel u; assuming that the system has a good timing synchronization mechanism, the base station can directly extract the delayed preamble signal l from the power level signal and the data signal and then detect the preamble signal and recover the data signal

[0081] According to Equation (14), the power level q received by the base station on channel u l carrying the delayed preamble signal is rewritten as

[0082]

[0083] wherein, the attenuation coefficient matrix of N users on the uplink channel u the joint selection matrix of N users for channel u and power level q l <000070> represents the delayed preamble matrix sent by N users.

[0084] Step 3b-2, the base station separates the preamble signals of each preamble sub-slot from the delayed preamble signal among them ​

[0085] Split the matrix C into groups of every J columns to obtain the leading sub-slots t, where t ∈ L t Received preamble signals on is

[0086]

[0087] where the matrix C t = C[:, (t - 1)J:tJ] represents the preamble matrix transmitted in the leading sub-slot t; represents the residual noise in the leading sub-slot t.

[0088] Step 3b-3, perform a correlation operation on the received preamble signal and the preamble matrix S to obtain an estimated value of the user's preamble selection vector, and then recover the user's preamble sequence p n .

[0089] The sum preamble signal received by the base station on the channel u and the power level q l is correlated with the preamble sequence matrix S, and the result is calculated as Let

[0090]

[0091] Let Equation (18) is simplified as

[0092]

[0093] where represents the selection matrix for N users to transmit preambles on the channel u at the power level q in the leading sub-slot t l ; represents the preamble selection vector of user n on the channel u and the power level q in the leading sub-slot t l ; if the k-th element of is 1, i.e., it means that user n transmits the preamble sequence s on the channel u at the power level q in the leading sub-slot t l ; k ; represents the noise matrix, which has the same distribution as , i.e.,

[0094] Assume that the channel attenuation coefficient matrix is known, multiply it by to obtain the estimated value of the preamble selection matrix l of N active users on the channel u and the power level q as is

[0095]

[0096] Among them, represents the Hermitian matrix of the matrix There is Using the rounding method to round the elements of to obtain its approximate estimated value Where represents the approximate estimated value of the vector That is, the selection vector for user n to transmit the preamble through channel u at power level q in the leading sub-slot t l of the preamble approximate estimated value.

[0097] According to the system model, each user only selects one preamble. Therefore, at most one element in any row of the preamble selection matrix is 1, and all other elements are 0. Considering the influence of noise, the estimated value of the label l of the preamble sent by user n through channel u at power level q in the leading sub-slot t can be calculated as

[0098]

[0099] Where represents the k-th element of the vector If represents that user n has sent the preamble sequence s through channel u at power level q in the leading sub-slot t l ; if k ; if represents that user n has not sent a signal through channel u at power level q in the leading sub-slot t l send a signal.

[0100] Considering N s leading sub-slots, the delayed preamble transmission pattern l sent by user n on channel u at power level q is calculated as

[0101]

[0102] Since each user only selects one channel and one power level to transmit the preamble sequence in a certain leading sub-slot, that is l ∈ I q , t ∈ T t , there is exactly one Therefore, the estimated value of the preamble serial number c n selected by user n can be calculated as

[0103]

[0104] Accordingly, the preamble sequence p sent by user n n The estimated value of can be calculated as

[0105]

[0106] The preamble matrix P sent by all N users is reconstructed as

[0107] Step 3b-4, detect the data signal And reconstruct the service data of each active user;

[0108] The sum data signal received by the base station on channel u and power level q l And the data signal Is expressed as

[0109]

[0110] Wherein, Represents the spreading signal matrix of N active users; Represents the spreading signal matrix sent by the user on channel u at power level q l Based on Equation (25), The estimated value of Is calculated as

[0111]

[0112] Thus, the estimated value of the spreading signal matrix X transmitted by all users is calculated as

[0113]

[0114] Wherein,

[0115] Using the preamble matrix sent by the user To despread That is, let each row of the matrix Perform a correlation operation with each row of And then merge and arrange. b = 1, 2,..., v, and the user data matrix can be reconstructed as Wherein Represents the reconstructed service data vector of user n, Represents the b-th service data of user n after reconstruction, and is calculated as

[0116]

[0117] 3) The user access conflict detection algorithm is described as follows:

[0118] Step 3c-1: Obtain the set of active users carried by each power level on all channels.

[0119] Let the set of users carried by power level q on channel u l be denote the user number that accesses channel u and transmits an uplink signal at power level q l , and calculate it as Calculate as

[0120]

[0121] where denotes the 0-norm of vector , that is, the number of non-zero elements in vector . Obviously, only users in the same set may have access conflicts, and the delay preamble transmission patterns of conflicting users must be the same.

[0122] Step 3c-2: Calculate the user pilot conflict pattern.

[0123] Define the pilot conflict pattern e of any two users n, n' in the set l on channel u and power level q u,l as the product of the delay preamble transmission patterns and of users n, n', that is

[0124]

[0125] Obviously, if the delay preambles of users n, n' are the same, e n,l (n, n') ≥ 1, otherwise, e n,l (n, n') = 0.

[0126] Step 3c-3: Judge user random access conflicts.

[0127] Based on Equation (30), design the following user access conflict judgment method:

[0128]

[0129] If e u,l (n, n') ≥ 1, users n, n' have access conflicts and both of them fail in competing for access.

[0130] If e u,l(n, n′) = 0, there is no conflict between users n and n′, and both of them succeed in competing for access.

[0131] Traverse the entire user set Then the access competition results of all users can be obtained.

[0132] Step 3c-4, the base station feeds back the access result to the user.

[0133] For users who succeed in competing for access, the base station feeds back an access success confirmation message to the user through the broadcast channel; for users who fail in competing for access, the base station feeds back an access failure message to the user; the user who receives the access success message continues to transmit the next frame of signal until all service data is transmitted and then goes silent; the user who receives the access failure message will try to access again after delaying for one frame time.

[0134] 4) The access performance analysis of the active users is as follows:

[0135] Define the user random access rate as the probability that the uplink signal of the active user is successfully received at the base station.

[0136] Step 3d-1, calculate the probability that the active user successfully obtains the access channel.

[0137] When active user n senses that the interference power level of at least one channel is not higher than the interference power threshold at this time, the user can obtain the access channel; therefore, the probability P that active user n successfully obtains the access channel ch is calculated as:

[0138]

[0139] where represents the probability that user n senses that the interference power level of channel u exceeds the interference threshold . According to the system model, once the access channel is selected, the uplink power level of the user can be directly determined according to the channel interference power. Let user n select channel u and power level q l to perform random access.

[0140] Step 3d-2, calculate the probability that the signal of the power level to which the active user belongs is correctly detected.

[0141] According to the SIC principle, if the signal on power level q l can be successfully detected, the signals on the power levels larger than it, i.e., q l+1 to q L can all be successfully detected. Therefore, the probability l that the signal carried on power level q on channel u can be successfully detected is calculated as

[0142]

[0143] wherein, represents the probability that the i-th power level on channel u can be correctly detected.

[0144] Step 3d-3: Calculate the probability that an active user uses a unique preamble on the selected channel and power level.

[0145] According to the system model, the total number of delayed preambles in one frame is KN s ones. If the number of elements in the set to which user n belongs is N u,l , the probability l that user n selects a unique delayed preamble on channel u and power level q can be calculated as

[0146]

[0147] where K represents the number of orthogonal preamble sequences.

[0148] Step 3d-4: Calculate the random access probability of an active user.

[0149] The conditional access probability l that active user n selects channel u and power level q is calculated as:

[0150]

[0151] Considering that all channels are the same and the channel interference power follows a uniform distribution, the probability that user n selects any channel u is the same, and the probability of selecting any power level q l is also the same. Therefore, the joint selection probability p l of user n for channel u and power level q u,l is calculated as

[0152]

[0153] The probability distribution of the number N l of users selecting channel u and power level q u,l is calculated as

[0154]

[0155] Taking the statistical average of the conditional access probability with respect to N u,l , 0 ≤ N u,l ≤ N, the random access rate Pr succ of the user is obtained as

[0156] (37)

[0158] Perform simulation research on the above CR-NOMA-PD license-free random access scheme. This research is based on Matlab software, considering the Rayleigh fading channel, and focuses on comparing the performance with the multi-power-level non-orthogonal multiple access (MP-NOMA) scheme proposed in the literature ISSN: 1001-506X.

[0159] Set the simulation parameters as follows: the number of base station antennas M = 8, the number of uplink channels N c = 4, the number of single-antenna terminals N u = 30, the terminal data packet length v = 8 bit, using the QPSK modulation method, the preamble sequence uses the Walsh sequence, the number of preambles K = 8, the preamble length J = 8, the number of preamble sub-slots N s = 30, the set of transmit power levels Q tr = {1, 3, 5} mW, the channel noise variance σ 2 = 1, the perceived interference power level obeys the uniform distribution of [0, 5] mW, the channel noise threshold the base station receiving sensitivity The combined number of preambles is 4. The simulation results are the average of 500 random experiments, and it is stipulated that the access probability is 0 when there is no terminal user.

[0160] Based on the above assumptions, the simulation results of the method described in the present invention are as Figure 5 、 Figure 6 shown.

[0161] Figure 5It is the relationship curve between the access probability of end users and the number of users under different numbers of preambles. As the number of users gradually increases, the access probability of the CR-NOMA-PD method first rises rapidly and then decreases slightly. The MP-NOMA method has a similar result. Moreover, the more the number of preambles, the higher the access probability of users, the greater the optimal system load, and the slower the access probability decreases caused by excessive increase of users; this means that the overload capacity of the system increases with the increase of the number of preambles. When the number of preambles is small and the system is heavily loaded, the performance of the CR-NOMA-PD method is significantly better than that of the MP-NOMA method. For example, given K = 16 preambles, when the number of users N = 50, that is, the overload rate is N / K = 3.125, the access probability of the MP-NOMA method is 0.68, while the access probability of the CR-NOMA-PD method reaches 0.86, an increase of 26.47%. Given K = 8 preambles, when the number of users N = 50, that is, the overload rate is N / K = 6.25, the access probability of the MP-NOMA method drops to 0.25, while the access probability of the CR-NOMA-PD method still reaches 0.77, nearly doubling. This shows that the CR-NOMA-PD method greatly improves the random access rate of users, and the performance advantage is more obvious with fewer preambles.

[0162] Figure 6 shows the influence of the channel and the number of users on the access probability of users. To ensure resource fairness, when the number of sensing channels N of the CR-NOMA method c changes, the number of cascaded preambles N of the MP-NOMA method p changes synchronously. As the number of channels increases, the access probability of the CR-NOMA-PD method increases rapidly first and then decreases slowly. The MP-NOMA method has a similar phenomenon but its decrease rate is faster. When the system is heavily loaded, the performance advantage of the CR-NOMA-PD method is more obvious. For example, given Nc = Np = 4 and the target access probability of 0.8, the system load of the MP-NOMA method is 8 users, while the system load of the CR-NOMA-PD method is 40 users, and the system load is increased by 4 times. This shows that the overload access performance of the CR-NOMA-PD method has been significantly improved, and multi-channel sensing and preamble delay transmission are more beneficial to random access performance, especially overload access.

[0163] The above is only the preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. Any equivalent modification or change made by those of ordinary skill in the art according to the disclosure of the present invention shall be included in the protection scope recorded in the claims.

Claims

1. A large-scale unlicensed random access method based on CR-NOMA-PD, characterized by: In this method, active users use a time frame structure to perform contention access transmission. The time frame includes three parts: a sensing time slot, a transmission time slot, and a feedback waiting time slot. The transmission time slot includes several preamble sub-time slots and one data sub-time slot. The specific steps of the access method are: Step 1: The active user senses the interference power of the channel in the sensing time slot to optimize the access channel and uplink transmission power; the active user independently senses the interference power level of all uplink channels, selects the channel with the lowest interference power level below the threshold as the uplink access channel, and selects a power level from the power level set as the uplink transmission power level according to the power level selection criterion; Step 2: The active user performs contention access and data transmission in the transmission time slot: first, a preamble sequence is randomly selected from the preamble set to spread the service data to form spread spectrum data; then, the active user transmits the preamble sequence to the base station through the selected channel at a selected power level and randomly delays a number of preamble sub-time slots, thereby forming a delayed preamble; Finally, the active user continues to transmit spread spectrum data to the base station in the data subslot; Step 3: The active user receives the access result fed back by the base station in the feedback waiting time slot; the base station uses frequency band filtering to obtain the received signal on each channel, and uses the sequential interference cancellation SIC technology to detect the signal at each power level; Perform preamble detection, data detection and conflict detection on signals of each power level, recover the user's uplink signal and obtain the access contention result, and then feed back the result to the user through access response.

2. According to claim 1, a large-scale unlicensed random access method based on CR-NOMA-PD is characterized in that: In step 1, the uplink access channel selection and uplink transmit power level selection are as follows: Consider a single-cell mMTC system scenario, where the base station is equipped with M antennas and N single-antenna users are evenly distributed in the cell. The user number set is The length of a single burst service data of each user is v bits; there are N c uplink channels and K orthogonal preamble sequences of length J serve the uplink random access of users, K≤J; let the uplink channel set and preamble sequence set be and s k represents the kth leading sequence, represents the set of preamble numbers; different preamble sequences are orthogonal to each other and the energy of each preamble sequence is 1, that is, any two different preamble sequences s are selected from the preamble set SC. k ,s k′ , k≠k', k' represents another leading index number, both have Define the leading matrix S = [s1,…,s K ] T , SS H =I K , where S H represents the Hermitian matrix of matrix S, I K represents the K-order unit matrix, represents a complex field with K rows and J columns; Active users perform contention access transmissions using a time frame consisting of sensing time slots T s , transmission time slot T t , Feedback waiting time slot T w Three parts, among which the transmission time slot is divided into N s preamble subslots and 1 data subslot, let Represents the set of preamble sub-time slot numbers; each active user first performs spectrum sensing to obtain channel state information before accessing the transmission, and then optimizes the access channel and transmission power level: Step 1a, specifying a user transmit power level set and dividing the interference power level segments; Let the uplink transmission power level set of active users be Q = {q1,…,q L }, the set contains L power levels, where q l represents the lth power level, I q A set of labels representing the transmit power levels; the power levels in set Q are arranged in ascending order, i.e. There is always q i j ;​ The channel interference power level is also evenly divided into L segments, and Respectively represent the starting point and end point of the i-th segment, i = 1, 2, ..., L, and the regulations in Indicates the interference power threshold of the accessible channel; have That is, the end point value of the i-th interference power level segment Less than the end point value of the jth interference power level segment Step 1b, the active user selects an uplink access channel according to the interference power level; Let active user n, n∈I a Perceive all N c The set of interference power levels of uplink channels is in represents the interference power level perceived by active user n on channel u; Active user n meets the interference power threshold, that is, The channel with the smallest current interference power level is selected as the best access channel. The uplink access channel of active user n Calculated as: If the active user perceives that the interference power level on all current channels is higher than the interference power threshold The active user tries to access again after a delay of one frame; Step 1c, the active user selects an uplink transmit power level; User n according to the selected channel The interference power level The interference interval determines the optimal uplink transmit power level Optimal transmit power level number l * Calculated as: Among them, l * ∈I q ; represents the i-th interference segment, and Represent the starting point and end point of the i-th interference segment, respectively, and are calculated as like Provisions * =1.

3. According to claim 2, a large-scale unlicensed random access method based on CR-NOMA-PD is characterized in that: In step 2, the active user transmits the preamble and data signals to the base station through the selected uplink channel at the selected transmit power level, and the delayed preamble and data transmission steps are specifically as follows: Step 2a-1, the active user randomly selects a preamble sequence from the K preamble sequences and randomly selects a sub-slot from the preamble sub-slot to generate a delayed preamble signal; Consider user n choosing a channel As the uplink channel, select the power level q l,n As the transmit power level, select c n preambles participate in access contention and in the preamble subslot t n The leading sequence is transmitted in n ∈L p ,n∈I a ,t n ∈L i ; Therefore, user n's leading selection vector a n and the leading time vector t n Respectively expressed as and 1 of them [x] is an indicator function. When the condition x is met, 1 [x] =1, otherwise 1 [x] =0; therefore, the leading sequence p of user n n and delayed preamble c n Expressed as in, And||p n || 2 =1; P=JN s Indicates the length of the delayed preamble signal; Step 2a-2, the active user uses the selected preamble sequence to spread its v-bit service data to generate a spreading sequence with a length of D=vJ; Let the business data of active user n be d n =[d1,d2,...,d k ,…,d v ], where d k ∈{-1,+1}; using p n Right n The data after spread spectrum x n Expressed as x n =[x1,x2,...x D ]=[d1p n ,d2p n ,...d v p n ] (5) in, ||x n || 2 =1, D = vJ represents the length of the spread spectrum data; Step 2a-3, the active user transmits the delayed preamble and the spread spectrum data to the base station in sequence through the selected uplink channel at the selected power level; Combining equations (2) and (3), a frame of signal z sent by active user n in the time domain is n for: make represents the channel selection vector of active user n, a frame of signal F sent by active user n in the two-dimensional space of time and frequency n Expressed as in, represents the transmission signal of active user n on channel u; 4. According to claim 3, a large-scale unlicensed random access method based on CR-NOMA-PD is characterized in that: In step 3, the base station first performs power level detection on the received signal, specifically: Step 3a-1: The base station receives a signal from N c The sum signal of the uplink channels; Consider that the uplink channel from active users to the base station experiences Rayleigh fading and additive Gaussian white noise, and the channel is quasi-static, that is, the channel attenuation coefficient and additive noise remain unchanged within one frame time; the one-frame signal G received by the base station is the sum signal of N users, expressed as: Among them, F n =z n u n A frame signal representing active user n; N represents the distance from user n to the base station c The channel attenuation coefficient matrix of the uplink channels; N represents the distance from user n to base station antenna m c The attenuation coefficient vector of the uplink channels is represents the attenuation coefficient of the u-th uplink channel from user n to base station antenna m, which obeys a Gaussian distribution with a mean of 0 and a variance of 1; N represents the channel additive white Gaussian noise matrix, Since the channel is quasi-static, the columns of N are identical and have a mean of 0 and a variance of σ. 2 Gaussian distribution of channel attenuation coefficient matrix H n It remains constant within one frame and has a quasi-orthogonal characteristic, that is, the uplink channel attenuation coefficient matrix H for any two users n and n′ is n and H n′ ,have in, Indicates H n The conjugate transposed matrix of , M represents the number of base station antennas; Step 3a-2: The base station performs bandpass filtering on the received uplink signal to obtain N c A received signal of an uplink channel; The base station receives a signal G that contains N c The received signal G is processed by N c The bandpass filter is used to obtain N c The received signal of the uplink channel is G u , which is expressed as in, N represents the access channel selected by user n; u represents the Gaussian white noise matrix on channel u, It has a mean of 0 and a variance of σ 2 I M Gaussian distribution of represents the channel attenuation coefficient vector from user n to base station on channel u, which is the matrix H n The u-th column vector of ; represents the set of users who select channel u; represents the attenuation coefficient from N users to the base station on channel u; represents the transmitted signals of N users on channel u, equation (10) can be rewritten as Step 3a-3, using sequential interference cancellation (SIC) technology to cancel the received signal G on each uplink channel u Detect each power level signal in order from high to low power level The base station detects each power level q in order from high to low l The signal carried on Calculate power level signal Signal-to-interference-noise ratio Right now in, Indicates signal The average power is calculated as Here, trace() means finding the trace of the matrix; and They represent the power level q on channel u respectively l and q j The set of hosted users, q j l ; N u,i and N u,j Respectively represent sets and The number of elements in, that is, the power level q on channel u l and q j The number of users carried, ​ judge Is it greater than the detection threshold of the base station receiver, i.e. the receiving sensitivity? Right now Is it established: If not, the power level q on channel u l ,q l-1 ,…,The signals carried on q1 cannot be detected, and the active users carried on these power levels fail to access; If so, the power level signal output by the base station is Estimated value of Then continue to detect the next power level signal on channel u After the power level signal detection on channel u is completed, the power level detection of the remaining channels is switched; When all channels are traversed, the signals of various power levels on all channels can be activated.

5. According to claim 4, a large-scale unlicensed random access method based on CR-NOMA-PD is characterized in that: In step 3, the base station further performs preamble detection on the power level signal detected on each channel to recover the active user data and determine the random access result; The leading sequence detection method is specifically as follows: Step 3b-1, the base station detects the power level signal Separate the preamble signal and data signal of active users; Will Decompose into in, Indicates that the power level carried by channel u detected by the base station is q l The leading signal of Indicates that the power level carried by channel u detected by the base station is q l Data signal, u∈CH, l∈I q ; and Respectively expressed as In the formula, and denote the residual white noise of the preamble sub-slot and data sub-slot on channel u, respectively, and have the same probability distribution; and They represent the power level q on channel u respectively l The delayed pilot signal and data signal of user n are carried; assuming that the system has a good timing synchronization mechanism, the base station Directly extract the delayed leading signal With data signal Then detect the leading signal And restore the data signal According to equation (14), the power level q received by the base station on channel u is l Delayed preamble signal carried Rewrite as Among them, the attenuation coefficient matrix of N users on the uplink channel u is N users have channel u and power level q l The joint selection matrix represents the delayed preamble matrix sent by N users; Step 3b-2, the base station receives the delayed preamble signal Separate the leading signal of each leading sub-time slot Split every J columns of matrix C into groups to obtain the leading sub-slot t, t∈L t The received preamble signal on for Among them, the matrix C t =C[:,(t-1)J:tJ] represents the leading matrix transmitted in the leading sub-time slot t; Preamble subslot t Residual noise in Step 3b-3, receiving the pilot signal Perform correlation operations with the leading matrix S to obtain the estimated value of the user's leading selection vector, and then restore the user's leading sequence p n ; Set the base station to channel u and power level q l The received and pilot signals Perform correlation operation with the leading sequence matrix S, and the result Calculated as make Formula (18) can be simplified as in, It means that N users transmit at power level q through channel u in the preamble subslot t. l Send the selection matrix of the preamble, represents user n in the leading subslot t, channel u and power level q l The leading selection vector on ; if The kth element of is 1, that is It means that user n transmits a signal at power level q through channel u in preamble subslot t. l The preamble sequence s is sent k ; represents the noise matrix, which is have the same distribution, that is Set the channel attenuation coefficient matrix Known, it is Multiply them together to get N active users in channel u and power level q l The leader selection matrix on Estimated value of for in, Representation Matrix The Hermitian matrix of Use rounding method to The elements of are rounded to get their approximate estimates in Representation vector An approximate estimate of the power level q of user n in the preamble subslot t. l Transmit preamble selection vector An approximate estimate of According to the system model, each user selects only one leader, so the leader selection matrix At most one element in any row of is 1, and all other elements are 0. Considering the influence of noise, user n transmits a signal at power level q through channel u in the leading sub-time slot t. l Send leading label Estimated value of Calculated as in, Representation vector The tth element of It means that user n transmits power level q through channel u in the preamble sub-time slot t. l The preamble sequence s is sent k ;like Indicates that user n did not pass through channel u at power level q in the leading sub-time slot t l Send a signal; Consider N s preamble subslots, user n transmits on channel u at power level q l Delayed preamble transmission pattern sent Calculated as Since each user only selects a channel and a power level to transmit the preamble sequence in a certain preamble subslot, that is, There is only one Therefore, the leading sequence number c selected by user n n Estimated value of Calculated as Accordingly, the preamble sequence p sent by user n is n Estimated value of Calculated as The transmission preamble matrix P of all N users is reconstructed as Step 3b-4, data signal Conduct detection and reconstruct the business data of each active user; The base station is in channel u and power level q l The received and data signals Expressed as in, represents the spread spectrum signal matrix of N active users; represents a user on channel u with power level q l The transmitted spread spectrum signal matrix; Based on formula (25), Estimated value of Calculated as Therefore, the estimated value of the spread spectrum signal matrix X transmitted by all users is calculated as in, Using the user's transmission preamble matrix right Despread, that is, let the matrix Each row of Each row of performs relevant operations and then merges and arranges, b = 1, 2, ..., v, that is, the reconstructed user data matrix is in represents the reconstructed business data vector of user n, represents the reconstructed b-th business data of user n, calculated as 6. According to claim 5, a large-scale unlicensed random access method based on CR-NOMA-PD is characterized in that: In step 3, the user conflict detection method is specifically as follows: Two or more users who select the same channel and use the same power level and the same preamble will have access conflicts; Step 3c-1, obtaining a set of active users carried by all channels at all power levels; Let the power level on channel u be q l The set of users carried is Denotes access channel u and power level q l The user serial number of the uplink signal transmission, Calculated as in, Representation vector The 0-norm of The number of non-zero elements in the same set Access conflicts will occur only for users in the must be the same; Step 3c-2, calculating the active user pilot conflict pattern; Defining a Collection Any two users n, n′ in channel u and power level q l Pilot conflict pattern on u,l (n,n′) is the delayed preamble transmission pattern of user n,n′ and The product of If the delayed preambles of users n and n′ are the same, e n,l (n,n′)≥1, otherwise, e n,l (n,n′)=0; Step 3c-3, determining random access conflicts of active users; Based on formula (30), the following active user access conflict judgment method is designed: 1) If e u,l (n,n′)≥1, active users n and n′ have access conflicts, and both of them fail to compete for access; 2) If e u,l (n,n′)=0, there is no access conflict between active users n and n′, and both users successfully compete for access; Traverse all user collections Obtain access competition results of all active users; Step 3c-4, the base station feeds back the access result to the active user; For users who successfully compete for access, the base station will feedback an access success confirmation message to the user through the broadcast channel; for users who fail to compete for access, the base station will feedback an access failure message to the user; the user who receives the access success message will continue to transmit the next frame signal until all service data is transmitted and then remain silent; the user who receives the access failure message will delay for one frame and try to access again.

7. According to claim 6, a large-scale unlicensed random access method based on CR-NOMA-PD is characterized in that: In step 3, the access performance analysis of active users is as follows: The user random access rate is defined as the probability that the uplink signal of an active user is successfully received at the base station; Step 3d-1, calculating the probability of an active user successfully acquiring an access channel; When active user n perceives that the interference power level of at least one channel is not higher than the interference power threshold When , the user can obtain access channel; Therefore, the probability P of active user n successfully acquiring access to the channel is ch Calculated as: in, represents the interference power level perceived by user n on channel u Exceeding the interference threshold According to the system model, once the access channel is selected, the user's uplink power level is directly determined by the channel interference power; let user n select channel u and power level q l performing random access; Step 3d-2, calculating the probability that the power level signal of the active user is correctly detected; According to the SIC principle, if the power level q l The signal above can be successfully detected, and the power level greater than it is, i.e., q l+1 to q L The signals on channel u can be successfully detected; the power level q on channel u l The probability that the signal carried can be successfully detected Calculated as in, represents the probability that the i-th power level on channel u can be correctly detected; Step 3d-3, calculating the probability that an active user uses a unique preamble on the selected channel and power level; According to the system model, the total number of delayed preambles in a frame is KN s If user n belongs to the set The number of elements in is N u,l , user n in channel u and power level q l The probability of selecting a unique delayed leader Calculated as Where K represents the number of orthogonal preamble sequences; Step 3d-4, calculating the random access probability of active users; Active user n chooses channel u and power level q l Conditional access probability Calculated as: Considering that all channels are the same and the channel interference power follows a uniform distribution, user n has the same probability of selecting any channel u and any power level q l The probability of user n for channel u and power level q is also the same; l The joint selection probability p u,l Calculated as Further, select channel u and power level q l Number of users N u,l The probability distribution of is calculated as Conditional Access Probability About N u,l , 0≤N u,l ≤N to find the statistical average and obtain the user's random access rate Pr succ for The user random access rate is equal to the ratio of the number of successfully accessed users to the total number of active users.

Images