A resource scheduling method for D2D communication in Cell-Free Massive MIMO systems based on mode selection and power allocation

Abstract

The application provides a resource scheduling method for D2D communication of a Cell-Free Massive MIMO system based on mode selection and power allocation. Step 1: assembling a system; step 2: constructing an optimization target for maximizing the transmission rate of the whole system under the condition of meeting the minimum rate requirement and power requirement of CFUE users and D2D users; step 3: according to the working mode of the D2D users, the D2D users are divided into optimization P1 working in a local routing mode and optimization P2 working in a D2D direct mode to solve step 2; step 4: the SCA method is used to solve the P1 and P2 optimizations respectively; step 5: based on the optimization results of the two working modes P1 and P2 solved, the working mode selection of the D2D users and the power allocation of the CFUE users and the D2D users are carried out. The application aims at the prior art that the Cell-Free Massive MIMO system is combined with D2D communication, only considers a working mode of the D2D users, i.e. the D2D direct mode, does not consider the problem that the D2D users should also have the ability to communicate with the AP, i.e. the local routing mode, and the suitable power allocation scheme can improve the system performance.

Description

Technical Field

[0001] This invention belongs to the field of communication technology, specifically relating to a resource scheduling method for D2D communication in a Cell-Free Massive MIMO system based on mode selection and power allocation. Background Technology

[0002] Cellular network architecture is currently the mainstream communication network architecture. It divides the wireless signal coverage area into multiple adjacent cells, each served by a single base station. However, this network architecture is prone to problems such as dropped calls due to handover during user movement, increased communication latency, and low user speeds at the cell edge due to the inherent network structure. This limits the development of high-speed, low-latency, and high-reliability communication technologies required by 5G and B5G. Therefore, the network needs to enter a cellular-free mode. Cell-Free Massive MIMO system is such a network facility. It eliminates the problems caused by cellular networks by deploying a large number of access points (APs) in the coverage area and using the same time-frequency resources to serve fewer users.

[0003] D2D (Device-to-Device Communication) is another end-to-end communication technology used to improve user transmission rates and reduce latency. It allows direct communication between devices, thus offering advantages such as those mentioned above. Furthermore, D2D communication can reduce the burden on the core network and is considered a promising communication technology in B5G, currently being widely researched in cellular networks. The introduction of D2D communication technology can improve the transmission rate of Cell-Free Massive MIMO systems, reduce the traffic burden passing through the Cell-Free Massive MIMO architecture, and decrease transmission loss in the fronthaul link.

[0004] To our knowledge, only four papers have studied the coexistence of Cell-Free Massive MIMO systems and D2D communication: First, "Downlink Achievable Rate of D2D Underlaid Cell-Free Massive MIMO Systems With Low-Resolution DACs," published in IEEE System Journal, Vol. 16, No. 3, 2022 by X. Qiao, Y. Zhang et al.; second, "Coexistence of D2D Communications and Cell-Free Massive MIMO Systems With Low Resolution ADC for Improved Throughput in Beyond-5G Networks," published in IEEE Transactions on Communications, Vol. 70, No. 2, 2022 by H. Masoumi, MJ Emadi et al.; and third, "Average Sum Rate of D2D Underlaid Multigroup Multicast Cell-Free Massive MIMO With...", published in IEEE Wireless Communications Letters on October 26, 2022 by M. Zhou, J. Li et al. "Multi-Antenna Users" and J.Li,A., November 1, 2022.Wan et al.'s publicly published paper in IEEE Access, "Downlink Analysis for the D2D Underlaid Multigroup Multicast Cell-Free Massive MIMO With Low-Resolution ADCs / DACs," only considers one operating mode for D2D users—the direct D2D mode. Furthermore, the latter two papers do not consider power allocation. Such assumptions lack consideration for real-world scenarios. Since communication via the AP remains the primary communication mode for users in the system, with D2D communication being secondary, D2D users with D2D communication capabilities should also have the ability to communicate with the AP, i.e., operate in local routing mode. A suitable power allocation scheme can improve system performance. Based on this, this work considers a resource scheduling method for joint mode selection and power allocation in a scenario where Cell-Free Massive MIMO systems and D2D communication coexist, where D2D users have two operating modes. This method aims to maximize the overall system transmission rate while meeting the minimum user rate and power requirements. Summary of the Invention

[0005] This invention provides a resource scheduling method for Cell-Free Massive MIMO systems oriented towards D2D communication based on mode selection and power allocation. In contrast, existing technologies that combine the two only consider one working mode of D2D users, namely the D2D direct mode, and do not consider that D2D users should also have the ability to communicate with the AP, i.e., the local routing mode. Furthermore, a suitable power allocation scheme can improve system performance.

[0006] This invention is achieved through the following technical solution:

[0007] A resource scheduling system for D2D communication in a Cell-Free Massive MIMO system based on mode selection and power allocation, the resource scheduling system including an access point (AP), a CFUE user, a pair of D2D users, and a central processing unit (CPU);

[0008] The AP is a single antenna, numbered M, and connected to the CPU via a fronthaul link. It can be used for receiving, sending, and processing information, and is equipped with low-resolution ADCs to achieve a trade-off between system deployment cost and performance.

[0009] The number of CFUE users is K, and each user is equipped with a single antenna, representing users in the system who can only communicate through the AP.

[0010] The pair of D2D users consists of a single-antenna transmitter and an N-antenna receiver, which can operate in D2D direct mode or in local routing mode via AP, depending on the system scheduling results.

[0011] Both CFUE users and D2D users operating in local routing mode are served by the corresponding AP clusters;

[0012] The CPU is used for receiving, sending, and processing information, as well as for resource scheduling of the entire system.

[0013] A resource scheduling method for D2D communication in a Cell-Free Massive MIMO system based on mode selection and power allocation, the resource scheduling method comprising the following steps:

[0014] Step 1: Assemble the system as described above;

[0015] Step 2: Based on the system assembled in Step 1, construct an optimization objective that maximizes the overall system transmission rate while meeting the minimum rate and power requirements of CFUE users and D2D users;

[0016] Step 3: Based on the working mode of D2D users, divide them into optimized P1 working in local routing mode and optimized P2 working in D2D direct mode to solve Step 2;

[0017] Step 4: Solve the optimizations of P1 and P2 in Step 3 using the SCA method;

[0018] Step 5: Based on the optimization results of the two working modes P1 and P2 obtained in Step 4, select the working mode for D2D users and allocate power between CFUE users and D2D users.

[0019] Furthermore, the optimization objective of step 2 can be expressed mathematically as follows:

[0020]

[0021] In the formula Let ξ represent the transmission rate of the k-th CFUE user, and let ρ represent the accuracy of the ADCs. c ,ρ d This indicates the maximum transmission power for CFUE and D2D users. η d This represents the power control coefficient for CFUE and D2D users. M represents the large-scale fading coefficient between the k-th CFUE and the D2D transmitter and the m-th AP. k This represents the set of APs serving the k-th CFUE, and N0 represents the noise power.

[0022] Furthermore, the M k The specific method involves sorting all large-scale fading coefficients connected to the k-th CFUE user in descending order, and taking the top |M... k A set of access points (APs) whose sum of large-scale fadings is greater than or equal to 95% of the sum of all large-scale fadings constitutes M. k ;

[0023] In the formula This represents the transmission rate when a D2D user is operating in local routing mode, in M. d This represents the set of APs serving D2D users, obtained using a method similar to the one described above;

[0024] In the formula This indicates the transmission rate of a D2D user operating in D2D direct mode. This represents the large-scale fading coefficient between the D2D transmitter and the D2D receiver, as well as the large-scale fading coefficient between the k-th CFUE user and the D2D receiver.

[0025] In the formula x d x represents the mode selection factor. d =1 indicates that the D2D user is working in local routing mode; otherwise, it indicates that the D2D user is working in D2D direct mode.

[0026] In the formula This indicates the minimum rate requirement for CFUE users and D2D users.

[0027] Furthermore, the optimization P1 of step 3 in local routing mode is as follows:

[0028]

[0029] The optimization P2 for step 3 in D2D direct mode is as follows:

[0030]

[0031] Furthermore, the specific optimization process for solving the optimization P1 in the local routing mode based on the SCA method mentioned in step 4 is as follows:

[0032] Step 4.1.P1: Set the maximum number of iterations N max and error tolerance ε;

[0033] Step 4.2.P1: Given initial values Set the initial iteration value n = 0;

[0034] Step 4.3.P1: Solve the following optimization problem using the fmincon function, where This represents the result of the nth iteration;

[0035]

[0036]

[0037]

[0038] in

[0039]

[0040] in for

[0041]

[0042] At point First-order Taylor expansion at the location;

[0043] Step 4.4.P1: Each time step 4.3.P1 is executed, the number of iterations n = n + 1.

[0044]

[0045] Until the number of iterations reaches N max or

[0046]

[0047] Furthermore, the specific optimization process for solving the optimization P2 in step 3 using the SCA method mentioned in step 4 is as follows:

[0048] Step 4.1.P2: Set the maximum number of iterations N max and error tolerance ε;

[0049] Step 4.2.P2: Given initial values Set the initial iteration value n = 0;

[0050] Step 4.3.P2: Solve the following optimization problem using the fmincon function, where This represents the result of the nth iteration:

[0051]

[0052]

[0053]

[0054] in

[0055]

[0056] in for

[0057]

[0058] At point First-order Taylor expansion at the location;

[0059] Step 4.4.P2: Each time step 4.3.P2 is executed, the number of iterations n = n + 1. Until the number of iterations reaches N max or

[0060] Further, step 5 specifically involves assuming that the maximum system speed corresponding to optimized P1 for D2D users operating in local routing mode is... Optimization of P2 system operating in D2D direct mode

[0061] The maximum value of the system rate is The final scheduling result is:

[0062]

[0063] in The corresponding system transmission rate corresponds to the working mode of D2D users and CFUE and D2D

[0064] The user's transmission power is the final result of mode selection and power allocation.

[0065] The beneficial effects of this invention are:

[0066] This invention considers a resource scheduling method for joint mode selection and power allocation in scenarios where Cell-Free Massive MIMO systems and D2D communication coexist, and D2D users have two operating modes. It overcomes the shortcomings of existing technologies combining Cell-Free Massive MIMO systems and D2D communication, which only consider one operating mode of D2D users, namely the D2D direct mode, and do not consider the local routing mode of communication between D2D users and APs. Furthermore, it comprehensively considers the power allocation scheme, which is more beneficial to improving system performance. Attached Figure Description

[0067] Figure 1 This is a schematic diagram of the structure of the present invention.

[0068] Figure 2 This is a graph showing the change of the iteration value of the present invention with the number of iterations.

[0069] Figure 3 This is a graph showing the change in the total system transmission rate of the present invention with the number of CFUEs.

[0070] Figure 4 The present invention provides curves showing the transmission rate versus D2D distance for different numbers of APs, wherein (a) is the curve showing the transmission rate versus D2D distance for 40 APs, (b) is the curve showing the transmission rate versus D2D distance for 60 APs, (c) is the curve showing the transmission rate versus D2D distance for 80 APs, and (d) is the curve showing the transmission rate versus D2D distance for 100 APs. Detailed Implementation

[0071] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0072] A resource scheduling system for D2D communication in a Cell-Free Massive MIMO system based on mode selection and power allocation, the resource scheduling system including an access point (AP), a CFUE user, a pair of D2D users, and a central processing unit (CPU);

[0073] The AP is a single antenna, numbered M, and connected to the CPU via a fronthaul link. It can be used for receiving, sending, and processing information, and is equipped with low-resolution ADCs to achieve a trade-off between system deployment cost and performance.

[0074] The number of CFUE users is K, and each user is equipped with a single antenna, representing users in the system who can only communicate through the AP.

[0075] The pair of D2D users consists of a single-antenna transmitter and an N-antenna receiver, which can operate in D2D direct mode or in local routing mode via AP, depending on the system scheduling results.

[0076] Both CFUE users and D2D users operating in local routing mode are served by the corresponding AP clusters;

[0077] The CPU is used for receiving, sending, and processing information, as well as for resource scheduling of the entire system.

[0078] A resource scheduling method for D2D communication in a Cell-Free Massive MIMO system based on mode selection and power allocation, the resource scheduling method comprising the following steps:

[0079] Step 1: Assemble the system as described above;

[0080] Step 2: Based on the system assembled in Step 1, construct an optimization objective that maximizes the overall system transmission rate while meeting the minimum rate and power requirements of CFUE users and D2D users;

[0081] Step 3: Based on the working mode of D2D users, divide them into optimized P1 working in local routing mode and optimized P2 working in D2D direct mode to solve Step 2;

[0082] Step 4: Solve the optimizations of P1 and P2 in Step 3 using the SCA method;

[0083] Step 5: Based on the optimization results of the two working modes P1 and P2 obtained in Step 4, select the working mode for D2D users and allocate power between CFUE users and D2D users.

[0084] A resource scheduling method for D2D communication in a Cell-Free Massive MIMO system based on mode selection and power allocation, wherein the optimization objective in step 2 can be expressed mathematically as follows:

[0085]

[0086] In the formula Let ξ represent the transmission rate of the k-th CFUE user, and let ρ represent the accuracy of the ADCs. c ,ρ d This indicates the maximum transmission power for CFUE and D2D users. η d This represents the power control coefficient for CFUE and D2D users. M represents the large-scale fading coefficient between the k-th CFUE and the D2D transmitter and the m-th AP. k This represents the set of APs serving the k-th CFUE, and N0 represents the noise power.

[0087] A resource scheduling method for D2D communication in a Cell-Free Massive MIMO system based on mode selection and power allocation, wherein the M k The specific method involves sorting all large-scale fading coefficients connected to the k-th CFUE user in descending order, and taking the top |M... k A set of access points (APs) whose sum of large-scale fadings is greater than or equal to 95% of the sum of all large-scale fadings constitutes M. k ;

[0088] In the formula This represents the transmission rate when a D2D user is operating in local routing mode, in M. d This represents the set of APs serving D2D users, obtained using a method similar to the one described above;

[0089] In the formula This indicates the transmission rate of a D2D user operating in D2D direct mode. This represents the large-scale fading coefficient between the D2D transmitter and the D2D receiver, as well as the large-scale fading coefficient between the k-th CFUE user and the D2D receiver.

[0090] In the formula x d x represents the mode selection factor. d =1 indicates that the D2D user is working in local routing mode; otherwise, it indicates that the D2D user is working in D2D direct mode.

[0091] In the formula This indicates the minimum rate requirement for CFUE users and D2D users.

[0092] A resource scheduling method for D2D communication in a Cell-Free Massive MIMO system based on mode selection and power allocation, wherein the optimization P1 in step 3 under local routing mode is as follows:

[0093]

[0094] The optimization P2 for step 3 in D2D direct mode is as follows:

[0095]

[0096] A resource scheduling method for D2D communication in a Cell-Free Massive MIMO system based on mode selection and power allocation, wherein the specific optimization process of solving the optimization P1 in the local routing mode of step 3 using the SCA method mentioned in step 4 is as follows:

[0097] Step 4.1.P1: Set the maximum number of iterations N max and error tolerance ε;

[0098] Step 4.2.P1: Given initial values Set the initial iteration value n = 0;

[0099] Step 4.3.P1: Solve the following optimization problem using the fmincon function, where This represents the result of the nth iteration;

[0100]

[0101]

[0102]

[0103]

[0104] in

[0105]

[0106] in for

[0107]

[0108] At point The first-order Taylor expansion at the location.

[0109] Step 4.4.P1: Each time step 4.3.P1 is executed, the number of iterations n = n + 1.

[0110]

[0111] Until the number of iterations reaches N max or

[0112]

[0113] A resource scheduling method for D2D communication in a Cell-Free Massive MIMO system based on mode selection and power allocation, wherein the specific optimization process of solving the optimization P2 in D2D direct mode based on the SCA method mentioned in step 4 is as follows:

[0114] Step 4.1.P2: Set the maximum number of iterations N max and error tolerance ε;

[0115] Step 4.2.P2: Given initial values Set the initial iteration value n = 0;

[0116] Step 4.3.P2: Solve the following optimization problem using the fmincon function, where This represents the result of the nth iteration:

[0117]

[0118]

[0119]

[0120]

[0121] in

[0122]

[0123] in for

[0124]

[0125] At point First-order Taylor expansion at the location;

[0126] Step 4.4.P2: Each time step 4.3.P2 is executed, the number of iterations n = n + 1. Until the number of iterations reaches N max or

[0127] A resource scheduling method for D2D communication in a Cell-Free Massive MIMO system based on mode selection and power allocation, wherein step 5 specifically involves assuming that the maximum system rate corresponding to optimized P1 for D2D users operating in local routing mode is... The maximum system speed corresponding to the optimized P2 operating in D2D direct mode is The final scheduling result is:

[0128]

[0129] in The corresponding system transmission rate, the working mode of the D2D user, the CFUE, and the transmission power of the D2D user constitute the final mode selection and power allocation result.

[0130] This embodiment considers a 1km×1km coverage area, with AP access points, CFUE users, and D2D transmitters evenly distributed within the coverage area. D2D receivers are located within a circle with a radius of 10-100m centered on the D2D transmitter. The large-scale fading coefficient includes path loss and shadow fading, and the model used is as follows (dB):

[0131]

[0132] in

[0133] L = 46.3 + 33.9log 10 (f)-13.82log 10 (h)-(1.1log 10 (f)-0.7)h u +(1.56log 10 (f)-0.8),

[0134] f represents the carrier frequency, and h represents the antenna height of the AP or D2D receiver user. u Indicates the user's antenna height, d a,bThis represents the distance between two points, where d1 and d0 represent reference distances, and X represents the distance between them. a,b =σ sh z ab This represents shadow fading (dB), where σ sh Indicates the standard deviation of shadow fading. Let represent the shadow fading coefficient, 0≤δ≤1 be a parameter, and a~N(0,1) and b~N(0,1) be independent random variables with covariance . Where d aa' d represents the distance between APs or between an AP and a D2D receiver. bb' d represents the distance between CFUE users or the distance between a CFUE user and the D2D transmitter. decorr This indicates the decorrelation distance. The main parameters used in the simulation are listed below:

[0135]

[0136]

[0137] Algorithm validity verification

[0138] This invention employs the SCA (Self-Controlled Algorithm) method for power control. The SCA method is an iterative algorithm that can reach an asymptotic state after a certain number of iterations. Therefore, the first step is to verify whether the SCA-based power control algorithm of this invention can achieve asymptotic results. If it can, the next step is to determine whether the number of iterations required to reach the asymptotic state is within an acceptable range. Simulation results are as follows: Figure 2 As shown in the figure, the D2D distance is 50m. It can be seen from the figure that the algorithm can asymptotically approach the target under different D2D user operating modes, different numbers of APs, and different numbers of CFUE users, achieving asymptoticity within 10 iterations. This demonstrates the effectiveness of the SCA method applied to the power control scheme of this invention.

[0139] Verification of the Importance of Mode Selection and Power Allocation

[0140] Because the scheme comprehensively considers mode selection and power allocation, the importance of comprehensively considering both is verified below, and the simulation results are as follows. Figure 3 As shown in the figure, the number of APs M = 60 and the D2D distance is 50m. It can be seen from the figure that considering both mode selection and power allocation comprehensively improves the overall system transmission rate compared to considering power allocation in only one operating mode, verifying the importance of combining the two. Similarly, it can be seen that as the number of CFUE users increases, regardless of the D2D user's operating mode, the overall system transmission rate increases with the number of CFUEs.

[0141] The impact of D2D distance and number of APs on verification

[0142] Figure 4 The graph illustrates the transmission rates for different AP numbers (40, 60, 80, and 100), including the total transmission rate when D2D users can only operate in AP mode (local routing mode), the total transmission rate when operating only in D2D mode, the total system transmission rate after mode selection, and the gain brought by selecting D2D mode. The graph shows that when the number of APs is constant, changes in D2D distance do not affect the total system transmission rate when operating only in AP mode, but they do affect the total system transmission rate when operating in D2D mode, decreasing as the D2D distance increases. Furthermore, comparing the four graphs showing the total transmission rates in the two modes with different AP numbers, it can be seen that as the number of APs increases, regardless of the D2D user's operating mode, the total system transmission rate increases with the number of APs. The mode selection results indicate that mode selection is more likely to choose the communication mode that maximizes the total system transmission rate. The gain curve of D2D mode shows that operating in direct D2D mode is beneficial when the D2D distance is short, while AP mode is more advantageous when the D2D distance is large, further illustrating the importance of mode selection.

Claims

1. A resource scheduling method for D2D communication in a Cell-Free Massive MIMO system based on mode selection and power allocation, characterized in that, The method is used in a resource scheduling system, which includes an access point (AP), CFUE users, a pair of D2D users, and a central processing unit (CPU). The AP is a single antenna, and the number of antennas is... M It is connected to the CPU via a fronthaul link and can be used for information reception, transmission and processing. It is also equipped with low-resolution ADCs to achieve a trade-off between system deployment cost and performance. The number of CFUE users is: K， A single antenna indicates that the system can only communicate through the AP. The D2D user pair consists of a single-antenna transmitter and a... N The antenna receiver can operate in D2D direct mode or in local routing mode via AP, depending on the system scheduling results. Both CFUE users and D2D users operating in local routing mode are served by the corresponding AP clusters; The CPU is used for receiving, sending, and processing information, as well as for resource scheduling of the entire system. The resource scheduling method includes the following steps: Step 1: Assemble the resource scheduling system; Step 2: Based on the system assembled in Step 1, construct an optimization objective that maximizes the overall system transmission rate while meeting the minimum rate and power requirements of CFUE users and D2D users; Step 3: Based on the working mode of D2D users, divide them into optimized P1 working in local routing mode and optimized P2 working in D2D direct mode to solve Step 2; Step 4: Solve the optimizations of P1 and P2 in Step 3 using the successive convex approximation SCA method; Step 5: Based on the optimization results of the two working modes P1 and P2 obtained in Step 4, select the working mode for D2D users and allocate power between CFUE users and D2D users.

2. The resource scheduling method for D2D communication in a Cell-Free Massive MIMO system based on mode selection and power allocation according to claim 1, characterized in that, The optimization objective of step 2 can be expressed mathematically as follows: (1) In the formula Indicates the first k The transmission rate of each CFUE user Indicates the accuracy of ADCs. This indicates the maximum transmission power for CFUE and D2D users. This represents the power control coefficient for CFUE and D2D users. Indicates the first k The first CFUE and D2D sender and the first m Large-scale fading coefficients between APs Indicates service to the k A set of APs for each CFUE. Indicates noise power; In the mathematical expression of the optimization objective The specific method is to derive all the results related to the first... k The large-scale fading coefficients of each CFUE user connected are sorted in descending order, and the top [number] are taken. An AP set is constituted if the sum of large-scale fadings is greater than or equal to 95% of the sum of all large-scale fadings. ; In the formula This indicates the transmission rate when a D2D user is operating in local routing mode. This represents the set of APs serving D2D users, and the method obtained is the same as described above. The methods used are the same; In the formula This indicates the transmission rate of a D2D user operating in D2D direct mode. Represents the large-scale fading coefficient between the D2D transmitter and the D2D receiver, and the first... k Large-scale fading coefficient between a CFUE user and a D2D receiver; In the formula Indicates the mode selection factor. This indicates that the D2D user is working in local routing mode; otherwise, it indicates that the D2D user is working in D2D direct mode. In the formula This indicates the minimum rate requirement for CFUE users and D2D users.

3. The resource scheduling method for D2D communication in a Cell-Free Massive MIMO system based on mode selection and power allocation according to claim 2, characterized in that, The optimization P1 of step 3 in local routing mode is as follows: (2) The optimization P2 for step 3 in D2D direct mode is as follows: (3)。 4. The resource scheduling method for D2D communication in a Cell-Free Massive MIMO system based on mode selection and power allocation according to claim 3, characterized in that, The specific optimization process for solving the optimization P1 in the local routing mode based on the SCA method mentioned in step 4 is as follows: Step 4.1.P1: Set the maximum number of iterations N max and error tolerance ; Step 4.2.P1: Given initial values Set the initial iteration value n=0; Step 4.3.P1: Solve the following optimization problem using the fmincon function, where This represents the result of the nth iteration; in in for At point First-order Taylor expansion at the location; Step 4.4.P1: Each time step 4.3.P1 is executed, the number of iterations n = n + 1. ， Until the number of iterations reaches N max or 。 5. A resource scheduling method for D2D communication in a Cell-Free Massive MIMO system based on mode selection and power allocation according to claim 4, characterized in that, The specific optimization process for solving the optimization P2 in step 3 using the SCA method mentioned in step 4 is as follows: Step 4.1.P2: Set the maximum number of iterations N max and error tolerance ; Step 4.2.P2: Given initial values Set the initial iteration value n=0; Step 4.3.P2: Solve the following optimization problem using the fmincon function, where This represents the result of the nth iteration: in in for At point First-order Taylor expansion at the location; Step 4.4.P2: Each time step 4.3.P2 is executed, the number of iterations n = n + 1. until the number of iterations reaches N. max or .

6. A resource scheduling method for D2D communication in a Cell-Free Massive MIMO system based on mode selection and power allocation according to claim 5, characterized in that, Step 5 specifically involves assuming that the maximum system speed corresponding to optimized P1 for a D2D user operating in local routing mode is... The maximum system speed corresponding to the optimized P2 operating in D2D direct mode is The final scheduling result is: (4) in The corresponding system transmission rate, the working mode of the D2D user, the CFUE, and the transmission power of the D2D user constitute the final mode selection and power allocation result.

Images