A method for optimizing performance of a dense cellular network based on NOMA and D2D

Abstract

The application relates to the technical field of wireless communication network optimization, and discloses a dense cellular network performance optimization method based on NOMA and D2D, which comprises the following steps: S1: acquiring real-time signal strength through a base station monitoring unit; S2: obtaining base station preset signal strength of a user terminal; S3: comparing the real-time signal strength of the specified base station acquired by the user terminal with the preset signal strength, if the real-time signal strength is greater than the preset signal strength, a cellular mode is selected; otherwise, step S4 is entered; S4: the user terminal searches for an HTCU which has successfully connected to the specified base station around, the HTCU is used as a D2D relay, data transmission is carried out through a two-hop D2D mode, otherwise, the user terminal is in an interrupted state, through deeply integrating NOMA and D2D communication technologies in the dense cellular network, network delay is reduced, the burden of the base station is reduced, and the local multiplexing efficiency of the frequency spectrum resources is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wireless communication network optimization, and particularly relates to a dense cellular network performance optimization method based on NOMA and D2D. BACKGROUND

[0002] With the rapid development of mobile technology and the global deployment of 5G networks, the use of intelligent terminal devices has entered an explosive growth; traditional cellular networks cannot provide high-quality networks for a large number of communication devices, so dense cellular networks are used to meet the network performance requirements of the 5G Internet of Everything era.

[0003] Dense cellular networks deploy dense small base stations under the base stations of cellular networks to reduce the load of large base stations and provide high-quality spectrum resources for users, so as to ensure the network transmission speed of a large number of communication devices.

[0004] Dense cellular networks are mainly used for human type communication users (HTCU) to perform human type communication, and with the gradual increase of the types of intelligent terminal devices, there are also a large number of machine type communication devices (MTCD) in the Internet of Everything scenario, if the machine type communication devices also directly connect with the base station through the cellular mode, the burden of the base station is increased, which may cause the network quality to decrease. SUMMARY

[0005] The present application aims to provide a dense cellular network performance optimization method based on NOMA and D2D, which solves the following technical problems:

[0006] How to reduce the burden of the base station and ensure the network quality of each user terminal.

[0007] The purpose of the present application can be achieved by the following technical solutions:

[0008] A dense cellular network performance optimization method based on NOMA and D2D is applied to a dense cellular network system, the dense cellular network system includes a plurality of base stations, a plurality of user terminals and a plurality of communication modes; the optimization method includes the following steps:

[0009] S1: Obtain the real-time signal strength of a specified base station by setting a base station monitoring unit on the user terminal;

[0010] S2: Obtain the preset signal strength of the base station of the user terminal by establishing a Dec-POMDP model;

[0011] S3: Compare the real-time signal strength of the specified base station obtained by the user terminal with the preset signal strength, if the real-time signal strength is greater than the preset signal strength, select the cellular mode, and the user terminal directly transmits data to the specified base station through the cellular link; otherwise, go to step S4;

[0012] S4: the user terminal searches for an HTCU around the user terminal which has successfully connected to the specified base station as a potential relay device, selects an HTCU closest to the real-time device and which has successfully connected to the specified base station from the searched HTCUs as a D2D relay, and performs data transmission through a two-hop D2D mode, wherein the user terminal first sends data to the selected HTCU relay through a D2D link, and the relay then forwards the data to the specified base station through a cellular link; otherwise, the user terminal is in an interrupted state.

[0013] As a further scheme of the present application, the user terminal comprises an HTCU and an MTCD.

[0014] As a further scheme of the present application, the specific method for obtaining the preset signal strength of the base station of the user terminal comprises the following steps:

[0015] S10: taking the base station as an intelligent agent, taking the increase or decrease operation of the preset signal strength as an action, taking the real-time signal strength of the base station as the state of the base station, and establishing a Dec-POMDP model by feeding back the environment reward to the base station;

[0016] S20: each base station initializes a Q table to store the Q values of all possible state-action pairs and updates the Q values;

[0017] S30: the base station selects the optimal action in the Q table with a probability of ε, and randomly selects an action with a probability of 1-ε;

[0018] S40: the reward r is calculated by the formula t ;

[0019] wherein, and are the rates of the jth HTCU to the ith base station link at the tth step and the (t-1)th step, respectively, and are the rates of the kth MTCD k to the ith base station link at the tth step and the (t-1)th step, respectively;

[0020] S50: repeating steps S30-S50 until a training end condition is reached; each base station finds the optimal preset signal strength.

[0021] As a further scheme of the present application, the determination process of the specified base station is as follows:

[0022] S100: each base station obtains the base station transmission data through a base station monitoring unit;

[0023] S200: each base station obtains a state evaluation index of each base station by analyzing the base station transmission data; ​

[0024] S300: determining a target area of the user terminal in a user terminal centric manner;

[0025] S400: numbering all base stations in the target area, and the number is i;

[0026] S500: the user terminal acquires state evaluation indexes of all base stations in the target area;

[0027] S600: the user terminal acquires connection indexes between the user terminal and all base stations in the target area by using the state evaluation indexes of all base stations and real-time distances between the user terminal and all base stations; and the base station corresponding to the maximum value in the connection indexes is determined as the specified base station.

[0028] As a further scheme of the present application, the data transmitted by the base station includes real-time delay time, real-time packet loss rate, real-time number of connected user terminals, and real-time average transmission rate.

[0029] As a further scheme of the present application, the state evaluation index of each base station is calculated by the following formula:

[0030]

[0031] The state evaluation index S of the base station numbered i is calculated by the following formula: i ;

[0032] Wherein, Q Link is the number of connected user terminals of the base station numbered i; Q0 is a preset number of connected user terminals of the base station numbered i; t s is real-time delay time of the base station numbered i; t0 is a preset delay time of the base station numbered i; is real-time packet loss rate of the base station numbered i; is a preset packet loss rate of the base station numbered i; v c is real-time average transmission rate of the connected user terminals of the base station numbered i; v0 is a preset average transmission rate of the connected user terminals of the base station numbered i; ε is a unit coefficient; C1 is a first preset constant; C2 is a second preset constant; C3 is a third preset constant; C4 is a fourth preset constant; γ1 is a first weight coefficient; γ2 is a second weight coefficient; γ3 is a third weight coefficient; γ4 is a fourth weight coefficient; and S0 is a basic state evaluation index.

[0033] As a further scheme of the present application, the connection index is calculated by the following formula:

[0034]

[0035] The connection index L i between the user terminal and the base station numbered i in the target area is calculated by the following formula:

[0036] Wherein, f(X) is a judgment function, when X>=0, f(X)=X; when X<0, f(X)=0; (X i , Y i , Z i ) is the coordinate of the base station numbered i; (X0, Y0, Z0) is the coordinate of the user terminal; R i is the preset coverage radius of the base station numbered i; sigma i is the building density of the user terminal and the base station numbered i.

[0037] As a further scheme of the application: the cellular mode includes OMA mode and NOMA mode; the two-hop D2D mode includes D2D-assisted OMA mode and D2D-assisted NOMA mode.

[0038] The beneficial effects of the application are:

[0039] (1) The application combines non-orthogonal multiple access (NOMA) and device-to-device (D2D) communication technology in dense cellular networks, D2D communication technology allows adjacent devices to exchange data directly without passing through the base station, which significantly reduces network delay, reduces the burden of the base station, and improves the local multiplexing efficiency of spectrum resources, ensuring the network quality of each user terminal; NOMA technology breaks the limitations of traditional orthogonal multiple access, allowing signals to be superimposed on non-orthogonal resources such as power domain, code domain, or spatial domain for transmission, enabling multiple users to share resources at the same time, frequency, or code channel, thereby significantly improving user connection density and spectrum utilization. By combining these two technologies, not only can the communication performance of the system be significantly improved, but also can effectively support large-scale device connection while meeting the different communication needs of HTCU and MTCD, effectively solving various technical problems that arise during integration, optimizing user mode selection, user grouping, and power allocation in dense cellular networks to improve the overall spectrum efficiency of the system, promoting the widespread application of 5G and future communication networks in the Internet of Things era, and providing users with a smoother and more intelligent communication experience. BRIEF DESCRIPTION OF DRAWINGS

[0040] The application will be further described below with reference to the accompanying drawings.

[0041] Figure 1 The method flowchart of an embodiment of the application. DETAILED DESCRIPTION

[0042] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described, obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative work are within the protection scope of the present application.

[0043] Please refer to Figure 1 As shown in the drawings, in one embodiment, a NOMA and D2D based dense cellular network performance optimization method is provided, applied to a dense cellular network system, the dense cellular network system including a plurality of base stations, a plurality of user terminals and a plurality of communication modes; the user terminal includes an HTCU and an MTCD; the optimization method includes the following steps:

[0044] S1: obtaining the real-time signal strength of the specified base station by setting a base station monitoring unit in the user terminal;

[0045] S2: obtaining the preset signal strength of the base station of the user terminal by establishing a Dec-POMDP model;

[0046] S3: comparing the real-time signal strength of the specified base station obtained by the user terminal with the preset signal strength, if the real-time signal strength is greater than the preset signal strength, selecting the cellular mode, the user terminal directly transmitting data to the specified base station through the cellular link; otherwise, entering step S4;

[0047] S4: the user terminal searches for the HTCU around it that has successfully connected to the specified base station as a potential relay device, from the searched HTCU, selecting the HTCU closest to the real-time device and successfully connected to the specified base station as the D2D relay, transmitting data through the two-hop D2D mode, the user terminal first sending data to the selected HTCU relay through the D2D link, the relay then forwarding the data to the specified base station through the cellular link; otherwise, the user terminal is in an interrupted state and cannot transmit data.

[0048] Through the technical solution, the cellular mode includes an OMA mode and a NOMA mode; the two-hop D2D mode includes a D2D-assisted OMA mode and a D2D-assisted NOMA mode; through deep integration of non-orthogonal multiple access (NOMA) and device-to-device (D2D) communication technology in a dense cellular network, the D2D communication technology allows adjacent devices to directly exchange data without passing through a base station, which significantly reduces network delay, relieves the burden of the base station, and improves the local multiplexing efficiency of the spectrum resource; the NOMA technology breaks the limitation of the traditional orthogonal multiple access, allows signals to be superimposed and transmitted on non-orthogonal resources such as the power domain, code domain, or space domain, realizes that multiple users share resources at the same time, frequency, or code channel, and greatly improves the user connection density and spectrum utilization; through the combination of the two technologies, the communication performance of the system can be significantly improved, and different communication requirements of the HTCU and the MTCD can be effectively supported, various technical problems in the integration process can be effectively solved, the user mode selection, user grouping, and power allocation in the dense cellular network are optimized to improve the overall spectrum efficiency of the system, promote the wide application of 5G and future communication networks in the Internet of Everything era, and bring users a smoother and more intelligent communication experience; any HTCU or MTCD selects the cellular mode only when the real-time signal strength from the designated base station is greater than the preset signal strength, otherwise, it selects the two-hop D2D mode, and selects a nearest HTCU (intermediate user) successfully connected to the base station as a relay to establish a D2D link; on the one hand, this communication mode can enable the HTCU or MTCD in the two-hop D2D mode to maintain a stable D2D link at a high average rate; on the other hand, from the perspective of energy efficiency, it can also reduce the UE (i.e., HTCU or MTCD) energy consumption in uplink transmission; if the cellular mode or the two-hop D2D mode is not selected, the HTCU or MTCD is in an interrupted state; the user can send data to the corresponding base station through a one-hop cellular link or a two-hop D2D link, wherein the user is the transmitter, and the base station and the HTCU as the D2D relay are the receivers; in the two-hop D2D uplink transmission process, the device first sends data to the nearby HTCU (D2D relay) through the D2D link, and then the HTCU forwards the data to the corresponding base station through the cellular link; the base and the user terminal are distributed using HPPP, the distribution of the HTCU and the MTCD is independent of the base, and is respectively modeled based on two independent HPPPs; based on the fact that the number of MTCDs is much larger than that of HTCUs in reality, and only a part of the MTCDs are active in each time slot, the present application considers that the MTCDs are randomly active in each time slot, and all the base, HTCU, and MTCD are equipped with a single antenna; the available channels are completely shared and reused in each cell.

[0049] As an embodiment of the present application, the specific method for obtaining the preset signal strength of the user terminal includes the following steps:

[0050] S10: Taking the base station as an intelligent agent, the base station takes the increase or decrease operation of the preset signal strength as an action, the real-time signal strength of the base station is the state of the base station, and the environment feedback reward is used to establish a Dec-POMDP model;

[0051] S20: Each base station initializes a Q table to store the Q values of all possible state-action pairs, and updates the Q values;

[0052] S30: The base station selects the optimal action in the Q table with a probability of ε, and randomly selects an action with a probability of 1-ε;

[0053] S40: The reward r is calculated by the formula t ;

[0054] Wherein, and are the rates of the jth HTCU to the ith base station link at the tth step and the t-1th step, respectively, and are the rates of the kth MTCD k to the ith base station link at the tth step and the t-1th step, respectively;

[0055] S50: Repeat steps S30-S50 until the training end condition is reached; each base station finds its optimal preset signal strength;

[0056] Through the above technical solution, the base station is taken as an intelligent agent, the base station takes the increase or decrease operation of the preset signal strength as an action, the real-time signal strength of the base station is the state of the base station, the environment feedback reward is used to establish a Dec-POMDP model: Then a Q table is initialized for each base station to store the Q values of all possible state-action pairs, and the Bellman equation is used to update the Q values; the base station selects the optimal action in the Q table with a probability of ε, and randomly selects an action with a probability of 1-ε; the reward r is calculated by the formula t ; is the rate difference of the jth HTCU to the ith base station link at the tth step and the t-1th step; is the rate difference of the kth MTCD k to the ith base station link at the tth step and the t-1th step; is the sum of the rate differences of all links at the tth step and the t-1th step; when the sum of the rate differences of all links at the tth step and the t-1th step ​​The reward is positive; otherwise, the reward is negative; steps S30-S50 are repeated until the training end condition is reached; each base station finds its optimal preset signal strength; by introducing an intermittent learning strategy, multi-agent reinforcement learning is applied to dense cellular networks to cope with the burstiness of MTCD traffic. Within each training step, the base station only learns and updates the preset signal strength based on local observation information at the beginning, and maintains this threshold in subsequent time slots until the end of the step to obtain a reward based on overall performance. This method can effectively reduce unnecessary adjustments, improve system stability and performance optimization efficiency;

[0057] It should be noted that the method for updating the Q value is an existing technology and will not be described in detail here;

[0058] As an embodiment of the present invention, the process of determining the designated base station is as follows:

[0059] S100: Each base station obtains base station transmission data through a base station monitoring unit;

[0060] S200: Each base station obtains a status assessment index of each base station by analyzing base station transmission data;

[0061] S300: Determine a target area of ​​the user terminal with the user terminal as the center;

[0062] S400: Number all base stations in the target area, and record the number as i;

[0063] S500: The user terminal obtains the status evaluation index of each base station in the target area;

[0064] S600: Obtaining a connection index between the user terminal and each base station by analyzing the status evaluation index of each base station and the real-time distance between the user terminal and each base station; determining the base station corresponding to the maximum value among the connection indexes as the designated base station;

[0065] Through the above technical solution, each base station in this embodiment obtains base station transmission data through the base station monitoring unit; each base station obtains the status evaluation index of each base station by analyzing the base station transmission data; then, with the user terminal as the center, the target area of ​​the user terminal is determined; all base stations in the target area are numbered, and the number is recorded as i; then the user terminal obtains the status evaluation index of each base station in the target area; finally, the connection index between the user terminal and each base station is obtained by comparing the status evaluation index of each base station and the real-time distance between the user terminal and each base station; the base station corresponding to the maximum value in the connection index is determined as the designated base station; since the current base stations are established relatively closely, there will be multiple base stations in the target area of ​​the user terminal. By calculating the connection index, it is ensured that the base station can be connected to a more suitable base station, thereby improving the quality of network connection.

[0066] As an embodiment of the present application, the base station transmitting data comprises real-time delay time, real-time packet loss rate, real-time number of connected user terminals and real-time average transmission rate;

[0067] It should be noted that the real-time delay time, real-time packet loss rate, real-time number of connected user terminals and real-time average transmission rate are obtained by a base station network monitoring device, and the specific obtaining process is prior art, which is not described in detail here.

[0068] As an embodiment of the present application, the state evaluation index of each base station is obtained by the formula:

[0069]

[0070] The state evaluation index S of the base station numbered i is calculated i ;

[0071] Wherein, Q Link is the number of connected user terminals of the base station numbered i; Q0 is the preset number of connected user terminals of the base station numbered i; t s is the real-time delay time of the base station numbered i; t0 is the preset delay time of the base station numbered i; is the real-time packet loss rate of the base station numbered i; is the preset packet loss rate of the base station numbered i; v c is the real-time average transmission rate of the connected user terminals of the base station numbered i; v0 is the preset average transmission rate of the connected user terminals of the base station numbered i; ε is a unit coefficient; C1 is a first preset constant; C2 is a second preset constant; C3 is a third preset constant; C4 is a fourth preset constant; γ1 is a first weight coefficient; γ2 is a second weight coefficient; γ3 is a third weight coefficient; γ4 is a fourth weight coefficient; S0 is a basic state evaluation index;

[0072] Through the above technical solution, the embodiment Q Link -Q0 is the first difference value between the number of connected user terminals of the base station numbered i and the preset number of connected user terminals; when the first difference value Q Link -Q0<0, it indicates that the number of connected user terminals of the base station numbered i does not exceed the preset number of connected user terminals, so the first difference value Q Link -Q0 is greater, the better the state of the base station, and the smaller the state evaluation index S i ; when the first difference value Q Link -Q0>0, it indicates that the number of connected user terminals of the base station numbered i exceeds the preset number of connected user terminals, so the first difference value Q Link -Q0 is greater, the worse the state of the base station, and the larger the state evaluation index S iThe greater t s -t0 is a second difference value between the real-time delay time of the base station numbered i and the preset delay time; when the second difference value t s When -t0<0, it indicates that the real-time delay time of the base station numbered i does not exceed the preset delay time, so the second difference value t s The greater the absolute value of -t0 is, the better the state of the base station is, and the smaller the state evaluation index S i The smaller the second difference value t s When -t0>0, it indicates that the real-time delay time of the base station numbered i exceeds the preset delay time, so the second difference value t s The greater -t0 is, the worse the state of the base station is, and the greater the state evaluation index S i The greater the second difference value t is a third difference value between the real-time packet loss rate of the base station numbered i and the preset packet loss rate; when the third difference value When -v0<0, it indicates that the real-time packet loss rate of the base station numbered i does not exceed the preset packet loss rate, so the third difference value The greater the absolute value of -v0 is, the better the state of the base station is, and the smaller the state evaluation index S i The smaller the third difference value When -v0>0, it indicates that the real-time packet loss rate of the base station numbered i exceeds the preset packet loss rate, so the third difference value The greater -v0 is, the worse the state of the base station is, and the greater the state evaluation index S i The greater the third difference value v c -v0 is a fourth difference value between the real-time average transmission rate of the base station numbered i and the preset average transmission rate; when the fourth difference value v c When -v0<0, it indicates that the real-time average transmission rate of the base station numbered i does not exceed the preset average transmission rate, so the fourth difference value v c The greater the absolute value of -v0 is, the better the state of the base station is, and the smaller the state evaluation index S i The smaller the fourth difference value v c When -v0>0, it indicates that the real-time average transmission rate of the base station numbered i exceeds the preset average transmission rate, so the fourth difference value v c The greater -v0 is, the worse the state of the base station is, and the greater the state evaluation index S i The greater the fourth difference value v

[0073] It should be noted that the preset number Q0 of connected user terminals of the base station numbered i, the preset delay time t0 of the base station numbered i, the preset packet loss rate The preset average transmission rate v0, the unit coefficient ε, the first preset constant C1, the second preset constant C2, the third preset constant C3, the fourth preset constant C4, the first weight coefficient γ1, the second weight coefficient γ2, the third weight coefficient γ3, the fourth weight coefficient γ4 and the basic status assessment index S0 of the connected user terminals of the base station numbered i are preset values, which are obtained based on experience and will not be described in detail here.

[0074] As an embodiment of the present invention, the connectivity index is obtained by the formula:

[0075]

[0076] Calculate the connection index L between the user terminal and the base station numbered i in the target area i ;

[0077] Among them, f(X) is the judgment function. When X≥0, f(X)=X; when X<0, f(X)=0; (X i , Y i , Z i ) are the coordinates of the base station numbered i; (X0, Y0, Z0) are the coordinates of the user terminal; R i is the preset coverage radius of the base station numbered i; σ i is the building density between user terminals and base station numbered i;

[0078] Through the above technical solution, this embodiment is the distance between the base station numbered i and the user terminal; is the ratio of the distance between the base station numbered i and the user terminal to the preset coverage radius of the base station numbered i; x in the judgment function f(X) is when When , it indicates the ratio of the distance between the base station numbered i and the user terminal to the preset coverage radius of the base station numbered i is greater than 1, the distance between the base station numbered i and the user terminal exceeds the preset coverage radius of the base station numbered i, so Connection index L i =0; when 1- When , it indicates the ratio of the distance between the base station numbered i and the user terminal to the preset coverage radius of the base station numbered i is less than or equal to 1, the distance between the base station numbered i and the user terminal is within the preset coverage radius of the base station numbered i, so The smaller the distance between the base station numbered i and the user terminal, the smaller the connection index L. iThe greater the building density σ of the user terminal and the base station numbered i i The specific analysis method is prior art and is not described here; the greater the building density σ of the user terminal and the base station numbered i i The greater the attenuation, and thus the connection index L i The smaller the distance between the base station numbered i and the user terminal, the smaller the connection index L i The greater the building density σ of the user terminal and the base station numbered i i The specific analysis method is prior art and is not described here; the greater the building density σ of the user terminal and the base station numbered i i The greater the attenuation, and thus the state evaluation index S i The greater the attenuation, and thus the state evaluation index S i The smaller the distance between the base station numbered i and the user terminal, the smaller the connection index L

[0079] It should be noted that the coordinates (X i , Y i , Z i ) of the base station numbered i, the coordinates (X0, Y0, Z0) of the user terminal, the building density σ of the user terminal and the base station numbered i i are obtained by prior art; the preset coverage radius R i of the base station numbered i is a preset value and is not described here.

[0080] The above describes one embodiment of the present application in detail, but the content described is only a preferred embodiment of the present application and cannot be considered as limiting the scope of the present application. Any equivalent changes and improvements made in the scope of the present application should still belong to the patent coverage of the present application.

Claims

1.A method for performance optimization of NOMA and D2D based dense cellular networks, applied to a dense cellular network system, the dense cellular network system comprising a plurality of base stations, a plurality of user terminals and a plurality of communication modes; characterized in that, The optimization method comprises the following steps: S1: acquiring real-time signal strength of a specified base station by setting a base station monitoring unit on a user terminal; S2: acquiring preset signal strength of the base station of the user terminal by establishing a Dec-POMDP model, wherein the user terminal comprises an HTCU and an MTCD, the HTCU is a human type communication user, and the MTCD is a machine type communication device; S3: comparing the real-time signal strength of the specified base station acquired by the user terminal with the preset signal strength, if the real-time signal strength is greater than the preset signal strength, selecting a cellular mode, and the user terminal directly transmitting data to the specified base station through a cellular link; otherwise, entering step S4; S4: searching, by the user terminal, surrounding HTCUs successfully connected to the specified base station as potential relay devices, selecting, from the searched HTCUs, an HTCU closest to the user terminal and successfully connected to the specified base station as a D2D relay, and transmitting data through a two-hop D2D mode, wherein the user terminal first sends data to the selected HTCU relay through a D2D link, and the relay then forwards the data to the specified base station through a cellular link; otherwise, the user terminal is in an interrupted state; The specific method for acquiring the preset signal strength of the base station of the user terminal comprises the following steps: S10: taking the base station as an agent, taking an increase or decrease operation of the preset signal strength as an action, taking real-time signal strength of the base station as a state of the base station, and establishing a Dec-POMDP model by feeding back a reward of an environment to the base station; S20: initializing a Q table for each base station, storing Q values of all possible state-action pairs, and updating the Q values; S30: selecting an optimal action in the Q table with an ε probability, and randomly selecting an action with a 1-ε probability; S40: Calculate the reward by the formula Calculate the reward ; wherein, and Rj,i(t) and Rj,i(t-1) are the rates of the jth HTCU to the ith base station link at the tth and t-1th steps, respectively, and Rk,i(t) and Rk,i(t-1) are the rates of the kth MTCD to the ith base station link at the tth and t-1th steps, respectively. S50: repeating steps S30-S50 until a training end condition is reached; and each base station finds an optimal preset signal strength. 2.The method of claim 1, wherein, The determination process of the specified base station is as follows: S100: acquiring, by the user terminal, base station transmission data through the base station monitoring unit; S200: acquiring, by the user terminal, a state evaluation index of each base station by analyzing the base station transmission data; S300: determining a target area of the user terminal as a center; S400: Number all base stations within the target region, and the number is recorded as ; S500: acquiring, by the user terminal, the state evaluation index of each base station in the target area; S600: acquiring a connection index of the user terminal and each base station by the state evaluation index of each base station and real-time distance between the user terminal and each base station; and determining a base station corresponding to a maximum value in the connection index as the specified base station. 3.The method of claim 2, wherein, The base station transmission data comprises real-time delay time, real-time packet loss rate, real-time number of connected user terminals, and real-time average transmission rate. 4.The method of claim 3, wherein, The state evaluation index of each base station is obtained by the following formula: ； A state evaluation index of a base station numbered ;​ wherein, the number of connected user terminals of a base station numbered ; a preset number of connected user terminals of a base station numbered ; a real-time delay time of a base station numbered ; a preset delay time of a base station numbered ; a real-time packet loss rate of a base station numbered ; a preset packet loss rate of a base station numbered ; a real-time average transmission rate of connected user terminals of a base station numbered ; a preset average transmission rate of connected user terminals of a base station numbered ; is a unit coefficient; is a first preset constant; is a second preset constant; is a third preset constant; is a fourth preset constant; is a first weight coefficient; is a second weight coefficient; is a third weight coefficient; is a fourth weight coefficient; is a basic status evaluation index. 5.The method for performance optimization of dense cellular networks based on NOMA and D2D of claim 4, wherein, The connection index is obtained by the following formula: ； calculating a connection index of the user terminal with a base station numbered in the target area ; wherein is a judging function, when , ; when , ; is the coordinate of the base station numbered ; is the coordinate of the user terminal; is the preset coverage radius of the base station numbered ; is the building density of the user terminal and the base station numbered . 6.The method of claim 1, wherein, The cellular mode comprises an OMA mode and a NOMA mode; and the two-hop D2D mode comprises a D2D assisted OMA mode and a D2D assisted NOMA mode.

Images