Method for maximizing total throughput of Internet of Things assisted by multiple unmanned aerial vehicles

Abstract

The invention belongs to the technical field of wireless transmission, and discloses a method for maximizing the total throughput of an internet of things assisted by multiple unmanned aerial vehicles, which comprises the following steps of: acquiring position information of cellular edge users and D2D (Device-to-Device) equipment; constructing a total throughput maximization problem of the multi-unmanned aerial vehicle assisted Internet of Things, and decomposing the maximization problem into two sub-problems; the relay selection and time distribution sub-problems of the D2D pair are solved; the three-dimensional deployment sub-problem of the unmanned aerial vehicle is solved; and alternately optimizing the two sub-problems until the objective function of the total throughput maximization problem of the multi-UAV-assisted Internet of Things converges, and obtaining a final optimization result. And according to the final optimization result, realizing the maximization of the total throughput of the multi-unmanned-aerial-vehicle-assisted Internet of Things. According to the invention, the total throughput of the multi-unmanned-aerial-vehicle-assisted Internet of Things is large.

Description

Technical Field

[0001] The present invention belongs to the technical field of wireless transmission, and in particular relates to a method for maximizing the total throughput of a multi-unmanned aerial vehicle-assisted Internet of Things. Background Art

[0002] With the rapid development of the Internet of Things, the number of IoT devices connected to cellular networks has grown significantly, and IoT devices have been widely used in society. However, due to the obstacles of long-distance communication and dense buildings, the services obtained by IoT devices at the edge of the cell are not satisfactory.

[0003] In order to improve the coverage and throughput performance of the Internet of Things, drones have attracted widespread attention due to their advantages of high maneuverability and flexibility. Specifically, multiple drones are deployed in the air as relays, with a large probability of line-of-sight transmission with ground-based IoT devices. In addition, in order to support large-scale connections between IoT devices and drones, non-orthogonal multiple access (NOMA) technology is used, allowing multiple users to transmit data on the same spectrum with different receiving power levels, further improving the performance of the Internet of Things.

[0004] In multi-drone-assisted IoT, the number of drones acting as relays is limited, and relay competition among IoT devices needs to be considered. At the same time, the three-dimensional deployment of drones will affect the channel fading of multiple links at the same time. Therefore, in order to maximize the total throughput of multi-drone-assisted IoT, it is crucial to reasonably optimize the relay selection of IoT devices, resource allocation, and three-dimensional deployment of drones. Summary of the invention

[0005] The object of the present invention is to provide a method for maximizing the total throughput of a multi-UAV-assisted Internet of Things, so as to maximize the total throughput of the multi-UAV-assisted Internet of Things while satisfying the minimum throughput constraint of the Internet of Things devices and the channel signal-to-noise ratio constraint.

[0006] To achieve the above object, the technical solution adopted by the present invention is:

[0007] A method for maximizing the total throughput of a multi-drone-assisted Internet of Things, the multi-drone-assisted Internet of Things comprising a multi-antenna base station, U single-antenna cellular edge users, M single-antenna drones, N pairs of D2D pairs, each pair of D2D pairs comprising a D2D transmitter and a D2D receiver, wherein the drone acts as a relay device, relaying data from the base station to the cellular edge user, and simultaneously relaying data from the D2D transmitter to the D2D receiver, the method for maximizing the total throughput of the multi-drone-assisted Internet of Things comprising:

[0008] Step 1: The base station obtains the location information of U cell edge users and N pairs of D2D pairs, divides the cell edge users into M clusters according to the location information, and designates one drone for each cluster to relay the data sent by the base station to the cell edge users of the cluster;

[0009] Step 2: Construct the total throughput maximization problem of multi-UAV-assisted IoT and decompose it into two sub-problems: the total throughput maximization problem of multi-UAV-assisted IoT under given three-dimensional deployment of UAVs and the total throughput maximization problem of multi-UAV-assisted IoT under given D2D relay selection and time allocation;

[0010] Step 3: To solve the problem of maximizing the total throughput of multi-UAV-assisted IoT under a given three-dimensional deployment of UAVs, the relay selection of D2D pairs, the time allocation of D2D transmitters in the NOMA transmission phase, and the time allocation of D2D receivers in the D2D transmission phase are optimized according to the current three-dimensional deployment of UAVs.

[0011] Step 4: Solve the problem of maximizing the total throughput of multi-UAV-assisted IoT given by D2D pair relay selection and time allocation, and optimize the three-dimensional deployment of UAVs according to the current D2D pair relay selection and time allocation;

[0012] Step 5: Repeat steps 3 and 4 to optimize the two sub-problems alternately until the objective function of the total throughput maximization problem of multi-UAV-assisted IoT converges;

[0013] Step 6: Maximize the total throughput of multi-UAV-assisted IoT based on the final optimization results, which include the three-dimensional deployment of UAVs, the relay selection of D2D pairs, the time allocation of D2D transmitters in the NOMA transmission phase, and the time allocation of D2D receivers in the D2D transmission phase.

[0014] Several optional methods are also provided below, but they are not intended to be additional limitations on the above-mentioned overall solution, but are merely further supplements or preferences. Under the premise that there are no technical or logical contradictions, each optional method can be combined with the above-mentioned overall solution separately, and multiple optional methods can also be combined.

[0015] Preferably, the total throughput maximization problem of constructing a multi-UAV assisted Internet of Things includes:

[0016] The throughput R of the transmission from the base station to the mth drone B,m yes:

[0017]

[0018] where m∈{1,2,...M}, τm,n represents the time for the nth D2D transmitter to transmit data to the mth UAV in the NOMA transmission phase, that is, the time allocation of the D2D transmitter in the NOMA transmission phase, W is the wireless channel bandwidth, P B is the base station’s transmit power, h B,m is the channel gain between the base station and the mth UAV, P N is the transmission power of the D2D transmitter, h n,m is the channel gain between the nth D2D transmitter and the mth UAV, σ 2 is the additive white Gaussian noise power;

[0019] The throughput R of the transmission from the nth D2D transmitter to the mth UAV n,m yes:

[0020]

[0021] The throughput R of the transmission from the mth UAV to the kth cell edge user of the corresponding cluster m,k yes:

[0022]

[0023] where k∈{1,2,...K m}, K m is the total number of cell edge users in the mth cluster, t m,k represents the transmission time of the mth UAV to the kth cell edge user of the corresponding cluster, P u represents the transmitting power of the UAV, h m,k represents the channel gain between the mth UAV and the kth cell edge user of the corresponding cluster;

[0024] The transmission throughput from the mth UAV to the nth D2D receiver is R m,n+N yes:

[0025]

[0026] Where n+n represents the nth D2D receiver, represents the time when the mth UAV transmits data to the nth D2D receiver in the D2D transmission phase, that is, the time allocation of the D2D receiver in the D2D transmission phase, h m,n+M represents the channel gain between the mth UAV and the nth D2D receiver;

[0027] The throughput R of the transmission from the base station to the kth cell edge user of the corresponding cluster through the relay of the mth drone is B,m,k yes:

[0028]

[0029] The throughput R of the transmission from the nth D2D transmitter to the nth D2D receiver via the relay of the mth UAV is n,m,n+N yes:

[0030] R n,m,n+N =min{R n,m ,R m,n+N = R n,m

[0031] The total throughput of multi-UAV-assisted IoT R sum for:

[0032]

[0033] Then the objective function of the total throughput maximization problem of multi-UAV-assisted IoT is maxR sum .

[0034] Preferably, the method of optimizing the relay selection of D2D pairs, the time allocation of D2D transmitters in the NOMA transmission phase, and the time allocation of D2D receivers in the D2D transmission phase according to the current three-dimensional deployment of drones includes:

[0035] Step 3.1: Based on the location information of the D2D pairs and the drones, calculate the channel gains between all D2D pairs and all drones. If the channel gain between the nth D2D transmitter or the nth D2D receiver and the mth drone does not meet the minimum channel gain constraint, the nth D2D pair cannot select the mth drone as a relay device. Update the channel gain h between the nth D2D transmitter and the mth drone. n,m =0;

[0036] Step 3.2: Set a probability vector P for each D2D transmitter n =(p n,1 ,p n,2 ,...,p n,M ), where p n,m represents the probability that the nth D2D transmitter selects the mth drone as a relay, initialized

[0037] Step 3.3: Each D2D transmitter receives a signal according to the probability vector P n Select the drone as the relay device and obtain the relay selection information;

[0038] Step 3.4: Based on the relay selection information obtained in step 3.3, use the greedy time allocation method to optimize the time allocation of D2D transmitters in the NOMA transmission phase: In the NOMA transmission phase, allocate time that satisfies the minimum throughput constraint to all D2D transmitters. For the mth drone, if Less than NOMA transmission phase time T 1 , then the NOMA transmission phase time T 1 The remaining time in is all allocated to the D2D transmitter with the largest channel gain with the mth UAV; otherwise, the time allocation τ of all D2D transmitters of the mth UAV is updated. m,n =0, where τ m,n It represents the time when the nth D2D transmitter transmits data to the mth UAV in the NOMA transmission phase;

[0039] Step 3.5: In the D2D transmission phase, according to R n,m =R n,m,n+N , calculate the time allocation of the D2D receiver, if Greater than the D2D transmission phase time θT 1 , then update the time allocation τ of all D2D transmitters of the mth UAV m,n =0; if Less than or equal to the D2D transmission phase time θT 1 , then the time allocation τ of all D2D transmitters of the mth UAV is selected m,n remains unchanged; among them, R n,m represents the transmission throughput from the nth D2D transmitter to the mth UAV, R n,m,n+N represents the throughput of transmission from the nth D2D transmitter to the nth D2D receiver through the relay of the mth UAV, It indicates the time when the mth UAV transmits data to the nth D2D receiver in the D2D transmission phase;

[0040] Step 3.6: Calculate the total throughput R of multi-UAV-assisted IoT under the current relay selection and time allocation scheme sum (t), and according to the current total throughput R sum (t) Update the probability vector of all D2D transmitters, t is the current iteration number: use R max Indicates the maximum throughput, the initial value is 0, if the current total throughput R sum (t) is greater than the maximum throughput R max , update R max =R sum (t), and the probability vector at the t+1th iteration is updated as:

[0041]

[0042] where p n,m (t+1) represents the probability that the nth D2D transmitter selects the mth UAV as a relay after update, p n,m (t) represents the probability that the nth D2D transmitter selects the mth UAV as a relay before updating, and a n (t) represents the UAV selected by the nth D2D transmitter in the tth iteration, ω 1 represents the first learning step, represents the normalized reward;

[0043] Otherwise, the probability vector at the t+1th iteration is updated as:

[0044]

[0045] where ω 2 represents the second learning step, m n,max represents the optimal relay selection of the nth D2D transmitter;

[0046] Step 3.7: Repeat steps 3.3 to 3.6 until the objective function of the total throughput maximization problem of multi-UAV-assisted IoT given the three-dimensional deployment of UAVs converges.

[0047] Preferably, the method of optimizing the three-dimensional deployment of drones according to the current D2D relay selection and time allocation includes:

[0048] The sub-problem of D2D relay selection and time allocation for the total throughput maximization of a given multi-UAV-assisted IoT is equivalently expressed as a three-dimensional deployment optimization problem for each UAV. The following optimization solution steps are performed for each UAV:

[0049] Step 4.1: According to the population size S and the chaotic iteration step, the whales are initialized based on Bernoulli shift chaos. The three-dimensional position of each whale represents the three-dimensional deployment position of the drone.

[0050] Step 4.2: Calculate the total throughput of the relay transmission through the drone as the fitness value of each whale, take the whale with the highest fitness value as the optimal whale, and record the three-dimensional position of the optimal whale. The fitness value calculation formula is as follows:

[0051]

[0052] Where R sum,m represents the total throughput of relay transmission through the mth UAV, K m is the total number of cell edge users in the mth cluster, R B,m,krepresents the throughput of transmission from the base station to the kth cell edge user of the corresponding cluster through the relay of the mth drone, R n,m represents the throughput of transmission from the nth D2D transmitter to the mth UAV;

[0053] Step 4.3: Update the adaptive convergence factor a according to the current iteration number t:

[0054]

[0055] where a max and a mim They represent the maximum and minimum values ​​of the adaptive convergence factor, w represents the adjustment factor, and t max Indicates the maximum number of iterations;

[0056] Update coefficients A and C according to the adaptive convergence factor a:

[0057] C=2r

[0058] A=2ar-a

[0059] Where r∈[0,1] is a random number;

[0060] Step 4.4: Generate a random number p∈[0,1). If p<0.5 and A≤1, use the shrinking and surrounding strategy to update the three-dimensional position of the whale:

[0061] X(t+1)=X * (t)-A|CX * (t)-X(t)|

[0062] Where X(t+1) represents the updated 3D position of the whale, X * (t) represents the 3D position of the optimal whale in the tth iteration, and X(t) represents the 3D position of the whale before the update;

[0063] If p<0.5 and A>1, the prey search strategy based on the Levy flight mechanism is used to update the three-dimensional position of the whale:

[0064]

[0065] Where X rand (t) represents the 3D position of the randomly selected whale in the t-th iteration, β represents the step size control parameter, represents point multiplication, Levy(s) represents the random search path of the prey search strategy based on the Levy flight mechanism and follows the Levy distribution;

[0066] If p ≥ 0.5, the bubble net attack strategy is used to update the whale's position:

[0067] X(t+1)=|X * (t)-X(t)|e bl cos(2πl)+X * (t)

[0068] Where b is a constant representing the shape of the spiral, and l∈[-1,1] is a random number;

[0069] Step 4.5: Calculate the reverse solution for all whales:

[0070] X′ i,j =λ(x min,j +x max,j )-X i,j

[0071] Where X i,j represents the reverse solution of the position of whale i in dimension j, λ∈[0,1] is a random number, X i,j represents the position of whale i in dimension j, x min,j and x max,j They represent the lower and upper bounds of the three-dimensional position of the whale in dimension j respectively;

[0072] Sort the three-dimensional positions of all whales and their reverse solutions according to their fitness values, and select the first S whales as the next generation population;

[0073] Step 4.6: Repeat steps 4.2 to 4.5 until the total throughput R of the relay transmission through the UAV is sum,m convergence.

[0074] Preferably, the step 5 alternately optimizes the two sub-problems until the objective function of the total throughput maximization problem of multi-UAV-assisted Internet of Things converges, including:

[0075] The two sub-problems are optimized alternately until the convergence values ​​of the objective functions of the two sub-problems are consistent, then the alternating optimization of the two sub-problems is terminated.

[0076] The present invention provides a method for maximizing the total throughput of a multi-UAV-assisted Internet of Things. A D2D device selects a relay device through a probability-based relay selection and a greedy time allocation method, thereby solving the relay competition problem among multiple devices. The relay selection probability vector of the D2D device is updated based on the total throughput, thereby achieving a larger total throughput. The UAVs are deployed in three dimensions through an improved whale optimization method, which introduces Bernoulli shift initialization, an adaptive convergence factor, a Levy flight mechanism, and an elite reverse learning mechanism. Compared with traditional optimization methods, a larger total throughput can be achieved. By alternately optimizing two sub-problems, the total throughput maximization problem of the multi-UAV-assisted Internet of Things can be well solved, thereby achieving a larger total throughput. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] Figure 1 This is a schematic diagram of the structure of the multi-UAV assisted Internet of Things of the present invention;

[0078] Figure 2 It is a schematic diagram of the time slot structure of the present invention;

[0079] Figure 3 The present invention is a flow chart of a method for maximizing the total throughput of a multi-UAV-assisted Internet of Things. DETAILED DESCRIPTION

[0080] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0081] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in the specification of the present invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention.

[0082] like Figure 1 As shown, the multi-drone assisted Internet of Things of this embodiment includes 1 multi-antenna base station (BS), U single-antenna cellular edge users, M single-antenna drones (UAVs), N pairs of device-to-device (D2D) pairs, each pair of D2D pairs includes 1 D2D transmitter and 1 D2D receiver, where the drone acts as a relay device, relaying data from the base station to the cellular edge user, and relaying data from the D2D transmitter to the D2D receiver. In order to avoid mutual interference between the relay transmissions of the M drones, the wireless channel bandwidth is divided into M orthogonal channels with bandwidths of W.

[0083] like Figure 2 As shown, each time slot T includes a NOMA transmission phase (duration T 1 ), the cell edge user transmission phase (duration is εT 1 ) and D2D transmission phase (duration is θT 1 ). In the NOMA transmission phase, the base station transmits data to the M drones through M orthogonal channels. At the same time, the D2D transmitters transmit data to the drones in turn. In the cellular edge user transmission phase, the drone relays the data received from the base station to the cellular edge users. In the D2D transmission phase, the drone relays the data received from the D2D transmitter to the D2D receiver.

[0084] like Figure 3 As shown, a method for maximizing the total throughput of a multi-UAV-assisted Internet of Things in this embodiment includes the following steps:

[0085] Step 1: The base station obtains the location information of U cell edge users and N pairs of D2D pairs, and divides the cell edge users into M clusters according to the location information. The mth cluster has K m Cell edge users are assigned to each cluster, and one drone is assigned to relay the data sent by the base station to the cell edge users of the cluster.

[0086] Step 2: Construct the total throughput maximization problem of multi-UAV-assisted Internet of Things, and decompose the total throughput maximization problem of multi-UAV-assisted Internet of Things into two sub-problems, namely the total throughput maximization problem of multi-UAV-assisted Internet of Things under given three-dimensional deployment of UAVs and the total throughput maximization problem of multi-UAV-assisted Internet of Things under given D2D relay selection and time allocation.

[0087] Set the binary variable α m,n Represents the relay selection of the D2D pair. If the nth D2D pair selects the mth drone as the relay device, then α m,n =1, otherwise α m,n = 0. Time variable τ m,n It represents the time when the nth D2D transmitter transmits data to the mth UAV in the NOMA transmission phase; represents the time for the mth UAV to transmit data to the nth D2D receiver in the D2D transmission phase. If the nth D2D pair selects other UAVs except the mth UAV, then τ m,n =0, In the NOMA transmission phase, the base station transmits the data of the cell edge user in the mth cluster to the mth drone through one antenna, and the D2D transmitter that selects the mth drone as a relay also transmits data to the mth drone in turn; since the received signal strength from the base station is usually greater than the received signal strength from the D2D transmitter, the mth drone first decodes the signal from the base station and then decodes the signal from the D2D transmitter. Among them, the total throughput maximization problem of multi-drone assisted IoT is constructed as follows:

[0088] The throughput R of the transmission from the base station to the mth drone B,m yes:

[0089]

[0090] where m∈{1,2,...M}, τ m,n represents the time for the nth D2D transmitter to transmit data to the mth UAV in the NOMA transmission phase, that is, the time allocation of the D2D transmitter in the NOMA transmission phase, W is the wireless channel bandwidth, P B is the base station’s transmit power, h B,m is the channel gain between the base station and the mth UAV, P N is the transmission power of the D2D transmitter, h n,m is the channel gain between the nth D2D transmitter and the mth UAV, σ 2 is the additive white Gaussian noise power.

[0091] The throughput R of the transmission from the nth D2D transmitter to the mth UAV n,m yes:

[0092]

[0093] The throughput R of the transmission from the mth UAV to the kth cell edge user of the corresponding cluster m,k yes:

[0094]

[0095] where k∈{1,2,...K m}, K m is the total number of cell edge users in the mth cluster, t m,k represents the transmission time of the mth UAV to the kth cell edge user of the corresponding cluster, P U represents the transmitting power of the UAV, h m,k represents the channel gain between the mth UAV and the kth cell edge user of the corresponding cluster.

[0096] The transmission throughput from the mth UAV to the nth D2D receiver is Rm,n+N yes:

[0097]

[0098] Where n+N represents the nth D2D receiver, represents the time when the mth UAV transmits data to the nth D2D receiver in the D2D transmission phase, that is, the time allocation of the D2D receiver in the D2D transmission phase, h m,n+N represents the channel gain between the mth UAV and the nth D2D receiver.

[0099] According to the barrel theory, the transmission throughput R from the base station to the kth cell edge user of the corresponding cluster through the relay of the mth drone is B,m,k yes:

[0100]

[0101] In order to successfully relay the data of the nth D2D transmitter to the nth D2D receiver, the throughput of the transmission from the nth D2D transmitter to the mth UAV must not exceed the throughput of the transmission from the mth UAV to the nth D2D receiver. The throughput R of the transmission from the nth D2D transmitter to the nth D2D receiver through the relay of the mth UAV is n,m,n+N yes:

[0102] R n,m,n+N =min{R n,m ,R m,n+N = R n,m

[0103] The total throughput of multi-UAV-assisted IoT R sum for:

[0104]

[0105] Then the objective function of the total throughput maximization problem of multi-UAV-assisted IoT is maxR sum .

[0106] Step 3: Aiming at solving the total throughput maximization problem of multi-UAV-assisted IoT under a given three-dimensional deployment of UAVs, according to the current three-dimensional deployment of UAVs, the relay selection of D2D pairs, the time allocation of D2D transmitters in the NOMA transmission phase, and the time allocation of D2D receivers in the D2D transmission phase are optimized to solve the relay selection and time allocation problems of D2D pairs.

[0107] This embodiment determines the relay selection and time allocation of D2D pairs under the constraints of the minimum throughput of D2D devices and the channel signal-to-noise ratio, and proposes a relay selection and greedy time allocation method based on random learning according to the current geographical location of the drone.

[0108] Step 3.1: Based on the location information of the D2D pairs and the drones, calculate the channel gains between all D2D pairs and all drones. If the channel gain between the nth D2D transmitter or the nth D2D receiver and the mth drone does not meet the minimum channel gain constraint, the nth D2D pair cannot select the mth drone as a relay device. Update the channel gain h between the nth D2D transmitter and the mth drone. n,m =0.

[0109] Step 3.2: Set a probability vector P for each D2D transmitter n =(p n,1 ,p n,2 ,...,p n,M ), where p n,m represents the probability that the nth D2D transmitter selects the mth drone as a relay, initialized

[0110] Step 3.3: Each D2D transmitter receives a signal according to the probability vector P n Select the drone as the relay device and obtain the relay selection information, which completes the binary variable α m,n Assignment.

[0111] Step 3.4: Based on the relay selection information obtained in step 3.3, use the greedy time allocation method to optimize the time allocation of D2D transmitters in the NOMA transmission phase: In the NOMA transmission phase, allocate time that satisfies the minimum throughput constraint to all D2D transmitters. For the mth drone, if Less than NOMA transmission phase time T 1 , then the NOMA transmission phase time T 1 The remaining time in is all allocated to the D2D transmitter with the largest channel gain with the mth UAV; otherwise, the time allocation τ of all D2D transmitters of the mth UAV is updated. m,n =0.

[0112] Step 3.5: In the D2D transmission phase, according to R n,m =R n,m,n+N , calculate the time allocation of the D2D receiver, if Greater than the D2D transmission phase time θT 1 , then update the time allocation τ of all D2D transmitters of the mth UAV m,n =0; if Less than or equal to the D2D transmission phase time θT 1 , then the time allocation τ of all D2D transmitters of the mth UAV is selected m,n Remain unchanged.

[0113] Step 3.6: Calculate the total throughput R of multi-UAV-assisted IoT under the current relay selection and time allocation scheme sum (t), and according to the current total throughput R sum (t) Update the probability vector of all D2D transmitters, t is the current iteration number: use R max Indicates the maximum throughput, the initial value is 0, if the current total throughput R sum (t) is greater than the maximum throughput R max , update R max =R sum (t), and the probability vector at the t+1th iteration is updated as:

[0114]

[0115] where p n,m (t+1) represents the probability that the nth D2D transmitter selects the mth UAV as a relay after update, p n,m (t) represents the probability that the nth D2D transmitter selects the mth UAV as a relay before updating, and a n (t) represents the UAV selected by the nth D2D transmitter in the tth iteration, ω 1 represents the first learning step, represents the normalized reward.

[0116] Otherwise, the probability vector at the t+1th iteration is updated as:

[0117]

[0118] where ω 2 represents the second learning step, m n,max It represents the optimal relay selection of the nth D2D transmitter. The relay selection under the current maximum throughput is the current optimal relay selection.

[0119] Step 3.7: Repeat steps 3.3 to 3.6 until the objective function of the total throughput maximization problem of multi-UAV-assisted IoT given the three-dimensional deployment of UAVs converges.

[0120] Step 4: To solve the total throughput maximization problem of multi-UAV-assisted IoT given by D2D pair relay selection and time allocation, optimize the three-dimensional deployment of UAVs according to the current D2D pair relay selection and time allocation to solve the three-dimensional deployment problem of UAVs.

[0121] Since the transmission between drones and IoT devices will not interfere with the transmission between other drones and IoT devices, the total throughput maximization problem of multi-drone-assisted IoT under the given relay selection and time allocation of D2D can be equivalently expressed as a three-dimensional deployment optimization problem for each drone. Therefore, this embodiment proposes an improved whale optimization method, which performs the following operations for the three-dimensional deployment of each drone:

[0122] Step 4.1: According to the population size S and the chaotic iteration step, the whales are initialized based on Bernoulli shift chaos. The three-dimensional position of each whale represents the three-dimensional deployment position of a drone.

[0123] Step 4.2: Calculate the total throughput of the relay transmission through the drone as the fitness value of each whale, and optimize the objective function R through a single drone sum,m To evaluate the quality of the solution, the objective function R sum,m The larger the value of , the higher the quality of the solution. Therefore, the whale with the highest fitness value is taken as the optimal whale, and the three-dimensional position of the optimal whale is recorded. The fitness value calculation formula is as follows:

[0124]

[0125] Where R sum,m represents the total throughput of relay transmission through the mth UAV, K m is the total number of cell edge users in the mth cluster, R B,m,k represents the throughput of transmission from the base station to the kth cell edge user of the corresponding cluster through the relay of the mth drone, R n,m represents the transmission throughput from the nth D2D transmitter to the mth UAV.

[0126] Step 4.3: Update the adaptive convergence factor a according to the current iteration number t:

[0127]

[0128] where a max and a mim They represent the maximum and minimum values ​​of the adaptive convergence factor, w represents the adjustment factor, and t max Indicates the maximum number of iterations;

[0129] Update coefficients A and C according to the adaptive convergence factor a:

[0130] C=2r

[0131] A=2ar-a

[0132] Where r∈[0,1] is a random number;

[0133] Step 4.4: Generate a random number p∈[0,1). If p<0.5 and A≤1, use the shrinking and surrounding strategy to update the three-dimensional position of the whale:

[0134] X(t+1)=X * (t)-A|CX * (t)-X(t)|

[0135] Where X(t+1) represents the updated 3D position of the whale, X * (t) represents the 3D position of the optimal whale in the tth iteration, and X(t) represents the 3D position of the whale before the update;

[0136] If p<0.5 and A>1, the prey search strategy based on the Levy flight mechanism is used to update the three-dimensional position of the whale:

[0137]

[0138] Where X rand (t) represents the 3D position of the randomly selected whale in the t-th iteration, β represents the step size control parameter, represents point multiplication, Levy(s) represents the random search path of the prey search strategy based on the Levy flight mechanism and follows the Levy distribution;

[0139] If p ≥ 0.5, the bubble net attack strategy is used to update the whale's position:

[0140] X(t+1)=|X * (t)-X(t)|e bl cos(2πl)+X * (t)

[0141] Where b is a constant representing the shape of the spiral, and l∈[-1,1] is a random number;

[0142] Step 4.5: Calculate the reverse solution for all whales:

[0143] X′ i,j =λ(x min,j +x max,j )-X i,j

[0144] Where X i,j represents the reverse solution of the position of whale i in dimension j, λ∈[0,1] is a random number, X i,j represents the position of whale i in dimension j, x min,j and x max,j They represent the lower and upper bounds of the three-dimensional position of the whale in dimension j respectively.

[0145] The three-dimensional positions of all whales and the reverse solutions of whales are sorted according to the fitness values, and the first S whales are selected as the next generation population.

[0146] Step 4.6: Repeat steps 4.2 to 4.5 until the objective function R sum,m convergence.

[0147] Step 5: Repeat steps 3 and 4 to optimize the two sub-problems alternately until the objective function of the total throughput maximization problem of multi-UAV-assisted IoT converges.

[0148] When the two sub-problems are optimized alternately, the convergence value of the objective function of the two sub-problems is the same, and then the alternate optimization of the two sub-problems is terminated. The convergence value of the objective function of the total throughput maximization problem of multi-UAV assisted Internet of Things under the given three-dimensional deployment of UAVs is the total throughput R of the multi-UAV assisted Internet of Things. sum , and the convergence value of the objective function of the total throughput maximization problem of multi-UAV assisted IoT given by D2D relay selection and time allocation is If the two are equal and no longer change, the overall optimization is considered complete.

[0149] Step 6: Based on the final optimization results, the total throughput of multi-UAV-assisted IoT is maximized. The optimization results include the three-dimensional deployment of UAVs, the relay selection of D2D pairs, the time allocation of D2D transmitters in the NOMA transmission phase, and the time allocation of D2D receivers in the D2D transmission phase.

[0150] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0151] The above-mentioned embodiments only express several implementation modes of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the invention. It should be pointed out that, for a person of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the attached claims.

Claims

1. A method for maximizing the total throughput of a multi-UAV-assisted Internet of Things, characterized in that: The multi-drone-assisted Internet of Things includes 1 multi-antenna base station, U single-antenna cellular edge users, M single-antenna drones, N pairs of D2D pairs, each pair of D2D pairs includes 1 D2D transmitter and 1 D2D receiver, wherein the drone acts as a relay device, relaying data from the base station to the cellular edge user, and at the same time relaying data from the D2D transmitter to the D2D receiver. The total throughput maximization method of the multi-drone-assisted Internet of Things includes: Step 1: The base station obtains the location information of U cell edge users and N pairs of D2D pairs, divides the cell edge users into M clusters according to the location information, and designates one drone for each cluster to relay the data sent by the base station to the cell edge users of the cluster; Step 2: Construct the total throughput maximization problem of multi-UAV-assisted IoT and decompose it into two sub-problems: the total throughput maximization problem of multi-UAV-assisted IoT under given three-dimensional deployment of UAVs and the total throughput maximization problem of multi-UAV-assisted IoT under given D2D relay selection and time allocation; Step 3: To solve the problem of maximizing the total throughput of multi-UAV-assisted IoT under a given three-dimensional deployment of UAVs, the relay selection of D2D pairs, the time allocation of D2D transmitters in the NOMA transmission phase, and the time allocation of D2D receivers in the D2D transmission phase are optimized according to the current three-dimensional deployment of UAVs. Step 4: Solve the problem of maximizing the total throughput of multi-UAV-assisted IoT given by D2D pair relay selection and time allocation, and optimize the three-dimensional deployment of UAVs according to the current D2D pair relay selection and time allocation; Step 5: Repeat steps 3 and 4 to optimize the two sub-problems alternately until the objective function of the total throughput maximization problem of multi-UAV-assisted IoT converges; Step 6: Maximize the total throughput of multi-UAV-assisted IoT based on the final optimization results, which include the three-dimensional deployment of UAVs, the relay selection of D2D pairs, the time allocation of D2D transmitters in the NOMA transmission phase, and the time allocation of D2D receivers in the D2D transmission phase.

2. The method for maximizing the total throughput of multi-UAV-assisted Internet of Things according to claim 1, characterized in that: The total throughput maximization problem of constructing a multi-UAV assisted Internet of Things includes: The throughput R of the transmission from the base station to the mth drone B，m yes: where m∈{1, 2, ...M}, τ m，n represents the time for the nth D2D transmitter to transmit data to the mth UAV in the NOMA transmission phase, that is, the time allocation of the D2D transmitter in the NOMA transmission phase, W is the wireless channel bandwidth, P B is the base station’s transmit power, h B,m is the channel gain between the base station and the mth UAV, P N is the transmission power of the D2D transmitter, h n，m is the channel gain between the nth D2D transmitter and the mth UAV, σ 2 is the additive white Gaussian noise power; The throughput R of the transmission from the nth D2D transmitter to the mth UAV n，m yes: The throughput R of the transmission from the mth UAV to the kth cell edge user of the corresponding cluster m,k yes: where k∈{1, 2, ...K m }, K m is the total number of cell edge users in the mth cluster, t m，k represents the transmission time of the mth UAV to the kth cell edge user of the corresponding cluster, P U represents the transmitting power of the UAV, h m,k represents the channel gain between the mth UAV and the kth cell edge user of the corresponding cluster; The transmission throughput R from the mth UAV to the nth D2D receiver m，n+N yes: Where n+N represents the nth D2D receiver, represents the time when the mth UAV transmits data to the nth D2D receiver in the D2D transmission phase, that is, the time allocation of the D2D receiver in the D2D transmission phase, h m,n+N represents the channel gain between the mth UAV and the nth D2D receiver; The throughput R of the transmission from the base station to the kth cell edge user of the corresponding cluster through the relay of the mth drone is B，m，k yes: The throughput R of the transmission from the nth D2D transmitter to the nth D2D receiver via the relay of the mth UAV is n，m，n+N yes: R n，m，n+N =min{R n，m ,R m,n+N }=R n，m The total throughput of multi-UAV-assisted IoT R sum for: Then the objective function of the total throughput maximization problem of multi-UAV-assisted IoT is maxR sum .

3. The method for maximizing the total throughput of multi-UAV-assisted Internet of Things according to claim 1, characterized in that: The method of optimizing the relay selection of D2D pairs, the time allocation of D2D transmitters in the NOMA transmission phase, and the time allocation of D2D receivers in the D2D transmission phase according to the current three-dimensional deployment of UAVs includes: Step 3.1: Based on the location information of the D2D pairs and the drones, calculate the channel gains between all D2D pairs and all drones. If the channel gain between the nth D2D transmitter or the nth D2D receiver and the mth drone does not meet the minimum channel gain constraint, the nth D2D pair cannot select the mth drone as a relay device. Update the channel gain h between the nth D2D transmitter and the mth drone. n，m =0; Step 3.2: Set a probability vector P for each D2D transmitter n =(p n，1 , p n，2 , ..., p n，M ), where p n，m represents the probability that the nth D2D transmitter selects the mth drone as a relay, initialized Step 3.3: Each D2D transmitter receives a signal according to the probability vector P n Select the drone as the relay device and obtain the relay selection information; Step 3.4: Based on the relay selection information obtained in step 3.3, use the greedy time allocation method to optimize the time allocation of D2D transmitters in the NOMA transmission phase: In the NOMA transmission phase, allocate time that satisfies the minimum throughput constraint to all D2D transmitters. For the mth drone, if If it is less than the NOMA transmission phase time T1, the remaining time in the NOMA transmission phase time T1 is allocated to the D2D transmitter with the largest channel gain with the mth UAV; otherwise, the time allocation τ of all D2D transmitters of the mth UAV is updated. m，n =0, where τ m，n It represents the time when the nth D2D transmitter transmits data to the mth UAV in the NOMA transmission phase; Step 3.5: In the D2D transmission phase, according to R n，m =R n，m，n+N , calculate the time allocation of the D2D receiver, if If it is greater than the D2D transmission phase time θT1, the time allocation τ of all D2D transmitters of the mth UAV is updated m，n =0; if If the D2D transmission phase time θT1 is less than or equal to the D2D transmission phase time θT1, the time allocation τ of all D2D transmitters of the mth UAV is selected. m，n remains unchanged; among them, R n，m represents the transmission throughput from the nth D2D transmitter to the mth UAV, R n，m，n+N represents the throughput of transmission from the nth D2D transmitter to the nth D2D receiver through the relay of the mth UAV, It indicates the time when the mth UAV transmits data to the nth D2D receiver in the D2D transmission phase; Step 3.6: Calculate the total throughput R of multi-UAV-assisted IoT under the current relay selection and time allocation scheme sum (t), and according to the current total throughput R sum (t) Update the probability vector of all D2D transmitters, t is the current iteration number: use R max Represents the maximum throughput, the initial value is 0, if the current total throughput R sum (t) is greater than the maximum throughput R max , update R max =R sum (t), and the probability vector at the t+1th iteration is updated as: where p n,m (t+1) represents the probability that the nth D2D transmitter selects the mth UAV as a relay after update, P n,m (t) represents the probability that the nth D2D transmitter selects the mth UAV as a relay before updating, and a n (t) represents the UAV selected by the nth D2D transmitter in the tth iteration, ω1 represents the first learning step, represents the normalized reward; Otherwise, the probability vector at the t+1th iteration is updated as: Where ω2 represents the second learning step, m n,max represents the optimal relay selection of the nth D2D transmitter; Step 3.7: Repeat steps 3.3 to 3.6 until the objective function of the total throughput maximization problem of multi-UAV-assisted IoT given the three-dimensional deployment of UAVs converges.

4. The method for maximizing the total throughput of multi-UAV-assisted Internet of Things according to claim 1, characterized in that: The method of optimizing the three-dimensional deployment of drones based on the current D2D relay selection and time allocation includes: The sub-problem of D2D relay selection and time allocation for the total throughput maximization of a given multi-UAV-assisted IoT is equivalently expressed as a three-dimensional deployment optimization problem for each UAV. The following optimization solution steps are performed for each UAV: Step 4.1: According to the population size S and the chaotic iteration step, the whales are initialized based on Bernoulli shift chaos. The three-dimensional position of each whale represents the three-dimensional deployment position of the drone. Step 4.2: Calculate the total throughput of the relay transmission through the drone as the fitness value of each whale, take the whale with the highest fitness value as the optimal whale, and record the three-dimensional position of the optimal whale. The fitness value calculation formula is as follows: Where R sum,m represents the total throughput of relay transmission through the mth UAV, K m is the total number of cell edge users in the mth cluster, R B，m，k represents the throughput of transmission from the base station to the kth cell edge user of the corresponding cluster through the relay of the mth drone, R n，m represents the throughput of the transmission from the nth D2D transmitter to the mth UAV; Step 4.3: Update the adaptive convergence factor a according to the current iteration number t: where a max and a mim They represent the maximum and minimum values ​​of the adaptive convergence factor, w represents the adjustment factor, and t max Indicates the maximum number of iterations; Update coefficients A and C according to the adaptive convergence factor a: C＝2r A=2ar-a Where r∈[0,1] is a random number; Step 4.4: Generate a random number p∈[0,1). If p<0.5 and A≤1, use the shrinking and encircling strategy to update the three-dimensional position of the whale: X(t+1)=X*(t)-A|CX*(t)-X(t)| Where X(t+1) represents the updated 3D position of the whale, X*(t) represents the 3D position of the optimal whale in the tth iteration, and X(t) represents the 3D position of the whale before the update; If p < 0.5 and A > 1, the prey search strategy based on the Levy flight mechanism is used to update the three-dimensional position of the whale: Where X rand (t) represents the 3D position of the randomly selected whale in the t-th iteration, β represents the step size control parameter, represents point multiplication, Levy(s) represents the random search path of the prey search strategy based on the Levy flight mechanism and follows the Levy distribution; If p ≥ 0.5, the bubble net attack strategy is used to update the whale's position: X(t+1)=|X * (t)-X(t)|e bl cos(2πl)+X * (t) Where b is a constant representing the shape of the spiral, and l∈[-1,1] is a random number; Step 4.5: Calculate the reverse solution for all whales: X′ i，j =λ(x min，j +x max，j )-X i，j Where X i，j represents the reverse solution of the position of whale i in dimension j, λ∈[0,1] is a random number, X i，j represents the position of whale i in dimension j, x min，j and x max,j They represent the lower and upper bounds of the three-dimensional position of the whale in dimension j respectively; Sort the three-dimensional positions of all whales and their reverse solutions according to their fitness values, and select the first S whales as the next generation population; Step 4.6: Repeat steps 4.2 to 4.5 until the total throughput R of the relay transmission through the UAV is sum，m convergence.

5. The method for maximizing the total throughput of multi-UAV-assisted Internet of Things according to claim 1, characterized in that: The step 5 alternately optimizes the two sub-problems until the objective function of the total throughput maximization problem of multi-UAV-assisted IoT converges, including: The two sub-problems are optimized alternately until the convergence values ​​of the objective functions of the two sub-problems are consistent, then the alternating optimization of the two sub-problems is terminated.

Images