Secret communication method combining unmanned aerial vehicle deployment, power distribution and coalition game

Abstract

The invention discloses a secure communication method combining unmanned aerial vehicle deployment, power distribution and coalition game, and belongs to the technical field of communication. The implementation method comprises the following steps: setting the number of subcarrier channels of the unmanned aerial vehicle as the number of alliances which can be formed, forming non-overlapping alliances by ground nodes, averagely distributing communication time slots to the nodes of the same alliance, and communicating with the unmanned aerial vehicle in a TDMA (Time Division Multiple Access) mode. A communication node in the multi-node alliance selects a node other than the communication node in the alliance as a cooperative interference person and sends the cooperative interference person to a ground eavesdropper, so that the alliance confidentiality and rate during communication are improved. And under the constraint of the cruising ability of the unmanned aerial vehicle, a double-layer optimization problem for improving the overall confidentiality and rate is constructed. In an upper layer optimization problem, a stable alliance formation is obtained based on a TCFG algorithm, and a non-convex optimization problem of a bottom layer is solved under the stable alliance formation, so that an unmanned aerial vehicle deployment result and a ground node power distribution result which enable the overall confidentiality and rate to be maximized are obtained, and the overall confidentiality and rate to be maximized are realized.

Description

Technical Field

[0001] The invention relates to a confidential communication method for joint unmanned aerial vehicle deployment, power allocation and alliance game, and belongs to the technical field of communication. Background Art

[0002] The International Telecommunication Union proposed six typical scenarios for the sixth generation of communications (6G), among which drones play an important role in achieving global stereoscopic coverage. Compared with ground networks, drone communication networks use the flexibility of drones to serve as relays or receivers for ground IoT devices at altitudes that allow line-of-sight (LoS) communications, thereby reducing fading and distortion caused by shadows and signal blocking. In addition, in hotspots with poor infrastructure such as forests, deserts, and oceans, traditional ground base stations may not be able to provide adequate services. With the development of wireless transmission technology, secure transmission has become an urgent problem to be solved in 6G wireless communications. There are a large number of intentional or unintentional eavesdroppers in drone communication networks, posing a major threat to the security of data transmission. Physical layer security (PLS) is a technology that uses the characteristics of the physical layer, such as the uniqueness and reciprocity of physical channel information, to achieve information encryption. As an alternative to cryptographic-based communication security technology, PLS does not require the design and distribution of keys, avoids complex algorithms, and is more suitable for future large-scale, decentralized 6G wireless communications. In order to further improve the PLS of communication systems, artificial noise (AN) has been widely studied as a collaborative interference technology. In the literature [1] (see: A. Li, Q. Wu, and R. Zhang, “UAV-enabled cooperative jamming for improving secrecy of ground wiretap channel,” IEEE Wireless Commun. Lett., vol. 8, no. 1, pp. 181–184, Feb. 2019), UAVs act as cooperative jammers to send AN to counter eavesdropping attacks. In the literature [2] (see: C. Han, L. Bai, T. Bai, and J. Choi, “Joint UAV deployment and power allocation for secure space-air-ground communications,” IEEE Trans. Commun., vol. 70, no. 10, pp. 6804–6818, Oct. 2022), UAVs act as both relays and cooperative jammers, sending AN to interfere with eavesdroppers while sending and receiving confidential data, and optimizing power to improve communication security.In addition, the literature [3] (see: H. Lei, H. Yang, K.-H. Park, IS Ansari, J. Jiang, and M.-S. Alouini, "Joint trajectory design and user scheduling for secure aerial underlay IoT systems," IEEE Internet Things J., vol. 10, no. 15, pp. 13 637–13 648, Aug. 2023.) developed a dual-UAV scenario in which one UAV transmits confidential data and the other UAV sends an AN to interfere with the eavesdropper. However, one UAV only acts as a cooperative jammer in a communication system, which wastes UAV resources. In addition, when a UAV acts as a relay and a cooperative jammer at the same time, it will cause additional energy consumption, such as UAV flight power consumption, antenna power consumption, and system electrostatic power consumption, and increase system complexity and the difficulty of executing tasks. Therefore, it is particularly important to find a mutually beneficial cooperative jamming strategy in UAV communication networks.

[0003] In a UAV-assisted communication network, the endurance problem of UAVs can be solved by providing laser charging services to UAVs via a low altitude platform (LAP). Special photovoltaic panels are installed on the UAVs, which receive lasers and convert them into electrical energy. Reference [4] (see: A. Ranjha and G. Kaddoum, “URLLC-enabled by laser powered UAV relay: A quasi-optimal design of resource allocation, trajectory planning and energy harvesting,” IEEE Trans. Veh. Technol., vol. 71, no. 1, pp. 753–765, Jan. 2022) considers using a laser transmitter to charge the relay UAV in the air, supporting the UAV to complete the flight using the collected laser energy. In addition, reference [4] gives an expression for the received signal strength of the UAV for energy collection on the laser link, and concludes that laser energy collection is less affected by the transceiver distance than RF energy collection. Summary of the invention

[0004] The technical problem to be solved by the confidential communication method of joint UAV deployment, power allocation and alliance game disclosed in the present invention is: the number of subcarrier channels available for UAV communication is set to the number of alliances, each ground node is assigned to a alliance, and each alliance occupies a subcarrier channel for uplink communication; the ground nodes in the same alliance communicate with the UAV in a time division multiple access TDMA mode and evenly allocate communication time slots, and the nodes in the multi-node alliance have dual identities, transmitting confidential information as senders in communication time slots, and being selected as or not as collaborative interferers to transmit artificial noise AN in non-communication time slots; within the communication time period, under the premise of ensuring the endurance of the UAV, the deployment of the UAV, the power allocation of the ground nodes and the alliance formation (Coalition Formulation) are jointly optimized to resist the eavesdropping of ground RF eavesdroppers, and maximize the sum of the confidentiality rates of all ground nodes, that is, the overall confidentiality and rate.

[0005] The objective of the present invention is achieved through the following technical solutions:

[0006] The present invention discloses a confidential communication method for joint UAV deployment, power allocation and alliance game, establishes a three-dimensional Cartesian coordinate system, and the ordinates of N ground nodes and a full-band RF eavesdropper are set to 0. A UAV with M subcarrier channels hovers over N (N>M) scattered ground nodes. While the ground node sends confidential information through the subcarrier channel, the eavesdropper launches an eavesdropping attack. In the T time period, the UAV endurance problem is solved by LAP laser charging. Special photovoltaic panels are installed on the UAV to receive laser and convert it into electrical energy. The position where the UAV hovers is deployed under the premise of ensuring endurance, so that it is as close to the ground node as possible to reduce the power loss caused by free space transmission. Since N>M, it is impossible for each node to completely occupy a UAV subcarrier channel in the T time period. The number of subcarrier channels is set to the number of alliances that can be formed, that is, M, and the ground nodes form M non-overlapping alliances, in which the nodes belonging to the same alliance communicate with the UAV in a time division multiple access TDMA communication mode and evenly distribute the communication time slots. A coordinated jamming strategy for uplink secure communication of nodes in a multi-node alliance is constructed. In a multi-node alliance, a node in a communication time slot selects a node in the same alliance as a coordinated jammer to send an AN to the eavesdropper, thereby improving the sum of the confidentiality rates of the ground nodes in the alliance within the T time period, namely, the alliance confidentiality sum rate. The channel model between the ground node and the UAV and the channel model between the ground node and the eavesdropper are established respectively, and the confidentiality rate between the UAV and the ground node is analyzed. Under the constraint of the UAV's endurance, a two-level optimization problem is constructed to improve the confidentiality sum rate of all ground nodes, namely, the overall confidentiality sum rate. In the upper optimization problem, all ground nodes form different alliances, and each node has the opportunity to change the alliance formation through "switching" or "exchanging" actions. In the bottom optimization problem, the UAV also acts as a receiver for the ground nodes in M ​​alliances, the position where the UAV hovers is deployed, and the transmission power and coordinated jamming power of the ground nodes in each alliance are redistributed. The block coordinate descent (BCD) method is used to solve the underlying optimization problem, and the underlying problem is decomposed into two non-convex sub-problems. Sub-problem 1: Optimize the UAV deployment under the given ground node power allocation; Sub-problem 2: Optimize the power allocation of ground nodes under the given UAV deployment. By introducing auxiliary variables and using the successive convex approximation (SCA) method, the two sub-problems are transformed into two convex function difference DC planning problems and a second-order cone SOC planning problem. The UAV deployment results are obtained by solving the two DC planning problems, and the power allocation results of the ground nodes are obtained by solving the SOC planning problem. Sub-problems 1 and 2 are iteratively solved so that the underlying problem converges in a finite number of iterations, and the UAV deployment results and ground node power allocation results that maximize the overall confidentiality and efficiency are obtained.In the upper-level optimization problem, based on the two-step coalition formation game TCFG algorithm, the coalition game method is used to change the coalition formation through the "switching" or "exchange" actions of the nodes until a stable coalition is formed. The underlying non-convex optimization problem is solved under the stable coalition formation to obtain the UAV deployment results and the power allocation results of the ground nodes that maximize the overall confidentiality and efficiency. The overall confidentiality and efficiency are maximized according to the UAV deployment results and the power allocation results of the ground nodes.

[0007] The confidential communication method for joint UAV deployment, power allocation and alliance game disclosed in the present invention comprises the following steps:

[0008] Step 1: Establish a three-dimensional Cartesian coordinate system, construct the received signal strength expression for the drone to collect energy in the laser link, and construct the conditions that need to be met to ensure the drone's endurance. Establish the channel model between the ground node and the drone and the channel model between the ground node and the eavesdropper. In the presence of an RF eavesdropper and non-communication nodes sending ANs to the eavesdropper, calculate the confidentiality rate of communication between the ground nodes and drones in each alliance, and establish the constraint equation for the power of the ground nodes in each alliance.

[0009] Step 1.1: Use E and U represent the set of ground nodes, ground eavesdroppers, and drones, respectively. The location of LAP is fixed and is represented by [x LAP ,y LAP ,z LAP ]; the horizontal coordinate of the nth node is represented by q n =[x n ,y n ], the horizontal coordinate of E is denoted as q E =[x E ,y E ], and their vertical coordinates are all 0. The horizontal and vertical coordinates of the drone are denoted as q U =[x U ,y U ] and z U .

[0010] Step 1.2: Construct the received signal strength expression for the drone to collect energy in the laser link, and construct the endurance conditions that need to be met to ensure the endurance of the drone. Considering that the limited onboard battery is not enough to support the propulsion and communication of the drone within the time period T, LAP provides laser charging services for the drone. Since the energy consumed by communication, baseband signal processing and thermal noise in the RF link is much smaller, only the energy consumed by the drone propulsion is considered. In the hovering state, the power consumed by the rotorcraft is expressed as

[0011] P U,hov =Pb +P i (1)

[0012] Where P b and P i are blade profile power and induced power respectively. LAP has a fixed transmission power P L , the laser signal strength received by the drone is

[0013]

[0014] in represents the distance between the drone and LAP, D is the initial size of the laser beam, χ is the optical efficiency of the joint transmitter and receiver, θ is the angular spread, A is the area of ​​the collection lens, and κ is the attenuation coefficient of the medium. The energy collection module on the drone has a constant laser power collection efficiency ω∈(0,1), and the laser power collected by the drone is expressed as

[0015]

[0016] P h is a about d UL A monotonically decreasing function. Both κ and θ are very small constant values, making laser power collection less affected by the transceiver distance than RF power collection. The conditions that need to be met for the drone to last are:

[0017] P h ≥P U,hov . (4)

[0018] Step 1.3: Establish the channel model between the ground node and the UAV and the channel model between the ground node and the eavesdropper. In the presence of an RF eavesdropper, the cooperative interference node in the non-transmission time slot sends AN to the eavesdropper, and the confidentiality rate of the communication between the ground node and the UAV in each alliance is calculated. Analyze the power distribution of nodes in the alliance. Assume that node For communication nodes, nodes The cooperative interference node selected for node n. The LoS model is used as the ground-to-air communication model, and the channel gain g between node n and the UAV is U,m [n] follows the free space path loss model. The channel gain g between node n and the eavesdropper is E,m [n] obeys the free space path loss model and small-scale Rayleigh fading distribution. U,m [n] and g E,m [n] are respectively represented as

[0019]

[0020] Where β0 is the channel gain between the ground node and the UAV with a reference distance of 1m, and β1 is the channel gain between the ground node and the eavesdropper with a reference distance of 1m. is a random variable with unit mean, describing an exponential distribution of Rayleigh fading. represents the distance between node n and the drone, represents the distance between node n and the eavesdropper. Under LoS propagation conditions, the communication rate between node n and the drone is

[0021]

[0022] where p s,m [n] represents the transmission power of node n, N 0,U represents the noise power on the drone, γ0=β0 / N 0,U The eavesdropping rate between node n and the ground eavesdropper is expressed as

[0023]

[0024] Where N 0,E represents the noise power of the eavesdropper, γ1=β1 / N 0,E Node n′ acts as a cooperative interferer of node n and sends a signal with power p to the eavesdropper. j,m [n], according to the Shannon capacity formula, the eavesdropping rate between node n and the ground eavesdropper is expressed as

[0025]

[0026] in represents the channel gain between node n′ and the eavesdropper.

[0027] When node n′ sends a jamming signal to the eavesdropper, the communication confidentiality rate between node n and the UAV is expressed as

[0028]

[0029] Among them, the operator [·] + To make equation (10) not smooth at zero, remove [·] + It does not affect the solution of the optimal value of formula (10) and can make formula (10) smooth at zero value. Formula (10) is transformed into:

[0030]

[0031] According to formulas (9), (10), and (11), the existence of cooperative interference is conducive to suppressing the eavesdropping rate and improving the confidentiality performance of node n.

[0032] Establish the power constraint equation of the ground nodes in each alliance. Assume that the average power of each node is P ave , the peak power is P peak , P ave ≤P peak ; For the mth alliance set up Contains N m nodes, total available power N m ×T×P ave is reallocated within a time period of T. and sequence Represents Alliance The combination of the transmission power and the coordinated interference power of the midpoint node. The power constraint equation in is expressed as

[0033]

[0034] Step 2: Use Represents the set of all alliances. All ground nodes form M non-overlapping alliances. According to the number of nodes in the alliance, the alliance is divided into three types: empty alliance, single-node alliance and multi-node alliance, and the maximum alliance confidentiality and rate of empty alliance, single-node alliance and multi-node alliance are defined. With the goal of maximizing the overall confidentiality and rate, under the constraints of drone endurance, ground node power, and the maximum number of nodes allowed in each alliance, an optimization problem of maximizing the overall confidentiality and rate is established.

[0035] Step 2.1: Use Represents the set of all alliances. All ground nodes form M non-overlapping alliances. According to the number of nodes in the alliance, the alliance is divided into three types: empty alliance, single-node alliance and multi-node alliance, and the maximum alliance confidentiality and rate of empty alliance, single-node alliance and multi-node alliance are defined.

[0036] There are no nodes in the empty coalition, and its maximum coalition confidentiality and rate is always zero.

[0037] The single node alliance There is only one node n in the network, which completely occupies a subcarrier channel in the T time period, but there is no cooperative interference node. The maximum alliance confidentiality and rate of a single-node alliance is given by g U,m [n] and g E,m [n] decide. U,m [n]>g E,m [n], maximum coalition confidentiality and rate On the contrary, the maximum alliance confidentiality and rate is 0, and the optimal behavior of node n is not to transmit data.

[0038] The multi-node alliance There are two or more nodes in the alliance. The nodes in the same alliance evenly distribute the communication time slots and communicate with the drones in TDMA mode. The node in the communication time slot selects a node in the same alliance as a collaborative jammer to send AN to the eavesdropper. Multi-node alliance The maximum coalition confidentiality and rate are obtained by optimizing the UAV deployment and the power of the ground nodes in step 3.2.

[0039] Step 2.2: With the goal of maximizing the overall confidentiality and rate, under the constraints of drone endurance, ground node power, and the maximum number of nodes allowed by each alliance, establish an optimization problem to maximize the overall confidentiality and rate.

[0040]

[0041] Among them, C max Indicates the maximum number of nodes allowed in a federation, Indicates the flight area of ​​the drone. and Respectively The set of transmit power and cooperative interference power of each node in.

[0042] Step 3: The optimization problem of maximizing overall confidentiality and rate established in step 2 It is equivalent to a two-level optimization problem. In the upper optimization problem, all ground nodes form M non-overlapping alliances, and each node has the opportunity to change the formation of the alliance through "switching" or "exchanging" actions. In the bottom optimization problem, the drone acts as the receiver of M alliances at the same time, the position where the drone hovers is deployed, and the transmission power and cooperative interference power of each node in each alliance are redistributed. The BCD method is used to solve the bottom optimization problem, and the bottom problem is decomposed into two non-convex sub-problems. Sub-problem 1: Optimize the deployment of drones under the given ground node power allocation; Sub-problem 2: Optimize the power allocation of ground nodes under the given drone deployment. By introducing auxiliary variables and using the SCA method, these two sub-problems are transformed into two DC planning problems and one SOC planning problem respectively. The drone deployment results are obtained by solving the two DC planning problems, and the power allocation results of the ground nodes are obtained by solving one SOC planning problem. Sub-problems 1 and 2 are iteratively solved so that the bottom problem converges in a finite number of iterations, and the drone deployment results and ground node power allocation results that maximize all overall confidentiality and rate are obtained. In the upper-level optimization problem, based on the TCFG algorithm, the alliance game method is used to change the alliance formation through the "switching" or "exchange" actions of the nodes until a stable alliance is formed. The underlying non-convex optimization problem is solved under the stable alliance formation to obtain the drone deployment results and the power allocation results of the ground nodes that maximize the overall confidentiality and efficiency. The overall confidentiality and efficiency are maximized according to the drone deployment results and the power allocation results of the ground nodes.

[0043] Step 3.1: In the upper optimization problem, all ground nodes form M non-overlapping coalitions, and each node has the opportunity to change the coalition formation through the "switch" or "exchange" action. Generate the initial coalition formation.

[0044] Step 3.1.1: The coalition formation optimization problem is expressed as

[0045]

[0046] In the formula Represents the maximum alliance confidentiality and rate under the current alliance formation.

[0047] Step 3.1.2: Generate the initial alliance formation and calculate the maximum overall confidentiality and rate under the alliance formation according to step 3.2. Set flag-switch←1 and flag-exchange←1.

[0048] Step 3.2: In the underlying optimization problem, the drone acts as a receiver for M alliances at the same time. The drone hovering position is deployed, and the transmission power and cooperative interference power of each node in each alliance are redistributed. The BCD method is used to solve the underlying optimization problem, and the underlying problem is decomposed into two non-convex sub-problems. Sub-problem 1: Optimize the deployment of drones under the given ground node power allocation; Sub-problem 2: Optimize the power allocation of ground nodes under the given drone deployment. By introducing auxiliary variables and using SCA technology, these two sub-problems are transformed into two DC planning problems and one SOC planning problem respectively. The drone deployment results are obtained by solving the two DC planning problems, and the power allocation results of the ground nodes are obtained by solving one SOC planning problem. Sub-problems 1 and 2 are iteratively solved so that the underlying problem converges in a finite number of iterations, and the drone deployment results and ground node power allocation results that maximize the overall confidentiality and rate are obtained.

[0049] Step 3.2.1: Based on the overall confidentiality and rate maximization problem established in step 2.2, based on the fixed alliance formation, the underlying optimization problem is expressed as

[0050]

[0051] Using the BCD method It is decomposed into two sub-problems: sub-problem 1: optimize the deployment of UAVs under the given power allocation of ground nodes; sub-problem 2: optimize the power allocation of ground nodes under the given UAV deployment. By introducing auxiliary variables and using the SCA method, these two sub-problems are transformed into two DC planning problems and one SOC planning problem respectively. The UAV deployment results are obtained by solving the two DC planning problems, and the power allocation results of the ground nodes are obtained by solving one SOC planning problem.

[0052] Step 3.2.2: Optimize the UAV deployment given the ground node power allocation. The optimization problem is expressed as

[0053]

[0054] where [x m [n],y m [n]] and p s,m [n] represents alliance The horizontal coordinates and transmission power of node n in . Since the objective function in equation (20) is non-convex and the constraint in equation (4) is non-convex, is a non-convex optimization problem. In order to transform (20) into a solvable form, an auxiliary variable is introduced and t relax the objective function in (20) and the constraint in (4) respectively. The objective function in (20) is converted into a concave function using the SCA method. The constraint in (4) is converted into two equivalent constraints and written in two convex function difference forms. Rewrite as

[0055]

[0056] (x U -x LAP ) 2 +(y U -y LAP ) 2 +(z U -z LAP ) 2 ≤t 2 , (twenty three)

[0057] (P b +P i )×(t·θ+D) 2 -P L ωAχe -κt ≤0, (24) (17)

[0059] The first-order Taylor expansion is used to deal with the non-concavity of the objective function in equation (21) and the non-concavity of the constraints in equations (23) and (24). It's about u m [n] and with the help of a first-order Taylor expansion, we get

[0060]

[0061] in is u m [n] is the value at the first iteration. In equations (23) and (24), t 2 and e -κt With the help of a first-order Taylor expansion, it is expanded into

[0062] t 2 ≥-(t (l) ) 2 +2tt (l) , (26)

[0063] and

[0064] e -κt ≥e -κt(l) -κe -κt(l) (tt (l) ). (27)

[0065] question was rewritten as

[0066]

[0067] st (x U -x LAP ) 2 +(y U -y LAP ) 2 +(z U -z LAP ) 2 +(t (l) ) 2 -2tt (l) ≤0, (29)

[0068]

[0069] The objective function in (28) is concave with respect to u, and the constraints in (29) and (30) are concave with respect to {x U ,y U ,z U ,u,t} is convex. Therefore, the convex optimization problem is solved by the interior point method Get the drone deployment results.

[0070] Step 3.2.3: Optimize the power allocation of ground nodes given the UAV deployment. The optimization problem is expressed as

[0071]

[0072] The objective function in formula (31) is a highly complex function containing multiple optimization variables and random variables. Considering the optimization problem In order to improve the robustness of the algorithm, the objective function in (31) is tightened and the negative term in (31) is taken to the maximum value, which is expressed as

[0073]

[0074] The second term of the objective function in formula (32) has random variables in both the numerator and the denominator. The objective function in formula (32) is regarded as a random variable For the convenience of solving, Expressed as

[0075]

[0076] in Jensen inequality is applied to deal with the “-max” in the objective function in equation (32). Specifically, the calculation of equation (32) The upper bound and When node n is the sender and node n′ is the cooperative interferer, find the maximum eavesdropping rate from node n to the eavesdropper. The concavity of

[0077]

[0078] for Its lower bound is based on Jensen inequality and ln(1+e x ) is obtained by

[0079]

[0080] in The parameter is An exponentially distributed random variable with

[0081]

[0082] Where ε is the Euler constant. Substituting equation (36) into equation (35), we get The lower bound of

[0083]

[0084] According to formulas (33)(34)(35)(36)(37), when n′ is When there are cooperative interference nodes in the network, the maximum eavesdropping rate from node n to the eavesdropper is

[0085]

[0086] Therefore, the optimization problem Re-expressed as

[0087]

[0088] The objective function in equation (39) is highly complex. To facilitate observation, the variable Transformed into an equivalent optimization problem, expressed as

[0089]

[0090] in and The non-convex constraints in (44), (45), and (46) can be equivalently written as

[0091]

[0092] The non-convex constraints in (47), (48), and (49) are linearized around the current point in the lth iteration to obtain

[0093] Therefore, the optimization problem Re-expressed as

[0094]

[0095] st(12),(13),(14),(41),(42),(43),(50),(51),(52)

[0096] Optimization Problem It is an optimization problem involving SOC constraints. CVX is used to solve it and obtain the power allocation result of the ground nodes.

[0097] Step 3.2.4: Use the optimization problem from Step 3.2.2 and Step 3.2.3 and Calculate the results of UAV deployment and ground node power allocation respectively, and solve iteratively and Until convergence, the results of UAV deployment and ground node power allocation are finally obtained.

[0098] Step 3.3: In the upper-level optimization problem, based on the TCFG algorithm, the alliance game method is used to change the alliance formation through the "switching" or "exchange" actions of the nodes until a stable alliance is formed, and the underlying non-convex optimization problem is solved under the stable alliance formation to obtain the UAV deployment results and the power allocation results of the ground nodes that maximize the overall confidentiality and efficiency. The overall confidentiality and efficiency are maximized according to the UAV deployment results and the power allocation results of the ground nodes.

[0099] Step 3.3.1: In the upper optimization problem, the framework and definition of the developed TCFG algorithm are as follows. The TCFG algorithm divides all nodes into M non-overlapping coalitions The alliance formation is updated using the two alliance rules of "switch partial order" and "exchange partial order". Each node follows the same alliance update rule and has its own alliance formation priority. After traversing all nodes, the alliance formation with the largest overall confidentiality and rate is the updated alliance formation. Consider node Current alliance formation Change node n from Switch to The utility is expressed as

[0100]

[0101] in is the coalition formed after node n moves. and Based on and The maximum overall confidentiality and rate are obtained. Therefore, u m,m′ [n]Describe the current alliance formation Next node n from Switch to The switching partial order is defined as follows. If there is a node and Alliance The following conditions are met:

[0102]

[0103] Remove node n from the alliance Switch to Consider two nodes and Current alliance formation The utility of the node exchange alliance is

[0104]

[0105] in is the alliance formed after node exchange. Therefore, v m,m′ [n,n′] describes the current alliance form The two nodes of the exchange alliance and The exchange partial order is defined as follows. If there are two nodes m′≠m satisfies the following conditions:

[0106]

[0107] Swap nodes n and n′ with the alliances they currently belong to.

[0108] Step 3.3.2: Traverse all nodes according to the switching partial order and the exchange partial order. In each traversal, select the alliance formation corresponding to the maximum value of the overall confidentiality and rate as the new alliance formation, and then perform the next traversal until the overall confidentiality and rate no longer increase. The specific operation is as follows. Traverse all nodes according to the switching partial order rule. The utility of the current alliance formation is For nodes node Traversing the Select Union and from Switch to Use step 3.2 to calculate the maximum overall confidentiality and rate under the new alliance formation Compare and And store the result u m,m′ [n]. If present Satisfy u m,m′ [n]>0, find the largest u m,m′ [n] The corresponding alliance is formed and the current alliance is updated, setting flag-switch←1; otherwise, setting flag-switch←0. Repeat the above steps until flag-switch←0. Traverse all nodes according to the exchange partial order rule. The utility of the current alliance is For nodes node Traversing Selections And exchange alliances with node n′. Use step 3.2 to calculate the maximum overall confidentiality and rate under the new alliance formation Compare and and store the result v m,m′ [n,n′]. If it exists Satisfy v m,m′ [n,n′]>0, find the largest v m,m′ The alliance corresponding to [n,n′] is formed and the current alliance formation is updated, and flag-exchange←1 is set; otherwise, flag-exchange←0 is set. Repeat the above steps until flag-exchange←0, that is, a stable alliance is formed. Under the stable alliance formation, step 3.2 is used to solve the underlying non-convex optimization problem, and the UAV deployment results and ground node power allocation results that maximize the overall confidentiality and rate are obtained. The overall confidentiality and rate are maximized according to the UAV deployment results and ground node power allocation results.

[0109] Beneficial effects:

[0110] 1. The present invention discloses a confidential communication method for joint UAV deployment, power allocation and alliance game, using Represents the set of all alliances. All ground nodes form M non-overlapping alliances. According to the number of nodes in the alliance, the alliance is divided into three types: empty alliance, single-node alliance and multi-node alliance, and the maximum alliance confidentiality and rate of empty alliance, single-node alliance and multi-node alliance are defined. With the goal of maximizing the overall confidentiality and rate, under the constraints of drone endurance, ground node power, and the maximum number of nodes allowed in each alliance, an optimization problem of maximizing the overall confidentiality and rate is established. The optimization problem of maximizing overall confidentiality and rate It is established as a two-layer optimization problem. The upper optimization problem is solved using the alliance game method, and the lower optimization problem is solved using the convex optimization method to obtain the alliance formation result, drone deployment result and ground node power allocation result that maximize the overall confidentiality and rate. According to the alliance formation result, drone deployment result and ground node power allocation result, the overall confidentiality and rate are maximized.

[0111] 2. The present invention discloses a method for joint UAV deployment, power allocation and alliance game confidential communication. Each alliance occupies a subcarrier channel of the UAV in full time slot. Each node in the alliance communicates with the UAV in TDMA mode and evenly allocates time slots, reducing the complexity of the optimization problem.

[0112] 3. The confidential communication method of the joint UAV deployment, power allocation and alliance game disclosed in the present invention constructs a collaborative interference strategy for the uplink secure communication of nodes in a multi-node alliance. The nodes in the communication time slot of the multi-node alliance select a node in the same alliance as a collaborative interferer to send an AN to the eavesdropper, thereby improving the confidentiality and rate of the alliance, thereby improving the overall confidentiality and rate.

[0113] 4. The present invention discloses a confidential communication method for joint UAV deployment, power allocation and alliance game. The UAV uses laser for aerial charging. Compared with the radio frequency charging solution, this solution is less affected by the transmission distance of the transceiver and is more suitable for wireless communication scenarios with long transmission distances. BRIEF DESCRIPTION OF THE DRAWINGS

[0114] Figure 1 is a schematic diagram of a three-dimensional Cartesian coordinate system and each channel in step 1 of the present invention;

[0115] Figure 2 It is a flow chart of the confidential communication method for joint UAV deployment, power allocation and alliance game disclosed in the present invention;

[0116] Figure 3 is the deployment of ground nodes, eavesdroppers, LAPs, and initial UAVs in this example;

[0117] Figure 4 It is the randomly generated initial alliance formation in this instance;

[0118] Figure 5 is the optimized UAV deployment in this example;

[0119] Figure 6 is the power allocation of the ground nodes obtained based on the initial alliance optimization in this example;

[0120] Figure 7 This is the stable alliance formation obtained by using the TCFG algorithm in this example. DETAILED DESCRIPTION

[0121] In order to better illustrate the purpose and advantages of the present invention, the invention is further described below with reference to the accompanying drawings and examples.

[0122] Embodiment 1:

[0123] In order to verify the feasibility of the method, based on Figure 1 The channel shown in the figure establishes a three-dimensional Cartesian coordinate system and defines the coordinates of the ground nodes, eavesdroppers, drones, and LAPs. Among them, 15 ground nodes are randomly distributed in the area with horizontal coordinates of [-500 500 -500 500]m, the coordinates of the eavesdropper are (500,500,0)m, the initial coordinates of the drone are (0,0,400)m, the deployable area of ​​the drone is [-500 500 100 -500 500 900]m, and the coordinates of the LAP are (0,0,1000)m.

[0124] like Figure 2 As shown, the confidential communication method for joint UAV deployment, power allocation and alliance game disclosed in this example has the following specific implementation steps:

[0125] Step 1: Establish a three-dimensional Cartesian coordinate system, construct the received signal strength expression for the drone to collect energy in the laser link, and construct the conditions that need to be met to ensure the drone's endurance. Establish the channel model between the ground node and the drone and the channel model between the ground node and the eavesdropper. In the presence of an RF eavesdropper and non-communication nodes sending ANs to the eavesdropper, calculate the confidentiality rate of communication between the ground nodes and drones in each alliance, and establish the constraint equation for the power of the ground nodes in each alliance.

[0126] In the hovering state, the power consumed by the rotorcraft is expressed as

[0127] P U,hov =P b +P i (58)

[0128] LAP has a fixed transmit power P L , the laser power collected by the UAV is expressed as

[0129]

[0130] The conditions that need to be met for drone endurance are:

[0131] P h ≥P U,hov . (60)

[0132] The channel gain g between node n and the UAV U,m [n] follows the free space path loss model, expressed as

[0133]

[0134] The channel gain g between node n and the eavesdropper E,m [n] obeys the free space path loss model and small-scale Rayleigh fading distribution, expressed as

[0135]

[0136] Under LoS propagation conditions, the data rate of communication between node n and the UAV is

[0137]

[0138] When node n′ acts as a cooperative interferer of node n, it sends power p to the eavesdropper. j,m [n] AN, the eavesdropping rate between node n and the eavesdropper is expressed as

[0139]

[0140] The confidentiality rate of communication between node n and the UAV is expressed as

[0141]

[0142] For the Alliance The power constraint of the ground node is expressed as

[0143]

[0144] Step 2: With the goal of maximizing the overall confidentiality and rate, establish an optimization problem to maximize the overall confidentiality and rate under the constraints of UAV endurance, ground node power, and the maximum number of nodes allowed by each alliance.

[0145]

[0146] Step 3: Optimize the problem It is considered as a two-layer optimization problem. In the upper optimization problem, all ground nodes form M non-overlapping alliances, and each node has the opportunity to change the alliance formation through "switching" or "exchanging" actions. In the bottom optimization problem, the drone acts as the receiver of M alliances at the same time, the drone hovering position is deployed, and the transmission power and cooperative interference power of each node in each alliance are redistributed. The BCD method is used to solve the bottom optimization problem, and the bottom problem is decomposed into two non-convex sub-problems. Sub-problem 1: Optimize the deployment of drones under the given ground node power allocation; Sub-problem 2: Optimize the power allocation of ground nodes under the given drone deployment. By introducing auxiliary variables and using SCA technology, these two sub-problems are transformed into two DC planning problems and one SOC planning problem respectively. The drone deployment results are obtained by solving the two DC planning problems, and the power allocation results of the ground nodes are obtained by solving one SOC planning problem. The iterative solution of sub-problems 1 and 2 makes the bottom problem converge in a finite number of iterations, and the drone deployment results and ground node power allocation results that maximize the overall confidentiality and efficiency are obtained. In the upper-level optimization problem, based on the TCFG algorithm, the alliance game method is used to change the alliance formation through the "switching" or "exchange" actions of the nodes until a stable alliance is formed. The underlying non-convex optimization problem is solved under the stable alliance formation to obtain the drone deployment results and the power allocation results of the ground nodes that maximize the overall confidentiality and efficiency. The overall confidentiality and efficiency are maximized according to the drone deployment results and the power allocation results of the ground nodes.

[0147] Step 3.1: In the upper optimization problem, all ground nodes form M non-overlapping coalitions, and each node has the opportunity to change the coalition formation through "switch" or "exchange" actions. Generate the initial coalition formation, set flag-switch←1 and flag-exchange←1. The coalition formation optimization problem is expressed as

[0148]

[0149] Step 3.2: In the underlying optimization problem, the drone acts as a receiver for M alliances at the same time. The drone hovering position is deployed, and the transmission power and cooperative interference power of each node in each alliance are redistributed. The BCD method is used to solve the underlying optimization problem, and the underlying problem is decomposed into two non-convex sub-problems. Sub-problem 1: Optimize the deployment of drones under the given ground node power allocation; Sub-problem 2: Optimize the power allocation of ground nodes under the given drone deployment. By introducing auxiliary variables and using SCA technology, these two sub-problems are transformed into two DC planning problems and one SOC planning problem respectively. The drone deployment results are obtained by solving the two DC planning problems, and the power allocation results of the ground nodes are obtained by solving one SOC planning problem. Sub-problems 1 and 2 are iteratively solved so that the underlying problem converges in a finite number of iterations, and the drone deployment results and ground node power allocation results that maximize the overall confidentiality and rate are obtained.

[0150] In the case of a fixed alliance, consider dealing with The underlying non-convex optimization problem is expressed as

[0151]

[0152] Sub-problem 1: Optimize UAV deployment given the power allocation of ground nodes. The optimization problem is expressed as

[0153]

[0154] (x U -x LAP ) 2 +(y U -y LAP ) 2 +(z U -z LAP ) 2 +(t (l) ) 2 -2tt (l) ≤0, (76)

[0155]

[0156] Solving convex optimization problems via interior point methods Get the drone deployment results.

[0157] Sub-problem 2: Optimize the power allocation of ground nodes given the deployment of UAVs to ensure the alliance For example, the optimization problem is expressed as

[0158]

[0159] Solving optimization problems using CVX Get the power distribution result of ground nodes.

[0160] Step 3.3: In the upper-level optimization problem, based on the TCFG algorithm, the alliance game method is used to change the alliance formation through the "switching" or "exchange" actions of the nodes until a stable alliance is formed, and the underlying non-convex optimization problem is solved under the stable alliance formation to obtain the UAV deployment results and the power allocation results of the ground nodes that maximize the overall confidentiality and efficiency. The overall confidentiality and efficiency are maximized according to the UAV deployment results and the power allocation results of the ground nodes.

[0161] All nodes are traversed according to the switching partial order and the exchange partial order. In each traversal, the alliance formation corresponding to the maximum value of the overall confidentiality and rate is selected as the new alliance formation, and then the next traversal is performed until the overall confidentiality and rate are no longer increased. The specific operation is as follows. All nodes are traversed according to the switching partial order rule. The utility of the current alliance formation is For nodes node Traversing the Select Union and from Switch to Use step 3.2 to calculate the maximum overall confidentiality and rate under the new alliance formation Compare and And store the result u m,m′ [n]. If present Satisfy u m,m′ [n]>0, find the largest u m,m′ [n] The corresponding alliance is formed and the current alliance is updated, setting flag-switch←1; otherwise, setting flag-switch←0. Repeat the above steps until flag-switch←0. Traverse all nodes according to the exchange partial order rule. The utility of the current alliance is For nodes node Traversing Selections And exchange alliances with node n′. Use step 3.2 to calculate the maximum overall confidentiality and rate under the new alliance formation Compare and and store the result v m,m′ [n,n′]. If it exists Satisfy v m,m′ [n,n′]>0, find the largest v m,m′The alliance corresponding to [n,n′] is formed and the current alliance formation is updated, and flag-exchange←1 is set; otherwise, flag-exchange←0 is set. Repeat the above steps until flag-exchange←0, that is, a stable alliance is formed. Under the stable alliance formation, step 3.2 is used to solve the underlying non-convex optimization problem, and the UAV deployment results and ground node power allocation results that maximize the overall confidentiality and rate are obtained. The overall confidentiality and rate are maximized according to the UAV deployment results and ground node power allocation results.

[0162] The specific description above further illustrates in detail the purpose, technical solutions and beneficial effects of the invention. It should be understood that the above is only a specific embodiment of the present invention, which is used to explain the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A confidential communication method for joint UAV deployment, power allocation and alliance game, characterized by: The following steps are included: Step 1: Establish a three-dimensional Cartesian coordinate system, construct the received signal strength expression for the drone to collect energy in the laser link, and construct the conditions that need to be met to ensure the drone's endurance; establish the channel model between the ground node and the drone and the channel model between N ground nodes and the eavesdropper; in the presence of RF eavesdroppers and non-communication nodes sending artificial noise AN to the eavesdropper, calculate the confidentiality rate of communication between the ground nodes and the drones in each alliance, and establish the constraint equation of the ground node power in each alliance; Step 2: Use Represents the set of all alliances; all ground nodes form M non-overlapping alliances; according to the number of nodes in the alliance, the alliance is divided into three types: empty alliance, single-node alliance and multi-node alliance, and the maximum alliance confidentiality and rate of empty alliance, single-node alliance and multi-node alliance are defined; with the goal of maximizing the overall confidentiality and rate, under the constraints of drone endurance, ground node power, and the maximum number of nodes allowed in each alliance, an optimization problem of maximizing the overall confidentiality and rate is established Step 3: The optimization problem of maximizing overall confidentiality and rate established in step 2 It is equivalent to a two-level optimization problem. In the upper-level optimization problem, all ground nodes form M non-overlapping alliances, and each node has the opportunity to change the formation of the alliance through "switching" or "exchanging" actions. In the underlying optimization problem, the drone acts as the receiver of M alliances at the same time, the position where the drone hovers is deployed, and the transmission power and cooperative interference power of each node in each alliance are redistributed. The block coordinate descent BCD method is used to solve the underlying optimization problem, and the underlying problem is decomposed into two non-convex sub-problems. Sub-problem 1: Optimize the deployment of drones under the given ground node power allocation; Sub-problem 2: Optimize the power allocation of ground nodes under the given drone deployment. Auxiliary variables are introduced and the successive convex approximation SCA method is used to transform the two sub-problems into two convex function difference DC planning problems and a second-order cone SOC planning problem. The drone deployment results are obtained by solving the two DC planning problems, and the power allocation results of the ground nodes are obtained by solving a SOC planning problem. Iteratively solve subproblems 1 and 2 so that the underlying problem converges in a finite number of iterations, and obtain the UAV deployment results and ground node power allocation results that maximize all overall confidentiality and efficiency; In the upper-level optimization problem, based on the two-step coalition formation game TCFG algorithm, the coalition game method is used to change the coalition formation through the "switching" or "exchange" actions of the nodes until a stable coalition formation is obtained. The underlying non-convex optimization problem is solved under the stable coalition formation to obtain the UAV deployment results and the power allocation results of the ground nodes that maximize the overall confidentiality and efficiency. The overall confidentiality and efficiency are maximized according to the UAV deployment results and the power allocation results of the ground nodes.

2. The method for secure communication of joint UAV deployment, power allocation and alliance game as claimed in claim 1, characterized in that: The implementation method of step one is: Step 1.1: Use E and U represent the set of ground nodes, ground eavesdroppers, and drones, respectively; the location of LAP is fixed, represented by [x LAP ,y LAP ,z LAP ]; the horizontal coordinate of the nth node is represented by q n =[x n ,y n ], the horizontal coordinate of E is denoted as q E =[x E ,y E ], and their vertical coordinates are all 0; the horizontal and vertical coordinates of the drone are represented by q U =[x U ,y U ] and z U ; Step 1.2: Construct the received signal strength expression of the drone for energy collection in the laser link, and construct the endurance conditions that need to be met to ensure the endurance of the drone; in the hovering state, the power consumed by the rotorcraft is expressed as P U,hov =P b +P i (1) Where P b and P i are blade profile power and induced power respectively; LAP has a fixed transmission power P L , the laser signal strength received by the drone is in represents the distance between the UAV and LAP, D is the initial size of the laser beam, χ is the optical efficiency of the joint transmitter-receiver, θ is the angular expansion, A is the area of ​​the collection lens, and κ is the attenuation coefficient of the medium; the energy collection module on the UAV has a constant laser power collection efficiency ω∈(0,1), and the laser power collected by the UAV is expressed as P h is a about d UL A monotonically decreasing function; κ and θ are both very small constant values, so that laser power collection is less affected by the transceiver distance than RF power collection; The conditions that the drone needs to meet for endurance are: P h ≥P U,hov . (4) Step 1.3: Establish the channel model between the ground node and the UAV and the channel model between the ground node and the eavesdropper; in the presence of an RF eavesdropper, the cooperative interference node in the non-transmission time slot sends AN to the eavesdropper, and calculates the confidentiality rate of the communication between the ground node and the UAV in each alliance; for a multi-node alliance Analyze the power distribution of nodes in the alliance; assume that the node For communication nodes, nodes n′≠n is the cooperative interference node selected by node n; the line-of-sight LoS model is used as the ground-to-air communication model, and the channel gain g between node n and the UAV U,m [n] follows the free space path loss model; the channel gain g between node n and the eavesdropper E,m [n] obeys the free space path loss model and small-scale Rayleigh fading distribution; g U,m [n] and g E,m [n] are respectively represented as Where β0 is the channel gain between the ground node and the UAV with a reference distance of 1m, and β1 is the channel gain between the ground node and the eavesdropper with a reference distance of 1m; is a random variable with unit mean, describing the exponential distribution of Rayleigh fading; represents the distance between node n and the drone, represents the distance between node n and the eavesdropper; under LoS propagation conditions, the communication rate between node n and the drone is where p s,m [n] represents the transmission power of node n, N 0,U represents the noise power on the drone, γ0=β0 / N 0,U ; The eavesdropping rate between node n and the ground eavesdropper is expressed as Where N 0,E represents the noise power of the eavesdropper, γ1=β1 / N 0,E Node n′ acts as a cooperative interferer of node n and sends a signal with power p to the eavesdropper. j,m [n], according to the Shannon capacity formula, the eavesdropping rate between node n and the ground eavesdropper is expressed as in represents the channel gain between node n′ and the eavesdropper; When node n′ sends a jamming signal to the eavesdropper, the communication confidentiality rate between node n and the UAV is expressed as Among them, the operator [·] + To make equation (10) not smooth at zero, remove [·] + It does not affect the solution of the optimal value of formula (10) and can make formula (10) smooth at zero value. Formula (10) is transformed into: According to formulas (9)(10)(11), the existence of cooperative interference is conducive to suppressing the eavesdropping rate and improving the confidentiality performance of node n; Establish the power constraint equation of the ground nodes in each alliance; let the average power of each node be P ave , the peak power is P peak , P ave ≤P peak ; For the mth alliance set up Contains N m nodes, total available power N m ×T×P ave is reallocated within a time period of T; the sequence and sequence Represents Alliance The set of transmit power and cooperative interference power of the mid-node; The power constraint equation in is expressed as 3. The method for secure communication of joint UAV deployment, power allocation and alliance game as claimed in claim 2, characterized in that: The implementation method of step 2 is: Step 2.1: Use Represents the set of all alliances; all ground nodes form M non-overlapping alliances; according to the number of nodes in the alliance, the alliance is divided into three types: empty alliance, single-node alliance and multi-node alliance, and the maximum alliance confidentiality and rate of empty alliance, single-node alliance and multi-node alliance are defined; There is no node in the empty alliance, and its maximum alliance confidentiality and rate is always zero; The single node alliance There is only one node n in the alliance, which completely occupies a subcarrier channel in the T time period, but there is no cooperative interference node; the maximum alliance confidentiality and rate of a single-node alliance is given by g U,m [n] and g E,m [n] decide; when U,m [n]>g E,m [n], maximum coalition confidentiality and rate On the contrary, the maximum alliance confidentiality and rate is 0, and the optimal behavior of node n is not to transmit data; The multi-node alliance There are two or more nodes in the alliance; nodes in the same alliance evenly distribute communication time slots and communicate with drones in time division multiple access (TDMA); nodes in the communication time slot select a node in the alliance as a collaborative jammer to send AN to the eavesdropper; multi-node alliance The maximum alliance confidentiality and rate of are obtained by optimizing the UAV deployment and the power of ground nodes in step 3.2; Step 2.2: With the goal of maximizing the overall confidentiality and rate, under the constraints of drone endurance, ground node power, and the maximum number of nodes allowed by each alliance, establish an optimization problem to maximize the overall confidentiality and rate. (4)，(12)，(13)，(14) Among them, C max Indicates the maximum number of nodes allowed in a federation, Indicates the flight area of ​​the drone. and Respectively The set of transmit power and cooperative interference power of each node in.

4. The method for secure communication of joint UAV deployment, power allocation and alliance game as claimed in claim 3, characterized in that: The implementation method of step three is: Step 3.1: In the upper optimization problem, all ground nodes form M non-overlapping coalitions, and each node has the opportunity to change the coalition formation through "switch" or "exchange" actions; generate the initial coalition formation; Step 3.2: In the underlying optimization problem, the drone acts as a receiver for M alliances at the same time, the drone hovering position is deployed, and the transmission power and cooperative interference power of each node in each alliance are redistributed; the BCD method is used to solve the underlying optimization problem, and the underlying problem is decomposed into two non-convex sub-problems. Sub-problem 1: Optimize the deployment of drones under the given ground node power allocation; Sub-problem 2: Optimize the power allocation of ground nodes under the given drone deployment; introduce auxiliary variables and use SCA technology to convert these two sub-problems into two DC planning problems and one SOC planning problem respectively. The drone deployment results are obtained by solving the two DC planning problems, and the power allocation results of the ground nodes are obtained by solving one SOC planning problem; Iteratively solve subproblems 1 and 2 so that the underlying problem converges in a finite number of iterations, and obtain the UAV deployment results and ground node power allocation results that maximize the overall confidentiality and efficiency; Step 3.3: In the upper-level optimization problem, based on the TCFG algorithm, the alliance game method is used to change the alliance formation through the "switching" or "exchange" actions of the nodes until a stable alliance formation is obtained, and the underlying non-convex optimization problem is solved under the stable alliance formation to obtain the UAV deployment results and the power allocation results of the ground nodes that maximize the overall confidentiality and efficiency. The overall confidentiality and efficiency are maximized according to the UAV deployment results and the power allocation results of the ground nodes.

5. The method for secure communication of joint UAV deployment, power allocation and alliance game as claimed in claim 4, characterized in that: Step 3.1 is implemented as follows: Step 3.1.1: The coalition formation optimization problem is expressed as st(16), In the formula Indicates the maximum alliance confidentiality and rate under the current alliance formation; Step 3.1.2: Generate the initial alliance formation, and calculate the maximum overall confidentiality and rate under the alliance formation according to step 3.2; set flag-switch←1 and flag-exchange←1.

6. The method for secure communication of joint UAV deployment, power allocation and alliance game as claimed in claim 5, characterized in that: Step 3.2 is implemented as follows: Step 3.2.1: Based on the overall confidentiality and rate maximization problem established in step 2.2, based on the fixed alliance formation, the underlying optimization problem is expressed as st(4),(12),(13),(14),(17) Using the BCD method It is decomposed into two sub-problems: sub-problem 1: optimize the UAV deployment under the given ground node power allocation; sub-problem 2: optimize the ground node power allocation under the given UAV deployment; introduce auxiliary variables and use the SCA method to transform these two sub-problems into two DC planning problems and one SOC planning problem respectively. The UAV deployment results are obtained by solving the two DC planning problems, and the ground node power allocation results are obtained by solving one SOC planning problem. Step 3.2.2: Optimize the UAV deployment given the ground node power allocation. The optimization problem is expressed as st(4),(17) where [x m [n],y m [n]] and p s,m [n] represents alliance The horizontal coordinates and transmission power of node n in ; Since the objective function in equation (20) is non-convex, the constraint in equation (4) is non-convex, so is a non-convex optimization problem; in order to transform (20) into a solvable form, an auxiliary variable is introduced and t relax the objective function in equation (20) and the constraint in equation (4) respectively; use the SCA method to transform the objective function in equation (20) into a concave function; transform the constraint in equation (4) into two equivalent constraints and write them in two convex function difference forms; Rewrite as (x U -x LAP ) 2 +(y U -y LAP ) 2 +(z U -z LAP ) 2 ≤t 2 , (23) (P b +P i )×(t·θ+D) 2 -P L ωAχe -κt ≤0, (24) (17) The first-order Taylor expansion is used to deal with the non-concavity of the objective function in equation (21) and the non-concavity of the constraints in equations (23) and (24); It's about u m [n] and with the help of a first-order Taylor expansion, we get in is u m [n] is the value at the first iteration; t in equations (23) and (24) 2 and e -κt With the help of a first-order Taylor expansion, it is expanded into t 2 ≥-(t (l) ) 2 +2tt (l) , (26) and question was rewritten as s.t.(x U -x LAP ) 2 +(y U -y LAP ) 2 +(z U -z LAP ) 2 +(t (l) ) 2 -2tt (l) ≤0, (29) (17)，(26) The objective function in (28) is concave with respect to u, and the constraints in (29) and (30) are concave with respect to {x U ,y U ,z U ,u,t} is convex; therefore, the convex optimization problem is solved by the interior point method Get the drone deployment results; Step 3.2.3: Optimize the power allocation of ground nodes given the UAV deployment; for the alliance The optimization problem is expressed as The objective function in formula (31) is a highly complex function containing multiple optimization variables and random variables; considering the optimization problem In order to improve the robustness of the algorithm, the objective function in (31) is tightened and the negative term in (31) is taken to the maximum value, which is expressed as The second term of the objective function in formula (32) has random variables in both the numerator and the denominator. The objective function in formula (32) is regarded as a random variable For the convenience of solving, Expressed as in Jensen’s inequality is applied to deal with the “-max” in the objective function in equation (32); specifically, the calculation of The upper bound and The lower bound of; when node n is the sender and node n′ is the cooperative interferer, solve the maximum eavesdropping rate from node n to the eavesdropper; due to the function The concavity of for Its lower bound is based on Jensen inequality and ln(1+e x ) is obtained by in The parameter is An exponentially distributed random variable with Where ε is the Euler constant; Substituting equation (36) into equation (35), we get The lower bound of According to formulas (33)(34)(35)(36)(37), when n′ is When there are cooperative interference nodes in the network, the maximum eavesdropping rate from node n to the eavesdropper is Therefore, the optimization problem Re-expressed as The objective function in equation (39) is highly complex; for ease of observation, the variable Transformed into an equivalent optimization problem, expressed as in and The non-convex constraints in (44), (45), and (46) can be equivalently written as The non-convex constraints in (47), (48), and (49) are linearized around the current point in the lth iteration to obtain Therefore, the optimization problem Re-expressed as st(12),(13),(14),(41),(42),(43),(50),(51),(52) Optimization Problem This is an optimization problem involving SOC constraints. We use CVX to solve it and get the power allocation result of the ground nodes. Step 3.2.4: Use the optimization problem from Step 3.2.2 and Step 3.2.3 and Calculate the results of UAV deployment and ground node power allocation respectively, and solve iteratively and Until convergence, the results of UAV deployment and ground node power allocation are finally obtained.

7. The method for secure communication of joint UAV deployment, power allocation and alliance game as claimed in claim 6, characterized in that: Step 3.3 is implemented as follows: Step 3.3.1: In the upper optimization problem, the framework and definition of the developed TCFG algorithm are as follows; the TCFG algorithm divides all nodes into M non-overlapping coalitions Use the two alliance rules of "switch partial order" and "exchange partial order" to update the alliance formation; each node follows the same alliance update rule and has its own alliance formation priority; after traversing all nodes, the alliance formation with the largest overall confidentiality and rate is the updated alliance formation; consider the node Current alliance formation Change node n from Switch to The utility is expressed as in is the coalition formed after node n moves; and Based on and The maximum overall confidentiality and rate obtained; therefore, u m,m′ [n]Describe the current alliance formation Next node n from Switch to The contribution of the switching partial order is defined as follows; if there is a node and Alliance m≠m′ satisfies the following conditions: Remove node n from the alliance Switch to Consider two nodes and m′≠m, the current alliance is formed The utility of the node exchange alliance is in is the alliance formed after node exchange; therefore, v m,m′ [n,n′] describes the current alliance form The two nodes of the exchange alliance and The contribution of the exchange partial order is defined as follows; if there are two nodes m′≠m satisfies the following conditions: Swap nodes n and n′ with the alliances they currently belong to; Step 3.3.2: Traverse all nodes according to the switching partial order and the exchange partial order; in each traversal, select the alliance formation corresponding to the maximum value of the overall confidentiality and rate as the new alliance formation, and then perform the next traversal until the overall confidentiality and rate no longer increase; the specific operation is as follows; traverse all nodes according to the switching partial order rule; the utility of the current alliance formation is For nodes node Traversing the Select Union m≠m′, and from Switch to Use step 3.2 to calculate the maximum overall confidentiality and rate under the new alliance formation Compare and And store the result u m,m′ [n]; if present m′≠m satisfies u m,m′ [n]>0, find the largest u m,m′ [n] The corresponding alliance is formed and the current alliance is updated, setting flag-switch←1; otherwise, setting flag-switch←0; repeating the above steps until flag-switch←0; traversing all nodes according to the exchange partial order rule; the utility of the current alliance is For nodes node Traversing Selections m′≠m and exchange alliances with node n′; use step 3.2 to calculate the maximum overall confidentiality and rate under the formation of the new alliance Compare and and store the result v m,m′ [n,n′]; if it exists m′≠m satisfies v m,m′ [n,n′]>0, find the largest v m,m′ The alliance corresponding to [n,n′] is formed and the current alliance formation is updated, and flag-exchange←1 is set; otherwise, flag-exchange←0 is set; Repeat the above steps until flag-exchange←0, that is, a stable alliance is formed; under the stable alliance formation, use step 3.2 to solve the underlying non-convex optimization problem to obtain the UAV deployment results and the power allocation results of the ground nodes that maximize the overall confidentiality and rate, and maximize the overall confidentiality and rate based on the UAV deployment results and the power allocation results of the ground nodes.

Images