Safety flux calculation integrated waveform and unmanned aerial vehicle 3D trajectory optimization design method

Abstract

The invention discloses an optimization design method for a safety, sensing and calculation integrated waveform and an unmanned aerial vehicle 3D trajectory, and the method comprises the steps: firstly constructing a communication system model, and constructing an optimization model with the calculation frequency of an unmanned aerial vehicle, the sensing and calculation integrated waveform and the 3D flight trajectory of the unmanned aerial vehicle as optimization variables by taking the maximization of a safety rate as a target; the optimization problem of communication performance, sensing performance, calculation performance and unmanned aerial vehicle energy consumption constraint is considered; then, based on a block coordinate descent method, an original optimization problem is decoupled into sub-problems about the calculation frequency of the unmanned aerial vehicle, the integrated beam forming of the common inductance calculation and the 3D flight path of the unmanned aerial vehicle, and convex approximation fitting is carried out on each non-convex sub-problem by using methods such as continuous convex approximation and first-order Taylor expansion; and finally, solving by adopting an iterative algorithm, and finding a global suboptimal solution of the whole optimization problem, an optimal unmanned aerial vehicle calculation frequency, a general inductance calculation integrated waveform and an unmanned aerial vehicle 3D flight path. According to the invention, while the communication security is ensured, the compromise maximization of the three-dimensional performance of the communication perception calculation is completed.

Description

Technical Field

[0001] This invention relates to the field of integrated sensing and computing, specifically to a method for designing safe integrated sensing and computing waveforms and 3D trajectories for unmanned aerial vehicles. Background Technology

[0002] The statements in this section are merely background information relating to this disclosure, and these statements may constitute prior art. In the process of developing this invention, the inventors discovered at least the following problems in the prior art.

[0003] Wireless sensing performance is an important development direction for future 6G, and the integration of communication, sensing, and computing technologies is a current research hotspot. As a core node of future mobile communication networks, drones have a wide range of applications and are attracting attention from all walks of life.

[0004] In the literature [C. Chen, J. Yao, M. Jin, and Q. Guo, "Beamforming and computing capacity allocation for ISAC-assisted secure mobile edge computing," IEEE Wireless Commun. Lett., pp. 1-1, Sep. 2024.], multiple users transmit uplink information to a base station equipped with a mobile edge computing server for computation, while a potential eavesdropper is simultaneously eavesdropping. The duplex base station sends radar signals to interfere with the potential eavesdropper, thus achieving secure communication.

[0005] The reference [Zhao Y, Wu Q, Chen W, et al. Multi-functional beamforming design for integrated sensing, communication, and computation[J]. IEEE Transactions on Communications, 2024, doi:10.1109 / tcomm.2024.3519551 1-1.] considers a multi-functional base station that simultaneously performs communication, sensing, and computation functions. It maximizes the performance of these three functions through multi-functional beamforming optimization.

[0006] In [N. Huang, C. Dou, Y. Wu, L. Qian, B. Lin, and H. Zhou, "Unmanned-aerial-vehicle-aided integrated sensing and computation with mobile-edge computing," IEEE Internet Things J., vol. 10, no. 19, pp. 16830-16844, Oct. 2023.], the authors used a drone to send radar signals to sense targets and offloaded the received reflected radar information to an edge server for computation.

[0007] However, according to the applicant's research, all of the above solutions have the following problems:

[0008] 1. Research on the integration of sensing and computing did not take into account information security.

[0009] 2. In the research on drone trajectory optimization integrating sensing and computing, the optimization of 3D trajectory was not considered.

[0010] 3. In the research on the integration of communication, sensing and computing, the performance boundaries of communication performance, sensing performance and computing performance have not been fully explored.

[0011] For example, patent application number 202311511305.4, entitled "A Design Method for UAV Trajectory and Beamforming in a Synchrotron Computing Resource Fusion Network," exemplifies this. It describes the problem of maximizing computational throughput under perception quality constraints by establishing a synchrotron computing resource fusion system with non-orthogonal multiple access and UAV assistance, and then uses deep learning methods to solve it. However, this type of patent does not consider information security, optimizes the 2D trajectory of the UAV, and lacks discussion on the performance boundaries of communication, perception, and computation.

[0012] The main problem addressed by this patent is how to explore the performance boundaries of communication, perception, and computing while ensuring secure communication, and how to fully leverage the flexible deployment capabilities of UAVs. Summary of the Invention

[0013] In view of the above problems, the purpose of this invention is to solve some of the problems in the prior art, or at least alleviate these problems.

[0014] A method for optimizing waveforms and 3D trajectory of a UAV that integrates sensing and computation for safety includes the following steps:

[0015] Construct a communication system model: Set up a drone S as an airborne base station, and initialize the flight trajectory of S to provide downlink communication services to K users on the ground. At the same time, it senses the target J and offloads some of the sensed data to a ground base station. Consider the existence of a potential eavesdropper E. In order to ensure the security of information transmission, the drone emits artificial noise to interfere with E and ensure the security of communication.

[0016] Constructing an optimized mathematical model: With the goal of maximizing the safe rate, an optimized mathematical model is constructed with the UAV's computing frequency, integrated waveform of communication, sensing and computing, and the UAV's 3D flight trajectory as optimization variables, taking into account communication performance, sensing performance, computing performance, and UAV energy consumption constraints.

[0017] For the aforementioned optimized mathematical model, the original problem is decoupled into subproblems based on the block coordinate descent method, which concern only the UAV's computational frequency, the integrated waveform of the synesthesia and computation, the UAV's horizontal flight trajectory, and the UAV's vertical flight trajectory.

[0018] Regarding the aforementioned sub-problem and The problem is transformed into a convex problem by using a convex approximation fitting method; where, in solving the subproblems... When dealing with the non-convex problem of calculating frequency and integrated waveform of syn-induction computing for UAVs, a positive semidefinite relaxation method that discards the rank-one constraint is used to solve the problem. Then, the solution is obtained by Gaussian randomization to obtain a solution that satisfies the rank-one constraint.

[0019] An iterative algorithm is used to solve for the drone's calculated frequency, integrated waveform of induction and computation, horizontal flight trajectory, and vertical flight trajectory.

[0020] Furthermore, the optimized mathematical model is as follows:

[0021]

[0022] C13:(q s [1],z[1])=(q I ,z I ),(q s [N],z[N])=(q F ,z F )

[0023] In the formula, F, Q s Z represents the optimization variable, where It is the computing frequency of the drone; It is a set of beamforming matrices integrating induction and computation, where W c,l [n] represents the beamforming matrix of the communication signal transmitted by the UAV, Wr [n] represents the beamforming matrix of the sensing signals transmitted by the UAV; It is the horizontal trajectory of the drone. It is the vertical trajectory of the drone; R s,k [n] is user U k At the receiving rate in time slot n, R e,k [n] is the value received by the eavesdropping E in time slot n and sent to user U. k The rate at which information is eavesdropped; Γ s,k ,Γ e,k The communication performance threshold and eavesdropping threshold of the UAV are defined respectively, where D[n] represents the amount of sensing data processed by the UAV, and R rad [n] represents the radar sensing mutual information rate, D min , f max The minimum amount of sensory data that a UAV can process, the minimum radar sensing mutual information rate, and the maximum computing frequency of a UAV are defined respectively. E represents the maximum transmit power of the drone. tra E fly E com , The power consumption for drone transmission, flight, computing, and battery capacity are defined respectively, δ. t The length of each flight time slot for the drone is defined. The maximum horizontal and vertical flight speeds of the drone are defined, z min z max q represents the minimum and maximum altitudes of the drone's flight. e q Δ The estimated location of the eavesdropping and the range of the eavesdropping are represented by (q). I ,z I ), (q F ,z F These are the starting and ending positions of the drone, respectively. and These represent the messages sent to U. k The signals from base station B and the eavesdropping signal E, and satisfy the following conditions: The corresponding transmission beamforming is represented as q s [n] represents the horizontal coordinate of the UAV S;

[0024] Among them, C1 to C2 are the signal-to-noise ratio constraints of the drone and the eavesdropper, which are requirements for secure transmission; C3 is the constraint of computing performance; C4 is the constraint of sensing performance; C5 limits the computing frequency of the drone; C6 to C7 are the constraints of beamforming of the integrated communication, sensing and computing technologies; C8 to C9 limit the transmission power and energy consumption of the drone; and C10 to C13 limit the speed, altitude, starting position and ending position of the drone.

[0025] The sub-problem for:

[0026]

[0027] Among them, H s,k [n]、H s,e [n] and H s,B [n] represents the distance from the drone to the user U. k The channel gain of E and ground base station B, Φ s,k [n]、Φ s,e [n] and Φ s,B [n] represents user U k Eavesdropping on interference and noise received by E and ground base station B, Γ B The threshold represents the computation speed, and tr represents the operation of finding the trace of a matrix;

[0028] The sub-problem for:

[0029]

[0030] The sub-problem for:

[0031]

[0032] Furthermore, the sub-problems Transforming it into a convex problem involves the following steps:

[0033] Replace C1.1.3 with the continuous convex approximation method.

[0034]

[0035] in,(·) (m) This represents the value in the m-th iteration. Introducing auxiliary variables a1 and a2, the objective function can be replaced with:

[0036]

[0037] When the rank-one constraint C7 is ignored, the entire optimization problem remains a non-convex problem. We first fix {a1, a2} to solve it. Re-fix Solve for {a1,a2} by iterating until convergence; then use Gaussian randomization to process the ignored constraint C7 and obtain an approximate solution that satisfies the rank-one condition.

[0038] Furthermore, the sub-problems The problem is transformed into a convex problem by introducing a series of relaxation variables and replacing the original non-convex constraints with convex constraints that are easier to solve using a continuous convex approximation method; C1-C4, C9, and the objective function can be re-expressed as:

[0039]

[0040] in, φ s,l [n], v1[n], v2[n], It is a new slack variable that has been introduced. It is expressed as the square of the distance from the drone to the corresponding location. This is expressed as the corresponding noise power; This refers to a constant value that is independent of the variable in the subproblem; and These are the corresponding auxiliary variables, and all are linear functions, where l∈{1,…,K,B,e,J}; C and D represent the relevant parameters in the probabilistic line-of-sight link, and P... b and P i These are the blade shape power and induced power of the drone in hovering state, U tip V represents the tip velocity of the rotor blades, v0 represents the average rotor-induced velocity when the UAV is hovering, d0, ρ, s, and A represent the fuselage drag ratio, air density, rotor rigidity, and rotor disk area, respectively, W represents the weight of the UAV, and V z [n] represents the vertical velocity of the drone.

[0041] Furthermore, the sub-problems In the case of a convex problem, for the non-convex constraint C2 and the objective function, the following equation can be used instead:

[0042]

[0043] in, It is a new slack variable. and These are the corresponding auxiliary variables, and they are all linear functions.

[0044] A module for implementing a method for optimizing waveform and 3D trajectory design of unmanned aerial vehicles (UAVs) through integrated sensing and computation for secure communication includes:

[0045] Initialization module: Set up UAV S as an airborne base station, and initialize the flight trajectory of S to provide downlink communication services to K users on the ground. At the same time, it senses the target J and offloads some of the sensed data to a ground base station. Considering the existence of a potential eavesdropper E, in order to ensure the security of information transmission, the UAV emits artificial noise to interfere with E and ensure the security of communication.

[0046] Mathematical model construction module: With the goal of maximizing safe speed, an optimization mathematical model is constructed, taking the UAV computing frequency, the integrated waveform of communication, sensing and computing, and the UAV 3D flight trajectory as optimization variables, and considering communication performance, sensing performance, computing performance, and UAV energy consumption constraints.

[0047] Mathematical Model Decoupling Module: For the optimized mathematical model, the original problem is decoupled into subproblems based on the block coordinate descent method, which only concern the UAV's calculation frequency, the integrated waveform of the synesthetic computation, the UAV's horizontal flight trajectory, and the UAV's vertical flight trajectory. For each non-convex subproblem, a convex approximation fitting method is used to convert it into a convex problem for solution. Specifically, when solving the non-convex problem concerning the integrated waveform of the synesthetic computation, a positive semidefinite relaxation method that discards the rank-one constraint is used for solution. Then, the obtained solution is randomized through Gaussian to obtain a solution that satisfies the rank-one constraint.

[0048] Iterative solution module: Uses iterative algorithms to solve for the UAV's calculated frequency, integrated waveform of induction and computation, horizontal flight trajectory, and vertical flight trajectory.

[0049] A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the security sensing and computing integrated waveform and UAV 3D trajectory optimization design method.

[0050] The present invention has the following beneficial effects:

[0051] This invention utilizes beamforming design to emit artificial noise towards potential eavesdroppers, ensuring communication security. Furthermore, by optimizing beamforming and the UAV's 3D trajectory, the performance of the UAV's integrated communication, sensing, and computing system is improved in multiple aspects, thereby maximizing the trade-off between communication security and the three-dimensional performance of communication, sensing, and computing. Attached Figure Description

[0052] Figure 1 This is a communication system model of the present invention;

[0053] Figure 2 This is a flowchart of the iterative algorithm proposed in this invention for solving optimization problems. Detailed Implementation

[0054] The present invention will be further described below with reference to the accompanying drawings. The embodiments of the present invention are only used to illustrate the present invention and not to limit the present invention. Various substitutions and modifications made based on ordinary technical knowledge and common practices in the art without departing from the technical concept of the present invention should be included within the scope of the present invention.

[0055] For multi-functional unmanned aerial vehicle (UAV) systems, in order to optimize the system's communication, sensing, and computing performance from multiple aspects, and to ensure the security and effectiveness of communication, this application... Figure 2 As shown, a system model is first constructed, including a communication model, a perception model, a computation model, and an energy consumption model. Then, several optimization problem models are proposed, including a communication-centric optimization problem, a perception-centric optimization problem, a computation-centric optimization problem, and an optimization problem considering trade-offs among the three performance aspects. For each optimization problem, the block coordinate descent method is used to decouple the original optimization problem into subproblems concerning the UAV's computational frequency, the integrated sensing waveform, and the UAV's 3D flight trajectory. A convex approximation is then performed on each non-convex subproblem. Finally, an iterative algorithm is used to find the global suboptimal solution for each optimization problem, as well as the optimal integrated sensing waveform and the UAV's 3D flight trajectory. The specific scheme is as follows.

[0056] S1: Construction of the communication model.

[0057] A drone S is set up as an airborne base station, and its flight trajectory is initialized to provide downlink communication services to K users on the ground. It also senses a target J and offloads some of the sensed data to a ground base station. Considering the existence of a potential eavesdropper E, to ensure the security of information transmission, the drone emits artificial noise to interfere with E, thus guaranteeing communication security.

[0058] like Figure 1 As shown, this invention mainly studies a method for designing waveforms and 3D trajectories of UAVs that integrates sensing and computation for secure communication. The UAV S flies from a starting point to a destination, providing downlink communication services to K ground users. Simultaneously, it transmits radar signals to the target for sensing and offloads some of the acquired sensing data to a ground base station for processing. We consider a 3D Cartesian coordinate system, where the horizontal coordinates of the ground communication users, the target J, and the UAV S are q. k =[x k ,y k ] T ,q J =[x J ,y J ] T ,q s [n] = [x] s [n],y s [n] T。 [] T The transpose operation of a matrix is ​​used. Since the coordinates of E cannot be known precisely in reality, we assume that E lies within a known range, i.e. in, The center of the range is indicated by Δ, which represents the radius of the area and is much smaller than the distance from the drone to the eavesdropping point. Two points are configured such that M = M x ×M y The uniform UPA of the root antenna is used to receive and transmit signals, where M x and M y This indicates the number of antennas on the x-axis and y-axis, with the antenna spacing being half a wavelength.

[0059] The conjugate transpose of the radar transmitting array steering vector can be expressed as:

[0060]

[0061] Indicates S to user U k The distance between base station B, eavesdropping device E, and sensing target J. Among them, Kronecker product operation, () H This represents the conjugate transpose operation of a matrix, where j represents the imaginary sign, and x represents the conjugate transpose of a matrix. l and y l q represents the x and y coordinates of the corresponding node. l The z-coordinate represents the horizontal coordinate of the corresponding node. s [n] is defined as the altitude of the drone.

[0062] The signal transmitted from UAV S in time slot n can be represented as:

[0063]

[0064] The radar signal is represented as x. r,m [n] and satisfy Radar beamforming is represented as This represents the operation of finding the expected value, where the radar beamforming matrix is ​​represented as W. r [n] = [w r,1 [n],…,w r,M [n]], the radar transmitted signal is represented as x r [n] = [x] r,1 [n];…;x r,M [n]]. Furthermore... and These represent the messages sent to U. k The signals from base station B and the eavesdropping signal E, and satisfy the following conditions: The corresponding transmission beamforming is represented as Without loss of generality, it is assumed that communication and radar signals are statistically independent.

[0065] Assume all communication and sensing channel models are probabilistic line-of-sight links. S and U k The probability that there is a probabilistic line-of-sight link between base station B, eavesdropping E, and sensing target J is:

[0066]

[0067] Where C and D are constant parameters. This indicates the elevation angle. Additionally, the probability of a non-probabilistic line-of-sight link occurring is...

[0068] The path loss of the communication channel under probabilistic line-of-sight (LOS) and non-probabilistic line-of-sight (NOS) conditions can be modeled as follows: and Where β0 represents the channel gain at a unit distance d0 = 1m. α < 1 represents the additional fading coefficient under non-probabilistic line-of-sight link conditions. The communication channel model set in n time slots can be modeled as follows:

[0069]

[0070] in, Nonprobabilistic line-of-sight link parameters It is a complex Gaussian random vector with zero mean and a unit covariance matrix.

[0071] User U k The signal-to-interference-plus-noise ratio can be expressed as:

[0072]

[0073] Among them, w c,i This represents the beamforming vector sent to other users.

[0074] The signal-to-interference-plus-noise ratio (SIR) of ground base station B can be expressed as:

[0075]

[0076] in, They represent U respectively k The noise power received by B, and user U k The receiving rate of base station B can be expressed as R s,k [n] = log2(1+γ) s,k [n]), R s,B [n] = log2(1+γ) s,B [n]).

[0077] Assuming that the eavesdropping device E is only interested in the information sent to the communication user and not in the offloaded sensing data, the offloaded sensing data will become part of the interference received by the eavesdropping device E. This signal-to-interference-plus-noise ratio (SIR) can be expressed as:

[0078]

[0079] in, This represents the noise power received by eavesdropping device E, which is transmitted to user U. k The rate of information is represented by R. e,k [n] = log2(1+γ) e,k [n]).

[0080] The UAV transmits radar signals towards the target and the reflected signals are received by the UAV. The Doppler shift caused by the UAV's motion can be set as a constant within a time slot and can be well compensated for. The sensing channel can be represented as:

[0081]

[0082] Where ξ represents the radar cross-sectional area, The elements are complex Gaussian random variables with zero mean and unit variance. The radar receiver signal-to-noise ratio of S can be expressed as:

[0083]

[0084] in, This represents the noise power received by S.

[0085] S2: Construct an optimized mathematical model.

[0086] With the goal of maximizing safe speed, an optimization mathematical model is constructed, taking the UAV computing frequency, the integrated waveform of communication, sensing and computing, and the UAV 3D flight trajectory as optimization variables, and considering constraints such as communication performance, sensing performance, computing performance, and UAV energy consumption.

[0087] The optimized mathematical model is as follows:

[0088]

[0089] C13:(q s [1],z[1])=(q I ,z I ),(q s [N],z[N])=(q F ,z F )

[0090] In the formula, F, Qs Z represents the optimization variable, where It is the computing frequency of the drone; It is a set of beamforming matrices integrating induction and computation, where W c,l [n] represents the beamforming matrix of the communication signal transmitted by the UAV, W r [n] represents the beamforming matrix of the sensing signals transmitted by the UAV; It is the horizontal trajectory of the drone. It is the vertical trajectory of the drone; R s,k [n] is user U k At the receiving rate in time slot n, R e,k [n] is the value received by the eavesdropping E in time slot n and sent to user U. k The rate at which information is eavesdropped; Γ s,k ,Γ e,k The communication performance threshold and eavesdropping threshold of the UAV are defined respectively, where D[n] represents the amount of sensing data processed by the UAV, and R rad [n] represents the radar sensing mutual information rate, D min , f max The minimum amount of sensory data that a UAV can process, the minimum radar sensing mutual information rate, and the maximum computing frequency of a UAV are defined respectively. E represents the maximum transmit power of the drone. tra E fly E com , The power consumption for drone transmission, flight, computing, and battery capacity are defined respectively, δ. t The length of each flight time slot for the drone is defined. The maximum horizontal and vertical flight speeds of the drone are defined, z min z max q represents the minimum and maximum altitudes of the drone's flight. e q Δ The estimated location of the eavesdropping and the range of the eavesdropping are represented by (q). I ,z I ), (q F ,z F These are the starting and ending positions of the drone, respectively. and These represent the messages sent to U. k The signals from base station B and the eavesdropping signal E, and satisfy the following conditions: The corresponding transmission beamforming is represented as q s [n] represents the horizontal coordinate of the UAV S;

[0091] Among them, C1 to C2 are the signal-to-noise ratio constraints of the drone and the eavesdropper, which are requirements for secure transmission; C3 is the constraint of computing performance; C4 is the constraint of sensing performance; C5 limits the computing frequency of the drone; C6 to C7 are the constraints of beamforming of the integrated communication, sensing and computing technologies; C8 to C9 limit the transmission power and energy consumption of the drone; and C10 to C13 limit the speed, altitude, starting position and ending position of the drone.

[0092] The amount of data D[n] processed by the UAV in each time slot can be represented as:

[0093] D[n]=min{(R s,B [n]+f s [n] / F s )δ t D J}

[0094] Among them, F s (cycles / bit) is the number of CPU cycles required by S to compute 1 bit of data, D J The amount of data sampled by the radar sensing data is a constant.

[0095] To address the non-convexity and high coupling of the optimization problem corresponding to the system model, an approximate fitting method based on block coordinate descent algorithm and continuous convex approximation was adopted to decouple and transform the original optimization problem.

[0096] The solution to the problem of this invention will be based on the following steps:

[0097] S3: Decouple the optimization mathematical model into subproblems

[0098] For the aforementioned optimized mathematical model, the original problem is decoupled based on the block coordinate descent method into a problem that only concerns the UAV's calculation frequency, the integrated waveform of the synaptic and computational systems, the UAV's horizontal flight trajectory, and the UAV's vertical flight trajectory.

[0099] The sub-problem for:

[0100]

[0101] Among them, H s,k [n]、H s,e [n] and H s,B [n] represents the distance from the drone to the user U. k The channel gain of E and ground base station B, Φ s,k [n]、Φ s,e [n] and Φ s,B [n] represents user U kEavesdropping on interference and noise received by E and ground base station B, Γ B The threshold represents the computation speed, and tr represents the operation of finding the trace of a matrix;

[0102] The sub-problem for:

[0103]

[0104] The sub-problem for:

[0105]

[0106] S4: Transform a non-convex subproblem into a convex problem and solve it.

[0107] Regarding the sub-problems and Both problems involve non-convex constraints and non-concave objective functions. Therefore, we need to use continuous convex approximation methods to introduce slack variables and perform convex approximation fitting on the non-convex constraints, thereby transforming the original problem into a solvable convex optimization problem.

[0108] S41: Addressing Subproblems The objective function and constraint C1.1.3 are non-convex. To achieve a better solution, a continuous convex approximation method is used to replace C1.1.3 with...

[0109]

[0110] in,(·) (m) This represents the value in the m-th iteration. Introducing auxiliary variables a1 and a2, the objective function can be replaced with:

[0111]

[0112] When the rank-one constraint C7 is ignored, the entire optimization problem remains a non-convex problem. We first fix {a1, a2} to solve it. Re-fix Solve for {a1,a2} by iterating until convergence; then use Gaussian randomization to process the ignored constraint C7 and obtain an approximate solution that satisfies the rank-one condition.

[0113] S42: Addressing Subproblems To address non-convex constraints such as C1-C4 and C9, a series of relaxation variables are introduced, and a continuous convex approximation method is used to replace the original non-convex constraints with convex constraints that are easier to solve. Therefore, C1-C4, C9, and the objective function can be re-expressed as follows:

[0114]

[0115]

[0116] in, φ s,l [n], v1[n], v2[n], It is a new slack variable that has been introduced. It is expressed as the square of the distance from the drone to the corresponding location. This is expressed as the corresponding noise power. This refers to a constant value that is independent of the variable in the subproblem; and These are the corresponding auxiliary variables, and all are linear functions, where l∈{1,…,K,B,e,J}. C and D represent the relevant parameters in the probabilistic line-of-sight link, and P... b and P i These are the blade shape power and induced power of the drone in hovering state, U tip V represents the tip velocity of the rotor blades; v0 represents the average rotor induced velocity when the UAV is hovering; d0, ρ, s, and A represent the fuselage drag ratio, air density, rotor rigidity, and rotor disk area, respectively; W represents the weight of the UAV, V z [n] represents the vertical velocity of the drone.

[0117] S43: Addressing Subproblems Using the solution A similar method is used to solve the problem. The difference is that for the non-convex constraint C2 and the objective function, the following formula needs to be used instead:

[0118]

[0119] in, It is a new slack variable. and These are the corresponding auxiliary variables, and they are all linear functions.

[0120] Through the above operations, the sub-problem and They have all been transformed into solvable convex optimization problems, which can be solved using the convex optimization solution tool CVX.

[0121] S5: Use iterative algorithms to solve the convex optimization problem of the transformation, thereby finding the global suboptimal solution of the entire optimization problem and the optimal UAV calculation frequency, integrated waveform of induction and computation, and UAV 3D flight trajectory.

[0122] The iterative algorithm flow is designed to connect the above convex optimization problems in sequence. The specific steps are shown in Table 1.

[0123] Table 1

[0124]

[0125] This invention can be represented by the following modules:

[0126] A module for implementing a method for optimizing waveform and 3D trajectory design of unmanned aerial vehicles (UAVs) through integrated sensing and computation for secure communication includes:

[0127] Initialization module: Set up UAV S as an airborne base station, and initialize the flight trajectory of S to provide downlink communication services to K users on the ground. At the same time, it senses the target J and offloads some of the sensed data to a ground base station. Considering the existence of a potential eavesdropper E, in order to ensure the security of information transmission, the UAV emits artificial noise to interfere with E and ensure the security of communication.

[0128] Mathematical model construction module: With the goal of maximizing safe speed, an optimization mathematical model is constructed, taking the UAV computing frequency, the integrated waveform of communication, sensing and computing, and the UAV 3D flight trajectory as optimization variables, and considering communication performance, sensing performance, computing performance, and UAV energy consumption constraints.

[0129] Mathematical Model Decoupling Module: For the optimized mathematical model, the original problem is decoupled into subproblems based on the block coordinate descent method, which only concern the UAV's calculation frequency, the integrated waveform of the synesthetic computation, the UAV's horizontal flight trajectory, and the UAV's vertical flight trajectory. For each non-convex subproblem, a convex approximation fitting method is used to convert it into a convex problem for solution. Specifically, when solving the non-convex problem concerning the integrated waveform of the synesthetic computation, a positive semidefinite relaxation method that discards the rank-one constraint is used for solution. Then, the obtained solution is randomized through Gaussian to obtain a solution that satisfies the rank-one constraint.

[0130] Iterative solution module: Uses iterative algorithms to solve for the UAV's calculated frequency, integrated waveform of induction and computation, horizontal flight trajectory, and vertical flight trajectory.

[0131] A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the security sensing and computing integrated waveform and UAV 3D trajectory optimization design method.

[0132] Step one of this invention first constructs a UAV-assisted wireless communication system model based on the fundamental theories of wireless communication and physical layer security. Steps two, three, and four refine the mathematical derivation and problem-solving analysis during the model building process. Addressing the non-convexity and high coupling of the optimization problem corresponding to this system model, an approximate fitting method based on block coordinate descent and continuous convex approximation is used to decouple and transform the original optimization problem. Step five, based on iterative thinking, iterates the sub-problems obtained in steps two, three, and four until the total user energy consumption continuously approaches a constant value. This constant value is the minimum total user energy consumption required by this invention.

[0133] This invention addresses the challenges of multi-functional unmanned aerial vehicle (UAV) systems that require consideration of various performance, energy consumption, and security aspects. It leverages the flexible deployment capabilities of UAVs and optimizes beamforming to enhance communication and sensing performance. Furthermore, it utilizes artificial noise to disrupt potential eavesdroppers and improves computational performance through data offloading. By integrating UAVs with physical layer security, communication, sensing, and computing, and optimizing beamforming and 3D trajectory design, this invention demonstrates high innovation and uniqueness.

Claims

1. A method for optimizing the design of waveforms and 3D trajectories of unmanned aerial vehicles (UAVs) using a combination of sensing and computation, characterized in that: Includes the following steps: Construct a communication system model: Set up a drone S as an airborne base station, and initialize the flight trajectory of S to provide downlink communication services to K users on the ground. At the same time, it senses the target J and offloads some of the sensed data to a ground base station. Consider the existence of a potential eavesdropper E. In order to ensure the security of information transmission, the drone emits artificial noise to interfere with E and ensure the security of communication. Constructing an optimized mathematical model: With the goal of maximizing the safe rate, an optimized mathematical model is constructed with the UAV's computing frequency, integrated waveform of communication, sensing and computing, and the UAV's 3D flight trajectory as optimization variables, taking into account communication performance, sensing performance, computing performance, and UAV energy consumption constraints. For the aforementioned optimized mathematical model, the original problem is decoupled into subproblems based on the block coordinate descent method, which only concern the UAV's calculation frequency, the integrated waveform of the synesthesia and computation, the UAV's horizontal flight trajectory, and the UAV's vertical flight trajectory. Regarding the sub-problem and The problem is transformed into a convex problem by using a convex approximation fitting method; where, in solving the subproblems... When dealing with the non-convex problem of calculating frequency and integrated waveform of syn-induction computing for UAVs, a positive semidefinite relaxation method that discards the rank-one constraint is used to solve the problem. Then, the solution is obtained by Gaussian randomization to obtain a solution that satisfies the rank-one constraint. An iterative algorithm is used to solve for the drone's calculated frequency, integrated waveform of induction and computation, horizontal flight trajectory, and vertical flight trajectory.

2. The integrated waveform and UAV 3D trajectory optimization design method for safety sensing and calculation as described in claim 1, characterized in that, The optimized mathematical model is as follows: In the formula, F, Q s Z represents the optimization variable, where It is the drone's computing frequency; It is a set of beamforming matrices integrating induction and computation, where W c,l [n] represents the beamforming matrix of the communication signal transmitted by the UAV, W r [n] represents the beamforming matrix of the sensing signals transmitted by the UAV; It is the horizontal trajectory of the drone. It is the vertical trajectory of the drone; R s,k [n] is user U k At the receiving rate in time slot n, R e,k [n] is the value received by the eavesdropping E in time slot n and sent to user U. k The rate at which information is eavesdropped; Γ s,k ,Γ e,k The communication performance threshold and eavesdropping threshold of the UAV are defined respectively, where D[n] represents the amount of sensing data processed by the UAV, and R rad [n] represents the radar sensing mutual information rate, D min , f max The minimum amount of sensory data that a UAV can process, the minimum radar sensing mutual information rate, and the maximum computing frequency of a UAV are defined respectively. E represents the maximum transmit power of the drone. tra E fly E com , The power consumption for drone transmission, flight, computing, and battery capacity are defined respectively, δ. t The length of each flight time slot for the drone is defined. The maximum horizontal and vertical flight speeds of the drone are defined, z min z max q represents the minimum and maximum altitudes of the drone's flight. e q Δ The estimated location of the eavesdropping and the range of the eavesdropping are represented by (q). I ,z I ), (q F ,z F These are the starting and ending positions of the drone, respectively. and These represent being sent to U. k The signals from base station B and the eavesdropping signal E, and satisfy the following conditions: The corresponding transmission beamforming is represented as q s [n] represents the horizontal coordinate of the UAV S; Among them, C1 to C2 are the signal-to-noise ratio constraints of the drone and the eavesdropper, which are requirements for secure transmission; C3 is the constraint of computing performance; C4 is the constraint of sensing performance; C5 limits the computing frequency of the drone; C6 to C7 are the constraints of beamforming of the integrated communication, sensing and computing technologies; C8 to C9 limit the transmission power and energy consumption of the drone; and C10 to C13 limit the speed, altitude, starting position and ending position of the drone.

3. The integrated waveform and UAV 3D trajectory optimization design method for safety sensing and calculation as described in claim 2, characterized in that, The sub-problem for: Among them, H s,k [n]、H s,e [n] and H s,B [n] represents the distance from the drone to the user U. k The channel gain of E and ground base station B, Φ s,k [n]、Φ s,e [n] and Φ s,B [n] represents user U k Eavesdropping on interference and noise received by E and ground base station B, Γ B The threshold represents the computation speed, and tr represents the operation of finding the trace of a matrix; The sub-problem for: The sub-problem for:

4. The integrated waveform and UAV 3D trajectory optimization design method for safety sensing and computation as described in claim 3, characterized in that, Subproblems Transforming it into a convex problem involves the following steps: Replace C1.1.3 with the continuous convex approximation method. in,(·) (m) This represents the value in the m-th iteration. Introducing auxiliary variables a1 and a2, the objective function can be replaced with: When the rank-one constraint C7 is ignored, the entire optimization problem remains a non-convex problem. We first fix {a1, a2} to solve it. Re-fix Solve for {a1,a2} by iterating until convergence; then use Gaussian randomization to process the ignored constraint C7 and obtain an approximate solution that satisfies the rank-one condition.

5. The integrated waveform and UAV 3D trajectory optimization design method for safety sensing and calculation as described in claim 3, characterized in that, Subproblems The problem is transformed into a convex problem by introducing a series of relaxation variables and replacing the original non-convex constraints with convex constraints that are easier to solve using a continuous convex approximation method; C1-C4, C9, and the objective function can be re-expressed as: in, φ s,l [n], v1[n], v2[n], It is a new slack variable that has been introduced. σ is expressed as the square of the distance from the drone to the corresponding location. l 2 This is expressed as the corresponding noise power; This refers to a constant value that is independent of the variable in the subproblem; and These are the corresponding auxiliary variables, and all are linear functions, where l∈{1,…,K,B,e,J}; C and D represent the relevant parameters in the probabilistic line-of-sight link, and P... b and P i These are the blade shape power and induced power of the drone in hovering state, U tip V represents the tip velocity of the rotor blades, v0 represents the average rotor-induced velocity when the UAV is hovering, d0, ρ, s, and A represent the fuselage drag ratio, air density, rotor rigidity, and rotor disk area, respectively, W represents the weight of the UAV, and V z [n] represents the vertical velocity of the drone.

6. The integrated waveform and UAV 3D trajectory optimization design method for safety sensing and computation as described in claim 3, characterized in that, Subproblems In the case of a convex problem, for the non-convex constraint C2 and the objective function, the following equation can be used instead: in, It is a new slack variable. and These are the corresponding auxiliary variables, and they are all linear functions.

7. A module for implementing a method for optimizing the design of waveforms and 3D trajectories of unmanned aerial vehicles (UAVs) through integrated sensing and computation for safety, characterized in that... include: Initialization module: Set up UAV S as an airborne base station, and initialize the flight trajectory of S to provide downlink communication services to K users on the ground. At the same time, it senses the target J and offloads some of the sensed data to a ground base station. Considering the existence of a potential eavesdropper E, in order to ensure the security of information transmission, the UAV emits artificial noise to interfere with E and ensure the security of communication. Mathematical model construction module: With the goal of maximizing safe speed, an optimization mathematical model is constructed, taking the UAV computing frequency, the integrated waveform of communication, sensing and computing, and the UAV 3D flight trajectory as optimization variables, and considering communication performance, sensing performance, computing performance, and UAV energy consumption constraints. Mathematical Model Decoupling Module: For the optimized mathematical model, the original problem is decoupled into subproblems based on the block coordinate descent method, which only concern the UAV's calculation frequency, the integrated waveform of the synesthetic computation, the UAV's horizontal flight trajectory, and the UAV's vertical flight trajectory. For each non-convex subproblem, a convex approximation fitting method is used to convert it into a convex problem for solution. Specifically, when solving the non-convex problem concerning the integrated waveform of the synesthetic computation, a positive semidefinite relaxation method that discards the rank-one constraint is used for solution. Then, the obtained solution is randomized through Gaussian to obtain a solution that satisfies the rank-one constraint. Iterative solution module: Uses iterative algorithms to solve for the UAV's calculated frequency, integrated waveform of induction and computation, horizontal flight trajectory, and vertical flight trajectory.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the integrated waveform and UAV 3D trajectory optimization design method according to any one of claims 1 to 6.

Images