Multi-uav cooperative covert communication and resource allocation method
By optimizing the task offloading ratio, UAV computing resource allocation, and active interference power, and combining particle swarm optimization and block coordinate descent methods, the contradiction between covert communication and low-latency task processing in multi-UAV collaborative scenarios was resolved, achieving improvements in security, timeliness, and collaboration.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- INFORMATION & COMM CO OF STATE GRID XINJIANG ELECTRIC POWER CO LTD
- Filing Date
- 2026-05-08
- Publication Date
- 2026-07-31
AI Technical Summary
Existing technologies struggle to simultaneously meet the needs of covert communication and low-latency task processing in multi-drone collaborative scenarios, resulting in high system complexity and insufficient security and timeliness.
By jointly optimizing the task offloading ratio, the UAV computing resource allocation ratio, and the active interference power, and combining the particle swarm optimization algorithm and the block coordinate descent method, the three-dimensional deployment position of multiple UAVs is optimized, achieving synergistic optimization of communication security and computing efficiency.
It effectively reduces task processing latency while meeting covert communication constraints, improves system security, timeliness and collaboration, and enhances eavesdropping detection and suppression capabilities as well as task offloading flexibility.
Smart Images
Figure CN122496868A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of aerial computing technology for unmanned aerial vehicles (UAVs), and in particular to a method for covert communication and resource allocation among multiple UAVs. Background Technology
[0002] With the convergence of wireless communication, edge computing, and UAV platform technologies, UAVs, with their high mobility, flexible deployment, wide coverage, and excellent line-of-sight transmission capabilities, have gradually become important carriers for aerial information collection, aerial computing, and aerial network services. In applications such as smart inspection, emergency response, low-altitude sensing, and intelligent monitoring, utilizing UAVs for data collection, communication relay, and task processing has become a crucial technological direction in the field of UAV aerial computing.
[0003] In existing technologies, aerial edge computing typically involves UAVs carrying communication and computing modules. During flight, these modules establish wireless links with multiple target nodes to perform data reception, task offloading, computation processing, result transmission, resource allocation, and trajectory adjustment operations, thereby enhancing the system's real-time processing capabilities and service flexibility. Meanwhile, given the openness, strong broadcasting nature, and vulnerability to eavesdropping inherent in aerial wireless links, existing research has introduced physical layer security or covert communication mechanisms into UAV communication systems. These mechanisms utilize transmit power control, flight position optimization, artificial noise injection, and coordinated interference to improve the security and anti-detection capabilities of aerial communication.
[0004] However, existing technical solutions often focus on a single aspect, such as improving task processing performance or optimizing secure data transmission performance, while neglecting the coordinated design of the task offloading process and communication concealment. On the one hand, while introducing active interference can suppress eavesdropping, the increased interference power negatively impacts the communication rate of the task offloading link, leading to increased task processing latency. On the other hand, in multi-UAV collaborative scenarios, task matching, resource competition, coordinated interference, and location deployment among nodes are interdependent, significantly increasing system complexity. Under conditions of increased UAV numbers and expanded service range, existing solutions struggle to simultaneously meet the requirements for concealed communication and low-latency task processing, resulting in performance deficiencies in security, timeliness, and coordination.
[0005] Therefore, there is an urgent need in related technologies for a way to improve communication security and computational efficiency in scenarios involving multi-UAV collaborative active interference and security task offloading. Summary of the Invention
[0006] Therefore, it is necessary to provide a method for covert communication and resource allocation of multi-UAVs in collaborative active interference and security task offloading scenarios, which can improve communication security and computational efficiency in these scenarios.
[0007] Firstly, this application provides a method for multi-UAV cooperative covert communication and resource allocation. The method includes: Initialize a multi-drone secure edge computing scenario consisting of multiple edge computing drones, multiple ground sensors, and one malicious drone, and establish a ground-to-air communication model, a covert communication detection model, and a task processing latency model; Under the premise of satisfying covert communication constraints, with the goal of minimizing the total task processing latency, the task offloading ratio, the UAV computing resource allocation ratio, the active interference power, and the three-dimensional deployment position of multiple UAVs are jointly optimized. Among them, given the three-dimensional deployment position of multiple UAVs, the block coordinate descent method is used to alternately and iteratively solve the task offloading ratio, computing resource allocation ratio, and active interference power to obtain the optimal task offloading ratio, optimal computing resource allocation ratio, and optimal active interference power under the current deployment position. The particle swarm optimization algorithm is used to perform a global search for the three-dimensional deployment positions of multiple UAVs. At each particle position, the block coordinate descent method is used to evaluate the system performance. The three-dimensional deployment positions of multiple UAVs with the minimum total task processing latency are obtained, along with the corresponding task offloading ratio, computational resource allocation ratio, and active interference power.
[0008] Optionally, in one embodiment of this application, in the multi-UAV secure edge computing scenario, the edge computing UAV has computing resources and variable active interference power, the ground sensor has local computing resources, transmission power and data to be processed, and the malicious UAV is located in three-dimensional space and performs energy detection on the communication between the ground sensor and the edge computing UAV.
[0009] Optionally, in one embodiment of this application, the process of calculating the task unloading ratio includes: Given the active interference power of UAVs, the allocation ratio of computing resources, and the three-dimensional deployment locations of multiple UAVs, the local processing latency of ground sensors and the processing latency of UAVs are constructed as piecewise functions with respect to the task offloading ratio. The optimal task offloading ratio of each ground sensor is determined based on the condition that the piecewise function reaches a minimum value where the local processing latency and the UAV processing latency are equal.
[0010] Optionally, in one embodiment of this application, the process of calculating the resource allocation ratio includes: Given the active interference power of UAVs, the task offloading ratio, and the three-dimensional deployment locations of multiple UAVs, the part of the total task processing delay related to the computing resource allocation ratio is constructed as a convex optimization problem. Under the conditions of satisfying the total computing resource constraints of UAVs and the value constraints of each allocation ratio, the computing resource allocation ratio of each edge computing UAV to each ground sensor is solved using convex optimization tools.
[0011] Optionally, in one embodiment of this application, the process of solving the active interference power includes: Given the task unloading ratio, computational resource allocation ratio, and the three-dimensional deployment positions of multiple UAVs, the objective function related to active interference power in the total task processing delay is subjected to continuous convex approximation. The non-convex terms are expanded by first-order Taylor at the current iteration point to obtain an affine approximation function. The covert communication constraint is transformed into a convex constraint form with respect to active interference power. Within the upper and lower bounds of the active interference power constraint, the optimal active interference power of each UAV is solved using convex optimization tools.
[0012] Optionally, in one embodiment of this application, the process of alternately iterating using the block coordinate descent method includes: Initialize the 3D deployment locations of multiple drones and the maximum number of iterations; Update the task offload ratio given the active interference power and the computing resource allocation ratio; Update the computing resource allocation ratio given the active interference power and task offload ratio; Update the active interference power given the task offload ratio and computing resource allocation ratio; Repeat the above update steps until the preset iteration termination condition is met, and output the optimal task unloading ratio, computational resource allocation ratio and active interference power at the current deployment location.
[0013] Optionally, in one embodiment of this application, the step of using a particle swarm optimization algorithm to perform a global search for the three-dimensional deployment locations of multiple UAVs, and evaluating the system performance at each particle position using a block coordinate descent method, includes: Initialize the number of drones, the number of ground sensors, the location of ground sensors, the concealment constraint value, the particle swarm population size, and the maximum number of iterations; Randomly initialize the position and velocity of each particle in the population, and each particle's position corresponds to a set of three-dimensional deployment positions of multiple UAVs; For each particle, the block coordinate descent method is called to evaluate its corresponding particle performance; Update the individual optimal position and the global optimal position based on the evaluation results; Iteratively update particle velocity and position until the maximum number of iterations is reached, and output the multi-UAV 3D deployment position corresponding to the global optimal position, as well as the task unloading ratio, computational resource allocation ratio and active interference power obtained by the block coordinate descent method at that deployment position.
[0014] Optionally, in one embodiment of this application, the covert communication constraint is constructed based on the total detection error rate of the malicious drone, and a lower bound of the total detection error rate is established using relative entropy divergence, such that the relative entropy divergence is no greater than twice the square of a preset covertness threshold.
[0015] Optionally, in one embodiment of this application, the air-to-ground communication model includes a probability-weighted average path loss for line-of-sight and non-line-of-sight communication, and determines the wireless channel gain and communication rate from the ground sensor to the edge computing UAV based on the path loss.
[0016] Optionally, in one embodiment of this application, the total task processing latency is composed of the local calculation latency of the ground sensor and the task unloading and calculation latency of the UAV side. After the optimal task unloading ratio of the ground sensor is determined, the total task processing latency is simplified to the sum of the UAV side latency.
[0017] The aforementioned method for multi-UAV collaborative covert communication and resource allocation addresses scenarios involving active interference and secure task offloading by cooperating with multiple UAVs. By endowing UAVs with dual functions of communication and computation, the method enables them to inject artificial noise to suppress eavesdropping and improve communication covertness, while also participating in task offloading as distributed computing nodes. This achieves synergistic optimization of communication security and task processing. Compared to existing technologies, this invention overcomes the limitations of focusing solely on optimizing task processing performance or secure transmission performance. It effectively reduces task processing latency and improves overall system efficiency while meeting communication constraints. Furthermore, this invention proposes a hybrid optimization framework combining Particle Swarm Optimization (PSO) and Block Coordinate Descent (BCD), which can jointly solve for UAV collaborative interference, resource allocation, and task processing, thus better adapting to application environments with increasing UAV numbers and expanded service ranges. Through the collaborative cooperation among multiple UAVs, the system's ability to suppress eavesdropping detection is enhanced, as well as the flexibility and load-sharing capacity of task offloading and computation processing are improved. This minimizes task processing latency while meeting covert communication constraints, significantly improving the system's security, timeliness, and collaboration. Attached Figure Description
[0018] Figure 1 This is a flowchart illustrating a method for covert communication and resource allocation among multiple unmanned aerial vehicles (UAVs) in one embodiment. Figure 2 This is a schematic diagram illustrating an applicable scenario for a multi-UAV collaborative covert communication and resource allocation method in one embodiment; Figure 3 This is a schematic diagram of an iterative optimization algorithm based on block coordinate descent (BCD) in one embodiment; Figure 4 This is a schematic diagram of a multi-UAV deployment algorithm based on particle swarm optimization (PSO) in one embodiment; Figure 5 This is a convergence graph of a multi-UAV deployment algorithm based on particle swarm optimization (PSO) in one embodiment; Figure 6 This is a comparison chart of the total task processing latency under different numbers of drones in one embodiment; Figure 7 This is a comparison chart of communication latency under different covert communication constraints in one embodiment; Figure 8 This is a comparison chart of active interference power under different covert communication constraints in one embodiment; Figure 9 This is a performance analysis chart showing the task status, offload ratio, and processing latency of ground sensors in one embodiment. Figure 10 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation
[0019] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.
[0020] In one embodiment, such as Figure 1 As shown, a method for multi-UAV cooperative covert communication and resource allocation is provided, including the following steps: S101: Initialize a multi-drone secure edge computing scenario consisting of multiple edge computing drones, multiple ground sensors, and one malicious drone, and establish a ground-to-air communication model, a covert communication detection model, and a task processing latency model.
[0021] In one embodiment of this application, in the multi-UAV secure edge computing scenario, the edge computing UAV has computing resources and variable active interference power, the ground sensor has local computing resources, transmission power and data to be processed, and the malicious UAV is located in three-dimensional space and performs energy detection on the communication between the ground sensor and the edge computing UAV.
[0022] In this embodiment of the application, initialization is as follows: Figure 2 The multi-drone secure edge computing scenario shown includes multiple edge drones, multiple ground sensors, and a single malicious drone, Willie. The multi-edge computing drone set is represented as... All drones have the same computing resources and active variable interference power The location information of each drone can be represented as The set of ground sensors is represented as Ground sensors in various regions have computing resources Transmission power Data to be processed and location information The location information of the malicious drone Willie can be represented as: .
[0023] Establishing a ground-to-air (G2A) communication model: In one embodiment of this application, the ground-to-air communication model includes a probability-weighted average path loss for line-of-sight (LAS) and non-LAS communication, and determines the wireless channel gain and communication rate from the ground sensor to the edge computing UAV based on the path loss.
[0024] The possibilities of line-of-sight (LoS) and non-line-of-sight (NLoS) communication are discussed. The path losses for LoS and NLoS can be calculated using the following formulas:
[0025]
[0026] in, , It is the system carrier frequency. It is a drone With sensors distance, It's the speed of light. and These are the additional attenuation factors caused by LoS and NLoS connections, respectively.
[0027] The probability of a LoS connection is:
[0028] in, and These are environmentally relevant constants; The elevation angle is given. Therefore, the average path loss for communication between the UAV and the sensor is:
[0029] Get sensor To drones The wireless channel gain is:
[0030] sensor To drones The communication rate is:
[0031] in, It is the noise power of the receiver. It is the sensor's transmission power. It is the channel bandwidth.
[0032] Establish a covert communication detection model: Considering the malicious drone Willie, an energy detection method is used to determine whether sensor-drone communication exists by comparing the received signal with a preset threshold. The eavesdropper receives the signal. It can be represented as:
[0033] in, and These indicate whether the ground sensor is sending communication signals.
[0034] In one embodiment of this application, the covert communication constraint is constructed based on the total detection error rate of the malicious drone, and a lower bound on the total detection error rate is established using relative entropy divergence, such that the relative entropy divergence is no greater than twice the square of a preset covertness threshold.
[0035] Given that Willie has equal prior probabilities, we use the total detection error rate to measure Willie's detection performance:
[0036] in, This indicates the false alarm probability, suggesting that Willie misjudged the sensor's signal transmission, even though the sensor wasn't actually communicating. Similarly, The false negative probability is defined, which represents the situation where Willie fails to detect the transmitted signal despite the sensor communicating. and It is a binary decision used to infer the presence of sensor transmission. To minimize... The likelihood test (LRT) was used for optimal testing, which can be expressed as:
[0037] in, This indicates the number of observation samples used by Willie during the detection. and They are and The likelihood function under the given conditions. and They can be defined as follows:
[0038]
[0039] In summary, the total error coefficient contains an incomplete function, making it difficult to solve directly. Therefore, the relative entropy (KL) divergence is used. As its lower limit:
[0040] In the formula, Indicates from arrive The KL divergence can be defined as:
[0041]
[0042] Furthermore, constraints on the requirements for covert communication can be derived:
[0043] in, The system's concealment requirements were defined. .
[0044] Establish a task processing latency model: In one embodiment of this application, the total task processing latency is composed of the local calculation latency of the ground sensor and the task unloading and calculation latency of the UAV side. After the optimal task unloading ratio of the ground sensor is determined, the total task processing latency is simplified to the sum of the UAV side latency.
[0045] Consider a scenario where a drone assists in processing data from ground sensors. The sensors offload a portion of their local data to the drone for processing. Two parameters need to be defined here: one is the proportion of data offloaded. The other is the computing resources required to process unit data. Then, the local task computation delay is obtained:
[0046] Task unloading latency and computation latency can be calculated using the following formulas respectively: ,
[0047] in, It is a drone Assigned to ground users Computing resources.
[0048] In one embodiment of this application, consider a ground sensor randomly distributed in A multi-drone scenario in the region. The size of the data to be processed generated by each grounding sensor is within... Between these points, the unit data computing power requirement is fixed at All sensors were configured with the same computing and transmission power, and were set to... and The drones are all isomorphic, each possessing... With its computing power, it can emit active jamming signals, with jamming power at... Optional within the range. For communication, the channel bandwidth is set to... The line-of-sight (LoS) attenuation is set to Non-line-of-sight (NLoS) attenuation is set to This embodiment simulates the position, calculation, and communication parameters of the UAV and ground sensors, which is more in line with engineering practice, and the simulation is based on the embodiment.
[0049] S102: Under the premise of satisfying the covert communication constraints, with the goal of minimizing the total task processing latency, jointly optimize the task offloading ratio, the UAV computing resource allocation ratio, the active interference power, and the three-dimensional deployment position of multiple UAVs; wherein, given the three-dimensional deployment position of multiple UAVs, the block coordinate descent method is used to alternately and iteratively solve the task offloading ratio, the computing resource allocation ratio, and the active interference power to obtain the optimal task offloading ratio, the optimal computing resource allocation ratio, and the optimal active interference power under the current deployment position.
[0050] In one embodiment of this application, the process of calculating the task unloading ratio includes: Given the active interference power of UAVs, the allocation ratio of computing resources, and the three-dimensional deployment locations of multiple UAVs, the local processing latency of ground sensors and the processing latency of UAVs are constructed as piecewise functions with respect to the task offloading ratio. The optimal task offloading ratio of each ground sensor is determined based on the condition that the piecewise function reaches a minimum value where the local processing latency and the UAV processing latency are equal.
[0051] In one embodiment of this application, given the active jamming power of the UAV Calculate the resource allocation ratio and 3D deployment location of drones In this case, the proportion of mathematical analysis tasks unloaded The issue of minimizing the total latency of related task processing. (Compared to the task unloading ratio) The problem of minimizing the total latency of related task processing can be formalized as follows:
[0052] For ground sensors The local computation delay can be expressed as ,in The latency of edge computing for drones can be expressed as: ,in . and All are constants, and the local calculation delay varies. Monotonically decreases, while the drone's computational latency increases. It increases monotonically. Therefore, the entire function is a piecewise convex function, and the critical point where the two are equal is the minimum point of the function. Each ground sensor can obtain its optimal task offload ratio:
[0053] Ultimately, the optimal task offloading ratio can be obtained, and given a fixed task offloading ratio, the segmented optimization objective can be simplified to UAV latency optimization:
[0054] In one embodiment of this application, the process of calculating the resource allocation ratio includes: Given the active interference power of UAVs, the task offloading ratio, and the three-dimensional deployment locations of multiple UAVs, the part of the total task processing delay related to the computing resource allocation ratio is constructed as a convex optimization problem. Under the conditions of satisfying the total computing resource constraints of UAVs and the value constraints of each allocation ratio, the computing resource allocation ratio of each edge computing UAV to each ground sensor is solved using convex optimization tools.
[0055] In one embodiment of this application, given the active jamming power of the UAV Task uninstallation ratio and 3D deployment location of drones In this case, mathematical analysis calculates the resource allocation ratio. The problem involves minimizing the total latency of related task processing, in relation to the allocation ratio of computing resources. The problem of minimizing the total latency of related task processing can be formalized as follows:
[0056] Simplifying the relevant constant terms of the objective function yields:
[0057] in, Since the drone's location, transmission power, and interference power are all determined, the transmission rate is not affected by the allocation of computing resources. It is a constant; Since the offloading ratio of ground sensor tasks in various regions has been determined, It is a constant. Based on this, each term... All are about Convex functions can be solved using the existing Convex Optimization Toolbox (CVX).
[0058] In one embodiment of this application, the process of solving for the active interference power includes: Given the task unloading ratio, computational resource allocation ratio, and the three-dimensional deployment positions of multiple UAVs, the objective function related to active interference power in the total task processing delay is subjected to continuous convex approximation. The non-convex terms are expanded by first-order Taylor at the current iteration point to obtain an affine approximation function. The covert communication constraint is transformed into a convex constraint form with respect to active interference power. Within the upper and lower bounds of the active interference power constraint, the optimal active interference power of each UAV is solved using convex optimization tools.
[0059] In one embodiment of this application, given a computing resource allocation ratio Task uninstallation ratio and 3D deployment location of drones In the case of mathematical analysis of the active interference power of UAVs The problem of minimizing the total processing latency of related tasks and its covert communication constraints. The active interference power of UAVs. The related problem of minimizing the total processing latency of tasks and its covert communication constraints can be formalized as follows:
[0060]
[0061] in, , and .Notice, , , ,and . , and All are constants.
[0062] Based on the objective function, define the function.
[0063] Differentiation yields
[0064] Using the Continuous Convex Approximation (SCA) technique, in Perform a first-order Taylor expansion nearby:
[0065] It can be written as ,in, For scalar variables, This is a constant term. Therefore, we can obtain... is a standard affine function that can be solved directly using CVX.
[0066] Constraints This stems from KL divergence (or false detection probability) control:
[0067] Based on this, the function is defined as follows:
[0068] in, , , ,and .
[0069] right Taking the first and second derivatives of the function, we get:
[0070]
[0071] therefore, The function behaves as a monotonically decreasing convex function. Furthermore, since... It is about Standard affine function, It can be solved directly using CVX.
[0072] Constraints The range of active interference power for the UAV is limited, which is a standard bounded constraint and is itself a convex set; CVX naturally supports this type of constraint. Through the mathematical analysis of the objective function and constraints above, it is proven that the active interference power of the UAV can be directly obtained using CVX tools. The approximate optimal solution.
[0073] In one embodiment of this application, the process of alternately iterating and solving using the block coordinate descent method includes: S201: Initialize the 3D deployment locations of multiple UAVs and the maximum number of iterations.
[0074] S202: Update the task offload ratio given the active interference power and the computing resource allocation ratio.
[0075] S203: Update the computational resource allocation ratio given the active interference power and task offloading ratio.
[0076] S204: Update active interference power given the task offloading ratio and computational resource allocation ratio.
[0077] S205: Repeat the above update steps until the preset iteration termination condition is met, and output the optimal task unloading ratio, computational resource allocation ratio and active interference power at the current deployment location.
[0078] In one embodiment of this application, such as Figure 3As shown, the optimal UAV active interference power under the current state is solved iteratively using a block coordinate descent (BCD) method. Calculate the resource allocation ratio and task uninstallation ratio First, initialize the 3D deployment location of the drone. and maximum number of iterations Subsequently, given the active jamming power of the UAV... and computing resource allocation ratio Under these circumstances, the current optimal task unloading ratio is obtained through updating. Subsequently, given the active jamming power of the UAV... and task uninstallation ratio In this case, the current optimal computing resource allocation ratio is updated. Then, given the proportion of computing resources allocated... and task uninstallation ratio Under these circumstances, the current optimal active interference power of the UAV is obtained through updating. Repeat the above update steps until the preset iteration termination condition is met to obtain the final UAV active interference power. Calculate the resource allocation ratio and task uninstallation ratio .
[0079] S103: The particle swarm optimization algorithm is used to perform a global search for the three-dimensional deployment positions of multiple UAVs, and the block coordinate descent method is used to evaluate the system performance at each particle position. The three-dimensional deployment positions of multiple UAVs with the minimum total task processing latency are obtained, along with the corresponding task offloading ratio, computational resource allocation ratio, and active interference power.
[0080] In one embodiment of this application, the step of using a particle swarm optimization algorithm to perform a global search for the three-dimensional deployment locations of multiple UAVs, and evaluating the system performance at each particle position using a block coordinate descent method, includes: S301: Initialize the number of drones, the number of ground sensors, the location of ground sensors, the concealment constraint value, the particle swarm population size, and the maximum number of iterations.
[0081] S302: Randomly initialize the position and velocity of each particle in the population. Each particle's position corresponds to a set of three-dimensional deployment positions of multiple UAVs.
[0082] S303: For each particle, call the block coordinate descent method to evaluate its corresponding particle performance.
[0083] S304: Update the individual optimal position and the global optimal position based on the evaluation results.
[0084] S305: Iteratively update particle velocity and position until the maximum number of iterations is reached, and output the multi-UAV 3D deployment position corresponding to the global optimal position, as well as the task unloading ratio, computational resource allocation ratio and active interference power obtained by the block coordinate descent method at that deployment position.
[0085] In one embodiment of this application, such as Figure 4 As shown, the particle swarm optimization (PSO) algorithm is used to optimize the 3D deployment locations of multiple UAVs. And obtain the optimal UAV launch power. Calculate the resource allocation ratio Task uninstallation ratio and 3D deployment location of drones To minimize the total task processing latency, the first step is to initialize the number of drones. Number of ground sensors Ground sensor location Hidden constraints PSO population size Maximum number of PSO iterations Then, based on the PSO population size, the particle positions within the population are initialized. and speed Each particle position corresponds to a set of three-dimensional deployment positions for multiple UAVs. For each particle, the block coordinate descent method is used to evaluate its corresponding particle performance, and the individual optimal value is set based on the evaluation results. To obtain the globally optimal result Then, update the internal particle positions. and speed .
[0086]
[0087]
[0088] in, and It is the PSO acceleration factor. and It is a random number in the interval [0,1]. It is the inertial weight, which usually decreases linearly with the evolutionary process.
[0089] The updated internal particle performance is evaluated by calling the block coordinate descent method. If the updated internal particle performance is better than the current individual optimal result, that is... Update the individual optimal result Iterate through the PSO population, updating particle velocity and position until the population size is traversed. The updated global optimum of the population is obtained. If the updated population-optimal result is better than the global-optimal result, that is... Update the global optimal result The system iterates through the loop to update the globally optimal result until the maximum number of iterations is reached. Output the final drone transmit power Calculate the resource allocation ratio Task uninstallation ratio and 3D deployment location of drones .
[0090] In one embodiment of this application, according to Figure 5 As shown, Figure 5 Convergence performance graphs for the proposed Particle Swarm Optimization (PSO) and Genetic Algorithm (GA) schemes are presented. It can be observed that the PSO scheme exhibits better convergence performance than the GA scheme. In PSO, particles rapidly approach the optimal region through a dual approach of individual and global optimization. While GA relies on random selection, crossover, and mutation, and possesses global search capabilities, its local search efficiency is low, and its initial iterations are typically slow.
[0091] In one embodiment of this application, according to Figure 6 As shown, Figure 6 The results comparing task processing latency under different schemes are presented. The figures show that the particle swarm optimization (PSO) scheme proposed in this application consistently achieves the lowest overall latency regardless of the number of drones. When the number of drones is small, the PSO-based scheme, the genetic algorithm (GA)-based scheme, and the block coordinate descent (BCD) method all perform well and have similar overall latency. This is because the three schemes share similar optimization principles, each embedding a BCD iterative optimization module to iteratively optimize drone transmit power, computational resource allocation ratio, and task offloading ratio. However, as the number of drones increases, the PSO-based scheme demonstrates better overall latency optimization, resulting in improved resource utilization.
[0092] In one embodiment of this application, according to Figure 7 As shown, Figure 7 The constraints of covert communication were evaluated. The impact on system communication latency. Since communication latency and overall latency are not on the same order of magnitude... Figure 6 The impact of covert communication constraints was specifically analyzed only at the system communication latency level. It can be observed that as the covert communication constraints are relaxed, the communication latency gradually decreases, such as... and This indicates that covert communication constraints do indeed have a negative impact on communication rate. It should be noted that the optimal communication rate scheme performs best in the results, because its optimization objective focuses on the communication process itself, while neglecting the allocation of computational resources between the ground sensors and the UAV.
[0093] In one embodiment of this application, according to Figure 8 As shown, Figure 8 The constraints of covert communication were evaluated. The impact on the active jamming power of UAVs. As can be seen from the figure, as the constraints of covert communication are relaxed, the jamming power gradually decreases, such as... and This indicates that the designed dual-role UAV can indeed exert its active jamming effect and suppress eavesdropping. It should be noted that the optimal communication rate scheme performs best in the results. This is because the optimization objective of this scheme focuses on the communication process, while neglecting the allocation of computing resources between the ground sensors and the UAV.
[0094] In one embodiment of this application, according to Figure 9 As shown, Figure 9 Numerical analysis was performed on the performance of the ground sensors in the scheme. As can be seen from the figure, this application has a good effect on ensuring the fairness of services provided by various ground sensors. Each ground sensor can offload data relatively evenly, allocate matching computing resources, improve overall resource efficiency, optimize data flow and task allocation between edge nodes, and avoid local network congestion.
[0095] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.
[0096] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 10As shown, the computer device includes a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When executed by the processor, the computer program implements a method for multi-UAV cooperative covert communication and resource allocation. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.
[0097] Those skilled in the art will understand that Figure 10 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.
[0098] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.
[0099] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.
[0100] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.
[0101] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.
[0102] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.
[0103] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0104] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.
Claims
1. A method for multi-UAV cooperative covert communication and resource allocation, characterized in that, The method includes: Initialize a multi-drone secure edge computing scenario consisting of multiple edge computing drones, multiple ground sensors, and one malicious drone, and establish a ground-to-air communication model, a covert communication detection model, and a task processing latency model; Under the premise of satisfying covert communication constraints, with the goal of minimizing the total task processing latency, the task offloading ratio, the UAV computing resource allocation ratio, the active interference power, and the three-dimensional deployment position of multiple UAVs are jointly optimized. Among them, given the three-dimensional deployment position of multiple UAVs, the block coordinate descent method is used to alternately and iteratively solve the task offloading ratio, computing resource allocation ratio, and active interference power to obtain the optimal task offloading ratio, optimal computing resource allocation ratio, and optimal active interference power under the current deployment position. The particle swarm optimization algorithm is used to perform a global search for the three-dimensional deployment positions of multiple UAVs. At each particle position, the block coordinate descent method is used to evaluate the system performance. The three-dimensional deployment positions of multiple UAVs with the minimum total task processing latency are obtained, along with the corresponding task offloading ratio, computational resource allocation ratio, and active interference power.
2. The method for multi-UAV cooperative covert communication and resource allocation according to claim 1, characterized in that, In the multi-drone secure edge computing scenario, the edge computing drone has computing resources and variable active interference power, the ground sensor has local computing resources, transmission power and data to be processed, and the malicious drone is located in three-dimensional space and performs energy detection on the communication between the ground sensor and the edge computing drone.
3. The method for multi-UAV cooperative covert communication and resource allocation according to claim 1, characterized in that, The process of calculating the task unloading ratio includes: Given the active interference power of UAVs, the allocation ratio of computing resources, and the three-dimensional deployment locations of multiple UAVs, the local processing latency of ground sensors and the processing latency of UAVs are constructed as piecewise functions with respect to the task offloading ratio. The optimal task offloading ratio of each ground sensor is determined based on the condition that the piecewise function reaches a minimum value where the local processing latency and the UAV processing latency are equal.
4. The method for multi-UAV cooperative covert communication and resource allocation according to claim 1, characterized in that, The process of calculating the resource allocation ratio includes: Given the active interference power of UAVs, the task offloading ratio, and the three-dimensional deployment locations of multiple UAVs, the part of the total task processing delay related to the computing resource allocation ratio is constructed as a convex optimization problem. Under the conditions of satisfying the total computing resource constraints of UAVs and the value constraints of each allocation ratio, the computing resource allocation ratio of each edge computing UAV to each ground sensor is solved using convex optimization tools.
5. A method for multi-UAV cooperative covert communication and resource allocation according to claim 1, characterized in that, The process of solving for the active interference power includes: Given the task unloading ratio, computational resource allocation ratio, and the three-dimensional deployment positions of multiple UAVs, the objective function related to active interference power in the total task processing delay is subjected to continuous convex approximation. The non-convex terms are expanded by first-order Taylor at the current iteration point to obtain an affine approximation function. The covert communication constraint is transformed into a convex constraint form with respect to active interference power. Within the upper and lower bounds of the active interference power constraint, the optimal active interference power of each UAV is solved using convex optimization tools.
6. A method for multi-UAV cooperative covert communication and resource allocation according to claim 1, characterized in that, The process of solving the problem using the block coordinate descent method with alternating iterative steps includes: Initialize the 3D deployment locations of multiple drones and the maximum number of iterations; Update the task offload ratio given the active interference power and the computing resource allocation ratio; Update the computing resource allocation ratio given the active interference power and task offload ratio; Update the active interference power given the task offload ratio and computing resource allocation ratio; Repeat the above update steps until the preset iteration termination condition is met, and output the optimal task unloading ratio, computational resource allocation ratio and active interference power at the current deployment location.
7. A method for multi-UAV cooperative covert communication and resource allocation according to claim 1, characterized in that, The step of using particle swarm optimization to perform a global search for the three-dimensional deployment positions of multiple UAVs, and evaluating the system performance at each particle position using a block coordinate descent method, includes: Initialize the number of drones, the number of ground sensors, the location of ground sensors, the concealment constraint value, the particle swarm population size, and the maximum number of iterations; Randomly initialize the position and velocity of each particle in the population, and each particle's position corresponds to a set of three-dimensional deployment positions of multiple UAVs; For each particle, the block coordinate descent method is called to evaluate its corresponding particle performance; Update the individual optimal position and the global optimal position based on the evaluation results; Iteratively update particle velocity and position until the maximum number of iterations is reached, and output the multi-UAV 3D deployment position corresponding to the global optimal position, as well as the task unloading ratio, computational resource allocation ratio and active interference power obtained by the block coordinate descent method at that deployment position.
8. A method for multi-UAV cooperative covert communication and resource allocation according to any one of claims 1 to 7, characterized in that, The covert communication constraint is constructed based on the total detection error rate of malicious drones, and a lower bound for the total detection error rate is established using relative entropy divergence, so that the relative entropy divergence is no greater than twice the square of the preset covertness threshold.
9. A method for multi-UAV cooperative covert communication and resource allocation according to any one of claims 1 to 7, characterized in that, The ground-to-air communication model includes a probability-weighted average path loss for line-of-sight and non-line-of-sight communication, and determines the wireless channel gain and communication rate from the ground sensor to the edge computing UAV based on this path loss.
10. A method for multi-UAV cooperative covert communication and resource allocation according to any one of claims 1 to 7, characterized in that, The total task processing latency is composed of the local calculation latency of the ground sensor and the task unloading and calculation latency of the UAV side. Once the optimal task unloading ratio of the ground sensor is determined, the total task processing latency is simplified to the sum of the UAV side latency.