A new freshness control method and system for unmanned aerial vehicle (uav) car networking based on double age of information (aoi)

Abstract

The application discloses a new freshness control method and system for unmanned aerial vehicle (UAV) Internet of Vehicles (IoV) based on double age of information (AoI), and relates to the technical field of wireless communication, IoV, space-air-ground integrated network, and intelligent optimization and reinforcement learning. L The frame observation vector is constructed by the latest L The time sequence state composed of the latest The policy network and the value network enhanced by the GRU are iteratively updated by using the proximal policy optimization (PPO) algorithm, and a joint control policy for perception and communication suitable for a dynamic UAV IoV scene is obtained, so as to solve the problem that it is difficult to simultaneously consider state perception, secure transmission, information freshness, UAV maneuvering control and long-term dynamic optimization in the UAV IoV state updating scene.

Description

Technical Field

[0001] This invention belongs to the interdisciplinary fields of wireless communication, vehicle networking, air-space-ground integrated networks, intelligent optimization and reinforcement learning, and specifically relates to a method and system for controlling the confidentiality and freshness of UAV vehicle networking based on dual AoI. Background Technology

[0002] With the rapid development of intelligent transportation systems, autonomous driving, and low-altitude intelligent network technologies, the Internet of Vehicles (IoV), as a crucial technological support for vehicle state perception, collaborative decision-making, and safety control, has become a key component of the next-generation intelligent transportation system. In IoV scenarios, vehicles, roadside infrastructure, aerial platforms, and cloud-edge control nodes need to continuously exchange various information such as location, speed, road conditions, and task control to support applications like collaborative driving, traffic scheduling, risk warning, and emergency response. However, in these state-updating services, system performance depends not only on successful data transmission but also on the freshness of the information received at the receiving end. Traditional evaluation methods, centered on throughput, latency, and bit error rate, are insufficient to accurately depict the essential real-time requirements of such services. Therefore, Age of Information (AoI) has gradually become an important indicator for measuring the timeliness performance of state-updating systems. On the other hand, to adapt to the flexible networking needs of hotspot road sections, emergency areas, and scenarios with insufficient ground coverage, unmanned aerial vehicles (UAVs) are gradually being introduced into IoV systems as dynamically mobile aerial service nodes, providing on-demand perception, downlink transmission, and aerial collaborative control capabilities for ground vehicles. Based on this, integrated sensing and communication (ISAC) technology enables the same aerial platform to simultaneously undertake environmental perception and data transmission tasks, thus providing a unified technical framework for UAV-assisted IoV. However, as system scale and task complexity continue to increase, traditional communication systems are gradually revealing problems such as limited multi-user interference management capabilities, insufficient adaptability to secure transmission, and weak long-term dynamic control capabilities, making it difficult to meet the higher requirements for "coordinated optimization of perception, communication, security, and freshness" in complex vehicle-to-everything (V2X) scenarios.

[0003] To address the aforementioned issues, Rate-Splitting Multiple Access (RSMA) technology has received widespread attention in recent years. RSMA divides user messages into public and private streams and combines this with interference cancellation mechanisms to enable multiple users to share spectrum resources, achieving a balance between spectral efficiency, interference management capabilities, and transmission robustness under heterogeneous channel conditions. Meanwhile, research on Physical Layer Security (PLS) for IoV (Internet of Vehicles) scenarios has also continued to develop. PLS utilizes mechanisms such as channel differentiation, beamforming, and resource allocation to improve the security margin between legitimate and eavesdropping links without relying on higher-level encryption, thereby enhancing the system's anti-eavesdropping capabilities in open wireless environments. Therefore, combining RSMA, PLS, UAV mobility control, and ISAC mechanisms to form a joint optimization scheme for state update services is considered an effective technical path to simultaneously improve the spectral efficiency, confidentiality, and freshness control capabilities of IoV systems. Furthermore, since there is a clear temporal coupling relationship between UAV trajectory, perception success rate, downlink secure transmission success rate and energy consumption in this type of system, relying solely on static optimization methods is no longer suitable for highly dynamic scenarios, and long-term sequential control methods based on deep reinforcement learning have gradually become a research hotspot.

[0004] Existing research on UAV-assisted IoV, ISAC, and RSMA joint control systems primarily focuses on improving system performance in areas such as information freshness, perception performance, spectrum utilization, and secure transmission efficiency. Existing research generally agrees that introducing a UAV platform into an IoV system not only enhances flexible coverage in complex traffic scenarios but also enables simultaneous state perception and information transmission via the ISAC mechanism, providing technical support for building highly dynamic and real-time intelligent transportation communication systems. Building upon this foundation, to further improve system performance, existing research generally focuses on optimization modeling around trajectory design, power allocation, beamforming, and learning-based control strategies. Joint resource allocation schemes among UAVs, ground vehicles, and auxiliary nodes have been designed for different system objectives, resulting in significant improvements in AoI, throughput, perception performance, and energy efficiency.For example, the paper "Proximal Policy Optimization With Adaptive Weight Adjustment for Airborne Integrated Sensing and Communication" establishes a reinforcement learning control framework for airborne ISAC scenarios, introduces AoI into the optimization objective, and uses Proximal Policy Optimization (PPO) and adaptive weight adjustment mechanisms to jointly optimize the control strategy of the airborne platform, thereby improving the system's convergence performance and novelty. The paper "Trajectory Design and Resource Allocation for RSMA-UAV Enabled Integrated Sensing and Communication Networks" focuses on the combined RSMA and UAV-ISAC scenario, establishing a joint trajectory design and resource allocation model, and improving the system's integrated communication and sensing performance by optimizing UAV maneuvering and multi-user transmission resource configuration. The paper "Research on Integrated Sensing, Communication Resource Allocation, and Digital Twin Placement Based on Digital Twin in IoV" focuses on the digital twin IoV scenario, optimizing the modeling of sensing, communication resource configuration, and digital twin deployment, thereby improving resource utilization efficiency and state mapping capabilities in the IoV system. The paper "Age of Information and Energy" "Efficiency Optimization in RIS-Assisted Vehicular Edge Computing via Hierarchical Deep Reinforcement Learning" introduces a hierarchical deep reinforcement learning framework in the vehicle-to-everything (V2X) edge computing scenario to jointly optimize AoI (Aspect-Oriented Intelligence) and energy efficiency, thereby improving the system's long-term scheduling capability in dynamic environments.

[0005] Meanwhile, significant progress has been made in research on the combination of RSMA and PLS. Existing research generally agrees that in multi-user heterogeneous wireless networks, incorporating public and private flow structures into secure transmission design not only enhances the link's resistance to eavesdropping but also provides more flexible resource adjustment freedom in complex interference environments. Building on this foundation, many studies have further developed modeling and analysis around beamforming, rate allocation, and dynamic control strategies. For example, the paper "RSMA-Enhanced Physical Layer Security for ISAC Systems" addresses the secure transmission problem in ISAC systems by establishing an RSMA-enhanced physical layer security optimization model and improving system security performance through joint design of beam and rate allocation. The paper "Dynamic Resource Management for RSMA-Enabled UAV Networks With Mobile Users" studies the online resource management problem in RSMA-enabled UAV networks for mobile user scenarios, improving communication efficiency in dynamic environments through joint control of resource allocation strategies. The paper "Decoupled AssociationWith Rate Splitting Multiple Access in UAV-Assisted Cellular Networks Using Multi-Agent Deep Reinforcement Learning" further introduces multi-agent deep reinforcement learning into UAV-assisted RSMA cellular networks, co-optimizing association and transmission control to enhance the system's adaptive decision-making capabilities in complex dynamic scenarios. Overall, existing research has verified the application value of RSMA, UAV control, and learning-based optimization in complex wireless scenarios from different perspectives, providing important references for building efficient and intelligent IoV communication control systems.

[0006] Although existing research has extensively explored the application of UAV-assisted IoV, ISAC, and RSMA technologies in state update and secure transmission scenarios, and proposed various solutions such as AoI optimization, throughput improvement, perception enhancement, and secure transmission improvement, certain limitations still exist. Specifically, most existing works focus on joint optimization around a single or partially coupled objective, failing to fully consider the essential difference between "whether the sensing side successfully generates the latest state information" and "whether the communication side successfully and securely delivers the state information." For state update services, perception failure means that the new state has not been generated at all, while secure transmission failure means that although the new state has been generated, the vehicle-side information has not been effectively updated. The impact mechanisms of these two types of events on system freshness are different, but existing research usually does not explicitly distinguish and model them. In addition, some studies only optimize under ideal channel conditions, static threshold settings, or fixed control parameters, failing to fully consider the joint time-varying impact of vehicle position changes, eavesdropper trajectory changes, and UAV maneuvering behavior on the sensing link and secure link in high-mobility IoV scenarios. Furthermore, while existing studies have employed reinforcement learning or online optimization methods, they have mostly focused on single-level tasks such as trajectory control, power scheduling, or rate optimization. They lack a unified long-term decision design for common flow allocation, public / private beam adjustment, energy constraints, and dual freshness states, making it difficult to simultaneously achieve perception success, secure communication success, and long-term AoI optimality in complex dynamic environments.

[0007] In summary, while existing research on UAV-assisted IoV, ISAC, and RSMA joint control systems has made some progress in terms of spectral efficiency, sensing performance, security rate, and dynamic resource scheduling, it still has significant shortcomings in the "UAV vehicle-to-everything (AoI) state update scenario with mobile eavesdroppers." Especially in state update operations, the system not only requires information to be transmitted, but also that the information be "perceived, delivered securely, and updated promptly." Existing optimization strategies struggle to simultaneously characterize and optimize two different AoI processes: sensing refresh freshness and communication refresh freshness. Therefore, for IoV scenarios with mobile eavesdroppers, parallel sensing and communication, and limited UAV trajectory and energy, further research is urgently needed on joint dynamic optimization methods for dual AoI, RSMA secure transmission, and phased learning control. This would allow for optimal freshness performance of UAV vehicle-to-everything (VoI) systems in complex dynamic environments, meeting the urgent needs of future intelligent transportation systems for efficient, reliable, and secure state updates. Summary of the Invention

[0008] To address the shortcomings of existing technologies, this invention provides a method and system for controlling the confidentiality and freshness of UAV-vehicle networking based on dual AoI, in order to solve the problem that it is difficult to simultaneously consider state perception, secure transmission, information freshness, UAV maneuver control, and long-term dynamic optimization in UAV-vehicle networking status update scenarios.

[0009] The technical solution of this invention is as follows: On the one hand, this invention provides a method for controlling the confidentiality and freshness of unmanned aerial vehicle (UAV) vehicle-to-everything (V2X) networks based on dual AoI (Autonomous Internet of Things) technology, comprising the following steps: A drone-assisted security ISAC system is constructed, and an initial scenario model and control strategy constraints for the drone-assisted security ISAC system are established, and initialization parameters are set. The control strategy constraints include flight constraints for legitimate drones, flight trajectory constraints for mobile eavesdropping drones, sensing power constraints, communication power constraints, common stream transmission constraints, and remaining energy constraints. Recent acquisition of drone-assisted confidentiality ISAC system Frame observation vectors, and construct from the nearest Temporal state composed of frame observation vectors ; A gated cyclic unit-enhanced policy network and value network are constructed. Based on the initial scenario model, control policy constraints, and initialization parameters of the UAV-assisted confidential ISAC system, the GRU-enhanced policy network and value network are iteratively updated using the near-end policy optimization (PPO) algorithm to obtain a perception and communication joint control strategy suitable for dynamic UAV-vehicle network scenarios. The gated cyclic unit-enhanced policy network is used to adjust the policy network according to the time-series state. Output the original continuous action vector The value network enhanced by the gated loop unit is used to determine the time-series state. Output the current state value .

[0010] Furthermore, the specific steps for constructing the UAV-assisted confidentiality ISAC system, establishing the initial scenario model and control strategy constraints of the UAV-assisted confidentiality ISAC system, and setting initialization parameters include: A1: Construct a drone-assisted security ISAC system for vehicle-to-everything (V2X) status update scenarios, including a legitimate drone, a mobile eavesdropping drone, and... There are 10 ground vehicles, of which the legal drones are equipped with M antennas, while all ground vehicles and mobile eavesdropping drones are equipped with a single antenna. The set of ground vehicles is represented as: (1); in, Indicates a collection of ground vehicles; A2: Discretize the task cycle of the UAV-assisted security ISAC system into... In each time slot, establish the initial scenario model and control strategy constraints of the UAV-assisted confidential ISAC system; The length of the time slot is Each time slot is divided into a first half perception phase and a second half communication phase. In the first half perception phase, a legitimate UAV performs state perception on vehicles on all ground, completing the perception execution. In the second half communication phase, a legitimate UAV uses rate split multiple access to perform confidential state update communication on vehicles on all ground, completing the communication execution. The initial scenario model of the UAV-assisted confidentiality ISAC system includes the following: A legal drone flies at a fixed altitude, and its first... Position vector of each time slot : (2); in, Indicates the first A time-slot-based legal UAV in a horizontal two-dimensional coordinate system Position components along the axial direction; Indicates the first A legal UAV in a time slot travels along a horizontal two-dimensional coordinate system. Position components along the axial direction; Indicates the time slot index; Indicates the flight altitude of a legal drone; Indicates transpose; the horizontal two-dimensional coordinate system, The axial direction is defined as the direction along which the road extends. The axial direction is defined as perpendicular to the direction of road extension; A legal drone flies at a fixed altitude, and its first... The position vector update relationship for each time slot is as follows: (3); (4); in, Indicates that legal drones are in the Position vector of each time slot Indicates that legal drones are in the Position increment within a time slot; Indicates that legal drones are in the Flight speed per time slot Indicates that legal drones are in the The movement direction angle of each time slot; A legal drone flies at a fixed altitude, and its first... The position vector update relationship for each time slot is as follows: (3); (4); in, Indicates that legal drones are in the Position vector of each time slot Indicates that legal drones are in the Position increment within a time slot; Indicates that legal drones are in the Flight speed per time slot Indicates that legal drones are in the The movement direction angle of each time slot; The mobile eavesdropping drone flies along a circular trajectory at a fixed altitude, and its trajectory is on the... Position vector of each time slot Represented as: (5); in, Indicates the first A mobile eavesdropping drone in a time slot travels along a horizontal two-dimensional coordinate system. Position components in the axial direction Indicates the first A mobile eavesdropping drone in a time slot travels along a horizontal two-dimensional coordinate system. Position components in the axial direction The flight altitude of the mobile eavesdropping drone; No. The ground vehicle in Position vector of each time slot Represented as: (6); No. The ground vehicle in The position vector update relationship for each time slot is represented as follows: (7); in, Indicates the first The lane where each ground vehicle is located is in a horizontal two-dimensional coordinate system. Position components in the axial direction Indicates the first The first time slot The lane where each ground vehicle is located is along the horizontal two-dimensional coordinate system. Position components in the axial direction Indicates the first The ground vehicle in The speed of travel in each time slot; The flight constraints for the legal drones are as follows: (8); (9); in, This indicates the maximum permissible flight speed for a legal drone during flight. Indicates the feasible service area for legal drones; The flight trajectory constraints of the mobile eavesdropping drone are as follows: (10); (11); (12); in, and This indicates the center coordinates of the circular trajectory of the mobile eavesdropping drone. This represents the orbital radius of the circular trajectory of a mobile eavesdropping drone. This indicates the initial phase of the mobile eavesdropping drone. This indicates the period of the mobile eavesdropping drone's flight along a circular trajectory. This represents the angular velocity of the mobile eavesdropping drone as it flies along a circular trajectory. The sensing power constraint is: (13); in, Indicates the first The first time slot faces the first The sensing beam vector of a ground vehicle This indicates the transmit power corresponding to the sensing beam vector. This indicates the maximum transmit power that a legal drone is allowed to use within a single time slot; The communication power constraint is: (14); in, Indicates the first The common stream transmit beam vector of each time slot, Indicates the first The first time slot faces the first Private stream transmit beam vector of a ground vehicle Indicates the common stream transmit power. Indicates the first The transmit power of the private stream corresponding to each ground vehicle; The common stream transmission constraint is: (15); (16); in, Indicates the first The time slot is allocated to the first The proportion of public flow resources for each ground vehicle; The remaining energy constraint is: (17); in, The legal drone LU is in the Total energy consumption within each time slot This indicates the maximum initial energy budget pre-set for a legal drone; The total energy consumption includes flight propulsion energy consumption, sensing phase launch energy consumption, and communication phase launch energy consumption, namely: (18); in, Indicates the first Flight propulsion energy consumption per time slot, Indicates the first Energy consumption during the sensing phase of each time slot. Indicates the first The energy consumption of the communication phase in each time slot; the remaining energy constraint is used to ensure that the cumulative energy consumption of the legitimate UAV during the entire mission cycle does not exceed its maximum initial energy budget. A3: Set initialization parameters, including setting the initial values ​​of dual AoI states, the initial remaining energy budget of legal UAVs, course phase variables, and perception and communication related parameters; The initial values ​​of the dual AoI states include the initial values ​​of the sensing AoI state and the initial values ​​of the communication AoI state; Definition of the first The ground vehicle in The perceived AoI state value and the communication AoI state value for each time slot are respectively and Initial value of AoI state perception Initial values ​​of AoI communication state Set to: (19); (20); in, Indicates the first The AoI state value of a ground vehicle at the initial moment. Indicates the first The communication AoI state value of each ground vehicle at the initial moment; definition The legality of drones is indicated in the first... The remaining energy budget for each time slot will be used to determine the initial remaining energy budget for legitimate drones. Set to: (twenty one); Set course stage variables ,in, Indicates the training round index; when hour, =1 indicates that it is in the first course stage; when hour, =2 indicates that it is in the second course stage; when hour, =3 indicates that the student is in the third stage of the curriculum; among which... and This indicates the preset threshold for dividing the course into stages; Configure sensing and communication related parameters, including sensing thresholds. Public flow confidentiality threshold Private Stream Confidentiality Threshold Perceived AoI weights And communication AoI weights and maximum transmission power .

[0011] Furthermore, the acquisition of drone-assisted confidentiality ISAC systems has recently become a focus. Frame observation vectors, and construct from the nearest Temporal state composed of frame observation vectors The specific steps include: B1: In the In each time slot, the current environmental observation information of the UAV-assisted confidential ISAC system is collected to construct a single-frame observation vector. ; Single-frame observation vector Defined as: (twenty two); in, Represents the observation subvector of a legal UAV. Indicates the first The observation subvector of a ground vehicle. This represents the observation subvector of the mobile eavesdropping drone; The observation subvector of a legitimate UAV is defined as: (twenty three); No. The observation subvector of each ground vehicle is defined as: (twenty four); in, Indicates that legal drones are in the The first time slot for the first Estimates of the location of each ground vehicle; Indicates the first The ground vehicle in The successful perception status of each time slot; Indicates the first The time slot is allocated to the first The common flow transmission rate of each ground vehicle; Indicates the first The ground vehicle in Each time slot corresponds to the secure transmission rate of the private stream; definition Indicates the first The first time slot for the first A successful perception indicator for a ground vehicle; when At, it indicates the time. The first time slot for the first One ground vehicle successfully sensed the situation; when At, it indicates the time. The first time slot for the first One ground vehicle failed to sense the incident; a legitimate drone was in the [missing information]. The first time slot for the first Estimated location of each ground vehicle The update rule is: when hour, ;when hour, ,in, Indicates according to the first The position increment is predicted by the historical position, speed and direction of motion information of a ground vehicle in the previous time slot; The observation subvector of a mobile eavesdropping drone is defined as: (25); in, Indicates that mobile eavesdropping drones are in the first Each time slot corresponds to the received signal-to-interference-plus-noise ratio of the common stream. Indicates that mobile eavesdropping drones are used against the first The ground vehicle in Each time slot corresponds to the received signal-to-interference-plus-noise ratio of the private stream; B2: Based on single-frame observation vectors Construction by the most recent Temporal state composed of frame observation vectors ; recent The temporal state composed of frame observation vectors is as follows: (26); If the dimension of the single-frame observation vector is Then we have: (27); in, Indicates the legal drone observation dimensions. This represents the observation dimension of a single ground vehicle. This indicates the observation dimension of the mobile eavesdropping drone.

[0012] Furthermore, a gated recurrent unit-enhanced policy network and value network are constructed. Based on the initial scenario model, control policy constraints, and initialization parameters of the UAV-assisted confidential ISAC system, the near-end policy optimization (PPO) algorithm is used to iteratively update the GRU-enhanced policy network and value network, resulting in a perception and communication joint control strategy suitable for dynamic UAV-vehicle networking scenarios. Specific steps include: C1: Construct a policy network and value network enhanced by gated recurrent units; The policy network and value network enhanced by the gated recurrent unit both adopt a neural network structure containing a GRU temporal coding layer and a fully connected output layer; C2: A policy network enhanced with gated recurrent units based on temporal states. Output the original continuous action vector ; definition Indicates the first The hidden state of the GRU temporal coding layer in the nth slot, let the nth slot be... The hidden state update relationship corresponding to the frame observation vector input into the GRU temporal coding layer is as follows: (28); in, Indicates the frame number and , Indicates the GRU time-series coding layer. Indicates the first The frame observation vector is input into the GRU temporal coding layer and then outputs the corresponding hidden state. The observation vector from the last frame is input into the GRU temporal coding layer, and the corresponding hidden state is output. This serves as a temporal feature representation of the current time slot, and is further mapped to a shared feature vector: (29); in, Represents a shared feature mapping function; Fully connected output layer based on shared feature vectors Output the original continuous action vector ; The original continuous action vector Defined as: (30); in, This represents the normalized flight speed of a legitimate drone. Indicates the amount of adjustment to the flight direction of a legal drone; Indicates facing the first Common flow allocation parameters for each ground vehicle; Represents the original variable of total power during the sensing phase; Indicates facing the first Power allocation parameters for the perception phase of a ground vehicle; Indicates the first Ground vehicle sensing beam control parameters; Represents the original variable of total power during the communication phase; The original variable representing the proportion of common flow power; Indicates the common beam control parameters; Indicates facing the first The original variables for private flow power allocation of each ground vehicle; Indicates the first Private flow beam control parameters for each ground vehicle; C3: Output the current state value using a value network enhanced by a gated loop unit. It is used for advantage function estimation in the Proximal Policy Optimization (PPO) algorithm and parameter updates of policy and value networks enhanced by gated recurrent units. The state value is expressed as follows: (31); in, Represents the value mapping function; C4: Based on the course stage of the current training round The original continuous action vector is subjected to a phased action mask and a feasible mapping to obtain a feasible control quantity that satisfies the control strategy constraints. C5: Based on feasible control variables that satisfy control strategy constraints, in the first half of the time slot's perception phase, state perception is performed on vehicles on the ground in each location, and the success of the perception execution is determined to obtain a perception success flag. ; Legitimate drones utilize sensing beams during the first half of the sensing phase in each time slot. For the first For a ground vehicle, the interference and noise term during the perception phase... Defined as: (48); in, Indicates noise power; express The conjugate transpose of; The corresponding sensing signal-to-interference-plus-noise ratio is defined as: (49); When the The perceived signal-to-interference-plus-noise ratio of each ground vehicle meets the perception threshold requirement. When the perception is successfully executed, it is determined that the result is obtained. When the first The perceived signal-to-interference-plus-noise ratio of each ground vehicle meets the perception threshold requirement. When the perception execution fails, it is determined that the result is... ; C6: Based on feasible control variables, in the latter half of the communication phase, perform confidentiality state update communication based on RSMA and determine whether the public and private streams were successfully transmitted during the confidentiality state update communication process to obtain the communication success flag. ; During the latter half of the communication phase of each time slot, the legitimate UAV uses RSMA to perform secure state update communication and transmits signals. Represented as: (50); in, Indicates a common flow signal. Indicates the first Private flow signal of a ground vehicle; For the For each ground vehicle, the received signal-to-interference-plus-noise ratio (SIR) for public and private flows are as follows: (51); (52); in, Indicates the first The received signal-to-interference-plus-noise ratio corresponding to the public flow of ground vehicles; Indicates the first The received signal-to-interference-plus-noise ratio corresponding to the private flow of each ground vehicle; Indicates the first Interference and noise terms during the decoding of public flow by ground vehicles Indicates the first Interference and noise terms when decoding private streams by ground vehicles; Public flow total confidentiality rate Defined as: (53); in, Indicates the common stream transmission rate. This indicates the signal-to-interference-plus-noise ratio (SIR) received by a mobile eavesdropping drone over a public stream. This indicates the operation of obtaining a non-negative value; Assigned to the Public flow confidentiality rate of individual ground vehicles for: (54); No. The private flow confidentiality rate of a ground vehicle is defined as: (55); in, Indicates that mobile eavesdropping drones are used against the first The received signal-to-interference-plus-noise ratio of a private flow of ground vehicles; when At that time, it was believed that the first The public flow transmission of ground vehicles was successful; when At that time, it was believed that the first The private stream transmission of a ground vehicle is successful; the determination is made only if both decoding of the public stream and decoding of the private stream are successful. Communication with the ground vehicle was successfully executed; successful communication is indicated by... Otherwise ; C7: Based on perceived success indicators and communication success indicator Update the perceived AoI status value and the communication AoI status value, remaining energy and instant reward value; Update # The perceived AoI state value of each ground vehicle is: (56); Update # The communication AoI status value of each ground vehicle is: (57); Remaining energy Updated to: (58); C8: Value based on the current state Perform advantage function estimation and update the parameters of the GRU-enhanced policy network and value network; C9: Increment the current training round by 1 and determine whether the preset maximum number of training rounds has been reached. If so, end the training and output the final perception and communication joint control strategy suitable for dynamic UAV vehicle networking scenarios, consisting of feasible control quantities that satisfy the control strategy constraints. If not, determine whether to switch the course stage. If to switch, switch the course stage and return to C2; otherwise, return directly to C2.

[0013] Furthermore, the specific steps of C4 include: C4.1: Based on the course stage to which the current training round belongs. The original continuous action vector is subjected to a phased action masking and then decoded to obtain the decoded continuous action vector. Let the total dimension of the original continuous actions be... Define the set of actions that can be released in each course stage as follows: and And satisfy: (32); The phased action mask vector is further defined as follows: (33); in, Reflects the first The action dimension in the first... Whether it is activated in each training round; when hour, , indicating the first The action dimension in the first... Activated in one training round; otherwise , indicating the first The action dimension in the first... Not activated in any training round; When the The action dimension in the first... When activated in the training round, the decoded action component is equal to the original continuous action vector. The action dimension in the first... The motion components in the training rounds; when the first... The action dimension in the first... If the action component is not activated in a training round, the decoded action component is equal to the default action value for that training phase. ; C4.2: Perform a feasible mapping on each action component in the decoded continuous action vector to obtain feasible control quantities that satisfy the control policy constraints, including legal UAV maneuver control quantities, perception control quantities, and communication control quantities; The legal drone maneuver control quantities include the legal drone in the... Flight speed in time slots and legal drones in the first time slot The movement direction angle of each time slot; Map the normalized flight speed of the legitimate drone to the flight speed of the legitimate drone in the first... Flight speed per time slot: (34); Map the flight direction adjustment amount of the legitimate drone to the legitimate drone in the [missing information]. The movement direction angle of each time slot: (35); in, This indicates that the target range mapped by the legitimate drone flight direction adjustment is... ; The sensing control quantity includes the total power during the sensing phase, and the power directed towards the first... The power ratio of the perception phase for each ground vehicle and the first Sensing beams for a ground vehicle; Map the original variable of total power in the sensing phase to the total power in the sensing phase: (36); in, The total power during the sensing phase. Represents the Sigmoid mapping function; Will be directed towards the first The power allocation parameters for the perception phase of the ground vehicle are mapped to the first... Power percentage of each ground vehicle in the perception phase: (37); (38); in, Indicates facing the first The percentage of power allocated to the perception phase for each ground vehicle. Indicates assignment to the first The actual perceived power of a ground vehicle. Indicates facing the first Sensing power allocation parameters for each ground vehicle For the index variable in the normalization summation process; The first The control parameters of the ground vehicle sensing beam are mapped to the first... Sensing beams for each ground vehicle: (39); in, Indicates the first Sensing beams for each ground vehicle. Indicates the first Legal channel directions for each ground vehicle ， This represents the scaling factor for the sensing beam residual. This indicates a normalization operation; The communication control quantities include the total power during the communication phase, the first... The allocation ratio of common current power to total power in each time slot communication phase, common beam, and orientation towards the first time slot communication phase. Private current power ratio, private current power, and corresponding number of ground vehicles Private beams for each ground vehicle; Map the original variable of total power during the communication phase to the total power during the communication phase: (40); in, Total power during the communication phase; Will be directed towards the first The common flow allocation parameters for the ground vehicle are mapped to the parameters allocated to the first ground vehicle. The actual public flow proportion of each ground vehicle is used to characterize the allocation to the first... The share of public flow resources for each ground vehicle: (41); Map the original variable of common current power ratio to the first The proportion of common stream power in the total power of the communication phase in each time slot: (42); in, For the first The proportion of common stream power in each time slot communication phase to the total power of the communication phase; Constructing common current power for: (43); Common beam The construction is as follows: (44); in, Indicates the average legal channel direction. This represents the scaling factor for the common beam residual; Will be directed towards the first The original variables for the private flow power allocation of the ground vehicle are mapped to the first... Private current power percentage of each ground vehicle: (45); in, Indicates facing the first The percentage of private flow power for each ground vehicle Indicates facing the first The original variables for private flow power allocation of each ground vehicle; Construct the first Private current power of each ground vehicle for: (46); Corresponding to the Private beam for each ground vehicle The construction is as follows: (47); in, Indicates the first Private beam reference direction for each ground vehicle ， This represents the private beam residual scaling factor.

[0014] On the other hand, the present invention also provides a drone-vehicle network security freshness control system based on dual AoI, comprising: The system initialization module is used to build the UAV-assisted confidentiality ISAC system, establish the initial scenario model and control strategy constraints of the UAV-assisted confidentiality ISAC system, and set the initialization parameters. The timing state construction module is used to obtain the latest information from the UAV-assisted confidentiality ISAC system. Frame observation vectors, and construct from the nearest Temporal state composed of frame observation vectors ; The control strategy module is used to construct a policy network and value network enhanced by gated recurrent units. Based on the initial scenario model, control strategy constraints and initialization parameters of the UAV-assisted confidential ISAC system, the GRU-enhanced policy network and value network are iteratively updated using the near-end policy optimization (PPO) algorithm to obtain a joint perception and communication control strategy suitable for dynamic UAV vehicle-to-everything (V2X) scenarios.

[0015] Thirdly, this application proposes an electronic device, including: one or more processors, and a memory, the memory being used to store instructions, which, when executed by the one or more processors, cause the one or more processors to execute the aforementioned method for controlling the confidentiality and freshness of unmanned aerial vehicle networking based on dual AoI.

[0016] Fourthly, this application proposes a computer-readable storage medium storing executable instructions that, when executed, cause a processor to perform the aforementioned method for controlling the confidentiality and freshness of unmanned aerial vehicle (UAV) vehicle networking based on dual AoI.

[0017] Fifthly, this application proposes a computer program product, including a computer program or instructions, which, when executed by a processor, implements the aforementioned method for controlling the confidentiality and freshness of unmanned aerial vehicle (UAV) vehicle networking based on dual AoI.

[0018] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention provides a method and system for controlling the security freshness of UAV-vehicle networking based on dual AoI (Aspect-of-Information) technology. By constructing an integrated perception and communication system encompassing the UAV platform, vehicle-mounted users, sensing targets, and mobile eavesdropping nodes, a joint optimization mechanism is proposed for the collaborative design of UAV trajectory control, perception resource allocation, communication power allocation, and public and private flow transmission. This mechanism effectively balances environmental perception freshness, information transmission freshness, and secure communication performance in complex and highly dynamic environments. The proposed method aims to reduce dual AoI and improve the system's long-term security service capability. By introducing a phased curriculum reinforcement learning mechanism and combining it with GRU for feature extraction of time-series observation information, the system can achieve stable decision-making and adaptive control under partially observable, highly maneuverable, and strongly coupled constraints, significantly improving the UAV's mission execution efficiency in continuous time slots. Simultaneously, based on the RSMA secure transmission mechanism, this invention coordinates the transmission of public and private information flows and combines time-division control between the perception and communication phases to effectively suppress performance loss caused by mobile eavesdropping nodes to secure communication, enhancing the system's anti-eavesdropping capability and security reliability in open wireless environments. Simulation results demonstrate that the present invention achieves superior freshness control and stable secure transmission performance under various vehicle motion states, eavesdropping threat intensities, energy budgets, and mission durations. Compared to traditional methods that optimize only communication rates, single AoI metrics, or static resource configurations, the present invention achieves multi-objective collaborative optimization under conditions of simultaneous constraints on perception, communication, security, and energy. It exhibits stronger environmental adaptability, control robustness, and engineering practical value, making it particularly suitable for applications in 6G, ITS, and low-altitude intelligent networks that demand high levels of real-time perception, low-latency transmission, and secure information interaction. Attached Figure Description

[0019] Figure 1 A flowchart illustrating a method for controlling the confidentiality and freshness of unmanned aerial vehicle (UAV) vehicle-to-everything (V2X) networks based on dual AoIs, provided in an embodiment of the present invention; Figure 2 A model diagram of the UAV-assisted security ISAC system provided in an embodiment of the present invention; Figure 3 The simulation diagram of the total reward convergence curve during the training process of a UAV vehicle-to-everything (V2X) security freshness control method based on dual AoI provided in an embodiment of the present invention. Figure 4 This is a simulation diagram comparing the performance of the UAV-assisted confidentiality ISAC system under different multiple access and transmission schemes provided in the embodiments of the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the following embodiments are only used to illustrate the technical concept, control process and implementation of the present invention, and are not intended to limit the scope of protection of the present invention. The embodiments are only some embodiments of the present invention, not all embodiments. The present invention focuses on key embodiments. Without departing from the core idea of ​​the present invention, those skilled in the art can adjust or make equivalent substitutions to the parameter settings, network size, training rounds and constraints according to specific application scenarios. Other embodiments obtained without creative effort are all within the scope of protection of the present invention.

[0021] Example 1: like Figure 1 and Figure 2 As shown, this embodiment addresses the scenario of integrated perception and communication in UAV-assisted vehicle-to-everything (V2X) networks, providing a method for controlling the confidentiality and freshness of UAV-based V2X networks based on dual AoIs. The specific steps are as follows: S1: Construct a UAV-assisted security ISAC (Integrated Sensing and Communication) system, and establish an initial scenario model and control strategy constraints for the UAV-assisted security ISAC system, and set initialization parameters; the control strategy constraints are used for subsequent feasible control quantity mapping, perception success determination, communication success determination and remaining energy update, including flight constraints of legitimate UAVs, flight trajectory constraints of mobile eavesdropping UAVs, perception power constraints, communication power constraints, common flow transmission constraints and remaining energy constraints; S1.1: Construct a drone-assisted security ISAC system for vehicle-to-everything (V2X) status update scenarios, including a legitimate drone (LU), a mobile eavesdropping drone (EU), and... There are 10 ground vehicles (GVs), of which the legal drone (LU) is equipped with M antennas, while each ground vehicle (GV) and mobile eavesdropping drone (EU) is equipped with a single antenna. The set of ground vehicles is represented as: (1); in, Indicates a collection of ground vehicles; In this embodiment, the drone-assisted security ISAC system mainly includes a legitimate drone with sensing and communication capabilities, multiple ground vehicles in the road area to be served, and at least one mobile eavesdropping drone. The legitimate UAV with sensing and communication capabilities acts as an aerial intelligent node, simultaneously undertaking environmental perception, downlink information transmission, and security assurance tasks during flight; the ground vehicle acts as a legitimate service recipient, and its position changes dynamically over time; the mobile eavesdropping node passively listens to the downlink signals of the legitimate UAV in an open wireless environment, thereby posing a threat to the secure transmission performance of the UAV-assisted confidentiality ISAC system. S1.2: Discretize the mission cycle of the UAV-assisted security ISAC system into... In each time slot, establish the initial scenario model and control strategy constraints of the UAV-assisted confidential ISAC system; The length of the time slot is Each time slot is divided into a first half perception phase and a second half communication phase. In the first half perception phase, a legitimate UAV performs state perception on vehicles on all ground, completing the perception execution. In the second half communication phase, a legitimate UAV uses Rate Split Multiple Access (RSMA) to perform confidential state update communication on vehicles on all ground, completing the communication execution. The initial scenario model of the UAV-assisted confidentiality ISAC system includes the following: A legal drone flies at a fixed altitude, and its first... Position vector of each time slot : (2); in, Indicates the first A time-slot-based legal UAV in a horizontal two-dimensional coordinate system Position components along the axial direction; Indicates the first A legal UAV in a time slot travels along a horizontal two-dimensional coordinate system. Position components along the axial direction; Indicates the time slot index; Indicates the flight altitude of a legal drone; Indicates transpose; the horizontal two-dimensional coordinate system, The axial direction is defined as the direction along which the road extends. The axial direction is defined as perpendicular to the direction of road extension; A legal drone flies at a fixed altitude, and its first... The position vector update relationship for each time slot is as follows: (3); (4); in, Indicates that legal drones are in the Position vector of each time slot Indicates that legal drones are in the Position increment within a time slot; Indicates that legal drones are in the Flight speed per time slot Indicates that legal drones are in the The movement direction angle of each time slot; In this embodiment, a legitimate drone is used to perform integrated perception and communication tasks in a vehicle-to-everything (V2X) scenario, including collecting environmental status information and transmitting relevant data to ground vehicles; The mobile eavesdropping drone flies along a circular trajectory at a fixed altitude, and its trajectory is on the... Position vector of each time slot Represented as: (5); in, Indicates the first A mobile eavesdropping drone in a time slot travels along a horizontal two-dimensional coordinate system. Position components in the axial direction Indicates the first A mobile eavesdropping drone in a time slot travels along a horizontal two-dimensional coordinate system. Position components in the axial direction The flight altitude of the mobile eavesdropping drone; No. The ground vehicle in Position vector of each time slot Represented as: (6); No. The ground vehicle in The position vector update relationship for each time slot is represented as follows: (7); in, Indicates the first The lane where each ground vehicle is located is in a horizontal two-dimensional coordinate system. Position components in the axial direction Indicates the first The first time slot The lane where each ground vehicle is located is along the horizontal two-dimensional coordinate system. Position components in the axial direction Indicates the first The ground vehicle in The speed of travel in each time slot; The flight constraints for the legal drones are as follows: (8); (9); in, This indicates the maximum permissible flight speed for a legal drone during flight. Indicates the feasible service area for legal drones; The flight trajectory constraints of the mobile eavesdropping drone are as follows: (10); (11); (12); in, and This indicates the center coordinates of the circular trajectory of the mobile eavesdropping drone. This represents the orbital radius of the circular trajectory of a mobile eavesdropping drone. This indicates the initial phase of the mobile eavesdropping drone. This indicates the period of the mobile eavesdropping drone's flight along a circular trajectory. This represents the angular velocity of the mobile eavesdropping drone as it flies along a circular trajectory. Based on the above constraints on the flight of legal drones and the flight trajectory of mobile eavesdropping drones, the drone-assisted security ISAC system also needs to meet the constraints on sensing power, communication power, common stream transmission, and remaining energy. The sensing power constraint is: (13); in, Indicates the first The first time slot faces the first The sensing beam vector of a ground vehicle This indicates the transmit power corresponding to the sensing beam vector. This indicates the maximum transmit power that a legal drone is allowed to use within a single time slot; The communication power constraint is: (14); in, Indicates the first The common stream transmit beam vector of each time slot, Indicates the first The first time slot faces the first Private stream transmit beam vector of a ground vehicle Indicates the common stream transmit power. Indicates the first The transmit power of each ground vehicle corresponds to a private stream. The communication power constraint is used to limit the combined transmit power of the public stream and all private streams.

[0022] The common stream transmission constraint is: (15); (16); in, Indicates the first The time slot is allocated to the first The proportion of public flow resources allocated to each ground vehicle. The above constraints ensure that the proportion of public flow resources allocated to each ground vehicle is non-negative, and that the sum of the proportions of public flow resources allocated to all ground vehicles in the same time slot is 1.

[0023] The remaining energy constraint is: (17); in, The legal drone LU is in the Total energy consumption within each time slot This indicates the maximum initial energy budget pre-set for a legal drone; The total energy consumption includes flight propulsion energy consumption, sensing phase launch energy consumption, and communication phase launch energy consumption, namely: (18); in, Indicates the first Flight propulsion energy consumption per time slot, Indicates the first Energy consumption during the sensing phase of each time slot. Indicates the first The energy consumption of the communication phase in each time slot; the remaining energy constraint is used to ensure that the cumulative energy consumption of the legitimate UAV in the entire mission cycle does not exceed its maximum initial energy budget; S1.3: Set initialization parameters, including setting the initial value of the dual AoI (information freshness) state, the initial remaining energy budget of the legitimate UAV (LU), course phase variables and perception and communication related parameters; The initial values ​​of the dual AoI states include the initial values ​​of the sensing AoI state and the initial values ​​of the communication AoI state; To reflect the dual requirements of information timeliness for UAV-assisted security ISAC systems, this embodiment defines perception AoI state values ​​and communication AoI state values ​​to characterize the freshness of environmental state updates and the freshness of effective security information received by ground vehicles, and uses both as important objectives for near-end strategy optimization; defining the first The ground vehicle in The perceived AoI state value and the communication AoI state value for each time slot are respectively and Initial value of AoI state perception Initial values ​​of AoI communication state Set to: (19); (20); in, Indicates the first The AoI state value of a ground vehicle at the initial moment. Indicates the first The communication AoI state value of each ground vehicle at the initial moment; definition The legality of drones is indicated in the first... The remaining energy budget for each time slot will be used to determine the initial remaining energy budget for legitimate drones. Set to: (twenty one); To support subsequent phased training, course phase variables are set. ,in, Indicates the training round index; when hour, =1 indicates that the program is in the first stage of the course, used to train basic maneuver control strategies for legal drones; when hour, =2 indicates that it is in the second course stage, used to train perception control strategies and common flow transmission control strategies based on the basic maneuver control strategies of legal UAVs; when hour, =3 indicates that the program is in the third stage of the course, used to train private stream transmission control strategies and joint sensing and communication control strategies; where... and This indicates the preset threshold for dividing the course into stages; In this embodiment, after establishing the initial system scenario model through the above operations, and completing the initial values ​​of the dual AoI states, the initial remaining energy budget of the legal UAV, and the course stage variable settings, the perception and communication related parameters, including the perception threshold, are further set. Public flow confidentiality threshold Private Stream Confidentiality Threshold Perceived AoI weights And communication AoI weights and maximum transmission power ; S2: Recent access to the drone-assisted confidential ISAC system Frame observation vectors, and construct from the nearest Temporal state composed of frame observation vectors ; S2.1: In the first In each time slot, the current environmental observation information of the UAV-assisted confidential ISAC system is collected to construct a single-frame observation vector. ; Single-frame observation vector Defined as: (twenty two); in, Represents the observation subvector of a legal UAV. Indicates the first The observation subvector of a ground vehicle. This represents the observation subvector of the mobile eavesdropping drone; The observation subvector of a legitimate UAV is defined as: (twenty three); No. The observation subvector of each ground vehicle is defined as: (twenty four); in, Indicates that legal drones are in the The first time slot for the first Estimates of the location of each ground vehicle; Indicates the first The ground vehicle in The successful perception status of each time slot reflects the current state of perception update. Indicates the first The time slot is allocated to the first The common flow transmission rate of each ground vehicle; Indicates the first The ground vehicle in Each time slot corresponds to the secure transmission rate of the private stream, reflecting the current secure status update communication status; Given the intermittent nature of legitimate UAVs acquiring the status of ground vehicles, the following method is used to correct and update the legitimate UAV's data in the [number]th [period]. The first time slot for the first Estimated location of each ground vehicle: definition Indicates the first The first time slot for the first A successful perception indicator for a ground vehicle; when At, it indicates the time. The first time slot for the first One ground vehicle successfully sensed the situation; when At, it indicates the time. The first time slot for the first One ground vehicle failed to sense the incident; a legitimate drone was in the [missing information]. The first time slot for the first Estimated location of each ground vehicle The update rule is: when hour, ;when hour, ,in, Indicates according to the first The position increment is predicted by the historical position, speed and direction of motion information of a ground vehicle in the previous time slot; The observation subvector of a mobile eavesdropping drone is defined as: (25); in, Indicates that mobile eavesdropping drones are in the first Each time slot corresponds to the received signal-to-interference-plus-noise ratio of the common stream. Indicates that mobile eavesdropping drones are used against the first The ground vehicle in Each time slot corresponds to the received signal-to-interference-plus-noise ratio of the private stream; S2.2: Based on single-frame observation vectors Construction by the most recent Temporal state composed of frame observation vectors ; To preserve the evolution of ground vehicle movement, mobile eavesdropping drone movement, perception AoI state values, and communication AoI state values, as well as the temporal correlation of link state changes corresponding to public and private flows, the most recent... The temporal state composed of frame observation vectors is as follows: (26); If the dimension of the single-frame observation vector is Then we have: (27); in, Indicates the legal drone observation dimensions. This represents the observation dimension of a single ground vehicle. Indicates the observation dimension of the mobile eavesdropping drone; In this embodiment, the legitimate UAV receives current environmental observation information at the beginning of each time slot. The environmental observation information includes at least the historical position, speed, and direction of movement of ground vehicles, the estimated position of the mobile eavesdropping UAV, the legitimate UAV's own position vector and remaining energy budget, the perception success flag and communication success flag corresponding to ground vehicle users in the previous time slot, and the perception AoI state value and communication AoI state value corresponding to each ground vehicle, to support the phased training of subsequent courses. Through the above steps, a unified mapping from the current operating state of the UAV-assisted confidentiality ISAC system to the temporal state is completed, enabling the policy network and value network of the subsequent gated recurrent unit (GRU) enhancement to extract effective temporal state information in partially observable and highly dynamic environments. S3: Construct a gated recurrent unit (GRU) enhanced policy network and value network. Based on the initial scenario model, control policy constraints, and initialization parameters of the UAV-assisted confidential ISAC system, iteratively update the GRU-enhanced policy network and value network using the near-end policy optimization (PPO) algorithm to obtain a perception and communication joint control strategy suitable for dynamic UAV-vehicle network scenarios. The gated recurrent unit (GRU) enhanced policy network is used to adjust the policy network according to the time-series state. Output the original continuous action vector The value network enhanced by the gated recurrent unit (GRU) is used to determine the time-series state. Output the current state value ; S3.1: Construct a policy network and value network enhanced by gated recurrent units (GRUs); In this embodiment, both the policy network and the value network enhanced by the gated recurrent unit (GRU) can be implemented using a neural network structure containing a GRU temporal coding layer and a fully connected output layer as an example. S3.2: Policy network enhanced with gated recurrent units (GRUs) based on temporal state Output the original continuous action vector ; For the policy network enhanced by the gated recurrent unit (GRU), the temporal state is... The input GRU temporal coding layer performs temporal coding on the movement of ground vehicles, changes in the trajectory of mobile eavesdropping drones, changes in the link states of public flows and private flows, and changes in the state values ​​of sensing AoI and communication AoI over time slots. This data is then further mapped to obtain a shared feature vector. ; Specifically: Definition Indicates the first The hidden state of the GRU temporal coding layer in the nth slot, let the nth slot be... The hidden state update relationship corresponding to the frame observation vector input into the GRU temporal coding layer is as follows: (28); in, Indicates the frame number and , Indicates the GRU time-series coding layer. Indicates the first The frame observation vector is input into the GRU temporal coding layer and then outputs the corresponding hidden state. The observation vector from the last frame is input into the GRU temporal coding layer, and the corresponding hidden state is output. This serves as a temporal feature representation of the current time slot, and is further mapped to a shared feature vector: (29); in, Represents a shared feature mapping function; In this embodiment, in terms of time-series state processing, the observation vectors of multiple recent consecutive time slots are arranged in chronological order to form a time-series state and input into the GRU time-series coding layer for feature extraction. The memory capability of the GRU time-series coding layer for historical position, speed and direction information is used to characterize the movement pattern of ground vehicles, the position change of mobile eavesdropping drones and their corresponding eavesdropping link status changes, as well as the time-series change characteristics of the perception AoI state value and the communication AoI state value increasing or updating with time slots, thereby generating a hidden state that is more suitable for continuous control tasks. Then, the fully connected output layer is based on the shared feature vector. Output the original continuous action vector ; The original continuous action vector Defined as: (30); in, This represents the normalized flight speed of a legitimate drone. Indicates the amount of adjustment to the flight direction of a legal drone; Indicates facing the first Common flow allocation parameters for each ground vehicle; Represents the original variable of total power during the sensing phase; Indicates facing the first Power allocation parameters for the perception phase of a ground vehicle; Indicates the first Ground vehicle sensing beam control parameters; Represents the original variable of total power during the communication phase; The original variable representing the proportion of common flow power; Indicates the common beam control parameters; Indicates facing the first The original variables for private flow power allocation of each ground vehicle; Indicates the first Private flow beam control parameters for each ground vehicle; In this way, this embodiment unifies the trajectory optimization, perception control and secure communication control of traditional distributed design into the same decision-making framework, and realizes the coordinated adjustment of multi-dimensional coupled resources. In this embodiment, considering the characteristics of high-speed movement of ground vehicles, rapid channel fluctuations, and incomplete observations in the vehicle-to-everything (V2X) scenario, this embodiment does not adopt a static decision-making method that relies solely on information from a single moment. Instead, it arranges the observation vectors of multiple recent consecutive time slots into a temporal state in chronological order to preserve the dynamic evolution characteristics of the UAV-assisted confidential ISAC system. Through the above steps, the mapping from the temporal state to the original continuous action vectors of joint control is realized, providing a foundation for subsequent course stage selection and feasible mapping. S3.3: Output the current state value using a value network enhanced by a gated recurrent unit (GRU). It is used for advantage function estimation in the Proximal Policy Optimization (PPO) algorithm and parameter updates of policy and value networks enhanced by gated recurrent units (GRUs). The state value is expressed as follows: (31); in, Represents the value mapping function; S3.4: Based on the course phase of the current training round The original continuous action vector is subjected to a phased action mask and a feasible mapping to obtain a feasible control quantity that satisfies the control strategy constraints. S3.4.1: Based on the course stage to which the current training round belongs. The original continuous action vector is subjected to a phased action masking and then decoded to obtain the decoded continuous action vector. The training round refers to the round in which the perception and communication joint control strategy of the UAV-assisted confidentiality ISAC system is trained by reinforcement learning based on the near-end policy optimization PPO algorithm. Let the total dimension of the original continuous actions be... Define the set of actions that can be released in each course stage as follows: and And satisfy: (32); The phased action mask vector is further defined as follows: (33); in, Reflects the first The action dimension in the first... Whether it is activated in each training round; when hour, , indicating the first The action dimension in the first... Activated in one training round; otherwise , indicating the first The action dimension in the first... Not activated in any training round; When the The action dimension in the first... When activated in the training round, the decoded action component is equal to the original continuous action vector. The action dimension in the first... The motion components in the training rounds; when the first... The action dimension in the first... If the action component is not activated in a training round, the decoded action component is equal to the default action value for that training phase. ; S3.4.2: Perform a feasible mapping on each action component in the decoded continuous action vector to obtain feasible control quantities that satisfy the control strategy constraints, including legal UAV maneuver control quantities, perception control quantities, and communication control quantities; The legal drone maneuver control quantities include the legal drone in the... Flight speed in time slots and legal drones in the first time slot The movement direction angle of each time slot; Map the normalized flight speed of the legitimate drone to the flight speed of the legitimate drone in the first... Flight speed per time slot: (34); Map the flight direction adjustment amount of the legitimate drone to the legitimate drone in the [missing information]. The movement direction angle of each time slot: (35); in, This indicates that the target range mapped by the legitimate drone flight direction adjustment is... ; The sensing control quantity includes the total power during the sensing phase, and the power directed towards the first... The power ratio of the perception phase for each ground vehicle and the first Sensing beams for a ground vehicle; Map the original variable of total power in the sensing phase to the total power in the sensing phase: (36); in, The total power during the sensing phase. This represents the Sigmoid mapping function, used to map the original variable of total power during the sensing phase. Mapped to the interval (0,1), thus increasing the total power of the sensing phase. Satisfy the total perceived power mapping constraint; Will be directed towards the first The power allocation parameters for the perception phase of the ground vehicle are mapped to the first... Power percentage of each ground vehicle in the perception phase: (37); (38); in, Indicates facing the first The percentage of power allocated to the perception phase for each ground vehicle. Indicates assignment to the first The actual perceived power of a ground vehicle. Indicates facing the first Sensing power allocation parameters for each ground vehicle For the index variable in the normalization summation process; The first The control parameters of the ground vehicle sensing beam are mapped to the first... Sensing beams for each ground vehicle: (39); in, Indicates the first Sensing beams for each ground vehicle. Indicates the first Legal channel directions for each ground vehicle ， This represents the scaling factor for the sensing beam residual. This indicates a normalization operation; The communication control quantities include the total power during the communication phase, the first... The allocation ratio of common current power to total power in each time slot communication phase, common beam, and orientation towards the first time slot communication phase. Private current power ratio, private current power, and corresponding number of ground vehicles Private beams for each ground vehicle; Map the original variable of total power during the communication phase to the total power during the communication phase: (40); in, Total power during the communication phase; Will be directed towards the first The common flow allocation parameters for the ground vehicle are mapped to the parameters allocated to the first ground vehicle. The actual public flow proportion of each ground vehicle is used to characterize the allocation to the first... The share of public flow resources for each ground vehicle: (41); Map the original variable of common current power ratio to the first The proportion of common stream power in the total power of the communication phase in each time slot: (42); in, For the first The proportion of common stream power in each time slot communication phase to the total power of the communication phase; Constructing common current power for: (43); Common beam The construction is as follows: (44); in, Indicates the average legal channel direction. This represents the scaling factor for the common beam residual; Will be directed towards the first The original variables for the private flow power allocation of the ground vehicle are mapped to the first... Private current power percentage of each ground vehicle: (45); in, Indicates facing the first The percentage of private flow power for each ground vehicle Indicates facing the first The original variables for private flow power allocation of each ground vehicle; Construct the first Private current power of each ground vehicle for: (46); Corresponding to the Private beam for each ground vehicle The construction is as follows: (47); in, Indicates the first Private beam reference direction for each ground vehicle ， This represents the private beam residual scaling factor; In this embodiment, the decoded action components are feasiblely mapped using a phased curriculum reinforcement learning strategy. This strategy involves the controller gradually learning optimal behavior in a complex, coupled environment, progressing from easy to difficult. Specifically, in the initial training phase, only action dimensions related to basic UAV maneuver control are opened, allowing the controller to prioritize mastering basic capabilities such as trajectory maintenance, approaching the service area, and avoiding unfavorable locations. In subsequent training phases, higher-dimensional actions, such as perception and communication power constraints and RSMA-based collaborative control of public and private flow transmission constraints, are gradually introduced. For action dimensions that are not yet open, they are fixed to preset safety values ​​by using action masks, thus avoiding oscillations or difficulty in convergence of the phased curriculum reinforcement learning strategy due to an excessively large search space in the early stages of training. This phased training method is particularly suitable for UAV-vehicle networking scenarios characterized by high maneuverability, strong coupling, and multiple constraints, effectively improving learning stability and the convergence efficiency of the phased curriculum reinforcement learning strategy. Through the above steps, the original continuous action vectors output by the GRU-enhanced policy network are transformed into feasible control quantities that meet the requirements of the course stage and the constraints of the control policy. The first stage mainly completes the UAV flight constraints and perception power constraints. The second stage further completes the communication power constraints and the transmission constraints of the common flow. The third stage releases the transmission constraints of the private flow and performs fine control of the private flow to provide feasible inputs for subsequent perception execution and communication execution.

[0024] S3.5: Based on feasible control variables that satisfy control strategy constraints, in the first half of the time slot, state perception is performed on vehicles on the ground in each location, and the success of perception execution is determined to obtain a perception success flag. ; Legitimate drones utilize sensing beams during the first half of the sensing phase in each time slot. For the first For a ground vehicle, the interference and noise term during the perception phase... Defined as: (48); in, Indicates noise power; express The conjugate transpose of; The corresponding sensing signal-to-interference-plus-noise ratio is defined as: (49); When the The perceived signal-to-interference-plus-noise ratio of each ground vehicle meets the perception threshold requirement. When the perception is successfully executed, it is determined that the result is obtained. When the first The perceived signal-to-interference-plus-noise ratio of each ground vehicle meets the perception threshold requirement. When the perception execution fails, it is determined that the result is... ; In this embodiment, during the perception phase, the UAV detects the target area or the relevant status of ground vehicles based on the total power, power ratio, and perception beam control parameters output by the current phased course reinforcement learning strategy. It also determines the success of the perception task in this time slot based on the received echo quality. When the perception signal-to-interference-plus-noise ratio (SIR) meets a preset perception threshold, the corresponding ground vehicle user is determined to have successfully perceived the system, a perception success flag is obtained, and the perception AoI status value of the corresponding ground vehicle user is set to a smaller value during subsequent perception AoI updates. When the perception SIR does not meet the preset perception threshold, the corresponding ground vehicle user is determined to have failed to perceive the system, and the perception AoI status value of the corresponding ground vehicle user is set to a smaller value during subsequent perception AoI updates. The perceived AoI status value increases based on the previous time slot; thus, the UAV-assisted confidential ISAC system can characterize the aging degree of the perceived information of the corresponding ground vehicle in real time through the perceived AoI status value, and prompt subsequent decisions to pay more attention to ground vehicles that have not been effectively perceived and updated for a long time; compared with the traditional method that only optimizes perception accuracy, this embodiment introduces the perceived AoI status value as a time-sensitive indicator, enabling the UAV to prioritize the service continuity of ground vehicles or targets with poor information freshness in a limited remaining energy budget and complex environment; through the above steps, the perception execution of the first half of the time slot is completed, and a perception success indicator is provided for the communication execution and perceived AoI status value update of the second half; S3.6: Based on feasible control variables, in the latter half of the communication phase, perform confidentiality state update communication based on RSMA and determine whether the public and private streams were successfully transmitted during the confidentiality state update communication process to obtain the communication success flag. ; During the latter half of the communication phase of each time slot, the legitimate UAV uses RSMA to perform secure state update communication and transmits signals. Represented as: (50); in, Indicates a common flow signal. Indicates the first Private flow signal of a ground vehicle; For the For each ground vehicle, the received signal-to-interference-plus-noise ratio (SIR) for public and private flows are as follows: (51); (52); in, Indicates the first The received signal-to-interference-plus-noise ratio corresponding to the public flow of ground vehicles; Indicates the first The received signal-to-interference-plus-noise ratio corresponding to the private flow of each ground vehicle; Indicates the first Interference and noise terms during the decoding of public flow by ground vehicles Indicates the first Interference and noise terms when decoding private streams by ground vehicles; Public flow total confidentiality rate Defined as: (53); in, Indicates the common stream transmission rate. This indicates the signal-to-interference-plus-noise ratio (SIR) received by a mobile eavesdropping drone over a public stream. This indicates the operation of obtaining a non-negative value; Assigned to the Public flow confidentiality rate of individual ground vehicles for: (54); No. The private flow confidentiality rate of a ground vehicle is defined as: (55); in, Indicates that mobile eavesdropping drones are used against the first The received signal-to-interference-plus-noise ratio of a private flow of ground vehicles; when At that time, it was believed that the first The public flow transmission of ground vehicles was successful; when At that time, it was believed that the first The private stream transmission of a ground vehicle is successful; the determination is made only if both decoding of the public stream and decoding of the private stream are successful. Communication with the ground vehicle was successfully executed; successful communication is indicated by... Otherwise ; In this embodiment, during the communication phase, the UAV sends downlink information to multiple ground vehicles based on the RSMA mechanism. Specifically, the information to be sent is divided into a public information stream and private information streams for each user, and then superimposed and sent through corresponding sensing beams. Each ground vehicle first decodes the public stream, and then decodes its own private stream. Mobile eavesdropping nodes may monitor the public or private streams. To ensure effective and secure transmission, this embodiment considers not only whether the ground vehicle can complete the correct decoding when determining whether the communication is successful, but also whether the security rates of its corresponding public and private streams meet the preset security requirements. Through the above steps, the secure status update communication based on RSMA in the second half of the time slot is completed, and the communication success flags of each ground vehicle are obtained. S3.7: Based on perceived success indicators and communication success indicator Update the perceived AoI status value and the communication AoI status value, remaining energy and instant reward value; Update # The perceived AoI state value of each ground vehicle is: (56); Update # The communication AoI status value of each ground vehicle is: (57); Specifically, when the perception success flag or communication success flag value is 1, its corresponding AoI status value is reset to 1. When the perception success flag and communication success flag values ​​are 0, the corresponding AoI status value increases based on the AoI status value of the previous time slot. ; In this embodiment, when a user completes a valid and secure information reception in the current time slot, its communication AoI status value is refreshed; otherwise, its communication AoI status value continues to increase. By coupling the decoding of public and private streams and the AoI status value update mechanism during the confidentiality status update communication process, this embodiment can simultaneously address the two key issues of "whether the message is transmitted" and "whether it is transmitted securely," thereby avoiding the shortcomings of traditional designs that only focus on throughput or confidentiality capacity and cannot reflect the timeliness of information. Remaining energy Updated to: (58); After updating the perception AoI state value and communication AoI state value and the remaining energy in the current time slot, the perception and communication performance of the UAV-assisted confidentiality ISAC system in this time slot is further comprehensively evaluated, and the evaluation results are used to form the instant reward value of the current time slot for subsequent near-end policy optimization PPO algorithm updates. The instant reward value is mainly determined by combining factors such as the perceived AoI status value, the communication AoI status value, the execution of confidentiality status update communication security, energy consumption, action stability, and the satisfaction of control strategy constraints. In this embodiment, when the UAV successfully completes state awareness and secure communication in the current time slot, and the dual AoI state values ​​are low, energy consumption is low, and control actions are relatively stable, the immediate reward value is relatively good; when perception fails, communication fails, dual AoI is high, energy consumption is high, or there is a violation of control policy constraints, the immediate reward value is relatively poor. By introducing the above-mentioned immediate reward value, the policy network enhanced with GRU can take into account information freshness control, continuous energy utilization, flight trajectory rationality, and system safety service performance during long-term training.

[0025] S3.8: Value based on the current state Perform advantage function estimation and update the parameters of the GRU-enhanced policy network and value network; When the UAV-assisted confidentiality ISAC system is in training mode, the timing state of the current time slot, the original continuous action vector, the instant reward value, and the timing state of the next time slot are written into the rollout buffer to form a state transition sample; the state transition sample includes the current state, action, instant reward, and next state. The current state value output by the value network enhanced by GRU The advantage function is estimated by combining state transition samples, and the parameters of the GRU-enhanced policy network and value network are updated based on the advantage function estimation results and the PPO objective function of the near-end policy optimization (PPO) algorithm. Specifically, when the rollout buffer reaches a preset length, the objective function of the PPO algorithm is optimized based on the near-end policy, the current state value evaluation result output by the GRU-enhanced value network is used, and the generalized dominance estimation method is used to estimate the dominance function of the state transition sample; further, the parameters of the GRU encoding layer and the fully connected output layer in the GRU-enhanced policy network and value network are updated using the state transition sample and the dominance function estimation result. S3.9: Increment the current training round by 1 and determine whether the preset maximum number of training rounds has been reached. If yes, end the training and output the final perception and communication joint control strategy suitable for dynamic UAV vehicle networking scenarios, consisting of feasible control quantities that satisfy the control strategy constraints. If no limit has been reached, determine whether to switch the course stage. If to switch, switch the course stage and return to S3.2; otherwise, return directly to S3.2. During training, the current round number is continuously monitored. and stage boundaries , The relationship; when the training rounds reach or Switch the corresponding course stage variable at that time. Simultaneously adjust the action release set, action mask vector, perception thresholds for each stage, public flow confidentiality thresholds, private flow confidentiality thresholds, as well as perception AoI weights and communication AoI weights. In this embodiment, the training process uses the Proximal Policy Optimization (PPO) algorithm to iteratively update the GRU-enhanced policy network and value network. Combined with the course stage variable switching rules, the task difficulty is gradually increased, so that the control policy gradually transitions from basic maneuver control and perception control to public flow confidentiality control and private flow fine control, and finally obtains a perception and communication joint control policy suitable for dynamic UAV vehicle networking scenarios. In this embodiment, multiple ground vehicles within a road area can be configured to move along predetermined lanes, with their speeds varying within a given range. Meanwhile, mobile eavesdropping drone nodes move along independent trajectories in the vicinity of the road. Legitimate drones perform continuous time-slot control within a fixed flight altitude range, constrained by maximum flight speed, initial remaining energy budget, and single-slot transmission power limits. After training, the drone-assisted security ISAC system deploys the learned control strategies into the online decision-making module, enabling the drone to autonomously adjust its flight trajectory, sensing resources, and communication resources based on real-time observations. Thus, even under conditions of rapidly changing ground vehicle positions, dynamically increasing eavesdropping threats, and limited resources in the drone-assisted security ISAC system, a consistently high level of sensing and communication freshness can be maintained.

[0026] like Figure 3 and Figure 4 As shown, this embodiment validates the proposed method in a simulation environment, including performance comparisons of RSMA, NOMA, and other comparative schemes. Simulation results demonstrate that the proposed method exhibits excellent total training reward convergence characteristics during training, gradually forming a stable and effective control strategy with increasing training rounds. Furthermore, compared to comparative methods employing NOMA or other benchmark transmission mechanisms, this embodiment demonstrates superior performance in dual AoI control, long-term secure service capabilities, and robustness under complex dynamic environments. This indicates that the present invention can establish an effective collaborative relationship between UAV maneuver control, perception updates, and RSMA secure transmission, making it suitable for UAV-vehicle networking application scenarios with high requirements for real-time performance, security, and continuous service capabilities.

[0027] Example 2 A dual-AoI-based drone-vehicle network security freshness control system includes: The system initialization module is used to build the UAV-assisted secure ISAC (Integrated Sensing and Communication) system, establish the initial scenario model and control strategy constraints of the UAV-assisted secure ISAC system, and set the initialization parameters. The timing state construction module is used to obtain the latest information from the UAV-assisted confidentiality ISAC system. Frame observation vectors, and construct from the nearest Temporal state composed of frame observation vectors ; The control strategy module is used to construct a policy network and value network enhanced by gated recurrent units (GRUs). Based on the initial scenario model, control strategy constraints and initialization parameters of the UAV-assisted confidential ISAC system, the module uses the near-end policy optimization (PPO) algorithm to iteratively update the GRU-enhanced policy network and value network, thereby obtaining a joint perception and communication control strategy suitable for dynamic UAV vehicle-to-everything (V2X) scenarios.

[0028] Example 3: This embodiment proposes an electronic device, including: one or more processors, and a memory, wherein the memory is used to store instructions, and when the instructions are executed by the one or more processors, the one or more processors execute the aforementioned method for controlling the confidentiality and freshness of unmanned aerial vehicle networking based on dual AoI.

[0029] The electronic device can be a mobile phone, computer, or tablet computer, etc., and includes a memory and a processor. The memory stores a computer program, which, when executed by the processor, implements a method for controlling the confidentiality and freshness of unmanned aerial vehicle (UAV) vehicle networking based on dual AoI as described in the embodiment. It is understood that the electronic device may also include input / output (I / O) interfaces and communication components.

[0030] The processor is used to execute all or part of the steps in the dual-AoI-based UAV vehicle-to-everything (V2X) confidentiality and freshness control method described in the above embodiments. The memory is used to store various types of data, which may include, for example, instructions for any application or method in the electronic device, as well as application-related data.

[0031] The processor can be implemented as an Application Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Programmable Logic Device (PLD), Field Programmable Gate Array (FPGA), controller, microcontroller, microprocessor, or other electronic components, and is used to execute the dual AoI-based UAV vehicle networking confidentiality freshness control method described in the above embodiments.

[0032] Example 5: This embodiment proposes a computer-readable storage medium that stores executable instructions. When these instructions are executed, if they are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium.

[0033] The computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the dual AoI-based drone vehicle networking confidentiality freshness control method described in various embodiments of this application.

[0034] The aforementioned storage media include: flash memory, hard disks, multimedia cards, card-type memory (e.g., SD (Secure Digital Memory Card) or DX (Memory Data Register, MDR) memory), random access memory (RAM), static random-access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic storage, disks, optical discs, servers, APP (Application) app stores, and other media capable of storing program verification codes. These media store computer programs, which, when executed by a processor, can implement the various steps of the aforementioned dual-AoI-based UAV vehicle-to-everything (UAV) security and freshness control method.

[0035] Example 6: This embodiment proposes a computer program product, including a computer program or instructions, which, when executed by a processor, implements the aforementioned method for controlling the confidentiality and freshness of unmanned aerial vehicle (UAV) vehicle networks based on dual AoI.

[0036] Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or part of the technical solution, can be embodied in the form of a computer program product.

[0037] The various embodiments in this application are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0038] The scope of protection of this application is not limited to the embodiments described above. Obviously, those skilled in the art can make various modifications and variations to this disclosure without departing from the scope and spirit of this disclosure. If such modifications and variations fall within the scope of this disclosure and its equivalents, then the intent of this disclosure also includes these modifications and variations.

Claims

1. A method for controlling the confidentiality and freshness of unmanned aerial vehicle (UAV) vehicle-to-everything (V2X) networks based on dual AoIs, characterized in that, Includes the following steps: A drone-assisted security ISAC system is constructed, and an initial scenario model and control strategy constraints for the drone-assisted security ISAC system are established, and initialization parameters are set. The control strategy constraints include flight constraints for legitimate drones, flight trajectory constraints for mobile eavesdropping drones, sensing power constraints, communication power constraints, common stream transmission constraints, and remaining energy constraints. Recent acquisition of drone-assisted confidentiality ISAC system Frame observation vectors, and construct from the nearest Temporal state composed of frame observation vectors ; A gated cyclic unit-enhanced policy network and value network are constructed. Based on the initial scenario model, control policy constraints, and initialization parameters of the UAV-assisted confidential ISAC system, the GRU-enhanced policy network and value network are iteratively updated using the near-end policy optimization (PPO) algorithm to obtain a joint perception and communication control policy suitable for dynamic UAV-vehicle network scenarios. The gated cyclic unit-enhanced policy network is used to adjust the policy network according to the time-series state. Output the original continuous action vector The value network enhanced by the gated loop unit is used to determine the time-series state. Output the current state value .

2. The method for controlling the confidentiality and freshness of unmanned aerial vehicle (UAV) vehicle networking based on dual AoIs according to claim 1, characterized in that, The construction of the UAV-assisted security ISAC system, the establishment of the initial scenario model and control strategy constraints of the UAV-assisted security ISAC system, and the setting of initialization parameters include the following specific steps: A1: Construct a drone-assisted security ISAC system for vehicle-to-everything (V2X) status update scenarios, including a legitimate drone, a mobile eavesdropping drone, and... There are 10 ground vehicles, of which the legal drones are equipped with M antennas, while all ground vehicles and mobile eavesdropping drones are equipped with a single antenna. The set of ground vehicles is represented as: （1）； in, Indicates a collection of ground vehicles; A2: Discretize the task cycle of the UAV-assisted security ISAC system into... In each time slot, establish the initial scenario model and control strategy constraints of the UAV-assisted confidential ISAC system; The length of the time slot is Each time slot is divided into a first half perception phase and a second half communication phase. In the first half perception phase, a legitimate UAV performs state perception on vehicles on all ground, completing the perception execution. In the second half communication phase, a legitimate UAV uses rate split multiple access to perform confidential state update communication on vehicles on all ground, completing the communication execution. The initial scenario model of the UAV-assisted confidentiality ISAC system includes the following: A legal drone flies at a fixed altitude, and its first... Position vector of each time slot : （2）； in, Indicates the first A time-slot-based legal UAV in a horizontal two-dimensional coordinate system Position components along the axial direction; Indicates the first A legal UAV in a time slot travels along a horizontal two-dimensional coordinate system. Position components along the axial direction; Indicates the time slot index; Indicates the flight altitude of a legal drone; Indicates transpose; the horizontal two-dimensional coordinate system, The axial direction is defined as the direction along which the road extends. The axial direction is defined as perpendicular to the direction of road extension; A legal drone flies at a fixed altitude, and its first... The position vector update relationship for each time slot is as follows: （3）； （4）； in, Indicates that legal drones are in the Position vector of each time slot Indicates that legal drones are in the Position increment within a time slot; Indicates that legal drones are in the Flight speed per time slot Indicates that legal drones are in the The movement direction angle of each time slot; A legal drone flies at a fixed altitude, and its first... The position vector update relationship for each time slot is as follows: （3）； （4）； in, Indicates that legal drones are in the Position vector of each time slot Indicates that legal drones are in the Position increment within a time slot; Indicates that legal drones are in the Flight speed per time slot Indicates that legal drones are in the The movement direction angle of each time slot; The mobile eavesdropping drone flies along a circular trajectory at a fixed altitude, and its trajectory is on the... Position vector of each time slot Represented as: （5）； in, Indicates the first A mobile eavesdropping drone in a time slot travels along a horizontal two-dimensional coordinate system. Position components in the axial direction Indicates the first A mobile eavesdropping drone in a time slot travels along a horizontal two-dimensional coordinate system. Position components in the axial direction The flight altitude of the mobile eavesdropping drone; No. The ground vehicle in Position vector of each time slot Represented as: （6）； No. The ground vehicle in The position vector update relationship for each time slot is represented as follows: （7）； in, Indicates the first The lane where each ground vehicle is located is in a horizontal two-dimensional coordinate system. Position components in the axial direction Indicates the first The first time slot The lane where each ground vehicle is located is along the horizontal two-dimensional coordinate system. Position components in the axial direction Indicates the first The ground vehicle in The speed of travel in each time slot; The flight constraints for the legal drones are as follows: （8）； （9）； in, This indicates the maximum permissible flight speed for a legal drone during flight. Indicates the feasible service area for legal drones; The flight trajectory constraints of the mobile eavesdropping drone are as follows: （10）； （11）； （12）； in, and This indicates the center coordinates of the circular trajectory of the mobile eavesdropping drone. This represents the orbital radius of the circular trajectory of a mobile eavesdropping drone. This indicates the initial phase of the mobile eavesdropping drone. This indicates the period of the mobile eavesdropping drone's flight along a circular trajectory. This represents the angular velocity of the mobile eavesdropping drone as it flies along a circular trajectory. The sensing power constraint is: （13）； in, Indicates the first The first time slot faces the first The sensing beam vector of a ground vehicle This indicates the transmit power corresponding to the sensing beam vector. This indicates the maximum transmit power that a legal drone is allowed to use within a single time slot; The communication power constraint is: （14）； in, Indicates the first The common stream transmit beam vector of each time slot, Indicates the first The first time slot faces the first Private stream transmit beam vector of a ground vehicle Indicates the common stream transmit power. Indicates the first The transmit power of the private stream corresponding to each ground vehicle; The common stream transmission constraint is: （15）； （16）； in, Indicates the first The time slot is allocated to the first The proportion of public flow resources for each ground vehicle; the remaining energy constraint is: （17）； in, The legal drone LU is in the Total energy consumption within each time slot This indicates the maximum initial energy budget pre-set for a legal drone; The total energy consumption includes flight propulsion energy consumption, sensing phase launch energy consumption, and communication phase launch energy consumption, namely: （18）； in, Indicates the first Flight propulsion energy consumption per time slot, Indicates the first Energy consumption of transmission during the sensing phase of each time slot Indicates the first The energy consumption of the communication phase in each time slot; the remaining energy constraint is used to ensure that the cumulative energy consumption of the legitimate UAV during the entire mission cycle does not exceed its maximum initial energy budget. A3: Set initialization parameters, including setting the initial values ​​of dual AoI states, the initial remaining energy budget of legal UAVs, course phase variables, and perception and communication related parameters; The initial values ​​of the dual AoI states include the initial values ​​of the sensing AoI state and the initial values ​​of the communication AoI state; Definition of the first The ground vehicle in The perceived AoI state value and the communication AoI state value for each time slot are respectively and Initial value of AoI state perception Initial values ​​of AoI communication state Set to: （19）； （20）； in, Indicates the first The AoI state value of a ground vehicle at the initial moment. Indicates the first The communication AoI state value of each ground vehicle at the initial moment; definition The legality of drones is indicated in the first... The remaining energy budget for each time slot will be used to determine the initial remaining energy budget for legitimate drones. Set to: （21）； Set course stage variables ,in, Indicates the training round index; when hour, =1 indicates that it is in the first course stage; when hour, =2 indicates that it is in the second course stage; when hour, =3 indicates that the student is in the third stage of the curriculum; among which... and This indicates the preset threshold for dividing the course into stages; Configure sensing and communication related parameters, including sensing thresholds. Public flow confidentiality threshold Private Stream Confidentiality Threshold Perceived AoI weights And communication AoI weights and maximum transmission power .

3. The method for controlling the confidentiality and freshness of unmanned aerial vehicle (UAV) vehicle networking based on dual AoIs according to claim 1, characterized in that, The recently acquired drone-assisted confidentiality ISAC system Frame observation vectors, and construct from the nearest Temporal state composed of frame observation vectors The specific steps include: B1: In the In each time slot, the current environmental observation information of the UAV-assisted confidential ISAC system is collected to construct a single-frame observation vector. ; Single-frame observation vector Defined as: （22）； in, Represents the observation subvector of a legal UAV. Indicates the first The observation subvector of a ground vehicle. This represents the observation subvector of the mobile eavesdropping drone; The observation subvector of a legitimate UAV is defined as: （23）； No. The observation subvector of each ground vehicle is defined as: （24）； in, Indicates that legal drones are in the The first time slot for the first Estimates of the location of each ground vehicle; Indicates the first The ground vehicle in The successful perception status of each time slot; Indicates the first The time slot is allocated to the first The common flow transmission rate of each ground vehicle; Indicates the first The ground vehicle in Each time slot corresponds to the secure transmission rate of the private stream; definition Indicates the first The first time slot for the first A successful perception indicator for a ground vehicle; when At, it indicates the time. The first time slot for the first One ground vehicle successfully sensed the situation; when At, it indicates the time. The first time slot for the first One ground vehicle failed to sense the incident; a legitimate drone was in the [missing information]. The first time slot for the first Estimated location of each ground vehicle The update rule is: when hour, ;when hour, ,in, Indicates according to the first The position increment is predicted by the historical position, speed and direction of motion information of a ground vehicle in the previous time slot; The observation subvector of a mobile eavesdropping drone is defined as: （25）； in, Indicates that mobile eavesdropping drones are in the first Each time slot corresponds to the received signal-to-interference-plus-noise ratio of the common stream. Indicates that mobile eavesdropping drones are used against the first The ground vehicle in Each time slot corresponds to the received signal-to-interference-plus-noise ratio of the private stream; B2: Based on single-frame observation vectors Construction by the most recent Temporal state composed of frame observation vectors ; recent The temporal state composed of frame observation vectors is as follows: （26）； If the dimension of the single-frame observation vector is Then we have: （27）； in, Indicates the legal drone observation dimensions. This represents the observation dimension of a single ground vehicle. This indicates the observation dimension of the mobile eavesdropping drone.

4. The method for controlling the confidentiality and freshness of unmanned aerial vehicle (UAV) vehicle networking based on dual AoIs according to claim 1, characterized in that, The process involves constructing a gated recurrent unit (GRU)-enhanced policy network and value network. Based on the initial scenario model, control policy constraints, and initialization parameters of the UAV-assisted secure ISAC system, the GRU-enhanced policy network and value network are iteratively updated using the near-end policy optimization (PPO) algorithm to obtain a perception and communication joint control strategy suitable for dynamic UAV-vehicle networking scenarios. Specific steps include: C1: Construct a policy network and value network enhanced by gated recurrent units; The policy network and value network enhanced by the gated recurrent unit both adopt a neural network structure containing a GRU temporal coding layer and a fully connected output layer; C2: A policy network enhanced with gated recurrent units based on temporal states. Output the original continuous action vector ; definition Indicates the first The hidden state of the GRU temporal coding layer in the nth slot, let the nth slot be... The hidden state update relationship corresponding to the frame observation vector input into the GRU temporal coding layer is as follows: （28）； in, Indicates the frame number and , Indicates the GRU time-series coding layer. Indicates the first The frame observation vector is input into the GRU temporal coding layer and then outputs the corresponding hidden state. The observation vector from the last frame is input into the GRU temporal coding layer, and the corresponding hidden state is output. This serves as a temporal feature representation of the current time slot, and is further mapped to a shared feature vector: （29）； in, Represents a shared feature mapping function; Fully connected output layer based on shared feature vectors Output the original continuous action vector ; The original continuous action vector Defined as: （30）； in, This represents the normalized flight speed of a legitimate drone. Indicates the amount of adjustment to the flight direction of a legal drone; Indicates facing the first Common flow allocation parameters for each ground vehicle; Represents the original variable of total power during the sensing phase; Indicates facing the first Power allocation parameters for the perception phase of a ground vehicle; Indicates the first Ground vehicle sensing beam control parameters; Represents the original variable of total power during the communication phase; The original variable representing the proportion of common flow power; Indicates the common beam control parameters; Indicates facing the first The original variables for private flow power allocation of each ground vehicle; Indicates the first Private flow beam control parameters for each ground vehicle; C3: Output the current state value using a value network enhanced by a gated loop unit. It is used for advantage function estimation in the Proximal Policy Optimization (PPO) algorithm and parameter updates of policy and value networks enhanced by gated recurrent units. The state value is expressed as follows: （31）； in, Represents the value mapping function; C4: Based on the course stage of the current training round The original continuous action vector is subjected to a phased action mask and a feasible mapping to obtain a feasible control quantity that satisfies the control strategy constraints. C5: Based on feasible control variables that satisfy control strategy constraints, in the first half of the time slot's perception phase, state perception is performed on vehicles on the ground in each location, and the success of the perception execution is determined to obtain a perception success flag. ; Legitimate drones utilize sensing beams during the first half of the sensing phase in each time slot. For the first For a ground vehicle, the interference and noise term during the perception phase... Defined as: （48）； in, Indicates noise power; express The conjugate transpose of; The corresponding sensing signal-to-interference-plus-noise ratio is defined as: （49）； When the The perceived signal-to-interference-plus-noise ratio of each ground vehicle meets the perception threshold requirement. When the perception is successfully executed, it is determined that the result is obtained. When the first The perceived signal-to-interference-plus-noise ratio of each ground vehicle meets the perception threshold requirement. When the perception execution fails, it is determined that the result is... ; C6: Based on feasible control variables, in the latter half of the communication phase, perform confidentiality state update communication based on RSMA and determine whether the public and private streams were successfully transmitted during the confidentiality state update communication process to obtain the communication success flag. ; During the latter half of the communication phase of each time slot, the legitimate UAV uses RSMA to perform secure state update communication and transmits signals. Represented as: （50）； in, Indicates a common flow signal. Indicates the first Private flow signal of a ground vehicle; For the For each ground vehicle, the received signal-to-interference-plus-noise ratio (SIR) for public and private flows are as follows: （51）； （52）； in, Indicates the first The received signal-to-interference-plus-noise ratio corresponding to the public flow of ground vehicles; Indicates the first The received signal-to-interference-plus-noise ratio corresponding to the private flow of each ground vehicle; Indicates the first Interference and noise terms during the decoding of public flow by ground vehicles Indicates the first Interference and noise terms when decoding private streams by ground vehicles; Public flow total confidentiality rate Defined as: （53）； in, Indicates the common stream transmission rate. This indicates the signal-to-interference-plus-noise ratio (SIR) received by a mobile eavesdropping drone over a public stream. This indicates the operation of obtaining a non-negative value; Assigned to the Public flow confidentiality rate of individual ground vehicles for: （54）； No. The private flow confidentiality rate of a ground vehicle is defined as: （55）； in, Indicates that mobile eavesdropping drones are used against the first The received signal-to-interference-plus-noise ratio of a private flow of ground vehicles; when At that time, it was believed that the first The public flow transmission of ground vehicles was successful; when At that time, it was believed that the first The private stream transmission of a ground vehicle is successful; the determination is made only if both decoding of the public stream and decoding of the private stream are successful. Communication with the ground vehicle was successfully executed; successful communication is indicated by... Otherwise ; C7: Based on perceived success indicators and communication success indicator Update the perceived AoI status value and the communication AoI status value, remaining energy and instant reward value; Update # The perceived AoI state value of each ground vehicle is: （56）； Update # The communication AoI status value of each ground vehicle is: （57）； Remaining energy Updated to: （58）； C8: Value based on the current state Perform advantage function estimation and update the parameters of the GRU-enhanced policy network and value network; C9: Increment the current training round by 1 and determine whether the preset maximum number of training rounds has been reached. If so, end the training and output the final perception and communication joint control strategy suitable for dynamic UAV vehicle networking scenarios, consisting of feasible control quantities that satisfy the control strategy constraints. If not, determine whether to switch the course stage. If to switch, switch the course stage and return to C2; otherwise, return directly to C2.

5. The method described in claim 4, which constructs a gated recurrent unit-enhanced policy network and value network, and based on the initial scenario model, control policy constraints, and initialization parameters of the UAV-assisted confidentiality ISAC system, iteratively updates the GRU-enhanced policy network and value network using the near-end policy optimization (PPO) algorithm to obtain a perception and communication joint control strategy suitable for dynamic UAV vehicle-to-everything (V2X) scenarios, is characterized in that... The specific steps for C4 include: C4.1: Based on the course stage to which the current training round belongs. The original continuous action vector is subjected to a phased action masking and then decoded to obtain the decoded continuous action vector. Let the total dimension of the original continuous actions be... Define the set of actions that can be released in each course stage as follows: and And satisfy: （32）； The phased action mask vector is further defined as follows: （33）； in, Reflects the first The action dimension in the first... Whether it is activated in each training round; when hour, , indicating the first The action dimension in the first... Activated in one training round; otherwise , indicating the first The action dimension in the first... Not activated in any training round; When the The action dimension in the first... When activated in the training round, the decoded action component is equal to the original continuous action vector. The action dimension in the first... The motion components in each training round; when the first... The action dimension in the first... If the action component is not activated in a training round, the decoded action component is equal to the default action value for that training phase. ; C4.2: Perform a feasible mapping on each action component in the decoded continuous action vector to obtain feasible control quantities that satisfy the control policy constraints, including legal UAV maneuver control quantities, perception control quantities, and communication control quantities; The legal drone maneuver control quantities include the legal drone in the... Flight speed in time slots and legal drones in the first time slot The movement direction angle of each time slot; Map the normalized flight speed of the legitimate drone to the flight speed of the legitimate drone in the first... Flight speed per time slot: （34）； Map the flight direction adjustment amount of the legitimate drone to the legitimate drone in the [missing information]. The movement direction angle of each time slot: （35）； in, This indicates that the target range mapped by the legitimate drone flight direction adjustment is... ; The sensing control quantity includes the total power during the sensing phase, and the power directed towards the first... The power ratio of the perception phase for each ground vehicle and the first Sensing beams for a ground vehicle; Map the original variable of total power in the sensing phase to the total power in the sensing phase: （36）； in, The total power during the sensing phase. Represents the Sigmoid mapping function; Will be directed towards the first The power allocation parameters for the perception phase of the ground vehicle are mapped to the first... Power percentage of each ground vehicle in the perception phase: （37）； （38）； in, Indicates facing the first The percentage of power allocated to the perception phase for each ground vehicle. Indicates assignment to the first The actual perceived power of a ground vehicle. Indicates facing the first Sensing power allocation parameters for each ground vehicle For the index variable in the normalization summation process; The first The control parameters of the ground vehicle sensing beam are mapped to the first... Sensing beams for each ground vehicle: （39）； in, Indicates the first Sensing beams for each ground vehicle. Indicates the first Legal channel directions for each ground vehicle ， This represents the scaling factor for the sensing beam residual. This indicates a normalization operation; The communication control quantities include the total power during the communication phase, the first... The allocation ratio of common current power to total power in each time slot communication phase, common beam, and orientation towards the first time slot communication phase. Private current power ratio, private current power and corresponding number of ground vehicles Private beams for each ground vehicle; Map the original variable of total power during the communication phase to the total power during the communication phase: （40）； in, Total power during the communication phase; Will be directed towards the first The common flow allocation parameters for the ground vehicle are mapped to the parameters allocated to the first ground vehicle. The actual public flow proportion of each ground vehicle is used to characterize the allocation to the first... The share of public flow resources for each ground vehicle: （41）； Map the original variable of common current power ratio to the first The proportion of common stream power in the total power of the communication phase in each time slot: （42）； in, For the first The proportion of common stream power in each time slot communication phase to the total power of the communication phase; Constructing common current power for: （43）； Common beam The construction is as follows: （44）； in, Indicates the average legal channel direction. This represents the scaling factor for the common beam residual; Will be directed towards the first The original variables for the private flow power allocation of the ground vehicle are mapped to the first... Private current power percentage of each ground vehicle: （45）； in, Indicates facing the first The percentage of private flow power for each ground vehicle Indicates facing the first The original variables for private flow power allocation of each ground vehicle; Construct the first Private current power of each ground vehicle for: （46）； Corresponding to the Private beam for each ground vehicle The construction is as follows: （47）； in, Indicates the first Private beam reference direction for each ground vehicle ， This represents the private beam residual scaling factor.

6. A dual-AoI-based UAV-vehicle network security freshness control system, used to implement the dual-AoI-based UAV-vehicle network security freshness control method according to any one of claims 1-5, characterized in that, include: The system initialization module is used to build the UAV-assisted confidentiality ISAC system, establish the initial scenario model and control strategy constraints of the UAV-assisted confidentiality ISAC system, and set the initialization parameters. The timing state construction module is used to obtain the latest information from the UAV-assisted confidentiality ISAC system. Frame observation vectors, and construct from the nearest Temporal state composed of frame observation vectors ; The control strategy module is used to construct a policy network and value network enhanced by gated recurrent units. Based on the initial scenario model, control strategy constraints and initialization parameters of the UAV-assisted confidential ISAC system, the GRU-enhanced policy network and value network are iteratively updated using the near-end policy optimization (PPO) algorithm to obtain a joint perception and communication control strategy suitable for dynamic UAV vehicle-to-everything (V2X) scenarios.

7. An electronic device, characterized in that, It includes one or more processors and a memory for storing instructions, which, when executed by the one or more processors, cause the one or more processors to perform the confidentiality and freshness control method for unmanned aerial vehicle networking based on dual AoI as described in any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, It stores executable instructions that, when executed, cause the processor to perform the confidentiality and freshness control method for unmanned aerial vehicle networking based on dual AoI as described in any one of claims 1-5.

9. A computer program product, characterized in that, Includes a computer program or instructions that, when executed by a processor, implement the confidentiality and freshness control method for unmanned aerial vehicle (UAV) vehicle networking based on dual AoI as described in any one of claims 1-5.

Images