CN-118678325-B - Unmanned aerial vehicle hidden communication method with high capacity and low energy consumption

CN118678325BCN 118678325 BCN118678325 BCN 118678325BCN-118678325-B

Abstract

The invention discloses a high-capacity low-energy-consumption unmanned aerial vehicle covert communication method, which belongs to the technical field of unmanned aerial vehicle path planning and covert communication and comprises the following steps of S1, constructing a system model of unmanned aerial vehicle covert communication, S2, constructing an unmanned aerial vehicle covert communication model, S3, constructing a motion energy consumption model of an unmanned aerial vehicle, S4, finding a task target of the unmanned aerial vehicle to find the optimal transmitting power and navigation track based on the system model, the communication model and the motion energy consumption model, maximizing the detection error probability of an eavesdropper, constructing an optimization problem and constraint conditions based on the optimal transmitting power and the navigation track, and S5, solving the optimization problem by utilizing a multi-target depth deterministic strategy gradient algorithm MODDPG. According to the invention, on the premise that the communication link between the sender and the receiver is not detected by an eavesdropper, the best pareto of link throughput and unmanned aerial vehicle movement energy consumption is realized, and the security performance, the communication performance and the energy-saving performance are ensured for unmanned aerial vehicle task development.

Inventors

WANG JINGJING
BAI LIN
CHEN JIANRUI
WANG JIAXING
REN PENGFEI
ZHANG MENGYUAN

Assignees

北京航空航天大学

Dates

Publication Date: 20260508
Application Date: 20240430

Claims (1)

1. A high-capacity low-energy-consumption unmanned aerial vehicle concealed communication method is characterized by comprising the following steps of: s1, constructing a system model of unmanned aerial vehicle hidden communication; s2, constructing a concealed communication model of the unmanned aerial vehicle; s3, constructing a motion energy consumption model of the unmanned aerial vehicle; s4, based on the system model, the communication model and the motion energy consumption model, finding a task target of the unmanned aerial vehicle to find the optimal transmitting power and navigation track, maximizing the detection error probability of an eavesdropper, and constructing an optimization problem and constraint conditions according to the optimal transmitting power and navigation track; s5, solving the optimization problem by utilizing a multi-target depth deterministic strategy gradient algorithm MODDPG; in the system model of unmanned aerial vehicle covert communication, the starting point of the unmanned aerial vehicle is set as follows The position of the target point is The unmanned plane is positioned as And the location of the eavesdropper is And recording that the total time spent by the unmanned aerial vehicle from the starting point to the target point is Will take the total time Equally divided into Duration of time of Is used for the transmission of data in the very short time slots of (a), at each time slot The internal-view unmanned aerial vehicle moves linearly at uniform speed, in the first place When the time slots are used, the distance between the unmanned aerial vehicle and the ground base station is recorded as The distance between the unmanned plane and the eavesdropper is ; The unmanned aerial vehicle covert communication model specifically comprises: The method comprises the steps that a channel model between an unmanned aerial vehicle and a base station is set to be an LoS channel, the unmanned aerial vehicle is set to adopt a block fading channel in signal transmission, and the channel gain of unmanned aerial vehicle signals is kept unchanged in the same block; In the movement process of the unmanned aerial vehicle, the unmanned aerial vehicle can select whether to communicate with a ground base station or not in different time slots, and use Transmitting signals to ground base station on behalf of unmanned aerial vehicle by Indicating that the unmanned aerial vehicle is not in communication with the ground base station; when the drone communicates with the ground base station, the drone maps the transmitted message to a codeword Wherein For the number of channels used, in the first First received by the ground base station of each time slot The signals of the strip channel are: Wherein the method comprises the steps of 、 And Respectively represent the first In the first time slot The transmit power gain of the stripe channel, the transmit signal and the channel gain of the drone to the ground base station, In order to receive the noise signal at the point, And The size of (2) is respectively subjected to Gaussian distribution And ; The eavesdropper monitors the signal energy in the area by adopting the energy meter, and judges whether the unmanned aerial vehicle is communicating with the ground base station according to the signal-to-noise ratio of the received signal The first time slot eavesdropper receives The signals of the strip channel are: Wherein, the For noise signals at eavesdroppers, the magnitude follows a gaussian distribution ; The eavesdropper judges whether the unmanned aerial vehicle is communicating with the ground base station by detecting whether the signal-to-noise ratio of the received signal exceeds a set power threshold, if the signal-to-noise ratio is larger than the threshold, the unmanned aerial vehicle is judged to be transmitting signals to the ground base station, if the signal-to-noise ratio is smaller than the threshold, the unmanned aerial vehicle is judged to not transmit signals, the eavesdropper adopts a maximum likelihood ratio detection method to minimize the detection error, and the detection error is expressed as: Wherein the method comprises the steps of Is at the first An eavesdropper within a slot receives the sum of the signals from all channels, Is a detection threshold set by an eavesdropper; By relative entropy Constructing probability constraints for unmanned aerial vehicle communications not detected by eavesdroppers, wherein And Respectively is And Let the maximum likelihood function of the received signal of the lower eavesdropper be expressed as: ; . Relative entropy Representing And Assuming that the distance between the distribution probabilities of the signals received by the eavesdropper is reduced, the eavesdropper cannot distinguish whether the sender sends the signals or not, and therefore hidden communication is achieved; Assuming that the eavesdropper has unmanned aerial vehicle transmitting power And noise power Is a priori information of (1), then the optimal threshold value set by the eavesdropper The method comprises the following steps: Wherein, the Is the signal to noise ratio of the signal at the eavesdropper: ; the adoption of the relative entropy to construct probability constraint that unmanned aerial vehicle communication is not detected by an eavesdropper is as follows: Wherein the method comprises the steps of The method comprises the following steps: Wherein the method comprises the steps of Probability of being detected by an eavesdropper for unmanned aerial vehicle communication; the unmanned aerial vehicle's motion energy consumption model is as follows: in each time slot, the unmanned aerial vehicle is assumed to be in a quasi-static equilibrium state, and the speed of the unmanned aerial vehicle is assumed to be kept unchanged in each time slot; Record unmanned plane in the first The speed in each time slot is The propulsion energy consumption of the unmanned aerial vehicle is a linear sum of the horizontal propulsion energy consumption, the vertical propulsion energy consumption and the profile energy consumption related to the fluid resistance, wherein: First, the Horizontal propulsion energy consumption in a time slot Expressed as: Wherein the method comprises the steps of Is the weight of the unmanned aerial vehicle, And Representing the mass and gravitational acceleration of the unmanned aerial vehicle; Is the cross-sectional area of the unmanned aerial vehicle in the direction of motion, In order to achieve a mass density of air, For the length of each slot; First, the Vertical propulsion energy consumption in a time slot Expressed as: the profile energy consumption associated with the fluid resistance is: Wherein the method comprises the steps of Is the relationship between the fluid resistance of the unmanned aerial vehicle and the physical structure thereof, For the speed of the drone relative to air, Is the wind velocity, which is the wind speed, Representing the resistance coefficient; Unmanned plane No. Total energy consumption in a time slot The method comprises the following steps: Wherein, the , And ; The optimization problem and constraint conditions are as follows: And Two optimization objectives in planning a path for an unmanned aerial vehicle, respectively representing maximizing unmanned aerial vehicle communication throughput and minimizing unmanned aerial vehicle movement energy consumption, wherein , Constraint for decoding error probability Is a decoding error probability constraint at the receiver Limiting the transmitting power of each channel of the unmanned aerial vehicle, and restricting Limiting the concealment requirement between the unmanned aerial vehicle and the ground base station, and limiting the condition And Limiting the maximum displacement and speed variation of the drone within each time slot, respectively, wherein Representing the maximum acceleration that can be achieved; In step S5, the path planning and transmission power control of the unmanned aerial vehicle is modeled as a finite markov decision process MDP problem, the unmanned aerial vehicle relies on interactions with the environment to adjust its actions and learn the optimal strategy, its state space, action space and rewarding functions are as follows: state space: ; Action space: ; Bonus function: , wherein, And Corresponding to maximization of effective throughput and minimization of unmanned energy consumption, expressed as: And For two auxiliary bonus functions: The communication safety performance of the unmanned aerial vehicle is represented, Reflecting the penalty for longer path lengths; The 4 bonus functions are weighted differently and recorded as The complete prize is noted as: ; the MODDPG algorithm comprises two network structures, namely an Actor network and a Critic network, wherein the two networks are composed of an online network and a target network, and the online Actor network is formed by specifying a main strategy Mapping the observed state to an action, online reviewer network estimate Wherein And Is a parameter of two online networks; two target networks employing actor-critter architecture, target value is calculated by freezing parameters of the target networks before updating, parameters of the target networks And Copying from an online actor-criticizer network in an initialization stage, and randomly extracting a small batch of samples from an experience playback pool when updating network parameters; Converting elements of a bonus vector into a scalar weighted sum using a linear weighting method, the weighting being given by taking into account preferences between a plurality of targets, constraints Adding a time-varying attenuated noise to actor strategy Based on the transformation from the empirical playback pool, the policy objective function is: the step of optimizing the online commentator network is to first calculate a target value given by the online commentator network and a target value given by the online commentator network The difference between the values, then a gradient descent method is used to minimize the loss function, which is defined as the mean square error MSE of the difference: the optimization objective of the online critics network is to minimize MSE; using online comments given by home networks Value then calculate policies for online actor networks Is a gradient of (2): the optimization goal of the online actor network is to maximize the gradient; The MODDPG algorithm steps are as follows: S51, inputting weight parameter vector ; S52, randomly initializing online actor network parameters And online critics network parameters Initializing network parameters of a target actor And target critics network parameters : , Initializing experience playback pool Small batch size Discount factor Search for noise Learning factors for target actor networks and critics networks And ; S53, obtaining initial observation state ; S54, at each step, according to the current state and noise Select and implement actions ; S55, executing action And observe rewards The next time state ; S56, will be Store in experience playback pool ; S57 from Random decimation in Small lot data; S58, calculating an objective function value for each data : S59 by minimizing the loss function Updating online critics network parameters by maximizing policy gradients Updating network parameters of online actors; S510, updating the target network parameters: S511, exploring noise attenuation: ; s512, adding 1 to the step length, and returning to the step S54 until the step length is the maximum; s513, returning to the step S53 to retrain until the maximum training times are reached.

Description

Unmanned aerial vehicle hidden communication method with high capacity and low energy consumption Technical Field The invention belongs to the technical field of unmanned aerial vehicle path planning and covert communication, and particularly relates to a high-capacity and low-energy-consumption unmanned aerial vehicle covert communication method. Background To facilitate the construction of space-to-ground networks, next-generation space networks aided by Unmanned Aerial Vehicles (UAVs) have been widely used in a variety of scenarios, such as acquiring ecosystem monitoring data, emergency rescue, natural resource exploration, plant monitoring, providing navigational assistance, communication coverage, and the like. However, the fragile wireless communication environment (time-varying and open channels, strong ambient noise, severe signal attenuation) presents significant challenges to establishing a stable unmanned aerial vehicle assisted communication system. The unstable wireless communication link gives potential eavesdroppers the opportunity to intercept, in which case a malicious eavesdropper can easily detect the signal leakage of the unmanned aerial vehicle, obtain sensitive information, including control commands, task targets and transmission content. At present, most researches consider an anti-eavesdropping mechanism in a static scene, and do not consider eavesdropping threats suffered by the unmanned aerial vehicle when performing a flight task, and how to plan the flight track and a signal sending strategy of the unmanned aerial vehicle is a key problem for protecting the unmanned aerial vehicle from eavesdropping by an eavesdropper in the whole flight process. Existing path planning algorithms includeAlgorithms, fast random tree (Rapidly Exploring Random Tree, RRT) algorithms, dijkstra algorithms, genetic algorithms, ant colony algorithms, and the like. These commonly used path planning algorithms can only be used for realizing obstacle avoidance or shortest path, and cannot be applied to task scenes with higher constraint conditions and non-convex constraint conditions, such as unmanned energy consumption and communication throughput optimization. The prior hidden communication method comprises 1) hidden power control, 2) hidden waveform design, 3) hidden signal modulation and 4) hidden frequency/time jump. The hidden power control prevents eavesdropping by adaptively changing the transmission power to blend with noise on the eavesdropper's channel, the hidden waveform design includes Direct Sequence Spread Spectrum (DSSS) technology, etc., through bandwidth expansion, reduces the power spectral density to prevent eavesdropping by the eavesdropper, the hidden signal modulation hides the communication link by expanding the bandwidth, such as Orthogonal Frequency Division Multiplexing (OFDM), the design principle of the time/frequency hopping technology is to dynamically change the transmission frequency or time during communication, and a shared hopping pattern is employed between transceivers to prevent eavesdropping by the eavesdropper. While the above-described techniques provide different ways of implementing covert communications, they are not suitable for use in dynamic environments, particularly in establishing a high capacity secure link between a mobile drone and a ground base station. Disclosure of Invention Therefore, the invention aims to provide a high-capacity low-energy-consumption unmanned aerial vehicle concealed communication method, which aims to plan the track of an unmanned aerial vehicle and complete the tasks of detecting, surveying, measuring and the like of a target area. In order to achieve the above purpose, the present invention provides the following technical solutions: A high-capacity low-energy-consumption unmanned aerial vehicle hidden communication method comprises the following steps: s1, constructing a system model of unmanned aerial vehicle hidden communication; s2, constructing a concealed communication model of the unmanned aerial vehicle; s3, constructing a motion energy consumption model of the unmanned aerial vehicle; s4, based on the system model, the communication model and the motion energy consumption model, finding a task target of the unmanned aerial vehicle to find the optimal transmitting power and navigation track, maximizing the detection error probability of an eavesdropper, and constructing an optimization problem and constraint conditions according to the optimal transmitting power and navigation track; and S5, solving the optimization problem by utilizing a multi-target depth deterministic strategy gradient algorithm MODDPG. Further, in the system model of unmanned aerial vehicle covert communication, the position of the starting point of the unmanned aerial vehicle is set as followsThe position of the target point isThe unmanned plane is positioned asAnd the location of the eavesdropper isAnd recording that the total time spent by the u