An energy-saving and concealed communication method for three-dimensional wireless sensor networks based on drone assistance

Abstract

The present invention discloses an energy-saving covert communication method for a three-dimensional wireless sensor network assisted by an unmanned aerial vehicle, comprising the following steps: (1) determining channel information between sensor nodes and relay unmanned aerial vehicles of the three-dimensional wireless sensor network; (2) deriving power constraints for covert communication between sensor nodes and relay unmanned aerial vehicles in a scenario of limited code length; (3) determining the transmission power of the sensor nodes according to the constraints; (4) alternately optimizing wake-up scheduling variables and unmanned aerial vehicle trajectories according to a gradient descent algorithm to solve the energy required for the three-dimensional wireless sensor network; the present invention ensures the concealment and security of data during transmission and prevents data from being intercepted and cracked.

Description

Technical Field

[0001] The present invention relates to the field of communication and information engineering technology, and in particular to an energy-saving and concealed communication method of a three-dimensional wireless sensor network assisted by an unmanned aerial vehicle. Background Art

[0002] Wireless sensor networks have been widely used in environmental monitoring, military reconnaissance, and smart cities. However, traditional communication methods have problems such as high energy consumption, long transmission distance, and data leakage. Especially in the military and security monitoring fields involving sensitive information, data confidentiality is particularly important. Existing covert communication methods usually rely on complex encryption algorithms and protocols. Although they can provide a certain degree of security, they often consume a lot of computing resources and electricity, which is a huge challenge for sensor nodes with limited energy. In order to solve these problems, we need more efficient and energy-saving covert communication technologies to ensure the security of information transmission.

[0003] Covert communication technology aims to ensure the confidentiality and security of information during transmission and prevent data from being intercepted and cracked. Although traditional covert communication methods can provide a certain degree of security, they often rely on complex encryption algorithms and protocols, which consume a lot of computing resources and power. Therefore, how to effectively implement covert communication in sensor nodes with limited energy has become a major problem. Summary of the invention

[0004] Purpose of the invention: The purpose of the present invention is to provide an energy-saving and covert communication method for a three-dimensional wireless sensor network assisted by a drone, so as to solve the problem of low energy utilization efficiency of wireless sensor networks in covert communication.

[0005] Technical solution: The energy-saving and concealed communication method of a three-dimensional wireless sensor network assisted by a drone described in the present invention comprises the following steps:

[0006] (1) Determine the channel information between sensor nodes and relay drones in a 3D wireless sensor network;

[0007] (2) In the scenario of limited code length, the power constraints for covert communication between sensor nodes and relay drones are derived;

[0008] (3) Determine the transmit power of the sensor node based on the constraints;

[0009] (4) Based on the gradient descent algorithm, the wake-up scheduling variables and UAV trajectories are alternately optimized to solve the energy required for the three-dimensional wireless sensor network.

[0010] Furthermore, step (1) is as follows: Assume that the wireless sensor network consists of K sensor nodes, and the position of each sensor node k in the three-dimensional space is: ;

[0011] The initial position S and the final position E of the relay drone U are set as follows: , ; The location of the monitoring drone W is ; The maximum data receiving time of the relay drone U is T. The relay drone U is in motion during the time period t. The position of the relay drone U at this time is expressed as ; Assume that the channel coefficient between the relay drone U and the sensor network node k is Follow the large-scale path loss coefficient and the small-scale fading coefficient but ;

[0012] Assume that the large-scale path loss coefficient between the air-to-ground link for:

[0013] ;

[0014] in, and Represent the air-to-ground path loss index of LoS and NLoS respectively, represents the LoS path loss coefficient per unit length, represents the excess path loss coefficient of the NLoS channel, represents the distance between the relay UAV and the sensor network node k;

[0015] For small-scale fading, consider the Rice fading of LoS channel and the Rayleigh fading of NLoS channel. Described as:

[0016] ;

[0017] in, represents the Rice factor, represents the LoS channel component, is the NLoS Rayleigh fading component that satisfies the complex Gaussian distribution;

[0018] The probabilities of the LoS and NLoS components are then expressed as:

[0019] ;

[0020] in, and is the environmental parameter, is the pitch angle of sensor node k relative to the relay UAV U;

[0021] Assuming that the sensor node k is located in a hidden position in a high-rise building, the channel coefficient between the sensor node k and the relay drone U is It is subject to large-scale NLoS path loss and small-scale quasi-static Rayleigh fading, that is, Among them, , Satisfies complex Gaussian distribution;

[0022] The ground-to-air channel between the sensor node k and the monitoring drone W also experiences large-scale NLoS path loss and small-scale quasi-static Rayleigh fading, so the channel coefficient ,in , Satisfies complex Gaussian distribution.

[0023] Furthermore, step (2) is as follows: Assume that the state of sensor node k remains unchanged in the current time slot t, at a fixed transmission power In the case of sensor node k, After precoding, the signal received by the relay drone U is:

[0024] ;

[0025] in, represents the complex Gaussian white noise received at the relay drone U;

[0026] The transmission rate between the sensor node k and the relay UAV U is approximately calculated as:

[0027] ;

[0028] in, represents the maximum permissible decoding error probability of a given relay, is the inverse Q function, represents the signal-to-noise ratio SNR, n represents the code length of signal transmission; the effective transmission bit from sensor node k to relay drone U is:

[0029] , where n is the code length of signal transmission, represents the probability of decoding error; since the sensor node k has two states in each time slot t, namely, the sending state and , so the signal received by the monitoring drone is:

[0030] ;

[0031] in, represents the complex Gaussian white noise received by the monitoring drone at W;

[0032] According to the signal detection principle, when the average power of M received signals is greater than a specific threshold, it is determined that the sensor node has sent a signal, otherwise it has not sent a signal; the signal judgment of the monitoring drone W is given by the following formula:

[0033] ;

[0034] in, It represents the average power of the signal received by the listener. represents the monitoring threshold, and Respectively represent the decision of whether the signal has been sent or not. If it exceeds the threshold, it is judged as , otherwise ; According to the judgment results, the monitoring of UAV W produces two kinds of errors, namely false alarm and missed detection; the probabilities of false alarm and missed detection are respectively expressed as and To express;

[0035] According to the two error conditions, the total error detection probability is set to , assuming in, is an arbitrarily given small constant; the received signal is and The likelihood function formula is as follows:

[0036] ;

[0037] ;

[0038] in, It is the signal received by the monitoring drone;

[0039] Use KL divergence to constrain the total probability of false detection, that is

[0040] ;

[0041] in, represents the total variation distance between two probabilities, for and The KL divergence between them gives the constraint condition for measuring the concealment performance as:

[0042] .

[0043] Furthermore, the formula of step (3) is as follows:

[0044] ;

[0045] in, It is the power of the monitoring drone in order to maximize the effective transmission bits while meeting the constraints of covert communication.

[0046] Furthermore, step (4) is as follows: First, for K sensor nodes, let the wake-up schedule X and the wake-up schedule variable of each sensor node in time slot t be ,when When , the sensor node k is in the awake state; when When the sensor node k is in a silent state, the trajectory Q of the relay drone U is used in each time slot t. Then, the wake-up schedule X of the wireless sensor network and the trajectory Q of the relay drone are jointly optimized to minimize the energy consumption of the wireless sensor network while ensuring the target data volume. Can be sent and collected.

[0047] Furthermore, the joint optimization method is as follows: Set the minimum time interval to , the maximum distance that the relay drone U can move per unit time is The energy consumed by the sensor per unit time in the wake-up state is , the link spectrum efficiency is ,in is the maximum moving speed of the relay UAV U, and B is the channel bandwidth. By optimizing the wake-up scheduling variables and the trajectory of the relay UAV U, the initial optimization problem P1 is:

[0048] ;

[0049] Among them, P1 is the initial optimization problem, constraint C1 is the communication energy constraint, constraint C2 is the transmission rate constraint, constraint C3 is the wake-up state constraint, constraint C4 is the wake-up state binary constraint, and constraints C5 and C6 are the distance and initial position constraints of the relay drone U; the initialization problem is converted into a mathematical problem for solution.

[0050] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages: it can extend the service life of the wireless sensor network and solve the problem of high energy consumption in the traditional sensor network. At the same time, the covert communication between the designed wireless sensor network and the relay drone can ensure the concealment and security of the data during the transmission process, and prevent the data from being intercepted and cracked. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 is a schematic diagram of the present invention;

[0052] Figure 2 is a flow chart of the present invention;

[0053] Figure 3 is a flow chart of the covert communication algorithm of the present invention;

[0054] Figure 4 It is a flow chart of the gradient descent algorithm of the present invention. DETAILED DESCRIPTION

[0055] The technical solution of the present invention is further described below in conjunction with the accompanying drawings.

[0056] like Figure 1 As shown, an embodiment of the present invention provides an energy-saving and concealed communication method for a three-dimensional wireless sensor network assisted by a drone, comprising the following steps:

[0057] Step 1: Determine the channel information between the sensor nodes and the relay drone of the three-dimensional wireless sensor network. The details are as follows:

[0058] Assume that the wireless sensor network consists of K sensor nodes, and the position of each sensor node k in the three-dimensional space is ;

[0059] The initial position S and the final position E of the relay drone U are set to be: , ; The location of the monitoring drone W is ; The maximum data receiving time of the relay drone U is T. The relay drone U is in motion during the time period t. The position of the relay drone U at this time is expressed as ; Assume that the channel coefficient between the relay drone U and the sensor network node k is Follow the large-scale path loss coefficient and the small-scale fading coefficient but ;

[0060] Assume that the large-scale path loss coefficient between the air-to-ground link for:

[0061] ;

[0062] in, and Represent the air-to-ground path loss index of LoS and NLoS respectively, represents the LoS path loss coefficient per unit length, represents the excess path loss coefficient of the NLoS channel, represents the distance between the relay UAV and the sensor network node k;

[0063] For small-scale fading, consider the Rice fading of LoS channel and the Rayleigh fading of NLoS channel. Described as:

[0064] ;

[0065] in, represents the Rice factor, represents the LoS channel component, is the NLoS Rayleigh fading component that satisfies the complex Gaussian distribution;

[0066] The probabilities of the LoS and NLoS components are then expressed as:

[0067] ;

[0068] in, and is the environmental parameter, is the pitch angle of sensor node k relative to the relay UAV U;

[0069] For sensor node k, its purpose is to hide its information transmission. Assuming that the sensor node k is located in a hidden position in a high-rise building, the channel coefficient between sensor node k and relay drone U is It is subject to large-scale NLoS path loss and small-scale quasi-static Rayleigh fading, that is, Among them, , Satisfies complex Gaussian distribution;

[0070] The ground-to-air channel between the sensor node k and the monitoring drone W also experiences large-scale NLoS path loss and small-scale quasi-static Rayleigh fading, so the channel coefficient ,in , Satisfies complex Gaussian distribution.

[0071] Step 2: In the scenario of limited code length, derive the power constraints for covert communication between sensor nodes and relay drones. Figure 2 As shown in Figure 2, the details are as follows: Once a sensor node decides whether to transmit, the state of the sensor node remains unchanged in the current time slot t, so at a fixed transmission power In the case of sensor node k, the sensor node k pairs of signals After precoding, the signal received by the relay drone U is:

[0072] ;

[0073] in, represents the complex Gaussian white noise received at the relay drone U;

[0074] because The code length is finite, so considering the decoding error, the transmission rate between the sensor node k and the relay drone U is approximately calculated as:

[0075] ;

[0076] in, represents the maximum permissible decoding error probability of a given relay, is the inverse Q function, represents the signal-to-noise ratio SNR, and n represents the code length of signal transmission;

[0077] Due to the non-negligible decoding error probability of the finite code length, the effective transmission bits from the sensor node k to the relay drone U are:

[0078] , where n is the code length of signal transmission, represents the probability of decoding error; since the sensor node k has two states in each time slot t, namely, the sending state and , so the signal received by the monitoring drone is:

[0079] ;

[0080] in, represents the complex Gaussian white noise received by the monitoring drone at W;

[0081] According to the signal detection principle, when the average power of M received signals is greater than a specific threshold, it is determined that the sensor node has sent a signal, otherwise it has not sent a signal; the signal judgment of the monitoring drone W is given by the following formula:

[0082] ;

[0083] in, It represents the average power of the signal received by the listener. represents the monitoring threshold, and Respectively represent the decision of whether the signal has been sent or not. If it exceeds the threshold, it is judged as , otherwise ; According to the judgment results, the monitoring drone W produces two kinds of errors, namely false alarm and missed detection; false alarm means that the sensor node k actually did not send a signal, but the monitoring drone W misjudged the signal; missed detection means that the sensor node k sent a signal, but the monitoring drone W thought it did not send it. The probabilities of false alarm and missed detection are respectively expressed as and To express;

[0084] According to the two error conditions, the total error detection probability is set to , assuming in, is an arbitrarily given small constant; the received signal is and The likelihood function formula is as follows:

[0085] ;

[0086] ;

[0087] in, It is the signal received by the monitoring drone;

[0088] Use KL divergence to constrain the total probability of false detection, that is

[0089] ;

[0090] in, represents the total variation distance between two probabilities, for and The KL divergence between them gives the constraint condition for measuring the concealment performance as:

[0091] .

[0092] Step 3: According to the constraints, determine the transmit power of the sensor node. Figure 3 As shown, the details are as follows: According to the relevant theory of relative entropy, that is:

[0093] ;

[0094] The hidden constraints in this scenario can be obtained as follows:

[0095] ,

[0096] right about Find the first-order derivative:

[0097] ;

[0098] because ,so ,also Satisfy the requirement of decoding error probability, so ;

[0099] right about Find the first-order derivative: ;

[0100] From the above formula, we can know that: ,therefore about Monotonically increasing, the power of sensor node k The higher the effective transmission bit It will also be higher;

[0101] right about Find the first-order derivative:

[0102] ;

[0103] because , so the power of sensor node k Increase, will also increase, so it is necessary to reduce the power of sensor node k To better meet the covert communication constraints between sensor node k and relay UAV U;

[0104] Since the covert communication constraints are met and increase the effective transmission bit There is a conflict, and the effective transmission bit is increased as much as possible under the constraints of covert communication. So the power of sensor node k ;

[0105] Step 4: According to the gradient descent algorithm, alternately optimize the wake-up scheduling variables and the UAV trajectory to solve the energy required for the three-dimensional wireless sensor network. Figure 4 As shown, the details are as follows:

[0106] For K sensor nodes, the wake-up schedule X and the wake-up schedule variable of each sensor node in time slot t are set as ,when When , the sensor node k is in the awake state; when When t, the sensor node k is in a silent state; for the relay drone U, its trajectory Q is used in each time slot t. By jointly optimizing the wake-up schedule X of the wireless sensor network and the trajectory Q of the relay drone, the energy consumption of the wireless sensor network can be minimized while ensuring the target data volume. can be sent and collected; as follows:

[0107] The communication link between sensor node k and relay UAV U is a quasi-static block fading channel. The channel remains unchanged in each time block, so the minimum time interval is set to , the maximum distance that the relay drone U can move per unit time is , the energy consumed by the sensor per unit time in the wake-up state is , the link spectrum efficiency is ,in is the maximum moving speed of the relay UAV U, and B is the channel bandwidth;

[0108] By optimizing the wake-up scheduling variables and the trajectory of the relay drone U, the initial optimization problem P1 is set as:

[0109] ;

[0110] Among them, P1 is the initial optimization problem, constraint C1 is the communication energy constraint, constraint C2 is the transmission rate constraint, constraint C3 is the wake-up state constraint, constraint C4 is the wake-up state binary constraint, and constraints C5 and C6 are the distance and initial position constraints of the relay UAV U;

[0111] Since the optimization problem P1 is a mixed integer non-convex optimization problem, the gradient descent algorithm is used to transform the initial optimization problem into two intermediate optimization problems P2 and P3. The wake-up scheduling variable X of the sensor node and the trajectory Q of the relay drone U are optimized respectively, and multiple iterations are repeated until the conditions are met and then derived; the details are as follows:

[0112] First, relax the binary constraints of C4 to , and then the relaxed problem is iteratively solved based on the block coordinate descent technique. Therefore, for any given trajectory Q, the relaxed wake-up schedule variable X can be obtained by solving the linear programming P2:

[0113] ;

[0114] At the same time, for any given wake-up scheduling variable X, the trajectory Q of the drone is optimized to solve the optimization problem P3. The goal is to make the communication throughput of all sensor nodes reach the maximum weighted minimum:

[0115] ;

[0116] in, represents the weighted minimum value of the communication throughput that needs to be maximized. By solving P3, the trajectory Q of the relay drone U can be obtained; however, due to The non-convexity of leads to P3 being a non-convex problem. Therefore, we need to use the successive convex approximation algorithm to Perform a first-order Taylor expansion to obtain an effective approximate function , ensuring convergence to at least a local optimal solution, and solving P4 to obtain the solution of P3:

[0117] ;

[0118] because It decreases during the iteration process, so the result obtained by the successive convex approximation algorithm is convergent; therefore, the solution to the initial optimization problem P1 is obtained by solving problems P2 and P3 to alternately optimize the wake-up schedule X and trajectory Q.

Claims

1. An energy-saving and concealed communication method for three-dimensional wireless sensor networks assisted by drones. It is characterized in that The following steps are involved: (1) Determine the channel information between the sensor nodes and the relay drone of the three-dimensional wireless sensor network; the details are as follows: Assume that the wireless sensor network consists of K sensor nodes, and the position of each sensor node k in the three-dimensional space is: ; The initial position S and the final position E of the relay drone U are set as follows: , ; The location of the monitoring drone W is ; The maximum data receiving time of the relay drone U is T. The relay drone U is in motion during the time period t. The position of the relay drone U at this time is expressed as ; Assume that the channel coefficient between the relay drone U and the sensor network node k is Follow the large-scale path loss coefficient and the small-scale fading coefficient but ; Assume that the large-scale path loss coefficient between the air-to-ground link for: ； in, and Represent the air-to-ground path loss index of LoS and NLoS respectively, represents the LoS path loss coefficient per unit length, represents the excess path loss coefficient of the NLoS channel, represents the distance between the relay UAV and the sensor network node k; For small-scale fading, consider the Rice fading of LoS channel and the Rayleigh fading of NLoS channel. Described as: ； in, represents the Rice factor, represents the LoS channel component, is the NLoS Rayleigh fading component that satisfies the complex Gaussian distribution; The probabilities of the LoS and NLoS components are then expressed as: ； in, and is the environmental parameter, is the pitch angle of sensor node k relative to the relay UAV U; Assuming that the sensor node k is located in a hidden position in a high-rise building, the channel coefficient between the sensor node k and the relay drone U is It is subject to large-scale NLoS path loss and small-scale quasi-static Rayleigh fading, that is, Among them, , Satisfies complex Gaussian distribution; The ground-to-air channel between the sensor node k and the monitoring UAV W also experiences large-scale NLoS path loss and small-scale quasi-static Rayleigh fading, so the channel coefficient ,in , Satisfies complex Gaussian distribution; (2) In the scenario of limited code length, the power constraint conditions for covert communication between sensor nodes and relay drones are derived as follows: Assume that the state of sensor node k remains unchanged in the current time slot t, and the transmission power is fixed. In the case of sensor node k, After precoding, the signal received by the relay drone U is: ； in, represents the complex Gaussian white noise received at the relay drone U; The transmission rate between the sensor node k and the relay UAV U is approximately calculated as: ； in, represents the maximum permissible decoding error probability of a given relay, is the inverse Q function, represents the signal-to-noise ratio SNR, n represents the code length of signal transmission; the effective transmission bit from sensor node k to relay drone U is: , where n is the code length of signal transmission, represents the probability of decoding error; since the sensor node k has two states in each time slot t, namely, the sending state and , so the signal received by the monitoring drone is: ； in, represents the complex Gaussian white noise received by the monitoring drone at W; According to the signal detection principle, when the average power of M received signals is greater than a specific threshold, it is determined that the sensor node has sent a signal, otherwise it has not sent a signal; the signal judgment of the monitoring drone W is given by the following formula: ； in, It represents the average power of the signal received by the listener. represents the monitoring threshold, and Respectively represent the decision of whether the signal has been sent or not. If it exceeds the threshold, it is judged as , otherwise ; According to the judgment results, the monitoring of UAV W produces two kinds of errors, namely false alarm and missed detection; the probabilities of false alarm and missed detection are respectively expressed as and To express; According to the two error conditions, the total error detection probability is set to , assuming in, is an arbitrarily given small constant; the received signal is and The likelihood function formula is as follows: ； ； in, It is the signal received by the monitoring drone; Use KL divergence to constrain the total probability of false detection, that is ； in, for and The KL divergence between them gives the constraint condition for measuring the concealment performance as: ； (3) According to the constraints, determine the transmit power of the sensor node; the formula is as follows: ； in, It is the power of the monitoring drone to maximize the effective transmission bits under the constraints of covert communication; (4) According to the gradient descent algorithm, the wake-up scheduling variables and the UAV trajectory are optimized alternately to solve the energy required for the three-dimensional wireless sensor network; the details are as follows: First, for K sensor nodes, let the wake-up schedule X and the wake-up scheduling variable of each sensor node in time slot t be ,when When , the sensor node k is in the awake state; when When the sensor node k is in a silent state, the trajectory Q of the relay drone U is used in each time slot t. Then, the wake-up schedule X of the wireless sensor network and the trajectory Q of the relay drone are jointly optimized to minimize the energy consumption of the wireless sensor network while ensuring the target data volume. can be sent and collected; The joint optimization method is as follows: Set the minimum time interval to , the maximum distance that the relay drone U can move per unit time is The energy consumed by the sensor per unit time in the wake-up state is , the link spectrum efficiency is ,in is the maximum moving speed of the relay UAV U, and B is the channel bandwidth. By optimizing the wake-up scheduling variables and the trajectory of the relay UAV U, the initial optimization problem P1 is: ； Among them, P1 is the initial optimization problem, constraint C1 is the communication energy constraint, constraint C2 is the transmission rate constraint, constraint C3 is the wake-up state constraint, constraint C4 is the wake-up state binary constraint, and constraints C5 and C6 are the distance and initial position constraints of the relay drone U. The initialization problem is converted into a mathematical problem for solution, as follows: First, relax the binary constraints of C4 to , and then the relaxed problem is iteratively solved based on the block coordinate descent technique. For any given trajectory Q, the relaxed wake-up schedule variable X is obtained by solving the linear programming P2: ； At the same time, for any given wake-up scheduling variable X, the trajectory Q of the drone is optimized to solve the optimization problem P3 so that the communication throughput of all sensor nodes reaches the maximum weighted minimum: ； in, represents the weighted minimum value of the communication throughput that needs to be maximized. By solving P3, the trajectory Q of the relay drone U can be obtained; since The non-convexity of leads to P3 being a non-convex problem. Using the successive convex approximation algorithm, Perform a first-order Taylor expansion to obtain the approximate function , we can get the solution of P3 by solving P4: 。

Images