Unmanned aerial vehicle deployment optimization method in unmanned aerial vehicle assisted wireless communication
By introducing adaptive mutations and cross operators into the differential evolution algorithm, the number and position of drone hover points are optimized, and the problem of difficulty in simultaneous optimization in the existing technology is solved, and high-energy-efficient communication is achieved.
Patent Information
- Application Number
- CN202410470965.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-04-18
- Publication Date
- 2024-11-01
AI Technical Summary
The existing drone deployment optimization technology is difficult to optimize the number and position of hover points at the same time, resulting in suboptimal deployment. Traditional methods are difficult to find the optimal solution when facing a variable-dimensional search space, and the introduced auxiliary variables increase individuals. length, may result in dimensional limitations.
Adaptive mutation and cross operators are introduced based on the traditional differential evolution algorithm, and the number and position of drone hover points are adjusted simultaneously by optimizing the objective function, and adaptive learning strategy selection and update method and dimension update strategy are adopted to dynamically adjust population size to find the optimal deployment.
A better drone deployment solution is realized, reducing the energy consumption of drones, improving communication efficiency, and avoiding the dimension limitations introduced by auxiliary variables in traditional methods.
Smart Images

Figure CN118890629A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of unmanned aerial vehicle (UAV)-assisted wireless communication, and in particular relates to a method for optimizing the deployment of UAVs in UAV-assisted wireless communication. Background Art
[0002] Since the base station cannot be changed after installation, it is very difficult for the base station to communicate directly with the remote user equipment, which will directly affect the service life of the system. Relays can be used to improve throughput and reliability and expand coverage. In wireless networks, relays can be divided into fixed relays and mobile relays. In fixed relays, the relay position is static, whereas in mobile relays, mobile relays with uncertain positions are used. Compared with traditional fixed relays, drones have the characteristics of high mobility, low cost, and deployment flexibility. As a mobile relay, drones have received widespread attention in recent years.
[0003] Despite the many advantages of drones, drone-assisted communications in relay networks have encountered challenges due to limited onboard energy. Specifically, drones require energy not only to communicate with ground users, but also to maintain flight, and the slow progress of battery technology has severely limited the service life of drones. Therefore, it is crucial to achieve energy-efficient communications in drone-assisted relay networks.
[0004] In order to effectively use drones as mobile relays to assist wireless communications, the deployment of drones must be optimized. If deployed properly, energy-efficient communication can be maximized. In fact, the deployment optimization problem of drones has been studied before. For example, He et al. studied the joint height and beamwidth optimization problem in downlink multicast, downlink broadcast, or uplink multiple access. Alzenad et al. optimized the deployment of drones with the goal of covering the maximum number of ground users with the minimum transmission power. Then the deployment optimization under different quality of service requirements was studied. Fan et al. studied the deployment of drones in a drone-assisted relay system, using drones as relays for data transmission between source nodes and destination nodes. Mozaffari et al. designed the optimal deployment of multiple drones for collecting data from IoT devices. Du et al. studied the deployment of drones in downlink communication aggregation scenarios. Yong Wang et al. studied the optimized deployment in drone-assisted IoT data collection systems.
[0005] In most existing drone deployment optimization technologies, the number of drone hovering points is preset, which means that these methods only optimize the position of the hovering points. However, if the preset number is not optimal, it is likely to lead to suboptimal deployment. If you want to optimize both the number and position of hovering points, since the number of hovering points is variable, the length of each individual in the population is not fixed. In this case, it is necessary to modify the traditional fixed-length mutation and crossover operators in evolutionary algorithms. To this end, the following attempts have been made in existing technologies: (1) Designing special mutation and crossover operators, but this method faces a variable-dimensional search space. Since search spaces of different dimensions have different optimal solutions, this will guide the search in a chaotic way. (2) Using auxiliary variables to convert variable-length individuals into fixed-length individuals. In this method, the length of each individual in the population is fixed, but a set of auxiliary variables are introduced to control the activation of the original variables. However, the introduction of auxiliary variables increases the length of the individual. Therefore, this method may suffer from dimensional limitations. Summary of the invention
[0006] The present invention provides a method for optimizing the deployment of drones in drone-assisted wireless communications. The main purpose of the present invention is to simultaneously optimize the number and positions of hovering points of drones, obtain a better deployment plan, and achieve energy-efficient communication.
[0007] Another object of the present invention is to introduce adaptive mutation and crossover operators based on the traditional differential evolution algorithm in the process of optimizing the number and position of the hovering points of the drone, so as to obtain better optimization effect and further reduce the energy consumption of the drone.
[0008] The technical solution provided by the present invention is:
[0009] A method for optimizing the deployment of drones in drone-assisted wireless communications comprises the following steps:
[0010] Step 1: Obtain the ground base station location, user location and drone flight trajectory;
[0011] Step 2: Determine the initial number of drone hovering points and the initial position of each drone hovering point;
[0012] Step 3: Determine the optimization objective function, and optimize the number of drone hovering points and the position of each drone hovering point according to the optimization objective function to obtain the optimal deployment plan;
[0013] The optimal deployment scheme includes: the optimal number of drone hovering points and the optimal position of each drone hovering point;
[0014] The optimization objective function is:
[0015] In the formula, Indicates the hovering energy consumption of the drone, Describes the energy consumption of UAV data transmission, φ represents the weight between the energy consumption of UAV hovering and the energy consumption of data transmission; They represent the horizontal and vertical coordinates of the drone at the hovering point m, respectively, and M represents the number of hovering points;
[0016] Step 4: The drone hovers according to the optimal deployment plan, receives base station data and transmits data to the user.
[0017] Preferably, in step 2, the initial number of drone hovering points is equal to the number of users.
[0018] Preferably, the hovering energy consumption of the drone is calculated by the following formula:
[0019]
[0020] Among them, P h Indicates the hovering power of the drone, P B and P I They represent the blade power and induced power of the UAV when it is in hovering state; represents the hovering time of the drone at the hovering point m, represents the time required for data to be transmitted from the base station to the drone hovering at the hovering point m, represents the time required for data to be transmitted from the drone at hovering point m to user k; a k,m The variable representing the association between user k and hover point m.
[0021] Preferably, the energy consumption of the drone data transmission is calculated by the following formula:
[0022]
[0023] in, represents the time required for data to be transmitted from the drone hovering at hovering point m to user k, P UAV,k represents the transmission power of the UAV to transmit data to user k, a k,m represents the association variable between user k and hover point m, and K represents the number of users.
[0024] Preferably, the time required for data to be transmitted from the base station to the drone hovering at the hovering point m is calculated by the following formula:
[0025]
[0026] in, is the signal-to-noise ratio of the drone at the hovering point m, is the transmission power of the base station to transmit data to the drone, B is the system bandwidth, and D k is the amount of data that the base station needs to transmit to user k; is the data transmission rate between the base station and the drone.
[0027] Preferably, the time required for data to be transmitted from the drone hovering at the hovering point m to the user k is calculated by the following formula:
[0028]
[0029] in, is the ratio of the channel to noise during the data transmission stage from the UAV to user k, is the transmission power of the UAV to transmit data to user k, is the data transmission rate between the UAV and user k.
[0030] Preferably, in step 3, the number of drone hovering points and the position of each drone hovering point are optimized at the same time, including the following steps:
[0031] Step 1, initialize the parent population;
[0032] Among them, the number of individuals in the initialized parent population is the initial number of drone hovering points, and each individual represents the coordinates of the initial position of a drone hovering point;
[0033] Step 2: Update each individual in the parent population one by one, and use the set of updated individuals as the offspring population;
[0034] Step 3: merge the first individual in the offspring population into the parent population as the first new population; replace a randomly selected individual in the parent population with the first individual in the offspring population, and use the replaced parent population as the second new population; randomly remove an individual from the parent population as the third new population;
[0035] If there is a feasible population among the first new population, the second new population and the third new population, the feasible population is evaluated, and the feasible population that has the greatest improvement on the optimization target relative to the parent population is selected as the new parent population;
[0036] Among them, if the feasible population has no improvement on the optimization target compared with the parent population, but the third population is a feasible population and has the same optimization target value as the parent population, then the third population is selected as the new parent population;
[0037] Repeat the process in step 3 for the second to last individuals in the offspring population to obtain the parent population for the next iteration;
[0038] Step 4: Repeat steps 2 to 3 until the iteration termination condition is met, and take the deployment plan corresponding to the parent population finally obtained as the optimal deployment plan.
[0039] Preferably, in step 2, an update method for each individual in the parent population is selected by using an adaptive learning strategy probability:
[0040] If p≤p1, the individual adopts the first strategy for update: v i =x r0 +F·(x r1 -x r2 );
[0041] Otherwise, the individual adopts the second strategy to update: i =x r0 +F·(x r1 -x r2 )+F·(x r3 -x r4 );
[0042] Where p is a random number between [0,1], p1 is the probability of using the first strategy for updating, and F is the scaling factor; v i is the individual x in the parent population i The new individual generated after the corresponding strategy update, x r0 、x r1 、x r2 、x r3 and x r4 is the parent population different from x i of individuals.
[0043] Preferably, during the first update, p1=0.5;
[0044] Each subsequent update of p1 is calculated by the following formula:
[0045]
[0046] Among them, s1 and s2 are the cumulative number of successful entry into the next generation using the first strategy and the second strategy before this update, respectively; f1 and f2 are the cumulative number of unsuccessful entry into the next generation using the first strategy and the second strategy before this update, respectively.
[0047] Preferably, in said step 2, it further comprises:
[0048] The update strategy for each dimension in each individual is selected through the adaptive strategy:
[0049]
[0050] Among them, u iis the individual entering the offspring population, j represents the dimension of the individual; CR is a random number that conforms to the normal distribution (CRa, 0.1), CRa is the judgment threshold, rand j is the random number for dimension determination, j rand is a random integer, j rand ∈[1,D], D is the dimension of the individual.
[0051] The beneficial effects of the present invention are:
[0052] The method for optimizing the deployment of drones in drone-assisted wireless communications provided by the present invention can simultaneously optimize the number and positions of the hovering points of the drones, obtain a more optimal deployment plan, and achieve energy-efficient communication.
[0053] The drone deployment optimization method in drone-assisted wireless communication provided by the present invention introduces adaptive mutation and crossover operators on the basis of traditional differential evolution algorithms in the process of optimizing the number and positions of the drone's hovering points, which can obtain better optimization effects and further reduce the energy consumption of the drone. BRIEF DESCRIPTION OF THE DRAWINGS
[0054] Figure 1 This is a schematic diagram of a wireless communication network scenario in which the drone described in the present invention acts as a mobile relay.
[0055] Figure 2 This is a flow chart of optimizing the number and positions of hovering points using an improved differential evolution algorithm according to the present invention.
[0056] Figure 3 This is a diagram of the population encoding method described in the present invention. DETAILED DESCRIPTION
[0057] The present invention is further described in detail below in conjunction with the accompanying drawings so that those skilled in the art can implement the invention with reference to the description.
[0058] The present invention is applied to wireless communication network scenarios using drones as mobile relays, such as Figure 1As shown, a base station (BS) is fixed on the ground, and many users (UE / User) are randomly distributed in a rectangular area centered on the base station. A rotary-wing UAV (UVA) flies at a fixed height around a fixed circular trajectory centered on the base station, and the UAV can hover at different hovering points on the fixed circular trajectory. At each hovering point, the UAV handles data reception and transmission tasks in two stages. The first stage receives the signal from the base station, and the second stage decodes and re-encodes the signal and sends it to the corresponding user. In order to minimize the transmission energy, each user selects the hovering point closest to the UAV to receive data. In this scenario, the UAV relay deployment optimization problem is formulated, and the energy consumption of the system is minimized by optimizing the deployment of the UAV (including the location and number of the UAV hovering points).
[0059] In this embodiment, the scenario is set that users are randomly distributed in a square area centered on the base station (the middle area between the inner square and the outer square, where the side length of the inner square is r b is 750m, and the side length of the outer square is r c =1500m), a rotor drone flies at a fixed height (height h is 100m) around the base station with a radius of r u Flying in a fixed circular trajectory.
[0060] The radius r of the fixed circular trajectory u The determination is related to the following factors:
[0061] Communication coverage requirements: The radius of the circular trajectory should be large enough to cover all ground users in the target communication area. The size of the communication coverage depends on the design requirements of the communication system and the distribution of ground users.
[0062] Signal strength attenuation: The signal transmission range of the drone is affected by signal attenuation, so the radius of the circular trajectory should take into account the signal strength attenuation factor to ensure that the communication signal can reach sufficient strength throughout the coverage area.
[0063] Flight altitude: When the drone flies in a circular trajectory around the base station, the flight altitude will also affect the radius of the circular trajectory. When the flight altitude is high, the radius of the circular trajectory can be appropriately reduced; conversely, when the flight altitude is low, the radius of the circular trajectory may need to be increased to ensure the integrity and stability of the coverage range.
[0064] Topographic and environmental factors: Environmental factors such as terrain undulations, buildings and obstacles can also affect the transmission of communication signals. These topographic and environmental factors need to be considered when determining the radius of the circular trajectory to ensure the stability and reliability of the communication signal.
[0065] Flight capabilities and resource limitations: The flight capabilities and resource limitations of the drone will also affect the radius of the circular trajectory. The drone’s endurance, flight speed, and other performance parameters need to be considered to determine the appropriate circular trajectory radius while ensuring that the drone can effectively perform its mission and maintain a stable communication link.
[0066] In this embodiment, the radius r of the fixed circular trajectory u Set to 450m.
[0067] like Figure 2 As shown in the figure, the entire deployment optimization method is to optimize the position and number of the hovering points of the UAV at the same time based on the improved differential evolution algorithm. The specific process is as follows.
[0068] (1) Initialize the population P (initialize the position and number of drone hovering points)
[0069] like Figure 3 As shown, the length of each individual in the population is fixed to 2, and each individual represents the coordinates of a hovering point (one dimension represents the horizontal coordinate and the other dimension represents the vertical coordinate). The population is recorded as P
[0070] The number of populations (number of hovering points) is equal to the number of ground users at the time of initialization, and the number of hovering points is dynamically optimized in the subsequent evolution process. Each individual (hovering point coordinate) is a randomly generated point on the circular trajectory set previously. After initializing the population P, it is necessary to verify whether the population satisfies the constraint that each hovering point can only communicate with N users at the same time. If not, the population P is regenerated until the constraint is satisfied.
[0071] The setting of N here needs to consider the following factors:
[0072] Limited communication resources: When a drone is used as a relay base station, its communication resources may be limited, such as limited wireless spectrum, bandwidth, etc. Therefore, in order to ensure communication quality and resource utilization efficiency, it is necessary to limit the number of users connected at the same time.
[0073] Signal interference: When a drone communicates with too many users at the same time, it may cause signal interference and mutual interference, affecting the communication quality and stability. Therefore, limiting the number of users connected at the same time can reduce the possibility of signal interference.
[0074] Processing capacity limitation: When the drone acts as a relay base station, it needs to handle communication requests from multiple users, data forwarding and other tasks. Due to the limited processing capacity of the drone, processing too many user requests at the same time may lead to performance degradation, or even communication delays or interruptions.
[0075] In this embodiment, N is set to 20.
[0076] (2) Generate the offspring population Q using adaptive mutation and crossover operators
[0077] In order to generate a new population Q, the adaptive learning strategy probabilistically selects one of the two learning methods (i.e., v i =x r0 +F·(x r1 -x r2 ) represents "DE / rand / 1" and v i =x r0 +F·(x r1 -x r2 )+F·(x r3 -x r4 ) represents "DE / rand / 2" and applies it to the current population P. The probability of each individual using the "DE / rand / 1" strategy is expressed as p1, and the probability of another strategy is expressed as p2=1-p1. In the above formula, x represents an individual in population P, and the indicators r0, r1, r2, r3, and r4 are unique integers selected freely from [1, NP] (NP represents the total number of individuals in population P) that are different from i, that is, x r0 、x r1 、x r2 、x r3 and x r4 is a species in population P different from x i individuals; F is the scaling factor, F is a random value between (0, 1) and obeys the normal distribution N(0.5, 0.3). Initially, set p1=p2=0.5; if p (p is a number between [0, 1] randomly generated each time) is less than or equal to p1, the individuals updated this time adopt the "DE / rand / 1" strategy; otherwise, the individuals updated this time use the "DE / rand / 2" strategy. After evaluating the newly generated trial vectors, the number of times the "DE / rand / 1" and "DE / rand / 2" strategies successfully enter the next generation are recorded as s1 and s2, respectively. The counts of trial vectors discarded by each strategy are marked as f1 and f2, respectively. These counts are accumulated in a defined algebra, called the learning period (LP). Subsequently, p1 and p2 are recalculated as follows:
[0078]
[0079] p2=1-p1
[0080] Among them, the probabilities of the two strategies are recalculated after LP. The values of s1, s2, f1, and f2 are reset at each update to avoid the side effects accumulated in the previous learning stage.
[0081] As a preferred method, the update strategy of each dimension in each individual is controlled by the parameter CR. The main strategies are as follows:
[0082]
[0083] u i are individuals that may enter the new population Q, j represents the dimension of the individual; u i The generation of x i (the i-th individual in population P) and v i (The aforementioned use of x i Related to the new individuals generated by the mutation strategy), specifically u i The generation of is shown in the above formula; where D is the dimension of the individual (in the present invention, D is 2, one dimension represents the horizontal coordinate, and the other dimension represents the vertical coordinate), where j needs to be operated from 1 to D according to the above formula, j rand is a random integer in [1,D] for all j, in order to ensure that u i At least one-dimensional data and x i different.
[0084] For each individual u i Generate different CRs. The generation rules are described below. Generate a random number rand for each j j belongs to [0,1].
[0085] We assume that CR is normally distributed in the range of mean CRa and standard deviation 0.1 (CR~N(CRa,0.1)).
[0086] Initially CRa is set to 0.5, and different CR values that conform to this normal distribution are generated for each current individual. The relevant CR values of each generation of test vectors that successfully enter the next generation are recorded in CRs. After the specified number of generations (i.e., the above-mentioned LP), the average value CRa of the CR normal distribution is recalculated based on the CR values corresponding to all recorded successful test vectors during this period (the values stored in CRs). The above process is repeated for the mean and standard deviation of this new normal distribution of 0.1 (CR~N(CRa,0.1), where CRa is the newly calculated mean). Please note that once the normal distribution mean is recalculated, the record of successful CR values (i.e., the values stored in CRs) will be cleared to avoid possible inappropriate long-term accumulation effects.
[0087] In the differential evolution algorithm, the control parameters CR and F are critical parameters, which follow a normal distribution, where CR~N(CRa,0.1). The CR values associated with the test vectors are recorded for each generation, and the CR values that are successfully advanced to the next generation are recorded in CRs. After a defined number of generations (i.e., the above LP), we recalculate the average CRa based on the values stored in CRs. It is worth noting that after recalculating the average, the data in CRs are deleted to prevent potential inappropriate accumulation effects. We combine the above learning strategy and parameter adaptation scheme into the DE crossover process.
[0088] (3) Use the offspring population Q to update the population P to generate new populations P1, P2, and P3
[0089] In the present invention, the main idea is to update at most one individual in P in each update, which not only dynamically changes the population size but also maintains a stable search. After the offspring population Q is generated, the first individual in Q is used to generate the following two new populations:
[0090] (a) Merge the first individual in Q into P to generate a new population P1.
[0091] (b) The first individual in Q replaces a randomly selected individual in P to produce another new population P2.
[0092] In addition, after randomly removing an individual from P, a third new population P3 is generated.
[0093] (4) Evaluate the feasibility of populations P1, P2, and P3
[0094] The association between the kth user and the drone at the mth hovering point is represented by the binary variable a k,m Indicates that a k,m =1 means that the drone at the mth hovering point transmits data to the kth user; otherwise, a k,m = 0. Each user can only communicate with one drone hovering point. Considering the limitation of system bandwidth, the drone can communicate with up to N users at each hovering point at the same time. Evaluate whether the population P1 P2 P3 meets the above constraints. If so, it means that the population is feasible, otherwise, it is not feasible.
[0095] (5) Evaluating the fitness of feasible populations
[0096] If there is at least one feasible population among P1 P2 P3, the fitness value of the feasible population is evaluated.
[0097] (a) Calculate the distance between the user and the drone’s hovering point
[0098]
[0099] Where (x k ,y k ,0) and They represent the coordinates of the k-th user and the m-th hovering point of the drone respectively.
[0100] (b) Calculate the channel power gain between the base station and the UAV and between the UAV and the user based on the probabilistic line-of-sight communication model
[0101]
[0102]
[0103] The reference channel coefficient at a distance of 1 meter is expressed as β0 = (4πf c / c) -2 , where c and f c are the speed of light and the carrier frequency respectively.
[0104] (c) Calculate the data transmission rate between the base station and the drone and between the drone and the user
[0105]
[0106]
[0107] The ratio of the channel to noise in the two stages at the hovering point m can be expressed as and The noise power in bandwidth B can be expressed as σ 2 =N0B, where N0 represents the noise power spectral density. From the above formula, we can see that the data transmission rate between the drone and the user is inversely proportional to the distance between them. In order to achieve energy-efficient communication, each user chooses the hovering point closest to the drone to receive data.
[0108] (d) Calculate the energy consumption of drone communication (data transmission energy consumption)
[0109] We assume that the base station transmits D to the kth terminal k The time required to transmit the data from the base station to the drone hovering at point m is The time required to transmit data from a drone hovering at point m to the kth user is When the drone receives data from the base station, it is mainly the base station that is working, so the energy consumption of the drone for receiving data is very small and can be ignored. Therefore, when the drone transmits data to the associated user, the main consideration is the energy consumption of the drone. The energy consumption of the drone for communication can be expressed as
[0110] (e) Calculate the drone hovering time
[0111] The drone hovers at each hovering point to transmit data to the relevant user. It will move to the next hovering point only after collecting the data sent by the base station and sending it to the corresponding user. Therefore, the hovering time of the mth hovering point consists of two components, namely the time to collect data and the time to transmit data. The hovering time of the drone at the mth hovering point can be expressed as
[0112]
[0113] (f) Calculation of hovering energy consumption
[0114]
[0115] Among them, P h Indicates the drone hovering power P h =P B +P I , where P B and P I They respectively represent the blade power and induced power in the hovering state, and their values are fixed.
[0116] Finally, the overall optimization objective function is designed as follows
[0117]
[0118] Among them, φ≥0 represents the weight between the drone hovering energy consumption and the data transmission energy consumption.
[0119] Based on the above objective function, the fitness of the feasible populations in P1 P2 P3 is calculated respectively, and then P is replaced with the feasible population that has the greatest improvement on the performance of P. Note that if there is no performance improvement, but the feasible fitness value of P3 is equal to the fitness value of P, then P is replaced with P3. This means that the individual removed from P has no effect on the performance; therefore, this individual is redundant.
[0120] The remaining individuals in Q cycle through this process one after another.
[0121] Repeat (2)-(5) until the iteration termination condition is met.
[0122] In this embodiment, the set termination condition is reaching a set number of iterations.
[0123] The proposed method (SaDEVPS) was compared with the method proposed by Yong Wang (using the mutation and crossover strategy of the traditional differential evolution algorithm):
[0124] The parameters are set as follows:
[0125] In our simulation, each algorithm will terminate after 100,000 fitness evaluations (i.e., the maximum number of iterations is 100,000), each algorithm will be run 30 times independently in each instance, and these results are the average of these iteration results. u , a circular trajectory with the center at (0,0,0), where r u Assuming it is 450m, the user is located in a positive direction-shaped area, whose inner edge length is r b , the outer edge length is r c , where r c Set to 1500m, r b Set to r c / 2. The flight altitude of the drone is 100 meters. In addition, Di(i∈K) is randomly distributed in [1,103]MB, N is set to 20, the communication bandwidth B is set to 1MHz, the spectral density value of the noise power is set to N0=-174dBm / Hz, P UAV,k =0.1W, P BS,UAV =1W
[0126] The performance of the algorithm is evaluated on 5 instances with different numbers of users, i.e., the number of users K = {100, 200, 300, 400, 500}.
[0127] The final experimental results are shown in Table 1.
[0128] Table 1
[0129]
[0130] It can be seen from Table 1 that under different user scales, the optimization method provided by the present invention can effectively reduce the energy consumption of the drone compared with the optimization method proposed by Yong Wang (using the mutation and crossover strategy of the traditional differential evolution algorithm).
[0131] Although the embodiments of the present invention have been disclosed as above, they are not limited to the applications listed in the specification and the implementation modes, and they can be fully applied to various fields suitable for the present invention. For those familiar with the art, additional modifications can be easily implemented. Therefore, without departing from the general concept defined by the claims and the scope of equivalents, the present invention is not limited to the specific details and the illustrations shown and described herein.
Claims
1. A method for optimizing drone deployment in drone-assisted wireless communication, characterized in that: The steps include: Step 1: Obtain the ground base station location, user location and drone flight trajectory; Step 2: Determine the initial number of drone hovering points and the initial position of each drone hovering point; Step 3: Determine the optimization objective function, and optimize the number of drone hovering points and the position of each drone hovering point according to the optimization objective function to obtain the optimal deployment plan; The optimal deployment scheme includes: the optimal number of drone hovering points and the optimal position of each drone hovering point; The optimization objective function is: In the formula, Indicates the hovering energy consumption of the drone, Describes the energy consumption of UAV data transmission, φ represents the weight between the energy consumption of UAV hovering and the energy consumption of data transmission; They represent the horizontal and vertical coordinates of the drone at the hovering point m, respectively, and M represents the number of hovering points; Step 4: The drone hovers according to the optimal deployment plan, receives base station data and transmits data to the user.
2. The method for optimizing drone deployment in drone-assisted wireless communication according to claim 1, characterized in that: In step 2, the initial number of drone hovering points is equal to the number of users.
3. The method for optimizing drone deployment in drone-assisted wireless communication according to claim 2, characterized in that: The hovering energy consumption of the drone is calculated by the following formula: Among them, P h Indicates the hovering power of the drone, P B and P I They represent the blade power and induced power of the UAV when it is in hovering state respectively; represents the hovering time of the drone at the hovering point m, represents the time required for data to be transmitted from the base station to the drone hovering at the hovering point m, represents the time required for data to be transmitted from the drone at hovering point m to user k; a k,m The variable representing the association between user k and hover point m.
4. The method for optimizing drone deployment in drone-assisted wireless communication according to claim 3, characterized in that: The energy consumption of the UAV data transmission is calculated by the following formula: in, represents the time required for data to be transmitted from the drone hovering at hovering point m to user k, P UAV,k represents the transmission power of the UAV to transmit data to user k, a k,m represents the association variable between user k and hover point m, and K represents the number of users.
5. The method for optimizing drone deployment in drone-assisted wireless communication according to claim 4, characterized in that: The time required for data to be transmitted from the base station to the drone hovering at the hovering point m is calculated by the following formula: in, is the signal-to-noise ratio of the drone at the hovering point m, is the transmission power of the base station to transmit data to the drone, B is the system bandwidth, and D k is the amount of data that the base station needs to transmit to user k; is the data transmission rate between the base station and the drone.
6. The method for optimizing the deployment of drones in drone-assisted wireless communications according to claim 5, characterized in that: The time required for data to be transmitted from a drone hovering at hovering point m to user k is calculated as follows: in, is the ratio of the channel to noise during the data transmission stage from the UAV to user k, is the transmission power of the UAV to transmit data to user k, is the data transmission rate between the UAV and user k.
7. The method for optimizing the deployment of drones in drone-assisted wireless communications according to any one of claims 1 to 6, characterized in that: In step 3, the number of drone hovering points and the position of each drone hovering point are optimized at the same time, including the following steps: Step 1, initialize the parent population; Among them, the number of individuals in the initialized parent population is the initial number of drone hovering points, and each individual represents the coordinates of the initial position of a drone hovering point; Step 2: Update each individual in the parent population one by one, and use the set of updated individuals as the offspring population; Step 3: merge the first individual in the offspring population into the parent population as the first new population; replace a randomly selected individual in the parent population with the first individual in the offspring population, and use the replaced parent population as the second new population; randomly remove an individual from the parent population as the third new population; If there is a feasible population among the first new population, the second new population and the third new population, the feasible population is evaluated, and the feasible population that has the greatest improvement on the optimization target relative to the parent population is selected as the new parent population; Among them, if the feasible population has no improvement on the optimization target compared with the parent population, but the third population is a feasible population and has the same optimization target value as the parent population, then the third population is selected as the new parent population; Repeat the process in step 3 for the second to last individuals in the offspring population to obtain the parent population for the next iteration; Step 4: Repeat steps 2 to 3 until the iteration termination condition is met, and take the deployment plan corresponding to the parent population finally obtained as the optimal deployment plan.
8. The method for optimizing drone deployment in drone-assisted wireless communication according to claim 7, characterized in that: In step 2, an adaptive learning strategy is used to probabilistically select an update method for each individual in the parent population: If p≤p1, the individual adopts the first strategy for update: v i =x r0 +F·(x r1 -x r2 ); Otherwise, the individual adopts the second strategy to update: i =x r0 +F·(x r1 -x r2 )+F·(x r3 -x r4 ); Where p is a random number between [0,1], p1 is the probability of using the first strategy for updating, and F is the scaling factor; v i is the individual x in the parent population i The new individual generated after the corresponding strategy update, x r0 、x r1 、x r2 、x r3 and x r4 is the parent population different from x i of individuals.
9. The method for optimizing drone deployment in drone-assisted wireless communication according to claim 8, characterized in that: During the first update, p1=0.5; Each subsequent update of p1 is calculated by the following formula: Among them, s1 and s2 are the cumulative number of successful entry into the next generation using the first strategy and the second strategy before this update, respectively; f1 and f2 are the cumulative number of unsuccessful entry into the next generation using the first strategy and the second strategy before this update, respectively.
10. The method for optimizing drone deployment in drone-assisted wireless communication according to claim 9, characterized in that: In step 2, it also includes: The update strategy for each dimension in each individual is selected through the adaptive strategy: Among them, u i is the individual entering the offspring population, j represents the dimension of the individual; CR is a random number that conforms to the normal distribution (CRa, 0.1), CRa is the judgment threshold, rand j is the random number for dimension determination, j rand is a random integer, j rand ∈[1,D], D is the dimension of the individual.
Citation Information
Cited By
Unmanned aerial vehicle satellite base station data transmission method and unmanned aerial vehicle satellite base station
CN119997035A
Unmanned aerial vehicle base station deployment method and system based on GNN and differential evolution
CN121619581A
