Ferry cruise implementation method based on unmanned aerial vehicle swarm intelligence

Abstract

The invention discloses an intelligent ferry cruise implementation method based on an unmanned aerial vehicle group, and aims to establish an unmanned aerial vehicle group ferry cruise network in order to solve the problem of how to realize information intercommunication and comprehensive integration among combat areas under the condition that a mobile base station in a field battlefield is damaged. And deploying the ferry unmanned aerial vehicle to cruise among the regions and communicate with the relay unmanned aerial vehicle in each region, releasing computing resources in real time or collecting data and uploading the data to the low-orbit satellite, and realizing the optimal total time delay of the ferry system. The method comprises the steps that a ferry unmanned aerial vehicle selects a path through the concentration of volatile information factors, the information factors on a high-quality path are accumulated step by step after iteration until the optimal path leading to a target is finally obtained, the hovering position of the ferry unmanned aerial vehicle during transmission with a relay unmanned aerial vehicle is optimized, and the transmission speed of the ferry unmanned aerial vehicle is improved. A calculation method of an optimized hovering position is given in a three-dimensional space through similar attributes, finally, an optimal cruise path of the ferry unmanned aerial vehicle and the unmanned aerial vehicle charging station is achieved, and efficient transmission and integration of battlefield information are achieved.

Description

Technical Field

[0001] The present invention relates to the fields of artificial intelligence, swarm intelligence technology, unmanned aerial vehicle (UAV) cluster communication, unmanned aerial vehicle (UAV) collaborative computing, unmanned aerial vehicle (UAV) regional combat simulation, "satellite-ground-no" integration technology, and ferry cruise technology, and specifically to a method for realizing ferry cruise based on unmanned aerial vehicle (UAV) swarm intelligence. Background Art

[0002] Unmanned aerial vehicles, or UAVs for short, are unmanned aircraft controlled by radio remote control or self-provided programs. With their streamlined structure, excellent maneuverability, and compact size, UAVs have shown significant advantages in military and civilian fields. When carrying payloads such as cameras, sensors, target detectors, and communication equipment, UAVs can still efficiently cope with complex and changing environmental challenges, enabling UAVs to achieve many novel applications, such as cargo transportation, surveillance, remote sensing, environmental monitoring, and traffic control.

[0003] With the rapid development and innovation of UAV technology, the importance of UAV in the field of military reconnaissance has become increasingly prominent. The types of UAVs are becoming increasingly diverse, and the scope of their missions continues to expand. From the initial simple tasks of reconnaissance and surveillance in safe airspace, they have gradually developed to complex combat tasks such as penetration and strikes in confrontational airspace, achieving a leapfrog development. At the same time, the battlefield environment is developing in the direction of networking, informatization and systematization, showing the characteristics of high dynamics, strong confrontation and extreme complexity. In this context, due to its limited payload capacity, a single UAV is often unable to perform complex tasks such as large-area surveillance and multi-target attacks alone. For this reason, the combat style of UAVs is undergoing changes, and UAV cluster combat has become an important trend in the future UAV combat mode. How to allocate the functions of each UAV and determine the tasks and flight paths of each UAV is of irreplaceable importance for the clustering and intelligent combat of UAVs on the battlefield.

[0004] The background of this study is that in a battlefield environment, a ferry drone cruises between bases in a ground war zone, communicates with the relay drone in each base, obtains real-time war zone information, and uploads data to a near-earth satellite after cruising to achieve timely processing and integration of information. The cruise calculation method of the ferry drone is implemented using swarm intelligence, using animal group behavior to simulate drone cruising behavior, and treating the entire group as a system to achieve mutual benefit and win-win results.

[0005] The traditional drone swarm combat mode relies on the command and control of ground base stations, but in the harsh environment of the field battlefield, the deployment of ground base stations is extremely limited or even non-existent. Therefore, in the absence of base stations, the present invention constructs a drone swarm ferry cruising network. By deploying ferry drones to cruise between bases and communicate with relay drones in the bases, data is collected and quickly uploaded to the low-orbit satellite LEO, thereby achieving efficient information exchange between different combat bases and comprehensive integration of key battlefield information. Summary of the invention

[0006] The purpose of the present invention is to address the defects and shortcomings of the above-mentioned prior art and propose a method for realizing ferry cruising based on drone swarm intelligence. The method is applied to scenarios in which drone swarms are widely used for combat in the future. If an effective method for ferry cruising based on drone swarm intelligence is selected, it is allowed to use medium and low-orbit satellites or Starlink to collaboratively solve the problems of efficient information exchange between different combat bases and comprehensive integration of key battlefield information in the absence of ground mobile base stations.

[0007] The technical solution adopted by the present invention to solve the technical problem is: a method for realizing ferry cruising based on the intelligence of drone swarms, the method comprising the following steps:

[0008] S1. Establish a drone swarm ferry cruising network, deploy relay drones and ferry drones, and set their coordinate parameters and cruising paths.

[0009] S2. Analyze the delay composition of the drone swarm ferry cruise system, including flight delay, data transmission delay and data upload delay, and obtain the calculation method of the total system delay.

[0010] S3. With the goal of minimizing the total system delay, a system optimization model is established, and the flight trajectory of the drone ferry cruise network is optimized through the intelligent algorithm of drone swarm cruise path planning.

[0011] S4. Based on the solved optimal R-UAV access sequence, the hovering position of the ferry UAV is further optimized to obtain the optimal trajectory calculation method based on the hovering position of the UAV, which serves as the final UAV swarm ferry cruising intelligent algorithm.

[0012] A. The steps to establish a drone swarm ferry cruise network are as follows:

[0013] A1. Relay UAV (R-UAV): In the harsh environment of the field battlefield, the entire battlefield is divided into multiple battle zones, and an R-UAV is deployed in each battle zone to hover in the air. Set the coordinate parameters of the relay UAV R-UAV.

[0014] A2. Ferry UAV (F-UAV): Deploy a F-UAV Charging Station (FCS) and a F-UAV throughout the battlefield. The F-UAV departs from the FCS and follows the preset flight path to visit each base circle one by one, and transmits data with the R-UAV until it cruises all the battle zones, returns to the FCS, and transmits all the collected data to the Low Earth Orbit (LEO) control center, preparing to start a new round of cruise. Set the coordinate parameters of the F-UAV and the hovering distance between the F-UAV and the R-UAV.

[0015] A3. Cruise path: Considering that only one relay UAV is deployed in each battle zone, the cruise path of the battle zone can be regarded as the visit order of F-UAV to R-UAV. Set the path set of the road section between any two bases.

[0016] B. The calculation method of the total system delay is as follows:

[0017] The total delay of the system is mainly composed of the flight delay of the F-UAV, the data transmission delay between the F-UAV and the R-UAV, and the data upload delay when the F-UAV uploads data to LEO. The calculation methods of each delay are introduced below.

[0018] B1. Flight delay: Decompose the flight delay of F-UAV into the flight delay between each R-UAV:

[0019]

[0020] In the formula, It represents the flight time taken by the F-UAV from the starting point FCS to the first R-UAV to be visited. Indicates that F-UAV is relaying drones and The flight time required between Indicates the time it takes for the F-UAV to return to the FCS after the cruise ends.

[0021] Assume that F-UAV is flying at a constant speed, and the flight speed is v f , so the flight delay of the drone is proportional to its flight path, T f It can be calculated by formula (2):

[0022]

[0023] The numerator is the total cruising distance. Assuming that the F-UAV travels along the shortest straight line between each point, each component is expressed as Euclidean distance, which is calculated by formulas (3), (4), and (5):

[0024]

[0025]

[0026]

[0027] in, These are the hovering positions of the F-UAVs in various battle zones.

[0028] B2. Data transmission delay

[0029] F-UAV collects data delay in various war zones The total amount of data cache in the corresponding area and the communication rate between F-UAV and R-UAV Related, expressed as:

[0030]

[0031] Communication rate It can be calculated by the following formula:

[0032]

[0033] in, represents the channel bandwidth, Relay drone The maximum transmission power, N0 is the noise power spectrum density, and g0 is the channel power gain at the reference distance. hour, However, this situation is unlikely to happen in reality. Considering the safety accidents such as collision when UAVs are too close, the hovering distance between F-UAV and R-UAV should be greater than a certain threshold d safe .

[0034] Therefore, the data transmission of the system T d It can be expressed as:

[0035]

[0036] B3. Data upload delay

[0037] When the F-UAV completes the cruise and returns to the FCS, it needs to upload all the collected data to the low-orbit satellite LEO. F,L =h L , d F,L is the distance between F-UAV and LEO, h L is the altitude of LEO.

[0038] Assuming that the channel does not change with time during transmission, is the received power of the ferry UAV, R F,L is the uplink data rate achievable on LEO, which can be expressed as:

[0039]

[0040]

[0041] In the formula, is the transmission power from the UAV to LEO, are the transmitting and receiving antenna gains from the UAV to LEO, respectively. L is the distance between LEO and the UAV. s is the wavelength of the LEO-UAV link. F,L , τ and ε are the bandwidth, noise temperature and Boltzmann constant allocated for the LEO-Sat link, respectively.

[0042] Therefore, the upload delay of F-UAV is T u It can be expressed as:

[0043]

[0044] C. The steps to establish the overall system model are as follows:

[0045] The total system delay T can be expressed as:

[0046] T=T f +T d +T u (12)

[0047] The optimization goal is to minimize the total system delay:

[0048] min T(13)

[0049] Combining the constraints of the ferry drone cruising, we can get:

[0050] min T f +T d +T u (14)

[0051]

[0052] In the formula, constraint 1 limits the flight delay of the UAV, T cmax Indicates the maximum flight time after a single charge, flight delay T f Not to exceed T cmax , and cannot be 0; Constraint 2 represents the hovering distance d between F-UAV and R-UAV mShould be greater than a certain threshold d safe ; Constraint 3 means that the war zone number visited each time is different; Constraint 4 means that the sum of the total data buffer of each war zone is not less than the total data volume of the system.

[0053] The key point to be optimized in the present invention is to determine a flight trajectory of the F-UAV so that the flight delay is the shortest, and the key point of flight trajectory planning is to find the set of F-UAV's visit sequence to each battle zone. The following is the path planning of the drone swarm intelligent algorithm:

[0054] In the path planning problem of F-UAV, each UAV is regarded as an F-UAV and the path search is completed, such as Figure 2 As shown, the specific algorithm steps are as follows:

[0055] C1. Initialize the parameters of the drone swarm intelligent algorithm, including the drone swarm size pop, the maximum number of iterations max_iter, the information factor volatility coefficient ρ, the information factor total concentration Q, the information factor heuristic factor α, and the heuristic function value factor β.

[0056] C2. Initialize the drone group information, set the charging station FCS as the access start and end point of the drone group, and set the charging station d safe The hovering point at is set as the intermediate node that the drone group needs to visit, and the sequence set of the drone group visiting R-UAV is randomly generated, which is called the path space. The information factor concentration τ between the hovering point i and the hovering point j is calculated. ij Set them to 1 and calculate the Euclidean distance d between each pair of hovering points ij , and then according to the formula

[0057]

[0058] Calculate the heuristic function value η ij (t). Initialize the taboo table tabu of the drone swarm to an empty set.

[0059] C3, the UAV starts from FCS and goes to the first R-UAV in the path space, and then according to the state transfer formula

[0060]

[0061] Select the next R-UAV to visit, add the corresponding serial number of the visited R-UAV to the taboo table tabu, and update the corresponding path length. is the probability that drone k transfers from node i to node j, allowed k is the set of next-hop nodes that can be selected by drone k. α and β are the concentrations of the influencing information factors τ ij(t) and the heuristic function value η ij (t) The importance weight factor for the drone to select the next destination node. When the parameter α increases, the drone k is more inclined to choose the node with high information factor concentration as the direction of advance; when the parameter β increases, the drone is more likely to move towards its expected degree η ij (t) High nodes move.

[0062] C4: Determine whether all UAVs have visited all R-UAVs to complete the data collection task. If there are still R-UAVs that have not been visited, return to C3; if all R-UAVs have been traversed, return to FCS and continue to execute the following steps.

[0063] C5. In order to prevent the algorithm from falling into a local optimal solution due to excessive concentration of information factors, we introduce an information factor volatilization mechanism to ensure that the information factor dissipates at a certain rate. The update strategy for information factor concentration is as follows:

[0064] τ ij (t+1)=(1-ρ)τ ij (t)+Δτ ij (t) (17)

[0065] Where ρ is the volatility coefficient of the information factor, ρ∈(0,1), Δτ ij (t) is the information factor concentration increment on the path between node i and node j during this time period. Assuming there are m drones in total, the increment is released and accumulated by the m drones:

[0066]

[0067] It represents the information factor concentration released by drone k at time t. The calculation formula is as follows:

[0068]

[0069] Q is the total concentration of information factor, l k is the total distance traveled by UAV k in this cycle. Therefore, this step needs to find the shortest R-UAV access path in this cycle, save the corresponding UAV ID, and correct the information factor concentration τ between two R-UAVs according to the information factor volatilization mechanism. ij .

[0070] C6. Determine whether the number of loops reaches the maximum number of iterations. If not, increase the number of iterations by one and return to step 3 to continue execution. If reached, exit the iteration process and output the calculated optimal R-UAV access sequence set and the shortest path length.

[0071] D. Model optimization based on the drone’s hovering position:

[0072] When establishing the above model, this method simplifies the hovering position of F-UAV and assumes that F-UAV is directly above R-UAV. safe However, considering the distance cost during the flight, it is necessary to further optimize the hovering position of the F-UAV in order to plan a shorter flight path.

[0073] The change of the hovering position of the F-UAV will affect the flight distance of the F-UAV between the battle zones, which may affect the solution of the optimal access sequence set. This method uses the optimal R-UAV access sequence solved by the UAV swarm intelligent algorithm based on the F-UAV cruise path planning, and adjusts the hovering position of the F-UAV on this basis.

[0074] D1. Based on the F-UAV cruise path planning, the UAV swarm intelligent algorithm obtains the optimal visit sequence set S of R-UAV.

[0075] D2. In order to obtain the shortest flight distance and minimize the flight distance cost, the flight trajectory of F-UAV between any two adjacent hovering points is a straight line. Based on the shortest hovering distance constraint between F-UAV and R-UAV, the preliminary path planning of F-UAV can be obtained as follows: Figure 3 As shown in the figure, the F-UAV in each battle zone is located at the sphere centered on the R-UAV, with the shortest hovering distance d safe On a sphere with a radius of d, the hovering distance of F-UAV and R-UAV in each battle zone is guaranteed to be the shortest distance d safe .

[0076] D3. Definition (x i ,y i ,h i ) is the hovering position of F-UAV in the ith battle zone, is the R-UAV position of the next battle zone that the F-UAV is heading to, (x i+1 ,y i+1 ,h i+1 ) represents the hovering position of the unoptimized F-UAV in the i+1th battle zone, (x' i+1 ,y' i+1 ,h' i+1 ) represents the hovering position of the optimized F-UAV in the i+1th battle zone, d i,i+1 is the distance between the (optimized) hovering point of the F-UAV in the i-th battle zone and the i+1th R-UAV. The geometric relationship between the different hovering points of the F-UAV and the R-UAV is as follows: Figure 4 As shown, according to the knowledge of similar triangles:

[0077]

[0078] According to formula (20), we can solve:

[0079]

[0080] Similarly, the other two coordinate components of the optimized hover position can be expressed as:

[0081]

[0082]

[0083] After the F-UAV departs from the FCS, it first selects the target theater according to the determined optimal visit sequence set. Then, the adjusted hovering position is calculated using the above optimization method to ensure that the F-UAV can fly to each theater along the shortest and most efficient path, thus realizing the optimal flight path planning.

[0084] Beneficial effects:

[0085] 1. The present invention is applied to scenarios where drone swarms are widely used for combat in the future. If a method for realizing ferry cruising based on the intelligence of drone swarms is effectively selected, it is allowed to use medium and low-orbit satellites or Starlink to collaboratively solve the problem of efficient information exchange between different combat bases and comprehensive integration of key battlefield information in the absence of ground mobile base stations.

[0086] 2. The present invention establishes a drone swarm ferry cruising network, deploys ferry drones to cruise between regions and communicate with relay drones in each region, releases computing resources in real time or collects data and uploads it to low-orbit satellites, and well implements the optimal implementation method for the total delay of the ferry system.

[0087] 3. The present invention includes a method for a ferry UAV to select a path by the concentration of volatile information factors, and gradually accumulates information factors on high-quality paths after iteration until the optimal route to the target is finally obtained, and optimizes the hovering position of the ferry UAV during transmission with the relay UAV. A calculation method for the optimized hovering position is given in three-dimensional space by similar attributes, and finally the optimal cruising path of the ferry UAV and the UAV charging station is achieved, which well realizes the efficient transmission and integration of battlefield information. BRIEF DESCRIPTION OF THE DRAWINGS

[0088] Figure 1 This is a schematic diagram of the drone swarm’s intelligent ferry cruise network.

[0089] Figure 2 It is the flow chart of the intelligent algorithm of drone swarm.

[0090] Figure 3 This is a schematic diagram of the optimized hovering position of the F-UAV.

[0091] Figure 4 It is a schematic diagram of the geometric relationship between the F-UAV (after optimization) hovering point and the R-UAV.

[0092] Figure 5 It is the optimization path of the drone swarm intelligent algorithm.

[0093] Figure 6 It is a hover position optimization diagram.

[0094] Figure 7 This is the hover position optimization diagram (top view).

[0095] Figure 8 It is the iterative process of the drone swarm intelligent algorithm.

[0096] Fig. 9 This is a comparison chart of three algorithms for solving the F-UAV cruise path length. DETAILED DESCRIPTION

[0097] The invention will be further described in detail below in conjunction with the accompanying drawings.

[0098] like Figure 1-Figure 9 The present invention provides a method for implementing ferry cruising based on drone swarm intelligence, the method comprising the following steps:

[0099] (1) Establishing a drone swarm ferry cruising network:

[0100] (1.1) Relay UAV (R-UAV): In the harsh environment of the field battlefield, such as Figure 1 As shown in Figure 1, the entire battlefield is divided into L battle zones, and an R-UAV is deployed in each battle zone and hovers in the sky. Assume that the relay UAV RU in the lth battle zone l , l∈{1,2,…,L} is deployed at the core location of the theater, and the relay drone RU l The position is recorded as

[0101] (1.2) Ferry UAV (F-UAV): Deploy a F-UAV Charging Station (FCS) throughout the battlefield, with the location recorded as It mainly provides charging services for F-UAV. A round of cruise of F-UAV is set to start from FCS and follow the preset flight path to visit each base circle one by one. When F-UAV flies to the communication range of R-UAV, it will hover stably and transmit data with R-UAV. After the transmission is completed, F-UAV will set off for the next war zone and repeat the above process until all war zones are cruised. Finally, F-UAV returns to FCS and transmits all collected data to the Low Earth Orbit (LEO) control center, preparing to start a new round of cruise. It is assumed that during data transmission, F-UAV hovers directly above R-UAV. safe Place.

[0102] (1.3) Cruise path: Considering that only one relay UAV is deployed in each battle zone, the cruise path of the battle zone can be regarded as the visit order of F-UAV to R-UAV. Assuming the planned cruise path m represents the visit sequence number of the R-UAV, m∈{1,2,…,L}. The F-UAV departs from the FCS and arrives at the mth battle zone in the given flight path, and then communicates with the relay UAV. distance The cruise will hover at the destination and start data transmission. After the cruise, all information will be uploaded to LEO.

[0103] (2) Total system delay:

[0104] The total delay of the system is mainly composed of the flight delay of the F-UAV, the data transmission delay between the F-UAV and the R-UAV, and the data upload delay when the F-UAV uploads data to LEO. The calculation methods of each delay are introduced below.

[0105] (2.1) Flight delay: The flight delay of F-UAV is decomposed into the flight delay between each R-UAV:

[0106]

[0107] In the formula, It represents the flight time taken by the F-UAV from the starting point FCS to the first R-UAV to be visited. Indicates that F-UAV is relaying drones and The flight time required between Indicates the time it takes for the F-UAV to return to the FCS after the cruise ends.

[0108] Assume that F-UAV is flying at a constant speed, and the flight speed is v f , so the flight delay of the drone is proportional to its flight path, T f It can be calculated by formula (2):

[0109]

[0110] The numerator is the total cruising distance. Assuming that the F-UAV travels along the shortest straight line between each point, each component is expressed as Euclidean distance, which is calculated by formulas (3), (4), and (5):

[0111]

[0112] in, These are the hovering positions of the F-UAVs in various battle zones.

[0113] (2.2) Data transmission delay

[0114] F-UAV collects data delay in various war zones The total amount of data cache in the corresponding area and the communication rate between F-UAV and R-UAV Related, expressed as:

[0115]

[0116] Communication rate It can be calculated by the following formula:

[0117]

[0118] in, represents the channel bandwidth, Relay drone The maximum transmission power, N0 is the noise power spectrum density, and g0 is the channel power gain at the reference distance. hour, However, this situation is unlikely to happen in reality. Considering the safety accidents such as collision when UAVs are too close, the hovering distance between F-UAV and R-UAV should be greater than a certain threshold d safe .

[0119] Therefore, the data transmission of the system T d It can be expressed as:

[0120]

[0121] (2.3) Data upload delay

[0122] When the F-UAV completes the cruise and returns to the FCS, it needs to upload all the collected data to the low-orbit satellite LEO. F,L =h L , d F,L is the distance between F-UAV and LEO, hL is the altitude of LEO.

[0123] Assuming that the channel does not change with time during transmission, is the received power of the ferry UAV, R F,L is the uplink data rate achievable on LEO, which can be expressed as:

[0124]

[0125]

[0126] In the formula, is the transmission power from the UAV to LEO, are the transmitting and receiving antenna gains from the UAV to LEO, respectively. L is the distance between LEO and the UAV. s is the wavelength of the LEO-UAV link. F,L , τ and ε are the bandwidth, noise temperature and Boltzmann constant allocated for the LEO-Sat link, respectively.

[0127] Therefore, the upload delay of F-UAV is T u It can be expressed as:

[0128]

[0129] (3) System model establishment:

[0130] The total system delay T can be expressed as:

[0131] T=T f +T d +T u (12)

[0132] The optimization goal is to minimize the total system delay:

[0133] min T(13) combined with the constraints of the ferry UAV cruise, we can get:

[0134] min T f +T d +T u

[0135]

[0136] In the formula, constraint 1 limits the flight delay of the UAV, T cmax Indicates the maximum flight time after a single charge, flight delay T f Not to exceed T cmax , and cannot be 0; Constraint 2 represents the hovering distance d between F-UAV and R-UAVm Should be greater than a certain threshold d safe ; Constraint 3 means that the war zone number visited each time is different; Constraint 4 means that the sum of the total data buffer of each war zone is not less than the total data volume of the system.

[0137] From formula (8), we can see that the data transmission delay is related to the data cache capacity and data transmission rate of the R-UAV in each battle zone. When the hovering distance is determined to be the safe distance d safe When T d Total data cache for all R-UAVs only The latency of data upload to LEO is T u Therefore, the key to optimizing this system is to determine a flight trajectory of the F-UAV that minimizes the flight delay, and the focus of flight trajectory planning is to find the set of F-UAV's visit sequences to each battle zone. The following is the path planning of the drone swarm intelligent algorithm:

[0138] The working theory of UAV Swarm Intelligent Optimization (USIO) is as follows: in the process of executing the task, each drone in the group will leave a volatile information factor on the path it has traveled. This factor will gradually dissipate over time. This process is the update of the information factor. The drones choose the next direction of travel based on the concentration of the information factor on the path. The higher the concentration, the greater the probability of the path being selected. Through this mechanism, the drone swarm can gradually accumulate information factors on high-quality paths after multiple iterations until the best path to the target is finally found.

[0139] In the path planning problem of F-UAV, each UAV is regarded as an F-UAV and the path search is completed, such as Figure 2 As shown, the specific algorithm steps are as follows:

[0140] (3.1) Initialization parameters: Initialize the parameters of the drone swarm intelligent algorithm, including the drone swarm size pop, the maximum number of iterations max_iter, the information factor volatility coefficient ρ, the information factor total concentration Q, the information factor heuristic factor α, and the heuristic function value factor β.

[0141] (3.2) Initialize the UAV swarm information: Set the charging station FCS as the starting and ending point of the UAV swarm’s visit, and set the charging station FCS directly above each R-UAV safe The hovering point at is set as the intermediate node that the drone group needs to visit, and the sequence set of the drone group visiting R-UAV is randomly generated, which is called the path space. The information factor concentration τ between the hovering point i and the hovering point j is calculated. ij Set them to 1 and calculate the Euclidean distance d between each pair of hovering pointsij , and then according to the formula

[0142]

[0143] Calculate the heuristic function value η ij (t). Initialize the taboo table tabu of the drone swarm to an empty set.

[0144] (3.3) The UAV starts from the FCS and goes to the first R-UAV in the path space, and then according to the state transition formula

[0145]

[0146] Select the next R-UAV to visit, add the corresponding serial number of the visited R-UAV to the taboo table tabu, and update the corresponding path length. is the probability that drone k transfers from node i to node j, allowed k is the set of next-hop nodes that can be selected by drone k. α and β are the concentrations of the influencing information factors τ ij (t) and the heuristic function value η ij (t) The importance weight factor for the drone to select the next destination node. When the parameter α increases, the drone k is more inclined to choose the node with high information factor concentration as the direction of advance; when the parameter β increases, the drone is more likely to move towards its expected degree η ij (t) High nodes move.

[0147] (3.4) Determine whether all UAVs have visited all R-UAVs to complete the data collection task. If there are still R-UAVs that have not been visited, return to step 3; if all R-UAVs have been traversed, return to FCS and continue to execute the following steps.

[0148] (3.5) In order to prevent the algorithm from falling into a local optimal solution due to excessive concentration of information factors, an information factor volatilization mechanism is introduced to ensure that the information factor dissipates at a certain rate. The update strategy of information factor concentration is as follows:

[0149] τ ij (t+1)=(1-ρ)τ ij (t)+Δτ ij (t) (17)

[0150] Where ρ is the volatility coefficient of the information factor, ρ∈(0,1), Δτ ij (t) is the information factor concentration increment on the path between node i and node j during this time period. Assuming there are m drones in total, the increment is released and accumulated by the m drones:

[0151]

[0152] It represents the information factor concentration released by drone k at time t. The calculation formula is as follows:

[0153]

[0154] Q is the total concentration of information factor, l k is the total distance traveled by UAV k in this cycle. Therefore, this step needs to find the shortest R-UAV access path in this cycle, save the corresponding UAV ID, and correct the information factor concentration τ between two R-UAVs according to the information factor volatilization mechanism. ij .

[0155] (3.6) Determine whether the number of loops reaches the maximum number of iterations max_iter. If not, increase the number of iterations by one and return to step 3 to continue execution. If reached, exit the iteration process and output the calculated optimal R-UAV access sequence set and the shortest path length.

[0156] (4) Model optimization based on the hovering position of the drone:

[0157] When establishing the above model, the present invention simplifies the hovering position of the F-UAV and assumes that the F-UAV is directly above the R-UAV. safe However, considering the distance cost during the flight, it is necessary to further optimize the hovering position of the F-UAV in order to plan a shorter flight path.

[0158] The change of the hovering position of the F-UAV will affect the flight distance of the F-UAV between the battle zones, which may affect the solution of the optimal access sequence set. We use the optimal R-UAV access sequence solved by the UAV swarm intelligent algorithm based on the F-UAV cruise path planning, and adjust the hovering position of the F-UAV on this basis.

[0159] (4.1) Based on the UAV swarm intelligent algorithm for F-UAV cruise path planning, the optimal visit sequence set S of R-UAV is obtained.

[0160] (4.2) In order to obtain the shortest flight distance and minimize the flight distance cost, the flight trajectory of F-UAV between any two adjacent hovering points is a straight line. Based on the shortest hovering distance constraint between F-UAV and R-UAV, the preliminary path planning of F-UAV can be obtained as follows: Figure 3 As shown in the figure, the F-UAV in each battle zone is located at the sphere centered on the R-UAV, with the shortest hovering distance d safe On a sphere with a radius of d, the hovering distance of F-UAV and R-UAV in each battle zone is guaranteed to be the shortest distance dsafe .

[0161] (4.3)Define (x i ,y i ,h i ) is the hovering position of F-UAV in the ith battle zone, is the R-UAV position of the next battle zone that the F-UAV is heading to, (x i+1 ,y i+1 ,h i+1 ) represents the hovering position of the unoptimized F-UAV in the i+1th battle zone, (x' i+1 ,y' i+1 ,h' i+1 ) represents the hovering position of the optimized F-UAV in the i+1th battle zone, d i,i+1 is the distance between the (optimized) hovering point of the F-UAV in the i-th battle zone and the i+1th R-UAV. The geometric relationship between the different hovering points of the F-UAV and the R-UAV is as follows: Figure 4 As shown, according to the knowledge of similar triangles:

[0162]

[0163] According to formula (20), we can solve:

[0164]

[0165] Similarly, the other two coordinate components of the optimized hover position can be expressed as:

[0166]

[0167]

[0168] After the F-UAV departs from the FCS, it first selects the target theater according to the determined optimal visit sequence set. Then, the adjusted hovering position is calculated using the above optimization method to ensure that the F-UAV can fly to each theater along the shortest and most efficient path, thus realizing the optimal flight path planning.

[0169] (5) Analysis of simulation results:

[0170] The above algorithm is simulated and analyzed using MATLAB to prove the effectiveness of the UAV Swarm Intelligent Optimization (USIO) algorithm. Consider a 10km×10km combat area, including 15 combat bases, and each combat base has a relay UAV RU l Location As shown in Table 1, the data storage capacity Datal ∈(4,6)Gbit, the channel bandwidth B of R-UAV l =10MHz, maximum transmission power p l =10dB,l∈{1,2,…,L}.

[0171] The rest of the simulation parameters are set as follows: The flight speed v of F-UAV f =15m / s, safe communication distance d between F-UAV and R-UAV safe =100m, channel gain power at reference distance g0 = -50dB, noise power spectrum density N0 = -174dBm / Hz, transmission power from F-UAV to LEO Transmit and receive antenna gain The height of LEO h L =550km, frequency f of LEO-UAV link s =2.4GHz, LEO-Sat link allocation bandwidth B F,L =40MHz, noise temperature τ = 1000 and Boltzmann constant ε = 1.38 × 10 -23 , the coordinates of FCS are set to (0,0,0). Thus, the optimal cruising path of F-UAV under the UAV swarm intelligent algorithm is obtained as follows: Figure 5 shown.

[0172] Based on the UAV swarm intelligent algorithm, the algorithm is optimized based on the hovering position of the ferry UAV. Here we show the simulation results when the number of bases in the war zone is 15. Figure 6 and Figure 7 It can be seen that through optimization, the hovering position of F-UAV is no longer directly above R-UAV. safe Instead, we choose R-UAV as the center of the sphere, d safe is an optimal position on the sphere of radius , and this position will iteratively change with the position of the previous R-UAV visited by the F-UAV.

[0173] In the comparative analysis of computing performance, since the R-UAV selected in each theater base represents the theater to exchange data with the F-UAV, Figure 8 The iterative process of the algorithm for solving the optimal sequence set of F-UAV visiting R-UAV when the number of bases in different war zones is 10, 15, 20, and 25 is shown. It can be seen that the convergence of the algorithm (expressed by the CPU calculation unit time) slows down with the increase in the number of war bases. However, through comparison with existing conventional intelligent algorithms, it can be seen that the convergence of the swarm intelligence ferry cruise calculation method proposed in the present invention is faster and more effective.

[0174] like Fig. 9As shown in the figure, after the hovering position optimization algorithm, the total cruising path of F-UAV becomes 33005m, which is 4.7% shorter than the traditional intelligent algorithm path. At this time, the calculated flight delay of F-UAV is 2200.3s, the data interaction delay between F-UAV and R-UAV is 176.9s, the delay of F-UAV uploading data to LEO is 66.5s, and the total delay is 2443.7s. Among them, the cruising delay of F-UAV accounts for 90% of the total system delay and is the most critical factor affecting system performance, which once again verifies the necessity of optimizing flight delay.

[0175] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

[0176] Those skilled in the art should understand that the present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0177] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0178] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1The steps for the functions specified in one or more boxes.

[0179] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for implementing ferry cruising based on drone swarm intelligence, characterized in that: include: S1. Establish a drone group ferry cruising network, deploy relay drones and ferry drones, and set their coordinate parameters and cruising paths; S2. Analyze the delay composition of the drone group ferry cruise system, including flight delay, data transmission delay and data upload delay, and obtain the calculation method of the total system delay; S3. With the goal of minimizing the total system delay, a system optimization model is established to optimize the flight trajectory of the drone ferry cruise network through an intelligent algorithm for drone swarm cruise path planning; S4. Based on the solved optimal R-UAV access sequence, the hovering position of the ferry UAV is further optimized to obtain the optimal trajectory calculation method based on the hovering position of the UAV, which serves as the final UAV swarm ferry cruising intelligent algorithm.

2. The method for realizing ferry cruising based on drone swarm intelligence according to claim 1 is characterized in that: The specific steps of step S1 are as follows: (a) Relay UAV (R-UAV): In the harsh environment of a field battlefield, the entire battlefield is divided into multiple battle zones. An R-UAV is deployed in each battle zone and hovers in the air. The coordinate parameters of the relay UAV R-UAV are set; (b) Ferry UAV (F-UAV): A F-UAV Charging Station (FCS) and an F-UAV are deployed throughout the battlefield. The F-UAV departs from the FCS and follows the preset flight path to visit each base circle one by one, and transmits data with the R-UAV until it cruises all the battle zones. It returns to the FCS and transmits all the collected data to the Low Earth Orbit (LEO) control center, preparing to start a new round of cruise, setting the coordinate parameters of the F-UAV and the hovering distance between the F-UAV and the R-UAV. (c) Cruise path: Considering that only one relay UAV is deployed in each battle zone, the cruise path of the battle zone can be regarded as the visit order of F-UAV to R-UAV, and the path set of the road section between any two bases is set.

3. The method for realizing ferry cruising based on drone swarm intelligence according to claim 1 is characterized in that: The specific steps of step S2 are as follows: (a) Flight delay: The flight delay of F-UAV is decomposed into the flight delay between each R-UAV: In the formula, It represents the flight time taken by the F-UAV from the starting point FCS to the first R-UAV to be visited. Indicates that F-UAV is relaying drones and The flight time required between It indicates the time taken by the F-UAV to return to the FCS after the cruise ends; Assume that F-UAV is flying at a constant speed, and the flight speed is v f , so the flight delay of the drone is proportional to its flight path, T f It can be calculated by formula (2): The numerator is the total cruising distance. Assuming that the F-UAV travels along the shortest straight line between each point, each component is expressed as Euclidean distance, which is calculated by formulas (3), (4), and (5): in, It is the hovering position of F-UAV in each battle zone; (b) Data transmission delay F-UAV collects data delay in various war zones The total amount of data cache in the corresponding area and the communication rate between F-UAV and R-UAV Related, expressed as: Communication rate It can be calculated by the following formula: in, represents the channel bandwidth, Relay drone The maximum transmission power, N0 is the noise power spectrum density, g0 is the channel power gain at the reference distance, when hour, However, this situation is unlikely to happen in reality. Considering the safety accidents such as collision when UAVs are infinitely close, the hovering distance between F-UAV and R-UAV should be greater than a certain threshold d. safe ; Therefore, the data transmission of the system T d It can be expressed as: (c) Data upload delay When the F-UAV completes the cruise and returns to the FCS, all the collected data needs to be uploaded to the low-orbit satellite LEO. F,L =h L , d F,L is the distance between F-UAV and LEO, h L is the altitude of LEO; Assuming that the channel does not change with time during transmission, is the received power of the ferry UAV, R F,L is the uplink data rate achievable on LEO, which can be expressed as: In the formula, is the transmission power from the UAV to LEO, are the transmitting and receiving antenna gains from the UAV to LEO, h L is the distance between LEO and the UAV, λ s is the wavelength of the LEO-UAV link, B F,L , τ and ε are the bandwidth, noise temperature and Boltzmann constant of the LEO-Sat link allocation, respectively; Therefore, the upload delay of F-UAV is T u It can be expressed as: (d) Total system delay The total system delay T can be expressed as: T=T f +T d +T u (12)。 4. The method for realizing ferry cruising based on drone swarm intelligence according to claim 1 is characterized in that: The specific steps of step S3 are as follows: (a) Establish optimization objectives and constraints, and build a system optimization model: The optimization goal is to minimize the total system delay: min T(13) combined with the constraints of the ferry UAV cruise, we can get: min T f +T d +T u (14) In formula (14), constraint 1 limits the flight delay of the UAV, T cmax Indicates the maximum flight time after a single charge, flight delay T f Not to exceed T cmax , and cannot be 0; Constraint 2 represents the hovering distance d between F-UAV and R-UAV m Should be greater than a certain threshold d safe ;Constraint 3 means that the war zone number is different each time it is visited; Constraint 4 indicates that the sum of the total data buffers of each theater is not less than the total data volume of the system; (b) Establish an intelligent algorithm for drone group cruise path planning to determine the flight trajectory of the F-UAV so that the flight delay is minimized. The key point of flight trajectory planning is to find the set of F-UAV's visit sequence to each battle zone.

5. The method for realizing ferry cruising based on drone swarm intelligence according to claim 1 is characterized in that: The specific steps of step S4 are as follows: (a) Based on the F-UAV cruise path planning, the UAV swarm intelligent algorithm obtains the optimal visit sequence set S of R-UAV; (b) In order to obtain the shortest flight distance and minimize the flight distance cost, the flight trajectory of the F-UAV between any two adjacent hovering points is a straight line. Based on the shortest hovering distance constraint between the F-UAV and the R-UAV, the preliminary path planning of the F-UAV is obtained. In each battle zone in the figure, the F-UAV is located at a sphere with the R-UAV as the center and the shortest hovering distance d safe On a sphere with a radius of d, the hovering distance of F-UAV and R-UAV in each battle zone is guaranteed to be the shortest distance d safe ; (c) Definition (x i ,y i ,h i ) is the hovering position of F-UAV in the ith battle zone, is the R-UAV position of the next battle zone that the F-UAV is heading to, (x i+1 ,y i+1 ,h i+1 ) represents the hovering position of the unoptimized F-UAV in the i+1th battle zone, (x i ' +1 ,y i ' +1 ,h i ' +1 ) represents the hovering position of the optimized F-UAV in the i+1th battle zone, d i,i+1 is the distance between the (optimized) hovering point of the F-UAV in the i-th battle zone and the i+1-th R-UAV. The geometric relationship between different hovering points of the F-UAV and the R-UAV is shown in Figure 4. According to the knowledge of similar triangles, we can know that: According to formula (15), we can solve: Similarly, the other two coordinate components of the optimized hover position can be expressed as: After the UAV departs from the FCS, it first selects the target theater based on the determined optimal visit order set. Then, the adjusted hovering position is calculated using the above optimization method to ensure that the F-UAV can fly to each theater along the shortest and most efficient path, thus realizing the optimal flight path planning.

6. The method for realizing ferry cruising based on drone swarm intelligence according to claim 4 is characterized in that: The specific steps of drone swarm cruising path planning are as follows: (a) Initialize the parameters of the drone swarm intelligent algorithm, including the drone swarm size pop, the maximum number of iterations max_iter, the information factor volatility coefficient ρ, the information factor total concentration Q, the information factor heuristic factor α, and the heuristic function value factor β; (b) Initialize the UAV swarm information, set the charging station FCS as the starting and ending point of the UAV swarm, and set the charging station d directly above each R-UAV safe The hovering point at is set as the intermediate node that the drone group needs to visit, and a random sequence set of drone group visits R-UAV is generated, which is called path space. The information factor concentration τ between hovering point i and hovering point j is calculated. ij Set them to 1 and calculate the Euclidean distance d between each pair of hovering points ij , and then according to the formula: Calculate the heuristic function value η ij (t), initialize the taboo table tabu of the drone swarm to an empty set; (c) The UAV starts from the FCS and goes to the first R-UAV in the path space, and then according to the state transition formula Select the next R-UAV to visit, add the corresponding serial number of the visited R-UAV to the taboo table tabu, and update the corresponding path length, where, is the probability that drone k transfers from node i to node j, allowed k is the set of next-hop nodes that can be selected by UAV k, α and β are the concentrations of the influencing information factors τ ij (t) and the heuristic function value η ij (t) For the importance weight factor of the drone's selection of the next destination node, when the parameter α increases, the drone k is more inclined to choose the node with high information factor concentration as the forward direction; when the parameter β increases, the drone is more likely to move towards its expected degree η ij (t) high node movement; (d) Determine whether all UAVs have visited all R-UAVs to complete the data collection task. If there are still R-UAVs that have not been visited, return to C3; if all R-UAVs have been traversed, return to FCS and continue to execute the following steps; (e) In order to prevent the algorithm from falling into a local optimal solution due to excessive information factor concentration, an information factor volatilization mechanism is introduced to ensure that the information factor dissipates at a certain rate. The update strategy of the information factor concentration is as follows: t ij (t+1)=(1-ρ)τ ij (t)+Δτ ij (t) (21) Where ρ is the volatility coefficient of the information factor, ρ∈(0,1), Δτ ij (t) is the information factor concentration increment on the path between node i and node j during this time period. Assuming there are m drones in total, the increment is released and accumulated by the m drones: It represents the information factor concentration released by drone k at time t. The calculation formula is as follows: Q is the total concentration of information factor, l k is the total distance traveled by UAV k in this cycle, find the shortest R-UAV access path in this cycle, save the corresponding UAV ID, and correct the information factor concentration τ between two R-UAVs according to the information factor volatilization mechanism ij ; (f) Determine whether the number of loops reaches the maximum number of iterations. If not, increase the number of iterations by one and return to step 3 to continue execution. If reached, exit the iteration process and output the calculated optimal R-UAV access sequence set and the shortest path length.

7. The method for realizing ferry cruising based on drone swarm intelligence according to claim 4 is characterized in that: F-UAV is directly above R-UAV safe Hovering at a certain position and transmitting data with R-UAV.

8. A ferry cruise computing system based on drone swarm intelligence, characterized by: The information preprocessing module is configured to: capture the initial geographic information data on the ground by Beidou satellite navigation, and timely capture and summarize the basic geographic coordinate information of the current UAV by a low-Earth orbit satellite (LEO) as the basic data and boundary conditions for subsequent resource scheduling plans; The resource scheduling module is configured to: analyze the local resource status and motion status of the drone group in each cycle, divide the current regional scheduling into different combat mission areas and designated relay drones (R-UAVs), and report to the scheduling decision center, thereby providing the cruise path and hovering position information of the ferry drone (F-UAV) through a ferry cruise implementation method based on drone group intelligence; The charging scheduling module is configured to: take into account the energy consumption uncertainty caused by the path uncertainty of the ferry drone cruise, set up the charging scheduling module, and obtain the best endurance charging location information for the ferry drone charging station (FCS) through a ferry cruise implementation method based on drone swarm intelligence.

9. An electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, characterized in that the instructions are executed by the at least one processor so that the at least one processor can execute the method described in any one of claims 1-8.

Images