WO2020147456A1 - 无人机路径优化方法、设备及存储介质 - Google Patents

无人机路径优化方法、设备及存储介质 Download PDF

Info

Publication number
WO2020147456A1
WO2020147456A1 PCT/CN2019/124036 CN2019124036W WO2020147456A1 WO 2020147456 A1 WO2020147456 A1 WO 2020147456A1 CN 2019124036 W CN2019124036 W CN 2019124036W WO 2020147456 A1 WO2020147456 A1 WO 2020147456A1
Authority
WO
WIPO (PCT)
Prior art keywords
wireless multicast
energy consumption
multicast user
hovering
flight path
Prior art date
Application number
PCT/CN2019/124036
Other languages
English (en)
French (fr)
Inventor
许文俊
邓昶
张平
张治�
高晖
冯志勇
Original Assignee
北京邮电大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 北京邮电大学 filed Critical 北京邮电大学
Publication of WO2020147456A1 publication Critical patent/WO2020147456A1/zh
Priority to US17/023,813 priority Critical patent/US11663920B2/en

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G5/00—Traffic control systems for aircraft, e.g. air-traffic control [ATC]
    • G08G5/0004—Transmission of traffic-related information to or from an aircraft
    • G08G5/0013—Transmission of traffic-related information to or from an aircraft with a ground station
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/20—Instruments for performing navigational calculations
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B64—AIRCRAFT; AVIATION; COSMONAUTICS
    • B64C—AEROPLANES; HELICOPTERS
    • B64C39/00—Aircraft not otherwise provided for
    • B64C39/02—Aircraft not otherwise provided for characterised by special use
    • B64C39/024—Aircraft not otherwise provided for characterised by special use of the remote controlled vehicle type, i.e. RPV
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/26—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
    • G01C21/34—Route searching; Route guidance
    • G01C21/3453—Special cost functions, i.e. other than distance or default speed limit of road segments
    • G01C21/3469—Fuel consumption; Energy use; Emission aspects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G06Q10/047—Optimisation of routes or paths, e.g. travelling salesman problem
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G5/00—Traffic control systems for aircraft, e.g. air-traffic control [ATC]
    • G08G5/003—Flight plan management
    • G08G5/0034—Assembly of a flight plan
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G5/00—Traffic control systems for aircraft, e.g. air-traffic control [ATC]
    • G08G5/0047—Navigation or guidance aids for a single aircraft
    • G08G5/0069—Navigation or guidance aids for a single aircraft specially adapted for an unmanned aircraft
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L12/00—Data switching networks
    • H04L12/02—Details
    • H04L12/16—Arrangements for providing special services to substations
    • H04L12/18—Arrangements for providing special services to substations for broadcast or conference, e.g. multicast
    • H04L12/185—Arrangements for providing special services to substations for broadcast or conference, e.g. multicast with management of multicast group membership
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/02—Services making use of location information
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/06—Selective distribution of broadcast services, e.g. multimedia broadcast multicast service [MBMS]; Services to user groups; One-way selective calling services
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/06—Selective distribution of broadcast services, e.g. multimedia broadcast multicast service [MBMS]; Services to user groups; One-way selective calling services
    • H04W4/08—User group management
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B64—AIRCRAFT; AVIATION; COSMONAUTICS
    • B64U—UNMANNED AERIAL VEHICLES [UAV]; EQUIPMENT THEREFOR
    • B64U2101/00—UAVs specially adapted for particular uses or applications
    • B64U2101/20—UAVs specially adapted for particular uses or applications for use as communications relays, e.g. high-altitude platforms
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B64—AIRCRAFT; AVIATION; COSMONAUTICS
    • B64U—UNMANNED AERIAL VEHICLES [UAV]; EQUIPMENT THEREFOR
    • B64U2201/00—UAVs characterised by their flight controls
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B64—AIRCRAFT; AVIATION; COSMONAUTICS
    • B64U—UNMANNED AERIAL VEHICLES [UAV]; EQUIPMENT THEREFOR
    • B64U2201/00—UAVs characterised by their flight controls
    • B64U2201/20—Remote controls
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/004—Artificial life, i.e. computing arrangements simulating life
    • G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00—Computing arrangements using knowledge-based models
    • G06N5/01—Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/90—Services for handling of emergency or hazardous situations, e.g. earthquake and tsunami warning systems [ETWS]
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Definitions

  • This application relates to the field of UAV communication technology, specifically a method, equipment and storage medium for UAV path optimization.
  • UAV communication has also become a research focus in the field of wireless communication.
  • UAV communication systems have the advantages of low cost, fast deployment, high flexibility, excellent channel conditions, and mobile controllability. Therefore, UAV communication will gradually become indispensable in the future wireless communication field. The missing part.
  • UAV wireless communication is not perfect and faces many challenges. Among them, insufficient energy is one of the biggest factors restricting the development of UAV communication.
  • wireless multicast has become a mature technology for multimedia transmission, which can effectively improve spectrum utilization.
  • traditional wireless multicast technology faces a very challenging problem, that is, the transmission rate of multicast is essentially limited by the worst user channel conditions in the multicast group.
  • UAV communication to achieve wireless multicast has become a research hotspot.
  • the use of UAV communication to achieve wireless multicast can solve the problem of limited wireless multicast transmission rate by taking advantage of the low cost of the UAV communication system, controllable mobility and excellent channel conditions.
  • it can also rely on multicast.
  • the advantage of transmitting data to multiple users at the same time effectively reduces the time and energy consumption required by the drone to perform transmission tasks, and achieves mutual benefit and win-win results.
  • this application proposes a UAV path optimization method, equipment and storage medium, which can reduce the energy consumption of the UAV by optimizing the hovering position and flight path of the UAV.
  • the UAV path optimization method proposed in this application includes:
  • determining the number of wireless multicast user groups includes: using a trained support vector machine model SVM to determine the number of wireless multicast user groups according to the number of wireless multicast users, geographic location, and file size to be transmitted.
  • determining the number of wireless multicast user groups includes:
  • K* of the number of each wireless multicast user group use the k-means algorithm to divide the number of users N into K* wireless multicast user groups, and calculate the average profile coefficient ⁇ * corresponding to K*;
  • the K * corresponding to the smallest predicted total energy consumption E K * is taken as the number K of the wireless multicast user group.
  • the training SVM model includes:
  • Each set of training data includes the number of users, the user's geographic location, the size of the file to be transmitted, and the number of wireless multicast user groups K;
  • the average contour coefficient of each group of data is calculated from the number of users in each group of training data, the user's geographic location, and the number of wireless multicast user groups K;
  • the X group includes the number of users, the split number K of the wireless multicast user group, the file size to be transmitted and the data of the average profile coefficient, and the corresponding X group only contains the data of the final total energy consumption;
  • the data including the number of users, the number of wireless multicast user groups K, the file size to be transmitted, and the average contour coefficient are used as input, and the data including the final total energy consumption is used as the output, and the SVM model is trained.
  • dividing the wireless multicast users into multiple wireless multicast user groups according to the number of the wireless multicast user groups includes: dividing the wireless multicast users into multiple wireless multicast user groups using k-means clustering algorithm k-means algorithm .
  • determining the hovering position of the drone corresponding to each wireless multicast user group includes: using the minimum circle coverage method to respectively determine the hovering position of the drone corresponding to each wireless multicast user group.
  • using the minimum circle coverage method to determine the hovering position of the drone corresponding to each wireless multicast user group includes:
  • A take any of a wireless multicast subscriber group, the wireless multicast user from the user group contains three users of arbitrarily taken N a, N b, N c ;
  • N d is a circle or in the circumference, then ask the smallest circle is a circle, the circle center for the wireless multicast group corresponding to a user hovering position; otherwise, N a, N b, Choose 3 points from N c and N d to make a circle containing these 4 points the smallest, and make these 3 points become new Na , N b , N c , and return to B.
  • determining the shortest flight path connecting the hovering position includes: obtaining the shortest flight path by using any one of genetic algorithm, particle swarm algorithm, ant colony algorithm, simulated annealing algorithm and neural network algorithm;
  • Determining the shortest flight path connecting the optimized hovering position includes: obtaining the shortest flight path by using any one of genetic algorithm, particle swarm algorithm, ant colony algorithm, simulated annealing algorithm, and neural network algorithm.
  • obtaining the shortest flight path by using a genetic algorithm includes:
  • a selection probability is set for each path according to the distance of each path in the L paths. The smaller the distance, the greater the selection probability is set, and L selections are made. Choose a path from L paths for each selection, and then proceed to the next step with the selected path;
  • each path is reversed in turn with a preset probability, and the previous path is replaced until all L paths have been selected without repeated selection, resulting in a new generation Population
  • determining the first total energy consumption of the drone according to the size of the file to be transmitted, the hovering position, and the flight path includes:
  • the total energy consumption of the drone is determined by the determined communication energy consumption, hovering energy consumption and flight energy consumption.
  • optimizing each of the hovering positions one by one includes: optimizing the hovering positions one by one by using an interior point method.
  • This application also provides a UAV path optimization device, including:
  • the user group quantity determination module is used to determine the number of wireless multicast user groups according to the number of wireless multicast users, the geographic location of each wireless multicast user, and the file size to be transmitted;
  • the grouping module is used to divide the wireless multicast users into multiple wireless multicast user groups according to the determined number of wireless multicast user groups;
  • the hovering position determination module is used to determine the hovering position of the drone corresponding to each wireless multicast user group;
  • the flight path determination module is configured to obtain the shortest flight path connecting the hovering position according to the determined hovering position of the drone corresponding to each wireless multicast user group;
  • the first energy consumption determination module is configured to determine the first total energy consumption of the drone according to the size, hover position, and flight path of the file to be transmitted;
  • the optimization module is used to optimize each hover position one by one to obtain multiple optimized hover positions, and determine the shortest flight connecting the multiple optimized hover positions according to the multiple optimized hover positions path;
  • the second energy consumption determination module is used to determine the second total energy consumption of the drone according to the size of the file to be transmitted, the optimized hovering position, and the flight path;
  • the comparison module is used to compare the first total energy consumption with the second total energy consumption, and if the absolute value of the difference between the two is greater than a preset threshold value, instruct the optimization module to perform the next step based on the current round of optimization.
  • One round of optimization if the absolute value of the difference between the two is less than or equal to the preset threshold, the optimized hover position and flight path are determined as the hover position and flight path of the drone.
  • the present application also provides a UAV path optimization device, including: at least one processor; and a memory communicatively connected with the at least one processor; wherein the memory stores storage that can be executed by the at least one processor The instructions are executed by the at least one processor, so that the at least one processor can execute the aforementioned UAV path optimization method.
  • the present application also provides a computer-readable storage medium on which computer instructions are stored, and the above-mentioned UAV path optimization method is realized when the processor executes the above-mentioned computer instructions.
  • This application can fully consider the number of users and geographical distribution, and adopt a data-driven method to adaptively split users into several multicast groups, and find a suitable hovering position in each multicast group.
  • the UAV can transmit files at this location at a higher transmission rate, so as to reduce the energy consumption of the UAV hovering and transmission.
  • this application can design the flight trajectory of the drone, find a shortest flight path through all hovering positions and return to the starting position, to reduce the flight energy consumption of the drone, and ultimately enable the drone to take into account mechanical Energy and communication energy consumption, with the minimum total energy consumption to complete the multicast service transmission task.
  • FIG. 1 is a schematic diagram of an implementation scenario of using drone communication to realize wireless multicast according to an embodiment of the application
  • FIG. 2 is a schematic flow chart of a method for optimizing UAV path according to an embodiment of the application
  • FIG. 3 is an example of the flow of the UAV path optimization method according to an embodiment of the application.
  • FIG. 5 is a schematic flowchart of a method for calculating the number of wireless multicast user groups K according to an embodiment of the application
  • FIG. 6 is a schematic flow chart of the method for determining the hovering position of the drone by adopting the minimum circle coverage method according to an embodiment of the application;
  • FIG. 7 is a schematic flowchart of a method for determining the shortest flight path by using a genetic algorithm according to an embodiment of the present invention.
  • FIG. 8 is a schematic structural diagram of a UAV path optimization device according to an embodiment of the application.
  • FIG. 9 is a schematic diagram of the hardware structure of the UAV path optimization device according to an embodiment of the application.
  • the current research on the UAV multicast communication system is mainly to improve the communication performance between the UAV and the user by designing the flight trajectory of the UAV.
  • the existing UAV flight trajectory design usually does not consider the energy consumption of the UAV when performing tasks, so that the problem of communication interruption due to insufficient energy in the communication process of the UAV frequently occurs.
  • FIG. 1 shows a schematic diagram of an implementation scenario of using drone communication to implement wireless multicast according to an embodiment of the present application.
  • the local communication infrastructure is destroyed by natural disasters, or there is no ground communication infrastructure in remote areas, it is necessary to use UAV 1 to multicast and transmit general files and perform wireless multicast tasks.
  • the machine 1 needs to provide a wireless multicast service to a wireless multicast user set 2 composed of multiple wireless multicast users 4.
  • the multicast transmission rate of wireless communication depends on the channel condition of the user with the worst channel
  • it is also possible to further target the wireless multicast user 4 in the wireless multicast user set 2 Grouping is performed to obtain multiple wireless multicast user groups 3, and the above-mentioned UAV 1 provides wireless multicast services for each wireless multicast user group.
  • some embodiments of the present application propose a method for optimizing the path of the drone.
  • This method can first divide multiple wireless multicast users into multiple wireless multicast user groups; then, find a hovering point in each wireless multicast user group; finally, find another one that passes through the multiple wireless multicast users
  • the shortest flight path of the hovering point of the group minimizes the total energy consumed by the drone to complete the multicast transmission task, thereby reducing the total energy consumed by the drone to complete the multicast transmission task, that is, reducing the total energy consumption.
  • the purpose of improving energy efficiency is a method for optimizing the path of the drone.
  • the aforementioned total energy consumption may include mechanical energy consumption and communication energy consumption.
  • mechanical energy consumption can include the energy consumed by the drone hovering at the hovering position, called hovering energy consumption; and the energy consumed by the drone flying between the hovering positions, called flight energy consumption.
  • communication energy consumption refers to the energy consumed by the drone for data transmission.
  • Figure 2 shows the flow of the UAV path optimization method described in the embodiment of the present application.
  • the method may be executed by other computing devices that communicate with the drone.
  • the method may include:
  • Step 201 Obtain the number of wireless multicast users, the geographic location of each wireless multicast user, and the size of the file to be transmitted.
  • the number of wireless multicast users, the geographic location, and the file size to be transmitted can be obtained through positioning methods such as GPS and the files that need to be transmitted in the task.
  • Step 202 Determine the number of wireless multicast user groups according to the number of wireless multicast users, the geographic location of each wireless multicast user, and the file size to be transmitted.
  • a trained support vector machine (SVM, Support Vector Machine) model may be used to determine the number of wireless multicast user groups according to the number of wireless multicast users, geographic locations, and file sizes to be transmitted.
  • SVM Support Vector Machine
  • a trained neural network may also be used to determine the number of wireless multicast user groups based on the number of wireless multicast users, their geographic locations, and the size of files to be transmitted.
  • Step 203 Divide the wireless multicast users into multiple wireless multicast user groups according to the determined number of wireless multicast user groups.
  • a k-means clustering algorithm (k-means) algorithm or other clustering algorithms may be used to divide wireless multicast users into multiple wireless multicast user groups.
  • Step 204 Determine the hovering position of the drone corresponding to each wireless multicast user group.
  • the minimum circle coverage method may be used to obtain the hovering positions corresponding to the aforementioned multiple invalid multicast user groups respectively.
  • Step 205 Obtain the shortest flight path connecting the above hovering position according to the determined hovering position of the drone corresponding to each wireless multicast user group.
  • any one of genetic algorithm, particle swarm algorithm, ant colony algorithm, simulated annealing algorithm, and neural network algorithm may be used to obtain the shortest flight path.
  • Step 206 Determine the first total energy consumption of the drone according to the size of the file to be transmitted, the hovering position, and the flight path.
  • the communication energy consumption and hovering energy consumption of the drone can be determined according to the size of the file to be transmitted and the hovering position.
  • the time required to transmit the file and the transmission power of the drone can be determined according to the size of the file to be transmitted and the hovering position, so that the communication energy consumption and hovering energy consumption of the drone can be determined.
  • the flight energy consumption of the drone can also be determined according to the above flight path.
  • the total energy consumption of the drone is determined by the determined communication energy consumption, hovering energy consumption and flight energy consumption.
  • Step 207 optimizing each of the above-mentioned hovering positions one by one to obtain multiple optimized hovering positions.
  • the interior point method or other nonlinear programming solution methods can be used to optimize the hovering positions mentioned above one by one.
  • Step 208 Determine the shortest flight path connecting the multiple optimized hovering positions according to the multiple optimized hovering positions.
  • the foregoing step 208 may be implemented using the same method as the foregoing step 205.
  • Step 209 Determine the second total energy consumption of the drone according to the size of the file to be transmitted, the optimized hovering position, and the flight path.
  • Step 210 Compare the above-mentioned first total energy consumption with the second total energy consumption, and if the absolute value of the difference between the two is greater than the preset threshold, return to the above-mentioned step 207 on the basis of this round of optimization and proceed to the next round Optimization; if the absolute value of the difference between the two is less than or equal to the preset threshold, step 211 is executed.
  • the first total energy consumption may be set as the second total energy consumption, and then the step 207 may be returned.
  • Step 211 Determine the optimized hover position and flight path as the hover position and flight path of the drone.
  • Fig. 3 shows a specific example of the flow of the UAV path optimization method according to an embodiment of the present application. As shown in Figure 3, the hovering positions and flight paths of the drone can be determined through the following steps.
  • Step S1 Obtain the number N of wireless multicast users, the geographic location W n of each wireless multicast user, and the file size S to be transmitted;
  • Step S2 Use the trained SVM model and the number of users N, the geographic location W n and the file size S to be transmitted to calculate the number of wireless multicast user groups K, and use the k-means algorithm to split the number of wireless multicast users N Form K wireless multicast user groups, referred to as wireless multicast user groups;
  • Step S3 Use the minimum circle coverage method to obtain the K hovering positions corresponding to the K wireless multicast user groups one-to-one, and use the genetic algorithm to obtain the shortest flight path connecting the K hovering positions.
  • the file size S and The hovering position determines the communication energy consumption and hovering energy consumption of the UAV, the flight energy consumption is determined by the flight path, and the total energy consumption of the UAV is composed of the communication energy consumption, hovering energy consumption and flight energy consumption Consumption
  • Step S4 Use the interior point method to optimize the hovering positions one by one to obtain the optimized hovering position, use the genetic algorithm to obtain the shortest flight path connecting the K optimized hovering positions, and use the optimized hovering position And the total energy consumption of the drone after the flight path is optimized;
  • Step S5 Compare the optimized total energy consumption of the UAV with the total energy consumption of the UAV before optimization. If the absolute value of the difference between the two is greater than the set threshold, repeat the steps based on the current round of optimization S4 performs the next round of optimization; if the absolute value of the difference between the two is less than or equal to the set threshold, the optimized hovering position and flight path of the drone are the final hovering position and final flight path.
  • the quality of the channel conditions mainly depends on the distance between the wireless multicast users and the UAV.
  • the quality of the channel conditions mainly depends on the distance between the wireless multicast users and the UAV.
  • reduce UAV transmission time and energy consumption in order to increase the UAV multicast transmission rate.
  • the wireless multicast user set 2 and the geographic locations of the wireless multicast users in it first find the optimal number of wireless multicast user groups K ; Then the wireless multicast user set is split into K wireless multicast user groups; and then an initial hovering position is determined relative to each wireless multicast user group. In this initial hovering position, when the drone transmits files to users in the group, the transmission rate is faster and the communication energy consumption is lower.
  • the transmission time is short, that is, the hovering time is short.
  • the hovering energy consumption is reduced.
  • the shortest flight route connecting the K hovering positions is determined, so as to minimize the flight energy consumption. Since the total energy consumption of the UAV is basically equal to the sum of the communication energy consumption, hovering energy consumption and flight energy consumption, the total energy consumption of the UAV can be determined after determining the hovering position and flight path of the UAV.
  • the above hovering position can be further optimized, and then the optimized shortest flight path is determined from the optimized hovering position, and finally the lowest total energy consumption of the UAV after optimization is determined, and the lowest one is finally determined.
  • the hovering position and flight path of the drone corresponding to the total energy consumption of the drone.
  • This application splits multiple wireless multicast users into multiple wireless multicast user groups adaptively, find a suitable hovering point in each wireless multicast user group, and find the shortest flight past these hovering points The path minimizes the total energy consumed by a UAV to complete the multicast transmission task, even if the UAV's energy utilization efficiency is optimal.
  • Figure 4 shows the training method of the SVM model described in some embodiments of the present application.
  • the training method of the SVM model includes:
  • Step 401 Simulate and generate X sets of training data.
  • Each set of training data includes the number of users, the geographic location of the users, the file size to be transmitted, and the number K of wireless multicast user groups.
  • Step 402 Calculate the average contour coefficient of each group of data from the number of users in each group of training data, the user's geographic location, and the number K of wireless multicast user groups.
  • Step 403 Split the number of users in each group into K wireless multicast user groups.
  • the k-means algorithm may be used to split the number of users in each group into K wireless multicast user groups.
  • Step 404 Obtain K hovering positions corresponding to K wireless multicast user groups one-to-one.
  • the minimum circle coverage method may be used to obtain K hovering positions corresponding to K wireless multicast user groups one-to-one.
  • Step 405 Obtain the shortest flight path connecting the above K hovering positions.
  • a genetic algorithm may be used to obtain the shortest flight path connecting the above K hovering positions.
  • Step 406 Determine the total energy consumption of the drone according to the file size, hovering position, and flight path that need to be transmitted.
  • the communication energy consumption and hovering energy consumption of the drone can be determined from the file size to be transmitted and the hovering position; further, the flight path of the drone can be determined by the flight path. Energy consumption; Finally, the total energy consumption of the UAV is determined by the communication energy consumption, hovering energy consumption and flight energy consumption.
  • Step 407 optimizing the hovering position and the flight path to obtain the final hovering position and the final flight path, and the final total energy consumption corresponding to the final hovering position and the final flight path.
  • Step 408 Determine that the X group includes the number of users, the split number K of the wireless multicast user group, the file size to be transmitted and the data of the average contour coefficient, and the corresponding X group only includes the data of the final total energy consumption.
  • Step 409 Take the data including the number of users, the number of wireless multicast user groups K, the file size to be transmitted, and the average contour coefficient as input, and the data including the final total energy consumption as the output, and train the SVM model.
  • the step of optimizing the hover position and flight path in step 407 may include:
  • Step 4071 Use the interior point method to optimize the K hover positions one by one to obtain K optimized hover positions, and use genetic algorithm to obtain the shortest flight path connecting the K optimized hover positions, and obtain the optimized Total energy consumption;
  • Step 4072 Compare the total energy consumption after optimization with the total energy consumption before optimization. If the absolute value of the difference between the two is greater than the set threshold, repeat step 4071 on the basis of this optimization for the next round of optimization ; If the absolute value of the difference between the two is less than or equal to the set threshold, the optimized hover position and flight path are the final hover position and final flight path, and the optimized total energy consumption is the final total energy consumption.
  • the specific meaning of the above-mentioned on the basis of this optimization may refer to determining the total energy consumption after this optimization as the total energy consumption before the next round of optimization optimization; and determining K after optimization
  • the hovering positions of are used as the K hovering positions before the next round of optimization.
  • training data can be acquired first.
  • the relevant data can be simulated based on actual conditions.
  • this application will simulate and generate X sets of data, where the value of X is not less than 100000, and each set of data includes the number of users, user geographic location, required file size, and number of wireless multicast user groups And other attributes. Then, calculate the average contour coefficient of each group of data.
  • the average contour coefficient is usually used to evaluate the effect of k-means clustering.
  • the average contour coefficient can be calculated from the contour coefficient of each user.
  • the profile coefficient ⁇ (i) of user i can be calculated by the following formula (1):
  • b(i) represents the average distance between user i and all users in other wireless multicast user groups; a(i) represents the average distance between user i and all other users in its own wireless multicast user group.
  • the average contour coefficient is the average of the sum of the contour coefficients of all users. Since the average profile coefficient reflects the degree of aggregation of the same wireless multicast user group and the degree of dispersion of different wireless multicast user groups, it also indirectly reflects the geographical distribution of users; the final energy of each group of data can be determined by the method described above Calculated.
  • the data is simply processed to generate X groups of data including the number of users, the number of wireless multicast user groups, the file size to be transmitted and the average profile coefficient, and the corresponding X groups only include the final energy consumption data.
  • an SVM prediction model can be trained.
  • the calculation method for the number K of wireless multicast user groups may be as shown in FIG. 5, including:
  • Step 501 Determine the possible value K* of the number of wireless multicast user groups according to the number of users N.
  • K* is from 1 to N, that is, K* ⁇ [1,N].
  • Step 502 According to the possible value K* of the number of each specific wireless multicast user group, use the k-means algorithm to divide the number of users N into K* wireless multicast user groups, and calculate the average contour coefficient corresponding to K* ⁇ *.
  • Step 503 Input the number of users N, the file size S to be transmitted, the possible value K* of the number of specific wireless multicast user groups, and the average contour coefficient ⁇ * corresponding to K* into the trained SVM model to obtain the The predicted total energy consumption E K * corresponding to the possible value K* of the number of broadcast user groups.
  • the SVM model can obtain the predicted total energy consumption E K * of a UAV for each K* in the current scene.
  • Step 504 Use K * corresponding to the smallest predicted total energy consumption E K * as the number K of the above-mentioned wireless multicast user group.
  • P t represents the transmission power of the drone
  • H represents the height of the drone from the ground
  • ⁇ 2 represents the power of the additional Gaussian white noise
  • ⁇ 0 represents the channel power gain at 1 m from the drone
  • B represents the drone The bandwidth used for multicast transmission.
  • time T k used by the drone to transmit data in the wireless multicast user group k can be represented by the following formula (3):
  • S represents the file size that the drone needs to transmit.
  • P h the mechanical power of the drone when hovering.
  • the drone needs to fly over these hovering positions and return in turn. Assuming the flying distance is D, then The time spent by the drone V represents the flight rate of the drone.
  • the specific solution method can use the minimum circle coverage method to determine the initial hovering position u k of the UAV, that is, use a circle with the smallest radius to cover all users belonging to the same wireless multicast user group.
  • the center of the circle is, That is the initial hovering position u k
  • the specific steps can be shown in Figure 6, including:
  • Step 601 take any of a wireless multicast user group, the wireless multicast user from the user group contains three users of arbitrarily taken N a, N b, N c .
  • Step 602 make a containing N a, N b, N of the three smallest circle c.
  • Step 603 Find the point N d farthest from the center of the minimum circle among other users in the wireless multicast user group.
  • Step 604 If N d is already inside or on the circle, go to step 605; otherwise, go to step 606.
  • Step 605 The circle is the desired circle, and the center of the circle is the hovering position corresponding to the wireless multicast user group, and the algorithm ends.
  • Step 606 the N a, N b, N c , N d selected three points, so that a circle containing the four points generated by them to a minimum and to make three points becomes the new N a, N b, N c , return to step 602 and step 603.
  • FIG. 7 shows the flow of the method for determining the foregoing shortest flight path by using genetic algorithm according to an embodiment of the present invention. As shown in Figure 7, the method may include:
  • Step 701 randomly initialize L flight paths as the primary population, calculate the distance of the L paths, and sort the L path distances;
  • Step 702 According to the roulette strategy, a selection probability is set for each path according to the distance of each path in the L paths, where the path with the smaller distance sets the greater the selection probability, and makes L selections , Select a path from L paths for each selection, and proceed to the next step with the selected path;
  • Step 703 From the L paths obtained in step 702, choose two arbitrarily to cross paths with a preset probability and replace the previous path until all L paths have been selected, and the selection will not be repeated;
  • Step 704 From the L paths in the crossover process, perform a reversal of a certain path for each path in turn with a preset probability, and replace the previous path until all L paths have been selected, and the selection will not be repeated. Get a new generation of population;
  • Step 705 Bring the new-generation population back into the process from step 701 to step 704 and repeat it cyclically, setting the maximum number of cyclic repetitions until the obtained shortest flight path no longer changes or the number of repetitions meets the set number of times.
  • Q k represents the total energy consumed by the drone in other K-1 wireless multicast user groups.
  • the remaining hover positions can be optimized in turn.
  • the algorithm is considered to have converged, and the hovering position and flight path are the final solutions. Otherwise, a new round of iterative optimization is performed until convergence.
  • the UAV path optimization method described in this application fully considers the number and geographic distribution of users, and adopts a data-driven method to adaptively split users into several wireless multicast user groups, and each wireless multicast user group Find a suitable hovering position in the user group, so that the drone can transmit files at a higher transmission rate at this position, so as to reduce the hovering and transmission energy consumption of the drone, and fly the drone.
  • the trajectory is designed to find the shortest flight path through all hovering positions and return to the starting position to reduce the flight energy consumption of the UAV, and ultimately enable the UAV to take into account both mechanical energy and communication energy consumption, with the smallest total energy Consumption to complete the task of multicast service transmission.
  • This application significantly improves the use value of UAV energy and the efficiency of multicast transmission by designing the flight trajectory of the UAV.
  • This application splits users into adaptive wireless multicast user groups, so that the drone can select an appropriate transmission mode according to specific scenarios, and reduces the time and energy spent by the drone to perform tasks.
  • This application trains an SVM model, which can directly predict the appropriate number of wireless multicast user groups based on the number of users, geographic locations, and the size of the files to be transmitted, which reduces the computational complexity.
  • This application uses a data-driven method for modeling, and can output an accurate prediction value of the number of wireless multicast user groups in real time based on the model, which can greatly improve the efficiency of determining the number of wireless multicast user groups, and is extremely practical.
  • an embodiment of an electronic device for optimizing the path of a communication drone is provided.
  • an embodiment of the present application provides a UAV path optimization device.
  • Fig. 8 shows the structure of the UAV path optimization device according to an embodiment of the present application.
  • the aforementioned UAV path optimization equipment includes:
  • the information obtaining module 801 is used to obtain the number of wireless multicast users, the geographic location of each wireless multicast user, and the size of the file to be transmitted.
  • the user group quantity determining module 802 is configured to determine the number of wireless multicast user groups according to the number of wireless multicast users, the geographic location of each wireless multicast user, and the file size to be transmitted.
  • the grouping module 803 is configured to divide the wireless multicast users into multiple wireless multicast user groups according to the determined number of wireless multicast user groups.
  • the hovering position determining module 804 is used to determine the hovering position of the drone corresponding to each wireless multicast user group.
  • the flight path determination module 805 is configured to obtain the shortest flight path connecting the above hovering positions according to the determined hovering positions of the drone corresponding to each wireless multicast user group.
  • the first energy consumption determination module 806 is configured to determine the first total energy consumption of the drone according to the size of the file to be transmitted, the hovering position, and the flight path.
  • the optimization module 807 is used to optimize each of the above-mentioned hovering positions one by one to obtain multiple optimized hovering positions, and determine the shortest connecting the above-mentioned multiple optimized hovering positions according to the above-mentioned multiple optimized hovering positions Flight path.
  • the second energy consumption determining module 808 is configured to determine the second total energy consumption of the drone according to the size of the file to be transmitted, the optimized hovering position, and the flight path.
  • the comparison module 809 is used to compare the above-mentioned first total energy consumption with the second total energy consumption, and if the absolute value of the difference between the two is greater than a preset threshold value, instruct the optimization module 807 to perform the optimization based on the current round of optimization The next round of optimization; if the absolute value of the difference between the two is less than or equal to the preset threshold, the optimized hover position and flight path are determined as the hover position and flight path of the drone.
  • the above-mentioned user group quantity determining module 802 may use the method shown in FIG. 5 to determine the number of wireless multicast user groups.
  • the above-mentioned hovering position determining module 804 may use the method shown in FIG. 6 to determine the hovering position of the drone corresponding to each wireless multicast user group.
  • the flight path determination module 805 may use the method shown in FIG. 7 to determine the shortest flight path.
  • the grouping module 803, the first energy consumption determination module 806, the optimization module 807, the second energy consumption determination module 808, and the comparison module 809 can also be implemented by the specific method described in the above-mentioned UAV path optimization method, and will not be repeated here. .
  • Fig. 9 shows the hardware structure of the UAV path optimization device provided by an embodiment of the present application.
  • the above-mentioned electronic device for UAV path optimization includes at least one processor 902; and a memory 904 communicatively connected to the at least one processor; wherein, the memory 904 stores data that can be used by the An instruction executed by the at least one processor 902, the instruction being executed by the at least one processor, so that the at least one processor can execute any one of the methods described above.
  • the electronic device includes a processor 902 and a memory 904, and may also include an input device and an output device.
  • the processor, memory, input device, and output device may be connected by a bus or in other ways.
  • the memory can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as the calculation migration of the mobile terminal program in the embodiments of the present application
  • the program instruction/module corresponding to the method.
  • the processor executes various functional applications and data processing of the server by running non-volatile software programs, instructions, and modules stored in the memory, that is, realizing the UAV path optimization method of the above method embodiment.
  • the memory may include a storage program area and a storage data area, where the storage program area can store an operating system and an application program required by at least one function; the storage data area can store data created according to the use of the mobile terminal program's computing migration device, etc. .
  • the memory may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other non-volatile solid-state storage devices.
  • the memory may optionally include a memory remotely provided with respect to the processor. Examples of the above network include but are not limited to the Internet, intranet, local area network, mobile communication network, and combinations thereof.
  • the input device can receive inputted number or character information, and generate key signal input related to user settings and function control of the computing migration device of the mobile terminal program.
  • the output device may include a display device such as a display screen.
  • the one or more modules are stored in the memory, and when executed by the processor, the calculation migration method of the mobile terminal program in any of the foregoing method embodiments is executed. Any embodiment of the electronic device that executes the calculation migration method of the mobile terminal program can achieve the same or similar effect as any of the foregoing method embodiments.
  • the program can be stored in a computer readable storage medium. When executed, it may include the procedures of the above-mentioned method embodiments.
  • the storage medium can be a magnetic disk, an optical disc, a read-only memory (Read-Only Memory, ROM), or a random access memory (Random Access Memory, RAM).
  • the embodiment of the computer program can achieve the same or similar effect as any of the aforementioned method embodiments.
  • the method according to the present disclosure may also be implemented as a computer program executed by a CPU, and the computer program may be stored in a computer-readable storage medium.
  • the computer program is executed by the CPU, the above-mentioned functions defined in the method of the present disclosure are executed.
  • the above method steps and system units can also be implemented using a controller and a computer-readable storage medium for storing a computer program that enables the controller to implement the above steps or unit functions.
  • the well-known power/ground connections with integrated circuit (IC) chips and other components may or may not be shown in the drawings provided.
  • the devices may be shown in the form of block diagrams in order to avoid making the application difficult to understand, and this also takes into account the fact that the details of the implementation of these block diagram devices are highly dependent on the platform on which the application will be implemented (ie These details should be completely within the understanding of those skilled in the art).
  • specific details for example, a circuit
  • DRAM dynamic RAM

Abstract

一种无人机（1）路径优化方法，包括：获得用户集合（2）及其地理位置；将用户集合（2）分成K个无线多播用户组（3）；确定与K个无线多播用户组（3）对应的K个初始悬停位置，并确定连接K个初始悬停位置的最短的初始飞行路径；对无人机（1）初始悬停位置进行逐个优化，得到与K个无线多播用户组（3）相对应的K个最终悬停位置及最终飞行路径。还提供了一种无人机路径优化设备和计算机可读存储介质。

Description

无人机路径优化方法、设备及存储介质
本申请基于申请号为201910041506.X，申请日为2019年1月16日的中国专利申请提出，并要求该中国专利申请的优先权，该中国专利申请的全部内容在此引入本申请作为参考。
技术领域
本申请涉及无人机通信技术领域，具体的说是一种无人机路径优化方法、设备及存储介质。
背景技术
随着无人驾驶飞行器(简称为无人机)近年来在工业界的广泛应用，无人机通信也已经成为无线通信领域研究的热点。相比于传统的地面通信系统，无人机通信系统具有成本低，部署快，灵活性高，信道条件优良，移动可控等优点，因此无人机通信也将逐渐成为未来无线通信领域不可或缺的一部分。然而无人机无线通信并不完美，面临着许多挑战，其中能量不足问题是制约无人机通信发展的最大因素之一。
另一方面，无线多播已经成为用于多媒体传输的成熟技术，可以有效提高频谱利用率。然而传统的无线多播技术面临着一个十分具有挑战性的问题，即多播的传输速率本质上受限于多播组中最差的用户信道条件。
针对无线多播领域面临的问题以及结合当前无人机通信系统的优势和挑战，利用无人机通信实现无线多播已成为一个研究热点。利用无人机通信实现无线多播一方面可以借助无人机通信系统成本低，移动可控及信道条件优良等优点来解决无线多播传输速率受限的问题，另一方面也借助多播能够同时给多个用户传输数据的优势，有效降低无人机执行传输任务所需要的时间和能量消耗，达到互利共赢的目的。
发明内容
根据以上现有技术的不足，本申请提出了一种无人机路径优化方法、设备及存储介质，可以通过优化无人机悬停位置以及飞行路径，降低无人机的能耗。
本申请提出的无人机路径优化方法包括：
根据获取的无线多播用户的数量、各个无线多播用户的地理位置以及需要传输的文件大小确定无线多播用户组的数量；
根据所述无线多播用户组的数量将无线多播用户分成多个无线多播用户组；
确定无人机对应每个无线多播用户组的悬停位置以及连接所述悬停位置的最短飞行路径；
根据所述需要传输的文件的大小、所述悬停位置以及所述飞行路径确定无人机的 第一总能耗；
对所述各个悬停位置进行逐个优化，得到多个优化后的悬停位置以及连接所述多个优化后的悬停位置的最短的飞行路径；
根据所述需要传输的文件的大小、所述优化后的悬停位置以及所述飞行路径确定无人机的第二总能耗；以及
将所述第一总能耗与第二总能耗相比较，如果两者的差的绝对值大于预先设定的阈值，则在本轮优化的基础上返回至对所述各个悬停位置进行逐个优化的步骤进行下一轮优化；如果两者差的绝对值小于或等于预先设定的阈值，则将优化后的悬停位置及飞行路径确定为无人机的悬停位置和飞行路径。
其中，确定无线多播用户组的数量包括：利用训练好的支持向量机模型SVM，根据所述无线多播用户的数量、地理位置以及需要传输的文件大小确定无线多播用户组的数量。
其中，确定无线多播用户组的数量包括：
根据用户数目N，确定无线多播用户组数量的可能值K*；
根据每一个无线多播用户组数量的可能值K*，使用k-means算法将用户数目N分成K*个无线多播用户组，并计算出与K*对应的平均轮廓系数λ*；
将用户数目N，需传输的文件大小S，无线多播用户组数量的可能值K*以及与K*对应的平均轮廓系数λ*输入训练好的SVM模型，得到无线多播用户组数量的可能值K*对应的预测总能耗E K*；以及
将最小的预测总能耗E K*所对应的K*作为所述无线多播用户组的数量K。
其中，训练SVM模型包括：
模拟生成X组训练数据，每组训练数据包括用户数目、用户地理位置、需要传输的文件大小和无线多播用户组数目K；
由每组训练数据中的用户数目、用户地理位置以及无线多播用户组数目K计算得到每组数据的平均轮廓系数；
将每组中的用户数目分裂成K个无线多播用户组；
获得与K个无线多播用户组一一对应的K个悬停位置以及连接所述K个悬停位置的最短的飞行路径；
根据所述需要传输的文件大小与悬停位置以及飞行路径确定无人机的总能耗；
对所述悬停位置和飞行路径进行优化，从而得到最终悬停位置和最终飞行路径，以及与最终悬停位置和最终飞行路径所对应的最终总能耗；
确定X组包括用户数目、无线多播用户组分裂数目K、需要传输的文件大小和平均轮廓系数的数据以及对应的X组只包含最终总能耗的数据；以及
将包含用户数目、无线多播用户组数量K、需要传输的文件大小和平均轮廓系数的数据作为输入，将包含最终总能耗的数据作为输出，训练得到SVM模型。
其中，根据所述无线多播用户组的数量将无线多播用户分成多个无线多播用户组 包括：利用k均值聚类算法k-means算法将无线多播用户分成多个无线多播用户组。
其中，确定无人机对应每个无线多播用户组的悬停位置包括：利用最小圆覆盖法分别确定无人机对应每个无线多播用户组的悬停位置。
其中，利用最小圆覆盖法分别确定无人机对应每个无线多播用户组的悬停位置包括：
A、任取一个无线多播用户组，从所述无线多播用户组里包含的用户中随意取出三个用户N a、N b、N c；
B、作一个包含N a、N b、N c三点的最小圆；
C、在所述无线多播用户组其他用户中找出距离所述最小圆圆心最远的点N d；
D、若N d已在圆内或圆周上，则所述最小圆为所求的圆，此圆的圆心为该无线多播用户组对应的悬停位置；否则，在N a、N b、N c、N d中选3个点，使由它们生成的一个包含这4个点的圆为最小，并令这3点成为新的N a、N b、N c，返回B。
其中，确定连接所述悬停位置的最短飞行路径包括：利用遗传算法、粒子群算法、蚁群算法、模拟退火算法及神经网络算法中的任意一种获得所述最短的飞行路径；
确定连接所述优化后的悬停位置的最短飞行路径包括：利用遗传算法、粒子群算法、蚁群算法、模拟退火算法及神经网络算法中的任意一种获得所述最短的飞行路径。
其中，利用遗传算法获得所述最短的飞行路径包括：
a、随机初始化L条飞行路径作为初代种群，计算所述L条路径的距离，并对L条路径距离进行排序；
b、根据轮盘赌策略，根据L条路径中每条路径的距离，为每条路径分别设定一个选择概率，其中距离越小的路径设定的选择概率越大，并做L次选择，每次选择均从L条路径中选择一条路径，被选中的路径再进行下一步操作；
c、从所述L条路径中，依次任取两条以预设概率进行路径交叉并取代之前的路径，直至L条路径都被选择过，且不会重复选择；
d、从交叉过程的L条路径中，依次对每一条路径以预设概率进行路径的反转，并取代之前的路径，直至L条路径都被选择过，且不会重复选择，得到新一代种群；
e、将新一代种群重新带入a至d的过程并循环重复，设定循环重复的最大次数，直到获得的最短飞行路径不再变化或重复次数满足设定的次数。
其中，根据所述需要传输的文件的大小、所述悬停位置以及所述飞行路径确定无人机的第一总能耗包括：
根据所述需要传输的文件大小与所述悬停位置确定传输文件所需的时间以及无人机的发射功率；
根据所述传输文件所需的时间以及无人机的发射功率确定无人机的通信能耗和悬停能耗；
根据所述飞行路径确定无人机的飞行能耗；以及
由确定的通信能耗、悬停能耗以及飞行能耗确定无人机的总能耗。
其中，对所述各个悬停位置进行逐个优化包括：利用内点法对所述悬停位置进行逐个优化。
本申请还提供了一种无人机路径优化设备，包括：
用户组数量确定模块，用于根据无线多播用户的数量、各个无线多播用户的地理位置以及需要传输的文件大小确定无线多播用户组的数量；
分组模块，用于根据确定的无线多播用户组的数量将无线多播用户分成多个无线多播用户组；
悬停位置确定模块，用于确定无人机对应每个无线多播用户组的悬停位置；
飞行路径确定模块，用于根据确定的无人机对应每个无线多播用户组的悬停位置获得连接所述悬停位置的最短飞行路径；
第一能耗确定模块，用于根据所述需要传输的文件的大小、悬停位置以及飞行路径确定无人机的第一总能耗；
优化模块，用于对所述各个悬停位置进行逐个优化，得到多个优化后的悬停位置，以及根据多个优化后的悬停位置确定连接多个优化后的悬停位置的最短的飞行路径；
第二能耗确定模块，用于根据需要传输的文件的大小、优化后的悬停位置以及所述飞行路径确定无人机的第二总能耗；以及
比较模块，用于将所述第一总能耗与第二总能耗相比较，如果两者的差的绝对值大于预先设定的阈值，则在本轮优化的基础上指示优化模块进行下一轮优化；如果两者差的绝对值小于或等于预先设定的阈值，则将优化后的悬停位置及飞行路径确定为无人机的悬停位置和飞行路径。
本申请还提供了一种无人机路径优化设备，包括：至少一个处理器；以及与所述至少一个处理器通信连接的存储器；其中，所述存储器存储有可被所述至少一个处理器执行的指令，所述指令被所述至少一个处理器执行，以使所述至少一个处理器能够执行上述无人机路径优化方法。
本申请还提供了种计算机可读存储介质，其上存储有计算机指令，在处理器执行上述计算机指令时实现上述的无人机路径优化方法。
本申请可以充分考虑用户的数目和地理分布情况，采用数据驱动的方法，将用户自适应的分裂成若干个多播组，并在每个多播组中找到一个合适的悬停位置，以供无人机能够在该位置以较高的传输速率来传输文件，达到减少无人机悬停和传输能量消耗的目的。
更进一步，本申请可以对无人机飞行轨迹进行设计，找到一条经过所有悬停位置并返回起始位置的最短飞行路径，来降低无人机的飞行能量消耗，最终使得无人机能够兼顾机械能量和通信能量消耗，以最小的总能量消耗来完成多播业务传输任务。
附图说明
为了更清楚地说明本申请实施例或现有技术中的技术方案，下面将对实施例或现 有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本申请的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1为本申请实施例所述利用无人机通信实现无线多播的实现场景示意图；
图2为本申请实施例所述的无人机路径优化方法流程示意图；
图3为本申请实施例所述的无人机路径优化方法流程示例；
图4为本申请实施例所述的SVM模型训练方法的流程图；
图5为本申请实施例所述的无线多播用户组数目K的计算方法流程示意图；
图6为本申请实施例所述的采用最小圆覆盖法确定所述无人机悬停位置的方法流程示意图；
图7为本发明实施例所述的采用遗传算法确定最短飞行路径的方法流程示意图；
图8为本申请实施例所述的无人机路径优化设备的结构示意图；以及
图9为本申请实施例所述的无人机路径优化设备的硬件结构示意图。
具体实施方式
如前所述，利用无人机通信实现无线多播已成为一个研究热点。目前对于无人机多播通信系统的研究主要是通过对无人机飞行轨迹的设计，来提高无人机与用户间的通信性能。然而，现有的无人机飞行轨迹设计通常都不会考虑无人机执行任务时的能量消耗问题，从而频繁出现无人机在通信过程中由于能量不足而导致通信中断的问题。
图1显示了本申请实施例所述的利用无人机通信实现无线多播的实现场景示意图。如图1所示，当地面通信基础设施被自然灾害破坏，或偏远地区还没有地面通信基础设施的情况下，需要利用无人机1来多播传输通用文件，执行无线多播任务的无人机1需要向由多个无线多播用户4组成的无线多播用户集合2提供无线多播服务。同时，由于无线通信的多播传输速率取决于信道最差的用户的信道条件，为了提高无线多播服务的质量和效率，还可以进一步对上述无线多播用户集合2中的无线多播用户4进行分组，得到多个无线多播用户组3，由上述无人机1分别为各个无线多播用户组提供无线多播服务。
考虑到无人机能量消耗的问题，本申请的一些实施例提出一种无人机路径优化方法。该方法首先可以将多个无线多播用户分成多个无线多播用户组；然后，在每个无线多播用户组内找到一个悬停点；最后，再找到一条经过上述多个无线多播用户组的悬停点的最短飞行路径，使得无人机完成多播传输任务所消耗的总能量最小，从而达到降低无人机完成多播传输任务所消耗的总能量，也即降低总能耗，提高能量利用效率的目的。
在本申请的实施例中，上述总能耗可以包括机械能耗以及通信能耗。其中，机械能耗又可以包括无人机在悬停位置处盘旋所消耗的能量，称为悬停能耗；以及无人机在悬停位置间飞行所消耗的能量，称为飞行能耗。其中，通信能耗是指无人机进行数 据传输所消耗的能量。
下面结合附图对本申请的实施例作进一步详细的说明，以帮助本领域技术人员对本申请的发明构思、技术方案有更完整、准确和深入的理解。
图2显示了本申请实施例所述的无人机路径优化方法流程。在本申请的实施例中，该方法可以由与无人机通信的其他计算设备执行。如图2所示，该方法可以包括：
步骤201：获得无线多播用户的数量、各个无线多播用户的地理位置以及需要传输的文件的大小。
在本申请的实施例中，可以通过GPS等定位方法及任务具体需要传输的文件获得无线多播用户的数量、地理位置以及需要传输的文件大小。
步骤202：根据无线多播用户的数量、各个无线多播用户的地理位置以及需要传输的文件大小确定无线多播用户组的数量。
在本申请的一些实施例中，可以利用训练好的支持向量机(SVM，Support Vector Machine)模型根据无线多播用户的数量、地理位置以及需要传输的文件大小确定无线多播用户组的数量。
在本申请的另一些实施例中，还可以利用训练好的神经网络根据无线多播用户的数量、地理位置以及需要传输的文件大小确定无线多播用户组的数量。
步骤203：根据确定的无线多播用户组的数量将无线多播用户分成多个无线多播用户组。
在本申请的实施例中，可以利用k均值聚类算法(k-means)算法或其他聚类算法将将无线多播用户分成多个无线多播用户组。
步骤204：确定无人机对应每个无线多播用户组的悬停位置。
在本申请的实施例中，可以利用最小圆覆盖法分别获得与上述多个无效多播用户组一一对应的悬停位置。
步骤205：根据确定的无人机对应每个无线多播用户组的悬停位置获得连接上述悬停位置的最短飞行路径。
在本申请的实施例中，可以利用遗传算法、粒子群算法、蚁群算法、模拟退火算法及神经网络算法中的任意一种获得上述最短的飞行路径。
步骤206：根据上述需要传输的文件的大小、上述悬停位置以及上述飞行路径确定无人机的第一总能耗。
在本申请的实施例中，可以根据上述需要传输的文件的大小与上述悬停位置确定无人机的通信能耗和悬停能耗。其中，根据上述需要传输的文件的大小与上述悬停位置可以确定传输文件所需的时间以及无人机的发射功率，从而可以确定无人机的通信能耗和悬停能耗。进一步，还可以根据上述飞行路径确定无人机的飞行能耗。最后，再由确定的通信能耗、悬停能耗以及飞行能耗确定无人机的总能耗。
步骤207：对上述各个悬停位置进行逐个优化，得到多个优化后的悬停位置。
在本申请的实施例中，可以利用内点法或者其他非线性规划解决方法对上述悬停 位置进行逐个优化。
步骤208：根据上述多个优化后的悬停位置确定连接上述多个优化后的悬停位置的最短的飞行路径。
在本申请的实施例中，上述步骤208可以采用与上述步骤205相同的方法实现。
步骤209：根据上述需要传输的文件的大小、上述优化后的悬停位置以及上述飞行路径确定无人机的第二总能耗。
步骤210：将上述第一总能耗与第二总能耗相比较，如果两者的差的绝对值大于预先设定的阈值，则在本轮优化的基础上返回上述步骤207进行下一轮优化；如果两者差的绝对值小于或等于预先设定的阈值，则执行步骤211。
在本申请的实施例中，如果两者的差的绝对值大于预先设定的阈值，则可以首先将上述第一总能耗设置为上述第二总能耗，然后返回上述步骤207。
步骤211：将优化后的悬停位置及飞行路径确定为无人机的悬停位置和飞行路径。
图3给出了本申请一个实施例所述的无人机路径优化方法流程的一个具体示例。如图3所示，可以通过如下步骤确定无人机的各个悬停位置和飞行路径。
步骤S1、获得无线多播用户数目N、各个无线多播用户的地理位置W n以及需要传输的文件大小S；
步骤S2、利用训练好的SVM模型和用户数目N、地理位置W n及需传输的文件大小S计算得到无线多播用户组的数量K，使用k-means算法将无线多播用户的数目N分裂成K个无线多播用户组，简称无线多播用户组；
步骤S3、利用最小圆覆盖法获得与K个无线多播用户组一一对应的K个悬停位置，使用遗传算法获得连接K个悬停位置的最短的飞行路径，由所述文件大小S与所述悬停位置确定无人机的通信能耗和悬停能耗，由所述飞行路径确定飞行能耗，由所述通信能耗、悬停能耗以及飞行能耗构成无人机总能耗；
步骤S4、利用内点法对悬停位置进行逐个优化，得到优化后的悬停位置，使用遗传算法获得连接K个优化后的悬停位置的最短的飞行路径，并由优化后的悬停位置与飞行路径得到优化后的无人机总能耗；
步骤S5、将优化后的无人机总能耗与优化前的无人机总能耗相比较，如果两者的差的绝对值大于设定的阈值，则在本轮优化的基础上重复步骤S4进行下一轮优化；如果两者差的绝对值小于或等于设定的阈值，则优化后的无人机悬停位置及飞行路径即为最终悬停位置和最终飞行路径。
由于无人机多播传输速率受限于信道条件最差的用户，而信道条件的好坏主要取决于无线多播用户到无人机的距离。为提升无人机多播传输速率，减少无人机传输时间及能量消耗。在本申请的实施例中，如图1、2和3所示，在获得无线多播用户集合2及其中无线多播用户的地理位置后，首先求出最优的无线多播用户组数目K；然后将无线多播用户集合分裂成K个无线多播用户组；再然后相对于每个无线多播用户组均确定一个初始悬停位置。在该初始悬停位置下，无人机向位于该组内的用户传送文件 时，传输速率较快，通信能耗较低，同时由于传输速率快，因此传输时间短，即悬停时间短，进而使悬停能耗降低。当与K个无线多播用户组一一对应的K个悬停位置确定后，再确定连接该K个悬停位置的最短飞行路线，从而使飞行能耗最低。由于无人机的总能耗基本等于通信能耗、悬停能耗和飞行能耗之和，因此，当确定无人机悬停位置和飞行路径后，即可确定无人机总能耗。接下来，可以进一步对上述悬停位置进行优化，再由优化后得到的悬停位置确定优化后的最短飞行路径，最终确定优化后最低的无人机总能耗，并最终确定与该最低的无人机总能耗相对应的无人机的悬停位置与飞行路径。
本申请将多个无线多播用户自适应地分裂成多个无线多播用户组，在每个无线多播用户组里找到一个合适的悬停点，并找到一条经过这些悬停点的最短飞行路径，使得一架无人机完成多播传输任务所消耗的总能量最小，也即使无人机的能量利用效率达到最优。
下面将结合附图详细说明上述SVM模型的训练方法。图4显示了本申请一些实施例所述的SVM模型的训练方法。如图4所示，该SVM模型的训练方法包括：
步骤401：模拟生成X组训练数据，每组训练数据包括用户数目、用户地理位置、需要传输的文件大小和无线多播用户组数目K。
步骤402：由每组训练数据中的用户数目、用户地理位置以及无线多播用户组数目K计算得到每组数据的平均轮廓系数。
步骤403：将每组中的用户数目分裂成K个无线多播用户组。
在本申请的实施例中，可以利用k-means算法将每组中的用户数目分裂成K个无线多播用户组。
步骤404：获得与K个无线多播用户组一一对应的K个悬停位置。
在本申请的实施例中，可以利用最小圆覆盖法获得与K个无线多播用户组一一对应的K个悬停位置。
步骤405：获得连接上述K个悬停位置的最短的飞行路径。
在本申请的实施例中，可以使用遗传算法获得连接上述K个悬停位置的最短的飞行路径。
步骤406：根据上述需要传输的文件大小与悬停位置以及飞行路径确定无人机的总能耗。
在本发明的实施例中，首先，可以由上述需要传输的文件大小与上述悬停位置确定无人机的通信能耗和悬停能耗；进一步，可以由上述飞行路径确定无人机的飞行能耗；最后，再由通信能耗、悬停能耗以及飞行能耗确定无人机的总能耗。
步骤407：对悬停位置和飞行路径进行优化，从而得到最终悬停位置和最终飞行路径，以及与最终悬停位置和最终飞行路径所对应的最终总能耗。
步骤408：确定X组包括用户数目、无线多播用户组分裂数目K、需要传输的文件大小和平均轮廓系数的数据以及对应的X组只包含最终总能耗的数据。
步骤409：将包含用户数目、无线多播用户组数量K、需要传输的文件大小和平均轮廓系数的数据作为输入，将包含最终总能耗的数据作为输出，训练得到SVM模型。
具体的，在本申请的实施例中，在SVM模型训练过程中，上述步骤407中对悬停位置及飞行路径进行优化的步骤可以包括：
步骤4071：利用内点法对K个悬停位置进行逐个优化，得到K个优化后的悬停位置，使用遗传算法获得连接K个优化后的悬停位置的最短的飞行路径，得到优化后的总能耗；
步骤4072：将优化后的总能耗与优化前的总能耗相比较，如果两者的差的绝对值大于设定的阈值，则在本次优化的基础上重复步骤4071进行下一轮优化；如果两者差的绝对值小于或等于设定的阈值，则优化后的悬停位置及飞行路径即为最终悬停位置和最终飞行路径，优化后的总能耗即为最终总能耗。
在本申请的实施例中，上述在本次优化的基础上的具体含义可以是指将本次优化后的总能耗确定为下一轮优化优化前的总能耗；以及将K个优化后的悬停位置作为下一轮优化优化前的K个悬停位置。
具体的，在本申请的实施例中，首先可以获取训练数据。
在实际应用中，可以根据实际情况，模拟产生相关数据。为了达到满意的精度，本申请将模拟产生X组数据，其中，X的取值不低于100000，且每一组数据包括用户数目、用户地理位置、所需文件大小和无线多播用户组数量等属性。然后，计算出上述每组数据的平均轮廓系数。其中，平均轮廓系数通常被用来评价k-means聚类效果的好坏，当用户数目、位置及无线多播用户组数量确定后，平均轮廓系数可以通过各个用户的轮廓系数计算得到。其中，用户i的轮廓系数λ(i)可以通过如下公式(1)计算得到：
Figure PCTCN2019124036-appb-000001
其中，b(i)表示用户i到其他无线多播用户组内所有用户的距离的平均值；a(i)表示用户i到自身无线多播用户组内其他所有用户的距离的平均值。
在本申请的实施例中，平均轮廓系数是所有用户的轮廓系数的和的平均。由于平均轮廓系数反映了同一无线多播用户组的聚合程度，以及不同无线多播用户组的离散程度，也间接反应了用户的地理分布情况；每组数据的最终能量可以通过上文所述方法计算得到。
然后，对数据进行简单的处理，产生X组包括用户数目、无线多播用户组数量、所需传输的文件大小和平均轮廓系数的数据和与之对应的X组只包括最终能量消耗的数据，将前者作为输入数据，后者作为输出数据，可以训练出一个SVM的预测模型。
当遇到一个新的场景时，即无人机需要给若干数目的用户传输一个通用文件时， 获取该场景的相关属性，即获取用户的数目N、地理位置W n、所需传输的文件大小S，并针对不同无线多播用户组数量K∈[1,N]，计算出其平均轮廓系数，并将上述数据代入之前训练好的SVM预测模型，即可预测出在不同的无线多播用户组数量K∈[1,N]下的能量消耗，也即总能耗E K,比较找到最小的总能耗所对应的无线多播用户组数量即可。
作为本申请的一种可选的实施方式，上述无线多播用户组数目K的计算方法可以如图5所示，包括：
步骤501：根据用户数目N，确定无线多播用户组数量的可能值K*。
具体地，K*的取值是从1到N，也即K*∈[1,N]。
步骤502：根据每一个具体的无线多播用户组数量的可能值K*，使用k-means算法将用户数目N分成K*个无线多播用户组，并计算出与K*对应的平均轮廓系数λ*。
步骤503：将用户数目N，需传输的文件大小S，具体的无线多播用户组数量的可能值K*以及与K*对应的平均轮廓系数λ*输入训练好的SVM模型，得到与无线多播用户组数量可能值K*对应的预测总能耗E K*。
由于K*的取值是从1到N，所以SVM模型对当前场景下的每一个K*都可以得到一个无人机的预测总能耗E K*。
步骤504：将最小的预测总能耗E K*所对应的K*作为上述无线多播用户组的数量K。
在本申请的实施例中，为了让无人机在每个无线多播用户组内可以拥有较高的数据传输速率，可以在每个无线多播用户组中先找一个传输数据的初始悬停位置u k＝(x k,y k,H)，当无人机飞到该位置时，将保持悬停并传输数据。根据香农公式，无人机在无线多播用户组k中的最大传输速率R k可以如下公式(2)所示：
Figure PCTCN2019124036-appb-000002
其中，P t表示无人机传输功率；H表示无人机距离地面的高度；σ 2表示附加高斯白噪声的功率；β 0表示距离无人机1m处的信道功率增益；B表示无人机多播传输所用的带宽。
进而无人机在无线多播用户组k中传输数据所用的时间T k可以如下公式(3)所示：
Figure PCTCN2019124036-appb-000003
其中，S表示无人机需要传输的文件大小。
如此，无人机在所有无线多播用户组中传输消耗的总时间为
Figure PCTCN2019124036-appb-000004
则无人机用来传输数据消耗的总能量为E t＝P tT t。不考虑误差的情况下，无人机在各个悬停位置的悬停总时间等于无人机传输信息的总时间，即T h＝T t，用来悬停消耗的总能量为E h＝P hT h，其中P h表示无人机悬停时的机械功率。此外，无人机需要依次飞过这些悬 停位置并返回，假设飞行距离为D，则
Figure PCTCN2019124036-appb-000005
则无人机飞行所消耗的时间
Figure PCTCN2019124036-appb-000006
其中V表示无人机的飞行速率，为更快完成传输任务，假定无人机以最大速率飞行，无人机飞行消耗的能量为E f＝P fT f，其中P f表示无人机飞行时消耗的机械功率。
显然，在本申请的实施例中，为最小化无人机执行任务消耗的总能量，需要找到合适的无线多播用户组数目，并找到每个无线多播用户组内合适的悬停位置以及合适的最终飞行路径，具体求解方法可以采用最小圆覆盖法确定所述无人机初始悬停位置u k即用一个半径最小的圆覆盖属于同一个无线多播用户组的所有用户，其圆心，即为初始悬停位置u k，具体步骤可以如图6所示，包括：
步骤601：任取一个无线多播用户组，从该无线多播用户组里包含的用户中随意取出三个用户N a、N b、N c。
步骤602：作一个包含N a、N b、N c三点的最小圆。
步骤603：在该无线多播用户组其他用户中找出距离上述最小圆圆心最远的点N d。
步骤604：若N d已在圆内或圆周上，则执行步骤605；否则，执行步骤606。
步骤605：该圆即为所求的圆，且此圆的圆心为该无线多播用户组对应的悬停位置，算法结束。
步骤606：在N a、N b、N c、N d中选3个点，使由它们生成的一个包含这4个点的圆为最小，并令这3点成为新的N a、N b、N c，返回步骤602和步骤603。
待确定全部的无线多播用户组的悬停位置后，无人机需要依次飞过这些悬停位置，并在悬停位置处传输文件，为节约无人机飞行能量，降低飞行能耗，需要找到一条通过所有悬停位置并返回起始位置的最短飞行路径，该问题等价于著名的旅行销售商问题，可用现有的方法求解，例如遗传算法、粒子群算法、蚁群算法、模拟退火算法及神经网络算法中的任意一种。图7显示了本发明实施例所述的采用遗传算法确定上述最短飞行路径的方法流程。如图7所示该方法可以包括：
步骤701：随机初始化L条飞行路径作为初代种群，计算这L条路径的距离，并对L条路径距离进行排序；
步骤702：根据轮盘赌策略，根据L条路径中每条路径的距离，为每条路径分别设定一个选择概率，其中距离越小的路径设定的选择概率越大，并做L次选择，每次选择均从L条路径中选择一条路径，被选中的路径再进行下一步操作；
步骤703：从步骤702得到的L条路径中，依次任取两条以预设概率进行路径交叉并取代之前的路径，直至L条路径都被选择过，且不会重复选择；
步骤704：从交叉过程的L条路径中，依次对每一条路径以预设概率进行某段路径的反转，并取代之前的路径，直至L条路径都被选择过，且不会重复选择，得到新一代种群；
步骤705：将新一代种群重新带入步骤701至步骤704的过程并循环重复，设定循环重复的最大次数，直到获得的最短飞行路径不再变化或重复次数满足设定的次数。
综上，初始的飞行路径确定后，无人机执行传输任务所消耗的总能量如公式(4)所示：
Figure PCTCN2019124036-appb-000007
由于在确定悬停位置时，只考虑了无人机悬停的机械能量消耗和传输时的通信能量消耗，没有考虑到经过所有悬停位置的飞行路径的长短，可能导致无人机飞行所消耗的能量过大，因此需要对各个无线多播用户组中的悬停位置进行了进一步优化，具体优化方法如下：
任取一个悬停位置u k，并固定其他悬停位置，根据上述总能量E的公式，优化问题的数学表达式可如下式(5)所示：
Figure PCTCN2019124036-appb-000008
其中，Q k表示无人机在其他K-1个无线多播用户组所消耗的总能量。
引入辅助变量η，令
Figure PCTCN2019124036-appb-000009
即
Figure PCTCN2019124036-appb-000010
在此情况下，优化问题变为：
Figure PCTCN2019124036-appb-000011
Figure PCTCN2019124036-appb-000012
这是一个非线性优化问题，可用一种求解非线性优化的算法，如内点法来优化这一悬停位置。同理，剩下来的悬停位置可依次获得优化。当所有悬停位置优化完成后，由于和初始位置相比发生了变化，因此需要根据之前的方法(遗传算法、粒子群算法、蚁群算法、模拟退火算法及神经网络算法中的任意一种)重新找一条经过这些悬停位置的最短飞行路径，然后求出无人机在新悬停位置和路径下完成传输任务的总能量消耗，若此能量消耗与之前能量消耗的差的绝对值小于设定的阈值，则认为算法已经收敛，此悬停位置及飞行路径即为最终确定的解，否则，进行新一轮的迭代优化，直至收敛。
本申请所述的无人机路径优化方法充分考虑了用户的数目和地理分布情况，采用 数据驱动的方法，将用户自适应的分裂成若干个无线多播用户组，并在每个无线多播用户组中找到一个合适的悬停位置，以供无人机能够在该位置以较高的传输速率来传输文件，达到减少无人机悬停和传输能量消耗的目的，并对无人机飞行轨迹进行设计，找到一条经过所有悬停位置并返回起始位置的最短飞行路径，来降低无人机的飞行能量消耗，最终使得无人机能够兼顾机械能量和通信能量消耗，以最小的总能量消耗来完成多播业务传输任务。
本申请通过对无人机的飞行轨迹进行设计，显著提升了无人机能量的使用价值和多播传输效率。
本申请通过对用户进行自适应无线多播用户组分裂，使得无人机可以根据具体场景选择合适的传输方式，减少了无人机执行任务所花费的时间和能量。
本申请训练了一个SVM模型，可以根据用户的数目、地理位置和所需传输文件的大小，直接预测出合适的无线多播用户组数目，降低了计算复杂度。
本申请利用数据驱动方法进行建模，可基于模型实时输出精确的无线多播用户组数目的预测值，可大幅提高确定无线多播用户组数目的效率，有极强的实用性。
作为本申请实施例的第二个方面，提供了一种优化通信无人机路径的电子设备的一个实施例。
对应上述无人机路径优化方法，本申请的实施例提供了一种无人机路径优化设备。图8示出了本申请实施例所述的无人机路径优化设备的结构。如图8所示，上述无人机路径优化设备包括：
信息获取模块801，用于获得无线多播用户的数量、各个无线多播用户的地理位置以及需要传输的文件的大小。
用户组数量确定模块802，用于根据无线多播用户的数量、各个无线多播用户的地理位置以及需要传输的文件大小确定无线多播用户组的数量。
分组模块803，用于根据确定的无线多播用户组的数量将无线多播用户分成多个无线多播用户组。
悬停位置确定模块804，用于确定无人机对应每个无线多播用户组的悬停位置。
飞行路径确定模块805，用于根据确定的无人机对应每个无线多播用户组的悬停位置获得连接上述悬停位置的最短飞行路径。
第一能耗确定模块806，用于根据上述需要传输的文件的大小、上述悬停位置以及上述飞行路径确定无人机的第一总能耗。
优化模块807，用于对上述各个悬停位置进行逐个优化，得到多个优化后的悬停位置，以及根据上述多个优化后的悬停位置确定连接上述多个优化后的悬停位置的最短的飞行路径。
第二能耗确定模块808，用于根据上述需要传输的文件的大小、上述优化后的悬停位置以及上述飞行路径确定无人机的第二总能耗。
比较模块809，用于将上述第一总能耗与第二总能耗相比较，如果两者的差的绝 对值大于预先设定的阈值，则在本轮优化的基础上指示优化模块807进行下一轮优化；如果两者差的绝对值小于或等于预先设定的阈值，则将优化后的悬停位置及飞行路径确定为无人机的悬停位置和飞行路径。
在本申请的实施例中，上述用户组数量确定模块802可以采用图5所示的方法确定无线多播用户组的数量。上述悬停位置确定模块804可以采用上述图6所示的方法确定无人机对应每个无线多播用户组的悬停位置。上述飞行路径确定模块805可以采用上述图7所示的方法确定最短飞行路径。上述分组模块803、第一能耗确定模块806、优化模块807、第二能耗确定模块808以及比较模块809也可以采用上述无人机路径优化方法所述的具体方法实现，在此不再赘述。
图9示出了本申请一个实施例提供的无人机路径优化设备的硬件结构。如图9所示，上述用于无人机路径优化的电子设备包括至少一个处理器902；以及与所述至少一个处理器通信连接的存储器904；其中，所述存储器904存储有可被所述至少一个处理器902执行的指令，所述指令被所述至少一个处理器执行，以使所述至少一个处理器能够执行如上所述任意一种方法。
该电子设备中包括一个处理器902以及一个存储器904，并还可以包括：输入装置和输出装置。处理器、存储器、输入装置和输出装置可以通过总线或者其他方式连接。存储器作为一种非易失性计算机可读存储介质，可用于存储非易失性软件程序、非易失性计算机可执行程序以及模块，如本申请实施例中的所述移动终端程序的计算迁移方法对应的程序指令/模块。处理器通过运行存储在存储器中的非易失性软件程序、指令以及模块，从而执行服务器的各种功能应用以及数据处理，即实现上述方法实施例的无人机路径优化方法。
存储器可以包括存储程序区和存储数据区，其中，存储程序区可存储操作系统、至少一个功能所需要的应用程序；存储数据区可存储根据移动终端程序的计算迁移装置的使用所创建的数据等。此外，存储器可以包括高速随机存取存储器，还可以包括非易失性存储器，例如至少一个磁盘存储器件、闪存器件、或其他非易失性固态存储器件。在一些实施例中，存储器可选包括相对于处理器远程设置的存储器。上述网络的实例包括但不限于互联网、企业内部网、局域网、移动通信网及其组合。
输入装置可接收输入的数字或字符信息，以及产生与移动终端程序的计算迁移装置的用户设置以及功能控制有关的键信号输入。输出装置可包括显示屏等显示设备。所述一个或者多个模块存储在所述存储器中，当被所述处理器执行时，执行上述任意方法实施例中的移动终端程序的计算迁移方法。所述执行所述移动终端程序的计算迁移方法的电子设备的任何一个实施例，可以达到与之对应的前述任意方法实施例相同或者相类似的效果。
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流程，是可以通过计算机程序来指令相关硬件来完成，所述的程序可存储于一计算机可读取存储介质中，该程序在执行时，可包括如上述各方法的实施例的流程。其中，所述的存储介 质可为磁碟、光盘、只读存储记忆体(Read-Only Memory，ROM)或随机存储记忆体(Random Access Memory，RAM)等。所述计算机程序的实施例，可以达到与之对应的前述任意方法实施例相同或者相类似的效果。
此外，根据本公开的方法还可以被实现为由CPU执行的计算机程序，该计算机程序可以存储在计算机可读存储介质中。在该计算机程序被CPU执行时，执行本公开的方法中限定的上述功能。
此外，上述方法步骤以及系统单元也可以利用控制器以及用于存储使得控制器实现上述步骤或单元功能的计算机程序的计算机可读存储介质实现。
所属领域的普通技术人员应当理解：以上任何实施例的讨论仅为示例性的，并非旨在暗示本公开的范围(包括权利要求)被限于这些例子；在本申请的思路下，以上实施例或者不同实施例中的技术特征之间也可以进行组合，步骤可以以任意顺序实现，并存在如上所述的本申请的不同方面的许多其它变化，为了简明它们没有在细节中提供。
本申请的实施例旨在涵盖落入所附权利要求的宽泛范围之内的所有这样的替换、修改和变型。因此，凡在本申请的精神和原则之内，所做的任何省略、修改、等同替换、改进等，均应包含在本申请的保护范围之内。
所属领域的普通技术人员应当理解：以上任何实施例的讨论仅为示例性的，并非旨在暗示本公开的范围(包括权利要求)被限于这些例子；在本申请的思路下，以上实施例或者不同实施例中的技术特征之间也可以进行组合，步骤可以以任意顺序实现，并存在如上所述的本申请的不同方面的许多其它变化，为了简明它们没有在细节中提供。
另外，为简化说明和讨论，并且为了不会使本申请难以理解，在所提供的附图中可以示出或可以不示出与集成电路(IC)芯片和其它部件的公知的电源/接地连接。此外，可以以框图的形式示出装置，以便避免使本申请难以理解，并且这也考虑了以下事实，即关于这些框图装置的实施方式的细节是高度取决于将要实施本申请的平台的(即，这些细节应当完全处于本领域技术人员的理解范围内)。在阐述了具体细节(例如，电路)以描述本申请的示例性实施例的情况下，对本领域技术人员来说显而易见的是，可以在没有这些具体细节的情况下或者这些具体细节有变化的情况下实施本申请。因此，这些描述应被认为是说明性的而不是限制性的。
尽管已经结合了本申请的具体实施例对本申请进行了描述，但是根据前面的描述，这些实施例的很多替换、修改和变型对本领域普通技术人员来说将是显而易见的。例如，其它存储器架构(例如，动态RAM(DRAM))可以使用所讨论的实施例。
本申请的实施例旨在涵盖落入所附权利要求的宽泛范围之内的所有这样的替换、修改和变型。因此，凡在本申请的精神和原则之内，所做的任何省略、修改、等同替换、改进等，均应包含在本申请的保护范围之内。

Claims (14)

  1. 一种无人机路径优化方法，其特征在于，包括：
    根据获取的无线多播用户的数量、各个无线多播用户的地理位置以及需要传输的文件大小确定无线多播用户组的数量；
    根据所述无线多播用户组的数量将无线多播用户分成多个无线多播用户组；
    确定无人机对应每个无线多播用户组的悬停位置以及连接所述悬停位置的最短飞行路径；
    根据所述需要传输的文件的大小、所述悬停位置以及所述飞行路径确定无人机的第一总能耗；
    对所述各个悬停位置进行逐个优化，得到多个优化后的悬停位置以及连接所述多个优化后的悬停位置的最短的飞行路径；
    根据所述需要传输的文件的大小、所述优化后的悬停位置以及所述飞行路径确定无人机的第二总能耗；以及
    将所述第一总能耗与第二总能耗相比较，如果两者的差的绝对值大于预先设定的阈值，则在本轮优化的基础上返回至对所述各个悬停位置进行逐个优化的步骤进行下一轮优化；如果两者差的绝对值小于或等于预先设定的阈值，则将优化后的悬停位置及飞行路径确定为无人机的悬停位置和飞行路径。
  2. 根据权利要求1所述的方法，其特征在于，所述确定无线多播用户组的数量包括：利用训练好的支持向量机模型SVM，根据所述无线多播用户的数量、地理位置以及需要传输的文件大小确定无线多播用户组的数量。
  3. 根据权利要求2所述的方法，其特征在于，所述确定无线多播用户组的数量包括：
    根据用户数目N，确定无线多播用户组数量的可能值K*；
    根据每一个无线多播用户组数量的可能值K*，使用k-means算法将用户数目N分成K*个无线多播用户组，并计算出与K*对应的平均轮廓系数λ*；
    将用户数目N，需传输的文件大小S，无线多播用户组数量的可能值K*以及与K*对应的平均轮廓系数λ*输入训练好的SVM模型，得到无线多播用户组数量的可能值K*对应的预测总能耗E K*；以及
    将最小的预测总能耗E K*所对应的K*作为所述无线多播用户组的数量K。
  4. 根据权利要求2所述的方法，其特征在于，训练SVM模型包括：
    模拟生成X组训练数据，每组训练数据包括用户数目、用户地理位置、需要传输的文件大小和无线多播用户组数目K；
    由每组训练数据中的用户数目、用户地理位置以及无线多播用户组数目K计算得到每组数据的平均轮廓系数；
    将每组中的用户数目分裂成K个无线多播用户组；
    获得与K个无线多播用户组一一对应的K个悬停位置以及连接所述K个悬停位置的最短的飞行路径；
    根据所述需要传输的文件大小与悬停位置以及飞行路径确定无人机的总能耗；
    对所述悬停位置和飞行路径进行优化，从而得到最终悬停位置和最终飞行路径，以及与最终悬停位置和最终飞行路径所对应的最终总能耗；
    确定X组包括用户数目、无线多播用户组分裂数目K、需要传输的文件大小和平均轮廓系数的数据以及对应的X组只包含最终总能耗的数据；以及
    将包含用户数目、无线多播用户组数量K、需要传输的文件大小和平均轮廓系数的数据作为输入，将包含最终总能耗的数据作为输出，训练得到SVM模型。
  5. 根据权利要求1所述的方法，其特征在于，所述根据所述无线多播用户组的数量将无线多播用户分成多个无线多播用户组包括：利用k均值聚类算法k-means算法将无线多播用户分成多个无线多播用户组。
  6. 根据权利要求1所述的方法，其特征在于，所述确定无人机对应每个无线多播用户组的悬停位置包括：利用最小圆覆盖法分别确定无人机对应每个无线多播用户组的悬停位置。
  7. 根据权利要求6所述的方法，其特征在于，所述利用最小圆覆盖法分别确定无人机对应每个无线多播用户组的悬停位置包括：
    A、任取一个无线多播用户组，从所述无线多播用户组里包含的用户中随意取出三个用户N a、N b、N c；
    B、作一个包含N a、N b、N c三点的最小圆；
    C、在所述无线多播用户组其他用户中找出距离所述最小圆圆心最远的点N d；
    D、若N d已在圆内或圆周上，则所述最小圆为所求的圆，此圆的圆心为该无线多播用户组对应的悬停位置；否则，在N a、N b、N c、N d中选3个点，使由它们生成的一个包含这4个点的圆为最小，并令这3点成为新的N a、N b、N c，返回B。
  8. 根据权利要求1所述的方法，其特征在于，所述确定连接所述悬停位置的最短飞行路径包括：利用遗传算法、粒子群算法、蚁群算法、模拟退火算法及神经网络算法中的任意一种获得所述最短的飞行路径；
    所述确定连接所述优化后的悬停位置的最短飞行路径包括：利用遗传算法、粒子群算法、蚁群算法、模拟退火算法及神经网络算法中的任意一种获得所述最短的飞行路径。
  9. 根据权利要求8所述的方法，其特征在于，所述利用遗传算法获得所述最短的飞行路径包括：
    a、随机初始化L条飞行路径作为初代种群，计算所述L条路径的距离，并对L条路径距离进行排序；
    b、根据轮盘赌策略，根据L条路径中每条路径的距离，为每条路径分别设定一个选择概率，其中距离越小的路径设定的选择概率越大，并做L次选择，每次选择均 从L条路径中选择一条路径，被选中的路径再进行下一步操作；
    c、从所述L条路径中，依次任取两条以预设概率进行路径交叉并取代之前的路径，直至L条路径都被选择过，且不会重复选择；
    d、从交叉过程的L条路径中，依次对每一条路径以预设概率进行路径的反转，并取代之前的路径，直至L条路径都被选择过，且不会重复选择，得到新一代种群；
    e、将新一代种群重新带入a至d的过程并循环重复，设定循环重复的最大次数，直到获得的最短飞行路径不再变化或重复次数满足设定的次数。
  10. 根据权利要求1所述的方法，其特征在于，所述根据所述需要传输的文件的大小、所述悬停位置以及所述飞行路径确定无人机的第一总能耗包括：
    根据所述需要传输的文件大小与所述悬停位置确定传输文件所需的时间以及无人机的发射功率；
    根据所述传输文件所需的时间以及无人机的发射功率确定无人机的通信能耗和悬停能耗；
    根据所述飞行路径确定无人机的飞行能耗；以及
    由确定的通信能耗、悬停能耗以及飞行能耗确定无人机的总能耗。
  11. 根据权利要求1所述的方法，其特征在于，所述对所述各个悬停位置进行逐个优化包括：利用内点法对所述悬停位置进行逐个优化。
  12. 一种无人机路径优化设备，其特征在于，包括：
    用户组数量确定模块，用于根据无线多播用户的数量、各个无线多播用户的地理位置以及需要传输的文件大小确定无线多播用户组的数量；
    分组模块，用于根据确定的无线多播用户组的数量将无线多播用户分成多个无线多播用户组；
    悬停位置确定模块，用于确定无人机对应每个无线多播用户组的悬停位置；
    飞行路径确定模块，用于根据确定的无人机对应每个无线多播用户组的悬停位置获得连接所述悬停位置的最短飞行路径；
    第一能耗确定模块，用于根据所述需要传输的文件的大小、悬停位置以及飞行路径确定无人机的第一总能耗；
    优化模块，用于对所述各个悬停位置进行逐个优化，得到多个优化后的悬停位置，以及根据多个优化后的悬停位置确定连接多个优化后的悬停位置的最短的飞行路径；
    第二能耗确定模块，用于根据需要传输的文件的大小、优化后的悬停位置以及所述飞行路径确定无人机的第二总能耗；以及
    比较模块，用于将所述第一总能耗与第二总能耗相比较，如果两者的差的绝对值大于预先设定的阈值，则在本轮优化的基础上指示优化模块进行下一轮优化；如果两者差的绝对值小于或等于预先设定的阈值，则将优化后的悬停位置及飞行路径确定为无人机的悬停位置和飞行路径。
  13. 一种无人机路径优化设备，其特征在于，包括：至少一个处理器；以及与所述至少一个处理器通信连接的存储器；其中，所述存储器存储有可被所述至少一个处理器执行的指令，所述指令被所述至少一个处理器执行，以使所述至少一个处理器能够执行如权利要求1至11任意一项所述的无人机路径优化方法。
  14. 一种计算机可读存储介质，其特征在于，其上存储有计算机指令，在处理器执行上述计算机指令时实现如权利要求1至11任意一项所述的无人机路径优化方法。
PCT/CN2019/124036 2019-01-16 2019-12-09 无人机路径优化方法、设备及存储介质 WO2020147456A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US17/023,813 US11663920B2 (en) 2019-01-16 2020-09-17 Method and device of path optimization for UAV, and storage medium thereof

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910041506.XA CN109682380B (zh) 2019-01-16 2019-01-16 一种通信无人机路径优化方法及设备
CN201910041506.X 2019-01-16

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US17/023,813 Continuation US11663920B2 (en) 2019-01-16 2020-09-17 Method and device of path optimization for UAV, and storage medium thereof

Publications (1)

Publication Number Publication Date
WO2020147456A1 true WO2020147456A1 (zh) 2020-07-23

Family

ID=66193304

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/124036 WO2020147456A1 (zh) 2019-01-16 2019-12-09 无人机路径优化方法、设备及存储介质

Country Status (3)

Country Link
US (1) US11663920B2 (zh)
CN (1) CN109682380B (zh)
WO (1) WO2020147456A1 (zh)

Families Citing this family (30)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109682380B (zh) * 2019-01-16 2020-01-10 北京邮电大学 一种通信无人机路径优化方法及设备
CN110190888B (zh) * 2019-05-06 2020-04-17 北京邮电大学 一种基于信息时效性的无人机信息收集方法及装置
CN110290542B (zh) * 2019-06-28 2020-09-25 清华大学 海上无人机通信覆盖优化方法及系统
CN110426039B (zh) * 2019-07-04 2023-06-09 中国人民解放军陆军工程大学 基于任务完成时间最小化的多无人机路径规划方法
CN110634331A (zh) * 2019-09-12 2019-12-31 广州大学 基于负载均衡的多uav路径设计方法
CN110809244B (zh) * 2019-11-07 2023-04-18 腾讯科技（深圳）有限公司 一种数据传输方法及相关设备
CN111193537B (zh) * 2019-12-20 2022-03-04 中移（杭州）信息技术有限公司 优化飞行设备传输数据时的能量消耗方法、系统及装置
CN111711974B (zh) * 2020-07-10 2023-03-17 河南科技大学 基于轨迹预测的飞行自组织网络能量有效型机会路由方法
CN112256054B (zh) * 2020-10-09 2022-03-29 北京邮电大学 一种无人机轨迹规划方法及装置
CN112564767B (zh) * 2020-11-30 2021-09-21 北京邮电大学 无人机网络中基于自组织优化协作的连续覆盖方法
CN112511197B (zh) * 2020-12-01 2022-12-23 南京工业大学 基于深度强化学习的无人机辅助弹性视频多播方法
CN112731967B (zh) * 2020-12-24 2021-11-19 中科院计算技术研究所大数据研究院 一种基于聚类和遗传算法的多无人机协同任务规划方法
CN112784963B (zh) * 2021-01-22 2022-07-01 重庆邮电大学 基于模拟退火优化bp神经网络的室内外无缝定位方法
CN112911534B (zh) * 2021-01-29 2022-03-29 大连理工大学 高能效无人机绿色数据采集系统设计方法
CN113050453B (zh) * 2021-03-25 2022-02-01 四川大学 一种固定翼无人机通信过程能量优化方法
CN113203417B (zh) * 2021-04-06 2022-08-19 广东安恒电力科技有限公司 基于头脑风暴优化算法的无人机巡检路径规划方法
CN113034981B (zh) * 2021-04-14 2022-07-08 北京航空航天大学 一种不确定信道环境中多中继无人机航迹规划方法、系统及存储介质
US20220380042A1 (en) * 2021-05-06 2022-12-01 Board Of Regents, The University Of Texas System Intercepting or surrounding a group of hostile uavs with a net
CN113364495B (zh) * 2021-05-25 2022-08-05 西安交通大学 一种多无人机轨迹和智能反射面相移联合优化方法及系统
CN113395706B (zh) * 2021-06-17 2022-07-22 郑州大学 基于粒子群算法的异构无人机三维空间部署方法及设备
CN113518361B (zh) * 2021-06-18 2022-06-17 中南林业科技大学 一种无人机辅助无线通信的轨迹优化方法、设备及介质
CN113325875B (zh) * 2021-06-21 2022-06-07 西安电子科技大学 一种实现无人机数目最小化的无人机路径规划方法
CN113592148B (zh) * 2021-07-01 2024-03-15 合肥工业大学 改进车辆与无人机配送路线的优化方法及系统
CN113825145B (zh) * 2021-09-15 2022-08-12 云南大学 一种面向用户体验的无人机系统服务方法及系统
CN114265363B (zh) * 2021-12-17 2024-02-20 中科航迈数控软件(深圳)有限公司 一种数控机床加工路径智能优化方法及其系统
CN114337785A (zh) * 2021-12-30 2022-04-12 陕西锐远信息科技有限公司 太阳能无人机通信能量管理策略、系统、终端及存储介质
CN114944999B (zh) * 2022-04-08 2023-10-10 北京遥测技术研究所 一种基于gpu和粒子群优化算法的高速路由通信系统及方法
CN115118724A (zh) * 2022-06-23 2022-09-27 福州大学 基于蚁群算法的多无人机辅助边缘计算系统部署优化方法
CN115358497B (zh) * 2022-10-24 2023-03-10 湖南长理尚洋科技有限公司 基于gis技术的智慧全景巡河方法及系统
CN117371638B (zh) * 2023-12-06 2024-02-20 合肥工业大学 基于浣熊优化算法的多鱼雷罐多轨道行走路径优化方法

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170126309A1 (en) * 2015-10-30 2017-05-04 The Florida International University Board Of Trustees Cooperative clustering for enhancing mu-massive-miso-based uav communication
CN107017940A (zh) * 2017-04-25 2017-08-04 中国民航大学 无人机中继广播通信系统航迹优化方法
CN107450587A (zh) * 2017-09-19 2017-12-08 广东电网有限责任公司佛山供电局 一种无人机精细化巡检的智能飞行控制方法及系统
CN108834049A (zh) * 2018-06-15 2018-11-16 广东工业大学 无线供能通信网络及确定其工作状态的方法、装置
CN109213712A (zh) * 2018-09-06 2019-01-15 北京邮电大学 用于机器类通信系统的服务提供方法、装置及电子设备
CN109682380A (zh) * 2019-01-16 2019-04-26 北京邮电大学 一种通信无人机路径优化方法及设备

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3183707A4 (en) * 2014-08-21 2018-02-28 Uber Technologies Inc. Arranging a transport service for a user based on the estimated time of arrival of the user
US9489846B2 (en) * 2014-12-31 2016-11-08 Venuenext, Inc. Modifying directions to a parking lot associated with a venue based on traffic conditions proximate to the parking lot
US9685088B2 (en) * 2015-04-10 2017-06-20 Architecture Technology, Inc. Communication travel plan generation system
CN108521667B (zh) * 2018-03-07 2020-12-29 杭州电子科技大学 一种低传输能耗的无人机数据传输方法
CN108616303B (zh) * 2018-05-03 2020-10-23 广东工业大学 无人机基站通信系统的控制方法、系统、装置及存储介质
CN108924791B (zh) * 2018-07-13 2021-09-03 广东工业大学 一种无线通信方法、装置、设备及可读存储介质
CN108832998B (zh) * 2018-08-15 2021-04-13 中国人民解放军陆军工程大学 一种空地融合通信网络中的协同数据分发方法
CN108848465B (zh) * 2018-08-15 2020-10-30 中国人民解放军陆军工程大学 一种面向数据分发的无人机飞行轨迹与资源调度联合优化方法

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170126309A1 (en) * 2015-10-30 2017-05-04 The Florida International University Board Of Trustees Cooperative clustering for enhancing mu-massive-miso-based uav communication
CN107017940A (zh) * 2017-04-25 2017-08-04 中国民航大学 无人机中继广播通信系统航迹优化方法
CN107450587A (zh) * 2017-09-19 2017-12-08 广东电网有限责任公司佛山供电局 一种无人机精细化巡检的智能飞行控制方法及系统
CN108834049A (zh) * 2018-06-15 2018-11-16 广东工业大学 无线供能通信网络及确定其工作状态的方法、装置
CN109213712A (zh) * 2018-09-06 2019-01-15 北京邮电大学 用于机器类通信系统的服务提供方法、装置及电子设备
CN109682380A (zh) * 2019-01-16 2019-04-26 北京邮电大学 一种通信无人机路径优化方法及设备

Also Published As

Publication number Publication date
US20210003412A1 (en) 2021-01-07
CN109682380B (zh) 2020-01-10
CN109682380A (zh) 2019-04-26
US11663920B2 (en) 2023-05-30

Similar Documents

Publication Publication Date Title
WO2020147456A1 (zh) 无人机路径优化方法、设备及存储介质
CN112488322B (zh) 一种基于数据特征感知聚合的联邦学习模型训练方法
WO2023168824A1 (zh) 一种基于联邦学习的移动边缘缓存优化方法
CN113268341B (zh) 电网边缘计算任务的分配方法、装置、设备和存储介质
CN112512013B (zh) 基于学习剪枝的车联网移动边缘计算任务卸载方法及系统
CN113395654A (zh) 一种边缘计算系统的多无人机任务卸载和资源分配的方法
CN110428115A (zh) 基于深度强化学习的动态环境下的最大化系统效益方法
CN112118287A (zh) 基于交替方向乘子算法与移动边缘计算的网络资源优化调度决策方法
CN111953547B (zh) 一种基于业务的异构基站重叠分组及资源配置方法及装置
WO2021238305A1 (zh) 一种基于强化学习的通用分布式图处理方法及系统
WO2023020502A1 (zh) 数据处理方法及装置
CN113987692A (zh) 用于无人机和边缘计算服务器的深度神经网络分区方法
CN113163377B (zh) 一种无人机网络部署和资源分配方法及其装置
CN113419561A (zh) 一种分布式空中基站高动态部署方法
CN115802370A (zh) 一种通信方法及装置
CN115987375A (zh) 融合网络的电力设备关联与资源优化方法、系统及介质
CN115967430A (zh) 一种基于深度强化学习的成本最优空地网络任务卸载方法
CN112261628B (zh) 一种应用于d2d设备的内容边缘缓存架构方法
Pang et al. A computing resource scheduling strategy of massive IoT devices in the mobile edge computing environment
KR102416342B1 (ko) 3차원 셀룰러 네트워크에서 간헐적 인식 연합 학습 방법
CN111935824B (zh) 无线资源分配策略更新方法、装置、设备及存储介质
Cui et al. UAV-enabled MEC system: Locations and coverage optimization
EP3995998A1 (en) Method for selecting task network, system and method for determining actions based on sensing data
CN115065683B (zh) 基于车辆聚类的车辆边缘网络任务分配卸载方法
CN113326112B (zh) 一种基于块坐标下降法的多无人机任务卸载和迁移方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19909891

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19909891

Country of ref document: EP

Kind code of ref document: A1