UAV trajectory optimization and cache scheduling method and system based on image recognition

Abstract

The present invention relates to the field of image recognition technology, and specifically to a method and system for optimizing and caching unmanned aerial vehicle (UAV) trajectory based on image recognition. The method comprises: gridding an area to be served, and processing to obtain user density of each service grid; setting an initial height of the UAV; obtaining quality evaluation indexes of each communication link through processing, and setting an initial horizontal position of the UAV; constructing a wireless communication system model and a joint optimization algorithm, and iteratively optimizing content placement and UAV height through the joint optimization algorithm; processing to obtain user satisfaction evaluation values ​​of each dense area, and providing feedback on the optimization of UAV trajectory according to the user satisfaction evaluation values ​​of each dense area. The present invention provides a method and system for optimizing and caching unmanned aerial vehicle (UAV) trajectory based on image recognition, and effectively improves user satisfaction with content downloading by jointly optimizing content placement and UAV height.

Description

Technical Field

[0001] The present invention relates to the technical field of image recognition, and in particular to a method and system for optimizing the trajectory of an unmanned aerial vehicle and caching it based on image recognition. Background Art

[0002] With the development of communication technology, the role of drones in emergency rescue communications in remote areas has become increasingly prominent. Drones with caching capabilities and the ability to flexibly adjust their trajectories can penetrate deep into disaster areas, collect information and transmit it in a timely manner, providing a basis for rescue decisions.

[0003] For example, the invention patent with publication number: CN119380226A is a method and system for identifying defects in power transmission line fittings based on the collaborative work of drones and robots, which belongs to the field of image recognition technology, including: the robot receives first image data and the flight trajectory data of the drone, and synchronizes the timestamp; analyzes the flight trajectory data and outputs a first search signal; according to the first search signal, obtains the image data at the corresponding timestamp, and optimizes the obtained image data; combines with deep learning algorithms to predict and warn of trajectory deviations.

[0004] For example, the invention patent with publication number: CN116109950A is a low-altitude anti-UAV visual detection, identification and tracking method, which belongs to the field of UAV countermeasures. It includes: first collecting UAV samples, constructing target detection training sample sets, target fine-grained recognition training sample sets and target tracking data sets, then optimizing the framework of the target detection model to obtain an improved target detection model, detecting and locating the UAV in the actual visual image, and obtaining the global image of the UAV. Then, the target fine-grained recognition model is improved, and a scale-adaptive attention mechanism and joint probability prediction are introduced to extract and identify UAV components in the global image of the UAV. Finally, the UAV is tracked, and the trajectory and tracking video image of each model of UAV are recorded.

[0005] However, in the process of implementing the technical solution of the invention in the embodiments of the present application, the present application discovered that the above-mentioned technology has at least the following technical problems: the current research on drone networks has long focused on the service quality of the network, such as improving network throughput, reducing user energy consumption, etc., and often ignores the multidimensionality of user experience effects and the personalized needs of different users, resulting in poor actual user experience. Summary of the invention

[0006] In view of the deficiencies in the prior art, the present invention provides a method and system for optimizing the trajectory of a UAV and caching it based on image recognition, which can effectively solve the problems involved in the above-mentioned background technology.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: The first aspect of the present invention provides a drone trajectory optimization and cache scheduling method based on image recognition, including: S1, obtaining a high-definition image of the area to be served, gridding the area to be served to obtain each service grid, collecting the number of users in each service grid in real time, and processing to obtain the user density of each service grid.

[0008] S2. Jointly mark the adjacent grids with user density in each user-dense interval value as user-dense areas, obtain the geographic environment data of each user-dense area, and set the initial height of the drone based on the horizontal density of each user-dense area.

[0009] S3. Collect the communication link quality data between the UAV and the users in the user-dense area, obtain the quality evaluation index of each communication link after processing, and set the initial horizontal position of the UAV according to the quality evaluation index of each communication link combined with the signal strength between the ground base station and the user-dense area.

[0010] S4. Build a wireless communication system model and joint optimization algorithm including MBS, drones and users, and iteratively optimize content placement and drone height through the joint optimization algorithm.

[0011] S5. Collect user satisfaction data of each densely populated area, process it to obtain a user satisfaction evaluation value of each densely populated area, and provide feedback for drone trajectory optimization based on the user satisfaction evaluation value of each densely populated area.

[0012] As a further method, the processing obtains the user density of each service grid. The specific analysis process is: extracting the area of ​​the area to be served from the high-definition image of the area to be served, and dividing the area to be served into grids according to the preset service demand level of the area to be served, and marking them as each service grid.

[0013] According to the preset time collection points, the number of users in each service grid is collected at a fixed point to obtain the number of users in each service grid at each time collection point, and the user density of each service grid is obtained based on the service grid area.

[0014] As a further method, the initial height of the drone is set in combination with the horizontal density of each user-dense area. The specific analysis process is: extracting the user density standard value from the regional information database, and jointly marking adjacent grids with user density higher than or equal to the user density standard value as user-dense areas.

[0015] The geographical environment data of each user-dense area is extracted from the high-definition image of the area to be served, including the building height, altitude, ground slope and building density of each collection point. The building height, altitude and ground slope of each collection point are deviated from the average building height, average altitude and average ground slope of the user-dense area respectively. Then, each deviation value is compared with the average building height, average altitude and average ground slope of the user-dense area respectively. The ratio between the building density of the user-dense area and the critical value of the building density of the user-dense area is combined, and the influencing weight factor is introduced to obtain the geographical environment flatness evaluation index of each user-dense area.

[0016] The user density in each service grid in each user-dense area is subjected to a standard deviation operation to obtain an evaluation value of the uniformity of user distribution in each user-dense area.

[0017] The initial height of the drone is set according to the geographical environment flatness evaluation index and user distribution uniformity evaluation value of each user-dense area.

[0018] As a further method, the initial height of the drone is set according to the geographical environment flatness evaluation index and the user distribution uniformity evaluation value of each user-dense area. The specific process is: extracting from the regional information database a mapping set between the user distribution uniformity evaluation value of the user-dense area and the initial height of the drone and a mapping set between the geographical environment flatness evaluation index and the initial height of the drone, the mapping relationship is one-to-one correspondence, and obtaining two initial height setting values ​​of the drone according to the mapping set, taking the average value, and setting the average value as the initial height of the drone.

[0019] As a further method, the initial horizontal position of the UAV is set according to each communication link quality evaluation index combined with the signal strength between the ground base station and the user-dense area. The specific process is: deploy a number of monitoring position points between the user-dense area and the ground base station, and the monitoring position points are all located on the initial height line of the UAV. The signal strength between the UAV at each time monitoring point of each monitoring position point and the ground base station is collected, and the signal strength between the UAV at each time monitoring point of each monitoring position point and the ground base station is summed and averaged to obtain the average signal strength between the UAV and the ground base station.

[0020] The communication link quality data between users in user-dense areas and drones at each monitoring location are collected at preset time intervals, and the quality evaluation index of each communication link is obtained by combining the average signal strength between the drone at each monitoring location and the ground base station. The initial horizontal position of the drone is set according to the quality evaluation index of each communication link.

[0021] As a further method, the initial horizontal position of the UAV is set according to each communication link quality evaluation index. The specific process is: each communication link quality evaluation index is used to quantitatively evaluate the communication link quality between the UAV and the ground base station and between the UAV and the user equipment in the user-dense area when the UAV is at each monitoring position point. The communication link quality evaluation index is sorted in order from large to small, and the monitoring position point where the UAV is located when the communication link quality evaluation index is the largest is set as the initial horizontal position of the UAV.

[0022] As a further method, a wireless communication system model and a joint optimization algorithm including MBS, UAVs and users are constructed, and the specific analysis process is as follows: the wireless communication system model includes a network model and a channel model.

[0023] Among them, the network model includes MBS, drones and users. Drones and MBS share the spectrum, MBS users are primary users, and drone users are secondary users.

[0024] The channel model uses a free space signal fading model to reflect the channel gain between the drone and the user. The drone optimizes the user's download rate by adjusting its altitude.

[0025] The joint optimization algorithm includes an optimization target and a Stackelberg game model.

[0026] As a further method, the input related initial values, and then execute the optimized trajectory algorithm to optimize the content placement to obtain parameter one, which is substituted into the height adjustment formula to calculate parameter two, and then the parameter one and parameter two are combined and substituted into the pricing formula to obtain the price, and then the price is substituted into the height adjustment formula to determine the height of the drone, and this cycle is iterated until the changes in content placement and drone height meet the accuracy requirements.

[0027] As a further method, the feedback on the optimization of the drone trajectory is specifically analyzed as follows: user satisfaction data of each densely populated area is collected, and the comprehensive satisfaction evaluation index of each user in each densely populated area is obtained by processing.

[0028] The comprehensive satisfaction evaluation indicators of each user in each densely populated area are sorted in ascending order to obtain a user comprehensive satisfaction evaluation indicator sequence, the number of users participating in the feedback is counted, the user comprehensive satisfaction evaluation indicator sequence is processed to obtain a secondary sequence of user comprehensive satisfaction evaluation indicators, and the user comprehensive satisfaction evaluation indicators of the secondary sequence of user comprehensive satisfaction evaluation indicators are averaged to obtain the user comprehensive satisfaction evaluation indicator parameters.

[0029] The comprehensive satisfaction evaluation index of each user in each densely populated area is compared with the user comprehensive satisfaction evaluation index parameter, and the number of users in each densely populated area whose comprehensive satisfaction evaluation index is higher than or equal to the user comprehensive satisfaction evaluation index parameter is recorded and marked as the number of satisfied users. If the ratio of the number of satisfied users to the number of users is greater than or equal to the pass rate, the drone trajectory optimization is marked as qualified. Otherwise, the drone trajectory needs to be optimized again.

[0030] The second aspect of the present invention provides a UAV trajectory optimization and cache scheduling system based on image recognition, including: a UAV initial height setting module, a UAV initial horizontal position setting module, a UAV trajectory pre-adjustment module, a UAV trajectory optimization module and a comprehensive feedback module.

[0031] The drone initial height setting module is used to obtain high-definition images of the area to be served, grid the area to be served, obtain each service grid, collect the number of users in each service grid in real time, process to obtain the user density of each service grid, jointly mark adjacent grids with user density in each user-dense interval value as user-dense areas, obtain the geographical environment data of each user-dense area, and set the drone initial height based on the horizontal density of each user-dense area.

[0032] The module for setting the initial horizontal position of the UAV is used to collect the communication link quality data between each UAV and the users in each user-dense area, obtain the quality evaluation index of each communication link after processing, and set the initial horizontal position of each UAV according to the quality evaluation index of each communication link combined with the signal strength between the ground base station and each user-dense area.

[0033] The drone trajectory optimization module is used to build a wireless communication system model and joint optimization algorithm including MBS, drones and users, and iteratively optimize content placement and drone height through the joint optimization algorithm.

[0034] The comprehensive feedback module is used to collect user satisfaction data of each dense area, process it to obtain the user satisfaction evaluation value of each dense area, and provide feedback on the drone trajectory optimization according to the user satisfaction evaluation value of each dense area.

[0035] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects:

[0036] (1) The present invention provides a method and system for optimizing drone trajectory and caching based on image recognition, which fully considers the timeliness, affinity, regionality and other spatiotemporal distribution characteristics and rate requirements of the content required by users. By jointly optimizing content placement and drone altitude, the present invention effectively improves user satisfaction with content downloading, and achieves the goal of adjusting the drone altitude based on the Stackelberg game while ensuring that MBS users are not disturbed, thereby improving the service rate of drone users.

[0037] (2) The present invention extracts the area of ​​the service area and divides it into service grids, and performs differentiated processing based on the service demand level, so as to reasonably allocate resources according to the actual conditions of different grids. This avoids resource waste and improves resource utilization efficiency. Dynamically allocating drone resources according to user density is conducive to giving priority to serving areas with dense user populations, reducing user waiting time, and improving service response speed.

[0038] (3) The present invention comprehensively analyzes the received signal strength, bit error rate, signal-to-noise ratio, delay and packet loss rate. These parameters influence each other and comprehensively reflect the quality and performance of the communication link in the UAV-assisted ground communication system from different angles. It can accurately evaluate the status of the communication link, discover problems in time and take effective optimization measures to meet the requirements of communication quality in different application scenarios.

[0039] (4) This paper comprehensively analyzes user satisfaction under different research needs and application scenarios. This refined evaluation can provide more targeted suggestions and guidance for drone trajectory optimization and communication quality improvement, which helps to improve the overall performance and user experience of the drone system. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The present invention is further described using the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative work.

[0041] Figure 1 It is a schematic diagram of the method flow of the present invention.

[0042] Figure 2 It is a schematic diagram of system module connection of the present invention.

[0043] Figure 3 A schematic diagram of a wireless communication system model involved in an embodiment of the present invention. DETAILED DESCRIPTION

[0044] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0045] Reference Figure 1 As shown, the first aspect of the present invention provides a UAV trajectory optimization and cache scheduling method based on image recognition, including: S1, obtaining a high-definition image of the area to be served, gridding the area to be served to obtain each service grid, collecting the number of users in each service grid in real time, and processing to obtain the user density of each service grid.

[0046] Specifically, the user density of each service grid is obtained by processing, and the specific analysis process is: the area of ​​the area to be served is extracted from the high-definition image of the area to be served, and according to the preset service demand level of the area to be served, the area to be served is divided into grids and marked as each service grid.

[0047] According to the preset time collection points, the number of users in each service grid is collected at a fixed point to obtain the number of users in each service grid at each time collection point, and the user density of each service grid is obtained based on the service grid area.

[0048] In a specific embodiment, the numerical expression of the user density of each service grid is:

[0049] ; (1) Among them, Indicates The density of service grid users, Indicates The first time point of the collection The number of users in the service grid, Indicates the number of each time collection point, s, s represents the total number of time acquisition points, Indicates the number of each service grid, h, h represents the total number of service meshes, Represents the service grid area.

[0050] It should be explained that in this embodiment, by extracting the area of ​​the area to be served and dividing the service grid, and combining the service demand level for differentiated processing, resources can be reasonably allocated according to the actual situation of different grids. For example, increase resource investment in grids with high demand levels to avoid resource waste and improve resource utilization efficiency. Dynamically allocate drone resources according to user density, increase the frequency of drone services or cache more popular content in grids with high user density to improve service quality. By analyzing these data, we can understand the temporal and spatial changes in user needs, predict user demand trends, and provide a strong basis for optimizing service strategies and improving service quality.

[0051] S2. Jointly mark the adjacent grids with user density in each user-dense interval value as user-dense areas, obtain the geographic environment data of each user-dense area, and set the initial height of the drone based on the horizontal density of each user-dense area.

[0052] Specifically, the initial height of the drone is set based on the horizontal density of each user-dense area. The specific analysis process is as follows: the user density standard value is extracted from the regional information database, and adjacent grids with user density higher than or equal to the user density standard value are jointly marked as user-dense areas.

[0053] The geographical environment data of each user-dense area is extracted from the high-definition image of the area to be served, including the altitude, ground slope and building density of each collection point. The building height, altitude and ground slope of each collection point are taken as deviation values ​​from the average building height, average altitude and average ground slope of the user-dense area respectively. Then, each deviation value is compared with the average building height, average altitude and average ground slope of the user-dense area respectively. The ratio between the building density in the user-dense area and the critical value of the building density in the user-dense area is combined, and an influencing weight factor is introduced to obtain the geographical environment flatness evaluation index of each user-dense area.

[0054] The user density in each service grid in each user-dense area is subjected to a standard deviation operation to obtain an evaluation value of the uniformity of user distribution in each user-dense area.

[0055] The initial height of the drone is set according to the geographical environment flatness evaluation index and user distribution uniformity evaluation value of each user-dense area.

[0056] In a specific embodiment, the numerical expression of the user distribution uniformity evaluation value of each user-dense area is:

[0057] ; (2)

[0058] in, Indicates The evaluation value of user distribution uniformity in user-dense areas, Indicates The user-dense area The user density of r service grids, Indicates the number of each user-dense area, , m represents the total number of user-dense areas.

[0059] In a specific embodiment, the numerical expression of the geographical environment flatness evaluation index of each user-dense area is:

[0060] ; (3)

[0061] in, Indicates The geographical environment flatness evaluation index of the user-dense area, e represents a natural constant, Indicates User-dense areas The altitude of the collection points, Indicates the number of each collection point, , g represents the total number of collection points, Indicates User-dense areas The ground slope of each collection point, Indicates The building density in the user-dense area, represents the critical building density, Indicates the weight factor of the geographical environment flatness corresponding to the preset altitude. Indicates the weight factor of the geographical environment flatness corresponding to the preset terrain slope, Indicates the weight factor affecting the flatness of the geographical environment corresponding to the preset building density.

[0062] It should be explained that, in this embodiment, the critical building density is extracted from the regional information database.

[0063] It needs to be explained that when the deviation between the altitude of each collection point and the average value is smaller, the deviation between the terrain slope of each collection point and the average value is smaller, and the regional building density is lower, the corresponding regional geographical environment flatness assessment index is larger, indicating that the regional geographical environment is flatter.

[0064] It should be explained that in this embodiment Indicates the weight factor of the geographical environment flatness corresponding to the preset altitude. Indicates the weight factor of the geographical environment flatness corresponding to the preset terrain slope, The weight factors of geographical environment flatness impact corresponding to the preset building density respectively represent the numerical values ​​of the impact degree of the unit values ​​of altitude, ground slope and building density on geographical environment flatness. When used, they can be directly obtained from the regional information library, and their corresponding relationship can be a pre-set mapping relationship. For example, the altitude, ground slope and building density respectively form a mapping set with the geographical environment flatness impact weight factors corresponding to the preset altitude, ground slope and building density in the regional information library, and the real-time altitude, ground slope and building density are input into the mapping set to obtain the corresponding geographical environment flatness impact weight factors, wherein the mapping relationship can be one-to-one or many-to-one. The value range of the above-mentioned impact weight factors is between 0 and 1.

[0065] Specifically, a mapping set between the user distribution uniformity evaluation value in the user-dense area and the initial height of the UAV and a mapping set between the geographical environment flatness evaluation index and the initial height of the UAV are extracted from the regional information database. The mapping relationship is one-to-one. Two UAV initial height setting values ​​are obtained according to the mapping set, and the average value is taken, and the average value is set as the initial height of the UAV.

[0066] In a specific embodiment, the regional information database is used to store basic data of the area to be served, including user density standard values ​​and drone-assisted ground communication link quality assessment thresholds and data extracted from the regional information database in the above embodiments. GIS software can be used to analyze densely populated areas, and user density interval values ​​can be obtained in combination with census data. Field tests can be carried out using communication equipment carried by drones to obtain data such as signal strength and communication delay. Various sensors, such as cameras and traffic monitors, can also be deployed to collect data in real time.

[0067] S3. Collect the communication link quality data between the UAV and the users in the user-dense area, obtain the quality evaluation index of each communication link after processing, and set the initial horizontal position of the UAV according to the quality evaluation index of each communication link combined with the signal strength between the ground base station and the user-dense area.

[0068] Specifically, the initial horizontal position of the UAV is set according to the quality evaluation index of each communication link combined with the signal strength between the ground base station and the user-dense area. The specific process is: deploy a number of monitoring position points between the user-dense area and the ground base station, and the monitoring position points are all located on the initial height line of the UAV. The signal strength between the UAV and the ground base station at each monitoring point at each time is collected, and the signal strength between the UAV and the ground base station at each monitoring point at each monitoring point is summed and averaged to obtain the average signal strength between the UAV and the ground base station at each monitoring point at each monitoring point.

[0069] The communication link quality data between users in user-dense areas and drones at each monitoring location are collected at preset time intervals, and the quality evaluation index of each communication link is obtained by combining the average signal strength between the drone at each monitoring location and the ground base station. The initial horizontal position of the drone is set according to the quality evaluation index of each communication link.

[0070] Furthermore, the communication link quality data between the user and the drone includes: received signal strength, bit error rate, signal-to-noise ratio, delay and packet loss rate at each moment, and the critical received signal strength, critical bit error rate, critical signal-to-noise ratio, critical delay, critical packet loss rate and critical signal strength are extracted from the regional information database.

[0071] In a specific embodiment, the numerical expression of each communication link quality evaluation index is:

[0072] ; (4)

[0073] in, Indicates A communication link quality assessment index, Indicates the number of each communication link, , g represents the total number of communication links, Indicates The communication link is The received signal strength at a given moment, Indicates the number of each moment, , s represents the total number of moments, Indicates The communication link is The bit error rate at a moment, Indicates The communication link is The signal-to-noise ratio at a given moment, Indicates The communication link is A time delay of one moment, Indicates The communication link is The packet loss rate at a moment, Indicates The average signal strength of the communication link, Indicates the critical received signal strength, represents the critical bit error rate, represents the critical signal-to-noise ratio, represents the critical delay, represents the critical packet loss rate, represents the critical signal strength, Indicates the communication link quality impact characteristic factor corresponding to the preset received signal strength, Indicates the communication link quality impact characteristic factor corresponding to the preset bit error rate, Indicates the communication link quality impact characteristic factor corresponding to the preset signal-to-noise ratio, Indicates the communication link quality impact characteristic factor corresponding to the preset delay, Indicates the communication link quality impact characteristic factor corresponding to the preset packet loss rate, Indicates the communication link quality impact characteristic factor corresponding to the preset signal strength.

[0074] It needs to be explained that when the received signal strength is greater, the bit error rate is smaller, the signal-to-noise ratio is greater, the delay is lower, and the packet loss rate is smaller, the corresponding communication link quality evaluation index is larger, indicating that the communication link quality is better.

[0075] It should be explained that in this embodiment Indicates the communication link quality impact characteristic factor corresponding to the preset received signal strength, Indicates the communication link quality impact characteristic factor corresponding to the preset bit error rate, Indicates the communication link quality impact characteristic factor corresponding to the preset signal-to-noise ratio, Indicates the communication link quality impact characteristic factor corresponding to the preset delay, Indicates the communication link quality impact characteristic factor corresponding to the preset packet loss rate, The communication link quality influencing characteristic factors corresponding to the preset signal strength respectively represent the numerical values ​​of the influence degree of the received signal strength, bit error rate, signal-to-noise ratio, delay, packet loss rate and signal strength unit values ​​on the communication link quality. When used, they can be directly obtained from the regional information library, and their corresponding relationship can be a pre-set mapping relationship. For example, the received signal strength, bit error rate, signal-to-noise ratio, delay, packet loss rate and signal strength respectively form a mapping set with the communication link quality influencing characteristic factors corresponding to the received signal strength, bit error rate, signal-to-noise ratio, delay, packet loss rate and signal strength preset in the regional information library, and the real-time received signal strength, bit error rate, signal-to-noise ratio, delay, packet loss rate and signal strength are input into the mapping set to obtain the corresponding communication link quality influencing characteristic factors, wherein the mapping relationship can be one-to-one or many-to-one. The value range of the above-mentioned influencing characteristic factors is between 0 and 1.

[0076] It should be explained that in this embodiment, the received signal strength refers to the average strength of the signal received by all user devices in the dense area from the drone. The signal-to-noise ratio refers to the ratio of signal power to noise power. The bit error rate refers to the ratio of the number of error symbols in the transmission process to the total number of transmitted symbols. The delay refers to the average time it takes for the signal to be sent from the drone to all user devices in the dense area. The packet loss rate refers to the ratio of the number of data packets lost during the transmission process to the total number of data packets sent within a certain period of time.

[0077] It needs to be explained that in the UAV-assisted ground communication system, the received signal strength, bit error rate, signal-to-noise ratio, delay and packet loss rate are closely related. The received signal strength is positively correlated with the bit error rate and signal-to-noise ratio. A strong signal reduces the bit error rate and improves the signal-to-noise ratio. The signal-to-noise ratio is negatively correlated with the bit error rate. A high signal-to-noise ratio is conducive to reducing bit errors. The reduction of the received signal strength and signal-to-noise ratio will increase the delay due to retransmission and complex processing, and a high bit error rate will also increase the delay due to retransmission. A bit error rate that is too high and exceeds the error correction capability will lead to an increase in the packet loss rate, and when the delay is large, the data packet may be discarded due to timeout, increasing the packet loss rate. These parameters affect each other and comprehensively reflect the quality and performance of the communication link in the UAV-assisted ground communication system from different angles. Comprehensive analysis of these parameters can accurately evaluate the status of the communication link, promptly identify problems and take effective optimization measures to meet the requirements of communication quality in different application scenarios.

[0078] Furthermore, the initial horizontal position of the UAV is set according to each communication link quality evaluation index. The specific process is as follows: each communication link quality evaluation index is used to quantitatively evaluate the communication link quality between the UAV and the ground base station and between the UAV and the user equipment in the user-dense area when the UAV is at each monitoring position point. The communication link quality evaluation index is sorted in descending order, and the monitoring position point where the UAV is located when the communication link quality evaluation index is the largest is set as the initial horizontal position of the UAV.

[0079] S4. Build a wireless communication system model and joint optimization algorithm including MBS, drones and users, and iteratively optimize content placement and drone height through the joint optimization algorithm.

[0080] Specifically, a wireless communication system model including MBS, UAVs and users and a joint optimization algorithm are constructed, and the specific analysis process is as follows: the wireless communication system model includes a network model and a channel model.

[0081] Among them, the network model includes MBS, drones and users. Drones and MBS share the spectrum, MBS users are primary users, and drone users are secondary users.

[0082] The channel model uses a free space signal fading model to reflect the channel gain between the drone and the user. The drone optimizes the user's download rate by adjusting its altitude.

[0083] The joint optimization algorithm includes an optimization target and a Stackelberg game model.

[0084] It should be explained that the optimization objectives in this embodiment include optimizing the channel gain between the drone and the user by adjusting the altitude of the drone, and optimizing the content cache allocation on the drone according to the user's needs and the popularity of the content.

[0085] The Stackelberg game model includes leaders and followers. MBS users, as leaders of the game, should obtain resources first to ensure that their communication rate is not disturbed. UAV users, as followers, maximize their service rate by adjusting the altitude and content cache allocation of the UAV.

[0086] Furthermore, the content placement and the height of the drone are iteratively optimized through a joint optimization algorithm. The specific analysis process is as follows: input relevant initial values, execute the optimization trajectory algorithm to optimize the content placement to obtain parameter one, substitute it into the height adjustment formula to calculate parameter two, then combine parameter one and parameter two and substitute them into the pricing formula to obtain the price, and then substitute the price into the height adjustment formula to determine the height of the drone, and repeat this cycle until the changes in content placement and drone height meet the accuracy requirements.

[0087] like Figure 3 As shown, each UAV in the wireless communication system for UAV-assisted communication has a caching capability and can provide the content required by the user. In addition, the content cached by the UAV (Unmanned Aerial Vehicle) is distributed by the MBS (Macro Base Station) through the backhaul link. It is assumed that the UAV in this system is a UAV with vertical lift function, that is, the UAV can freely adjust its altitude to meet the needs of coverage and power control. In this system, there are MBS, UAVs and users. In order to improve spectrum efficiency, it is assumed that MBS and UAV share frequencies, and the sharing method is a master-slave relationship, that is, the MBS user is regarded as the primary user, and the UAV user is regarded as a secondary user, and the secondary user and the primary user share the frequency without interfering with each other. The main user is represented as ; , secondary user Expressed as In addition, the UAV collection is .

[0088] Since the LOS (Line-of-Sight propagation) propagation mode is mainly between the drone and the user, In the above equation, the channel gain between the UAV and the user can be expressed as a free space signal fading model:

[0089] ; (5)

[0090] in, Is a user and drones Between The distance during a time slot is the distance of the user To the drone The transmit power of the transmitted signal, is the path loss index. UAV and user The three-dimensional coordinates of and , so that Given by:

[0091] ; (6)

[0092] This embodiment only needs to adjust the vertical lift height of the UAV to meet the user's needs. Therefore, after the UAV is configured, the distance between the UAV and the user is only a function of time. Assume that the UAV uses TDMA (Time Division Multiple Access) for downlink communication. Therefore, the distance between the UAV and the user is only a function of time. There is no interruption in the download of content between users. The downlink rate can be expressed as:

[0093] ; (7)

[0094] The system model uses a master-slave model to achieve spectrum sharing. It will interfere with MBS users when communicating At this time, MBS and users The communication rate can be expressed as:

[0095] ; (8)

[0096] Because the system model takes into account the way drone users and MBS users use the shared spectrum, it is necessary to ensure that the transmission power of the drone is increased while ensuring that the MBS users are not interfered with, so as to improve the service rate of the drone users. This problem can be abstracted as a Stackelberg game. Specifically, since continuous power control is more difficult and drone altitude adjustment is easier to achieve, in the Stackelberg game, the person who gets resources first is called the leader, and the owner of the remaining resources becomes the follower. In this game, the leader's income can be expressed as:

[0097] ; (9)

[0098] in is the pricing vector, is the distance vector. In the case of limited interference tolerance of MBS, the price vector can be optimized by solving the following problem: :

[0099] ; (10)

[0100] ; (11)

[0101] UAV user benefits can be formulated as the following question:

[0102] ; (12)

[0103] ; (13)

[0104] ; (14)

[0105] in:

[0106] ; (15)

[0107] in is the repeated storage cost, which can be expressed as:

[0108] ; (16)

[0109] represents the similarity between nodes i and j, that is, different users have the same demand for the same content. Represents the cache capacity of the corresponding node. Display content The distribution ratio of The result of optimizing the placement of content can be expressed as:

[0110] ; (17)

[0111] It is popular content In time The probability of prevalence, Indicates the allocated value of the link bandwidth. represents the remaining cache capacity of the corresponding node. Here, m* and Γ* can be solved by formulas (31 and 32).

[0112] ; (31)

[0113] ; (32)

[0114] Furthermore, in order to improve user About content The probability of successful downloading needs to be reduced by adjusting the flight altitude of each drone. For the proposed Stackelberg game, the Stackelberg game equilibrium (SE) is defined as follows.

[0115] Definition 1: Let and is the solution to problem 1, point is the SE of the proposed Stackelberg game if the following conditions are met:

[0116] ; (18)

[0117] ; (19)

[0118] in, and The Nash equilibrium of the Stackelberg game can be obtained by finding the Nash equilibrium of the subgame. In the game, it is assumed that each player can get a working point. At this working point, each player cannot unilaterally increase his profit by changing his strategy.

[0119] Specifically, the process of optimizing the game problem is as follows: It can be proved that the objective function of problem 2 is The concave function of , the constraint is affine. Therefore, Problem 2 is a convex problem. Since the duality gap between Problem 2 and its dual optimization problem is zero, the problem is solvable.

[0120] ; (20)

[0121] ;(twenty one)

[0122] ;(twenty two)

[0123] Furthermore, the optimization problem in the game is solved by using a non-uniform pricing scheme.3

[0124] ;(twenty three)

[0125] ;(twenty four)

[0126] The solution to problem 3 is as follows. The detailed derivation is given in

[0127] ; (25)

[0128] ; (26)

[0129] The solution to problem 4 is as follows. The detailed derivation process is shown in Substituting the Lagrange multiplier into the objective function, the following results are obtained:

[0130] ; (33)

[0131] Then the KKT condition can be written as:

[0132] ; (34)

[0133] ; (35)

[0134] ; (36)

[0135] ; (37)

[0136] From formula (34), we can get: ,because so By solving equation (34), we have the following equation:

[0137] ; (38)

[0138] Assumptions So ,thereby , then it and is contradictory, so this assumption is not true, then only , substituting formula (39) into formula (38) yields the pricing:

[0139] ; (39)

[0140] That is, the pricing is as follows: ; (40)

[0141] According to equation (40), the range of cumulative interference can be obtained as follows.

[0142] ; (27)

[0143] Based on the above formula, we can get the maximum tolerance of MBS. Cumulative interference per user as follows:

[0144] ; (28)

[0145] Here, we assume that the distance adjustment granularity is ∆=dmax-dmin / V, then the number of players participating in the game is The cumulative interference threshold of the drones is:

[0146] ; (29)

[0147] The solution to Problem 4 gives the pricing expressed by equation (30):

[0148] ; (30)

[0149] Furthermore, the iterative process of the joint optimization algorithm is:

[0150] Input: ;

[0151] Output: ;

[0152] repeat;

[0153] Q1: Execution Optimization Trajectory algorithm;

[0154] Input: ;

[0155] Output: ;

[0156] R1: Satellite nodes periodically publish the content preferences of users in the region, which can be obtained ;

[0157] R2: if then;

[0158] R3: Gotoline;

[0159] R4: else;

[0160] renew value

[0161] endif;

[0162] R5: solve (41) and (43) to obtain the optimized trajectory;

[0163] R6: According to , calculate .

[0164] ; (41)

[0165] ; (42)

[0166] ; (43)

[0167] get ;

[0168] Q2: Substitute into the height adjustment formula (23) to obtain ;

[0169] Q3: Will and Substitute into equation (30) to obtain the pricing ;

[0170] Q4: Substituting into equation (23) we obtain ;

[0171] Q5: unti .

[0172] Execution of the algorithm: Optimization results The initial value of and the interference tolerance allowed by MBS are given in the joint optimization algorithm to obtain the number of drones participating in the game, and the pricing of each drone is obtained according to the pricing formula. Finally, the user's price is calculated according to formula (39). Given the initial power value, the initial height of the drone is calculated based on this initial value. In this algorithm, the two parameters and Used to control the accuracy of algorithm execution.

[0173] Algorithm Implementation and Complexity Analysis: Optimization The algorithmic complexity of the trajectory algorithm is In the game model of this paper, through the joint optimization of content placement and drone altitude, the relative execution time of the algorithm comes from the accuracy requirements of content placement and altitude adjustment. Therefore, the complexity of the joint optimization algorithm is The adjustment accuracy of the drone's altitude depends on the drone's manufacturing requirements, time slot The number of drones depends on the protocol design and the coverage design. Therefore, the time complexity of the joint optimization algorithm is simplified to .

[0174] S5. Collect user satisfaction data of each densely populated area, process it to obtain a user satisfaction evaluation value of each densely populated area, and provide feedback for drone trajectory optimization based on the user satisfaction evaluation value of each densely populated area.

[0175] Specifically, feedback is given to the optimization of the drone trajectory, and the specific analysis process is as follows: user satisfaction data of each densely populated area is collected, and the comprehensive satisfaction evaluation index of each user in each densely populated area is obtained through processing.

[0176] Specifically, the user satisfaction data refers to the overall satisfaction score of users in each densely populated area with the quality of assisted ground communications after drone trajectory optimization, including signal strength perception score, signal stability score, data transmission rate score, real-time perception score, response time score, video call quality score and image transmission quality score, and the value range is set to 1-5.

[0177] Furthermore, considering the correlation between parameters, the correlation coefficient matrix R is introduced to represent the correlation between parameter i and parameter j (the value range is -1 to 1, -1 indicates a complete negative correlation, 0 indicates no correlation, and 1 indicates a complete positive correlation). For example, there is a positive correlation between signal strength perception and signal stability, which may be a value close to 1.

[0178] Define scenario factors to dynamically adjust the weights of each parameter according to different usage scenarios. Assume that there are m usage scenarios, such as aerial photography, logistics distribution, communication relay, etc. The weights of each parameter are different in each scenario. For the kth scenario, the weight vector of each parameter is , and satisfies .

[0179] In a specific embodiment, the numerical expression of the comprehensive satisfaction evaluation index of each user in each densely populated area is:

[0180] ; (44)

[0181] in, represents the comprehensive satisfaction evaluation index of the yth user in the nth dense area, represents the satisfaction test score of the ith category of the yth user in the nth dense area, represents the j-th category satisfaction test score of the y-th user in the n-th dense area, is an adjustment coefficient used to control the influence of the association between parameters on the comprehensive score. = 0, the formula degenerates into a simple weighted average form. , When the parameters are correlated, they will have an impact on the comprehensive satisfaction evaluation index.

[0182] It should be explained that this embodiment comprehensively analyzes user satisfaction under different research needs and application scenarios. This refined evaluation can provide more targeted suggestions and guidance for drone trajectory optimization and communication quality improvement, which helps to improve the overall performance and user experience of the drone system.

[0183] The comprehensive satisfaction evaluation indicators of each user in each densely populated area are sorted in ascending order to obtain a user comprehensive satisfaction evaluation indicator sequence, the number of users participating in the feedback is counted, the user comprehensive satisfaction evaluation indicator sequence is processed to obtain a secondary sequence of user comprehensive satisfaction evaluation indicators, and the user comprehensive satisfaction evaluation indicators of the secondary sequence of user comprehensive satisfaction evaluation indicators are averaged to obtain the user comprehensive satisfaction evaluation indicator parameters.

[0184] The comprehensive satisfaction evaluation index of each user in each densely populated area is compared with the user comprehensive satisfaction evaluation index parameter, and the number of users in each densely populated area whose comprehensive satisfaction evaluation index is higher than or equal to the user comprehensive satisfaction evaluation index parameter is recorded and marked as the number of satisfied users. If the ratio of the number of satisfied users to the number of users is greater than or equal to the pass rate, the drone trajectory optimization is marked as qualified. Otherwise, the drone trajectory needs to be optimized again.

[0185] In a specific embodiment, the pass rate can be set to 60% or 80%.

[0186] It should be explained that this embodiment fully considers the timeliness, affinity, regionality and other spatiotemporal distribution characteristics and rate requirements of the content required by users, and effectively improves user satisfaction with content downloading by jointly optimizing content placement and drone altitude. It also achieves the adjustment of drone altitude based on Stackelberg game while ensuring that MBS users are not disturbed, thereby improving the service rate of drone users. It solves the problem of poor optimization performance caused by power control discontinuity, and provides new ideas and methods for improving network throughput and reducing user energy consumption. At the same time, this embodiment also analyzes the complexity of the algorithm, clarifies the relationship between the relative execution time of the algorithm and the content placement accuracy, altitude adjustment accuracy, etc., and provides valuable reference for subsequent research in algorithm optimization and practical application deployment, which will help promote the further development of drone networks in the direction of satisfying user service experience.

[0187] Reference Figure 2 As shown, the second aspect of the present invention provides a UAV trajectory optimization and cache scheduling method system based on image recognition, including: a UAV initial height setting module, a UAV initial horizontal position setting module, a UAV trajectory pre-adjustment module, a UAV trajectory optimization module and a comprehensive feedback module.

[0188] The drone initial height setting module is used to obtain high-definition images of the area to be served, grid the area to be served, obtain each service grid, collect the number of users in each service grid in real time, process to obtain the user density of each service grid, jointly mark adjacent grids with user density in each user-dense interval value as user-dense areas, obtain the geographical environment data of each user-dense area, and set the drone initial height based on the horizontal density of each user-dense area.

[0189] The module for setting the initial horizontal position of the UAV is used to collect the communication link quality data between the UAV and users in the user-dense area, obtain the quality evaluation index of each communication link after processing, and set the initial horizontal position of the UAV according to the quality evaluation index of each communication link combined with the signal strength between the ground base station and the user-dense area.

[0190] The drone trajectory optimization module is used to build a wireless communication system model and joint optimization algorithm including MBS, drones and users, and iteratively optimize content placement and drone height through the joint optimization algorithm.

[0191] The comprehensive feedback module is used to collect user satisfaction data of each dense area, process it to obtain the user satisfaction evaluation value of each dense area, and provide feedback on the drone trajectory optimization according to the user satisfaction evaluation value of each dense area.

[0192] The above contents are merely examples and explanations of the structure of the present invention. The technicians in this technical field may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the claims, they should all fall within the protection scope of the present invention.

Claims

1. The UAV trajectory optimization and cache scheduling method based on image recognition is characterized by: include: S1. Obtain a high-definition image of the area to be served, grid the area to be served, obtain each service grid, collect the number of users in each service grid in real time, and process to obtain the user density of each service grid; S2, jointly mark the adjacent grids with user density in each user-dense interval value as user-dense areas, obtain the geographical environment data of each user-dense area, and set the initial height of the drone based on the horizontal density of each user-dense area; S3, collecting communication link quality data between the UAV and users in the user-dense area, obtaining each communication link quality evaluation index after processing, and setting the initial horizontal position of the UAV according to each communication link quality evaluation index combined with the signal strength between the ground base station and the user-dense area; S4. Build a wireless communication system model and joint optimization algorithm including MBS, drones and users, and iteratively optimize content placement and drone height through the joint optimization algorithm; S5. Collect user satisfaction data of each dense area, process to obtain user satisfaction evaluation values ​​of each dense area, and provide feedback for drone trajectory optimization according to the user satisfaction evaluation values ​​of each dense area; The above processing obtains the user density of each service grid. The specific analysis process is as follows: The area of ​​the area to be served is extracted from the high-definition image of the area to be served, and the area to be served is divided into grids according to the preset service demand level of the area to be served, and marked as each service grid; According to the preset time collection point, the number of users in each service grid is collected at a fixed point to obtain the number of users in each service grid at each time collection point, and the user density of each service grid is obtained according to the area of ​​the service grid; The communication link quality data between the user and the drone includes: received signal strength, bit error rate, signal-to-noise ratio, delay and packet loss rate at each moment, and the critical received signal strength, critical bit error rate, critical signal-to-noise ratio, critical delay and critical packet loss rate and critical signal strength are extracted from the regional information database; The numerical expression of each communication link quality evaluation index is: ； in, Indicates A communication link quality assessment index, Indicates the number of each communication link, , g represents the total number of communication links, Indicates The communication link is The received signal strength at a given moment, Indicates the number of each moment, , s represents the total number of moments, Indicates The communication link is The bit error rate at a moment, Indicates The communication link is The signal-to-noise ratio at a given moment, Indicates The communication link is A time delay of one moment, Indicates The communication link is The packet loss rate at a moment, Indicates The average signal strength of the communication link, Indicates the critical received signal strength, represents the critical bit error rate, represents the critical signal-to-noise ratio, represents the critical delay, represents the critical packet loss rate, represents the critical signal strength, Indicates the communication link quality impact characteristic factor corresponding to the preset received signal strength, Indicates the communication link quality impact characteristic factor corresponding to the preset bit error rate, Indicates the communication link quality impact characteristic factor corresponding to the preset signal-to-noise ratio, Indicates the communication link quality impact characteristic factor corresponding to the preset delay, Indicates the communication link quality impact characteristic factor corresponding to the preset packet loss rate, Indicates the communication link quality impact characteristic factor corresponding to the preset signal strength.

2. The method for optimizing the trajectory of unmanned aerial vehicles and caching the unmanned aerial vehicles based on image recognition according to claim 1 is characterized in that: The initial height of the drone is set in combination with the horizontal density of each user-dense area. The specific analysis process is as follows: Extracting a user density standard value from a regional information database, and jointly marking adjacent grids whose user density is higher than or equal to the user density standard value as a user density area; Extract the geographical environment data of each user-dense area from the high-definition image of the area to be served, including the building height, altitude, ground slope and building density of each collection point, take the deviation value of the building height, altitude and ground slope of each collection point from the average building height, altitude and ground slope of the user-dense area, and then compare each deviation value with the average building height, altitude and ground slope of the user-dense area, respectively, and combine the ratio between the building density of the user-dense area and the critical value of the building density of the user-dense area, and introduce the influencing weight factor to obtain the geographical environment flatness evaluation index of each user-dense area; Performing a standard deviation operation on the user density in each service grid in each user-dense area to obtain an evaluation value of the uniformity of user distribution in each user-dense area; The initial height of the drone is set according to the geographical environment flatness evaluation index and user distribution uniformity evaluation value of each user-dense area.

3. The method for optimizing the trajectory and caching the drone based on image recognition according to claim 2 is characterized in that: The initial height of the drone is set according to the geographical environment flatness evaluation index and the user distribution uniformity evaluation value of each user-dense area. The specific process is as follows: A mapping set between the user distribution uniformity evaluation value in the user-dense area and the initial height of the drone and a mapping set between the geographical environment flatness evaluation index and the initial height of the drone are extracted from the regional information database. The mapping relationship is one-to-one. Two initial height setting values ​​of the drone are obtained according to the mapping set, and the average value is taken, and the average value is set as the initial height of the drone.

4. The method for optimizing the trajectory of a UAV and caching it based on image recognition according to claim 1 is characterized in that: The initial horizontal position of the UAV is set according to the quality evaluation index of each communication link combined with the signal strength between the ground base station and the user-dense area. The specific process is as follows: Deploy several monitoring positions between the user-dense area and the ground base station, wherein the monitoring positions are all located on the initial altitude line of the UAV, collect the signal strength between the UAV and the ground base station at each time monitoring point of each monitoring position, sum and average the signal strength between the UAV and the ground base station at each time monitoring point of each monitoring position, and obtain the average signal strength between the UAV and the ground base station; The communication link quality data between users in user-dense areas and drones at each monitoring location are collected at preset time intervals, and the quality evaluation index of each communication link is obtained by combining the average signal strength between the drone at each monitoring location and the ground base station. The initial horizontal position of the drone is set according to the quality evaluation index of each communication link.

5. The method for optimizing the trajectory and caching the drone based on image recognition according to claim 4 is characterized in that: The specific process of setting the initial horizontal position of the drone according to the quality evaluation index of each communication link is as follows: The communication link quality evaluation index is used to quantitatively evaluate the communication link quality between the drone and the ground base station and between the drone and the user equipment in the user-dense area when the drone is at each monitoring position. The communication link quality evaluation index is sorted in descending order, and the monitoring position point where the drone is located when the communication link quality evaluation index is the largest is set as the initial horizontal position of the drone.

6. The method for optimizing the trajectory and caching the drone based on image recognition according to claim 1 is characterized in that: The specific analysis process of constructing a wireless communication system model and a joint optimization algorithm including MBS, drones and users is as follows: The wireless communication system model includes a network model and a channel model; The network model includes MBS, drones and users. Drones and MBS share spectrum, MBS users are primary users, and drone users are secondary users. The channel model uses a free space signal fading model to reflect the channel gain between the drone and the user. The drone optimizes the user's download rate by adjusting its altitude. The joint optimization algorithm includes an optimization target and a Stackelberg game model.

7. The method for optimizing the trajectory of unmanned aerial vehicles and caching the unmanned aerial vehicles based on image recognition according to claim 6 is characterized in that: The joint optimization algorithm is used to iteratively optimize content placement and drone height. The specific analysis process is as follows: Input relevant initial values, and then execute the optimized trajectory algorithm to optimize content placement to obtain parameter one, substitute it into the height adjustment formula to calculate parameter two, then combine parameter one and parameter two and substitute them into the pricing formula to get the price, and then substitute the price into the height adjustment formula to determine the height of the drone, and repeat this cycle until the changes in content placement and drone height meet the accuracy requirements.

8. The method for optimizing the trajectory and caching the drone based on image recognition according to claim 1 is characterized by: The feedback of the optimization of the UAV trajectory is analyzed in detail as follows: Collect user satisfaction data in each densely populated area, and process it to obtain comprehensive satisfaction evaluation indicators for each user in each densely populated area; The comprehensive satisfaction evaluation index of each user in each densely populated area is sorted in ascending order to obtain a user comprehensive satisfaction evaluation index sequence, the number of users participating in the feedback is counted, the user comprehensive satisfaction evaluation index sequence is processed to obtain a secondary sequence of user comprehensive satisfaction evaluation indexes, and the user comprehensive satisfaction evaluation indexes of the secondary sequence of user comprehensive satisfaction evaluation indexes are averaged to obtain user comprehensive satisfaction evaluation index parameters; The comprehensive satisfaction evaluation index of each user in each densely populated area is compared with the user comprehensive satisfaction evaluation index parameter, and the number of users in each densely populated area whose comprehensive satisfaction evaluation index is higher than or equal to the user comprehensive satisfaction evaluation index parameter is recorded and marked as the number of satisfied users. If the ratio of the number of satisfied users to the number of users is greater than or equal to the pass rate, the drone trajectory optimization is marked as qualified. Otherwise, the drone trajectory needs to be optimized again.

9. A drone trajectory optimization and cache scheduling system based on image recognition, applying a drone trajectory optimization and cache scheduling method based on image recognition as claimed in any one of claims 1 to 8, characterized in that: include: The module for setting the initial altitude of the drone is used to obtain high-definition images of the area to be served, to grid the area to be served, to obtain each service grid, to collect the number of users in each service grid in real time, to process and obtain the user density of each service grid, to jointly mark the adjacent grids with user density in each user-dense interval as user-dense areas, to obtain the geographical environment data of each user-dense area, and to set the initial altitude of the drone in combination with the horizontal density of each user-dense area; The module for setting the initial horizontal position of the UAV is used to collect the communication link quality data between the UAV and the users in the user-dense area, obtain the quality evaluation index of each communication link after processing, and set the initial horizontal position of the UAV according to the quality evaluation index of each communication link combined with the signal strength between the ground base station and the user-dense area; The drone trajectory optimization module is used to build a wireless communication system model and joint optimization algorithm including MBS, drones and users, and iteratively optimize content placement and drone altitude through the joint optimization algorithm; The comprehensive feedback module is used to collect user satisfaction data of each dense area, process it to obtain the user satisfaction evaluation value of each dense area, and provide feedback on the drone trajectory optimization according to the user satisfaction evaluation value of each dense area.

Images