Unmanned aerial vehicle trajectory planning method and device for edge data recovery

Abstract

The invention discloses an unmanned aerial vehicle trajectory planning method and device for edge data recovery, relates to the field of navigation, and is used for improving the reliability of a data recovery task under the condition of ensuring the data recovery efficiency. The method comprises the following steps: acquiring edge node information of all target edge nodes; acquiring basic parameters and initial state information of the unmanned aerial vehicle, and further acquiring unit energy consumption of the unmanned aerial vehicle; calculating the flight energy consumption of the unmanned aerial vehicle between every two nodes, and constructing an edge node geographic information model; iterating the following steps until all the edge nodes are accessed to obtain a flight trajectory of the unmanned aerial vehicle: planning an access trajectory corresponding to the maximum number of access nodes by adopting a greedy strategy according to the geographic information model of the edge nodes and the state information of the unmanned aerial vehicle; and the unmanned aerial vehicle accesses the nodes according to the access track in sequence and updates the state information of the unmanned aerial vehicle. The data recovery reliability can be effectively improved, and the data recovery efficiency and the energy consumption efficiency of the system are ensured.

Description

Technical Field

[0001] The present invention relates to the field of navigation, and in particular to a method and device for unmanned aerial vehicle trajectory planning for edge data recovery. Background Art

[0002] With the booming development of edge computing, more and more tasks are being transferred from traditional cloud computing centers to edge nodes. However, this process also leads to the generation of a large amount of edge data, which often needs to be recovered and processed in a timely manner to meet the needs of various application scenarios. In many practical scenarios, network latency is an issue that cannot be ignored, especially in harsh environments, such as remote areas, natural disaster sites, or industrial parks, where the coverage of wireless networks is limited and even reliable connections cannot be established, resulting in huge challenges for data transmission. Therefore, how to efficiently recycle the data generated by edge nodes has become an urgent problem to be solved.

[0003] Considering these issues, drones have attracted widespread attention as a flexible and efficient data recovery tool. Drones are highly maneuverable and can perform data collection and recovery tasks in various complex environments. By deploying drones for data recovery, the limitations of traditional wireless network transmission can be overcome, ensuring that data can be delivered to the processing center in a timely and secure manner. However, in the process of using drones to recover data from multiple edge nodes, how to plan a path that makes the overall data collection time-saving and energy-efficient has become a hot topic in current research.

[0004] Existing UAV trajectory optimization algorithms usually focus on the shortest distance or minimum time of the path, such as the "A Path Planning Method for UAV in Data Collection Process" proposed in the Chinese patent document with publication number CN112902969A, which constructs a global path planning problem model for UAVs with the goal of minimizing the flight path of the UAV. These planning methods ignore the objective condition that the battery power cannot meet the energy consumption requirements of one-time data collection in actual scenarios. This neglect may cause the UAV to run out of power on the way back, making it unable to complete the mission, or even causing the UAV to crash, causing equipment damage and safety hazards. Therefore, in the trajectory planning process, the UAV's power status needs to be used as a constraint condition for trajectory decision-making to improve the availability and safety of the UAV. Summary of the invention

[0005] The purpose of the present invention is to provide a UAV trajectory planning method and device for edge data recovery to address all or part of the above-mentioned problems, so as to improve the reliability of data recovery tasks while ensuring data recovery efficiency.

[0006] The technical solution adopted by the present invention is as follows:

[0007] A UAV trajectory planning method for edge data recovery includes the following processes:

[0008] S1. Obtain edge node information of all target edge nodes;

[0009] S2. Obtain the basic parameters and initial state information of the UAV, and obtain the working power based on the basic parameters of the UAV;

[0010] S3. Based on the edge node information, basic parameters and working power of each edge node, the flight energy consumption of the drone between two nodes is calculated, and an edge node geographic information model is constructed;

[0011] Iterate steps S4-S5 until all edge nodes have been visited:

[0012] S4. Based on the edge node geographic information model and the status information of the drone, a greedy strategy is used to plan the access trajectory corresponding to the maximum number of access nodes, while ensuring that the drone can return to the starting point;

[0013] S5. The drone visits nodes in sequence according to the access trajectory, and updates the state information of the drone according to the edge node geographic information model.

[0014] Furthermore, the step S3 includes:

[0015] S3.1. Calculate the flight time of the drone between two nodes based on the edge node information of each edge node and the basic parameters of the drone.

[0016] S3.2, calculate the flight energy consumption of the drone between two nodes according to the unit energy consumption of the drone;

[0017] S3.3. Construct an energy consumption relationship library that expresses the energy consumption of drone flight between two nodes, and initialize the set of edge nodes that the drone has visited.

[0018] Further, the step S4 includes:

[0019] Under the constraint of the state information of the UAV, according to the edge node geographic information model, the maximum number of edge nodes that the UAV can continue to visit and eventually return to the starting point starting from each current optional edge node is obtained, and the edge node with the largest maximum number of visited edge nodes among the traversed optional edge nodes is selected as the next node, and the path with the largest maximum number of visited edge nodes corresponding to the node is combined with the path returning to the starting point as the planned access trajectory; if there is no optional edge node, the path returning to the starting point is used as the planned access trajectory.

[0020] Furthermore, the method for determining the optional edge nodes includes:

[0021] Filter out unvisited edge nodes from all target edge nodes;

[0022] Traverse the edge nodes that have not been visited. If the state information of the drone supports the drone starting from the current node to reach the currently traversed edge node and then returning to the starting point from the edge node, the currently traversed edge node is determined to be an optional edge node.

[0023] Furthermore, the method for obtaining the maximum number of access edge nodes includes:

[0024] Traverse each optional edge node, obtain all remaining optional edge nodes that can be accessed and returned to the starting point from the currently traversed edge node as the next node, then traverse each next node separately, and update the status information of the drone and the optional edge nodes when traversing the next node accordingly, obtain all remaining optional edge nodes that can be accessed and returned to the starting point as the next next node, and so on, until there are no more optional edge nodes; count the number of edge nodes in all paths passed from the currently traversed edge node, select the maximum value, and record the corresponding node access order.

[0025] Furthermore, when traversing each optional edge node, the remaining power of the drone after visiting the maximum number of edge nodes is calculated respectively; when there are more than one optional edge nodes with the largest number of visited edge nodes, the edge node with the least remaining power of the drone is selected as the next node.

[0026] Furthermore, when there is more than one optional edge node with the least remaining power of the drone, the next node is determined from the edge nodes with the least remaining power of the drone according to uniform distribution sampling.

[0027] Furthermore, the step S5 comprises:

[0028] The drone visits each node in the planned access trajectory in sequence. For each next node visited:

[0029] If the next node is the starting point, the drone's status information is updated to the optimal state;

[0030] If the next node is an edge node, the flight energy consumption from the current node to the next node is deducted based on the status information of the current node drone, and the next node is updated to the visited edge node set.

[0031] The present invention also provides a UAV trajectory planning device for edge data recovery, comprising a processor and a storage medium, wherein the processor runs a computer program in the storage medium to execute the above-mentioned UAV trajectory planning method for edge data recovery.

[0032] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0033] The present invention considers the starting point (rechargeable point) in the trajectory planning process, and adopts a greedy strategy-based algorithm to make trajectory point decisions under the constraints of the drone state information, which can effectively optimize the length and energy consumption of the trajectory path. This method not only ensures that the drone still has enough power to return to the charging station after completing the data recovery task, thereby effectively avoiding the problem of power exhaustion, but also ensures that each flight can access as many edge nodes as possible, taking into account the data recovery efficiency and the reliability of task execution, and by deciding on multiple access node trajectories at one time, it reduces the algorithm execution time and enhances the adaptability of the drone in complex environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The present invention will now be described by way of example with reference to the accompanying drawings, in which:

[0035] Figure 1 It is a flow chart of the UAV trajectory planning method for edge data recovery in an embodiment of the present application.

[0036] Figure 2 It is a node jump tree diagram when accessing the node path after planning in an embodiment of the present application.

[0037] Figure 3 , Figure 4 They are comparative experimental result diagrams of the completion rate and execution time of drone access to nodes under different algorithms provided in the embodiments of the present invention, wherein the algorithm of the present application is named EnerDrontic. DETAILED DESCRIPTION

[0038] All features disclosed in this specification, or steps in all methods or processes disclosed, except mutually exclusive features and / or steps, can be combined in any manner.

[0039] Any feature disclosed in this specification (including any additional claims and abstract), unless otherwise stated, may be replaced by other equivalent or alternative features having similar purposes. That is, unless otherwise stated, each feature is only an example of a series of equivalent or similar features.

[0040] In view of the fact that current drone trajectory planning algorithms usually focus on the shortest path or the shortest time, but ignore the potential risks caused by whether the drone battery power can meet the planned path, the embodiments of the present application provide a drone trajectory planning method for edge data recovery, which aims to ensure the reliability of data recovery tasks while recovering data as efficiently as possible.

[0041] like Figure 1 As shown, the drone trajectory planning method for edge data recovery proposed in the embodiment of the present application includes the following process:

[0042] S1. Obtain edge node information of all target edge nodes.

[0043] The so-called target edge nodes are the edge nodes that the drone wants to visit. These edge nodes usually have data to be recovered. Of course, there may also be edge nodes that do not have data to be recovered but are designated to be inspected. These target edge nodes must be specified before the drone trajectory is planned. Usually, a drone is responsible for data recovery within a given geographical location. For example, electronic fence technology can be used to determine all target edge nodes within the geographical location.

[0044] In some embodiments, the edge node information includes geographic information and status information of the edge node. The geographic information is used to indicate the geographic location of the edge node, including the latitude, longitude and altitude of the edge node, and the status information is used to indicate the amount of data to be recovered by the edge node. The geographic information is used to calculate the flight energy consumption between nodes, and the status information can be used to evaluate the recovery efficiency of the planned path.

[0045] In some possible implementations, P0 represents the charging point, and P0 is set as the starting point of the UAV flight. For edge nodes P within a given geographical location, i ,1≤i≤n, establish about P i The edge node information expression S i = {G i ,H i ,D i}, where i is the edge node identifier, n is the number of target edge nodes, G i Indicates the longitude and latitude of edge node i, H i represents the altitude of edge node i, D i Indicates the amount of data to be recovered by edge node i (this data can be ignored for access trajectory planning).

[0046] S2. Obtain the basic parameters and initial state information of the UAV, and obtain the working power based on the basic parameters of the UAV.

[0047] The basic parameters of the drone belong to the attribute information of the drone configuration, including the flight speed of the drone and the maximum capacity of the drone battery. The unit energy consumption of the drone can be calculated by the flight speed, and the remaining power of the drone in the best state can be obtained by the maximum capacity of the drone battery. The initial state information of the drone is the current remaining power of the drone and the set of edge nodes that the drone has visited. The current remaining power is used to constrain the path planning conditions to ensure that the drone has enough power to return to the starting point, and the set of edge nodes that have been visited is recorded to facilitate considering only the edge nodes that have not been visited when planning the flight trajectory.

[0048] In some possible implementations, the drone substrate parameters are expressed as U = {v, B *}, where v represents the flight speed of the drone, B * Indicates the maximum battery capacity of the drone. The set of edge nodes that the drone has visited is represented by set V. The current remaining battery power of the drone can be expressed using B i It means that when reaching the edge node P i remaining power when

[0049] S3. Based on the edge node information, basic parameters and working power of each edge node, the flight energy consumption of the drone between two nodes is calculated, and an edge node geographic information model is constructed.

[0050] The nodes of drone flight include starting points and edge nodes. By constructing an edge node geographic information model, it can provide decision-making support data for drone trajectory planning.

[0051] In some embodiments, step S3 includes the following sub-steps:

[0052] S3.1. Calculate the flight time of the drone between any two nodes based on the edge node information of each edge node and the basic parameters of the drone.

[0053] Since the edge node information of each edge node can determine the position of each edge node, the optimal flight trajectory q between every two nodes i and j can be obtained by combining the position and the flight dynamics of the drone i,j ,0≤i,j≤n,i≠j, n represents the number of edge nodes; and the flight time of the drone between every two nodes can be calculated based on the flight speed given by the basic parameters of the drone.

[0054] S3.2. Calculate the flight energy consumption of the drone between two nodes based on the unit energy consumption of the drone.

[0055] The flight energy consumption of the drone between two nodes can be calculated by multiplying the flight time between two nodes calculated in sub-step S3.1 by the working power of the drone.

[0056] S3.3. Construct an energy consumption relationship library that expresses the energy consumption of drone flight between two nodes, and initialize the set of edge nodes that the drone has visited.

[0057] In some possible implementations, the energy consumption relationship library is expressed in the form of a two-dimensional matrix. For example, the energy consumption relationship library is expressed using a two-dimensional matrix X:

[0058]

[0059] Among them, X i,j Represents the flight energy consumption of the UAV from node i to node j.

[0060] Initialize the set V to be an empty set.

[0061] After obtaining the data calculated or constructed in the above embodiments, the drone trajectory planning can be performed. By iterating step S4-step S5 until all edge nodes have been visited, the drone flight trajectory is obtained.

[0062] S4. Based on the edge node geographic information model and the status information of the drone, and on the premise of ensuring that the drone can return to the starting point, a greedy strategy is used to plan the access trajectory corresponding to the maximum number of visited nodes (in the current round).

[0063] The flight energy consumption between two nodes described by the edge node geographic information model, combined with the remaining power represented by the drone's status information, can be used to analyze whether the drone can return to the starting point after visiting the edge node.

[0064] In some embodiments, step S4 includes the following process:

[0065] Under the constraint of the state information of the UAV, according to the edge node geographic information model, the maximum number of edge nodes that the UAV can continue to visit and eventually return to the starting point starting from each current optional edge node is obtained, and the edge node with the largest maximum number of visited edge nodes among the traversed optional edge nodes is selected as the next node, and the path with the largest maximum number of visited edge nodes corresponding to the node is combined with the path returning to the starting point as the planned access trajectory; if there is no optional edge node, the path returning to the starting point is used as the planned access trajectory.

[0066] This step involves screening of optional edge nodes, calculation of the maximum number of access edge nodes for each optional edge node, and screening of the final access trajectory.

[0067] In some possible implementations, the method for determining an optional edge node includes:

[0068] Filter out unvisited edge nodes from all target edge nodes;

[0069] Traverse the edge nodes that have not been visited. If the state information of the drone supports the drone starting from the current node to reach the currently traversed edge node and then returning to the starting point from the edge node, the currently traversed edge node is determined to be an optional edge node.

[0070] In this implementation, assuming that the kth decision is currently being executed, P k , k∈{0,1,2,...,n} represents the current node, and uses Q k,available Indicates that from the current point P k An optional set of edge nodes to start from.

[0071] According to the flight energy consumption relationship matrix X between the nodes of the drone, all the nodes from the current point P k The optional edge nodes of the departure are put into the set Q k,available The optional edge node here indicates that the remaining power of the current drone supports the slave node P k Start, visit the edge node, and return to the starting point P0.

[0072] As a feasible implementation method, first obtain the edge nodes that the drone has not visited. After each edge node is left, the drone will add the left edge node to the visited edge node set V, using the set Q k,screen Expressed as the set of visited edge nodes, we have: Q k,screen =P * -P k -V, that is, from the target edge node set P * , remove the current edge node and the set V of visited edge nodes.

[0073] Then traverse the set Q k,screen The edge node P in s , if the remaining power of the drone meets B k -X k,s -X s,0 ≥0, then the edge node P s Put it into set Q k,available middle.

[0074] In some possible implementations, the method for obtaining the maximum number of access edge nodes includes:

[0075] Traverse each optional edge node, obtain all remaining optional edge nodes that can be accessed and returned to the starting point from the currently traversed edge node as the next node, then traverse each next node separately, and update the drone's status information and the optional edge nodes when traversing the next node accordingly, obtain all remaining optional edge nodes that can be accessed and returned to the starting point as the next next node, and so on, until there are no more optional edge nodes, respectively count the number of edge nodes in all paths passed from the currently traversed edge node, select the maximum value, and record the corresponding node access order.

[0076] As a feasible implementation method, the algorithm optPath for obtaining the maximum number of visited edge nodes is executed as follows:

[0077]

[0078] When traversing each optional edge node while planning the starting node of the access trajectory for the current node, such as Figure 2 As shown, the currently traversed optional edge node is taken as the root node, and starting from the root node, the optional edge node at the root node is taken as the first-level child node. These first-level child nodes may all be the next node starting from the currently traversed optional edge node. The flight energy consumption to reach different next nodes is different, so the state information of the drone needs to be updated correspondingly at each first-level child node, and the update of the state information of the drone will cause the optional edge nodes to change. Therefore, after updating the state information of the drone, the optional edge node set at each first-level child node needs to be updated. Similarly, each first-level child node is taken as the parent node, and its optional edge node is taken as its second-level child node. These second-level child nodes may be the next node starting from the first-level child node. At each second-level child node, the state information of the drone and the optional edge node also need to be updated. This cycle is repeated until the state information of the drone no longer supports continued access to the optional edge node, and only supports returning to the starting point, that is, there is no longer an optional edge node at a certain level of child nodes, and the edge node becomes a leaf node. The number of edge nodes passed by the longest path from the currently traversed optional edge node (ie, the root node) to all leaf nodes that it can reach is the maximum number of edge nodes visited by the currently traversed optional edge node.

[0079] As a feasible implementation, by traversing the set Q k,available The edge nodes in the traversal are repeatedly planned for the next node, the drone's status information and the optional edge nodes in a recursive manner until the drone's status information no longer supports accessing any edge nodes and then returns to the starting point. The maximum number of edge nodes visited starting from the traversed edge nodes is calculated, and the remaining power of the drone after the visit is saved accordingly.

[0080] The definition of "the farthest node without charging" means the maximum number of edge nodes that can be accessed from the current node without charging in the middle under the condition of given remaining power. k,available Each edge node in the traversal set Q has a maximum number of visited edge nodes, and the one with the largest number is the number of unfilled farthest nodes. k,available After each edge node in the set Q k,available The edge node with the largest number of uncharged farthest nodes is selected as the next node.

[0081] If there are multiple edge nodes with the same number of farthest nodes without charging, the edge node with the least remaining battery power of the drone will be selected as the next node.

[0082] If there are multiple edge nodes with the same number of farthest nodes without charging, and the remaining power of the drones is also the same, the next node is determined from the edge nodes with the least remaining power of these drones based on uniform distribution sampling.

[0083] If the current state information of the drone does not support accessing any unvisited edge nodes and then returning to the starting point, that is, the set Q k,available If it is empty, the starting point is used as the next node.

[0084] As a feasible approach, the algorithm EnerDrontic for planning access trajectories in the embodiment of the present application is executed as follows:

[0085]

[0086] S5. The drone visits nodes in sequence according to the access trajectory and updates the drone's status information based on the edge node geographic information model.

[0087] After step S4, each node that the drone is to reach in sequence can be determined. After each node the drone reaches, the remaining power will change, and the edge nodes that have been visited may also change, so both need to be updated.

[0088] In some embodiments, step S5 includes:

[0089] The drone visits each node in the planned access trajectory in sequence. For each next node visited:

[0090] If the next node is the starting point P0, the state information of the drone is updated to the optimal state, that is, B k =B * , the set V remains unchanged;

[0091] If the next node is an edge node P j(Assume that the current node is P i , the remaining power of the drone is B k-1 ), then based on the status information of the current node drone (especially the remaining power), deduct the distance from the current node to the next node P j Flight energy consumption: B k =B k-1 -X i,j , and the visited node P j Update to the visited edge node set V.

[0092] The embodiment of the present application also proposes a drone trajectory planning device for edge data recovery, the device includes a processor and a storage medium, the storage medium stores a computer program, and running the computer program can execute the drone trajectory planning method for edge data recovery in the above embodiment. The processor runs the computer program in the storage medium to execute the above drone trajectory planning method for edge data recovery.

[0093] Since the actual experiment related to drone trajectory planning takes a long time, the embodiment of the present application builds a drone trajectory planning simulation environment, and conducts comparative experiments of four methods in this simulation environment, and compares the experimental results. The experimental results are as follows: Figure 3 shown.

[0094] The embodiment of this application demonstrates the proposed algorithm and compares it with three other currently common general methods in the current scenario. The three general methods are:

[0095] (1) A* method: By defining a heuristic function to evaluate the cost of each node, the path with the minimum cost is selected. According to this method, the algorithm prioritizes the path most likely to lead to the target during the search process, ensuring that the optimal trajectory is quickly found in a complex environment.

[0096] (2) GGA method: In each generation, the best trajectory is selected for reproduction according to the fitness of the path. According to this method, individuals with higher path quality will be more likely to be selected to pass on their genes, thereby gradually optimizing the flight trajectory of the drone.

[0097] (3) ACO method: By releasing pheromones on the path to guide the choice of subsequent ants. According to this method, the concentration of pheromones with higher path quality will increase, thereby attracting more ants to choose this path, and finally forming the optimal flight trajectory of the drone.

[0098] See also Figure 3,The results show that compared with methods A*, GGA and ACO, the method proposed in this application can always maintain a 100% node access completion rate, while the node access completion rates of the other methods are very low when the battery capacity is small. This proves that the method provided by this application not only has sufficient reliability, but also has better performance in data recovery efficiency.

[0099] See also Figure 4 ,The results show that compared with methods A*, GGA and ACO, the method proposed in this application has the shortest execution time, while the other methods, especially GGA and ACO, take a long time.,This proves that the method provided in this application also has better performance in terms of execution efficiency.

[0100] The present invention is not limited to the above-mentioned specific embodiments, but extends to any new features or any new combination disclosed in this specification, as well as any new method or process steps or any new combination disclosed.

Claims

1. A UAV trajectory planning method for edge data recovery, characterized in that: The process includes: S1. Obtain edge node information of all target edge nodes; S2. Obtain the basic parameters and initial state information of the UAV, and obtain the working power based on the basic parameters of the UAV; S3. Based on the edge node information, basic parameters and working power of each edge node, the flight energy consumption of the drone between two nodes is calculated, and an edge node geographic information model is constructed; Iterate steps S4-S5 until all edge nodes have been visited: S4. Based on the edge node geographic information model and the status information of the drone, a greedy strategy is used to plan the access trajectory corresponding to the maximum number of access nodes, while ensuring that the drone can return to the starting point; S5. The drone visits nodes in sequence according to the access trajectory, and updates the state information of the drone according to the edge node geographic information model.

2. The method for UAV trajectory planning for edge data recovery according to claim 1, characterized in that: The step S3 comprises: S3.

1. Calculate the flight time of the drone between two nodes based on the edge node information of each edge node and the basic parameters of the drone. S3.2, calculate the flight energy consumption of the drone between two nodes according to the unit energy consumption of the drone; S3.

3. Construct an energy consumption relationship library that expresses the energy consumption of drone flight between two nodes, and initialize the set of edge nodes that the drone has visited.

3. The UAV trajectory planning method for edge data recovery according to claim 2, characterized in that: The step S4 comprises: Under the constraint of the state information of the UAV, according to the edge node geographic information model, the maximum number of edge nodes that the UAV can continue to visit and eventually return to the starting point starting from each current optional edge node is obtained, and the edge node with the largest maximum number of visited edge nodes among the traversed optional edge nodes is selected as the next node, and the path with the largest maximum number of visited edge nodes corresponding to the node is combined with the path returning to the starting point as the planned access trajectory; if there is no optional edge node, the path returning to the starting point is used as the planned access trajectory.

4. The method for UAV trajectory planning for edge data recovery according to claim 3, characterized in that: Methods for determining optional edge nodes include: Filter out unvisited edge nodes from all target edge nodes; Traverse the edge nodes that have not been visited. If the state information of the drone supports the drone starting from the current node to reach the currently traversed edge node and then returning to the starting point from the edge node, the currently traversed edge node is determined to be an optional edge node.

5. The method for UAV trajectory planning for edge data recovery according to claim 4, characterized in that: Methods for obtaining the maximum number of access edge nodes include: Traverse each optional edge node, obtain all remaining optional edge nodes that can be accessed and returned to the starting point from the currently traversed edge node as the next node, then traverse each next node separately, and update the status information of the drone and the optional edge nodes when traversing the next node accordingly, obtain all remaining optional edge nodes that can be accessed and returned to the starting point as the next next node, and so on, until there are no more optional edge nodes; count the number of edge nodes in all paths passed from the currently traversed edge node, select the maximum value, and record the corresponding node access order.

6. The method for UAV trajectory planning for edge data recovery according to claim 5, characterized in that: When traversing each optional edge node, the remaining power of the drone after visiting the maximum number of edge nodes is also calculated; when there are more than one optional edge nodes with the largest number of visited edge nodes, the edge node with the least remaining power of the drone is used as the next node.

7. The method for UAV trajectory planning for edge data recovery according to claim 6, characterized in that: When there is more than one optional edge node with the least remaining power of the drone, the next node is determined from the edge nodes with the least remaining power of the drone according to uniform distribution sampling.

8. The method for UAV trajectory planning for edge data recovery according to claim 2, characterized in that: The step S5 comprises: The drone visits each node in the planned access trajectory in sequence. For each next node visited: If the next node is the starting point, the drone's status information is updated to the optimal state; If the next node is an edge node, the flight energy consumption from the current node to the next node is deducted based on the status information of the current node drone, and the next node is updated to the visited edge node set.

9. A drone trajectory planning device for edge data recovery, comprising a processor and a storage medium, characterized in that: The processor runs the computer program in the storage medium to execute the drone trajectory planning method for edge data recovery as described in any one of claims 1-8.

Images