Unmanned aerial vehicle flight path planning method and device

Abstract

The invention discloses an unmanned aerial vehicle flight path planning method and device, and relates to the technical field of electric power field rescue. According to the main technical scheme of the invention, the method comprises the steps: firstly, obtaining the data of the current number of people at each preset disaster relief point and the data of a road network topological structure corresponding to the state of a distributed road network between the preset disaster relief points, and then carrying out the processing of the data information through a pre-trained preset personnel migration prediction model. And finally, according to the data of the target personnel existing at each preset disaster relief point at any specified moment and the geographic position data of each preset disaster relief point, clustering processing is carried out on the plurality of preset disaster relief points, so that the target personnel existing at each preset disaster relief point at any specified moment and the geographic position data of each preset disaster relief point are obtained. A disaster relief point set is correspondingly allocated to each different unmanned aerial vehicle, so that the flight path of one unmanned aerial vehicle is planned in the disaster relief set.

Description

Technical Field

[0001] The present application relates to the field of power field rescue technology, and in particular to a method and device for planning the flight path of an unmanned aerial vehicle. Background Art

[0002] During field power emergency rescue, drones can collect information about people gathered at the disaster relief site through the dedicated wireless communication equipment carried by rescue workers to assist the command center in assessing the disaster situation and guide subsequent rescue missions. At the same time, drones can also distribute medicines, food and other relief supplies to victims of damaged roads.

[0003] At present, with the increase in the number of disaster relief sites and to ensure the timeliness of disaster relief, it is usually necessary to send multiple drones to different disaster relief sites at the same time to perform rescue tasks such as material distribution and information collection. During the mission execution, each drone first carries the materials to the disaster relief site, then hovers to distribute the materials and collect data.

[0004] However, as the disaster relief operations continue and the roads are reopened, some disaster victims will move to safer places. The change in the number of people at the disaster relief site will lead to changes in the demand for material delivery and data collection. Since the main target of disaster relief is still personnel support, how to reasonably plan the flight paths of each drone to avoid waste of disaster relief costs and still meet the changing needs of the disaster relief site is a technical problem that needs to be solved urgently. Summary of the invention

[0005] In view of this, the present application provides a method and device for UAV flight path planning, the main purpose of which is not only to adapt to the changes in the number of people at the disaster relief site, but also to avoid waste of disaster relief costs, and to ensure the timeliness of each UAV covering each disaster relief point to meet the changing needs of the disaster relief point.

[0006] In order to achieve the above objectives, this application mainly provides the following technical solutions:

[0007] The first aspect of the present application provides a method for planning a flight path of an unmanned aerial vehicle, the method comprising:

[0008] Acquire data on the current number of people at each preset disaster relief point and road network topology data corresponding to the distribution road network status between each of the preset disaster relief points, wherein the road network topology data at least includes geographical location data of each of the preset disaster relief points;

[0009] The preset personnel migration prediction model is used to process the current number of people data and the road network topology data, and output the target personnel data predicted to exist at each preset disaster relief point at a specified time; the preset personnel migration prediction model is a model trained based on a time-series graph convolutional neural network, which is used to predict the dynamic changes of personnel data at the preset disaster relief points;

[0010] According to the target personnel data existing at each preset disaster relief point and the geographical location data corresponding to each preset disaster relief point, a plurality of preset disaster relief points are clustered to obtain a plurality of disaster relief point sets, each of which is assigned a corresponding drone to provide services;

[0011] In each of the disaster relief point sets, a flight path of the UAV is planned.

[0012] A second aspect of the present application provides a UAV flight path planning device, the device comprising:

[0013] An acquisition unit, used to acquire data on the current number of people at each preset disaster relief point and road network topology data corresponding to the distribution road network status between each of the preset disaster relief points, wherein the road network topology data at least includes geographical location data of each of the preset disaster relief points;

[0014] A first processing unit is used to process the current number of people data and the road network topology data using a preset personnel migration prediction model, and output target personnel data predicted to exist at each preset disaster relief point at a specified time; the preset personnel migration prediction model is a model trained based on a time-series graph convolutional neural network, and is used to predict dynamic changes in personnel data at the preset disaster relief points;

[0015] A second processing unit is used to perform clustering processing on the plurality of preset disaster relief points according to the target personnel data existing at each preset disaster relief point and the geographical location data corresponding to each preset disaster relief point, so as to obtain a plurality of disaster relief point sets, each of which is provided with a corresponding drone;

[0016] The determination unit is used to plan the flight path of the UAV in each set of disaster relief points.

[0017] A third aspect of the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the UAV flight path planning method as described above is implemented.

[0018] The fourth aspect of the present application provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the UAV flight path planning method as described above.

[0019] By means of the above technical solution, the technical solution provided by this application has at least the following advantages:

[0020] The present application provides a method and device for planning a flight path for a drone. The present application first obtains the current number of people data existing at each preset disaster relief point and the road network topology data corresponding to the distributed road network status between each preset disaster relief point, and then uses a pre-trained preset personnel migration prediction model to process the above data information, and outputs the target number of people data predicted to exist at each preset disaster relief point at any specified time. Since the preset personnel migration prediction model is trained based on a time-series graph convolutional neural network, and the time-series graph convolutional neural network has good performance in long-term traffic flow prediction accuracy in the transportation field, it can well predict the dynamic changes of personnel data at the preset disaster relief points. Finally, the present application clusters multiple preset disaster relief points according to the target personnel data existing at each preset disaster relief point at any specified time and the geographical location data of each preset disaster relief point, so as to achieve the allocation of a disaster relief point set to each different drone, thereby planning a flight path for a drone in the disaster relief set.

[0021] Compared with the technical problem of the prior art that it is difficult to effectively and reasonably arrange the flight paths of each drone due to the change of personnel at the disaster relief point caused by the transfer of disaster victims, the solution proposed in this application is to first assign a disaster relief point set to each drone in response to the change in the number of people, ensuring that one drone only serves one disaster relief point, and based on the fact that all drones can fully cover the disaster relief points, and secondly plan the optimal path for the drone in each disaster relief point set. Based on the solution provided by this application, it not only adapts to the change in the number of people at the disaster relief point, and avoids the waste of disaster relief costs, but also ensures that each drone covers each disaster relief point in a timely manner, meeting the changing needs of the disaster relief points.

[0022] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present application. Also, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0024] Figure 1 The embodiment of the present application shows a rescue scenario in which multiple drones perform resource delivery and data collection in emergency rescue;

[0025] Figure 2 A flow chart of a method for planning a flight path for a drone provided in an embodiment of the present application;

[0026] Figure 3 A flow chart of another method for planning a flight path for a drone provided in an embodiment of the present application;

[0027] Figure 4 A schematic diagram of the structure of a temporal graph convolutional neural network (T-GCN network) provided in an embodiment of the present application;

[0028] Figure 5 A schematic diagram of a UAV path planning solution framework provided in an embodiment of the present application;

[0029] Figure 6 A block diagram of a UAV flight path planning device provided in an embodiment of the present application;

[0030] Figure 7 A block diagram of another UAV flight path planning device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0031] The exemplary embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present application and to fully convey the scope of the present application to those skilled in the art.

[0032] During field power emergency rescue, drones can collect information about people gathered at the disaster relief site through the dedicated wireless communication equipment carried by rescue workers to assist the command center in assessing the disaster situation and guide subsequent rescue missions. At the same time, drones can also distribute medicines, food and other relief supplies to victims of damaged roads.

[0033] Now, the emergency command center in the disaster-stricken area has set up multiple disaster relief points to accommodate the victims based on public facilities such as schools and governments. Assume that the set of disaster relief points is N = {1,..,n,...,N}, and the coordinates of each disaster relief point n∈N are represented by a two-dimensional vector: [x n ,y n ] T ∈R 2 The traffic network between some disaster relief sites can be used normally, and the victims can be transferred between disaster relief sites through the road network. Therefore, the number of people in the disaster relief sites will change over time. Assume that the number of people in the disaster relief site n at the initial moment is π n init , the number of victims in the disaster relief point n at time t is π n,t . Let set E represent the set of roads between disaster relief points, and e n,n′∈E represents the roads between the disaster relief points n, n′∈N. Based on this, the geographical distribution of the disaster relief points in the entire disaster-stricken area can be represented by the graph G = (N, E).

[0034] like Figure 1 The rescue scene of multiple drones performing material delivery and data collection in emergency rescue is shown. The inventor found that because a large number of disaster-stricken users are temporarily accommodated in the disaster relief site, each disaster relief site has a demand for material distribution and data transmission, and the demand is affected by the number of disaster-stricken people gathered at the disaster relief site. The most important purpose of disaster relief is to ensure the safety of people's lives and sufficient material needs. Therefore, as the disaster relief operation continues and the roads are reopened, some disaster-stricken people will move to safer places, which will cause the number of people at the disaster relief site to change, which will inevitably affect the flight planning of drones.

[0035] Based on the above considerations, the present application embodiment provides a method for planning a flight path of a drone, such as Figure 2 As shown, the present application embodiment provides the following specific steps:

[0036] 101. Obtain data on the current number of people at each preset disaster relief point and road network topology data corresponding to the road network status distributed between each preset disaster relief point. The road network topology data at least includes geographical location data of each preset disaster relief point.

[0037] Among them, the current number of people data at each preset disaster relief point is the number of people data at each preset disaster relief point counted at a certain moment; the road network topology structure data corresponding to the distributed road network status between each preset disaster relief point includes at least: the geographical location data of each preset disaster relief point, the road network connection structure and road network status (such as blocked / unblocked) between any two different preset disaster relief points and other data information.

[0038] In the embodiment of the present application, the geographical distribution of each preset disaster relief point is known and fixed, but as the disaster relief operation continues and the roads are reopened, some disaster victims will move to safer places. This dynamic transfer of people is affected by the road conditions. Therefore, this step analyzes the dynamic changes in the number of people based on the above two parts of data information.

[0039] 102. The preset personnel migration prediction model is used to process the current number of people data and road network topology data, and output the target personnel data predicted to exist at each preset disaster relief point at a specified time; the preset personnel migration prediction model is a model trained based on a time series graph convolutional neural network, which is used to predict the dynamic changes of personnel data at the preset disaster relief points.

[0040] In the embodiment of the present application, the dynamic changes in the number of people at each preset disaster relief point are not only subject to road conditions, but also affected by time factors. It can be seen that it not only has a strong spatial correlation, but also has a strong temporal correlation. Therefore, the embodiment of the present application uses a time-series graph convolutional neural network (T-GCN network) to pre-train a preset personnel migration prediction model for dynamic changes in personnel data at preset disaster relief points. The purpose is to predict the target personnel data at each preset disaster relief point after migration at any time.

[0041] 103. Based on the target personnel data existing at each preset disaster relief point and the geographical location data corresponding to each preset disaster relief point, multiple preset disaster relief points are clustered to obtain multiple disaster relief point sets, and each disaster relief point set is assigned a corresponding drone to provide services.

[0042] 104. Plan the flight path of the drone at each disaster relief point.

[0043] After determining the number of people who have experienced migration, the target personnel data existing at each preset disaster relief point at any time is predicted, and then combined with the geographical location data corresponding to each preset disaster relief point, the embodiment of the present application can be, but is not limited to, using a clustering algorithm (such as the K-Means algorithm) to process a large number of preset disaster relief points distributed in different geographical locations to divide them into multiple disaster relief sets. In the process of this clustering process, the target personnel data at the preset disaster relief points at the specified time that is actually predicted is a weight, so that the number of people covered by each disaster relief point set is not much different; and the relative geographical location of each preset disaster relief point is another weight, so that the disaster relief point set is divided within a similar range, thereby taking into account the overall flight distance / time of the drone. In the above, while meeting the needs of a single person being rescued as much as possible, the flight cost is also taken into account.

[0044] After planning a set of disaster relief points for each drone, the optimal flight path of a drone in a disaster relief set can be planned by using, but not limited to, a minimum path algorithm.

[0045] As mentioned above, the embodiment of the present application provides a method for planning the flight path of a drone. Compared with the technical problem of the prior art that the personnel changes at the disaster relief point due to the transfer of the affected personnel make it difficult to effectively and reasonably arrange the flight paths of each drone, the solution proposed in this application is to first allocate a set of disaster relief points to each drone in response to the change in the number of people, to ensure that one drone only serves one disaster relief point, and based on the fact that all drones can fully cover the disaster relief points, and secondly plan the optimal path for the drone in each disaster relief point set. Based on the solution provided by this application, it not only adapts to the change in the number of people at the disaster relief point, and avoids the waste of disaster relief costs, but also ensures that each drone covers each disaster relief point in a timely manner, and meets the changing needs of the disaster relief point.

[0046] In some specific embodiments, the combination of Figure 1 Analysis shows that since a large number of disaster-stricken users are temporarily accommodated in the disaster relief sites, each disaster relief site has a demand for material distribution and data transmission, and the demand is affected by the number of disaster victims gathered at the disaster relief site. Suppose the data demand of disaster relief site n at time t is Material demand is They represent the average amount of data transmitted by the disaster victims and the average amount of materials required by the disaster victims, respectively. In order to quickly complete the rescue mission, multiple identical rescue drones K = {1, ..., k, ... K} depart from the same drone take-off point to serve different disaster relief points. After the drone arrives at a disaster relief point with materials, it hovers at a hovering point at a height of Dh above the disaster relief point to deliver the materials. During the distribution of materials, the drone also collects data information about the personnel at the disaster relief point. If the drone has not completed data collection after the material distribution is completed, it will continue to hover until the data collection task is completed.

[0047] Therefore, in addition to considering the impact of the migration of people at each preset disaster relief point, the embodiment of the present application further considers the changes in demand caused by the changes in the number of people at the preset disaster relief points, such as the number of materials delivered and the number of data collected. Since both the material delivery operation and the data collection operation require the consumption of drone energy, for example:

[0048] The energy loss of a drone includes the energy loss during propulsion flight, the energy loss during hovering, and the energy loss during data transmission. In particular, the energy loss during propulsion flight is also affected by the weight of the materials, and the energy consumption during hovering is mainly to ensure the time spent on the integrity of data collection at the preset disaster relief point, and the drone needs to remain hovering. It should be noted that compared with drone flight and hovering, the energy loss of data transmission is smaller, and the energy consumed is negligible. To simplify the analysis, the embodiment of this application does not consider the energy loss of drone turning, emergency stopping, acceleration and deceleration, etc.

[0049] Based on the above considerations, in some specific embodiments, in addition to allocating a set of disaster relief points to a drone, the embodiments of the present application also take into account the benefits of serving the disaster relief points (i.e., the energy loss of the drone) while optimizing the path of each drone when planning the flight path in a set of disaster relief points.

[0050] To explain it in more detail, the present application also provides another method for planning the flight path of a drone, such as Figure 3 As shown, the present application embodiment provides the following specific steps:

[0051] 201. Obtain data on the current number of people at each preset disaster relief point and road network topology data corresponding to the road network status distributed between each preset disaster relief point, wherein the road network topology data at least includes geographical location data of each preset disaster relief point.

[0052] In the embodiment of the present application, please refer to step 101 for the explanation of this step, which will not be repeated here.

[0053] 202. The preset personnel migration prediction model is used to process the current number of people data and road network topology data, and output the target personnel data predicted to exist at each preset disaster relief point at a specified time; the preset personnel migration prediction model is a model trained based on a time series graph convolutional neural network, which is used to predict the dynamic changes of personnel data at the preset disaster relief points.

[0054] This step may be, but is not limited to, refined to include the following: A1-A4;

[0055] A1. Input the current population data and road network topology data into the preset personnel migration prediction model;

[0056] A2. Use the graph convolutional network module in the preset personnel migration prediction model to process the road network topology data and capture the spatial features existing in the road network topology data;

[0057] A3, input the time series with spatial features into multiple gated recurrent unit modules for processing to capture the temporal features;

[0058] A4. Use the fully connected layer to process the time features and the current number of people at each preset disaster relief point, and output the target personnel data predicted to exist at each preset disaster relief point at the specified time.

[0059] As shown in A1-A4, the embodiment of the present application uses the T-GCN network to predict the number of people at the disaster relief point, while ensuring accuracy, and transforms the dynamic multi-UAV path planning problem into a static path planning problem for solution. Figure 4 A schematic diagram of the T-GCN network structure is shown, and a detailed explanation is given as follows:

[0060] like Figure 4 The input is the historical data of the number of people at the disaster relief site [x t-n ,x2,...,x t ] and the topological structure G = (N, E), where x t =[π 1,t-n ,π 2,t ,...,π N,t ]. The T-GCN network first uses the Graph Convolution Network (GCN) module to capture the topological structure of the road network between the disaster relief points and obtain spatial features. Secondly, the obtained time series with spatial features is input into the Gated Recurrent Unit (GRU) module, and dynamic changes are obtained through information transfer between units to capture temporal features. Finally, the number of people x at the disaster relief point after time t+T is obtained through the fully connected layer. t+T .

[0061] In some specific embodiments, the present application also uses the root mean square error (RMSE) as the loss function of the neural network to predict the number of people at the disaster relief point n at any time t. The loss function expression is as follows. Among them, the actual number of people at the disaster relief point π n,t It can be obtained from observation, and the RMSE expression is:

[0062]

[0063] Through model processing, the T-GCN network predicts the number of people who get the disaster relief point n at time t as follows:

[0064]

[0065] 203. Based on the target personnel data existing at each preset disaster relief point and the geographical location data corresponding to each preset disaster relief point, a plurality of preset disaster relief points are clustered to obtain a plurality of disaster relief point sets, and each disaster relief point set is assigned a corresponding drone to provide services.

[0066] 204. Plan the flight path of the drone at each disaster relief point.

[0067] As shown in 203-204, in this step, the embodiment of the present application can add a constraint condition and implement the planning of the flight path based on the constraint condition, which can include but is not limited to the following:

[0068] The constraints for planning the flight path include: the UAV starts from a starting position and returns to the starting position after traversing and visiting each preset disaster relief point in the disaster relief set; under the constraints, the geographic location data of each preset disaster relief point in the disaster relief point set is processed using a preset shortest path algorithm to obtain a visit order for traversing and visiting each preset disaster relief point; based on the constraints and the visit order, the flight path of the UAV is obtained.

[0069] Based on the above constraints, in order to ensure that the drone safely flies back to the starting point after completing a mission, the energy consumption of the drone needs to be considered when planning the drone path. Accordingly, the embodiment of the present application adopts the following 205-206 to provide an optimization solution.

[0070] 205. At the preset disaster relief point, control the drone to perform material distribution and data collection operations.

[0071] 206. Under the energy consumption of the drone caused by the material distribution operation and the data collection operation, the set of disaster relief points corresponding to each drone is revised and planned to obtain a new set of disaster relief points corresponding to the drone.

[0072] Specifically, after each UAV completes a flight mission, the remaining energy corresponding to each UAV is obtained. The remaining energy is the initial energy of the UAV when it takes off minus the flight propulsion energy consumed to support the material delivery operation and the energy consumed corresponding to the data operation. According to the remaining energy corresponding to the UAV, a preset greedy algorithm is used to adjust and re-divide the set of disaster relief points corresponding to each UAV.

[0073] Example 1: The embodiment of the present application uses clustering processing (such as the K-Means algorithm) to obtain the initial division of drone disaster relief points, that is, one drone is assigned a set of disaster relief points, and the shortest access path for each drone is planned and calculated based on the set of disaster relief points. The above can be used as the initial solution.

[0074] The K-Means algorithm clusters and divides the disaster relief points according to their geographical locations, and the shortest path algorithm optimizes the flight distance of a single drone, which is conducive to finding an initial solution that meets the energy constraint. However, considering the energy consumption of drones, the power of drones increases with the increase of load when the speed is constant, and the number of people at the disaster relief point also changes over time. Therefore, the solution with the shortest flight path does not necessarily obtain the optimal service benefit. In order to improve the initial solution, the algorithm alternately optimizes the disaster relief point division and single drone path planning subproblems through two layers of loops. In the outer loop, the greedy algorithm is used to optimize the disaster relief point division subproblem, and in the inner loop, the taboo search algorithm is used to optimize the single drone path planning subproblem until the termination condition is met.

[0075] Tabu search is a local search algorithm for solving combinatorial optimization problems. It is one of the most widely used metaheuristic algorithms for solving the traveling salesman problem and its variants. Therefore, it is suitable for solving the path planning problem of a single UAV with energy constraints in this study. By searching in the neighborhood of the current solution, the algorithm can gradually improve the quality of the solution, and record the solutions that have been visited within a period of time through the taboo table to prevent the same solution from being selected again in the short term.

[0076] Therefore, the embodiment of the present application first adopts an outer loop greedy algorithm to improve the overall energy efficiency ratio of the drone by re-dividing the disaster relief points. The specific operation is: each time, select the drone k′ with the most remaining energy and the drone k with the least remaining energy. Traverse the disaster relief points served by drone k, call the taboo search algorithm to calculate its total revenue after joining the service set of drone k′, and select the disaster relief point n with the largest increase in total revenue value to add to the service set of drone k′. The above implementation improves the division of the original disaster relief point set to obtain a new disaster relief point set.

[0077] 207. Re-plan the flight path of the drone based on the new disaster relief point assembly.

[0078] Specifically, according to the benefits of the UAV in performing material delivery operations and the benefits of performing data collection, a path planning evaluation function of the UAV under energy constraints is constructed. The energy constraint is to limit the energy of the UAV to be no less than zero when it returns to the starting position after completing a flight mission. Using the path planning evaluation function, a preset neighborhood movement strategy is adopted to replan the flight path of the UAV by exchanging the order of any two points on the current path.

[0079] Continuing with 206, Example 2, after obtaining the improved new disaster relief point set, the embodiment of the present application adopts the inner taboo search algorithm to find a path with a higher benefit value than the current single drone path under energy constraints. By designing the evaluation function and neighborhood movement strategy for single drone path planning under energy constraints, the optimization goal and possible improvement direction of the taboo search can be determined. The evaluation function of the taboo search for single drone path planning is:

[0080]

[0081] The function f is used to calculate the benefit value in formula (13). In order to make the UAV meet the energy limit, f = 0 when the energy limit is not met. The neighborhood movement strategy refers to the method of generating new solutions from existing solutions. In the single UAV path planning problem, the neighborhood movement strategy can be designed to exchange the current path s k The order of any two points:

[0082] s k ={n0,...,n k,i ,..,n k,j...,n0}

[0083] neibor(s k )={n0,...,n k,j ,..,n k,i ...,n0};

[0084] In the above, as shown in 201-207, the embodiment of the present application combines the migration of people at each disaster relief point and the energy consumption of the drone flight (two factors), such as Figure 5 The schematic diagram of the UAV path planning solution framework is shown. In the embodiment of the present application, the solution for planning the UAV path is divided into two sub-solutions: the disaster relief point division sub-problem and the single UAV path planning sub-problem; the disaster relief point division sub-problem is to assign each UAV to the set of disaster relief points it serves; the single UAV path planning sub-problem is to optimize the path of each UAV to increase the total revenue of its service disaster relief points. By alternately optimizing the two sub-solutions, the maximum total UAV service revenue can be obtained. This enables a better planning of drone routes while taking into account personnel migration, material delivery and data collection.

[0085] Furthermore, as a response to the above Figure 2 , 3 In order to realize the method shown in the figure, the embodiment of the present application provides a device for planning the flight path of a drone. The device embodiment corresponds to the aforementioned method embodiment. For ease of reading, the present device embodiment will not repeat the details of the aforementioned method embodiment one by one, but it should be clear that the device in this embodiment can correspond to all the contents of the aforementioned method embodiment. The device is used to achieve a better planning scheme for the drone path while taking into account the migration of personnel, the delivery of materials and the collection of data. Specifically, Figure 6 As shown, the device comprises:

[0086] An acquisition unit 31 is used to acquire data on the current number of people at each preset disaster relief point and road network topology data corresponding to the distribution road network status between each of the preset disaster relief points, wherein the road network topology data at least includes geographical location data of each of the preset disaster relief points;

[0087] The first processing unit 32 is used to process the current number of people data and the road network topology data using a preset personnel migration prediction model, and output the target personnel data predicted to exist at each preset disaster relief point at a specified time; the preset personnel migration prediction model is a model trained based on a time series graph convolutional neural network, and is used to predict the dynamic changes of personnel data at the preset disaster relief points;

[0088] The second processing unit 33 is used to perform clustering processing on the plurality of preset disaster relief points according to the target personnel data existing at each preset disaster relief point and the geographical location data corresponding to each preset disaster relief point, so as to obtain a plurality of disaster relief point sets, each of which is provided with a corresponding drone;

[0089] The determination unit 34 is used to plan the flight path of the UAV in each of the disaster relief point sets.

[0090] In some embodiments, Figure 7 As shown, the first processing unit 32 includes:

[0091] An input module 321, used for inputting the current number of people data and the road network topology data into the preset personnel migration prediction model;

[0092] A first processing module 322 is used to process the road network topology data using the graph convolutional network module in the preset personnel migration prediction model to capture the spatial features existing in the road network topology data;

[0093] A second processing module 323 is used to input the time series with the spatial features into a plurality of gated recurrent unit modules for processing to capture the time features;

[0094] The third processing module 324 is used to use the fully connected layer to process the time feature and the current number of people at each preset disaster relief point, and output the target personnel data predicted to exist at each preset disaster relief point at a specified time.

[0095] In some embodiments, Figure 7 As shown, the determining unit 34 includes:

[0096] A setting module 341 is used to set the constraint conditions for planning the flight path, wherein the constraint conditions include: the UAV starts from a starting position and returns to the starting position after traversing and visiting each of the preset disaster relief points in the disaster relief set;

[0097] A fourth processing module 342 is used to process the geographical location data of each preset disaster relief point in the set of disaster relief points by using a preset shortest path algorithm under the constraint conditions to obtain a visit sequence for traversing and visiting each preset disaster relief point;

[0098] The determination module 343 is used to obtain the flight path of the UAV according to the constraint conditions and the access sequence.

[0099] In some embodiments, Figure 7 As shown, the device also includes:

[0100] A control unit 35 is used to control the drone to perform material distribution operations and data collection operations at the preset disaster relief point;

[0101] A correction unit 36 ​​is used to correct the set of disaster relief points corresponding to each drone under the energy consumption of the drone caused by the material distribution operation and the data collection operation, so as to obtain a new set of disaster relief points corresponding to the drone;

[0102] The planning unit 37 is used to re-plan the flight path of the UAV according to the new disaster relief point set.

[0103] In some embodiments, the correction unit 36 ​​is specifically used to:

[0104] After each of the drones completes a flight mission, the remaining energy corresponding to each of the drones is obtained, where the remaining energy is the initial energy of the drone when it takes off minus the flight propulsion energy consumed to support the material delivery operation and the energy consumed corresponding to the data operation operation;

[0105] According to the residual energy corresponding to the drone, a preset greedy algorithm is used to adjust and re-divide the set of disaster relief points corresponding to each drone.

[0106] In some embodiments, the planning unit 37 is specifically used to:

[0107] According to the benefits of the drone in performing material delivery operations and data collection, a path planning evaluation function of the drone under energy constraints is constructed, wherein the energy constraint is to limit the energy of the drone to be not less than zero when it returns to the starting point after completing a flight mission;

[0108] The path planning evaluation function is used to replan the flight path of the UAV by adopting a preset neighborhood movement strategy in a manner of exchanging the order of any two points on the current path.

[0109] In summary, the UAV flight path planning device includes a processor and a memory. The above-mentioned acquisition unit, first processing unit, second processing unit and determination unit are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize corresponding functions.

[0110] The processor contains a kernel, which calls the corresponding program unit from the memory. One or more kernels can be set, and by adjusting the kernel parameters, a better planning scheme for the drone path can be achieved while taking into account personnel migration, material delivery and data collection.

[0111] An embodiment of the present application provides a storage medium having a program stored thereon, which implements a method for planning a flight path of a drone when the program is executed by a processor.

[0112] An embodiment of the present application provides a processor, which is used to run a program, wherein a method for planning a flight path of a drone is executed when the program is run.

[0113] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program that initializes the steps of the above UAV flight path planning method.

[0114] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0115] In a typical configuration, the device includes one or more processors (CPU), memory and bus. The device may also include input / output interface, network interface and the like.

[0116] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip. The memory is an example of a computer-readable medium.

[0117] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0118] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0119] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0120] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A method for planning a flight path of an unmanned aerial vehicle, characterized in that: The method comprises: Acquire data on the current number of people at each preset disaster relief point and road network topology data corresponding to the distribution road network status between each of the preset disaster relief points, wherein the road network topology data at least includes geographical location data of each of the preset disaster relief points; The preset personnel migration prediction model is used to process the current number of people data and the road network topology data, and output the target personnel data predicted to exist at each preset disaster relief point at a specified time; the preset personnel migration prediction model is a model trained based on a time-series graph convolutional neural network, which is used to predict the dynamic changes of personnel data at the preset disaster relief points; According to the target personnel data existing at each preset disaster relief point and the geographical location data corresponding to each preset disaster relief point, a plurality of preset disaster relief points are clustered to obtain a plurality of disaster relief point sets, each of which is assigned a corresponding drone to provide services; In each of the disaster relief point sets, a flight path of the UAV is planned.

2. The method according to claim 1, characterized in that: The method of using the preset personnel migration prediction model to process the current number of people data and the road network topology data, and outputting the target personnel data predicted to exist at each preset disaster relief point at a specified time, includes: Inputting the current number of people data and the road network topology data into the preset personnel migration prediction model; Processing the road network topology data using the graph convolutional network module in the preset personnel migration prediction model to capture the spatial features existing in the road network topology data; Inputting the time series with the spatial features into a plurality of gated recurrent unit modules for processing to capture the temporal features; The time feature and the current number of people at each of the preset disaster relief points are processed using a fully connected layer, and target personnel data predicted to exist at each of the preset disaster relief points at a specified time is output.

3. The method according to claim 1, characterized in that The step of planning the flight path of the UAV in each of the disaster relief point sets includes: Setting constraints for planning the flight path, wherein the constraints include: the UAV starts from a starting position and returns to the starting position after traversing and visiting each of the preset disaster relief points in the disaster relief set; Under the constraint conditions, using a preset shortest path algorithm to process the geographical location data of each of the preset disaster relief points in the disaster relief point set, and obtain a visit sequence for traversing and visiting each of the preset disaster relief points; According to the constraint conditions and the access sequence, a flight path of the UAV is obtained.

4. The method according to any one of claims 1 to 3, characterized in that At the preset disaster relief point, controlling the drone to perform material distribution operations and data collection operations; Under the energy consumption of the drone caused by the material distribution operation and the data collection operation, the set of disaster relief points corresponding to each drone is revised and planned to obtain a new set of disaster relief points corresponding to the drone; The flight path of the UAV is replanned according to the new set of disaster relief points.

5. The method according to claim 4, characterized in that The revised plan of the disaster relief point set corresponding to each drone includes: After each of the drones completes a flight mission, the remaining energy corresponding to each of the drones is obtained, where the remaining energy is the initial energy of the drone when it takes off minus the flight propulsion energy consumed to support the material delivery operation and the energy consumed corresponding to the data operation operation; According to the remaining energy corresponding to the drone, a preset greedy algorithm is used to adjust and re-divide the set of disaster relief points corresponding to each drone.

6. The method according to claim 5, characterized in that The replanning of the flight path of the UAV according to the new disaster relief point set includes: According to the benefits of the drone in performing material delivery operations and data collection, a path planning evaluation function of the drone under energy constraints is constructed, wherein the energy constraint is to limit the energy of the drone to be not less than zero when it returns to the starting point after completing a flight mission; The path planning evaluation function is used to replan the flight path of the UAV by adopting a preset neighborhood movement strategy in a manner of exchanging the order of any two points on the current path.

7. A UAV flight path planning device, characterized in that: The device comprises: An acquisition unit, used to acquire data on the current number of people at each preset disaster relief point and road network topology data corresponding to the distribution road network status between each of the preset disaster relief points, wherein the road network topology data at least includes geographical location data of each of the preset disaster relief points; A first processing unit is used to process the current number of people data and the road network topology data using a preset personnel migration prediction model, and output target personnel data predicted to exist at each preset disaster relief point at a specified time; the preset personnel migration prediction model is a model trained based on a time-series graph convolutional neural network, and is used to predict dynamic changes in personnel data at the preset disaster relief points; A second processing unit is used to perform clustering processing on the plurality of preset disaster relief points according to the target personnel data existing at each preset disaster relief point and the geographical location data corresponding to each preset disaster relief point, so as to obtain a plurality of disaster relief point sets, each of which is provided with a corresponding drone; The determination unit is used to plan the flight path of the UAV in each set of disaster relief points.

8. The device according to claim 7, characterized in that The first processing unit comprises: An input module, used for inputting the current number of people data and the road network topology data into the preset personnel migration prediction model; A first processing module, used to process the road network topology data using the graph convolutional network module in the preset personnel migration prediction model to capture the spatial features existing in the road network topology data; A second processing module, used for inputting the time series with the spatial features into a plurality of gated recurrent unit modules for processing to capture the time features; The third processing module is used to use the fully connected layer to process the time feature and the current number of people at each preset disaster relief point, and output the target personnel data predicted to exist at each preset disaster relief point at a specified time.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by the processor, the UAV flight path planning method according to any one of claims 1 to 6 is implemented.

10. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the UAV flight path planning method as described in any one of claims 1 to 6.

Images