Low-altitude route generation method, apparatus, equipment and storage medium

Abstract

This invention discloses a method, apparatus, device, and storage medium for generating low-altitude flight paths, belonging to the field of low-altitude flight technology. The method includes: acquiring a three-dimensional spatial model of the target area and its population density distribution; determining the path cost of each voxel based on its population density at the ground projection of each voxel in the three-dimensional spatial model, the obstacle status of each voxel, and the distance between each voxel and the target facility; and performing path planning with a first position in the three-dimensional spatial model as the starting point and a second position in the three-dimensional spatial model as the ending point, aiming to minimize the sum of the path costs of the voxels traversed between the first and second positions, thereby obtaining the target low-altitude flight path. This solution allows low-altitude aircraft to stay away from densely populated areas, avoid obstacles as much as possible, and fly according to the instructions of the target facility in the target area, while ensuring flight efficiency and safety.

Description

Technical Field

[0001] This invention relates to the field of low-altitude flight technology, and in particular to a method, apparatus, equipment and storage medium for generating low-altitude routes. Background Technology

[0002] In existing technologies, flight path planning for low-altitude aircraft (such as drones) mainly relies on geographic information and airspace restrictions. Currently, drone flights typically rely on Geographic Information Systems (GIS) and airspace management data to plan reasonable flight paths.

[0003] In existing technologies, urban environments are typically modeled as grid maps for path planning, and classic graph search algorithms (such as Dijkstra's algorithm) are used. These algorithms can find the shortest path in two-dimensional or simplified three-dimensional environments. However, low-altitude aircraft flying in urban airspace need to consider a large number of complex environmental factors. Even if the shortest path is found through graph search algorithms, it is difficult to meet the safety requirements for low-altitude flight.

[0004] Therefore, there is an urgent need for a method to generate low-altitude routes to guide the safe flight of aircraft on urban roads. Summary of the Invention

[0005] Therefore, it is necessary to provide a method, apparatus, equipment, and storage medium for generating low-altitude routes to address the above problems, while ensuring the flight efficiency and safety of the aircraft.

[0006] This application provides a method for generating low-altitude airways, the method comprising:

[0007] Obtain a three-dimensional spatial model of the target area and its population density distribution; the three-dimensional spatial model is marked with obstacles and target facilities.

[0008] Based on the population density of each voxel at the ground projection, the obstacle status of each voxel, and the distance between each voxel and the target facility in the three-dimensional spatial model, the path cost of each voxel is determined.

[0009] Using the first position in the three-dimensional spatial model as the starting point and the second position in the three-dimensional spatial model as the ending point, path planning is performed with the goal of minimizing the sum of the path costs of the voxels traversed between the first position and the second position to obtain the target low-altitude flight path.

[0010] In one optional implementation, determining the path cost of each voxel based on the population density at the ground projection of each voxel in the three-dimensional spatial model, the obstacle status of each voxel, and the distance of each voxel to the target facility includes:

[0011] Based on the population density of each voxel at the ground projection, obtain the population avoidance cost of each voxel;

[0012] Based on the positional relationship between each voxel and the obstacle, the obstacle cost for each voxel is obtained;

[0013] Based on the positional relationship between each voxel and the target facility, the facility affinity cost of each voxel is obtained;

[0014] The path cost of each voxel is obtained based on the population avoidance cost, the obstacle cost, and the facility affinity cost of each voxel.

[0015] In one alternative implementation, the population avoidance cost per voxel is positively correlated with the population density of each voxel at its ground projection.

[0016] In one alternative implementation, the target facility is a pre-defined road within the target area;

[0017] The process of obtaining the facility affinity cost for each voxel based on the positional relationship between each voxel and the target facility includes:

[0018] If the projection of the target voxel on the ground overlaps with the preset road, then the facility affinity cost of the target voxel is set to the first preset value.

[0019] If the projection of the target voxel on the ground does not overlap with the preset road, then a second preset value is determined based on the first distance between the target voxel and the preset road, and the second preset value is determined as the facility affinity cost of the target voxel; the second preset value is greater than the first preset value, and the facility affinity cost of the target voxel is positively correlated with the first distance.

[0020] In one optional implementation, obtaining the obstacle cost for each voxel based on a first distance between each voxel and the obstacle includes:

[0021] If the target voxel overlaps with the obstacle, the obstacle cost of the target voxel is set to a third preset value.

[0022] If the target voxel does not overlap with the obstacle, a fourth preset value is determined based on the second distance between the target voxel and the obstacle, and the fourth preset value is determined as the obstacle cost of the target voxel; the fourth preset value is less than the third preset value, and the obstacle cost of the target voxel is inversely correlated with the second distance.

[0023] In one optional implementation, the step of obtaining a target low-altitude flight path by using a first position in the three-dimensional spatial model as the starting point, a second position in the three-dimensional spatial model as the ending point, and minimizing the sum of the path costs of the voxels traversed between the first and second positions, includes:

[0024] The first position is added to the voxel set as the first intermediate voxel, and the voxel set is iteratively updated.

[0025] If the distance between a second intermediate voxel in the voxel set and the second position is less than a distance threshold, then the iterative update of the voxel set is stopped.

[0026] The target low-altitude flight path is obtained based on the candidate paths stored in the voxel set and the path connecting the second intermediate voxel to the second position; the voxel set includes candidate paths formed by connecting each intermediate voxel to its parent node.

[0027] The iterative update of the voxel set includes:

[0028] Candidate voxels are selected based on the path cost of each voxel in the 3D spatial model;

[0029] Add the candidate voxels to the voxel set;

[0030] Select the parent node of the candidate voxel from among the intermediate voxels, and establish a candidate path between the parent node of the candidate voxel and the candidate voxel.

[0031] In one alternative implementation, the step of selecting the parent node of the candidate voxel among the intermediate voxels includes:

[0032] Obtain the total path cost corresponding to the candidate voxel and each intermediate voxel; the total path cost is the sum of the path costs in the path formed by connecting the candidate voxel to the first position through the intermediate voxel;

[0033] The intermediate voxel with the lowest total path cost is selected as the parent node of the candidate voxel.

[0034] In one alternative implementation, the iterative update of the voxel set further includes:

[0035] If the candidate voxel is added to the voxel set, obtain the neighboring voxels within a specified range of the candidate voxel;

[0036] If, after connecting the neighboring voxel to the candidate voxel, the total cost of the path formed by the neighboring voxel connecting to the first position through the candidate voxel is minimized, then the parent node of the neighboring voxel is switched to the candidate voxel.

[0037] In one optional implementation, the step of selecting candidate voxels based on the path costs of each voxel in the three-dimensional spatial model includes:

[0038] Based on the path cost of each voxel in the three-dimensional spatial model, a first selection probability of each voxel in the three-dimensional spatial model is obtained; the first selection probability is inversely correlated with the path cost.

[0039] The candidate voxels are selected based on the selection probability of each voxel in the three-dimensional spatial model.

[0040] In one optional implementation, the step of selecting candidate voxels based on the path costs of each voxel in the three-dimensional spatial model includes:

[0041] Based on the path cost of each voxel in the three-dimensional spatial model, a second selection probability of each voxel in the three-dimensional spatial model is obtained; the second selection probability is inversely correlated with the path cost.

[0042] In the three-dimensional space model, candidate voxels are randomly selected, and the acceptance of a candidate voxel is determined based on the second selection probability of the candidate voxel.

[0043] If the candidate voxel is not accepted, a new candidate voxel is randomly selected until the randomly selected candidate voxel is accepted.

[0044] In an optional implementation, the method further includes:

[0045] Update the dynamic environmental parameters of the target area according to the specified period;

[0046] When the updated dynamic environment parameters meet the preset conditions, the path cost of each voxel in the three-dimensional spatial model is updated according to the updated dynamic environment parameters, and the target low-altitude route is re-planned according to the updated path cost of each voxel.

[0047] In one optional implementation, the preset conditions include at least one of the following:

[0048] The update magnitude of at least one type of data in the dynamic environment parameters is greater than the magnitude threshold.

[0049] Since the last path planning, the number of times the dynamic environment parameters have been updated has exceeded the threshold.

[0050] In one optional implementation, the step of re-planning the target low-altitude route based on the updated path costs of each voxel includes...

[0051] The updated voxel region is obtained by comparing the path cost of each voxel after the update with the path cost of each voxel before the update.

[0052] Among the target low-altitude routes in the target area, the target low-altitude routes that match the updated voxel region are selected for re-path planning.

[0053] This application also provides a low-altitude flight path generation device, the device comprising:

[0054] The parameter acquisition module is used to acquire a three-dimensional spatial model of the target area and the population density distribution; the three-dimensional spatial model is marked with obstacles and target facilities.

[0055] The cost calculation module is used to determine the path cost of each voxel based on the population density of each voxel at the ground projection, the obstacle status of each voxel, and the distance between each voxel and the target facility in the three-dimensional spatial model.

[0056] The path planning module is used to perform path planning with a first position in the three-dimensional spatial model as the starting point, a second position in the three-dimensional spatial model as the ending point, and the goal of minimizing the sum of the path costs of the voxels traversed between the first position and the second position, so as to obtain the target low-altitude route.

[0057] This application also provides an electronic device, the computer device including a processor and a memory, the memory storing at least one instruction, the at least one instruction being loaded and executed by the processor to implement the above-described low-altitude flight path generation method.

[0058] This application also provides a computer-readable storage medium storing at least one instruction, which is loaded and executed by a processor to implement the low-altitude flight path generation method described above.

[0059] Compared with the prior art, the technical solution provided in this application has the following advantages:

[0060] This application first obtains the target area where the low-altitude aircraft needs to fly, and acquires a three-dimensional spatial model and population density distribution within the target area. The three-dimensional spatial model is marked with obstacles and target facilities. At this point, the path cost of each voxel can be determined based on the population density at the ground projection of each voxel in the three-dimensional model, the obstacle status of each voxel, and the distance between each voxel and the target facility. The path cost of each voxel represents whether it is suitable for the low-altitude aircraft to fly to that voxel. Therefore, by minimizing the sum of each voxel in the path, a path planning is performed to obtain a relatively short and suitable low-altitude flight route. This allows the low-altitude aircraft to fly away from densely populated areas and avoid obstacles as much as possible when flying from the first position to the second position, and to fly according to the instructions of the target facilities in the target area, while ensuring flight efficiency and safety. Attached Figure Description

[0061] Figure 1 A schematic diagram of the architecture of a low-altitude flight system according to an embodiment of this application is shown.

[0062] Figure 2 shows a flowchart of a method for generating low-altitude routes according to an embodiment of the present invention.

[0063] Figure 3 A logical schematic diagram of a path planning algorithm according to an embodiment of this application is shown.

[0064] Figure 4 A flowchart of a method for generating low-altitude routes according to an embodiment of the present invention is shown.

[0065] Figure 5 A schematic diagram of multi-layer three-dimensional cost construction involved in an embodiment of this application is shown.

[0066] Figure 6 A flowchart of a path planning algorithm according to an embodiment of this application is shown.

[0067] Figure 7 The diagram illustrates the architecture and data flow of a low-altitude route generation system according to an embodiment of this application.

[0068] Figure 8 This is a schematic diagram of a low-altitude flight path generation device provided in an embodiment of this application.

[0069] Figure 9 This is a schematic diagram of the structure of an electronic device provided in an optional embodiment of the present invention. Detailed Implementation

[0070] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Many specific details are set forth in the following description to provide a thorough understanding of the present invention. However, the present invention can be practiced in many other ways different from those described herein, and those skilled in the art can make similar modifications without departing from the spirit of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0071] Figure 1 A schematic diagram of the architecture of a low-altitude flight system according to an embodiment of this application is shown. Figure 1 As shown, the low-altitude flight system in the target area 100 includes a control device 101 and a low-altitude aircraft 102.

[0072] Optionally, the low-altitude aircraft 102 can be a drone; or the low-altitude aircraft 102 can be a low-altitude manned aircraft, such as a helicopter or a flying car; the low-altitude aircraft 102 in this embodiment can achieve automatic flight according to a planned target path.

[0073] In this embodiment of the application, the control device may be a base station or a terminal device, thereby controlling the low-altitude aircraft 102 and establishing a communication connection with the low-altitude aircraft 102.

[0074] Taking the low-altitude aircraft 102 as an example, the control device 101 can be a terminal device, such as a mobile terminal or a fixed terminal device with wireless communication function.

[0075] Optionally, in this embodiment, the user can manually control the drone's movement state through the control device 101, such as flying in a certain direction, ascending, descending, and landing. The user can also control the drone's destination (e.g., by controlling the control device 101 and the drone's endpoint) Figure 1 At point B in the diagram, the control device 101 can perform path planning for the drone and control the drone to fly automatically from its current location at point A along the planned path.

[0076] In another alternative implementation, the path planning of the drone can be implemented by the processor inside the drone. For example, when the user sets the destination of the drone through the control device 101, the control device 101 sends the location information of the destination to the drone through the wireless network. After receiving the location of point B, the drone obtains the location information of its current location at point A through the positioning module, and automatically performs path planning. Then, it flies from point A to point B according to the planned path.

[0077] In one optional implementation, the control device 101 is a base station, such as a 4G or 5G network base station. In this case, the user can log in to the drone control server through a network device outside the target area and generate instructions to control the corresponding drone. After the server determines the drone that the user needs to control, it can communicate with the drone through the base station in the target area to control the drone.

[0078] Optionally, the aforementioned server may be a cloud server that provides cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms, etc.

[0079] Optionally, the wireless network described above uses standard communication technologies and / or protocols. The network is typically the Internet, but can also be any other network, including but not limited to any combination of local area networks (LANs), metropolitan area networks (MANs), wide area networks (WANs), mobile, wired or wireless networks, private networks, or virtual private networks (VPNs). In some embodiments, technologies and / or formats including Hypertext Markup Language (HTML), Extensible Markup Language (XML), etc., are used to represent data exchanged over the network. Furthermore, conventional encryption technologies such as Secure Sockets Layer (SSL), Transport Layer Security (TLS), VPNs, and Internet Protocol (IP) security can be used to encrypt all or some links. In other embodiments, customized and / or dedicated data communication technologies can be used to replace or supplement the aforementioned data communication technologies.

[0080] Combination Figure 2 As shown in Figure 2, a flowchart of a low-altitude flight path generation method according to an embodiment of the present invention is presented. This method is applied to a computer device, which can be, for example, […]. Figure 1 The control device 101 or the low-altitude aircraft 102 shown. For example... Figure 2 As shown, the method flow includes:

[0081] Step 201: Obtain the three-dimensional spatial model of the target area and the population density distribution.

[0082] The three-dimensional spatial model is marked with obstacles and target facilities.

[0083] In this embodiment of the application, the three-dimensional spatial model includes building data, terrain features, road data, and no-fly zone information for the entire target area.

[0084] Optionally, in this embodiment of the application, after the computer device obtains the three-dimensional spatial model, it can obtain building data and terrain features (such as the distribution of mountains, trees, etc.) in the target area. Then, based on the building data and terrain features, it can identify the buildings, mountains, trees, and other elements in the target area that may interfere with flight as obstacles, and mark the voxels in the three-dimensional spatial model that overlap with the above obstacles.

[0085] Furthermore, in this embodiment of the application, a no-fly zone may also be set in the target area. After the computer device obtains the three-dimensional spatial model, it can mark each voxel that is projected into the no-fly zone as having obstacles and being unable to fly.

[0086] Optionally, in this embodiment of the application, the target facility may be a preset road, such as a highway, railway, or waterway.

[0087] Step 202: Determine the path cost of each voxel based on the population density at the ground projection of each voxel in the three-dimensional spatial model, the obstacle status of each voxel, and the distance between each voxel and the target facility.

[0088] In this embodiment of the application, after obtaining the three-dimensional spatial model, the path cost of each voxel can be determined based on the population density distribution, obstacle distribution, and target facility distribution in the three-dimensional spatial model. The path cost actually refers to the risk brought about by a low-altitude aircraft passing through the voxel.

[0089] For example, when the population density at the ground projection of the target voxel is high, the risk of low-altitude aircraft flying over the area is obviously high. If a low-altitude aircraft has an accident, it is easy to cause related accidents. Therefore, low-altitude aircraft should avoid passing through densely populated areas as much as possible.

[0090] Similarly, if there is an obstacle at the location of the target voxel, or if the target voxel is close to an obstacle, then the risk of low-altitude aircraft passing through the target voxel is obviously greater. Therefore, low-altitude aircraft should also avoid passing through voxels where obstacles are located or nearby.

[0091] Correspondingly, since the three-dimensional space model also marks the target facilities, and these target facilities can be pre-set roads, the distribution of highways, railways or waterways usually avoids densely populated areas and obstacles. In addition, pre-set roads such as highways or railways also have a certain degree of visual characteristics. Low-altitude aircraft can fly based on target facilities, which can not only reduce the above two types of risks as much as possible, but also improve the stability of low-altitude aircraft flying along the route.

[0092] Step 203: Using the first position in the three-dimensional spatial model as the starting point and the second position in the three-dimensional spatial model as the ending point, perform path planning with the goal of minimizing the sum of the path costs of the voxels traversed between the first and second positions to obtain the target low-altitude flight path.

[0093] In this embodiment of the application, after the path cost of each voxel in the three-dimensional spatial model is divided, path planning can be performed based on the path cost of each voxel.

[0094] Optionally, in this embodiment, a corresponding path planning algorithm can be adopted as needed. For example, one of the existing algorithms such as RRT and Dijkstra can be used to directly plan the path, obtain the possible low-altitude routes, and then select the path with the lowest sum of path costs of the voxels passed through each path as the target low-altitude route.

[0095] The following example demonstrates the use of a specific path planning algorithm for path planning:

[0096] The computer device selects a first intermediate voxel based on the path cost of each voxel, starting from a first location; and there are no obstacles in the path connecting the first intermediate voxel to the first location.

[0097] Starting from the nth intermediate voxel, select the (n+1)th intermediate voxel based on the path cost of each voxel; and there are no obstacles in the path connecting the nth intermediate voxel and the (n+1)th intermediate voxel; n is greater than or equal to 1 and n is an integer;

[0098] If the distance between the (n+1)th intermediate voxel and the second position is less than the first threshold, and there are no obstacles in the path connecting the (n+1)th intermediate voxel and the second position, then candidate low-altitude routes are obtained based on the paths sequentially connected by the first position, the (n+1)th intermediate voxels, and the second position.

[0099] Iteratively execute the above steps to obtain k candidate low-altitude routes, and select the one with the lowest sum of path costs as the target low-altitude route.

[0100] Please refer to Figure 3 This illustrates a logical diagram of a path planning algorithm according to an embodiment of this application. Figure 3 As shown, the computer device first uses the first position 301 as the starting point and adds the starting point. Then, it selects the first intermediate voxel in the three-dimensional space model in a random manner. This random process is similar to the randomness in the RRT algorithm. However, in this embodiment, the probability of each voxel being selected as an intermediate voxel is determined based on the path cost of each voxel. That is, voxels with lower path costs are more likely to be selected as intermediate voxels. Therefore, in this embodiment, the selection of intermediate voxels based on the path cost of each voxel is actually a pseudo-random process based on weighted probability.

[0101] After obtaining the first intermediate element through a pseudo-random process, the computer device will determine whether there is an obstacle in the path connecting the first intermediate element and the first position. If there is an obstacle, it means that the path formed by the connection between the first intermediate element and the first position is impassable. At this time, the first intermediate element needs to be reselected until there is no obstacle in the path connecting the first intermediate element and the first position.

[0102] Then, the above process is iteratively executed to obtain n+1 intermediate voxels. The n+1 intermediate voxels are connected in sequence to form a path, and there are no obstacles between each path. Finally, the distance between the n+1 intermediate voxel and the second position is less than the first threshold, and there are no obstacles in the path connecting the n+1 intermediate voxel and the second position. At this time, the n+1 intermediate voxel can be directly connected to the second position. Combined with the path obtained by connecting the previous n+1 intermediate voxels and the first position in sequence, a candidate low-altitude route can be obtained.

[0103] However, it is clear that in the process of constructing the above-mentioned candidate low-altitude routes, although the path cost of selecting intermediate voxels is taken into account, there is still a strong degree of randomness. For example, the path cost of the selected intermediate voxels may be low, but the path formed by the intermediate voxels may be long or may pass through densely populated areas, resulting in a high path cost for the voxels in the path. Therefore, in order to reduce the possibility of the above situation, in the embodiments of this application, the computer device can repeatedly execute the pseudo-random path generation process to obtain multiple candidate low-altitude routes. Then, for each candidate low-altitude route, its path cost (which is the sum of the path costs of all voxels passed through by the route) is calculated, and the candidate low-altitude route with the minimum path cost is determined as the target low-altitude route.

[0104] The algorithm described above can make the path search process more inclined to unfold in low-cost areas by using weighted pseudo-random sampling based on voxel path cost. Finally, by selecting the route with the minimum sum of path costs from multiple candidate paths, it can effectively avoid high-cost areas in the path, making the final path more economical, safe, and smooth.

[0105] In summary, this application first obtains the target area where the low-altitude aircraft needs to fly, and then obtains a three-dimensional spatial model and population density distribution within the target area. The three-dimensional spatial model is marked with obstacles and target facilities. At this point, the path cost of each voxel can be determined based on the population density at the ground projection of each voxel in the three-dimensional model, the obstacle status of each voxel, and the distance between each voxel and the target facility. The path cost of each voxel represents whether it is suitable for the low-altitude aircraft to fly to that voxel. Therefore, by minimizing the sum of all voxels in the path, a relatively short and suitable low-altitude flight path is obtained. This allows the low-altitude aircraft to fly away from densely populated areas and avoid obstacles as much as possible when flying from the first position to the second position, and to fly according to the instructions of the target facilities in the target area, while ensuring flight efficiency and safety.

[0106] Combination Figure 4 As shown in Figure 4, a flowchart of a low-altitude flight path generation method according to an embodiment of the present invention is presented. This method is applied to applications such as... Figure 1 The control device 101 shown can optionally be a ground-based central server or a cloud platform, integrating data processing, cost map generation, and path planning algorithm modules. Through this system, UAM operators or airspace management agencies can input take-off and landing point coordinates, and the system can automatically generate one or more optimized routes that balance safety and efficiency for UAV fleets to perform passenger or cargo transport missions. Figure 4 As shown, the method flow includes:

[0107] Step 401: Obtain the three-dimensional spatial model and population density distribution of the target area.

[0108] In this embodiment of the application, the system first collects geospatial data required to construct the map from multiple data sources; wherein, the geospatial data includes a three-dimensional spatial model, a dynamic population density layer, an infrastructure corridor layer, and an airspace restriction layer;

[0109] The three-dimensional spatial model can be derived from commercial GIS platforms or LiDAR scans of urban three-dimensional buildings and terrain data, used to define physical obstacles. The dynamic population density layer combines published static census data with anonymized, aggregated mobile network signaling data provided by network operators to create a dynamic population distribution map reflecting real-time population movement (such as during peak commuting hours), thus characterizing the population density distribution within the target area. The infrastructure corridor layer can be derived from vector data of linear infrastructure such as road and railway networks provided by map service providers. The airspace restriction layer can be derived from no-fly zone data published by aviation regulatory agencies.

[0110] Step 402: Based on the population density of each voxel at the ground projection, the obstacle status of each voxel, and the distance between each voxel and the target facility in the three-dimensional spatial model, determine the path cost of each voxel.

[0111] In this embodiment, the population avoidance cost of each voxel is obtained based on the population density at the ground projection of each voxel; the obstacle cost of each voxel is obtained based on the positional relationship between each voxel and the obstacle; the facility affinity cost of each voxel is obtained based on the positional relationship between each voxel and the target facility; and the path cost of each voxel is obtained based on the population avoidance cost, the obstacle cost, and the facility affinity cost.

[0112] Furthermore, the population avoidance cost for each voxel is positively correlated with the population density at the ground projection of that voxel. In other words, for each voxel, the higher the population density of the ground area corresponding to that voxel, the greater the path cost for that voxel, thus guiding low-altitude flight paths to avoid high-density population areas as much as possible.

[0113] Furthermore, the target facility is a preset road in the target area; if the projection of the target voxel on the ground overlaps with the preset road, the facility affinity cost of the target voxel is set to a first preset value; if the projection of the target voxel on the ground does not overlap with the preset road, a second preset value is determined based on a first distance between the target voxel and the preset road, and the second preset value is determined as the facility affinity cost of the target voxel; the second preset value is greater than the first preset value, and the facility affinity cost of the target voxel is positively correlated with the first distance.

[0114] In other words, this application embodiment introduces facility affinity cost to make the path more inclined to be close to the ground road. For example, when a drone flies at low altitude along the road, if the voxel's projection on the ground overlaps with the road, a first preset value is set, indicating that the voxel is close to the road, which is safer or in line with the plan. If the voxel's projection on the ground does not overlap with the road, a larger second preset value is set according to the first distance between the voxel and the road. The second preset value is positively correlated with the distance, which means that the farther the voxel is from the road, the greater the cost.

[0115] Optionally, in this embodiment of the application, the projection of the voxel on the ground may be close to multiple roads, or close to multiple segments of a single road. Therefore, when calculating the second preset value, the closest distance between the voxel and each road can be used as the first distance between the voxel and the road.

[0116] Furthermore, if the target voxel overlaps with the obstacle, the obstacle cost of the target voxel is set to a third preset value; if the target voxel does not overlap with the obstacle, a fourth preset value is determined based on the second distance between the target voxel and the obstacle, and the fourth preset value is determined as the obstacle cost of the target voxel; the fourth preset value is less than the third preset value, and the obstacle cost of the target voxel is inversely correlated with the second distance.

[0117] In this embodiment of the application, if a voxel overlaps with an obstacle, the obstacle cost is set to a third preset value (optionally, the third preset value can be set to an extremely high value or even positive infinity), thereby introducing an obstacle cost to avoid the path crossing a building or no-fly zone.

[0118] Furthermore, if the voxel is too close to the obstacle, it may cause great risks to the flight path. Therefore, in this embodiment, if the voxel does not overlap with the obstacle, a fourth preset value can be determined based on the second distance between the voxel and the obstacle. The fourth preset value is lower than the third preset value, but the fourth preset value is inversely related to the second distance. That is to say, the closer the voxel is to the obstacle, the higher the cost, so that the path generation process can stay away from the obstacle as much as possible and reduce the risk of collision.

[0119] Figure 5 A schematic diagram of multi-layer three-dimensional cost construction involved in an embodiment of this application is shown. For example... Figure 5 As shown, in one optional implementation, the comprehensive cost value is calculated as follows:

[0120] This application embodiment can fuse the collected data into a three-dimensional voxel mesh of a three-dimensional spatial model, and calculate a comprehensive cost value for each voxel. (That is, path cost).

[0121] In order to calculate the overall cost value, the embodiments of this application need to consider the impact of obstacles, population density and infrastructure on each voxel at the same time. That is, for each voxel, it is necessary to calculate the population avoidance cost, facility affinity cost and obstacle cost.

[0122] Among them, population avoidance costs ( The definition of ) is as follows: mapping dynamic population density data onto the ground, voxels The value is positively correlated with the population density of the area directly below it, and an exponential function can be used to amplify the penalty for high-density areas.

[0123] Facility affinity cost ( The definition of ) is as follows: An "attraction field" is generated around linear features such as roads and railways. (Volumetric data) The value is calculated as a function of its distance to the nearest infrastructure; the closer the distance, the lower the cost (the more attractive it is).

[0124] Obstacle costs ( The definition of ) is as follows: For voxels that overlap with buildings, terrain, or no-fly zones, their cost Set it to infinity, making it impassable.

[0125] Comprehensive cost function: Finally, the total cost at any point x in space is calculated using the weighted sum formula:

[0126]

[0127] in, and These are adjustable weights, representing the degree of avoidance of population risks and the degree of affinity for infrastructure, respectively. It is a basic cost item related to distance. As a hard constraint (not reflected in the formula, but rather the cost is set to infinity when the voxel is in an obstacle area or no-fly zone), the above scheme sets population hotspots as high-cost areas, infrastructure areas as low-cost areas, and obstacles and no-fly zones as infinite-cost areas.

[0128] Step 403: Add the first position as the first intermediate voxel to the voxel set, and iteratively update the voxel set.

[0129] Optionally, in this embodiment of the application, the iterative update of the voxel set includes:

[0130] Candidate voxels are selected based on the path cost of each voxel in the 3D spatial model;

[0131] Add the candidate voxel to the voxel set;

[0132] Select the parent node of the candidate voxel from among the intermediate voxels, and establish a candidate path between the parent node of the candidate voxel and the candidate voxel.

[0133] In the embodiments of this application, each intermediate voxel in the voxel set can be considered as a node obtained by the path planning algorithm, that is, a node that the UAV may pass through. The path formed by connecting each node with its parent node is the flight path that the UAV may take.

[0134] Therefore, in the path planning process of this application embodiment, the first thing to determine is each intermediate voxel in the voxel set, and the parent voxel corresponding to the intermediate voxel (that is, the connection path between the intermediate voxel and other voxels).

[0135] Since the first position is the starting point of the path planning, it must be a node of the path tree obtained by the path planning. At this time, it can be added to the voxel set as the first intermediate voxel, and then the voxel set can be iterated based on the first intermediate voxel.

[0136] In one alternative implementation, a first selection probability of each voxel in the three-dimensional spatial model is obtained based on the path cost of each voxel in the three-dimensional spatial model; the first selection probability is inversely correlated with the path cost; and the candidate voxel is selected based on the selection probability of each voxel in the three-dimensional spatial model.

[0137] In this embodiment, after obtaining the first intermediate voxel, other voxels can be searched in the three-dimensional space model as child nodes of the first intermediate voxel to find the subsequent flight path of the UAV. Since the path cost of each voxel has been calculated in the above steps, and the path cost can be used to indicate whether the voxel is suitable for inclusion in the flight path, in this embodiment, candidate voxels can be selected directly based on the path cost of each voxel. For example, the selection probability of each voxel can be calculated based on the path cost of each voxel, and then candidate voxels can be randomly selected in the three-dimensional space model based on the selection probability of each voxel.

[0138] In another optional implementation, a second selection probability of each voxel in the three-dimensional spatial model is obtained based on the path cost of each voxel in the three-dimensional spatial model; the second selection probability is inversely correlated with the path cost; a candidate voxel is randomly selected in the three-dimensional spatial model, and it is determined whether the candidate voxel is accepted based on the second selection probability of the candidate voxel; if the candidate voxel is not accepted, a candidate voxel is randomly selected again until the randomly selected candidate voxel is accepted.

[0139] In this embodiment, candidate voxels can be directly determined from each voxel in the three-dimensional spatial model using a truly random method. Then, based on the second selection probability of the candidate voxels in the three-dimensional spatial model, it is determined whether the candidate voxel is accepted. If the second selection probability of the candidate voxel is high, it means that the path cost of the candidate voxel is low, and the candidate voxel is more suitable as a node in the flight path. If the second selection probability of the candidate voxel is low, it means that the path cost of the candidate voxel is low, and the candidate voxel is not very suitable as a node in the flight path. In this case, if the candidate voxel is rejected based on the second selection probability, other candidate voxels can be randomly selected again.

[0140] Optionally, in this embodiment, the path cost of each voxel can be normalized, and a first selection probability or a second selection probability of each voxel can be generated based on the normalized value of each voxel (for example, the normalized value can be directly used as the first selection probability or the second selection probability).

[0141] Once a candidate voxel is obtained, it can be formally added to the voxel set, which is equivalent to adding a new node in the path tree. Then, the parent node of the candidate voxel is selected from the existing intermediate voxels in the voxel set (for example, the one closest to the candidate voxel or the one with the lowest path cost between it and the candidate voxel) as the intermediate connection point between the candidate voxel and the first position, and a candidate path is established between the parent node of the candidate voxel and the candidate voxel.

[0142] Iterating through the above process gradually expands the paths that the voxel set can form, thus gradually connecting to the second position.

[0143] Optionally, in this embodiment, the parent node can be selected based on path cost. Specifically, the control device can obtain the total path cost corresponding to the candidate voxel and each intermediate voxel; the total path cost is the sum of the path costs in the path formed by the candidate voxel connecting to the first position through the intermediate voxel; and select the intermediate voxel with the smallest total path cost as the parent node of the candidate voxel.

[0144] Furthermore, if the candidate voxel is added to the voxel set, the neighboring voxels within a specified range of the candidate voxel are obtained; if the total cost of the path formed by the neighboring voxel connecting to the first position through the candidate voxel is the minimum after the neighboring voxel is connected to the candidate voxel, then the parent node of the neighboring voxel is switched to the candidate voxel.

[0145] When adding a new candidate voxel (equivalent to adding a new node) through the above steps, a "parent node" is selected to connect it. At this time, the path cost between the candidate voxel (new node) and the parent node is locally optimal, but the entire path tree may not be globally optimal.

[0146] Therefore, in this embodiment, if a candidate voxel has been determined to be added to the voxel set, all nodes that are close to the candidate voxel (i.e., intermediate voxels within the specified range of the candidate voxel) can be found in the set as potential optimizable objects. At this time, the control device needs to determine whether the path cost is lower for each neighboring voxel if it is tried to bypass the original parent node and instead connect to the starting point (i.e., the first position) through the candidate voxel. If it is lower, it means that it is better for the neighboring voxel to have the candidate voxel as its parent node, and the correspondence between voxels and parent nodes in the voxel set can be updated according to the better connection method.

[0147] In short, the logical framework of this path planning algorithm can be found by referring to... Figure 6 , Figure 6 A flowchart of a path planning algorithm according to an embodiment of this application is shown. Figure 6As shown, when performing path planning, the starting point and the ending point need to be set first. For example, the starting point can be the current location of the drone, and the ending point can be the three-dimensional spatial coordinates manually input by the user. Then, the control device can initialize the search tree according to the starting point and the ending point, such as building a three-dimensional spatial model and calculating the path cost of each voxel in it. Then, the control device can perform cost-based bias sampling, that is, obtain candidate voxels in the above way, and then obtain the optimal parent node corresponding to the candidate voxel in the nearest neighbor node through the cost metric based on path integral, and update the candidate node and its relationship with the optimal parent node in the path tree (i.e., the set of voxels mentioned above).

[0148] In addition, the control device also needs to select the neighboring nodes of the candidate voxels, use cost integration to optimize the correspondence between the neighboring nodes and the parent node, thereby performing a rewiring operation and updating the new node correspondence in the path tree.

[0149] If there is no voxel in the path tree that is close to the endpoint (e.g., the distance is less than the threshold), then iterate the above operation, re-execute cost-based bias sampling, and obtain new candidate voxels; if there is, then end the iteration process and directly output the optimal path.

[0150] Step 404: If the distance between a second intermediate voxel in the voxel set and the second position is less than a distance threshold, then stop iteratively updating the voxel set.

[0151] The above scheme iteratively adds selected candidate voxels to the voxel set, continuously adding nodes to the path tree and improving the path tree's exploration of the entire 3D spatial model. When the distance between a second intermediate voxel and the second position (i.e., the endpoint) in the voxel set is less than a distance threshold, it can be considered that the path tree has explored the vicinity of the endpoint. At this point, simply connecting the endpoint to the nearest second intermediate voxel completes the path tree's exploration process.

[0152] Furthermore, in the above scheme, each time a candidate voxel is added to the voxel set, the existing structure can be locally optimized. That is, the cost of the final generated path tree gradually converges to a solution close to the optimal solution without generating a large amount of computation, while ensuring computational efficiency and path planning effect.

[0153] Step 405: Obtain the target low-altitude flight path based on the candidate paths stored in the voxel set and the path connecting the second intermediate voxel to the second position.

[0154] The set of voxels includes candidate paths formed by connecting each intermediate voxel with its parent node. In other words, the set of voxels has already stored the connection relationship between each voxel and its parent node, forming a complete path tree. The second intermediate voxel is the voxel closest to the target endpoint (second position) and can be directly connected to the endpoint.

[0155] Therefore, from the first intermediate voxel to each of the progressively selected candidate voxels, the set of voxels retains the candidate paths connecting them. Thus, the control device can synthesize a complete, safe, and low-cost flight path from the candidate paths in the voxel set, and finally connect it to the second position to obtain the target low-altitude flight path.

[0156] Furthermore, in this embodiment, several target low-altitude routes can be pre-planned within the target area as needed, so that different low-altitude aircraft can directly fly according to the pre-planned target low-altitude routes when they have flight needs. Taking a flying car as an example, there may be multiple preset take-off and landing points in the target area. Therefore, the control equipment can pre-plan multiple target low-altitude routes according to the above steps. At this time, when the low-altitude aircraft has flight needs, real-time planning is not required, which significantly reduces the preparation time before flight navigation.

[0157] However, since obstacle information, population density and other data change in real time, the pre-planned flight path must be updated in real time to ensure the feasibility and safety of the flight path.

[0158] In this embodiment of the application, the control device first needs to update the dynamic environmental parameters of the target area according to a specified period. When the updated dynamic environmental parameters meet the preset conditions, the path cost of each voxel in the three-dimensional spatial model is updated according to the updated dynamic environmental parameters, so as to re-plan the path of the target low-altitude route according to the updated path cost of each voxel.

[0159] The aforementioned dynamic environmental parameters can include real-time population density, real-time no-fly zone directives, and moving obstacles (such as cranes or other obstacles that may affect low-altitude flight). Changes in these dynamic environmental parameters may cause the planned target low-altitude route to pass through high-population-density areas, leading to safety hazards, or pass through no-fly zones or obstacles, resulting in complete inaccessibility. Therefore, in this embodiment, the dynamic environmental parameters of the target area need to be updated at a specified period (e.g., the specified period is set to 10 seconds, 30 seconds, 1 minute, etc.). After the dynamic environmental parameters are updated, the path cost can be updated based on the updated dynamic environmental parameters, and the path can be replanned, ensuring the safety and passability of the target low-altitude route.

[0160] Furthermore, in this embodiment of the application, the preset conditions that need to be met for re-planning the route may include at least one of the following:

[0161] The update magnitude of at least one type of data in the dynamic environment parameters is greater than the magnitude threshold;

[0162] Since the last path planning, the number of times the dynamic environment parameters have been updated has exceeded the threshold.

[0163] In simple terms, not every update of dynamic environmental parameters will require a re-execution of path planning. For example, if the dynamic environmental parameters are updated very frequently, such as once every 10 seconds, then re-executing path planning every time the dynamic environmental parameters are updated will place an extremely high computational load on the control equipment. Furthermore, if the update magnitude of the dynamic environmental parameters is very small, such as only a slight change in population density distribution, then obviously there is no need to re-execute path planning.

[0164] Therefore, in this embodiment of the application, the magnitude or number of updates to the dynamic environmental parameters can be considered to determine whether path planning needs to be re-performed.

[0165] On the one hand, an amplitude threshold can be set, for example, greater than 10%. For example, for population density, the control device can obtain the extreme points of population density in the 3D spatial model and determine whether the change in population density at the extreme point since the last path planning exceeds the amplitude threshold to determine whether the population density update exceeds the update amplitude. For no-fly zones or obstacles, it can be determined whether the non-overlapping area between the voxel of the current obstacle and the voxel of the obstacle at the time of the last path planning exceeds 10%. If it does, it means that the obstacle has changed significantly and exceeded the update amplitude.

[0166] On the other hand, a threshold number of times can be set to indicate the update time. That is, even if the change of dynamic environmental parameters is small each time, but the dynamic environmental parameters have been updated a preset number of times (e.g., 10 times), and the path planning has not been re-performed, the control device can force the path planning to be re-performed, thereby improving the timeliness of the target low-altitude route.

[0167] Furthermore, in this embodiment, if there are many target low-altitude paths in the target area, even if the path planning is re-performed under the condition of meeting the preset conditions, it will still generate a huge computational load on the control equipment. Therefore, in this embodiment, after obtaining the updated voxel area by comparing the path cost of each voxel after the update with the path cost of each voxel before the update, the target low-altitude airways that match the updated voxel area can be selected from the target low-altitude airways in the target area to re-perform the path planning.

[0168] In other words, after the voxel region is updated, not all target low-altitude routes need to be updated. For example, if the voxel region is divided into four regions, A, B, C, and D, and the path cost of voxels in regions A and B has changed, while the voxels in regions B and D have not changed, then the path cost of target low-altitude routes that are completely surrounded by region D will not actually change. Therefore, they can be excluded, thereby reducing the load on the computer.

[0169] Therefore, in this embodiment, the updated voxel region enclosed by the voxels of the path cost update can be divided first (for example, each updated voxel is surrounded by the smallest fully connected region to obtain the updated voxel region); then it can be determined which of the target low-altitude routes intersect with the updated voxel region. If there is an intersection, it can be considered to match the updated voxel region and needs to be replanned.

[0170] Figure 7 This illustration shows the architecture and data flow diagram of a low-altitude flight path generation system according to an embodiment of this application. Figure 7 As shown in the embodiments of this application, the low-altitude flight path generation system first needs to acquire three-dimensional environmental building data, dynamic population data, infrastructure (such as road and waterway data), and airspace restriction no-fly zone data. After collecting the above multi-source data, the data is preprocessed to construct a multi-layer three-dimensional cost map (that is, according to...). Figure 4 The corresponding approach involves fusing the path cost of each voxel in the 3D spatial model with the voxel itself; then, a path search algorithm is executed based on the constructed 3D cost map to obtain the target low-altitude flight path. Finally, post-processing and smoothing operations are performed on the target low-altitude flight path to obtain the final flight path output, thereby instructing the UAV on low-altitude flight.

[0171] In summary, this application first obtains the target area where the low-altitude aircraft needs to fly, and then obtains a three-dimensional spatial model and population density distribution within the target area. The three-dimensional spatial model is marked with obstacles and target facilities. At this point, the path cost of each voxel can be determined based on the population density at the ground projection of each voxel in the three-dimensional model, the obstacle status of each voxel, and the distance between each voxel and the target facility. The path cost of each voxel represents whether it is suitable for the low-altitude aircraft to fly to that voxel. Therefore, by minimizing the sum of all voxels in the path, a relatively short and suitable low-altitude flight path is obtained. This allows the low-altitude aircraft to fly away from densely populated areas and avoid obstacles as much as possible when flying from the first position to the second position, and to fly according to the instructions of the target facilities in the target area, while ensuring flight efficiency and safety.

[0172] This application also provides a low-altitude flight path generation apparatus for implementing the above embodiments and preferred embodiments, which will not be repeated hereafter. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the apparatus described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0173] This application provides a low-altitude flight path generation device. Figure 8 This is a schematic diagram of a low-altitude flight path generation device provided in an embodiment of this application. The device includes:

[0174] The parameter acquisition module 801 is used to acquire a three-dimensional spatial model of the target area and the population density distribution; the three-dimensional spatial model is marked with obstacles and target facilities.

[0175] The cost calculation module 802 is used to determine the path cost of each voxel based on the population density of each voxel at the ground projection of each voxel in the three-dimensional spatial model, the obstacle status of each voxel, and the distance between each voxel and the target facility.

[0176] The path planning module 803 is used to perform path planning with the first position in the three-dimensional spatial model as the starting point, the second position in the three-dimensional spatial model as the ending point, and the goal of minimizing the sum of the path costs of the voxels traversed between the first position and the second position, so as to obtain the target low-altitude route.

[0177] In summary, this application first obtains the target area where the low-altitude aircraft needs to fly, and then obtains a three-dimensional spatial model and population density distribution within the target area. The three-dimensional spatial model is marked with obstacles and target facilities. At this point, the path cost of each voxel can be determined based on the population density at the ground projection of each voxel in the three-dimensional model, the obstacle status of each voxel, and the distance between each voxel and the target facility. The path cost of each voxel represents whether it is suitable for the low-altitude aircraft to fly to that voxel. Therefore, by minimizing the sum of all voxels in the path, a relatively short and suitable low-altitude flight path is obtained. This allows the low-altitude aircraft to fly away from densely populated areas and avoid obstacles as much as possible when flying from the first position to the second position, and to fly according to the instructions of the target facilities in the target area, while ensuring flight efficiency and safety.

[0178] Please see Figure 9 , Figure 9 This is a schematic diagram of the structure of an electronic device provided in an optional embodiment of the present invention. This electronic device can be a computer device used to implement the above-described apparatus. Figure 9As shown, the electronic device includes one or more processors 10, a memory 20, and interfaces for connecting the various components, including high-speed interfaces and low-speed interfaces. The various components communicate with each other via different buses and can be mounted on a common motherboard or otherwise installed as needed. The processor can process instructions executed within the electronic device, including instructions stored in or on memory to display graphical information of a GUI on external input / output devices (such as display devices coupled to the interfaces).

[0179] The processor 10 may further include a hardware chip. This hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The programmable logic device may be a complex programmable logic device (CAMP), a field-programmable gate array (FPGA), a general-purpose array logic (GDA), or any combination thereof.

[0180] The memory 20 stores instructions executable by at least one processor 10 to cause the at least one processor 10 to perform the method shown in the above embodiments.

[0181] The memory 20 may include a program storage area and a data storage area. The program storage area may store the operating system and application programs required for at least one function; the data storage area may store data used by the electronic device. Furthermore, the memory 20 may include high-speed random access memory (RAM), and may also include non-transient memory, such as at least one disk storage device, flash memory, or other non-transient solid-state storage device. The memory 20 may include volatile memory, such as RAM; the memory may also include non-volatile memory, such as flash memory, hard disk, or solid-state drive; the memory 20 may also include combinations of the above types of memory.

[0182] The electronic device also includes a communication interface 30 for communicating with other devices or communication networks.

[0183] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods shown in the above embodiments.

[0184] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0185] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A low altitude air route generation method characterized by comprising: The method comprises: acquiring a three-dimensional space model of a target area and a population density distribution; the three-dimensional space model is marked with obstacles and target facilities; the target facilities are preset roads in the target area; determining a path cost of each voxel based on a population density at a ground projection of each voxel in the three-dimensional space model, an obstacle state of each voxel, and a distance of each voxel to the target facilities; performing path planning with a first position in the three-dimensional space model as a starting point, a second position in the three-dimensional space model as an end point, and a sum of path costs of voxels passing between the first position and the second position being the lowest as a target to acquire a target low-altitude route; the determining of the path cost of each voxel based on the population density at the ground projection of each voxel in the three-dimensional space model, the obstacle state of each voxel, and the distance of each voxel to the target facilities comprises: acquiring a population avoidance cost of each voxel based on the population density at the ground projection of each voxel; acquiring an obstacle cost of each voxel based on a positional relationship between each voxel and the obstacles; if a projection of a target voxel on the ground overlaps with the preset roads, setting a facility affinity cost of the target voxel as a first preset value; if the projection of the target voxel on the ground does not overlap with the preset roads, determining a second preset value according to a first distance between the target voxel and the preset roads, and setting the second preset value as the facility affinity cost of the target voxel; the second preset value is greater than the first preset value, and the facility affinity cost of the target voxel is positively correlated with the first distance; acquiring the path cost of each voxel based on the population avoidance cost of each voxel, the obstacle cost of each voxel, and the facility affinity cost of each voxel.

2. The method of claim 1, wherein, The population avoidance cost of each voxel is positively correlated with the population density at the ground projection of each voxel.

3. The method of claim 1, wherein, the acquiring of the obstacle cost of each voxel based on the first distance between each voxel and the obstacles comprises: if a target voxel overlaps with the obstacles, setting an obstacle cost of the target voxel as a third preset value; if the target voxel does not overlap with the obstacles, determining a fourth preset value according to a second distance between the target voxel and the obstacles, and setting the fourth preset value as the obstacle cost of the target voxel; the fourth preset value is less than the third preset value, and the obstacle cost of the target voxel is inversely correlated with the second distance.

4. The method according to any one of claims 1 to 3, characterized in that, the acquiring of the target low-altitude route by performing path planning with the first position in the three-dimensional space model as the starting point, the second position in the three-dimensional space model as the end point, and the sum of the path costs of the voxels passing between the first position and the second position being the lowest as the target comprises: adding the first position as a first intermediate voxel to a voxel set, and iteratively updating the voxel set; stopping the iteratively updating of the voxel set when a second intermediate voxel in the voxel set has a distance to the second position less than a distance threshold. According to the candidate path stored in the voxel set and the path connected with the second intermediate voxel and the second position, a target low-altitude air route is obtained; the candidate path includes each intermediate voxel connected with a parent node in the voxel set; The iterative updating of the voxel set comprises: selecting a candidate voxel according to the path cost of each voxel in the three-dimensional space model; adding the candidate voxel to the voxel set; selecting a parent node of the candidate voxel from the intermediate voxels and establishing a candidate path between the parent node of the candidate voxel and the candidate voxel.

5. The method of claim 4, wherein, The selection of the parent node of the candidate voxel from the intermediate voxels comprises: obtaining a total path cost corresponding to the candidate voxel and each intermediate voxel; the total path cost is the sum of the path costs in the path formed by the connection of the candidate voxel to the first position through the intermediate voxels; selecting the intermediate voxel with the minimum total path cost as the parent node of the candidate voxel.

6. The method of claim 5, wherein, The iterative updating of the voxel set further comprises: if the candidate voxel is added to the voxel set, obtaining a neighboring voxel within a specified range of the candidate voxel; if the neighboring voxel has the minimum total path cost in the path formed by the connection of the neighboring voxel to the first position through the candidate voxel after being connected with the candidate voxel, switching the parent node of the neighboring voxel to the candidate voxel.

7. The method of claim 4, wherein, The selection of the candidate voxel according to the path cost of each voxel in the three-dimensional space model comprises: obtaining a first selection probability of each voxel in the three-dimensional space model according to the path cost of each voxel in the three-dimensional space model; the first selection probability is inversely related to the path cost; selecting the candidate voxel according to the selection probability of each voxel in the three-dimensional space model.

8. The method of claim 4, wherein, The selection of the candidate voxel according to the path cost of each voxel in the three-dimensional space model comprises: obtaining a second selection probability of each voxel in the three-dimensional space model according to the path cost of each voxel in the three-dimensional space model; the second selection probability is inversely related to the path cost; randomly selecting a candidate voxel in the three-dimensional space model and determining whether the candidate voxel is accepted according to the second selection probability of the candidate voxel; if the candidate voxel is not accepted, randomly selecting a candidate voxel again until the randomly selected candidate voxel is accepted.

9. The method according to any one of claims 1 to 3, characterized in that, The method further comprises: updating the dynamic environment parameter of the target area according to a specified period; when the updated dynamic environment parameter meets a preset condition, updating the path cost of each voxel in the three-dimensional space model according to the updated dynamic environment parameter and re-planning the target low-altitude air route according to the updated path cost of each voxel.

10. The method of claim 9, wherein, The preset condition comprises at least one of the following: the update amplitude of at least one type of data in the dynamic environment parameter is greater than an amplitude threshold; the number of updates of the dynamic environment parameter after the last path planning is greater than a number threshold.

11. The method of claim 10, wherein, The re-planning of the target low-altitude air route according to the updated path cost of each voxel comprises According to a comparison between the updated path cost of each voxel and the path cost of each voxel before the update, an updated voxel region is obtained; In each target low-altitude air route of the target region, a target low-altitude air route matched with the updated voxel region is selected to perform path planning again.

12. A low altitude air route generation device characterized by comprising: The device comprises: a parameter acquisition module configured to acquire a three-dimensional space model of a target region and a population density distribution, wherein the three-dimensional space model is marked with obstacles and target facilities, and the target facilities are preset roads in the target region; a cost calculation module configured to determine a path cost of each voxel based on a population density at a ground projection of each voxel, an obstacle state of each voxel, and a distance between each voxel and the target facilities; a path planning module configured to perform path planning with a first position in the three-dimensional space model as a starting point, a second position in the three-dimensional space model as an ending point, and a sum of path costs of voxels passed between the first position and the second position being the lowest as a target, to obtain a target low-altitude air route; the cost calculation module is further configured to: acquire a population avoidance cost of each voxel based on the population density at the ground projection of each voxel; acquire an obstacle cost of each voxel based on a positional relationship between each voxel and the obstacles; if a projection of a target voxel on the ground overlaps with the preset roads, set a facility affinity cost of the target voxel as a first preset value; if the projection of the target voxel on the ground does not overlap with the preset roads, determine a second preset value according to a first distance between the target voxel and the preset roads, and determine the second preset value as the facility affinity cost of the target voxel; the second preset value is greater than the first preset value, and the facility affinity cost of the target voxel is positively correlated with the first distance; acquire the path cost of each voxel based on the population avoidance cost of each voxel, the obstacle cost of each voxel, and the facility affinity cost of each voxel.

13. An electronic device, comprising: comprise: a memory and a processor, which are communicatively connected, the memory stores computer instructions, and the processor executes the computer instructions to perform the low-altitude air route generation method in any one of claims 1 to 11.

14. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer instructions, and the computer instructions are used to make a computer execute the low-altitude air route generation method in any one of claims 1 to 11.

Images