A collision risk prediction system and method for a drone

Abstract

The present application relates to a kind of unmanned aerial vehicle collision risk prediction system and method, system includes processor, processor is configured as: based on the contour range of unmanned aerial vehicle constructs regular first geometric body;Based on the three-dimensional model of the surrounding building of the planned route of unmanned aerial vehicle constructs regular second geometric body;First geometric body is overlaid along planned route with second geometric body and is tested to filter second geometric body with superposition condition;According to the motion trend of unmanned aerial vehicle, first geometric body is divided into several first sub-geometric body, the superposition range of first sub-geometric body and second geometric body is calculated, at least one superposition point in superposition range is in the three-dimensional model of building or edge condition, judge that unmanned aerial vehicle and building exist collision risk.For the defect that real-time collision risk prediction cannot be carried out to unmanned aerial vehicle carrying material, the present application significantly reduces the amount of calculation, so that unmanned aerial vehicle can be real-time feedback and the collision risk condition of building.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of unmanned aerial vehicle risk judgment, and particularly relates to a collision risk prediction system and method for unmanned aerial vehicles. BACKGROUND

[0002] Unmanned aerial vehicle groups are widely used in industrial plant patrolling. Especially in the initial stage of a fire, the advantage of the rapid action of unmanned aerial vehicles makes the fire more likely to be quickly eliminated or effectively delays the spread speed and range of the fire. In the process of extinguishing the fire, a large number of unmanned aerial vehicles need to transport or deliver fire-fighting materials. The ways of unmanned aerial vehicles carrying materials include two kinds: the first kind is to carry liquid materials through their own storage devices, such as carrying fire extinguishing liquid, fire extinguishing bombs, window breaking bombs, etc. The second kind is to hoist large solid materials through the grabbing assembly, such as carrying sand, sandbags, and sandy soil, etc. The unmanned aerial vehicles carrying fire-fighting materials change their flight width and weight. In the case of carrying fire-fighting materials, in order to quickly reach the fire point, the unmanned aerial vehicles need to rejudge whether the original flight route is available and re-plan the flight route according to the current flight width in the case of changing the flight width.

[0003] In the prior art, since the algorithm for calculating the collision risk in real time according to the flight width of the unmanned aerial vehicle is complex, has a large amount of calculation, and takes more time, it is unable to provide the unmanned aerial vehicle with collision risk prediction information in real time, so the flight route of the unmanned aerial vehicle is generally planned to be wider to make the unmanned aerial vehicle fly safely whether it is flying alone or carrying materials. However, the defects of such a setting include that in an emergency, the unmanned aerial vehicle needs to take a long detour to reach the fire point; in the case of heavy load, the flight height is reduced, the spatial distribution of the same height in the industrial plant changes, the flight inertia of the unmanned aerial vehicle is large and it is not easy to avoid obstacles in time.

[0004] At present, the patent application with the publication number CN113867391A discloses a method for unmanned aerial vehicle low-altitude safety early warning and monitoring based on digital twinning, which collects and pre-processes relevant data related to the operation of the unmanned aerial vehicle, including unmanned aerial vehicle operation trajectory and basic data of operation performance, unmanned aerial vehicle operation geographic environment data, unmanned aerial vehicle operation weather environment live and spatial position information data, and unmanned aerial vehicle operation restricted area data, etc.; combines a pre-trained neural network model to evaluate the low-altitude safety risk of the unmanned aerial vehicle, thereby realizing integrated monitoring of the unmanned aerial vehicle operation, completing the early warning of the operation risk, and timely taking evasive measures. However, the data calculation amount of the patent in operation is huge, which cannot be calculated by the chip carried by the unmanned aerial vehicle, nor can it be calculated in real time by the processor, so the patent is difficult to apply to real-time risk prediction in emergency situations, especially cannot meet the requirement for calculation speed.

[0005] It is a current demand for unmanned aerial vehicles to provide a collision risk prediction method that significantly reduces the amount of calculation, enabling the unmanned aerial vehicle to predict collision risks in real time based on its own hardware and change the planned route.

[0006] In addition, on the one hand, there are differences in the understanding of those skilled in the art; on the other hand, a large number of literatures and patents have been studied by the applicant when making the invention, but due to the limitations of space, all details and contents have not been listed in detail. However, this does not mean that the present invention does not have these prior art characteristics. On the contrary, the present invention already has all the characteristics of the prior art, and the applicant reserves the right to add relevant prior art in the background art. SUMMARY

[0007] When the unmanned aerial vehicle is used to carry objects, especially in emergency situations, the preset planned route of the unmanned aerial vehicle may change in height, width, etc. due to the influence of the objects, which may cause the unmanned aerial vehicle carrying the objects to collide with buildings when flying according to the preset planned route, posing a collision risk. The current system for planning the route of the unmanned aerial vehicle tends to directly plan a safer route that is farther away in order to avoid this situation. In an emergency, the unmanned aerial vehicle may delay arrival at the destination by taking a longer route, which is not conducive to the emergency transportation of goods.

[0008] The prior art has proposed a technical solution for collision detection based on geometric modeling to analyze the collision between dynamic models and static models. For example, the patent document with publication number CN116680858A discloses a three-dimensional collision detection and early warning method based on digital twinning, which includes obtaining a model to be detected and dividing it into static and dynamic models, classifying the dynamic models to distinguish the affiliation of each dynamic model; creating a collision space and dividing the collision space using the classic bisection method, then creating an envelope box and selecting an appropriate envelope box based on the shape of the model to be detected to envelope the model to be detected, so that the model to be detected is completely wrapped by the envelope box; using digital twinning technology to simulate the collision detection environment in real time and accurately, and finally performing collision detection and early warning. The envelope boxes in the divided collision space are traversed, and if there is a collision between the envelope boxes in the space, the collision early warning condition is met, triggering the collision early warning. However, in this technical solution, a small cube is used to contain all the objects that may collide in the scene, and the collision space at this time includes all potential collision elements in a general way. On the one hand, this significantly increases the amount of data processing required for collision calculation, which is contrary to the demand for significantly reducing the amount of calculation as implemented by the present invention. On the other hand, due to the existence of many idle areas between the collision space in the form of a cube and the actual collision elements, the positioning accuracy of the actual collision analysis is reduced, so it is not possible to quickly provide a safe planned route in an emergency.

[0009] In view of the deficiencies of the prior art, the present application discloses a collision risk prediction system for a UAV from one aspect, which comprises a processor configured to: construct a first geometric body based on the contour range of the UAV; construct a second geometric body based on the three-dimensional model of the surrounding buildings of the planned route of the UAV; superimpose test the first geometric body along the planned route with the second geometric body to screen the second geometric body with overlapping conditions; divide the first geometric body into a plurality of first sub-geometric bodies according to the movement trend of the UAV, calculate the overlapping range of the first sub-geometric body and the second geometric body, and determine that the UAV has a collision risk with the building if at least one overlapping point in the overlapping range is within the three-dimensional model of the building or the edge. Compared with the prior art, the present application can construct a second geometric body based on the three-dimensional model of the surrounding buildings of the planned route of the UAV, and use it as the collision analysis object between the first geometric body constructed based on the contour range of the UAV. Based on the above technical features, the problem to be solved by the present application can include: how to reduce the data processing amount of the UAV in the collision analysis process of the planned route. Specifically, due to the different spaces occupied by the buildings along the planned route, the existing collision detection method based on geometric modeling needs to perform different collision analysis processes for all buildings involved in the planned route, which leads to insufficient real-time performance of the collision analysis during the flight of the UAV, and the time-consuming is too long due to the large amount of data operation, which cannot guarantee the safety of the UAV during flight, especially in the case of changing flight state. The present application can calculate the collision risk of the planned route based on the size difference of the goods transported by the UAV in real time and adjust the planned route, so that the UAV can transport goods based on the nearest route and fly safely. Since the present application preliminarily determines the collision risk by superimposition test, the calculation amount is small, so it will not occupy a long calculation time, and a large amount of data operation is not required, so it can support the processor to perform real-time operation.

[0010] According to a preferred embodiment, the processor constructs a first geometric body in a regular manner by: receiving geometric parameters of materials collected by the UAV; confirming the UAV's carrying method of the materials; generating contour range parameters of the UAV based on the geometric parameters of the materials and the carrying method; and constructing the first geometric body based on the longest side of the UAV's contour in three dimensions. Compared with the prior art, the present invention can adjust the specific parameters of the first geometric body according to the information of the materials carried by the UAV. Based on the above-mentioned distinguishing technical features, the problem to be solved by the present invention may include: how to incorporate the information of the materials carried by the UAV into the scope of collision analysis to improve the accuracy of collision analysis and route planning. Specifically, due to the different materials carried, the collision analysis parameters that need to be considered during the flight of the UAV are also different. Even if carrying materials of the same mass, the space occupied by the materials is different depending on the type of materials, which in turn affects the safe flight of the UAV's planned route. However, the shape of the materials or packaging carried by the UAV is not completely regular. If collision analysis is performed separately for different carried materials, it will obviously increase the additional material modeling procedures and collision analysis calculations, thereby reducing the real-time performance of UAV collision risk warning analysis. Conversely, this invention simplifies the outline of the drone and its carried supplies into geometric shapes, which helps to simplify the initial calculation steps and reduce the amount of computation.

[0011] According to a preferred embodiment, the processor constructs the second geometry of the rule by: determining the height range of the building based on the 3D simulation scene and the planned route; determining the 3D model segments of the building based on the height range; and constructing the second geometry based on the longest side of the 3D model segments. This invention simplifies the 3D model segments of the building into cubic units based on flight altitude and the planned route, aiming to simplify the computational load and steps of the 3D model overlay test and quickly eliminate 3D model parts that do not pose a collision risk.

[0012] According to one preferred embodiment, the processor performs the superposition test in the following manner: predicting the dynamic situation of the first geometric body passing the second geometric body along the planned route, calculating the superposition situation of the first geometric body and the second geometric body, and the superposition situation includes intersection, tangency and inclusion. Compared with the prior art, the processor of the present application can dynamically analyze the superposition state of the first geometric body and the second geometric body during the flight of the first geometric body along the planned route. Based on the above technical features, the problem to be solved by the present application can include: how to dynamically analyze the superposition situation of the first geometric body and the second geometric body according to the planned route. Specifically, the existing collision detection method based on geometric modeling mainly determines the collision area or collision position through the static superposition of the model. However, during the flight of the unmanned aerial vehicle, different planned flight routes will cause the positional relationship between the unmanned aerial vehicle and the static building to change, and relying only on the static model cannot reflect the collision risk in this state. On the contrary, unlike static superposition, there are multiple angles to determine whether the unmanned aerial vehicle collides with the building. Therefore, testing the superposition situation according to the dynamic situation passing the second geometric body can fully obtain accurate superposition results.

[0013] According to one preferred embodiment, the contour range parameter of the unmanned aerial vehicle is generated according to the geometric parameters of the material and the carrying method of the material, and the first geometric body is constructed based on the maximum value of the sum of the contour range parameter and the safety range parameter of the unmanned aerial vehicle in the three-dimensional direction. The present application includes the safety range parameter of the unmanned aerial vehicle in the first geometric body, giving the unmanned aerial vehicle sufficient safety flight deviation space, and avoiding the risk of collision between the unmanned aerial vehicle and the building in the case of adverse wind and material swing.

[0014] According to one preferred embodiment, the processor is further configured to: divide the second geometry in which the overlap exists along a horizontal direction and / or a vertical direction to form a plurality of second sub-geometries before calculating the overlap range, and correct the size of the second sub-geometries based on the three-dimensional model segment of the building within the second sub-geometries; and perform an overlap test on the first sub-geometries and the second sub-geometries, and screen the second sub-geometries in which the overlap exists. Compared with the prior art described above, the present application can divide the second geometry into a plurality of second sub-geometries by means of segmentation. Based on the above distinguishing technical features, the problem to be solved by the present application can include: how to screen the area on the second geometry in which the collision risk exists, so as to reduce the data calculation amount of the collision analysis. Specifically, the second geometry actually refers to the space occupied by the entire building along the planned route, and the part of the space actually only occupies a small part of the entire space in which the unmanned aerial vehicle can produce contact or exist contact risk. If the data information of the second geometry is all included in the data processing process of the collision analysis, the data processing speed will be significantly reduced, resulting in a decrease in the real-time performance of the collision risk analysis. In order to further simplify the steps and calculation amount of the early-stage calculation, the present application segments the second geometry to simply exclude the area in which the collision risk does not exist, thereby reducing the calculation amount of the complex calculation of the subsequent overlap range. By dividing the second geometry, especially the second geometry of the building with a special structure, the present application can obviously exclude a large number of areas without collision risk, thereby reducing the calculation amount of the subsequent steps.

[0015] According to one preferred embodiment, the processor is further configured to: calculate the overlap range of the first sub-geometries and the second sub-geometries in which the overlap exists, and adjust the planned route based on the overlap range. The present application excludes the first sub-geometries and the second sub-geometries in which the collision risk does not exist, significantly reduces the calculation amount of the overlap range, shortens the calculation time, and enables the unmanned aerial vehicle to obtain the calculation feedback in real time.

[0016] The present application discloses, from another aspect, a collision risk prediction method for an unmanned aerial vehicle. The method comprises: constructing a regular first geometry based on the contour range of the unmanned aerial vehicle; constructing a regular second geometry based on the three-dimensional model of the surrounding building of the planned route of the unmanned aerial vehicle; performing an overlap test on the first geometry and the second geometry along the planned route to screen the second geometry in which the overlap exists; dividing the first geometry into a plurality of first sub-geometries according to the movement trend of the unmanned aerial vehicle, calculating the overlap range of the first sub-geometries and the second geometry, and judging that the unmanned aerial vehicle and the building exist a collision risk in the case that at least one overlap point in the overlap range is within the three-dimensional model of the building or on the edge of the building.

[0017] The collision risk prediction method of the present application reduces the calculation time and amount by a large number of simplified calculation steps of regular geometric bodies in advance, and excludes the volume part without collision risk, so that the calculation amount of the subsequent step requiring accurate calculation result is significantly reduced.

[0018] According to a preferred embodiment, the superposition test mode comprises: predicting the dynamic situation of the first geometric body passing the second geometric body along the planned route, calculating the superposition situation of the first geometric body and the second geometric body, and the superposition situation comprises intersection, tangency and inclusion.

[0019] The present application can avoid the phenomenon that the multi-angle superposition range of the UAV is ignored in the case of bypassing or passing through the building by predicting the dynamic situation of the UAV carrying the material to calculate the superposition range.

[0020] According to a preferred embodiment, the method further comprises: before calculating the superposition range, dividing the second geometric body with the superposition situation along the horizontal direction and / or the vertical direction to form a plurality of second sub-geometric bodies, and correcting the size of the second sub-geometric body based on the three-dimensional model segment of the building in the second sub-geometric body; and performing superposition test on the first sub-geometric body and the second sub-geometric body, and screening out the second sub-geometric body with the superposition situation. The present application can obviously exclude a large number of regions without collision risk by dividing the second geometric body, especially the second geometric body of the building with special structure, thereby reducing the calculation amount of the subsequent step. BRIEF DESCRIPTION OF DRAWINGS

[0021] Figure 1 is a schematic diagram of the collision risk state of the UAV provided by the present application;

[0022] Figure 2 is a schematic diagram of the superposition test of the first geometric body and the second geometric body provided by the present application;

[0023] Figure 3 is a schematic diagram of the steps of the collision risk prediction method of the UAV provided by the present application;

[0024] Figure 4 is a schematic diagram of the first sub-geometric body and the third sub-geometric body without collision situation provided by the present application;

[0025] Figure 5 is a schematic diagram of the first sub-geometric body and the third sub-geometric body without collision situation provided by the present application from the perspective angle;

[0026] Figure 6 is a schematic diagram of the first sub-geometric body and the third sub-geometric body with collision situation provided by the present application;

[0027] Figure 7is a schematic diagram of a top view angle of a collision between the first sub-geometric body and the third sub-geometric body of the present application;

[0028] Figure 8 is a schematic diagram of a situation in which a remote processor of the present application plans a route for a UAV;

[0029] Figure 9 is a schematic diagram of a top view angle of a situation in which a UAV of the present application plans a route.

[0030] List of reference signs

[0031] 100: first geometric body; 101: UAV; 102: material; 103: first sub-geometric body; 200: second geometric body; 201: building; 202: second sub-geometric body; 203: third sub-geometric body; 300: planned route; 301: first route; 302: second route; 303: third route; 400: fire point; 401: smoke range; 500: first building; 600: second building. DETAILED DESCRIPTION

[0032] The following will be described in detail with reference to the accompanying drawings.

[0033] Three-dimensional model: a mapping of a real object in a virtual space, wherein the mapping also covers the physical relationship, spatial relationship and various attributes of the real object with other real objects. The three-dimensional model involved in the present application is, for example, a building 201 and its related components in a virtual three-dimensional space. It should be noted that the three-dimensional model of the building 201 is not limited to the building 201 itself, but also includes the high-altitude components near the building 201. The high-altitude components are, for example, high-line pipelines, lines, extended air conditioner outdoor units, etc.

[0034] Geometric body: refers to a regular geometric structure with a rectangular cross section constructed based on three-dimensional coordinate axes in a three-dimensional virtual scene.

[0035] Building 201: when described in a three-dimensional simulation scene, the building 201 refers to a virtual building corresponding to a physical building and displayed in a three-dimensional model. In a three-dimensional virtual scene of a digital twin factory, the virtual building is consistent with the physical building, and therefore the virtual building is also referred to as the building 201.

[0036] Embodiment 1

[0037] In the prior art, the control system of the UAV 101 pre-plans the route for the UAV 101 to fly. When the UAV 101 carries large or heavy goods 102 in an emergency, the UAV 101 may collide with a building 201. At this time, if the UAV 101 temporarily changes the route, it also needs to spend time waiting for a new planned route 300. During the waiting period, the UAV 101 carrying heavy goods consumes a large amount of electrical energy. Even if the UAV 101 has an automatic obstacle avoidance device, it can only make the UAV 101 recognize the obstacle and hover to wait for new flight instructions; or the UAV 101 autonomously calculates and avoids the obstacle according to the image of the obstacle, which also prolongs the hovering time of the UAV 101. Moreover, the UAV 101 can only recognize and bypass the obstacle when it is near the obstacle, and cannot bypass it in advance. Temporary bypassing is extremely disadvantageous for the UAV 101 carrying heavy or large goods 102, not only increasing the electrical energy consumed by the UAV 101 when ascending, but also delaying the UAV 101 to reach the destination.

[0038] As described above, for an industrial plant, there are a large number of industrial pipelines in the high altitude. The UAV 101 carrying heavy goods needs to consume more energy in the high altitude where there are no buildings 201, which is not conducive to emergency rescue. The UAV 101 carrying heavy goods has large inertia, and the risk of temporary bypassing when encountering obstacles is also extremely high. Therefore, when the UAV 101 carries goods 102, temporarily planning or adjusting a shorter low-altitude route for the UAV 101 based on the condition of the goods 102 so that the UAV 101 reaches the destination in a shorter time is the purpose of the present application. Reducing the calculation amount of the collision risk of the UAV 101 so that the UAV 101 can be calculated and fed back with route information and collision risk information in real time is the core of the UAV 101 to shorten the flight time.

[0039] In view of the deficiencies of the prior art, the present application provides a collision risk prediction system and method for a UAV 101. The present application can also provide a digital twin factory UAV 101 route planning method and system for calibrating the flight route according to the collision risk. The present application can also provide a digital twin-based UAV 101 route adjustment system and method for adjusting the flight route according to the collision risk. The present application can also provide a UAV 101 goods 102 management system and method for managing the type, volume specification and flight route of the goods 102 carried by the UAV 101 according to the prediction result of the collision risk, so that the UAV 101 carrying different goods 102 can fly according to different planned routes 300. When the planned route 300 of the UAV 101 is designated and cannot be changed, the volume specification of the goods 102 carried by the UAV 101 can be limited according to the predicted collision risk, so that the UAV 101 can safely reach the destination in a shorter time.

[0040] In view of the deficiencies of the prior art, the present application discloses, in one aspect, a collision risk prediction system for a UAV 101, which comprises a remote processor in communication with the UAV. Since the UAV itself has limited power and computing capacity, it is necessary to use the remote processor to calculate and analyze the most appropriate planned route that can be used when carrying supplies. The limitation of the UAV flight is that it cannot fly backward directly, but only forward according to the route. Therefore, if the UAV flies to the vicinity of a building while carrying supplies and then plans a new route after finding that it cannot pass smoothly, the UAV will need to consume a large amount of energy to change the flight direction and plan a new route, which will undoubtedly waste the rescue time, energy and data processing function of the chip of the UAV in an emergency. Therefore, after carrying supplies, or after determining the size and carrying posture of the supplies, the UAV should send the data of the supplies and the planned route to the remote processor, which calculates and analyzes whether there is a flight risk when the UAV carries the supplies. If there is, the processor changes the planned route for the UAV and feeds it back to the UAV.

[0041] The number of processors is not limited and can be one, two or more. The processor is used to run the program code of the method of the present application. The processor of the present application can be a processor provided by the UAV 101 or a processor in communication connection with the UAV 101. The processor is preferably a device including an application-specific integrated chip, a server, a server group, etc. For example, the processor can refer to a ground control station of the UAV 101 or a large server group of a digital twin factory.

[0042] In the present application, the three-dimensional virtual scene refers to a three-dimensional scene of a digital twin factory existing in a real factory. The three-dimensional model in the three-dimensional scene is consistent with the physical building in reality.

[0043] As shown in Figure 1 and Figure 3 , the method of the present application comprises:

[0044] S110: constructing a first geometric body 100 based on the contour range of the UAV 101.

[0045] S120: constructing a second geometric body 200 based on the three-dimensional model of the surrounding buildings 201 of the planned route 300 of the UAV 101.

[0046] S130: superimposing the first geometric body 100 along the planned route 300 with the second geometric body 200 to screen the second geometric body 200 that exists in the superimposed condition.

[0047] Preferably, the processor performs the superposition test in the following way: predicting the dynamic situation of the first geometric body 100 passing the second geometric body 200 along the planned route 300, and calculating the superposition of the first geometric body 100 and the second geometric body 200. The superposition includes intersection, tangency and inclusion. Unlike the static superposition, the present application can determine whether the UAV 101 collides with the building 201 from multiple angles. Therefore, testing the superposition according to the dynamic situation of passing the second geometric body 200 can obtain an accurate superposition result.

[0048] S140: As shown in the figure, the first geometric body 100 is divided into a plurality of first sub-geometric bodies 103 according to the movement trend of the UAV. The superposition range of the first sub-geometric body 103 and the second geometric body 200 is calculated. If at least one superposition point in the superposition range is in the three-dimensional model of the building 201 or on the edge, it is determined that the UAV 101 has a collision risk with the building 201. Figure 2

[0049] In the case of the UAV carrying goods, the flight direction of the UAV is different, and the specific position of the first geometric body 100 composed of the UAV, the goods and the safety range is also different, which makes the collision risk of the first geometric body 100 and the second geometric body 200 completely different. Therefore, the first geometric body 100 is divided according to the flight trend of the UAV, and specifically, the first geometric body 100 is divided according to the flight direction of the UAV, so that the division line is perpendicular to the flight direction, which can avoid the collision between the UAV and the surrounding buildings and the goods carried by the UAV. If the first geometric body 100 is not divided according to the flight direction of the UAV, for example, the first geometric body 100 is divided according to the direction perpendicular to the flight direction of the UAV, then the division line is parallel to the flight direction, which makes the collision calculation of part of the first sub-geometric body 103 and the second geometric body 200 of the building invalid, and instead wastes the computing power of the remote processor.

[0050] The present application can calculate the collision risk of the planned route 300 and adjust the planned route 300 based on the difference in size of the goods 102 carried by the UAV 101 in real time, so that the UAV 101 can carry goods 102 based on a shorter route and fly safely. Since the present application preliminarily determines the collision risk by superposition test, the calculation amount is small, so it will not occupy a long calculation time, and it will not require a large amount of data operation, so it can support the processor to perform real-time operation.

[0051] Preferably, the step of constructing a regular first geometric body 100 by the processor includes:

[0052] S111: Receive the geometric parameters of the goods 102 collected by the UAV 101.

[0053] ​Preferably, when the unmanned aerial vehicle 101 hoists the material 102, the unmanned aerial vehicle 101 can measure the size of the material 102 according to the optical measurement component such as the infrared measurement instrument carried by itself. Alternatively, the code or identification of the material 102 is collected, and the processor retrieves the geometric parameters of the material 102 according to the code or identification. The geometric parameters include the length, width, height and other parameters of the material 102. Preferably, when the unmanned aerial vehicle 101 is provided with a hoisting tool, the unmanned aerial vehicle 101 can determine the distance between the material 102 and the base of the unmanned aerial vehicle 101 according to the length of the hoisting tool or the length change value.

[0054] S112: Confirm the carrying manner of the unmanned aerial vehicle 101 to the material 102.

[0055] The processor confirms the carrying manner of the material 102 according to the type of the unmanned aerial vehicle 101 or the hoisting information of the material 102 fed back by the unmanned aerial vehicle 101. For example, when the unmanned aerial vehicle 101 hoists the heavy material 102 such as a box or a bag, the carrying manner is hoisting, and the processor selects to recalculate the overall profile of the unmanned aerial vehicle 101 and the material 102 according to the hoisting manner. When the unmanned aerial vehicle 101 fills the material 102 such as fire extinguishing liquid or water through its own box, the appearance profile of the unmanned aerial vehicle 101 does not change. Therefore, the processor can determine whether the overall profile formed by the unmanned aerial vehicle 101 and the material 102 changes according to the type and carrying manner of the unmanned aerial vehicle 101.

[0056] S113: Generate the profile range parameters of the unmanned aerial vehicle 101 according to the geometric parameters of the material 102 and the carrying manner. A first geometric body 100 is constructed based on the longest side of the profile of the unmanned aerial vehicle 101 in the three-dimensional direction. The present application simplifies the profile of the unmanned aerial vehicle 101 and the material 102 carried thereby into a geometric body, which is beneficial to simplify the calculation steps and calculation amount in the early stage.

[0057] In the case that the unmanned aerial vehicle 101 carries the material 102, especially in the case that the unmanned aerial vehicle 101 hoists the material 102, the flight space required by the unmanned aerial vehicle 101 in the vertical direction increases. If the unmanned aerial vehicle 101 flies according to the original planned route 300 to a certain narrow space, the unmanned aerial vehicle 101 may not be able to pass through exactly. Therefore, the overall profile formed by the unmanned aerial vehicle 101 and the material 102 is set as a geometric body, which can cover the space required by the unmanned aerial vehicle 101 as a regular geometric body, so as to reduce the subsequent calculation amount.

[0058] Preferably, the step of constructing the regular first geometric body 100 by the processor can further be:

[0059] S114: Generate the profile range parameters of the unmanned aerial vehicle 101 according to the geometric parameters of the material 102 and the carrying manner of the material 102.

[0060] S115: constructing the first geometric body 100 based on the maximum value of the sum of the contour range parameter and the safety range parameter of the UAV 101 in the three-dimensional direction.

[0061] Since the UAV 101 flies in the air, there is a wind resistance effect. When the wind is against, the wind resistance is large, and when the wind is with, the wind resistance is small, so the UAV 101 has a jitter and slight deviation phenomenon when flying. The safety range parameter corresponding to the UAV 101 is set for this phenomenon. The safety range is included in the first geometric body 100, which expands the entity range of the UAV 101, but improves the safety of the UAV 101.

[0062] The safety range parameter of the UAV 101 is also included in the first geometric body 100, which gives the UAV 101 sufficient safe flight deviation space, avoiding the risk of collision between the UAV 101 and the building 201 in the case of adverse wind and material 102 swing.

[0063] Preferably, in the case of the UAV 101 bypassing or passing through the building 201, the first channel model is constructed based on the influence distance between the UAV 101 and the building 201 and the planned route 300 as the first geometric body 100. Compared with the case that the UAV 101 passes through the building 201 in a straight line, the UAV 101 bypasses or passes through the building 201 will exist multiple first geometric bodies. If multiple geometric bodies are used to perform superposition testing respectively, the calculation of redundant data will undoubtedly increase. Therefore, when the UAV 101 bypasses or passes through the building 201 in a non-linear manner, the first channel model is constructed based on the planned route 300, the contour range parameter and the safety range parameter of the UAV 101. Preferably, the cross section of the first channel model is a regular rectangular surface. The length of each side of the rectangular surface is the maximum value of the sum of the UAV 101 and the material 102, the safety range parameter in the corresponding direction.

[0064] The first channel model is used to perform superposition testing with the second geometric body 200, which can significantly reduce the calculation and processing of redundant data. Especially in the case that the UAV 101 needs to pass through the building 201, the passing space of the building 201 is limited, and whether the first geometric body 100 of the UAV 101 will collide with the building 201 or pass through smoothly needs to be accurately predicted.

[0065] According to one preferred embodiment, the step of the processor constructing the regular second geometric body 200 comprises:

[0066] S121: determining the height range of the building 201 based on the three-dimensional simulation scene and the planned route 300.

[0067] The building 201 is built from the ground, so the data of the three-dimensional model of the building 201 is more. If the superposition test is performed according to the data of the overall three-dimensional model of the building 201, a large amount of invalid data will be used, and a large amount of calculation will be increased. This is obviously inconsistent with the purpose of reducing the amount of calculation of the present application.

[0068] Therefore, the present application quickly intercepts the height range of the building 201 according to the planned route 300 and the flight height, and only calculates the effective height that can collide with the unmanned aerial vehicle 101 and the goods 102 carried thereby, so that the amount of data calculation can be obviously reduced.

[0069] S122: Determine the three-dimensional model segment of the building 201 based on the height range. In this step, only a small amount of contour data of the three-dimensional model is used, and the overall data of the three-dimensional model in the height range is not needed, so the amount of data calculation is small, and the calculation time is short.

[0070] S123: Construct the second geometric body 200 based on the longest side of each coordinate axis of the three-dimensional model segment. The advantage of this setting is that the three-dimensional model of the building 201 can be simplified. The calculation amount of the regular geometric body is less when the superposition test is performed, and the calculation time is short.

[0071] The present application simplifies the three-dimensional model segment of the building 201 to a cube based on the flight height and the planned route 300, so as to simplify the calculation amount and calculation steps of the superposition test of the three-dimensional model, and quickly exclude the three-dimensional model part that does not have a collision risk.

[0072] Preferably, for the linear building 201, the second geometric body 200 in the height range has a small difference with the actual contour of the building 201. At this time, the first sub-geometric body 103 and the second geometric body 200 are used to perform the superposition test, that is, when the distance between the first geometric body 100 and the second geometric body 200 is the shortest, the overlap of the first sub-geometric body 103 and the second geometric body 200 is calculated. Specifically, the point set of the superposition points intersected by the first sub-geometric body 103 and the second geometric body 200 is calculated. The calculation method can be quickly calculated by a simple data range comparison, a simple function algorithm of intersection point calculation, and a plurality of algorithms, which will not be described in detail here.

[0073] The set of superposition points constitutes a superposition range. The processor can determine the collision risk of the unmanned aerial vehicle 101 carrying the goods 102 and the building 201 according to the superposition range. The larger the volume of the superposition range is, the greater the collision risk is.

[0074] Preferably, the processor adjusts the flight posture and flight height of the UAV 101 when passing the building 201 based on the superposition range to eliminate the collision risk of the UAV 101. The advantage of the implementation of the present application is that for the case of small collision range, the UAV 101 can pass through successfully as long as the flight height and flight posture are adjusted based on the special situation. There is no need to reselect the planned route 300. The defect of reselecting the planned route 300 is that the calculation amount of re-planning the route 300 is large, the flight route is far, and the flight time is prolonged. These defects are not conducive to the UAV 101 carrying the supplies 102 in an emergency.

[0075] Preferably, for the structure in which part of the building 201 is in the height range, it is possible to be irregular. For example, the shape of many production buildings in the industrial park changes greatly, or a large number of mechanical structures are arranged so that the surface structure of the building 201 is irregular. Thus, a large amount of redundant space is contained in the second geometric body 200. In order to further reduce the data processing amount in the superposition test process, the present application also divides the second geometric body 200, thereby eliminating the part without collision risk.

[0076] Preferably, before calculating the superposition range, the processor further implements the following steps.

[0077] S150: The second geometric body 200 is divided to eliminate the building 201 region without collision risk, which specifically includes the following steps.

[0078] S151: As shown in the figure, the second geometric body 200 in which the superposition exists is divided in the horizontal direction and / or the vertical direction to form a plurality of second sub-geometric bodies 202. Figure 2

[0079] Preferably, it can be divided in the horizontal direction, or it can be divided in the vertical direction, or it can be divided in the horizontal direction and the vertical direction at the same time.

[0080] Preferably, when the distance between the vertical axis of the first geometric body 100 and the vertical axis of the second geometric body 200 is the shortest, the processor divides the second geometric body 200 in the vertical direction, that is, the superposition test is performed when the first geometric body 100 is closest to the second geometric body 200.

[0081] For the building 201 in which there is a significant change in the outline or bending in the second geometric body 200, the processor should further eliminate the part without collision risk to reduce the calculation amount in the later stage. Preferably, in the height range of the planned route 300, all the vertices of the two ends of the second geometric body 200 are connected and constructed to form a third geometric body 203, as shown in the figure. Figures 4 to 7 Figure 4 ​​As shown, all the vertices of the two ends of the second geometry 200 of the building 201 are connected and a third geometry 203 is formed. At this time, the third geometry 203 is a trapezoidal body.

[0082] The first geometry 100 and the third geometry 203 are subjected to the superposition test, and there are three special cases between the first geometry 100 and the third geometry 203:

[0083] First, the first geometry 100 is tangent to or overlaps with the third geometry 203, as shown in Figure 6 and Figure 7 .

[0084] Second, the first geometry 100 contains the third geometry 203.

[0085] Third, the first geometry 100 neither contains nor is tangent to or overlaps with the third geometry 203, as shown in Figure 4 and Figure 5 .

[0086] After the superposition test between the first geometry 100 and the third geometry 203, the processor retains the third geometry 203 which has the tangential, overlapping relationship and the containing relationship with the first geometry 100.

[0087] Alternatively, in the case where the first geometry 100 has been divided into a plurality of first sub-geometries 103, the processor retains the third geometry 203 which has the tangential, overlapping relationship and the containing relationship with the first sub-geometry 103.

[0088] In this way, a part of the second geometry 200 irrelevant to the first geometry 100 can be excluded, and the data processing amount of the subsequent point-by-point collision can be reduced.

[0089] Preferably, as shown in Figure 6 , in the case where the first geometry 100 is tangent to or overlaps with the third geometry 203, the second geometry 200 is divided into at least two geometric sub-bodies based on the bending characteristics of the second geometry 200. For example, the division (cutting) line is determined at the point where the second geometry 200 has the largest bending angle. The size of the geometric sub-body is corrected based on the contour of the three-dimensional model in the geometric sub-body. The first geometry 100 or the retained first sub-geometry 103 is subjected to the superposition test with the corrected geometric sub-body, and the geometric sub-body with the collision risk is selected and retained. In the subsequent step, the geometric sub-body is subjected to the high-precision calculation of the superposition range with the first sub-geometry 103. Such steps can be performed multiple times, which is beneficial to simply and quickly remove invalid data.

[0090] After the division and selection of the geometric sub-body, the probability of the three-dimensional model in the geometric sub-body having the collision risk is extremely high.

[0091] For the building 201 with changing profile, after cutting to form several second sub-geometric bodies 202, the second sub-geometric bodies 202 are far from the real profile of the building 201 therein. At this time, if the calculation of the superposition range is performed based on the current second sub-geometric bodies 202, there is still a large amount of calculation of invalid data.

[0092] S152: The size of the second sub-geometric body 202 is corrected based on the three-dimensional model segment of the building 201 therein. Through this step, the size of the second sub-geometric body 202 is corrected according to the real profile of the building 201, so that the processor can first roughly determine whether the second sub-geometric body 202 has a collision risk with the first sub-geometric body 103. Then, the second sub-geometric body 202 with the collision risk is screened out.

[0093] S153: The first sub-geometric body 103 is superposed with the second sub-geometric body 202 or the geometric sub-body, and the second sub-geometric body 202 or the geometric sub-body with the superposition is screened out.

[0094] In order to further simplify the steps and calculation amount of the early-stage calculation, the second geometric body 200 is cut to roughly exclude the regions without the collision risk, thereby reducing the calculation amount of the complex calculation of the subsequent superposition range.

[0095] According to one preferred embodiment, the processor is further configured to perform step S160.

[0096] S160: The superposition range of the first sub-geometric body 103 and the second sub-geometric body 202 or the geometric sub-body with the superposition is calculated, and the planned route 300 is adjusted based on the superposition range. The first sub-geometric body 103 and the second sub-geometric body 202 without the collision risk are excluded, the calculation amount of the superposition range is obviously reduced, the calculation time is shortened, and the unmanned aerial vehicle 101 can obtain the calculation feedback in real time.

[0097] For example, the first geometric body 100 and the second geometric body 200 are respectively represented as standard equations. The superposition points are calculated based on the parametric equation and the standard equation.

[0098] Preferably, the standard equation representing the first aggregate is:

[0099] Ax+By+Cz+D=0

[0100] Wherein, A, B, and C are respectively the coefficients of the three directions of the geometric body, and D is a constant.

[0101] The second geometric body 200 is depicted based on the parametric representation method, and the parametric equation thereof is:

[0102] P(t)=P1+t(P2-P1)

[0103] wherein P1 is the first end of the second geometric body 200, and P2 is the second end of the second geometric body 200.

[0104] The parametric equation of the second geometric body 200 is substituted into the standard equation of the first geometric body 100 to calculate the value of t. If 0≤t≤1, it is determined that the first geometric body 100 and the second geometric body 200 are superimposed. At this time, the coordinates of the specific superimposed point can be obtained by substituting t into the parametric equation P(t). If the value of t is not in the interval [0, 1], it is determined that the first geometric body 100 and the second geometric body 200 do not intersect.

[0105] Preferably, the processor stores a superimposition range threshold value, i.e., a volume threshold value, related to the adjustment of the planned route 300. In the case where the superimposition range is greater than the superimposition range threshold value, the processor reselects the planned route 300 for the unmanned aerial vehicle 101. In the case where the superimposition range is less than or equal to the superimposition range threshold value, the processor slightly adjusts the planned route 300 for the unmanned aerial vehicle 101, so that the unmanned aerial vehicle 101 can slow down or pass through the building 201 at a specified height and specified posture.

[0106] As shown in Figure 8 and Figure 9 , there are three planned routes from the location where the unmanned aerial vehicle 101 carries the supplies to the fire point 400, which are the first route 301, the second route 302, and the third route 303. The unmanned aerial vehicle 101 remotely sends the data of the first route 301, the second route 302, and the third route 303 to the processor, and the processor calculates and feeds back the optimal flight route. At the same time, the unmanned aerial vehicle 101 also sends the geometric parameters of the supplies and the information about the carrying situation to the processor. The first route 301 is the route for the unmanned aerial vehicle 101 to fly from one side of the first building body 500 to the fire point 400. The second route 302 is the route for the unmanned aerial vehicle 101 to carry the supplies to pass through the passage between the first building body 500 and the second building body 600. The dashed line segment in the second route 302 represents the route segment passing through the passage. The third route 303 is the route for the unmanned aerial vehicle 101 to carry the supplies to pass through the periphery of the second building body 600.

[0107] The processor selects the first building body 500 and the second building body 600 that have an impact on flight according to the position information of the fire point 400, the data of the first route 301, the second route 302, and the third route 303. The processor retrieves the geometric structure information of the buildings of the first building body 500 and the second building body 600, especially the contour information. The advantage of retrieving the contour information is that it can reduce the amount of data retrieval and analysis. The internal structure information of the first building body 500 and the second building body 600 is redundant data and does not need to be retrieved.

[0108] As shown in Figure 8 andFigure 9 As shown, the top view interface of the first building body 500 tends to be square. The top view interface of the second building body 600 tends to be L-shaped. There is a narrow passage between the first building body 500 and the second building body 600. Under the action of the wind, the smoke of the fire point 400 spreads to the side of the first building body 500 away from the passage, forming a smoke range 401. The processor collects the visibility in the smoke range 401 according to the three-dimensional simulation scene in the digital twin factory, or confirms the smoke concentration and its smoke range 401 according to the image data transmitted by the unmanned aerial vehicle performing the observation task. The processor determines whether to affect the planned route of the unmanned aerial vehicle according to the parking position of the fire vehicle of the fire point 400 and the type of the fire vehicle. The processor confirms the parking position of the fire vehicle, the type of the fire vehicle and its occupied space according to the image data transmitted by the unmanned aerial vehicle performing the observation task. Preferably, the processor calculates the passing probability of the first route 301 according to the smoke range 401 and its visibility. The first route 301 is excluded in the case that the smoke range 401 is large and affects the judgment of the unmanned aerial vehicle throwing materials and obstacle avoidance. Preferably, when the second route 302 is to be used, the processor predicts the collision between the unmanned aerial vehicle 101 and the passage between the first building body 500 and the second building body 600 and the ladder of the ladder fire vehicle according to the deployment height of the ladder fire vehicle. If the processor judges that there is a collision risk between the unmanned aerial vehicle 101 and the ladder of the ladder fire vehicle, and the electric energy of the unmanned aerial vehicle is not enough to support the unmanned aerial vehicle to climb and land in the air above the ladder with carrying materials, the processor judges that the second route 302 is not suitable for the unmanned aerial vehicle to fly with carrying materials. The processor calculates the electric energy consumption and the collision risk with the first building body 500 and the second building body 600 of the unmanned aerial vehicle 101 flying along the third route 303 with carrying materials. Although the third route is long, the unmanned aerial vehicle 101 does not need to carry materials to climb and land in the air, there is not much collision risk due to the obstruction of the ladder position, and the unmanned aerial vehicle 101 cannot judge the throwing position and throwing angle of the materials due to the influence of the smoke, so the third route 303 is more suitable for the current unmanned aerial vehicle 101 with carrying materials. The processor sends the third route 303 to the unmanned aerial vehicle 101. The remote processor of the present application can simultaneously meet the requests of the planning routes of multiple unmanned aerial vehicles due to the less data required by the analysis and calculation method and the short analysis time, and can feed back accurate route information to the request of the planning route of the unmanned aerial vehicle.

[0109] It should be noted that the above-mentioned embodiments illustrate rather than limit the application, and that those skilled in the art will be able to devise modifications which, though perhaps not specifically enumerated herein, fall within the scope of the application. Those skilled in the art will understand that the drawings, described above, and the claims, described below, are intended to be illustrative, and not restrictive, of the scope of the application. The true scope of the application is defined by the claims, and equivalents thereof. The specification and drawings are, accordingly, to be regarded as illustrative rather than restrictive. Several inventive concepts are described herein, such as "preferably," "according to one embodiment," or "optionally," each of which indicates that the corresponding paragraph discloses an independent concept, and the applicant reserves the right to file a divisional application based on each of these concepts.

Claims

1. A collision risk prediction system for a drone, the system comprising a remote processor in communication with the drone, characterized in that, The processor is configured to: construct a first geometric body (100) based on the contour range of the UAV (101); construct a second geometric body (200) based on the three-dimensional model of the surrounding building (201) of the planned route (300) of the UAV (101); superimpose the first geometric body (100) along the planned route (300) with the second geometric body (200) to test for the second geometric body (200) with an overlapping condition; divide the first geometric body (100) into a plurality of first sub-geometric bodies (103) according to the movement trend of the UAV, and calculate the overlapping range of the first sub-geometric body (103) and the second geometric body (200), if at least one overlapping point in the overlapping range is within or on the edge of the three-dimensional model of the building (201), determine that the UAV (101) has a collision risk with the building (201). 2.The collision risk prediction system of the UAV of claim 1, wherein, The manner in which the processor constructs the first geometric body (100) includes: receiving the geometric parameters of the material (102) collected by the UAV (101); confirming the carrying manner of the material (102) by the UAV (101); generating the contour range parameters of the UAV (101) according to the geometric parameters of the material (102) and the carrying manner, constructing the first geometric body (100) based on the longest side of the contour of the UAV (101) in the three-dimensional direction. 3.The collision risk prediction system of the UAV of claim 1 or 2, wherein, The manner in which the processor constructs the second geometric body (200) includes: determining the height range of the building (201) based on the three-dimensional simulation scene and the planned route (300); determining the three-dimensional model segment of the building (201) based on the height range; constructing the second geometric body (200) based on the longest side of the three-dimensional model segment. 4.The collision risk prediction system of the UAV of claim 1, wherein, The manner in which the processor performs the superimposition test includes: predicting the dynamic condition of the first geometric body (100) along the planned route (300) passing through the second geometric body (200), and calculating the overlapping condition of the first geometric body (100) and the second geometric body (200), the overlapping condition includes intersection, tangency and inclusion.

5. The collision risk prediction system of the UAV according to claim 2, wherein: generating the contour range parameters of the UAV (101) according to the geometric parameters of the material (102) and the carrying manner of the material (102), constructing the first geometric body (100) based on the maximum value of the sum of the contour range parameters and the safety range parameters of the UAV (101) in the three-dimensional direction. 6.The collision risk prediction system of the UAV of claim 1, wherein, The processor is further configured to: before calculating the overlapping range, dividing the second geometric body (200) with an overlapping condition along the horizontal direction and / or the vertical direction to form a plurality of second sub-geometric bodies (202), correcting the size of the second sub-geometric body (202) based on the three-dimensional model segment of the building (201) within the second sub-geometric body (202). The first sub-geometric body (103) is superimposed with the second sub-geometric body (202) to screen the second sub-geometric body (202) with superimposition. 7.The collision risk prediction system of drones of claim 6, wherein, The processor is further configured to: calculate the superimposition range of the first sub-geometric body (103) and the second sub-geometric body (202) with superimposition, adjust the planned route (300) based on the superimposition range. 8.A method for predicting collision risk of a UAV, the method comprising: The method comprises: constructing a regular first geometric body (100) based on the contour range of the unmanned aerial vehicle (101); constructing a regular second geometric body (200) based on the three-dimensional model of the surrounding building (201) of the planned route (300) of the unmanned aerial vehicle (101); superimposing the first geometric body (100) along the planned route (300) with the second geometric body (200) to screen the second geometric body (200) with superimposition; dividing the first geometric body (100) into a plurality of first sub-geometric bodies (103) according to the movement trend of the unmanned aerial vehicle, and calculating the superimposition range of the first sub-geometric body (103) and the second geometric body (200), if at least one superimposition point in the superimposition range is within or on the edge of the three-dimensional model of the building (201), determining that the unmanned aerial vehicle (101) has a collision risk with the building (201). 9.The method of claim 8, wherein, The superimposition test comprises: predicting the dynamic condition of the first geometric body (100) along the planned route (300) through the second geometric body (200), and calculating the superimposition condition of the first geometric body (100) and the second geometric body (200), The superimposition condition comprises intersection, tangency and inclusion. 10.The method of claim 8 or 9, wherein, The method further comprises: before calculating the superimposition range, dividing the second geometric body (200) with superimposition along the horizontal direction and / or the vertical direction to form a plurality of second sub-geometric bodies (202), correcting the size of the second sub-geometric body (202) based on the three-dimensional model segment of the building (201) within the second sub-geometric body (202); superimposing the first sub-geometric body (103) with the second sub-geometric body (202) to screen the second sub-geometric body (202) with superimposition.

Images