UAV Collision Warning Method Based on Distributed Parallel Computing and Partition Compensation

Abstract

The present invention discloses a UAV collision warning method based on distributed parallel computing and partition compensation. By improving the anti-collision warning algorithm and performing anti-collision warning analysis and calculation on UAVs in a parallel computing manner based on a distributed platform, the computational complexity is reduced, the calculation time is decreased, and the warning effect is improved. On this basis, a partition compensation scheme under boundary conditions is also proposed to perform partition compensation on UAVs located at the geographical grid boundaries, thereby improving the accuracy of collision warning on the premise of meeting real-time warning through distributed parallel computing.

Description

Technical Field

[0001] The present invention relates to the technical field of UAV collision warning, and particularly to a UAV collision warning method based on distributed parallel computing and partition compensation. Background Art

[0002] In recent years, UAVs have been widely used in both military and civilian fields. The increasing number of UAVs has brought challenges to flight safety. Anti-collision warning for UAVs in the air is one of the important measures to ensure the flight safety of UAVs. At present, the more widely used UAV anti-collision warning technology is the technology that calculates the distance between UAVs based on a geometric model for anti-collision warning. Its computational complexity is quadratic with the number of UAVs. When the number of simultaneously flying UAVs is large, its computational complexity will increase exponentially, the calculation time will be prolonged, and the anti-collision warning effect will be affected. Summary of the Invention

[0003] The present invention provides a UAV collision warning method based on distributed parallel computing and partition compensation. By improving the anti-collision warning algorithm and performing anti-collision warning analysis and calculation on UAVs in a parallel computing manner based on a distributed platform, the computational complexity is reduced, the calculation time is reduced, and the warning effect is improved. On this basis, a partition compensation scheme under boundary conditions is also proposed to perform partition compensation on UAVs located at the boundaries of geographical grids, thereby improving the accuracy of collision warning on the premise of meeting real-time warning through distributed parallel computing.

[0004] To achieve this purpose, the present invention adopts the following technical solutions:

[0005] Provide a UAV collision warning method based on distributed computing and partition compensation, the steps including:

[0006] S1, collecting flight data of the UAVs;

[0007] S2, calculating the partition flag value corresponding to the geographical grid currently flown into by the UAVs according to the collected flight data;

[0008] S3, matching the machines associated with the partition flag value;

[0009] S4, sending the flight data generated by each of the UAVs currently in the same geographical grid to the machine;

[0010] S5, the machine calculates the distances between each of the UAVs currently flying in the geographical grid based on the received flight data and using a pre-constructed simplified UAV distance calculation model, and issues a warning when it is determined that the collision warning condition is met.

[0011] Preferably, the flight data includes the longitude, latitude, and altitude of the flight position.

[0012] Preferably, the partition flag value is calculated by the following formula (1):

[0013] PartitionID = Lat × 10000007 + Lon Formula (1)

[0014] In Formula (1), Lat represents the longitude value of the left boundary of the geographic grid into which the UAV currently flies;

[0015] Lon represents the latitude value of the upper boundary of the geographic grid into which the UAV currently flies;

[0016] 10000007 is a multiplier factor.

[0017] Preferably, Lat and Lon are integer values.

[0018] Preferably, the method by which the machine calculates the distance between UAVs using the UAV distance calculation simplified model is expressed by the following formulas (2)-(3):

[0019]

[0020]

[0021] In Formulas (2) and (3), l represents the distance between the first UAV and the second UAV; Δh represents the elevation difference between the first UAV and the second UAV;

[0022] represents the distance between the first UAV and the second UAV on the surface of the Earth model sphere;

[0023] R represents the average radius of the Earth;

[0024] Δlat represents the coordinate latitude difference between the first UAV and the second UAV;

[0025] Δlon represents the coordinate longitude difference between the first UAV and the second UAV;

[0026] Lat1 and Lat2 respectively represent the latitude values of the first UAV and the second UAV.

[0027] Preferably, the collision warning condition is that l is less than the product of the relative flight speed of the first UAV and the second UAV and the warning time, and the warning time is 10 seconds.

[0028] Preferably, in step S5, the method steps for calculating the distances between the drones currently flying within the geographic grid include:

[0029] S51, the machine determines whether the drone is currently flying within the boundary buffer zone of the geographic grid.

[0030] If so, it proceeds to step S52;

[0031] If not, the machine itself calculates the distance between the drones.

[0032] S52, determine the type of the boundary buffer zone that the drone currently enters, and then perform corresponding zonal compensation on the drones entering the boundary buffer zone of the corresponding type according to a preset strategy.

[0033] S53, copy the flight data of the drone to the distributed node corresponding to the geographic grid where it falls after zonal compensation, and the distributed node calculates the distance between the drones.

[0034] Preferably, step S52 specifically includes the steps:

[0035] S521, determine the type of the boundary buffer zone that the drone currently enters;

[0036] S522, determine whether the drone currently enters the intersection area of two boundary buffer zones of different types.

[0037] If so, perform zonal compensation on the drone according to the second preset strategy;

[0038] If not, perform zonal compensation on the drone according to the first preset strategy.

[0039] Preferably, in step S52, when the drone currently enters a boundary buffer zone of a single type, the first preset strategy for performing zonal compensation on the drone is:

[0040] When the boundary buffer zone that the drone currently enters is the defined area of the upper boundary of the geographic grid, the drone is zoned and compensated through the following formula (4):

[0041]

[0042] When the boundary buffer zone that the drone currently enters is the defined area of the lower boundary of the geographic grid, the drone is zoned and compensated through the following formula (5):

[0043]

[0044] When the boundary buffer zone that the UAV currently enters is the left boundary defined area of the geographic grid, the UAV is compensated by the following formula (6):

[0045]

[0046] When the boundary buffer zone that the UAV currently enters is the right boundary defined area of the geographic grid, the UAV is compensated by the following formula (7):

[0047]

[0048] In formulas (4)-(7), lon and lat respectively represent the longitude value and latitude value of the geographic coordinates at which the UAV flies;

[0049] S up 、S bottom 、S left 、S right respectively represent the distances from the current flight position of the UAV to the upper boundary, lower boundary, left boundary, and right boundary of the geographic grid where it is located;

[0050] w represents the flight safety distance between UAVs.

[0051] Preferably, when the UAV currently enters boundary buffer zones of multiple types, the second compensation strategy for compensating the UAV is as follows:

[0052] When the UAV is currently flying within the intersection area of the right boundary defined area and the lower boundary defined area of the geographic grid, the UAV is compensated by the following formula (8):

[0053]

[0054] When the UAV is currently flying within the intersection area of the left boundary defined area and the lower boundary defined area of the geographic grid, the UAV is compensated by the following formula (9):

[0055]

[0056] When the UAV is currently flying within the intersection area of the upper boundary defined area and the right boundary defined area of the geographic grid, the UAV is compensated by the following formula (10):

[0057]

[0058] When the UAV is currently flying in the intersection area of the upper boundary defined area and the left boundary defined area of the geographical grid, the UAV is subjected to zonal compensation through the following formula (11):

[0059]

[0060] In formulas (8)-(11), lon and lat respectively represent the longitude value and latitude value of the geographical coordinates at which the UAV flies.

[0061] The present invention improves the anti-collision warning algorithm, and performs anti-collision warning analysis and calculation on the UAV in a parallel computing manner based on a distributed platform, reducing the computational complexity, reducing the computational time, and improving the warning effect. On this basis, a zonal compensation scheme under boundary conditions is also proposed to perform zonal compensation on the UAV located at the geographical grid boundary, improving the accuracy of collision warning on the premise of meeting real-time warning through distributed parallel computing. Description of the Drawings

[0062] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required to be used in the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and those of ordinary skill in the art can also obtain other drawings based on these drawings without creative efforts.

[0063] Figure 1 is the implementation flowchart of the UAV collision warning method based on distributed parallel computing and zonal compensation provided by the embodiments of the present invention;

[0064] Figure 2 is a schematic diagram of the sphere model constructed for calculating the distance of the UAV;

[0065] Figure 3 is an example diagram of dividing the earth into geographical grids;

[0066] Figure 4 is an example diagram of zonal calculation of the position where the UAV is located;

[0067] Figure 5 is an example diagram of the UAV flying at the boundary of a double area;

[0068] Figure 6 is an example diagram of the UAV flying in the double area buffer;

[0069] Figure 7 is an example diagram of the UAV flying in the four area buffer;

[0070] Figure 8It is the structural diagram of the UAV collision warning platform based on distributed parallel computing and partition compensation provided by the embodiments of the present invention. Detailed implementation manners

[0071] The technical solutions of the present invention will be further described below with reference to the accompanying drawings and through specific implementation manners.

[0072] Among them, the accompanying drawings are only for illustrative purposes, showing only schematic diagrams, not physical diagrams, and should not be construed as a limitation to this patent; in order to better illustrate the embodiments of the present invention, some components in the accompanying drawings will be omitted, enlarged or reduced, which do not represent the dimensions of the actual product; for those skilled in the art, it is understandable that some well-known structures and their descriptions in the accompanying drawings may be omitted.

[0073] In the accompanying drawings of the embodiments of the present invention, the same or similar reference numerals correspond to the same or similar components; in the description of the present invention, it should be understood that if terms such as "upper", "lower", "left", "right", "inner", "outer", etc. are used to indicate the orientation or positional relationship, it is based on the orientation or positional relationship shown in the accompanying drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, so the terms describing the positional relationship in the accompanying drawings are only for illustrative purposes and should not be construed as a limitation to this patent. For those of ordinary skill in the art, the specific meanings of the above terms can be understood according to specific circumstances.

[0074] In the description of the present invention, unless otherwise clearly specified and limited, if terms such as "connection" are used to indicate the connection relationship between components, this term should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or integrated; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two components or the interaction relationship between two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0075] The UAV collision warning method based on distributed computing and partition compensation provided by the embodiments of the present invention has carried out simulation experiments in the Spark distributed platform processing environment. The present invention calculates the partition flag value corresponding to the geographical grid where the current flight position of the UAV is located based on the UAV flight data, and then matches the corresponding machine based on the pre-established association relationship between the partition flag value and the corresponding machine in the distributed cluster. The machine thus matched performs the collision warning analysis calculation for the UAV, realizing the parallel analysis and calculation of the collision warning among many simultaneously flying UAVs, greatly reducing the calculation complexity of the UAV collision warning and shortening the calculation time. On this basis, a compensation function under boundary conditions is also designed to improve the accuracy of the UAV collision warning.

[0076] The UAV collision warning method based on distributed computing and partition compensation provided by the embodiments of the present invention, as Figure 1 shown, includes the steps:

[0077] S1. Collect the flight data of the UAV, including the longitude, latitude and altitude of the flight position;

[0078] S2. Calculate the partition flag value corresponding to the geographic grid currently entered by the UAV according to the collected flight data;

[0079] The geographic grid is in degrees. After the earth is divided into regions according to longitude and latitude, a geographic grid is established, and all positions on the earth are covered by the geographic grid. In this application, the division of the geographic grid based on longitude and latitude preferably adopts a segmentation scheme with a unit of 1 degree. The longitude range of each geographic grid is 1 degree, and the latitude range is also 1 degree. The distance spanned by each degree of longitude is about 111 kilometers, and the distance spanned by each degree of latitude decreases as the latitude increases. In adjacent regions, the shapes of each geographic grid are the same or similar. For example Figure 3 the shapes of adjacent geographic grids P1 and P2 are the same or similar. As Figure 3 shown in the figure, the area is divided into four geographic grid regions P1, P2, P3, and P4 by longitude lines lon1, lon2, lon3 and latitude lines lat1, lat2, lat3.

[0080] Each geographic grid has a corresponding longitude and latitude spatial coverage range. As long as the longitude and latitude information of the current flight of the UAV is collected, it can be determined which geographic grid the UAV currently enters.

[0081] The partition flag value (PartitionID) corresponding to the geographic grid is unique, that is, different geographic grids have a unique corresponding partition flag value. After collecting the longitude, latitude, and altitude of the current flight of the UAV, the partition flag value corresponding to the geographic grid where the UAV is currently located can be uniquely determined. The partition flag value is represented by an integer. The storage and calculation cost of computer integers is smaller than that of floating-point numbers, and the calculation speed is faster, which is beneficial to further improving the calculation speed of UAV distributed collision warning.

[0082] The partition flag value is calculated by the following formula (1):

[0083] PartitionID = Lat × 10000007 + Lon Formula (1)

[0084] In formula (1), Lat represents the longitude value of the left boundary of the geographic grid currently entered by the UAV (such as Figure 3 the longitude value of the left boundary "b1" of geographic grid P1 in the figure). For a schematic diagram of a UAV flying into geographic grid P1, please refer toFigure 4 ;

[0085] Lon represents the latitude value of the upper boundary of the geographical grid into which the UAV currently flies (such as Figure 3 the latitude value of the upper boundary "b2" of the geographical grid P1 in

[0086] Both Lat and Lon take integer values. By multiplying by 10000007, the uniqueness of the partition flag value is ensured.

[0087] Specifically, the UAV is partitioned according to its longitude, latitude, and altitude to calculate the geographical grid where the UAV is located;

[0088] Definition of the geographical grid to which it belongs: The partition flag value is calculated from the longitude, latitude, and altitude of the UAV, that is, the UAV belongs to the geographical grid corresponding to this partition flag value.

[0089] As Figure 4 shown, any UAV needs to be mapped into a geographical grid. The longitude and latitude information of the UAV is a floating-point value with four decimal places reserved. In order to map it into the geographical grid, the following steps are required:

[0090] First, numerical normalization is required. The purpose is to convert the floating-point longitude and latitude values into integer values for convenient subsequent calculations. The normalization formula is as follows:

[0091] lat1 = lat × 10000

[0092] lon1 = lon × 10000

[0093] where lat and lon are the real-time latitude and longitude of the UAV respectively, with four decimal places. After normalization, integer values expanded by 10000 times are obtained. lat1 and lon1 are the longitude value and latitude value after normalization.

[0094] Next, the longitude and latitude are taken modulo 10000 to obtain the integer part of the original longitude and latitude. The integer-taking formula is as follows:

[0095] lat2 = lat1 / 10000

[0096] lon2 = lon1 / 10000

[0097] Finally, a linear transformation function is used to obtain the mapped partition flag value. The transformation coefficient uses 10000007 here, which not only retains sufficient partition granularity for subdivision but also ensures the uniqueness of the partition. The linear transformation function is shown as follows:

[0098] PartitionID = lat2 × 10000007 + lon2

[0099] Through the above steps, any drone with known longitude and latitude information can obtain the partition flag value mapped by the drone.

[0100] S3. Match the machine associated with the partition flag value calculated in step S2.

[0101] After obtaining the unique partition flag value corresponding to the geographical grid into which the drone flies, according to the established association relationship between the partition flag value and the corresponding machine, match the machine used for distributed parallel collision warning analysis and calculation of the drone.

[0102] S4. Send the flight data generated by each drone in the same geographical grid at the current moment to the corresponding machine in the distributed cluster.

[0103] Each drone flying in the same geographical grid has the same PartitionID calculated based on formula (1). Each PartitionID has a unique corresponding machine, and one machine may correspond to multiple different PartitionIDs simultaneously (that is, one machine may be used for drone collision warning analysis and calculation of multiple different geographical grids). The flight data of each drone is sent to the corresponding machine in the distributed cluster.

[0104] S5. The machine calculates the distances between the drones currently flying in the geographical grid based on the received flight data and using the pre-constructed simplified model for calculating the distances between drones, and issues a warning when it determines that the collision warning condition is met.

[0105] The method by which the machine calculates the distances between drones using the simplified model for calculating the distances between drones is expressed by the following formulas (2)-(3):

[0106]

[0107]

[0108] In formulas (2) and (3), l represents the distance between the first drone and the second drone; Δh represents the elevation difference between the first drone and the second drone. represents the distance between the first drone and the second drone on the spherical surface of the earth model as shown in Figure 2 shown;

[0109] R represents the average radius of the earth, taking 6371.393 kilometers.

[0110] Δlat represents the difference in coordinate latitudes between the first drone and the second drone.

[0111] Δlon represents the difference in coordinate longitudes between the first drone and the second drone.

[0112] Lat1 and Lat2 represent the latitude values of the first drone and the second drone respectively.

[0113] When the machine determines that the distance l between the first drone and the second drone is less than the safe distance, it means that there is a collision risk between the two drones, and a collision warning is issued; when it is determined that the distance l is greater than or equal to the safe distance, it means that there is no collision risk between the two drones, and no collision warning is issued. In this application, the safe distance is taken as 10 times the relative flight speed of the first drone and the second drone, that is, 10 seconds of early warning time is reserved. For example, if the relative flight speed is 20 m / s, 200 m is taken as the safe distance.

[0114] Through simulation experiments, as shown in Table a below, for 10,000 drone flights, using the conventional geometric algorithm, the time required to complete the warning is 1.379 s, while the update frequency of the drone's geographical location is usually 1 Hz, which cannot meet the warning time requirement. After adopting the partition parallel computing scheme based on geographical grids in this application, when the number of partitions is 10, the computing time is shortened to 0.137 s, and when the number of partitions is 100, the computing time is further shortened to 0.024 s, which proves that the partition scheme can complete the warning calculation for drone anti-collision within milliseconds, greatly shortening the warning time.

[0115]

[0116] Table a

[0117] Drones located at the intersection of geographical grids may belong to different regions. For example Figure 5 drone A in is in geographical grid P1, and drone B is in geographical grid P2. If the distance between drone A and drone B is less than the safe distance, there is a risk of collision. However, if at this time, the machine corresponding to geographical grid P1 only calculates the distances between the drones currently within geographical grid P1, and the machine corresponding to geographical grid P2 only calculates the distances between the drones currently within geographical grid P2, it will result in the distance between drone A and drone B, which are at risk of collision and are respectively in geographical grids P1 and P2, being missed. For example, as Figure 6As shown, the distance between UAV A flying within geographic grid P1 and UAV B flying within geographic grid P2 is less than the safe distance, and there is a risk of collision between them. However, according to the parallel computing rule of the present application, assuming the partition flag value corresponding to P1 is PartitionID1 and the machine corresponding to PartitionID1 is distributed node 1, then distributed node 1 only calculates the distances between the UAVs flying within P1; assuming the partition flag value corresponding to P2 is PartitionID2 and the machine corresponding to PartitionID2 is distributed node 2, then distributed node 2 only calculates the distances between the UAVs flying within P2. Since the PartitionID values corresponding to UAV A and B are different at this time, the distributed cluster will not calculate the distance between A and B, thus avoiding omission of the collision warning calculation between A and B.

[0118] Therefore, in order to eliminate the collision risk caused by miscalculation in distributed parallel computing, the present invention provides a boundary compensation scheme, and the specific scheme is as follows:

[0119] First, a boundary buffer zone with a width of w (w is the safe distance between UAVs) is established between adjacent geographic grids as a cross-partition definition area. For example, Figure 6 a boundary buffer zone 100 with a width of w is established between adjacent geographic grids P1 and P2 and between P3 and P4 as shown, and a boundary buffer zone 200 with a width of w is established between geographic grids P1 and P3 and between P2 and P4.

[0120] Then, the flight data generated by each UAV is used to determine whether the UAV is currently flying within the boundary buffer zone of the geographic grid.

[0121] If so, further determine the type of the boundary buffer zone that the UAV currently enters, and then perform partition compensation on the UAV according to a preset strategy.

[0122] If not, the machine corresponding to the geographic grid where the UAV currently falls calculates the distance between the UAVs.

[0123] The types of the boundary buffer zone include the upper boundary definition area 10 as shown in Figure 6 (the "10" in Figure 6 is the upper boundary definition area of P4), the lower boundary definition area 20 (the "20" in Figure 6 is the lower boundary definition area of P2), the left boundary definition area 30 (the "30" in Figure 6 is the left boundary definition area of P2), and the right boundary definition area 40 (the "40" in Figure 6 is the right boundary definition area of P3). When the type of the boundary buffer zone that the UAV currently enters is a single type (for example, Figure 6The drone A in it only flies in the area 40 delimited by the right boundary. The first preset strategy for making zonal compensation for the drones falling into the corresponding boundary-delimited areas is as follows:

[0124] When the boundary buffer zone that the drone currently enters is the upper boundary-delimited area of the geographical grid, the drone is zonally compensated through the following formula (4):

[0125]

[0126] When the boundary buffer zone that the drone currently enters is the lower boundary-delimited area of the geographical grid, the drone is zonally compensated through the following formula (5):

[0127]

[0128] When the boundary buffer zone that the drone currently enters is the left boundary-delimited area of the geographical grid, the drone is zonally compensated through the following formula (6):

[0129]

[0130] When the boundary buffer zone that the drone currently enters is the right boundary-delimited area of the geographical grid, the drone is zonally compensated through the following formula (7):

[0131]

[0132] In formulas (4)-(7), lon and lat respectively represent the longitude value and latitude value of the geographical coordinates of the drone's flight;

[0133] S up 、S bottom 、S left 、S right respectively represent the distances from the current flight position of the drone to the upper boundary, lower boundary, left boundary, and right boundary of the geographical grid where it is located;

[0134] w represents the flight safety distance between drones.

[0135] It should be noted here that the longitude increases gradually from left to right, and the latitude increases gradually from bottom to top. For example Figure 6 in, the drone flying in the geographical grid P2 has a larger longitude value than the drone flying in the geographical grid P1, and the drone flying in P1 has a larger latitude value than the drone flying in P3.

[0136] However, when the drone enters the partition junction of the four adjacent geographical grids (as shown in Figure 7 in, the drone A enters the partition junction 300 of P1, P2, P3, and P4), there is a risk of collision with the drones in the diagonal partitions (as shown inFigure 7 Among them, there may be a risk of collision between the drone A in P1 and the drone D in P4). Therefore, it is necessary to further execute a diagonal partition compensation mechanism for the drone A. The second compensation strategy after adding the diagonal partition compensation mechanism to the drone is as follows:

[0137] When the drone is currently flying within the intersection area of the right boundary defined area and the lower boundary defined area of the geographic grid (for example Figure 7 the drone A in) is flying within the intersection area 201), the drone is partition compensated by the following formula (8):

[0138]

[0139] When the drone is currently flying within the intersection area of the left boundary defined area and the lower boundary defined area of the geographic grid (for example Figure 7 the drone B in) is flying within the intersection area 202), the drone is partition compensated by the following formula (9):

[0140]

[0141] When the drone is currently flying within the intersection area of the upper boundary defined area and the right boundary defined area of the geographic grid (for example Figure 7 the drone C in) is flying within the intersection area 203), the drone is partition compensated by the following formula (10):

[0142]

[0143] When the drone is currently flying within the intersection area of the upper boundary defined area and the left boundary defined area of the geographic grid (for example Figure 7 the drone D in) is flying within the intersection area 204), the drone is partition compensated by the following formula (11):

[0144]

[0145] In formulas (8)-(11), lon and lat respectively represent the longitude value and latitude value of the geographic coordinates of the drone's flight.

[0146] It should be noted here that in formulas (4)-(11), the purpose of multiplying lon by 10000007 is: the linear transformation coefficient uses the prime number 10000007, which not only retains sufficient partition granularity subdivision space but also ensures the uniqueness of the partition.

[0147] After completing the partition compensation for each UAV, finally, copy the flight data of the UAV to the distributed node corresponding to the geographical grid where the UAV falls after partition compensation. The distributed node calculates the distance between UAVs for this UAV, solving the problem that the distributed parallel computing method provided by the present invention may miss the calculation objects for collision warning.

[0148] The above steps solve the collision warning within the partition of the UAV under normal conditions (not in the boundary buffer zone) and the collision warning between partitions under special conditions (in the boundary buffer zone). The computing platform for collision warning is as Figure 8 shown, and it consists of three modules: a data access module, a partition scheduling module, and a distributed computing module. The data source is accessed through the Kafka message queue. Then, SparkCore (SparkCore is the core processing unit of the distributed platform) completes the functions of data partitioning and task scheduling. The data partitioning is determined by the partitioning function, and the task scheduling Scheduled by the Yarn scheduler, scheduling policy uses capacity scheduling, that is, the UAV data accessed from different sources on the cloud is divided into multiple queues, and each queue contains multiple computing nodes. After the scheduling is completed, it enters the last part. The data is distributed to the Executor nodes (computing nodes) in the cluster to execute the collision warning algorithm. Each Executor is responsible for all the computing tasks in a region and determines whether to execute the collision warning measures based on the distance between UAVs. It should be noted that the above specific implementation manners are only the preferred embodiments of the present invention and the applied technical principles. Those skilled in the art should understand that various modifications, equivalent replacements, changes, etc. can be made to the present invention. However, as long as these transformations do not deviate from the spirit of the present invention, they should be within the protection scope of the present invention. In addition, some terms used in the specification and claims of this application are not restrictive and are only for the convenience of description.

Claims

1. A method for collision warning of unmanned aerial vehicles based on distributed parallel computing and partition compensation, characterized in that the steps Including: S1, collecting the flight data of the drone; S2, calculating the partition flag value corresponding to the geographical grid into which the drone currently flies according to the collected flight data; S3, matching the machine associated with the partition flag value; S4, sending the flight data generated by each of the drones in the same geographical grid at the current moment to the machine; S5, the machine calculates the distances between the drones currently flying in the geographical grid based on the received flight data and uses a pre-constructed simplified model for calculating the distances between drones, and issues a warning when it is determined that the collision warning condition is met; In step S5, the method steps for calculating the distances between the drones currently flying in the geographical grid include: S51, the machine determines whether the drone is currently flying in the boundary buffer zone of the geographical grid, if so, proceed to step S52; if not, the machine itself calculates the distance between the drones; S52, determining the type of the boundary buffer zone into which the drone currently enters, and then performing corresponding partition compensation on the drone entering the boundary buffer zone of the corresponding type according to a preset strategy; S53, copying the flight data of the drone to the distributed node corresponding to the geographical grid into which it falls after partition compensation, and the distributed node calculates the distance between the drones.

2. The method for UAV collision warning based on distributed parallel computing and partition compensation according to claim 1, wherein The flight data includes the longitude, latitude, and altitude of the flight position.

3. The method for collision warning of unmanned aerial vehicles based on distributed parallel computing and partition compensation according to claim 1, characterized in that The partition flag value is calculated by the following formula (1): PartitionID = Lat × 10000007 + Lon Formula (1) In Formula (1), Lat represents the longitude value of the left boundary of the geographical grid into which the drone currently flies; Lon represents the latitude value of the upper boundary of the geographical grid into which the drone currently flies; 10000007 is a multiplier factor.

4. The method for collision warning of unmanned aerial vehicles based on distributed parallel computing and partition compensation according to claim 3, characterized in that, Lat and Lon are integer values.

5. The method for UAV collision warning based on distributed parallel computing and partition compensation according to claim 1, wherein The method by which the machine calculates the distance between the drones using the simplified model for calculating the distances between drones is expressed by the following formulas (2)-(3): In Formulas (2) and (3), l represents the distance between the first drone and the second drone; Δh represents the elevation difference between the first drone and the second drone; Indicates the distance between the first drone and the second drone on the surface of the spherical Earth model; R represents the average radius of the earth; Δlat represents the coordinate latitude difference between the first drone and the second drone; Δlon represents the coordinate longitude difference between the first drone and the second drone; Lat1 and Lat2 respectively represent the latitude values of the first drone and the second drone.

6. The method for collision warning of unmanned aerial vehicles based on distributed parallel computing and partition compensation according to claim 5, wherein The collision warning condition is that l is less than the product of the relative flight speed of the first drone and the second drone and the warning time, and the warning time is 10 seconds.

7. The method for collision warning of unmanned aerial vehicles based on distributed parallel computing and partition compensation according to claim 1, characterized in that Step S52 specifically includes the steps: S521, determining the type of the boundary buffer zone into which the drone currently enters; S522, determining whether the drone currently enters the intersection area of two boundary buffer zones of different types, if so, perform partition compensation on the drone according to the second preset strategy; Otherwise, perform zonal compensation on the UAV according to the first preset strategy.

8. The method for UAV collision warning based on distributed parallel computing and partition compensation according to claim 1 or 7, characterized in that, In step S52, when the UAV currently enters a boundary buffer zone of a single type, the first preset strategy for performing zonal compensation on the UAV is as follows: When the boundary buffer zone that the UAV currently enters is the upper boundary defined area of the geographical grid, perform zonal compensation on the UAV through the following formula (4): When the boundary buffer zone that the UAV currently enters is the lower boundary defined area of the geographical grid, perform zonal compensation on the UAV through the following formula (5): When the boundary buffer zone that the UAV currently enters is the left boundary defined area of the geographical grid, perform zonal compensation on the UAV through the following formula (6): When the boundary buffer zone that the UAV currently enters is the right boundary defined area of the geographical grid, perform zonal compensation on the UAV through the following formula (7): In formulas (4)-(7), lon and lat respectively represent the longitude value and latitude value of the geographical coordinates of the UAV flight; S up 、 S bottom 、 S left 、 S right respectively represent the distances of the current flight position of the drone from the upper boundary, lower boundary, left boundary, and right boundary of the geographical grid where it is located; w represents the flight safety distance between UAVs.

9. The method for collision warning of unmanned aerial vehicles based on distributed parallel computing and partition compensation according to claim 1 or 7, characterized in that, When the UAV currently enters boundary buffer zones of multiple types, the second compensation strategy for performing zonal compensation on the UAV is as follows: When the UAV is currently flying within the intersection area of the right boundary defined area and the lower boundary defined area of the geographical grid, perform zonal compensation on the UAV through the following formula (8): When the UAV is currently flying within the intersection area of the left boundary defined area and the lower boundary defined area of the geographical grid, perform zonal compensation on the UAV through the following formula (9): When the UAV is currently flying within the intersection area of the upper boundary defined area and the right boundary defined area of the geographical grid, perform zonal compensation on the UAV through the following formula (10): When the UAV is currently flying within the intersection area of the upper boundary defined area and the left boundary defined area of the geographical grid, perform zonal compensation on the UAV through the following formula (11): In formulas (8)-(11), lon and lat respectively represent the longitude value and latitude value of the geographical coordinates of the UAV flight.

Images