KR102067363B1 - Method for predicting and avoiding collision and conflict/conjunction among moving objects - Google Patents

Abstract

Disclosed are a collision prediction and avoidance method between moving objects. A collision prediction and avoidance method between moving objects includes: calculating at least two or more collision prediction objects by calculating a Voronoi diagram between a plurality of moving objects; Generating S1 which is a copy of any one of the collision prediction objects and placing the S1 at the same point as the one collision prediction object; Changing at least one of the moving direction and the moving speed of the S1 differently from the moving direction and the moving speed of the collision predicted object; Calculating whether S1 collides with the remaining collision predicted objects; And when it is determined that S1 does not collide with the remaining collision prediction objects, changing the moving direction and the moving speed of the one collision prediction object to the moving direction and the moving speed of the S1.

Description

Method for predicting and avoiding collision and conflict / conjunction among moving objects}

The present invention relates to a collision prediction method and a method of avoiding proximity between a plurality of moving objects, and more particularly, to model a plurality of moving objects in a two-dimensional or three-dimensional geometric shape, and a Voronoi diagram for the geometric shapes. And predicting and avoiding collisions and proximity of the geometric shapes using dynamic data of the geometric shapes and phase information of the Voronoi diagram.

Similar to many casualties caused by traffic accidents on land, collisions frequently occur between ships and many casualties occur. In addition, collisions often occur between aircraft, resulting in very high death rates. A typical example was the air collision in Germany over Uberlingen on July 1, 2002, killing all 71 passengers. There are also occasional collisions between planes moving on the runway from the ground. In recent years, the use of drones such as drones has become frequent, resulting in frequent collisions between drones and airliners or drones and helicopters (e.g., Quebec, Canada, drone / passenger collisions on October 16, 2017). As such, failure to accurately predict collisions or collisions between aircraft or vessels can lead to massive casualties.

As a method of predicting collision and proximity between a plurality of moving objects, a method of sampling the positions of objects in pairs at regular time intervals and checking whether the sampled positions intersect are used. However, this method has a risk of not detecting it even though there is a real proximity risk depending on the sampling method and frequency, and the amount of calculation must increase rapidly to reduce the risk. For example, when the number of sampling is M for N moving objects, the calculation amount of the conventional method is O (N ² M ² ). At this time, even if the frequency of sampling is slightly increased, the amount of calculation increases rapidly. Therefore, a large-scale supercomputer is required for the prediction of collision and proximity of moving objects.

Segment Tree [Bentley JL, Wood D. An optimal worst case algorithm for reporting intersections of rectangles. IEEE Transactions on Computers. 1980 Jul 1 (7): 571-7.] Is an efficient way to determine the collision of two moving objects (O (N log N) time to create a binary tree and to find the two nearest moving objects). Average time O (log N) hours). However, this method is specialized only for collisions between two moving bodies and depends on the coordinate system.

On the other hand, the method proposed in the present invention is a method independent of the coordinate system, and can efficiently analyze the proximity of two or three as well as any number of moving objects (preprocessing for collision and proximity prediction). O (N log N) time for) and the average O (log N) time for each subsequent event processing).

In addition, the method proposed in the present invention replays the situation of the collision and proximity through the event history for collision avoidance and other analysis even after the calculation for prediction of the collision and proximity is completed. This is possible but difficult to reproduce through the above method. Rapid reproducibility is an important feature in that you can repeat various analyzes for future situations.

In addition, there has been a study on how to calculate the Voronoid diagram for the moving point set without considering the size of each moving object [Roos T. Voronoi diagrams over dynamic scenes. Discrete Applied Mathematics. 1993 Jun 10; 43 (3): 243-59.]. However, this method is difficult to accurately predict the collision and proximity because the size of the moving body can not be accurately reflected.

The present invention provides a method for predicting and avoiding collisions and proximity of multiple moving bodies moving linearly with a small amount of calculation.

In addition, the present invention provides a method for predicting and avoiding collisions and proximity of a plurality of moving bodies by approximating the movement of a plurality of moving bodies moving in a nonlinear manner to a plurality of linear movements.

In addition, the present invention provides a method for accurately determining the collision and proximity of the moving objects without missing.

According to another aspect of the present invention, there is provided a collision prediction and avoidance method comprising: calculating at least two or more collision prediction objects by calculating a Voronoi diagram between a plurality of moving objects; Generating S1 which is a copy of any one of the collision prediction objects and placing the S1 at the same point as the one collision prediction object; Changing at least one of the moving direction and the moving speed of the S1 differently from the moving direction and the moving speed of the collision predicted object; Calculating whether S1 collides with the remaining collision predicted objects; And when it is determined that S1 does not collide with the remaining collision prediction objects, changing the moving direction and the moving speed of the one collision prediction object to the moving direction and the moving speed of the S1.

The method may further include generating S2, which is a copy of the collision prediction object, and placing the same at the same point as the collision prediction object; Changing at least one of the moving direction and the moving speed of the S2 differently from the moving direction and the moving speed of the collision predicted object and the moving direction and the moving speed of the S1; Calculating collision with the remaining collision prediction objects using the Voronoi diagram of each moment previously calculated for each of S1 and S2; The collision calculation step using the pre-calculated Voronoi diagram may be applied to each of S1 and S2 independently to perform parallel processing; And when it is determined that S1 and S2 do not collide with the remaining collision predicted objects, one of the moving direction and the moving speed of S1 and the moving direction and the moving speed of S2 is selected as the optimal solution, and the It is possible to change the moving direction and the moving speed of either collision prediction object.

According to the present invention, since the collision and proximity are determined only for the moving objects generating the flipping event and the collision event by applying the Voronoi diagram, the calculation amount for the collision and proximity prediction of the plurality of moving objects can be minimized.

In addition, according to the present invention, when the two-dimensional moving objects collide necessarily have the edge of the Voronoi diagram, and when the three-dimensional moving objects collide necessarily have the face of the Voronoi diagram, the edge of the Voronoi diagram (two-dimensional moving body Identifying objects that define collisions and proximity prediction problems or planes of Voronoi diagrams (collisions and proximity prediction problems between three-dimensional moving objects) can accurately predict the actual collision.

Further, according to the present invention, when three two-dimensional moving objects are close to each other, the vertices of the Voronoi diagram are moved between the moving objects, and when four or more moving objects are close to each other in two dimensions, the substructure of the Voronoi diagram is shown. Can be defined. In two dimensions, the Voronoi diagram consists of the vertices, edges, and faces of the Voronoi diagram (areas that are open or closed in a polygonal region), and the substructure of the Voronoi diagram is described above. A subset of Voronoi diagrams that include at least one of the vertices, edges, and faces of a noise diagram.

In the case of a three-dimensional moving object, similarly to two-dimensional, when three three-dimensional moving objects are close to each other, they have an edge of the Voronoi diagram between the moving objects. With the vertices of, the substructure of the Voronoi diagram can be defined even when five or more moving bodies are in close proximity. In three dimensions, the Voronoi diagram consists of the vertices, edges, faces, and cells of the Voronoi diagram, which can be opened or closed with a polyhedron bounded by the faces of the Voronoi diagram. The underlying structure of the noise diagram refers to a subset of boronogram diagrams including at least one of the vertices, edges, faces, and cells of the Voronoi diagram described above.

In this way, if the Voronoi diagram between the moving bodies is calculated, the collision between two moving bodies as well as the proximity between three, four or more moving bodies can be accurately and efficiently predicted. Therefore, the method proposed in the present invention can perform the analysis of the proximity of two or three as well as any number of moving objects most efficiently.

1 is a view showing a system for predicting and avoiding collision and proximity between moving bodies according to an embodiment of the present invention.
2 to 4 are diagrams illustrating a method in which the moving object modeling unit models the moving object.
5 is a flowchart illustrating a method of predicting and avoiding collision and proximity between different moving bodies in the first embodiment of the present invention.
6 is a diagram illustrating a step of calculating corners of a Voronoi diagram by calculating a Voronoi diagram between moving disks.
7 is a diagram illustrating a flipping event according to an embodiment of the present invention.
8 is a diagram illustrating a collision event according to an embodiment of the present invention.
FIG. 9 is a diagram for explaining information of a Voronoi diagram updated after elapse of a flipping event and a collision event time.
10 is a diagram illustrating a step of determining whether collisions between objects occur.
11 is a flowchart illustrating a method of predicting collision and proximity between moving bodies according to a second embodiment of the present invention.
12 is a diagram illustrating a step of calculating a plane, an edge, and a vertex of a Voronoi diagram by calculating a Voronoi diagram between moving spheres.
FIG. 13 illustrates an edge flipping event and a face flipping event according to the present invention.
FIG. 14 is a diagram illustrating a method of predicting a collision and proximity of a drone and providing an avoiding path according to an embodiment of the present invention.
FIG. 15 is a diagram for describing a method of predicting a collision and proximity of a drone and providing an avoiding path according to another embodiment of the present invention.
FIG. 16 is a diagram for describing a method of predicting collision and proximity of moving bodies for which a moving path is known according to one embodiment of the present invention.
17 is a diagram for describing a method of determining whether or not a moving object collides with each other.
FIG. 18 is a flowchart illustrating a method of predicting collision of moving objects for which a moving path is known according to another embodiment of the present invention.
19 is a view for explaining in detail the first collision determination step of the moving spheres described in FIG.
20 is a view for explaining in detail the second collision determination step of the moving spheres described in FIG.
FIG. 21 is a diagram for describing a third collision determination step of the moving spheres illustrated in FIG. 18 in detail.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. However, the technical idea of the present invention is not limited to the exemplary embodiments described herein and may be embodied in other forms. Rather, the embodiments introduced herein are provided to ensure that the disclosed contents are thorough and complete, and that the spirit of the present invention can be sufficiently delivered to those skilled in the art.

In the present specification, when a component is mentioned to be on another component, it means that it may be formed directly on the other component or a third component may be interposed therebetween. In addition, in the drawings, the thicknesses of films and regions are exaggerated for effective explanation of technical contents.

In addition, in various embodiments of the present specification, terms such as first, second, and third are used to describe various components, but these components should not be limited by these terms. These terms are only used to distinguish one component from another. Thus, what is referred to as a first component in one embodiment may be referred to as a second component in another embodiment. Each embodiment described and illustrated herein also includes its complementary embodiment. In addition, the term 'and / or' is used herein to include at least one of the components listed before and after.

In the specification, the singular encompasses the plural unless the context clearly indicates otherwise. In addition, the terms "comprise" or "having" are intended to indicate that there is a feature, number, step, element, or combination thereof described in the specification, and one or more other features or numbers, steps, configurations It should not be understood to exclude the possibility of the presence or the addition of elements or combinations thereof. In addition, the term "connection" is used herein to mean both indirectly connecting a plurality of components, and directly connecting.

In addition, in the following description of the present invention, if it is determined that a detailed description of a related known function or configuration may unnecessarily obscure the subject matter of the present invention, the detailed description thereof will be omitted.

The present invention provides a method for estimating a danger due to a proximity or collision situation between moving objects (hereinafter referred to as 'moving body') and avoiding a collision. Vehicles are all high speed or low speed vehicles, including aircraft, ships, submarines, satellites, drones, automobiles, etc., and their modes of operation include manned or unmanned. In addition to moving objects such as satellites orbiting around the earth, moving bodies are not only constant in moving paths such as aircraft, autonomous cars, drones, submarines, ships, etc. Contains objects that can be. The moving path of the moving body may include a linear path and a nonlinear path. According to the present invention, the nonlinear path can be approximated into a plurality of linear paths. The number of linear paths to be approximated is determined according to the user's setting, and it is assumed that the moving bodies move along the approximated linear paths. The moving body may move in a constant velocity linear motion in each section of the linear path and the linear paths described above.

In the present invention, in order to predict the collision between the moving objects and to avoid the collision, dynamic data of the moving object is required. The dynamic data may be measured through a global navigation satellite system (GNSS), and an appropriate method. Can be collected by

According to the embodiment, the dynamic data measured using the satellite navigation system is information including the instantaneous position and the moving speed of the moving object, the aircraft uses the Automatic Dependent Surveillance Broadcast (ADS-B), the vessel is the Automatic Identification Can be collected using System. Drones can be collected using wireless technology, such as Wi-Fi, or by using ADS-B. Autonomous vehicles can be collected using V2X, which is a communication link of a vehicular ad hoc network (VANET) that enables communication between vehicles.

1 is a diagram illustrating a system for predicting and avoiding collision and proximity between moving bodies according to an exemplary embodiment of the present invention, and FIGS. 2 to 4 are views illustrating a method for modeling a moving body by the moving modeling unit. 2 to 4 represent the moving object in two dimensions for convenience of understanding, it can be easily understood on the same principle in three-dimensional space.

First, referring to FIG. 1, the system 100 for predicting and avoiding collision and proximity between moving bodies includes a moving object information receiving unit 110, a moving object modeling unit 120, and a collision predicting unit 130.

The moving object information receiving unit 110 receives information from each moving object. The received information may include the type, shape, size, weight, moving path, instantaneous position, moving direction and speed of the moving object. Hereinafter, the moving path, the instantaneous position, the moving direction, and the speed of the moving object are called dynamic data of the moving object.

The moving object modeling unit 120 receives the information of the moving objects from the moving object information receiving unit 110 and models the moving objects in a predetermined shape. The moving object modeling unit 120 may model in two or three dimensions according to the type of moving object. According to the embodiment, since the ship and the vehicle can be expressed in the two-dimensional plane, it can be modeled in a two-dimensional shape. Aircrafts, drones, submarines, and satellites, on the other hand, can be modeled in three-dimensional shapes because they can represent movement in three-dimensional space. Of course, ships and cars can be modeled in three-dimensional shapes, but data throughput can be reduced by modeling in two-dimensional shapes.

Referring to FIG. 2, the moving object modeling unit 120 may generate two-dimensional or three-dimensional objects 11 by modeling the moving bodies 10 from the shape information of the moving bodies 10. The created objects 11 have a simplified geometric shape than the actual shape of the moving bodies. According to an embodiment, the vessel 10 may be modeled as an elliptical two-dimensional object 11.

The moving object modeling unit 120 may model the objects 11 as two-dimensional circles or three-dimensional spheres 12 from the generated size information of the objects 11. According to an embodiment, objects 11 of ships and cars may be modeled as two-dimensional circles 12, and objects 11 of aircraft, drones, submarines, and satellites may be modeled as three-dimensional spheres 12. have. The moving object modeling unit 120 may model the circumferential circle having the smallest diameter circumscribing the objects 11 as the three-dimensional sphere 12 or the circumferential sphere 12 having the smallest diameter circumscribing the objects 11.

The moving object modeling unit 120 may model the two-dimensional circle or the three-dimensional sphere 12 as the moving disk or the moving sphere 13 by reflecting uncertainties in the instantaneous positions and moving speeds of the moving bodies 10. Specifically, the moving object modeling unit 120 uses a moving disk using at least one of moving speeds of the moving objects 10, monitoring time intervals of the moving objects 10, and size information of the two-dimensional circles or the three-dimensional spheres 12. Alternatively, the moving sphere 13 may be modeled.

According to an embodiment, the moving object modeling unit 120 may model the two-dimensional circle 12 as a two-dimensional moving disk 13 by using Equation 1 below.

[Equation 1]

R = r + V _max t _p + δ

(Where R is the radius of the moving disks 13, r is the radius of the two-dimensional circle 12, V _max is the maximum velocity of the moving object 10, t _p is the prediction of collisions and proximity of the moving objects 10). Calculation time, δ is the safety factor)

According to another embodiment, the moving object modeling unit 120 may model the three-dimensional sphere 12 as the moving sphere 13 by using Equation 2 below.

[Equation 2]

R = r + V _max t _p + δ

(Where R is the radius of the moving spheres 13, r is the radius of the three-dimensional sphere 13, V _max is the maximum velocity of the moving object 10, t _p is the prediction of collisions and proximity of the moving objects 10). Calculation time, δ is the safety factor)

In the above-described information, and t _p is the prediction know the number of the moving object as the object of the collision, and close-up relationship, the number of the moving object is determined when the predetermined t _p. That is, the prediction time can be obtained by knowing the number of moving objects to be predicted. Alternatively, t _p may be appropriately assigned by the user.

Referring to FIG. 3, the moving object modeling unit 120 may model the two-dimensional objects 12 into a plurality of inscribed circles 21 to 25 inscribed therein. Specifically, the boundary line of the objects 12 is composed of a plurality of boundary region edges and boundary region vertices where they meet, and the edge of the Voronoi diagram is applied by applying the Voronoi diagram to the boundary region edges and the boundary region vertices. Create the vertices of the Voronoi diagram where they meet. Then, among the vertices of the Voronoi diagram, a vertex having the largest radius of contact with the boundary lines of the objects is selected, and a first inscribed circle 21 having the largest radius is created based on the vertex. Next, second inscribed circles 22 and 24 having a next larger radius and not overlapping with the first inscribed circle 21 are generated. In this order, the two-dimensional objects 12 may be modeled into a plurality of inscribed circles 21 to 25.

The moving object modeling unit 120 may model the three-dimensional objects 12 as a plurality of internal spheres 21 to 25 inscribed therein. The moving object modeling unit 120 may model the internal spheres 21 to 25 through the boundary area data of the 3D objects 12. Specifically, by applying the Voronoi diagram to the boundary region data of the three-dimensional objects 12, the first inscribed sphere 21 having the largest radius of contact with the boundary regions of the objects, and the next inscribed sphere having the next largest radius. Second inscribed spheres 22 and 23 that do not overlap with (21) are generated. In this way, the three-dimensional objects 12 can be modeled as a plurality of inscribed spheres 21 to 25 inscribed therein.

Referring to FIG. 4, when the moving object modeling unit 120 knows the turning position and angle of the moving body on the linear moving path in which the nonlinear moving path is approximated, the moving object modeling unit 120 may change the turning direction and the object 12 at the turning position and angle. It can be modeled as a geometry 14 that includes the numbers 12, 12a, 12b of all cases present. According to the embodiment, when the ship changes the direction, it is assumed that the direction can be changed only within a certain angle ± θ with respect to the existing traveling direction (V1) based on the center of gravity (C), the moving body modeling unit 120 ) May model the object as a geometric shape 14 including both the object 12 in the existing direction, the object 12a turned to + θ, and the object 12b turned to −θ.

In addition, the moving object modeling unit 120 may model a plurality of inscribed circles or inscribed spheres (not shown) that inscribe the geometric shapes 14. Modeling of the plurality of inscribed circles or inscribed spheres inscribed in the geometric shapes 14 may proceed in the same manner as described in FIG. 3.

The collision predicting unit 130 calculates a Voronoi diagram between the moving disks or the moving spheres, and predicts the collision and the proximity between the moving disks or the moving spheres. The collision predicting unit 130 predicts collision and proximity between moving disks or moving spheres in the monitoring time intervals t ₀ , t ₀ + Δt of the moving objects.

According to the first embodiment, the collision prediction unit 130 calculates the Voronoi diagram between the moving disks to calculate the corners of the Voronoi diagram. The collision prediction unit 130 calculates a flipping event and a collision event. A flipping event is an event in which at least one of the corners of a Voronoi diagram is converted back to an edge through a vertex, and a collision event is an event in which a collision is predicted between a pair of moving disks defining an edge of the Voronoi diagram. . The collision predicting unit 130 generates a collision event when it is predicted that a pair of moving disks defining an edge of the Voronoi diagram collides using dynamic data of the moving disks.

There may be several occurrences of the flipping event and the collision event, and the collision prediction unit 130 schedules the flipping event and the collision event in a time sequence in which the occurrence time is high.

The collision prediction unit 130 calculates whether or not an actual collision occurs between the mobile disks generating the event in the order of rapid occurrence of the event. The collision prediction unit 130 calculates whether or not the collision is performed by using dynamic data of the mobile disks. As a result of the calculation, it may be determined that an actual collision does not occur or a collision does not occur. When it is determined that an actual collision occurs, the collision prediction unit 130 modifies the dynamic data of the collision prediction moving disks so as to avoid collisions between the moving disks.

The collision prediction unit 130 updates the information of the Voronoi diagram after the elapse of the corresponding event time.

In more detail, the collision prediction unit 130 may update the following three pieces of information after the flipping event occurs.

A-1-1) the collision prediction unit 130 updates the topology information on the corners of the Voronoi diagram where the flipping event is generated and the corners of another Voronoi diagram connected thereto. The collision prediction unit 130 redefines the Voronoi diagram for the corner of the Voronoi diagram where the flipping event is generated and the corners of the other Voronoi diagram connected thereto.

A-1-2) the collision prediction unit 130 recalculates an occurrence time of the flipping event with respect to the corner of the Voronoi diagram where the flipping event occurs and the edges of another Voronoi diagram connected thereto.

A-1-3) the collision prediction unit 130 recalculates an occurrence time of a collision event between a pair of moving disks defining an edge of the Voronoi diagram in which the flipping event is generated.

In addition, the collision prediction unit 130 may update the following two pieces of information after the collision event occurs.

A-2-1) the collision prediction unit 130 collides with each other a pair of moving disks generating the collision event and the other moving disks defining the pair of moving disks and the corners of the Voronoi diagram. Recalculate the event occurrence time.

A-2-2) collision prediction unit 130 is a corner of a Voronoi diagram defined by a pair of mobile disks generating the collision event, and a Voronoi defined by the pair of mobile disks and other mobile disks. Recalculate the flipping event occurrence time for the corners of the diagram and the corners of the Voronoi diagram defined by the other moving disks.

After updating the above information of the Voronoi diagram, the collision prediction unit 130 schedules the recalculated flipping event and the recalculated collision event in the order of occurrence in the fastest time, and generates the events in the order of the occurrence occurring in the fastest time. Calculate the actual collision between the moving disks, and if it is determined that the actual collision occurs, modify the dynamic data of the mobile disks to avoid the collision.

The collision prediction unit 130 monitors the information of the flipping event and the Voronoi diagram in which the collision event is generated, the calculation of the actual collision between the mobile disks with updated information, and the collision avoidance process. ₀ , t ₀ + △ t) is repeated sequentially, and the monitoring time interval (t ₀ , t ₀ + △ t) of the moving object has elapsed, or flipped in the corresponding monitoring time interval (t ₀ , t ₀ + △ t) If neither an event nor a conflict event is found, the conflict calculation is terminated.

According to the second embodiment, the collision prediction unit 130 defines a Voronoi diagram between the moving spheres to calculate the plane, edge, and vertex of the Voronoi diagram. The collision prediction unit 130 calculates an edge flipping event, a face flipping event, and a collision event. The edge flipping event is an event in which the edges of the Voronoi diagram are converted into triangle-shaped faces through the vertices, and the face flipping event is an event in which the faces of the triangle-shaped Voronoi diagram are converted to edges through the vertices. Collision events are events in which collisions are predicted between a pair of moving spheres that define the faces of all types of Voronoi diagrams as well as triangles. The collision prediction unit 130 uses dynamic data of moving spheres, and generates a collision event when it is predicted that the pair of moving spheres collide.

An edge flipping event, a face flipping event, and a collision event may be generated several times, and the collision prediction unit 130 schedules the event occurrence time in a time order.

The collision prediction unit 130 calculates whether or not the actual collision between the moving spheres generating the event in the order of the rapid occurrence of the event. As a result of the calculation, it may be determined that an actual collision does not occur or a collision does not occur. When it is determined that the actual collision occurs, the collision prediction unit 130 modifies the dynamic data of the collision prediction moving spheres so as to avoid the collision.

The collision prediction unit 130 updates the information of the Voronoi diagram after the elapse of the corresponding event time.

In more detail, after the edge flipping event is generated, the collision prediction unit 130 may update the following four pieces of information.

B-1-1) the collision prediction unit 130 updates the phase information of the Voronoi diagram in which the edge flipping event is generated. The collision prediction unit 130 redefines the Voronoi diagram for the moving spheres defining the faces of the Voronoi diagram after the edge flipping event is generated and the corners of the Voronoi diagram that abut the vertices constituting the corner flipping event. .

B-1-2) The collision prediction unit 130 performs an edge flipping event with respect to the corners of the Voronoi diagram that abuts vertices of the triangular Voronoi diagram after the edge flipping event occurs. Recalculate the occurrence of. However, if the edge of the Voronoi diagram is included in the face of the other triangle Voronoi diagram, it is excluded from the calculation.

B-1-3) the collision prediction unit 130 recalculates an occurrence time of the plane flipping event with respect to the plane of the triangle-shaped Voronoi diagram newly generated by the occurrence of the edge flipping event. If there is a face of another triangle-shaped Voronoi diagram that is in contact with a vertex of the face of the triangular Voronoi diagram, the occurrence time of the face flipping event is recalculated for this face.

B-1-4) Recalculate collision event occurrence between the moving spheres that define the faces of all newly created Voronoi diagrams in B-1-2) above. Recalculation of collision event occurrences covers not only faces of triangular Voronoi diagrams, but also faces of other forms of Voronoi diagrams.

The collision prediction unit 130 may update the following four pieces of information after the face flipping event is generated.

B-2-1) the collision prediction unit 130 updates the phase information of the Voronoi diagram in which the plane flipping event is generated. The collision prediction unit 130 redefines a Voronoi diagram for the moving spheres that define the corners of the Voronoi diagram that abut the edges of the Voronoi diagram after the face flipping event occurs.

B-2-2) The collision prediction unit 130 recalculates the occurrence of the edge flipping event with respect to the edge of the Voronoi diagram where the face flipping event is generated and the edge of the Voronoi diagram connected thereto. However, if the edge of the Voronoi diagram is included in the face of the other triangle Voronoi diagram, it is excluded from the calculation.

B-2-3) The collision prediction unit 130 recalculates the occurrence of the plane flipping event for the planes of the Voronoi diagram connected to the edge of the Voronoi diagram in which the plane flipping event is generated.

B-2-4) The collision prediction unit 130 may determine the collision event between the moving spheres defining the plane of the Bonoroloy diagram generated after the recalculation of the edge flipping event occurrence according to B-2-2). Recalculate the occurrences.

In addition, the collision prediction unit 130 may update the following two pieces of information after a collision event occurs.

B-3-1) collision prediction unit 130 generates collision event generation for the pair of moving spheres generating the collision event and other moving spheres defining the pair of moving spheres and the surfaces of the Voronoi diagram. Recalculate

B-3-2) the collision prediction unit 130 is a plane of the Voronoi diagram that abuts vertices of the Voronoi diagram included in the planes of the Voronoi diagram defined by the pair of moving spheres and the other moving spheres. Or recalculate edge flipping events and face flipping events for the edges of the Voronoi diagram.

The collision prediction unit 130 schedules the recomputed edge flipping event, the surface flipping event, and the collision event in the order of occurrence time in the earliest order, and generates the events in the order of occurrence of the occurrence time in the earliest occurrence time. If it is determined that the actual collision occurs, and if it is determined that the actual collision occurs, modify the dynamic data of the collision prediction movement sphere to avoid the collision.

The collision prediction unit 130 repeats the update of the information of the Voronoi diagram described above, the calculation of the actual collision between the moving spheres with updated information, and the collision avoidance process sequentially during the monitoring time interval of the moving bodies, and the monitoring of the moving bodies. When the time interval elapses, or when the edge flipping event, the face flipping event, and the collision event are not found during the time interval, the actual collision prediction between the moving spheres is terminated.

Hereinafter, a method of predicting and avoiding collision and proximity between moving objects using a system for predicting and avoiding collision and proximity between a plurality of moving objects will be described. A method of predicting and avoiding collision and proximity between a plurality of moving objects is a first embodiment in which a plurality of moving objects are modeled in two dimensions and predicting a collision, and a second embodiment in which a plurality of moving objects is modeled in three dimensions and predicting a collision. The description will be made with an example.

5 is a flowchart illustrating a method of predicting and avoiding collision and proximity between a plurality of moving objects according to a first embodiment of the present invention.

Referring to FIG. 5, in the method of predicting and avoiding collision and proximity between a plurality of moving objects, generating objects by modeling shapes of a plurality of moving objects (S110), the object from the size information of the objects Modeling them to generate a two-dimensional circle (S120); Modeling the two-dimensional circles as moving disks by using at least one of a moving speed of the moving bodies, a monitoring time interval of the moving bodies, and size information of the two-dimensional circles (S130), and between the moving disks. Computing the noise diagram to calculate the corners of the Voronoi diagram (S140), and during the monitoring time interval of the moving object, calculating the actual collision between the moving disk (S150).

2 and 5, the generating of the objects 11 by modeling the shapes of each of the plurality of moving objects 10 (S110) may generate the objects 11 through the shape information of the moving objects 10. do.

Generating the two-dimensional circle 12 by modeling the objects 11 from the size information of the objects (11) (S120) is a two-dimensional circle 12 of the smallest diameter to circumscribe each of the objects (11) Create them.

The step S130 of modeling the two-dimensional circles 12 as moving disks 13 includes a moving speed of the moving bodies 10, a monitoring time interval of the moving bodies 10, and the two-dimensional circles 12. The two-dimensional circles 12 are modeled as moving disks 13 using at least one of the size information. According to an embodiment, the two-dimensional circles 12 may be modeled as the moving disks 13 by using Equation 1 described above.

6 is a diagram illustrating a step of calculating corners of a Voronoi diagram by calculating a Voronoi diagram between moving disks.

5 and 6, the calculating of the corners of the Voronoi diagram by calculating the Voronoi diagram between the moving disks may include calculating the Voronoi diagram between the plurality of moving disks 13a to 13g. Then, the edges (E1, E2, E3, ...) and vertices (V1, V2, ...) of the Voronoi diagram are calculated. The edges of the Voronoi diagram (E1, E2, E3, ...) are created between a pair of adjacent moving disks (13a to 13g), and the vertices (V1, V2, ...) of the Voronoi diagram are at least three Voronoi diagrams. Edges (E1, E2, E3, ...) meet and are generated.

Computing the actual collision between the mobile disk (13a to 13g) (S150), during the monitoring time interval of the moving object, calculates the flipping event and the collision event occurrence, and schedules the event occurrence time in the order of rapidity.

7 is a diagram illustrating a flipping event according to an embodiment of the present invention.

First, referring to FIG. 7A, the edge E2 of the Voronoi diagram is generated by the two moving disks 13b and 13c at the first time point t ₀ . Referring to (B), the edge E2 of the Voronoi diagram is converted from the second time point t ₀ <t ₁ <t ₀ + Δt to the vertex V _f as the moving disk 13a moves. . Referring to (C), the vertex V _f of the Voronoi diagram switches from the third time point t ₁ <t ₂ <t ₀ + Δt to the edge E4 as the moving disk 13a moves. do. As such, a flipping event occurs in which the edge E2 of the Voronoi diagram is converted back to the edge E4 through the vertex V _{f as the} moving disk 13a moves.

8 is a diagram illustrating a collision event according to an embodiment of the present invention.

Referring to FIG. 8A, the edge E2 of the Voronoi diagram is generated by the pair of moving disks 13a and 13c at the first time point t ₀ . Referring to (B), the collision of the two moving disks 13a and 13c at the second time point t ₀ <t ₁ <t ₀ + Δt according to the dynamic data of the pair of moving disks 13a and 13c This is predicted and a collision event is generated.

Although one flipping event and a collision event are shown in the figure, a plurality of flipping events and a collision event may occur simultaneously or in time series according to the number of moving disks and dynamic data.

The step of calculating the actual collision between the mobile disks 13a and 13c calculates whether or not the mobile disks 13a and 13c collide with each other in order of generating the corresponding event in the order of rapid event occurrence time. The calculation of the collision uses dynamic data of the moving disks 13a and 13c, and it can be determined that an actual collision does not occur or a collision does not occur. If it is determined that a collision occurs, the dynamic data of the collision prediction disks 13a and 13c are modified to avoid the collision.

FIG. 9 is a diagram for explaining information of a Voronoi diagram updated after elapse of a flipping event and a collision event time.

Referring to FIG. 9A, the information of the Voronoi diagram updated after the elapse of the flipping event time is the same as the above-described A-1-1) to A-1-3).

Specifically, the phase information is updated with respect to the edge E4 of the Voronoi diagram in which the corresponding event is generated and the edges E1, E3, E5, and E6 of the other Voronoi diagram connected to the flipping event. .

The time of occurrence of the flipping event is recalculated for the edge E4 of the Voronoi diagram and the edges E1, E3, E5, and E6 of the other Voronoi diagram connected thereto.

In addition, the time of occurrence of the collision event between the pair of moving disks 13a and 13d is recalculated with respect to the edge E4 of the Voronoi diagram.

Referring to FIG. 9B, the information of the Voronoi diagram updated after the collision event time elapses is the same as the above-described A-2-1) and A-2-2). Specifically, as the collision event time elapses, the moving disks 13a and 13c defining the edge E2 of the Voronoi diagram and the edge E2 of the Voronoi diagram are different from the other disks 13b, 13d to 13f. Recalculate the collision event occurrence time between the moving disks (13a to 13f) for the edges (E1, E3 to E7) of the Voronoi diagram, defined together.

And the edges of the Voronoi diagram defined by the moving disks 13a, 13c defining the edges E2 of the Voronoi diagram and the edges E2 of the Voronoi diagram together with the other disks 13b, 13d-13f. (E1, E3 through E7) and the edges E8, E9, E10 and E11 of the other Voronoi diagrams connected to them are recalculated for the flipping event occurrence time.

Computing the actual collision between the mobile disk (S150), after updating the information of the Voronoi diagram described above, scheduling the re-calculated flipping event and the re-calculated collision event in the order of occurrence time, It calculates the actual collision between the moving disks that generate the events in the fast order, and when it determines that the actual collision occurs, the dynamic data of the moving disks are corrected to avoid the collision.

The calculating of the actual collision between the mobile disks (S150) may include: updating the information of the Voronoi diagram in which the above-mentioned flipping event and the collision event have occurred, and calculating whether or not the actual collision between the mobile disks with updated information is performed; the monitoring time period for collision avoidance process _{(t 0, t 0 + △} t) sequentially and repeatedly, the moving object monitoring time period of _{(t 0, t 0 + △} t) has elapsed, or the monitoring time interval (t ₀ for If no flipping event or collision event is found at t ₀ + Δt), the collision calculation is terminated.

On the other hand, the method of predicting and avoiding collisions and proximity between a plurality of moving objects is a moving disk 13a to 13d in which a collision is predicted when an actual collision is calculated between the moving disks 13a to 13d generating the above-described event. Determining whether the objects of each other collide with each other. By determining whether the objects collide with each other, the possibility of the actual collision of the moving object is verified.

The determining of the collision between the objects according to an embodiment of the present disclosure calculates whether the objects actually collide from the moving direction and the moving speed of each of the moving disks during the monitoring time of the moving objects.

10 is a diagram illustrating a step of determining whether collisions between objects occur. Referring to FIG. 10, the method of determining whether or not the objects collide with each other may include modeling the objects 12a using a plurality of inscribed circles 21a to 25a inscribed in each of the objects 12a by the method described with reference to FIG. 3. The collision is calculated in units of inscribed circles 21a to 25a. Judging whether the collision in the inscribed circle (21a to 25a) unit, applying the moving direction and the moving speed (V1) of the moving object to each of the inscribed circle (21a to 25a) to predict the movement path of the inscribed circle (21a to 25a), other The collision path between the inscribed circles 21a to 25a and 21b to 25b of the two objects 12a and 12b is estimated by estimating the movement path of the object 12b and its inscribed circles 21b to 25b according to the moving information V2 of the moving object. To judge. When any one of the inscribed circles 21a to 25a constituting one object 12a collides with the inscribed circles 21b to 25b constituting the other object 12b, between the two objects 12a and 12b. You can predict that a collision will occur.

The determining of the collision between the objects according to another embodiment may include the same as described with reference to FIG. 4, when the position and angle of the moving object are known on the linear moving path in which the nonlinear moving path is approximated. And modeling the object with a geometric shape including the number of cases in which the object can be redirected within an angle, and calculating the collision of the geometric shapes by applying the moving direction and the moving speed of the moving object to the geometric shapes.

In addition, the geometric shapes may be modeled by a plurality of inscribed circles inscribed in each of the geometric shapes, and the collisions of the geometric shapes may be calculated based on the inscribed circle.

11 is a flowchart illustrating a method for predicting and avoiding collision and proximity between a plurality of moving objects according to a second embodiment of the present invention.

Referring to FIG. 11, in the method of predicting and avoiding collision and proximity between a plurality of moving objects, generating objects by modeling shapes of each of the plurality of moving objects (S210), and detecting the objects from the size information of the objects. Generating a 3D sphere by modeling (S220), modeling the 3D spheres as moving spheres using at least one of moving speeds of the moving bodies, a monitoring time interval of the moving bodies, and size information of the 3D spheres. (S230), calculating a Voronoi diagram between the moving spheres, calculating faces, edges, and vertices of the Voronoi diagram (S240), and during the monitoring time interval of the moving bodies, actual collisions between the moving spheres. Comprising a step (S250).

Generating objects by modeling shapes of each of the plurality of moving objects (S210) generates objects through shape information of the moving objects.

Generating the three-dimensional sphere by modeling the objects from the size information of the object (S220) generates a three-dimensional circumferential sphere of the minimum diameter circumscribe each of the objects.

The modeling of the three-dimensional spheres as moving spheres (S230) may include moving spheres of the three-dimensional spheres using at least one of a moving speed of the moving bodies, a monitoring time interval of the moving bodies, and size information of the three-dimensional spheres. Model with. According to an embodiment, the three-dimensional spheres are modeled as the moving spheres using Equation 2 described above.

Computing the Voronoi diagram between the moving spheres and calculating the faces, the edges, and the vertices of the Voronoi diagram (S240), as shown in FIG. 12, calculates the Voronoi diagram between the plurality of moving spheres S1 to S5. The surfaces P1, P2, ..., the edges E1 to E7, and the vertices V1 and V2 of the Voronoi diagram are calculated. The faces P1, P2,... Of the Voronoi diagram are created between two adjacent moving spheres S1-S5, and the edges E1-E7 of the Voronoi diagram have a face P1, at least two of the Voronoi diagrams. P2,... Are met and generated, and vertices V1, V2 of the Voronoi diagram are generated by meeting at least three edges E1-E7 of the Voronoi diagram.

Computing the actual collision between the moving spheres (S250), during the monitoring time interval of the moving object, calculates the occurrence of the corner flipping event, the plane flipping event, and the collision event and schedules the event occurrence time in the order of rapidity .

FIG. 13 illustrates an edge flipping event and a face flipping event according to the present invention.

First, referring to FIG. 13A, the edge E1 of the Voronoi diagram is generated by the planes of the three Voronoi diagrams at the first time point t ₀ . Referring to (B), the edge E1 of the Voronoi diagram moves from the second time point t ₀ <t ₁ <t ₀ + Δt to the vertex V _f as the movement spheres S1 to S5 move. Is switched. Referring to (C), the vertex V _f of the Voronoi diagram is a triangular plane (T ₁ <t ₂ <t ₀ + Δt) at the third time point t ₁ <t ₂ <t ₀ + Δt according to the movement of the moving spheres S1 to S5. Switch to P3).

On the contrary, at the first time point t ₀ , the moving spheres S1 to S5 form a plane P3 of the Voronoi diagram in the form of a triangle (FIG. At 2 o'clock (t ₀ <t ₁ <t ₀ + Δt), the plane P3 of the Voronoi diagram is converted to the vertex V _f (Fig. B), and the motion spheres S1 to S5 are shifted accordingly. At three views (t ₁ <t ₂ <t ₀ + Δt), an edge E1 of the Voronoi diagram and a face of the Voronoi diagram including the same may be generated (Fig. A).

As described above, the process of converting the edge E1 of the Voronoi diagram into the triangular plane P3 via the vertex V _f is defined as an edge flipping event, and the plane of the triangle Voronoi diagram ( The process of converting P3) to the edge E1 through the vertex V _f may be defined as a face flipping event.

Although one flipping event is shown in the figure, the number of flipping events may occur simultaneously or in time series according to the number of moving spheres S1 to S5 and the dynamic data.

Calculating the actual collision between the moving spheres (S250) calculates whether or not the actual collision of the moving spheres (S1 to S5) that generates the event in the order of the event time is fast. The calculation of the collision uses dynamic data of the moving spheres S1 to S5, and it may be determined that an actual collision does not occur or a collision does not occur. If it is determined that a collision occurs, the dynamic data of the collision prediction moving spheres is modified to avoid collision of the moving spheres.

The step of calculating the actual collision between the moving spheres updates the information of the Voronoi diagram after the elapse of the corresponding event time.

The update of the information of the Voronoi diagram in which the edge flipping event is generated proceeds to the contents of B-1-1) to B-1-4) described above, and the update of the information of the Voronoi diagram in which the face flipping event is generated. Is proceeded to the contents of the above B-2-1) to B-2-4), and the update of the information of the Voronoi diagram in which the collision event occurred is described in the above B-3-1) and B-3-2. Proceeds to).

Calculating the actual collision between the moving spheres (S250) is to schedule the re-calculated corner flipping event, the surface flipping event, and the collision event described above in the order of occurrence time, Calculate whether there is an actual collision between the moving spheres generating the event, and if it is determined that the actual collision occurs, modify the dynamic data of the collision predicted moving spheres to avoid the collision.

The calculating of the actual collision between the moving spheres (S250) includes updating the information on the Voronoi diagram described above, calculating whether or not the actual collision between the updated spheres is updated, and avoiding collision during the monitoring time interval of the moving objects. When the monitoring time interval of the moving objects elapses, or when no edge flipping event, a face flipping event, and a collision event are not found during the corresponding time interval, the actual collision prediction between the moving spheres is terminated.

On the other hand, the method of predicting and avoiding collision and proximity between a plurality of moving objects, when collision is predicted between the moving spheres generating an event, determining whether the objects of the moving spheres that are predicted to collide further collide. Include. By determining whether the objects collide with each other, the possibility of the actual collision of the moving object is verified.

The determining of the collision between the objects according to an embodiment of the present disclosure calculates whether the objects actually collide from the moving direction and the moving speed of each of the moving spheres during the monitoring time of the moving objects.

In the determining of whether the objects collide with each other according to an embodiment of the present disclosure, the objects are modeled using a plurality of inscribed spheres inscribed in each of the objects, and the collision is calculated based on inscribed spheres. The method of modeling objects with indirect spheres may proceed to the process described with reference to FIG. 3. The determination of whether the collision is in the unit of the inscribed sphere is the same as described in FIG. 8, by applying the moving direction and the moving speed of the moving object to each of the inscribed spheres to predict the inscribed paths of the inscribed spheres, and inscribed spheres of other objects in the expected path To determine whether or not When any one of the inscribed spheres constituting one object collides with the inscribed spheres constituting the other object, it can be predicted that a collision occurs between both objects.

In the determining of whether the objects collide with each other according to another embodiment (S260), when the non-linear movement path knows the redirection position and angle of the moving object on the approximated linear movement path, the same as described with reference to FIG. 4. The object can be modeled with a geometric shape that includes the number of cases the object can turn at the turning position and angle, and the collision direction can be calculated by applying the moving direction and the moving speed of the moving object to the geometric shapes. have.

In addition, the geometric shapes may be modeled using a plurality of inscribed spheres inscribed in each of the geometric shapes, and the collision of geometric shapes may be calculated in units of inscribed spheres.

In order to avoid this when it is predicted that two of the plurality of moving objects are proximate / crashing after? T time, one or all of them must be modified to avoid collision. Here, the equation of the movement path also includes the movement path and the magnitude of the speed according to time in the movement path. Therefore, the method of modifying the equation of the moving path may be a method of changing only one of the moving path and the moving speed or both the moving path and the moving speed. At this time, the movement path and the movement speed may include a nonlinearity. Given that many moving bodies are moving and at any time any moving object can modify the equation of the moving path, the optimal way of evading path finding the safest escape path is to make various candidate solutions and evaluate each of them quickly. Choose the most favorable solution. It is important to minimize the energy (fuel, battery, etc.) used for the moving object to be the target of the present invention. For example, in the case of satellites, the energy used for avoiding maneuvers is very expensive and in many cases directly linked to the life of the satellites, so careful selection is essential. Therefore, the concept of optimal design of evasion operation is needed. At this time, it is essential to define the time window to define the optimal state. Suppose we have a dynamic Voronoi diagram that covers the desired time window, and we decide that the satellite S we are interested in collides a little later, and soon we need to make a decision to dodge. In the present invention, one copy of S1, which is a copy of the moving object S, which is the main cause of avoidance triggering, is placed in the same position as the moving object in the dynamic Voronoi diagram already calculated. At this time, S1 is in a state where there is a difference in angle, a difference in speed, or a slight difference in both the angle and the speed compared to the moving body S. At this time, as time passes along the time window, S1 uses the dynamic Voronoi diagram at each moment to analyze the close relationship with other adjacent objects. This allows one evaluator to evaluate the average O (K) time. Where K is the number of events that exist across the time window of the dynamic Voronoi diagram. It is possible to generate another avoiding maneuver S2 different from S1. In this case, a solution can be obtained by using S1 and S2 independently in a given dynamic Voronoi diagram. That is, the algorithm for finding the optimal solution of the avoidance path can be parallelized based on the dynamic Voronoi diagram, and each of the avoidance paths can be evaluated in O (K), so the optimal solution can be found very quickly. Evaluate all candidate evasions and make the best answer for them. In addition, the AI function may be additionally used as a method for finding a good solution among various candidate solutions.

FIG. 14 is a diagram illustrating a method of predicting a collision and proximity of a drone and providing an avoiding path according to an embodiment of the present invention. Referring to FIG. 14, (A) to (E) show that a drone predicts a collision and sets an avoidance path through a method of predicting collision and proximity of a moving object according to the present invention, thereby forming a two-dimensional star in a three-dimensional space without collision The process of moving from two shapes to two circles is shown sequentially. In the present embodiment, 100 drones are used, and each of the moving spheres models a moving drone. When the drones are moved to the position E of the moving spheres at a time point t2 after a predetermined time at the position A of the moving spheres at a specific time point t1, the starting point and the end point of the specific drone are determined at this time. Drones move linearly between two points. Thus, calculating a dynamic Voronoi diagram between moving spheres during a time interval (t1, t2) allows collision prediction, and if this happens, stop one of the two spheres (or move it slowly) You can move it quickly to avoid collisions. If necessary, the spatial information in the Voronoi diagram can be used more actively to reflect more complicated situations such as changing the drone avoidance path from the linear path described above to the non-linear path.

FIG. 15 is a diagram for describing a method of predicting a collision of a drone and providing an avoiding path, according to another exemplary embodiment.

Referring to FIG. 15, (A) to (E) sequentially illustrate a process in which drones move three-dimensional space without collision by predicting collision and setting an avoidance path through a collision prediction method of a moving object according to the present invention. . In the present embodiment, 500 drones were used, and each of the moving spheres models a moving drone. The moving spheres change from a rabbit-shaped cluster to a hand-shaped cluster. When the drones are moved to the position E of the moving spheres at a time point t2 after a predetermined time at the position A of the moving spheres at a specific time point t1, the starting point and the end point of the specific drone are determined at this time. Drones move linearly between two points. Thus, calculating a dynamic Voronoi diagram between moving spheres during a time interval (t1, t2) allows collision prediction, and if this happens, stop one of the two spheres (or move it slowly) You can move it quickly to avoid collisions. If necessary, the spatial information in the Voronoi diagram can be used more actively to reflect more complicated situations such as changing the drone avoidance path from the linear path described above to the non-linear path.

Hereinafter, a method of predicting collision of moving objects for which a moving path is known according to an embodiment of the present invention will be described. In the present invention, a moving object known to the moving path is described as an example of a satellite for processing around the earth, the moving object known to the moving path is not limited to this can be applied to the collision prediction of various objects moving along a certain orbit. In the present invention, a three-dimensional moving sphere is described as an example, but the technical idea of the present invention can be equally applied to a two-dimensional moving disk. In addition, although the moving sphere is represented on two dimensions for convenience of understanding, it will be easily understood in three dimensions.

FIG. 16 is a diagram for describing a method of predicting a collision of moving bodies in which a moving path is known according to one embodiment of the present invention, and FIG. 16 is a diagram for describing a method of determining whether a moving body collides with each other.

First, referring to FIG. 16, the satellites may be modeled as moving spheres by the moving object modeling unit 120 described with reference to FIG. 1. The first to third moving spheres 210, 310, 410 process around the earth 1000 along the respective trajectories 220, 320, 420.

(A) shows the positions of the first to third moving spheres 210, 310, and 410 at the first reference time point, and (B) shows the position of the second reference time point when a predetermined time has elapsed from the first reference time point. Indicates the positions of the first to third moving spheres 210, 310, and 410, and (C) denotes the first to third moving spheres at the third reference time point when a preset time elapses from the second reference time point. The positions of 210, 310, and 410 are shown.

Referring to (A), the surfaces P1, P2, and P3 of the Voronoi diagram are calculated by calculating Voronoi diagrams for the first to third moving spheres 210, 310, and 410 at the first reference time point. do.

Among the first to third moving spheres 210, 310 and 410, the second and third moving spheres 310 and 410 nearest to the surface P3 of the Voronoi diagram are extracted as collision prediction objects.

The collision speed is determined by analyzing the moving speeds of the second and third moving spheres 310 and 410 extracted as the collision prediction objects.

Referring to FIG. 17, the second and third moving spheres 310 and 410 in which collision is expected may be set under the following conditions. The radius of the second moving sphere 310 is r _A, and the radius of the third moving sphere 410 is r _{B. The} position of the center point A at the first reference time point of the second moving sphere 310 is (A _x , A _y , A _z ), and the position of the center point B of the third moving sphere 410 at the first reference time point is (B _x , B _y , B _z ). The speed at which the second moving sphere 310 moves along the track is V _A = (v _x , v _y , v _z ), and the speed at which the third moving sphere 410 moves along the track is U _B = (u _x , u _y , u _z ).

In the above-described conditions, the estimated position of the center point A of the second moving sphere 310 is A (t) after t time elapses from the first reference time point, and is the center point B of the third moving sphere 410. ) Is defined as B (t), and whether or not the second moving sphere 310 and the third moving sphere 410 collide may be calculated through Equation 1 below.

[Equation 3]

(r _x is A _x -B _x , r _y is A _y -B _y , r _z is A _z -B _z , w _x is v _x -u _x , w _y is v _y -u _y , w _z is v _z- u _z )

d (t) is a distance between the second moving sphere 310 and the third moving sphere 410 after the time elapses from the first reference time point. If d (t) that converge to the sum of the radius (r _B) of the radius (r _A), and a third mobile opening (410) of the second moving sphere 310, the second moving sphere 310 and the third mobile It can be determined that the sphere 410 collides.

The collision point of the second moving sphere 310 and the third moving sphere 410 may be calculated through Equation 2 below.

[Equation 4]

According to an embodiment, it is determined that the second and third moving spheres 310 and 410 do not collide.

Referring again to FIG. 16B, the surfaces P1, P2, and P of the Voronoi diagram may be calculated by calculating Voronoi diagrams of the first to third moving spheres 210, 310, and 410 at the second reference time point. P3) is calculated.

Among the first to third moving spheres 210, 310 and 410, the second and third moving spheres 310 and 410 nearest to the planes P1, P2 and P3 of the Voronoi diagram are extracted as collision prediction objects. do.

By analyzing the moving speeds of the second and third moving spheres 310 and 410 extracted as the collision prediction objects, it is determined whether the collision is performed by the method described with reference to FIG. 17. According to an embodiment, it is determined that the second and third moving spheres 310 and 410 do not collide.

Referring to (C), the positions of the first to third moving spheres 210, 310, and 410 may be checked at the third reference time point. It can be seen that the second and third moving spheres 310 and 410 collide at the third reference time point.

18 is a flowchart illustrating a method of predicting collision of moving objects for which a moving path is known according to another embodiment of the present invention. In the present embodiment, as shown in FIG. 15, a method of predicting collision of satellites will be described as an example.

Referring to FIG. 18, a method of predicting collision of moving objects for which a moving path is known includes a first collision determination step S300 of moving spheres and a second determination step S4000 of moving spheres.

In the first collision determination step (S300) of the moving spheres, the first orbital approximation and the size correction step of the moving spheres (S310), the surface calculation step (S320) of the Voronoi diagram, the collision prediction object extraction step (S330), and the collision It includes a determination step (S340).

In the first orbital approximation and the size correction step S310 of the moving spheres, the trajectories of the moving spheres are approximated to n straight sections to generate first approximate orbits. The size of the moving spheres is first corrected by reflecting an error between the trajectories of the moving spheres and the first approximate orbits.

In order to determine whether or not there is a collision between the first-corrected moving spheres, the plane calculation step (S320) of the Voronoi diagram, the collision prediction object extraction step (S330), and whether there is a collision The determination step S340 is performed. The plane calculation step S320, the collision prediction object extraction step S330, and the collision determination step S340 of the Voronoi diagram may be performed in the same manner as described with reference to FIG. 11.

When it is determined that the collision between the moving spheres whose size is first corrected does not occur in the first collision determination step (S300) of the moving spheres, the determination of whether the moving spheres collide with each other over a predetermined time period (S300) Can be repeated.

When it is determined that collision between the moving spheres whose size is first corrected has occurred in the first collision determination step (S300) of the moving spheres, the second collision determination step (S400) of the moving spheres is performed. The second collision determination step (S400) of the moving spheres is the secondary orbital approximation and the size correction step of the moving spheres (S410), the surface calculation step (S420) of the Voronoi diagram, the collision prediction object extraction step (S430), and the collision It includes a determination step (S440).

The second orbital approximation and the size correction step (S410) of the moving spheres to approximate the trajectory of the moving spheres to m linear sections larger than n to generate second approximate orbits. The size of the moving spheres is secondarily corrected by reflecting an error between the moving spheres and the second approximate orbits. The size of the second-corrected moving spheres is smaller than the size of the first-corrected moving spheres.

In order to determine whether the collisions between the second-corrected moving spheres are corrected, the plane calculation step (S420), the collision prediction object extraction step (S430), and whether there is a collision in the Voronoi diagram for the second-corrected moving spheres The determination step S440 is performed.

When it is determined that the collision between the moving spheres whose size is second-corrected has not occurred in the secondary object collision determination step (S400), the secondary collision determination step (S400) of the moving spheres is repeatedly performed at predetermined periods. Can be performed.

When it is determined in step S400 of whether the collision of the moving spheres has occurred, the collision between the two spheres whose size is second-corrected is approximated, and the trajectories of the moving spheres are approximated to more than m straight sections. A third object collision determination that determines whether the spheres collide may be performed. Alternatively, according to another embodiment, the third collision determination of the moving spheres does not approximate the trajectory approximation and the size correction of the moving spheres, and determines whether the moving spheres collide with each other in the known actual orbit of the moving spheres. can do.

The third collision determination step of the moving spheres includes a plane calculation step, a collision prediction object extraction step, and a collision determination step of the Voronoi diagram for the moving spheres described above.

19 is a view for explaining in detail the first collision determination step of the moving spheres described in FIG. 18, FIG. 20 is a view for explaining in detail the second collision determination step of the moving spheres described in Figure 18, Figure 21 FIG. 18 is a diagram for describing a third collision determination step of the moving spheres described with reference to FIG. 18 in detail.

First, referring to FIG. 19, first to third moving spheres 210, 310, and 410 are approximated to five straight sections of the first to third trajectories 220, 320, and 420 at a first reference time point A. FIG. First approximate orbits 230, 330, and 430. The first to third reflecting errors between the first to third trajectories 220, 320, 420 and the first approximate trajectories 230, 330, 430 of the first to third moving spheres 210, 310, and 410. The size of the moving spheres 210, 310, and 410 is first corrected.

The surfaces P1, P2, and P3 of the Voronoi diagram are calculated by calculating the Voronoi diagrams of the first to third moving spheres 240, 340, and 440 whose sizes are first-corrected.

Among the first to third moving spheres 240, 340, and 440 of which the size is first-corrected, the second and third moving spheres 340 and 440 of which the size is first-corrected to the nearest surface P3 of the Voronoi diagram. ) Is extracted as collision prediction objects.

The collision speed is determined by analyzing the moving speeds of the second and third moving spheres 340 and 440 whose sizes extracted as the collision prediction objects are first corrected. According to the embodiment, it is determined that the second and third moving spheres 340 and 440 whose sizes are first corrected do not collide.

Thereafter, the collision determination between the first to third moving spheres 240, 340, and 440 whose sizes are first corrected may be repeatedly performed at predetermined periods.

At the second reference time point B, the first and second moving spheres 240 and 340 whose sizes are firstly corrected are extracted as collision prediction objects, and the first and second moving spheres whose first sizes are firstly corrected, The collision speed is determined by analyzing the moving speeds of the 240 and 340. According to an embodiment, it is determined that the first and second moving spheres 240 and 340 whose sizes are first corrected collide with each other.

Referring to FIG. 20A, when it is determined that the first and second moving spheres 240 and 340 whose sizes are first corrected collide with each other, the first to third trajectories 220, 320, and 420 may be moved to 10. The second approximate trajectories 250, 350, and 450 of the first to third moving spheres 210, 310, and 410 are generated by approximating to four straight sections. The size is first-order corrected by reflecting an error between the first to third trajectories 220, 320, 420 and the second approximate trajectories 250, 350, 450 of the first to third moving spheres 210, 310, and 410. Secondly correcting the sizes of the first to third moving spheres 240, 340, and 440.

The surfaces P1, P2, and P3 of the Voronoi diagram are calculated by calculating the Voronoi diagrams of the first to third moving spheres 260, 360, and 460 whose sizes are second-corrected.

Among the first to third moving spheres 260, 360, and 460 whose sizes are second-corrected, the first and second moving spheres 260 and 360 whose second-corrected sizes are the nearest to the plane P1 of the Voronoi diagram. ) Is extracted as collision prediction objects.

The collision speed is determined by analyzing the moving speeds of the first and second moving spheres 260 and 360 whose sizes extracted with the collision prediction objects are second-corrected. According to an embodiment, it is determined that the first and second moving spheres 260 and 360 whose sizes are second-corrected do not collide.

Thereafter, the collision determination between the first to third moving spheres 240, 340, and 440 whose sizes are second-corrected may be repeatedly performed at predetermined periods.

The second and third moving spheres 360 and 460 whose sizes are secondarily corrected at the third reference time point B are extracted to the collision prediction objects, and the second and third moving spheres whose sizes are secondarily corrected. By analyzing the moving speed of the 360, 460, it is determined whether the collision. According to the embodiment, it is determined that the second and third moving spheres 360 and 460 whose sizes are second-corrected collide with each other.

If it is determined that the second and third moving spheres 360 and 460 whose sizes are second-corrected collide with each other, a third collision determination step of the moving spheres is performed.

Referring to FIG. 21, at the third reference time point, second and third moving spheres 310 and 410 are extracted as the collision prediction objects, and moving speeds of the second and third moving spheres 310 and 410 are analyzed. To determine if there is a collision. According to an embodiment, it is determined that the second and third moving spheres 310 and 410 collide with each other.

Comparing the results of FIG. 20B with the results of FIG. 21, the second collision determination result of the moving spheres 310 and 410 determines whether the object collides according to the actual trajectory of the moving spheres 310 and 410. It can be seen that the results coincide with the results of the third collision determination of the spheres 310 and 410.

Therefore, the collision prediction method of the moving objects 210, 310, and 410 having known movement paths according to the embodiment of the present invention reduces the overall calculation amount required for collision prediction and improves the collision prediction speed, thereby improving the accuracy of collision prediction. Can be improved.

As mentioned above, although this invention was demonstrated in detail using the preferable embodiment, the scope of the present invention is not limited to a specific embodiment, Comprising: It should be interpreted by the attached Claim. In addition, those skilled in the art should understand that many modifications and variations are possible without departing from the scope of the present invention.

100: collision prediction and avoidance system between moving objects
110: moving object information receiving unit
120: moving object modeling unit
130: collision prediction unit
10: moving object
11: object
12: two-dimensional circle, three-dimensional sphere
13: moving disk, moving sphere
14: Geometric Formation
21 to 25: inscribed circle, inscribed sphere
E1, E2, E3, ...: edge of Voronoi diagram
V1, V2, ... vertex of Voronoi diagram
S1 to S5: moving sphere
P1, P2, P3: Faces of Voronoi Diagram
S: moving object
210, 310, 410: moving sphere
220, 320, 420: Orbit
230, 330, 430: approximate orbit
1000: earth

Claims (2)

Extracting at least two or more collision prediction objects by calculating a Voronoi diagram between a plurality of moving objects;
Generating S1 which is a copy of any one of the collision prediction objects and placing the S1 at the same point as the one collision prediction object;
Changing at least one of the moving direction and the moving speed of the S1 differently from the moving direction and the moving speed of the collision predicted object;
Calculating whether S1 collides with the remaining collision predicted objects; And
If it is determined that S1 does not collide with the remaining collision prediction objects, changing the moving direction and the moving speed of the one collision prediction object to the moving direction and the moving speed of the S1 between the moving objects Collision prediction and avoidance methods. The method of claim 1,
Generating an additional copy of the collision prediction object S2 and positioning the same at the same point as the collision prediction object;
Changing at least one of the moving direction and the moving speed of the S2 differently from the moving direction and the moving speed of the collision predicted object and the moving direction and the moving speed of the S1;
Calculating collision with the remaining collision prediction objects using the Voronoi diagram of each moment previously calculated for each of S1 and S2;
The collision calculation step using the pre-calculated Voronoi diagram may be applied to each of S1 and S2 independently to perform parallel processing; And
When it is determined that the S1 and the S2 do not collide with the remaining collision predicted objects, one of the moving direction and the moving speed of the S1 and the moving direction and the moving speed of the S2 is selected as the optimal solution. A collision prediction and avoidance method between moving objects that change to a moving direction and a moving speed of one collision predicted object.

Images