Anti-collision method and device, electronic equipment, storage medium and product

Abstract

The invention provides an anti-collision method and device, electronic equipment, a storage medium and a product. The method comprises the following steps: acquiring an observation state of a first flying object and a first prediction state of the first flying object; based on the observation state of the first flying object and the first prediction state of the first flying object, establishing a chasing action data model of the first flying object to judge whether the first flying object has a chasing action or not; in response to the fact that the first flying object has the chasing action, collision detection is carried out on the unmanned aerial vehicle in the first area, and the collision detection is used for detecting whether a collision risk exists between the first flying object and the unmanned aerial vehicle or not; and if there is a collision risk between the first flying object and the unmanned aerial vehicle, flying according to a first path obtained by a preset path planning algorithm to prevent the first flying object from colliding with the unmanned aerial vehicle.

Description

Technical Field

[0001] The present disclosure relates to the technical field of low-altitude security, and in particular, to a collision prevention method, device, electronic device, storage medium, and product. Background Art

[0002] In related UAV collision prevention solutions, UAV obstacle avoidance flight technology usually relies on a variety of sensing technologies and map data. For example, a three-dimensional elevation map provides height information of the terrain, enabling the UAV to identify and avoid obstacles on the ground; a visual sensing system captures environmental images through a camera and uses computer vision algorithms to analyze these images to detect static and dynamic obstacles; an infrared sensing system identifies obstacles by measuring the difference in heat emitted by objects and is suitable for low-light or night operation environments; an ultrasonic sensing system measures the distance to nearby objects by emitting and receiving ultrasonic waves and is suitable for close-range obstacle avoidance. However, in the above solutions, there is a problem that it is impossible to avoid the pursuit of moving objects such as birds. Summary of the Invention

[0003] The present disclosure provides a collision prevention method, device, electronic device, storage medium, and product to solve the problem in related technologies that it is impossible to avoid the pursuit of moving objects such as birds.

[0004] A first aspect embodiment of the present disclosure proposes a collision prevention method, which includes:

[0005] Obtain the observation state of a first flying object and the first prediction state of the first flying object;

[0006] Based on the observation state of the first flying object and the first prediction state of the first flying object, establish a pursuit action data model of the first flying object to determine whether the first flying object has a pursuit action;

[0007] In response to detecting that the first flying object has a pursuit action, perform a collision detection on the UAV in a first area, and the collision detection is used to detect whether there is a collision risk between the first flying object and the UAV;

[0008] If there is a collision risk between the first flying object and the UAV, fly along a first path obtained according to a preset path planning algorithm to prevent the first flying object from colliding with the UAV.

[0009] In an embodiment, obtaining the observation state of the first flying object and the first prediction state of the first flying object includes:

[0010] Obtain the observation states of the first flying object at at least two consecutive moments, and the observation state of the first flying object at least includes the position of the first flying object;

[0011] Determine the velocity of the first flying object based on the positions of the first flying object at the two consecutive moments;

[0012] Determine the second predicted state of the first flying object based on the velocity of the first flying object and a preset state transition matrix, where the second predicted state of the first flying object includes the velocity of the first flying object at the next moment and the position of the first flying object at the next moment;

[0013] Determine the first predicted state of the first flying object by using a preset state estimation algorithm based on the second predicted state of the first flying object and the observed state of the first flying object.

[0014] In one embodiment, determining the first predicted state of the first flying object by using a preset state estimation algorithm based on the second predicted state of the first flying object and the observed state of the first flying object includes:

[0015] Determine the predicted covariance of the first flying object and the observed noise covariance of the first flying object based on the observed state of the first flying object;

[0016] Determine the Kalman filter gain based on a preset observation matrix, the predicted covariance of the first flying object, and the observed noise covariance of the first flying object;

[0017] Determine the first predicted state of the first flying object based on the observed state of the first flying object, the second predicted state of the first flying object, and the Kalman filter gain.

[0018] In one embodiment, in response to detecting that the first flying object has a pursuit action, perform a collision detection on the unmanned aerial vehicles in the first area, including:

[0019] In response to detecting that the first flying object has a pursuit action, obtain the position of the first flying object, the position of the unmanned aerial vehicle, and a preset support function;

[0020] Determine the movement path of the first flying object and the movement path of the unmanned aerial vehicle based on the position of the first flying object, the position of the unmanned aerial vehicle, and the preset support function;

[0021] Calculate the Euclidean distance between the first flying object and the unmanned aerial vehicle based on the movement path of the first flying object and the movement path of the unmanned aerial vehicle, so as to perform a collision detection on the unmanned aerial vehicles in the first area.

[0022] In one embodiment, if there is a collision risk between the first flying object and the unmanned aerial vehicle, fly along the first path obtained according to a preset path planning algorithm to prevent the first flying object from colliding with the unmanned aerial vehicle, including:

[0023] If there is a collision risk between the first flying object and the drone, obtain the first coordinate of the drone, where the first coordinate of the drone is the coordinate when the drone is listed as a pursuit target by the first flying object;

[0024] Obtain a random coordinate point, use the first coordinate of the drone as the root node, and use the direction from the first coordinate of the drone to the random coordinate point as the growth direction to generate new nodes according to a preset step size;

[0025] If the path from the first coordinate of the drone to the new node does not intersect with the movement path of the first flying object, add the new node to the random expansion tree and stop iteratively generating new nodes until a preset condition is met;

[0026] Fly along the first path between the first coordinate of the drone and the target point coordinate to prevent the first flying object from colliding with the drone, where the first path is the path in the random expansion tree that traces back from the target point coordinate to the first coordinate of the drone.

[0027] In an embodiment, after flying along the first path between the first coordinate of the drone and the target point coordinate to prevent the first flying object from colliding with the drone, the method provided by the present disclosure further includes:

[0028] In response to the distance between the second coordinate of the drone and the coordinate of the first flying object being not less than a preset threshold, fly to the target point coordinate according to a preset path.

[0029] An embodiment of the second aspect of the present disclosure provides a collision prevention device, and the device includes:

[0030] An acquisition unit, configured to acquire the observation state of the first flying object and the first prediction state of the first flying object;

[0031] A model establishment unit, configured to establish a pursuit action data model of the first flying object based on the observation state of the first flying object and the first prediction state of the first flying object to determine whether the first flying object has a pursuit action;

[0032] A collision detection unit, configured to perform collision detection on the drones in the first area in response to detecting that the first flying object has a pursuit action, where the collision detection is used to detect whether there is a collision risk between the first flying object and the drones;

[0033] A path planning unit, configured to fly along the first path obtained according to a preset path planning algorithm to prevent the first flying object from colliding with the drone if there is a collision risk between the first flying object and the drone.

[0034] A third aspect embodiment of the present disclosure provides an electronic device, including:

[0035] At least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method described in the first aspect embodiment of the present disclosure.

[0036] A fourth aspect embodiment of the present disclosure provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to execute the method described in the first aspect embodiment of the present disclosure.

[0037] A fifth aspect embodiment of the present disclosure provides a computer program product, including a computer program, where the computer program, when executed by a processor, implements the method described in the first aspect embodiment of the present disclosure.

[0038] In summary, the present disclosure provides an anti-collision method, which includes: obtaining an observation state of a first flying object and a first prediction state of the first flying object; based on the observation state of the first flying object and the first prediction state of the first flying object, establishing a pursuit action data model of the first flying object to determine whether the first flying object has a pursuit action; in response to detecting that the first flying object has a pursuit action, performing a collision detection on unmanned aerial vehicles in a first area, where the collision detection is used to detect whether there is a collision risk between the first flying object and the unmanned aerial vehicles; if there is a collision risk between the first flying object and the unmanned aerial vehicles, flying along a first path obtained according to a preset path planning algorithm to prevent the first flying object from colliding with the unmanned aerial vehicles.

[0039] According to the solution provided by the present disclosure, by establishing a pursuit action data model of the first flying object based on the observation state of the first flying object and the first prediction state of the first flying object, it is possible to accurately determine whether the first flying object has a pursuit action. If it is detected that the first flying object has a pursuit action, a collision detection is performed on the unmanned aerial vehicles in the first area. If there is a collision risk between the first flying object and the unmanned aerial vehicles, flying along the first path obtained according to the preset path planning algorithm can prevent the first flying object from colliding with the unmanned aerial vehicles.

[0040] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The accompanying drawings herein are incorporated into the specification and constitute a part of this specification, showing embodiments consistent with the present disclosure, and are used together with the specification to explain the principles of the present disclosure, and do not constitute an improper limitation to the present disclosure.

[0042] Figure 1 Schematic flowchart of the anti-collision method provided by an embodiment of the present disclosure;

[0043] Figure 2 Schematic flowchart of the method for determining the first predicted state of a first flying object provided by an embodiment of the present disclosure;

[0044] Figure 3 Schematic flowchart of the method for determining the first predicted state of a first flying object provided by an embodiment of the present disclosure;

[0045] Figure 4 Schematic flowchart of the collision detection method provided by an embodiment of the present disclosure;

[0046] Figure 5 Effect diagram of the collision detection provided by an embodiment of the present disclosure;

[0047] Figure 6 Schematic flowchart of the method for obtaining the first path provided by an embodiment of the present disclosure;

[0048] Figure 7 Effect diagram of the first path obtained provided by an embodiment of the present disclosure;

[0049] Figure 8 Scene diagram of the anti-collision method provided by an embodiment of the present disclosure;

[0050] Figure 9 Schematic flowchart of the anti-collision method provided by an application example of the present disclosure;

[0051] Figure 10 Schematic structural diagram of the anti-collision device provided by an embodiment of the present disclosure;

[0052] Figure 11 Schematic diagram of the hardware composition structure of the electronic device provided by an embodiment of the present disclosure. Detailed implementation manners

[0053] The embodiments of the present disclosure will be described in detail below. Examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions from start to finish. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to explain the present disclosure, but should not be construed as limiting the present disclosure.

[0054] To facilitate better understanding of the technical solutions described in the embodiments of the present disclosure by those skilled in the art, the following explanations are made for the technical terms in the embodiments of the present disclosure before introducing the embodiments of the present disclosure.

[0055] Integrated Sensing And Communication (ISAC): The integration of communication and sensing systems, abbreviated as integrated communication and sensing.

[0056] Kalman Filter (KF) algorithm: Used to estimate the state of a linear dynamic system, which minimizes the covariance of the estimation error by combining prediction and measurement data to provide an optimal state estimate.

[0057] Gilbert-Johnson-Keerthi (GJK) algorithm: The GJK algorithm is an efficient algorithm for detecting whether two convex objects intersect. It determines whether a collision occurs by constructing the Minkowski difference and checking whether the origin lies within this difference set.

[0058] Rapidly-exploring Random Tree (RTT) algorithm: Generates a tree structure through random sampling to explore the free space and find a feasible path from the starting point to the target point.

[0059] In related UAV anti-collision solutions, UAV obstacle avoidance flight technology usually relies on various sensing technologies and map data. For example, a three-dimensional elevation map provides height information of the terrain, enabling the UAV to identify and avoid obstacles on the ground; a visual sensing system captures environmental images through a camera and uses computer vision algorithms to analyze these images to detect static and dynamic obstacles; an infrared sensing system identifies obstacles by measuring the difference in heat emitted by objects and is suitable for low-light or night operation environments; an ultrasonic sensing system measures the distance to nearby objects by emitting and receiving ultrasonic waves and is suitable for close-range obstacle avoidance. However, in the above solutions, there is a problem that it is impossible to avoid the pursuit of moving objects such as birds.

[0060] Narrow-sense integrated communication and sensing refers to a communication system with capabilities such as ranging, speed measurement, angle measurement, imaging, target detection, target tracking, and target recognition. It was also called "radar-communication integration" in the early days. Broad-sense integrated communication and sensing refers to a communication system that can sense the attributes and states of all services, networks, users, terminals, and environmental objects, and its sensing ability can exceed that of traditional radars.

[0061] As the 5G spectrum expands from traditional Sub6G to millimeter waves, the reduction in wavelength continuously improves the sensing ability. Therefore, in the second half of 5G, namely the 5G-Advanced stage (5G-A), integrated communication and sensing has been included in the standardization agenda. In future 6G, the spectrum will expand to terahertz, and the sensing ability will be further enhanced, bringing greater room for imagination.

[0062] "Low-altitude security" is a typical integrated communication and sensing application scenario. With the development of consumer drones (Unmanned Aerial Vehicles, UAVs), their applications in military, agricultural, logistics and other fields are becoming more and more widespread. Due to the difficulty of monitoring, the phenomenon of drones flying around randomly is becoming more and more serious. Although this is not a big problem for individuals, for some units that need to keep secrets, no matter how strict the ground security is, it cannot stop drones from flying in and out as if there is no one there, and it is simply too easy to take pictures in the air without permission.

[0063] The traditional manual control method cannot meet the flight requirements in its complex environment. Therefore, how to achieve autonomous obstacle avoidance flight of drones has become an important content of current research.

[0064] At the same time, in order to prevent problems such as leakage of secrets, collisions and noise caused by "unauthorized flight" of drones, it is necessary to deploy a low-altitude security system efficiently and at low cost. At present, there are various detection schemes in the drone security market, but they all face many limitations such as technology, efficiency and cost.

[0065] Conversely, based on the imaging, map construction and environment reconstruction capabilities provided by the integrated communication and sensing, the system can turn passive into active, send out drones to carry out activities such as reconnaissance and logistics delivery, and can perform automatic navigation and path planning in unknown environments according to the multi-station sensing capabilities.

[0066] However, at present, the drone obstacle avoidance flight technology generally relies on three-dimensional elevation maps, visual perception systems, infrared perception systems or ultrasonic perception systems, focusing on the research of obstacle avoidance flight route planning at the medium and long-distance navigation level. It cannot avoid the active impact of moving objects such as birds, nor can it avoid the simultaneous impact of different obstacles in multiple directions. Therefore, how to deal with the real complex environment is still the deficiency of the existing technology.

[0067] The following briefly introduces several solutions of anti-collision methods in related technologies:

[0068] Solution A: Solution A provides a drone countermeasure system, including: a target detection system, a target disposal system and a command and control system connected by communication; the present invention can use a variety of detection means to discover suspicious incoming targets, and determine the target characteristics based on the distances of the suspicious incoming targets under a variety of detection means, and then determine whether the suspicious incoming targets are incoming targets; once determined, the target disposal system can take appropriate means to strike the incoming targets. The present invention can quickly discover, quickly identify and quickly dispose of various types of "low, slow and small" targets such as drones, balloons, kites, sky lanterns, birds, etc. that threaten the safety of important targets in the threat protection area. The system can not only be used in a single node, but also be networked with multiple nodes, and the commander can remotely monitor the situation of the defense area through a handheld terminal and implement remote command, breaking through the geographical space limit. Using the present invention can improve the low-altitude defense and threat response capabilities.

[0069] Solution B: Solution B provides a method for obstacle avoidance flight of an unmanned aerial vehicle (UAV), which relates to the field of flight control and includes: detecting the neighborhood environment of the UAV in real time, detecting and calculating the velocity vector of the UAV; detecting multiple obstacles and calculating the approaching velocity vectors of the multiple obstacles respectively; calculating the anti-collision acceleration vector through the particle swarm algorithm; and driving the UAV to adjust its pose according to the anti-collision acceleration vector to avoid multiple obstacles. The present invention can detect obstacles in different directions simultaneously, sense the motion states of these multiple obstacles, then iteratively seek and optimize the acceleration solution for anti-collision according to its own motion state, and then make appropriate avoidance actions. It can effectively avoid the active impact of multiple-direction obstacles such as birds and cope with the complex real environment.

[0070] Solution C: Solution C provides a UAV that can prevent bird strikes, including a UAV body. A solar panel is installed on the top surface of the UAV body. A charge and discharge controller and a storage battery are arranged inside the UAV body. One end of the charge and discharge controller is connected to the solar panel, and the other end is connected to the storage battery. Radars are arranged around and on the upper and lower surfaces of the UAV body. Ultrasonic generators are arranged around and on the upper and lower surfaces of the UAV body. A CPU is arranged inside the UAV body. The output end of the radar is connected to the CPU, and the output end of the CPU is connected to the ultrasonic generator. The storage battery powers the ultrasonic generator. When the UAV detects a bird through the radar arranged on it, the CPU drives the ultrasonic generator to emit sound waves to drive away the bird. The ultrasonic generator is powered by solar energy, which can save energy, be clean and environmentally friendly. Red is the color that birds are most sensitive to and afraid of, so it can further prevent birds from approaching.

[0071] Solution D: Solution D provides an energy-saving path planning method for a UAV used for fan inspection, including the following steps: S1: Obtain the shooting viewpoints of the blade part and the hub part of the fan to be inspected to obtain a first viewpoint set; obtain the shooting viewpoints of the tower barrel part of the fan to be inspected to obtain a second viewpoint set; S2: Screen the first viewpoint set based on the greedy algorithm to obtain a third viewpoint set; S3: Perform path planning on the fourth viewpoint set based on the improved LKH and RTT* to obtain an initial inspection path, where the fourth viewpoint set is composed of the third viewpoint set and the second viewpoint set; S4: Convert the initial inspection path into a smooth inspection path, and at the same time control the UAV to fly along the smooth inspection path at a preset speed and take pictures at the passing shooting viewpoints. It can not only comprehensively cover the inspection of the fan but also has low energy consumption.

[0072] Solution E: Solution E provides an intelligent obstacle avoidance control method, device and unmanned aerial vehicle (UAV) based on dynamic obstacle perception. The method includes: acquiring acquisition data signals, performing data analysis to obtain obstacle information, using a perception algorithm and a tracking algorithm to fuse the obstacle information with the state information of the UAV, generating a model of the obstacle and performing dynamic perception and tracking, using a filter, a preset state estimation algorithm or a particle filter algorithm to estimate the state of the obstacle to obtain dynamic obstacle information, then evaluating the collision risk through a dynamic collision detection algorithm, using model prediction and simulation technology to simulate the interaction between the UAV and the obstacle, determining the collision risk assessment result, and generating a control signal according to the collision risk assessment result to control the movement of the UAV. This solves the problems that the accuracy and real-time performance of the existing intelligent obstacle avoidance control are poor, the perception accuracy of obstacles is low, which is not conducive to improving the safety and driving efficiency of intelligent UAVs.

[0073] In the above solutions, there are the following defects:

[0074] First of all, the recognition of the pursuit action of birds is not given.

[0075] Secondly, adding ultrasonic generators to the UAV body to emit sound waves to drive away birds undoubtedly increases the weight and power consumption of the UAV and reduces the endurance of the UAV.

[0076] Thirdly, the intelligent obstacle avoidance control method based on dynamic obstacle perception is executed on the UAV side, which increases the consumption of the UAV's CPU and algorithm resources, and at the same time does not recognize the pursuit actions of dynamic birds.

[0077] To solve the defects in the related technologies, the present disclosure establishes a pursuit action data model of the first flying object based on the observed state of the first flying object and the first predicted state of the first flying object, which can accurately determine whether the first flying object has a pursuit action. If it is detected that the first flying object has a pursuit action, collision detection is performed on the UAV within the first area. If there is a collision risk between the first flying object and the UAV, it flies along the first path obtained according to the preset path planning algorithm, which can prevent the first flying object from colliding with the UAV.

[0078] A collision avoidance method provided by an embodiment of the present disclosure can be applied to a base station deploying a 5G-A communication and sensing integrated device and realizing contiguous networking. The execution subject of the method can be a 5G-A communication and sensing integrated communication system.

[0079] The following further describes the present disclosure in detail with reference to the accompanying drawings and specific embodiments.

[0080] As Figure 1 shown, Figure 1Schematic flowchart of the anti-collision method provided by an embodiment of the present disclosure. The anti-collision method provided by an embodiment of the present disclosure includes the following steps:

[0081] Step 101, obtain the observation state of the first flying object and the first predicted state of the first flying object;

[0082] In one embodiment, the observation state of the first flying object can be obtained through a communication and sensing base station, or the observation state of the first flying object can be obtained through a camera.

[0083] In one embodiment, the observation state of the first flying object at least includes the observation position of the first flying object and the observation speed of the first flying object.

[0084] In one embodiment, the first flying object can be a bird, such as an eagle, or other flying objects.

[0085] In one embodiment, the first predicted state of the first flying object refers to the state of the first flying object obtained through prediction based on the state of the first flying object at the previous time.

[0086] Step 102, based on the observation state of the first flying object and the first predicted state of the first flying object, establish a pursuit action data model of the first flying object to determine whether the first flying object has a pursuit action;

[0087] In one embodiment, compare the observation state of the first flying object with the first predicted state of the first flying object. If there is a significant deviation between the observation state of the first flying object and the first predicted state of the first flying object, it can be determined that the first flying object is performing a hunting behavior. Taking the first flying object as a bird as an example, if the height of the bird drops suddenly, the bird hovers abnormally or turns sharply, it can be considered that the bird is performing a hunting behavior.

[0088] In one embodiment, before the training of the pursuit action data model of the first flying object, a bird feature model library can be constructed from the bird perception echoes captured by the communication and sensing base station. For example, through the analysis of data such as the difference in the movement trajectory of the bird's wings, a feature model of the bird's flapping wings and movement trajectory can be constructed, and the target features can be learned from the massive communication and sensing data through artificial intelligence algorithms to train the model to automatically identify and classify the target.

[0089] In one embodiment, in a 5G-A contiguous networking environment, bird data can be captured from multiple perspectives. Combining the bird feature database and the feature model of the bird's flapping wings and movement trajectory can greatly improve the reliability and accuracy of bird identification, and achieve precise identification of multiple targets. Specifically, the 5G-A communication and sensing integration technology can accurately distinguish birds from drones within a low altitude range of 600 meters.

[0090] Step 103, in response to detecting that the first flying object has a pursuit action, perform a collision detection on the drones within the first area, where the collision detection is used to detect whether there is a collision risk between the first flying object and the drones;

[0091] In one embodiment, the first area refers to a dangerous area where there is a collision risk.

[0092] In one embodiment, taking the first flying object as a bird for example, the first area refers to the area where the circle with the bird as the center and a preset length as the radius is located, where the preset length can be obtained from the historical experience length.

[0093] In one embodiment, the bird having a pursuit action indicates that the bird may pursue the drones within the area where the bird is located. Therefore, it is necessary to perform a collision detection on all the drones within the area where the bird is located.

[0094] Step 104, if there is a collision risk between the first flying object and the drones, fly along the first path obtained according to the preset path planning algorithm to prevent the first flying object from colliding with the drones.

[0095] In one embodiment, if there is no collision risk between the first flying object and the drones, fly along the preset path.

[0096] In one embodiment, the preset path planning algorithm can be the RTT algorithm, or the Probabilistic Roadmap (PRM) algorithm, or the A* (A-Star) algorithm, the Dijkstra algorithm, the Self-Collection algorithm, the Dynamic Window Approach algorithm, etc. In this application, the RTT algorithm is used as the preset path planning algorithm.

[0097] In one embodiment, the first path refers to the flight path of the drones determined according to the preset path planning algorithm.

[0098] In one embodiment, by flying along the first path obtained according to the preset path planning algorithm when there is a collision risk between the first flying object and the drones, the risk of the first flying object colliding with the drones can be avoided.

[0099] In one embodiment, as Figure 2 shown, step 102 includes:

[0100] Step 201, obtain the observation states of the first flying object at at least two consecutive moments, where the observation states of the first flying object at least include the position of the first flying object;

[0101] In one embodiment, a 5G-A integrated communication and sensing system can be used to obtain the observation states of a first flying object at at least two consecutive moments, such as p i =(x i ,y i ,z i ) and p i+1 =(x i+1 ,y i+1 ,z i+1 ) respectively represent the three-dimensional coordinates of the position of the flying bird at the i-th moment and the (i + 1)-th moment. Among them, x i ,y i ,z i respectively represent the longitude, latitude, and altitude of the position where the flying bird is located at the i-th moment.

[0102] Step 202: Based on the positions of the first flying object at the two consecutive moments, determine the speed of the first flying object;

[0103] In one embodiment, based on the positions p i and p i+1 of the flying bird at two consecutive moments, the calculation relationship between the position and the speed can be used to determine the speed v of the current flying bird.

[0104] In one embodiment, based on the position of the flying bird and the calculated speed, the state of the flying bird can be determined. The mathematical expression of the state of the flying bird is:

[0105]

[0106] where x, y, and z are the longitude, latitude, and altitude of the position where the flying bird is located, v is the flying speed of the current flying bird, and H is the observation matrix of the state of the flying bird.

[0107] Step 203: Based on the speed of the first flying object and a preset state transition matrix, determine the second predicted state of the first flying object. The second predicted state of the first flying object includes the speed of the first flying object at the next time and the position of the first flying object at the next time;

[0108] In one embodiment, the mathematical expression of the preset state transition matrix is:

[0109]

[0110] where Δt is the step size of the flying data acquisition period of the flying bird, and F is the state transition matrix.

[0111] In one embodiment, the state transition matrix is used to represent the state change of the system from one time step to the next time step.

[0112] In one embodiment, based on the speed of the first flying object and a preset state transition matrix, the speed of the first flying object at the next moment is determined, and based on the position of the first flying object and the preset state transition matrix, the position of the first flying object at the next moment is determined.

[0113] In one embodiment, since the positions of the first flying object at two consecutive moments are collected by a sensor, there are certain errors in the measurement of the sensor. The errors come from the accuracy of the sensor and environmental interference (such as radio interference, etc.). In addition, due to the uncertainty in the system model prediction process, the state of the first flying object at the next moment obtained through the state transition matrix is not accurate enough.

[0114] Step 204, based on the second predicted state of the first flying object and the observed state of the first flying object, use a preset state estimation algorithm to determine the first predicted state of the first flying object.

[0115] In one embodiment, the preset state estimation algorithm can be the KF algorithm, or the particle filter (PF), or a fusion algorithm (such as multi-sensor fusion), a fuzzy logic control (FLC) algorithm, an artificial potential fields (APF) algorithm, and a convolutional neural network algorithm in deep learning algorithms, etc. In this application, the preset state estimation algorithm is taken as an example of the KF algorithm.

[0116] In one embodiment, by using a preset state estimation algorithm based on the second predicted state of the first flying object and the observed state of the first flying object to determine the first predicted state of the first flying object, the problem that the first predicted state of the first flying object is inaccurate due to the observed state of the first flying object being affected by noise interference and the predicted state of the first flying object being affected by the uncertainty of the system model can be solved.

[0117] In one embodiment, as Figure 3 shown, using a preset state estimation algorithm based on the second predicted state of the first flying object and the observed state of the first flying object to determine the first predicted state of the first flying object includes:

[0118] Step 301, based on the observed state of the first flying object, determine the prediction covariance of the first flying object and the observation noise covariance of the first flying object;

[0119] In one embodiment, the prediction covariance of the first flying object is used to measure the uncertainty of the prediction state estimation of the first flying object.

[0120] In one embodiment, the observation noise covariance of the first flying object is used to describe the uncertainty of the sensor measurement value.

[0121] Step 302: Determine the Kalman filter gain based on a preset observation matrix, the prediction covariance of the first flying object, and the observation noise covariance of the first flying object.

[0122] In one embodiment, the preset observation matrix refers to the observation matrix in the aforementioned step 202.

[0123] In one embodiment, the mathematical expression for determining the Kalman filter gain is:

[0124] K t =P t|t1 H T (HP t|t1 H T +N) 1

[0125] where P t|t-1 is the prediction covariance of the first flying object, H is the preset observation matrix, H T is the transpose of the preset observation matrix, N is the observation noise covariance of the first flying object, and K t is the Kalman filter gain.

[0126] Step 303: Determine the first predicted state of the first flying object based on the observed state of the first flying object, the second predicted state of the first flying object, and the Kalman filter gain.

[0127] In one embodiment, the mathematical expression for determining the first predicted state of the first flying object is:

[0128] X t|t1 =X t|t1 K t (Z t -K t X t|t1 )

[0129] where X t|t-1 on the left side of the equation is the first predicted state of the first flying object, and X t|t-1 on the right side of the equation is the second predicted state of the first flying object (i.e., the predicted state of the first flying object determined using the state transition matrix), and Z t is the observed state of the first flying object at the current moment.

[0130] In one embodiment, before determining the first predicted state of the first flying object, it is also necessary to update the prediction covariance to obtain a more accurate predicted state, where the mathematical expression for updating the prediction covariance is:

[0131] Pt|t1 = FP t-1|t-1 F T Q

[0132] where Q is the process noise covariance, and P t-1|t-1 is the prediction covariance at time t - 1, F is the state transition matrix, and F T is the transpose of the state transition matrix, and P t|t-1 is the prediction covariance at time t. Specifically, the process noise covariance is the corresponding covariance calculated by statistically analyzing the process noise through analyzing historical data.

[0133] In one embodiment, before determining the first predicted state of the first flying object, it is also necessary to update the prediction covariance for the next state estimation. The mathematical expression for updating the prediction covariance is:

[0134] P t|t1 = FX t|t-1

[0135] where F is the state transition matrix and X t|t-1 is the first predicted state of the first flying object.

[0136] In one embodiment, as Figure 4 shown, step 103 includes:

[0137] Step 401, in response to detecting that the first flying object has a pursuit action, obtain the position of the first flying object, the position of the unmanned aerial vehicle, and a preset support function;

[0138] In one embodiment, the preset support function is used to calculate the farthest point of the first flying object and the unmanned aerial vehicle in a given direction.

[0139] In one embodiment, to simplify the operation, the bird and the unmanned aerial vehicle can be regarded as spheres with a radius of zero, that is, the collision detection of two points is set. The position of the unmanned aerial vehicle can be expressed as u i = (longitude i , latitude i , high i ), where longitude i , latitude i , high i are the longitude, latitude, and altitude of the unmanned aerial vehicle (UAV) at the i-th moment, respectively.

[0140] Step 402, based on the position of the first flying object, the position of the unmanned aerial vehicle, and the preset support function, determine the movement path of the first flying object and the movement path of the unmanned aerial vehicle;

[0141] In one embodiment, before determining the motion path of the first flying object and the motion path of the drone based on the position of the first flying object, the position of the drone, and a preset support function, the Minkowski difference is defined. Specifically, the mathematical expression for determining the Minkowski difference is as follows:

[0142] C = P - U = {p i - u i | p i ∈ P, u i ∈ U}

[0143] where C is the Minkowski difference, P is the set of the first flying objects, U is the set of drones, p i is the i-th first flying object in P, and u i is the i-th drone in the set of drones.

[0144] In one embodiment, the mathematical expression for determining the motion path of the first flying object is as follows:

[0145] S P (d) = argmax x∈P (x · d)

[0146] where d is the direction vector in the flying direction, x · d represents the dot product of the point p i in the direction of d, and S P (d) is the point that returns the farthest point of the Minkowski difference p i in the direction of d.

[0147] In one embodiment, the mathematical expression for determining the motion path of the drone is as follows:

[0148] S U (d) = argmax x∈U (x · d)

[0149] where S U (d) is the point that returns the farthest point of the Minkowski difference u i in the direction of d.

[0150] Step 403: Based on the motion path of the first flying object and the motion path of the drone, calculate the Euclidean distance between the first flying object and the drone to perform collision detection on the drones in the first area.

[0151] In one embodiment, the mathematical expression for the preset support function is as follows:

[0152] S(P, U, d) = argmax x∈P (x · d) + argmax x∈U (x · -d)

[0153] Among them, the support function S(P, U, d) represents the point p i when moving in the direction d, whether it will collide with the point u i Is there a case where the Euclidean distance between the support points in the opposite direction -d is close to 0, that is:

[0154] ‖S(P, U, d)‖≈0

[0155] If S(P, U, d) updates the direction d and the positions of the two points p i and u i through multiple iterations and the distance between them is not close to 0, it indicates that the two points in space, the bird and the drone, do not intersect, that is, no collision will occur. In one embodiment, if the Minkowski difference includes the origin, it means that the first flying object and the drone intersect, that is, the bird and the drone will collide.

[0156] In one embodiment, as Figure 5 shown Figure 5 is the effect diagram of collision detection provided by the embodiment of the present disclosure. Among them, the first three rows in the figure are the collision situations, and the last row is the non-collision situation.

[0157] In one embodiment, by responding to detecting that the first flying object has a pursuit action, obtaining the position of the first flying object, the position of the drone, and a preset support function; based on the position of the first flying object, the position of the drone, and the preset support function, determining the movement path of the first flying object and the movement path of the drone; based on the movement path of the first flying object and the movement path of the drone, calculating the Euclidean distance between the first flying object and the drone, using model prediction and simulation technology, simulating the interaction scenario between the drone and the bird, determining whether there is a possibility of collision, and finally calculating the collision probability and predicting the collision time according to the collision determination result. Specifically, the collision probability can be obtained by dividing the number of collisions occurring in the statistical simulation process by the total number of simulations, and the time when the collision will occur can be predicted according to the simulation time step and the time point when the collision occurs.

[0158] In one embodiment, as Figure 6 shown, step 104 includes:

[0159] Step 601, if there is a collision risk between the first flying object and the drone, obtain the first coordinate of the drone, and the first coordinate of the drone is the coordinate when the drone is listed as a pursuit target by the first flying object;

[0160] In one embodiment, if the coordinate of the drone when it is listed as a pursuit target by the first flying object is U t =(longitude t , latitude t , hight ), use the first coordinate of the drone as the starting point q start , that is, q start = U t = (longitude t , latitude t , high t ), and start generating a random expansion tree T with the starting point as the root node.

[0161] Step 602: Obtain a random coordinate point. Using the first coordinate of the drone as the root node and the direction from the first coordinate of the drone to the random coordinate point as the growth direction, generate a new node according to a preset step size;

[0162] In one embodiment, a random coordinate point q can be obtained in the area where the drone is located rand .

[0163] In one embodiment, the preset step size can be determined according to the historical experience step size or can be customized as needed. The shorter the preset step size, the greater the computational amount in the entire route planning, and the more frequent the route changes; conversely, the longer the preset step size, the less the computational amount and the fewer the route changes. There may be a problem of insufficient path accuracy, and it is easy to ignore narrow channels or obstacles. The preset step size needs to balance path accuracy and convergence speed and be adjusted according to the environmental complexity to improve the algorithm performance.

[0164] In one embodiment, find the point q rand nearest to q in the random expansion tree T near , and calculate the distance between the random coordinate point and the point nearest to the random coordinate point. Its mathematical expression is:

[0165]

[0166] In one embodiment, expand the preset step size from q rand to q near to obtain the coordinates of the new node. Among them, the mathematical expression for calculating the new node is:

[0167]

[0168] where Δ is the preset step size.

[0169] Step 603: If the path from the first coordinate of the drone to the new node does not intersect with the movement path of the first flying object, add the new node to the random expansion tree until the iteration to generate new nodes stops when the preset condition is met;

[0170] In one embodiment, check from the random coordinate point q rand to the nearest point qnear The path of the bird intersects with the position state of the bird or the obstacle. If the path is valid (no collision with the bird / no obstacle blocking), q new Added to the random expansion tree T, the obstacle can be identified by the synaesthesia integrated base station, that is, the location of the obstacle is dynamically refreshed based on the location data of the flying bird continuously reported by the base station.

[0171] In one embodiment, the preset condition may be reaching the maximum number of iterations or finding the end point q of the preset path of the drone. goal , a preset path refers to a predetermined flight path.

[0172] Step 604: Fly along a first path between the first coordinates of the drone and the target point coordinates to prevent the first flying object from colliding with the drone, wherein the first path is a complete path in the random expansion tree tracing back from the target point coordinates to the first coordinates of the drone.

[0173] In one embodiment, if Figure 7 As shown, Figure 7 This is a rendering of the first path obtained according to an embodiment of the present disclosure.

[0174] In one embodiment, when a drone faces an emergency situation of being chased by a bird, if the drone only obtains the first path, it will not only greatly increase its energy consumption, but also increase its algorithm processing burden. Through the 5G-A telesensory integrated base station, the low-altitude flight command and control (Command and Control, C2) technology provided to the drone takes over the flight route planning task. With its powerful computing resources and stable mains power supply, the drone will be able to execute the re-planned effective escape route more quickly, thereby significantly reducing the risk of being crashed or captured.

[0175] In one embodiment, after flying along a first path between the first coordinates of the drone and the coordinates of the target point to prevent the first flying object from colliding with the drone, the anti-collision method further includes:

[0176] In response to the distance between the second coordinates of the UAV and the coordinates of the first flying object being not less than a preset threshold, the UAV flies to the target point coordinates according to a preset path.

[0177] In one embodiment, after executing the flight route replanned based on the RRT algorithm, the UAV adjusts the flight trajectory to ensure that a safe and sufficiently large distance d is maintained between the UAV and the bird. min (d min ≥100 meters), thus effectively avoiding the tracking range of flying birds.

[0178] In one embodiment, the second coordinate of the drone refers to the coordinate during the flight of the drone along the first path.

[0179] In one embodiment, the size of the preset threshold can be determined from the relevant escape determination scheme or through multiple escape tests.

[0180] In one embodiment, the preset threshold is used to indicate the critical value for the drone to successfully escape. For example, the distance d max (d max ≥ 500 meters) between the second coordinate of the drone and the coordinate of the first flying object. When this condition is met, it is determined that the drone has successfully escaped the pursuit of the bird, and the preset threshold at this time is 500.

[0181] In one embodiment, due to the physical exhaustion of the bird, misjudgment, or adjustment of flight strategy of the bird, when the bird decelerates or even completely stops the pursuit action, the drone can successfully escape the pursuit of the bird. Further, after the drone successfully escapes the pursuit of the bird, it flies along the preset path.

[0182] In one embodiment, as Figure 8 shown, Figure 8 This is a schematic diagram of the scenario of the anti-collision method provided by the embodiments of the present disclosure, including obtaining the state of the bird and the state of the drone by using a 5G-A communication and sensing integrated base station.

[0183] In summary, the solution provided by the present disclosure:

[0184] First, by establishing a pursuit action data model of the first flying object based on the observed state of the first flying object and the first predicted state of the first flying object, it is possible to accurately determine whether the first flying object has a pursuit action. If it is detected that the first flying object has a pursuit action, collision detection is performed on the drones within the first area. If there is a collision risk between the first flying object and the drone, it flies along the first path obtained according to the preset path planning algorithm, which can prevent the first flying object from colliding with the drone.

[0185] Second, by using the preset state estimation algorithm to determine the first predicted state of the first flying object based on the second predicted state of the first flying object and the observed state of the first flying object, it is possible to solve the problem of inaccurate first predicted state of the first flying object caused by the interference of noise on the observed state of the first flying object and the influence of system model uncertainty on the predicted state of the first flying object.

[0186] The following further illustrates the anti-collision method provided by the present disclosure with an application example:

[0187] As Figure 9 shown, Figure 9Flow schematic diagram of the anti-collision method provided for the application example of the present disclosure. The anti-collision method provided for the application example of the present disclosure includes the following steps:

[0188] Step 901, obtain the observation states of the first flying object at at least two consecutive moments, where the observation states of the first flying object at least include the position of the first flying object;

[0189] Step 902, determine the speed of the first flying object based on the positions of the first flying object at the two consecutive moments;

[0190] Step 903, determine the second predicted state of the first flying object based on the speed of the first flying object and a preset state transition matrix;

[0191] Step 904, determine the prediction covariance of the first flying object and the observation noise covariance of the first flying object based on the observation state of the first flying object;

[0192] Step 905, determine the Kalman filter gain based on a preset observation matrix, the prediction covariance of the first flying object, and the observation noise covariance of the first flying object;

[0193] Step 906, determine the first predicted state of the first flying object based on the observation state of the first flying object, the second predicted state of the first flying object, and the Kalman filter gain;

[0194] Step 907, establish a pursuit action data model of the first flying object based on the observation state of the first flying object and the first predicted state of the first flying object to determine whether the first flying object has a pursuit action;

[0195] Step 908, in response to detecting that the first flying object has a pursuit action, obtain the position of the first flying object, the position of the unmanned aerial vehicle, and a preset support function;

[0196] Step 909, determine the movement path of the first flying object and the movement path of the unmanned aerial vehicle based on the position of the first flying object, the position of the unmanned aerial vehicle, and the preset support function;

[0197] Step 910, calculate the Euclidean distance between the first flying object and the unmanned aerial vehicle based on the movement path of the first flying object and the movement path of the unmanned aerial vehicle to perform collision detection on the unmanned aerial vehicle within the first area;

[0198] Step 911, if there is a collision risk between the first flying object and the unmanned aerial vehicle, obtain the first coordinate of the unmanned aerial vehicle;

[0199] Step 912: Obtain random coordinate points. Taking the first coordinate of the drone as the root node and the direction from the first coordinate of the drone to the random coordinate points as the growth direction, generate new nodes according to a preset step size.

[0200] Step 913: If the path from the first coordinate of the drone to the new node does not intersect with the movement path of the first flying object, add the new node to the random expansion tree and stop iteratively generating new nodes until a preset condition is met.

[0201] Step 914: Fly along the first path from the first coordinate of the drone to the target point coordinate to prevent the first flying object from colliding with the drone.

[0202] Step 915: In response to the distance between the second coordinate of the drone and the coordinate of the first flying object being not less than a preset threshold, fly to the target point coordinate according to a preset path.

[0203] To implement the anti-collision method provided by the embodiments of the present disclosure, the embodiments of the present disclosure further provide an anti-collision device, as Figure 10 shown. Figure 10 FIG. is a schematic structural diagram of the anti-collision device provided by the embodiments of the present disclosure. The anti-collision device 1000 includes:

[0204] An acquisition unit 1001, configured to acquire the observation state of the first flying object and the first prediction state of the first flying object.

[0205] A model establishment unit 1002, configured to establish a pursuit action data model of the first flying object based on the observation state of the first flying object and the first prediction state of the first flying object to determine whether the first flying object has a pursuit action.

[0206] A collision detection unit 1003, configured to, in response to detecting that the first flying object has a pursuit action, perform collision detection on the drones within the first area, where the collision detection is used to detect whether there is a collision risk between the first flying object and the drones.

[0207] A path planning unit 1004, configured to, if there is a collision risk between the first flying object and the drone, fly along the first path obtained according to a preset path planning algorithm to prevent the first flying object from colliding with the drone.

[0208] In an embodiment, the acquisition unit 1001 is specifically configured to:

[0209] Acquire the observation states of the first flying object at at least two consecutive moments, where the observation state of the first flying object at least includes the position of the first flying object.

[0210] Determine the speed of the first flying object based on the positions of the first flying object at the two consecutive moments;

[0211] Determine the second predicted state of the first flying object based on the speed of the first flying object and a preset state transition matrix, where the second predicted state of the first flying object includes the speed of the first flying object at the next time and the position of the first flying object at the next time;

[0212] Determine the first predicted state of the first flying object by using a preset state estimation algorithm based on the second predicted state of the first flying object and the observed state of the first flying object.

[0213] In one embodiment, the processing unit 1001 is specifically configured to:

[0214] Determine the predicted covariance of the first flying object and the observed noise covariance of the first flying object based on the observed state of the first flying object;

[0215] Determine the Kalman filter gain based on a preset observation matrix, the predicted covariance of the first flying object, and the observed noise covariance of the first flying object;

[0216] Determine the first predicted state of the first flying object based on the observed state of the first flying object, the second predicted state of the first flying object, and the Kalman filter gain.

[0217] In one embodiment, the collision detection unit 1003 is specifically configured to:

[0218] In response to detecting that the first flying object has a pursuit action, obtain the position of the first flying object, the position of the unmanned aerial vehicle, and a preset support function;

[0219] Determine the movement path of the first flying object and the movement path of the unmanned aerial vehicle based on the position of the first flying object, the position of the unmanned aerial vehicle, and the preset support function;

[0220] Calculate the Euclidean distance between the first flying object and the unmanned aerial vehicle based on the movement path of the first flying object and the movement path of the unmanned aerial vehicle to perform collision detection on the unmanned aerial vehicle within the first area.

[0221] In one embodiment, the path planning unit 1004 is specifically configured to:

[0222] If there is a collision risk between the first flying object and the unmanned aerial vehicle, obtain the first coordinate of the unmanned aerial vehicle, where the first coordinate of the unmanned aerial vehicle is the coordinate when the unmanned aerial vehicle is listed as a pursuit target by the first flying object;

[0223] Obtain random coordinate points. Using the first coordinate of the drone as the root node and the direction from the first coordinate of the drone to the random coordinate points as the growth direction, generate new nodes according to a preset step size.

[0224] If the path from the first coordinate of the drone to the new node does not intersect with the movement path of the first flying object, add the new node to the random expansion tree, and stop iteratively generating new nodes until a preset condition is met.

[0225] Fly along the first path between the first coordinate of the drone and the target point coordinate to prevent the first flying object from colliding with the drone. The first path is the path in the random expansion tree that traces back from the target point coordinate to the first coordinate of the drone.

[0226] In an embodiment, the anti-collision device 1000 further includes a preset path flight unit, and the preset path flight unit is specifically configured to:

[0227] In response to the distance between the second coordinate of the drone and the coordinate of the first flying object being not less than a preset threshold, fly to the target point coordinate according to a preset path.

[0228] It should be noted that when the anti-collision device provided in the above embodiment performs anti-collision, only the above-mentioned division of each program module is used for illustration. In actual applications, the above-mentioned processing can be allocated to different program modules according to needs, that is, the internal structure of the anti-collision device is divided into different program modules to complete all or part of the above-described processing. In addition, the anti-collision device provided in the above embodiment and the anti-collision method embodiment provided in the present disclosure belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0229] Figure 11 This is a schematic diagram of the hardware composition structure of the electronic device provided in the embodiment of the present disclosure. As Figure 11 shown, the electronic device 1100 includes at least one processor 1102; and a memory 1101 communicatively connected to at least one processor 1102; wherein, the memory 1101 stores instructions executable by at least one processor 1102, and the instructions are executed by at least one processor 1102 to implement the steps of the anti-collision method of the embodiment of the present disclosure.

[0230] Optionally, the electronic device may specifically be the anti-collision device of the embodiment of the present application, and the electronic device can implement the corresponding processes implemented by the anti-collision device in each method of the embodiment of the present application. For the sake of brevity, it will not be repeated here.

[0231] It can be understood that the electronic device further includes a communication interface 1103. Each component in the electronic device is coupled together through a bus system 1104. It can be understood that the bus system 1104 is used to realize the connection and communication between these components. In addition to the data bus, the bus system 1104 further includes a power bus, a control bus, and a status signal bus. However, for the sake of clear illustration, in Figure 11 all kinds of buses are labeled as the bus system 1104.

[0232] It can be understood that the memory 1101 can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM, Read Only Memory), a programmable read-only memory (PROM, Programmable Read-Only Memory), an erasable programmable read-only memory (EPROM, Erasable Programmable Read-Only Memory), an electrically erasable programmable read-only memory (EEPROM, Electrically Erasable Programmable Read-Only Memory), a ferromagnetic random access memory (FRAM, ferromagnetic random access memory), a flash memory (Flash Memory), a magnetic surface memory, an optical disc, or a compact disc read-only memory (CD-ROM, Compact Disc Read-Only Memory); the magnetic surface memory can be a disk memory or a tape memory. The volatile memory can be a random access memory (RAM, Random Access Memory), which is used as an external cache. By way of example but not limitation, many forms of RAM are available, such as a static random access memory (SRAM, Static Random Access Memory), a synchronous static random access memory (SSRAM, Synchronous Static Random Access Memory), a dynamic random access memory (DRAM, Dynamic Random Access Memory), a synchronous dynamic random access memory (SDRAM, Synchronous Dynamic Random Access Memory), a double data rate synchronous dynamic random access memory (DDR SDRAM, Double Data Rate Synchronous Dynamic Random Access Memory), an enhanced synchronous dynamic random access memory (ESDRAM, Enhanced Synchronous Dynamic Random Access Memory), a synchronous link dynamic random access memory (SLDRAM, SyncLink Dynamic Random Access Memory), a direct rambus random access memory (DRRAM, Direct Rambus Random Access Memory).The memory 1101 described in the embodiments of the present invention is intended to include, but is not limited to, these and any other suitable types of memories.

[0233] The method disclosed in the above embodiments of the present disclosure can be applied to, or implemented by, the processor 1102. The processor 1102 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the integrated logic circuit in the hardware of the processor 1102 or by instructions in the form of software. The above-mentioned processor 1102 may be a general-purpose processor, a DSP, or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The processor 1102 can implement or execute each method, step, and logic block diagram disclosed in the embodiments of the present invention. The general-purpose processor may be a microprocessor or any conventional processor, etc. Combining the steps of the method disclosed in the embodiments of the present invention can be directly embodied as being executed and completed by a hardware decoding processor, or by a combination of hardware and software modules in the decoding processor. The software module may be located in a storage medium, and this storage medium is located in the memory 1101. The processor 1102 reads the information in the memory 1101 and combines its hardware to complete the steps of the foregoing method.

[0234] In an exemplary embodiment, the electronic device may be implemented by one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), FPGAs, general-purpose processors, controllers, MCUs, microprocessors, or other electronic components, and is used to execute the foregoing method.

[0235] The embodiments of the present disclosure also provide a non-transitory computer-readable storage medium storing computer instructions, and the computer instructions are used to cause a computer to execute the steps of the anti-collision method in the embodiments of the present invention when executed.

[0236] Optionally, the computer-readable storage medium can be applied to the anti-collision device in the embodiments of the present application, and the computer instructions cause the computer to execute the corresponding processes implemented by the anti-collision device in each method of the embodiments of the present application. For the sake of brevity, it will not be elaborated here.

[0237] The embodiments of the present disclosure also provide a computer program product, including a computer program, and the computer program implements the steps of the anti-collision method provided in the embodiments of the present invention when executed by a processor.

[0238] In several embodiments provided by the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined, or can be integrated into another system, or some features can be ignored, or not executed. In addition, the couplings, direct couplings, or communication connections between the various components shown or discussed can be through some interfaces. The indirect couplings or communication connections of devices or units can be electrical, mechanical, or other forms.

[0239] The units described above as separate components may or may not be physically separated. The components shown as units may or may not be physical units, that is, they can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0240] In addition, each functional unit in the embodiments of the present invention can be all integrated in a processing unit, or each unit can be separately used as a unit, or two or more units can be integrated in a unit. The above integrated units can be implemented in the form of hardware, or in the form of hardware plus software functional units.

[0241] Those of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions. The foregoing program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps including the above method embodiments. The foregoing storage medium includes various media that can store program codes, such as removable storage devices, ROM, RAM, magnetic disks, or optical discs.

[0242] Alternatively, if the above integrated units of the present invention are implemented in the form of software function modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the embodiments of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods of the various embodiments of the present invention. The foregoing storage medium includes various media that can store program codes, such as removable storage devices, ROM, RAM, magnetic disks, or optical discs.

[0243] As described above, it is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims described above.

Claims

1. A collision prevention method, characterized in that: include: Acquiring an observed state of a first flying object and a first predicted state of the first flying object; Based on the observed state of the first flying object and the first predicted state of the first flying object, establishing a pursuit action data model of the first flying object to determine whether the first flying object has a pursuit action; In response to detecting that the first flying object has a pursuit action, performing a collision detection on the drones in the first area, wherein the collision detection is used to detect whether there is a collision risk between the first flying object and the drone; If there is a risk of collision between the first flying object and the UAV, the first flying object is flown according to a first path obtained by a preset path planning algorithm to prevent the first flying object from colliding with the UAV.

2. The method according to claim 1, characterized in that: The obtaining of the observed state of the first flying object and the first predicted state of the first flying object comprises: Acquire an observation state of a first flying object at at least two consecutive moments, wherein the observation state of the first flying object at least includes a position of the first flying object; Determining the speed of the first flying object based on the positions of the first flying object at the two consecutive moments; Determine a second predicted state of the first flying object based on the speed of the first flying object and a preset state transfer matrix, wherein the second predicted state of the first flying object includes the speed of the first flying object at a next time and the position of the first flying object at a next time; Based on the second predicted state of the first flying object and the observed state of the first flying object, a first predicted state of the first flying object is determined using a preset state estimation algorithm.

3. The method according to claim 2, characterized in that The method of determining the first predicted state of the first flying object by using a preset state estimation algorithm based on the second predicted state of the first flying object and the observed state of the first flying object includes: Determining a predicted covariance of the first flying object and an observed noise covariance of the first flying object based on the observed state of the first flying object; Determining a Kalman filter gain based on a preset observation matrix, a predicted covariance of the first flying object, and an observed noise covariance of the first flying object; Based on the observed state of the first flying object, the second predicted state of the first flying object and the Kalman filter gain, a first predicted state of the first flying object is determined.

4. The method according to claim 1, characterized in that: In response to detecting that the first flying object has a pursuit action, performing collision detection on the UAV in the first area includes: In response to detecting that the first flying object has a pursuit action, obtaining a position of the first flying object, a position of the drone, and a preset support function; Determining a motion path of the first flying object and a motion path of the drone based on the position of the first flying object, the position of the drone, and a preset support function; Based on the motion path of the first flying object and the motion path of the drone, the Euclidean distance between the first flying object and the drone is calculated to perform collision detection on the drone in the first area.

5. The method according to claim 1, characterized in that If there is a collision risk between the first flying object and the drone, flying according to a first path obtained by a preset path planning algorithm to prevent the first flying object from colliding with the drone, includes: If there is a collision risk between the first flying object and the UAV, obtaining the first coordinates of the UAV, where the first coordinates of the UAV are the coordinates of the UAV when it is listed as a pursuit target by the first flying object; Obtain a random coordinate point, take the first coordinate of the drone as the root node, take the direction from the first coordinate of the drone to the random coordinate point as the growth direction, and generate a new node according to a preset step size; If the path from the first coordinate of the drone to the new node does not intersect with the motion path of the first flying object, the new node is added to the random expansion tree until the preset condition is met and the iterative generation of new nodes is stopped; Fly along a first path between the first coordinates of the drone and the target point coordinates to prevent the first flying object from colliding with the drone, wherein the first path is a path in the random expansion tree that traces back from the target point coordinates to the first coordinates of the drone.

6. The method according to claim 5, characterized in that After flying along a first path between the first coordinates of the drone and the coordinates of the target point to prevent the first flying object from colliding with the drone, the method further includes: In response to the distance between the second coordinates of the UAV and the coordinates of the first flying object being not less than a preset threshold, the UAV flies to the target point coordinates according to a preset path.

7. An anti-collision device, characterized in that: include: An acquisition unit, configured to acquire an observed state of the first flying object and a first predicted state of the first flying object; a model building unit, configured to build a pursuit action data model of the first flying object based on the observed state of the first flying object and the first predicted state of the first flying object to determine whether the first flying object has a pursuit action; a collision detection unit, configured to perform a collision detection on the UAV in the first area in response to detecting that the first flying object has a pursuit action, wherein the collision detection is used to detect whether there is a collision risk between the first flying object and the UAV; A path planning unit is used to fly according to a first path obtained by a preset path planning algorithm to prevent the first flying object from colliding with the drone if there is a risk of collision between the first flying object and the drone.

8. An electronic device, characterized in that: include: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium storing computer instructions, characterized in that: The computer instructions are used to cause the computer to execute the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, wherein the computer program implements the method according to any one of claims 1 to 6 when executed by a processor.

Images