Wireless Ultraviolet Light Cooperative Swarm UAV Multi-Target Tracking Method

Abstract

The present invention discloses a method for multi-target tracking of wireless ultraviolet light cooperative swarm drones, including: obtaining the relative positions and speeds of adjacent pursuit drones and the local drone through the ultraviolet light transceiver device carried by each pursuit drone; obtaining the current position of the target drone and predicting its trajectory; dividing the pursuit drones into coalition blocks with the same number as the target drones; under the condition that the predicted trajectory of the target drone meets the constraint conditions for successful pursuit by the pursuit drones, surrounding the target drone, and during the surrounding process, the pursuit drones adjust their speeds in real time according to the relative positions and speeds of adjacent pursuit drones and the local drone. Ensure the accuracy of the target drone's trajectory and improve the tracking and pursuit efficiency.

Description

Technical Field

[0001] The present invention belongs to the technical field of unmanned aerial vehicle (UAV) tracking methods, and relates to a multi-target tracking method for wireless ultraviolet light cooperative swarm UAVs. Background Art

[0002] In recent years, the status of unmanned combat systems in war has become increasingly prominent. The research on unmanned combat systems has gradually moved from theory to the battlefield. The war conflicts between some countries in recent years have proved that the role played by unmanned combat systems on the battlefield cannot be underestimated. As an advanced form of unmanned combat - unmanned swarm combat, it will inevitably have greater destructive power in war. The existing UAV tracking methods have relatively low tracking accuracy and cannot guarantee the pursuit efficiency. Summary of the Invention

[0003] The purpose of the present invention is to provide a multi-target tracking method for wireless ultraviolet light cooperative swarm UAVs, which solves the problem of relatively low tracking accuracy in the prior art.

[0004] The technical solution adopted by the present invention is that the multi-target tracking method for wireless ultraviolet light cooperative swarm UAVs includes the following steps:

[0005] Step 1: Obtain the relative position and speed between adjacent pursuit UAVs and the local UAV through the ultraviolet light transceiver device carried by each pursuit UAV;

[0006] Step 2: Obtain the current position of the target UAV and predict its trajectory;

[0007] Step 3: Divide the pursuit UAVs into coalition blocks with the same number as the target UAVs;

[0008] Step 4: Under the condition that the predicted trajectory of the target UAV satisfies the constraint conditions for the successful pursuit of the pursuit UAVs, conduct an encirclement and capture of the target UAV. During the encirclement and capture process, the pursuit UAVs adjust their speeds in real time according to the relative position and speed between adjacent pursuit UAVs and the local UAV.

[0009] The characteristics of the present invention also lie in:

[0010] Step 2 includes the following steps:

[0011] Step 2.1: Obtain the bounding box of the target UAV through the 3D target detection module carried by each pursuit UAV. Assume that n target UAVs are detected by the pursuit UAV at time t Each target UAV is represented by (x, y, z, l, w, h, θ, s), where (x, y, z) represents the position of the target UAV in space, (l, w, h) represents the size of the bounding box, and s represents the confidence level;

[0012] Step 2.2: Predict the trajectory of the target UAV. The trajectory of the i-th UAV at time t is expressed as:

[0013]

[0014] Among them, v x , v y , v z Indicates the speed of the target drone in three dimensions,

[0015] x est =x+v x ;

[0016] y est =y+v y ;

[0017] z est =z+v z ;

[0018] Step 2.3: Each target drone and trajectory Perform matching to obtain the corresponding trajectory of each target drone;

[0019] Step 2.4: Update the trajectory of the successfully matched target drone to obtain all updated trajectories at time t Then, according to the Bayes rule, the updated trajectory of each target drone is obtained:

[0020] T t k =(x',y',z',θ',l',w',h',s',v x ',v y ',v z '), k∈{1,2,...,w t}.

[0021] Step 3 includes the following steps:

[0022] Step 3.1: Based on the dual graph formed by the pursuit UAV and the target UAV, remove the edge with the smallest weight in the dual graph;

[0023] Step 3.2: Check if there is any node without any connection. If not, go back to step 3.1. Otherwise, mark the pursuer connected by the edge removed in this cycle as the alliance leader.

[0024] Step 3.3: Repeat steps 3.1-3.2 until each target drone is assigned to a pursuit drone coalition block.

[0025] The constraints in step 4 include:

[0026] Condition 1: The predicted trajectory T of the target UAV t k Passes through the Apollonius circle where any pursuing UAV is located;

[0027] Condition 2: The predicted trajectory T of the target UAV t k Falls within the convex polygon formed by the coalition blocks of the pursuing UAVs, and the adjacent pursuit - escape Apollonius circles formed by the target UAV and the pursuing UAVs satisfy pairwise intersection.

[0028] The ultraviolet light transceiver device is a hemispherical MIMO communication system.

[0029] The beneficial effects of the present invention are as follows: In the wireless ultraviolet light cooperative swarm UAV multi - target tracking method of the present invention, each UAV is equipped with an ultraviolet light transceiver device, and the use of the wireless ultraviolet light LED transceiver device enables real - time interaction of information between UAVs, improving the stability and reliability of communication between multiple UAVs and ensuring the smooth progress of the pursuit process; The 3D multi - target tracking method is used to track the target UAV cluster, and the target UAV is surrounded after meeting the encirclement conditions, ensuring the accuracy of the target UAV trajectory and improving the tracking efficiency; The pursuing UAV swarm is divided into coalition blocks with the same number as the target UAVs to ensure the pursuit efficiency. Brief Description of the Drawings

[0030] Figure 1 It is a schematic structural diagram of the ultraviolet light transceiver device in the wireless ultraviolet light cooperative swarm UAV multi - target tracking method of the present invention;

[0031] Figure 2 It is a bipartite graph formed by the pursuing UAVs and the target UAVs in the wireless ultraviolet light cooperative swarm UAV multi - target tracking method of the present invention;

[0032] Figure 3 It is a schematic diagram of multiple UAVs pursuing a single UAV in the wireless ultraviolet light cooperative swarm UAV multi - target tracking method of the present invention;

[0033] Figure 4 It is another schematic diagram of multiple UAVs pursuing a single UAV in the wireless ultraviolet light cooperative swarm UAV multi - target tracking method of the present invention. Detailed Embodiment

[0034] The present invention will be described in detail below in conjunction with the drawings and specific embodiments.

[0035] The wireless ultraviolet light cooperative swarm UAV multi - target tracking method includes the following steps:

[0036] Step 1: Obtain the relative positions and speeds of adjacent pursuing UAVs and the local UAV through the ultraviolet light transceiver device carried by each pursuing UAV;

[0037] As Figure 1 shown, the ultraviolet light transceiver device is a hemispherical MIMO communication system. During the process of multiple UAVs surrounding and capturing an enemy aircraft, different wavelengths of ultraviolet light emitted by the LED array of the airborne hemispherical MIMO structure are used to enable real-time interaction of information between UAVs through signal transmission and reception, ensuring the effectiveness of information transmission between UAVs and improving the stability and reliability of communication between multiple UAVs. The surface of each hemispherical LED array is divided into M latitudinal directions and N longitudinal directions, and LEDs are installed at the intersection of longitude and latitude, with omnidirectional reception at the top. The Figure 1 two hemispherical LED array devices in Figure 1 are respectively carried on two UAVs. Each ultraviolet light device has a corresponding signal sending end and signal receiving end. The UAV determines its relative position, speed, etc. with respect to its own aircraft during the circum-navigation process by receiving the signals sent by adjacent UAVs. Each LED has its own coding, which consists of latitudinal coding and longitudinal coding. This coding reflects the direction of the LED emission beam relative to the node itself; each LED can be independently controlled and can independently send information to complete the tasks of information sending and receiving.

[0038] Step 2: Use a 3D multi-target tracking method to track and predict the target UAV, obtain the current position of the target UAV, and predict its trajectory;

[0039] Step 2.1: Obtain the bounding box of the target UAV through the 3D target detection module carried by each pursuing UAV. Assume that n target UAVs are detected by the pursuing UAV at time t Each target UAV is represented by (x, y, z, l, w, h, θ, s), where (x, y, z) represents the position of the target UAV in space, (l, w, h) represents the size of the bounding box, and s represents the confidence level;

[0040] Step 2.2: Use a constant-speed Kalman filter to predict the trajectory of the detected target UAV. The trajectory of the i-th UAV at time t is expressed as:

[0041]

[0042] where, v x , v y , v z represent the speeds of the target UAV in three-dimensional directions,

[0043] x est = x + v x ;

[0044] y est = y + v y ;

[0045] z est= z + v z ;

[0046] Step 2.3. Use the Hungarian algorithm to calculate the 3D IOU (Intersection over Union) between each target UAV and the trajectory and perform matching to obtain the corresponding trajectory for each target UAV; After data association, four sets can be obtained:

[0047]

[0048]

[0049]

[0050]

[0051] Among them, T match and D match represent the correctly matched trajectory and detection result respectively, T umatch and D umatch represent the unmatched trajectory and detection result respectively, T match and T umatch are complementary subsets of T est , D match and D umatch are complementary subsets of D t .

[0052] Step 2.4. Update the trajectories of the target UAVs with successful matches to obtain all the updated trajectories at time t Then, according to Bayes' rule, obtain the updated trajectory of each target UAV:

[0053] T t k = (x', y', z', θ', l', w', h', s', v x ', v y ', v z '), k ∈ {1, 2,..., w t}, T t k is and 's weighted average.

[0054] Step 3. Divide the pursuit UAVs into coalition blocks with the same number as the target UAVs; Assume there are k target UAVs and n pursuit UAVs, so the n pursuit UAVs will be divided into k pursuit coalitions, and the n pursuit UAVs will perform one-on-one pursuit of the k target UAVs;

[0055] Step 3.1: According to the bipartite graph formed by the pursuit drones and the target drones, as Figure 2 shown, remove the edge with the minimum weight (the benefit of the pursuit drone connected to the target drone by the edge in the bipartite graph) in the bipartite graph;

[0056] Step 3.2: Check whether there are any nodes in the bipartite graph that have no connections. If not, go back to Step 3.1. Otherwise, mark the pursuers connected by the edges removed in this loop as coalition leaders;

[0057] Step 3.3: Repeat Steps 3.1 - 3.2 until each target drone is assigned to a pursuit drone coalition block.

[0058] Step 4: Under the condition that the predicted trajectories of the target drones satisfy two constraints for the successful pursuit by the pursuit drones, conduct a round-up of the target drones. During the round-up process, the pursuit drones adjust their speeds in real time according to the relative positions and speeds of adjacent pursuit drones and themselves.

[0059] As Figure 3 shown, E represents the position of the target drone e, and P1, P2…, P n respectively represent the positions of n pursuit drones p1, p2, …p n The circles O1, O2, …, O n respectively represent the centers of the pursuit Apollonius circles formed by p1, p2, …p n and e. To achieve no successful escape path for the target drone, it is necessary to make the adjacent circles among O1, O2, …, O n tangent or intersect pairwise. Therefore, the constraints include:

[0060] Condition 1: The predicted trajectory T t k of the target drone passes through the interior of the Apollonius circle where any pursuit drone is located;

[0061] Condition 2: As Figure 4 shown, where E represents the position of the target drone e, and P1, P2, P3, P4 respectively represent the positions of 4 pursuit drones p1, p2, p3, p4. The predicted trajectory T t k of the target drone falls within the convex polygon formed by the pursuit drone coalition block (4 pursuit drones), and the adjacent pursuit - escape Apollonius circles formed by the target drone and the pursuit drones satisfy pairwise intersection or tangency.

[0062] Through the above method, in the multi-target tracking method of the wireless ultraviolet light cooperative swarm drones of the present invention, each drone is equipped with an ultraviolet light transceiver device. The wireless ultraviolet light LED transceiver device enables real-time interaction of information between drones, improving the stability and reliability of communication between multiple drones and ensuring the smooth progress of the pursuit process. The 3D multi-target tracking method is used to track the target drone swarm, and the target drones are surrounded after meeting the encirclement conditions, ensuring the accuracy of the target drone trajectory and improving the tracking efficiency. The pursuit drone swarm is divided into coalition blocks with the same number as the target drones, ensuring the pursuit efficiency.

Claims

1. A multi-target tracking method for wireless ultraviolet collaborative swarm drones, characterized in that, It includes the following steps: Step 1: Obtain the relative positions and speeds of adjacent pursuit drones and the local drone through the ultraviolet transceiver devices carried by each pursuit drone; Step 2: Obtain the current position of the target drone and predict its trajectory. Specifically: Step 2.1: Obtain the bounding boxes of the target drones through the 3D target detection module carried by each pursuing drone. Assume that at time t, the pursuing drone detects n target drones , and each target drone is represented by . Among them, represents the position of the target drone in space, represents the size of the bounding box, and s represents the confidence level; Step 2.2: Predict the trajectory of the target drone. The trajectory of the i-th drone at time t is expressed as: ； Among them, represents the speed of the target UAV in three-dimensional directions. Step 2.

3. Match each target UAV with the trajectory to obtain the corresponding trajectory of each target UAV; Step 2.4: Update the trajectories of the successfully matched target UAVs to obtain all the updated trajectories at time t , and then, according to the Bayes rule, obtain the updated trajectory of each target UAV: ； Step 3: Divide the pursuit drones into coalition blocks with the same number as the target drones; Step 4: Under the condition that the predicted trajectory of the target drone meets the constraint conditions for successful pursuit by the pursuit drones, conduct an encirclement and capture of the target drone. During the encirclement and capture process, the pursuit drones adjust their speeds in real time according to the relative positions and speeds of adjacent pursuit drones and the local drone; The constraint conditions include: Condition 1: The predicted trajectory of the target UAV passes through the Apollonius circle where any pursuing UAV is located; Condition 2, Predicted Trajectory of the Target UAV falls inside the convex polygon formed by the pursuit UAV coalition blocks, and the adjacent pursuit-evasion Apollonius circles formed by the target UAV and the pursuit UAVs intersect pairwise.

2. The multi-target tracking method of the wireless ultraviolet light cooperative swarm drones according to claim 1, wherein, Step 3 includes the following steps: Step 3.1: According to the bipartite graph formed by the pursuit drones and the target drones, remove the edge with the smallest weight in the bipartite graph; Step 3.2: Check whether there are any nodes without any connections. If not, return to Step 3.

1. Otherwise, mark the pursuers connected by the edges removed in this loop as coalition leaders; Step 3.3: Repeat Steps 3.1 - 3.2 until each target drone is assigned to a pursuit drone coalition block.

3. The multi-target tracking method for wireless ultraviolet collaborative swarm drones according to claim 1, characterized in that The ultraviolet transceiver device is a hemispherical MIMO communication system.

Images