Distributed heuristic unmanned cluster brain-like swarm intelligence fusion search and capture method

Abstract

The present invention discloses a distributed heuristic unmanned cluster brain-like swarm intelligence fusion search and capture method, comprising: step 1: completing single target detection on the basis of a Yolov5 network, and predicting the position of a next frame of target on the basis of a DeepSort algorithm; step 2, on the basis of the distance difference and the area difference, determining, by mapping, navigation end point coordinates formula (1) of an intelligent agent in a physical space; step 3, constructing a diffusion-contraction coordinated chaotic mapping algorithm, and mapping the update of the navigation end point coordinates into a uniformly distributed chaotic space; step 4: on the basis of target detection confidence decision information and regional position information of all intelligent agents, fusing and generating global search information formula (3) of a search space formula (2); step 5: on the basis of the intelligent agent global search information, fusing various sensing data in a multi-agent system, so as to estimate target position information; and step 6: performing measurement on the basis of a real-time position of the intelligent agent and an obstacle position. The present invention has high capture efficiency and can accurately update a capture end point of the intelligent agent to complete a task of encircling a moving target.

Description

A Distributed-Inspired Unmanned Cluster Brain-Inspired Swarm Intelligence Fusion Search and Encirclement Method Technical Field

[0001] This invention relates to a distributed, heuristic-inspired, unmanned swarm brain-like swarm intelligence fusion search and capture method, and more particularly to a swarm intelligence decision-making technology using heuristic optimization algorithms, belonging to the field of artificial intelligence and swarm intelligence technology. Background Technology

[0002] In the research of multi-agent search technology, both domestic and international studies mainly focus on the pre-division of target areas and the setting of search routes for agents to complete the collaborative search task. However, the search task is dynamic and random, and in unmanned combat scenarios, environmental and map information is unknown, making it impossible to pre-set search routes. Decisions must be made in real time based on real-time information during the search process. In the research of multi-agent encirclement technology, international research mainly focuses on the optimization of control algorithms. For example, control strategies based on deep reinforcement learning have achieved significant results in improving the efficiency of agent collaborative encirclement. Domestic research mainly focuses on improving agent path planning algorithms. Path optimization based on genetic algorithms performs well in solving path planning problems in search and encirclement. By simulating natural selection and genetic mechanisms in biological evolution, genetic algorithms can effectively search for the optimal path, improving the efficiency and success rate of search and encirclement. In addition, domestic research also focuses on collaborative control strategies among agents, such as collaborative planning and collaborative execution, to improve the overall collaborative efficiency of search and encirclement.

[0003] Distributed-inspired swarm intelligence fusion technology is one of the main methods for multi-agent collaborative search and capture. However, in unmanned combat scenarios, it has the characteristics of wide coverage, high difficulty in search and capture, and difficulty in target location. At the same time, in unmanned combat scenarios, due to the high mobility of intelligent agents, problems such as motion ambiguity are easily generated, resulting in low target detection efficiency. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings and deficiencies of existing technologies by proposing a distributed-inspired, brain-like swarm intelligence fusion search and capture method for unmanned swarms. First, based on YOLOv5s target detection technology, it addresses the issue of real-time target movement by using predicted trajectories to track dynamic targets. Second, to address the problem of large search areas in unmanned combat scenarios, it improves the chaotic mapping algorithm based on the principle of region diffusion and contraction, mapping the search endpoint of unmanned agents in a coordinated diffusion and contraction manner. A decision-level brain-like fusion mechanism is then used to complete the decision fusion of multiple agents, reducing redundant region searches. Finally, to address the problem of low efficiency in capturing moving targets, a multi-agent information-level brain-like fusion capture mechanism is proposed, based on the role update mechanism of the Grey Wolf optimization algorithm and combined with dynamic Voronoi region partitioning. This mechanism accurately updates the capture endpoints of the agents, completing the task of encircling moving targets.

[0005] The technical solution adopted by this invention to solve its technical problem is: a distributed-inspired unmanned cluster brain-like swarm intelligence fusion search and capture method, which includes the following steps:

[0006] Step 1: Perform single target detection based on the Yolov5 network, and combine it with the DeepSort algorithm to predict the possible location of the target in the next frame, and calculate the normalized distance difference diff between the points in the predicted bounding box and the central axis. det The normalized area difference (diff) between the predicted bounding box area and the defined imaging area. area .

[0007] Step 2: Map the coordinates of the agent's navigation endpoint in physical space based on the distance difference and area difference.

[0008] Step 3: Construct a chaotic mapping algorithm that coordinates diffusion and contraction to map the update of the navigation endpoint coordinates to a uniformly distributed chaotic space.

[0009] Step 4: Based on the target detection confidence decision information and region location information of all agents, fuse them to generate the search space. Global search information Based on global search information, the dynamic decision-making agent performs both expansion and contraction search tasks.

[0010] Step 5: Based on the global search information of the agents, integrate various sensor data from the multi-agent system, estimate the target location information, divide the encirclement area based on the dynamic Voronoi diagram, and accurately determine the encirclement position of the agents.

[0011] Step 6: Based on the agent's real-time position and the obstacle's position, and simultaneously detect the target object V based on the Voronoi diagram. goal Area of ​​the figure and boundaries, if area Less than the threshold If all boundaries are adjacent to a Voronoi diagram, then the encirclement ends.

[0012] Furthermore, in step 1, single target detection is performed based on the Yolov5 network, and the DeepSort algorithm is used to predict the possible location of the target in the next frame. Specifically, this includes: the predicted target position (x' det ,y' det ,width' det ,height' det The following is an example: (x') det ,y' det ,width' det ,height' det = DeepSort(x) det ,y det ,width det height det )

[0013] Among them, (x det ,y det ,width det height det The YOLOv5 object detection bounding box is defined by its center point and dimensions. Then, the normalized distance difference (diff) between the center point of the bounding box and the central axis is calculated based on the image width. det :

[0014] Simultaneously calculate the normalized area difference between the area of ​​the detection box and the defined imaging area:

[0015] Among them, width img and height img These are the width and height of the image, respectively.

[0016] Furthermore, in step 2, the coordinates of the intelligent agent's navigation endpoint in physical space are mapped based on the distance difference and area difference. Specifically, this includes: first, estimating the distance the agent needs to move forward and the distance it needs to travel to the left and right based on the normalized distance difference and area difference.

[0017] in, and Δ represents the maximum forward and left / right movement distances, respectively; Δ is the minimum real number, indicating that the agent stops moving forward and backward when the detection box area is greater than the optimal area, while ensuring the feasibility of subsequent angle calculations; and the mapped navigation endpoint coordinates in the agent's global coordinate system. The specific expression is:

[0018] in, and The real-time coordinates of the intelligent agent. The deflection angle of the agent in the global coordinate system.

[0019] Furthermore, in step 3, constructing a diffusion-contraction coordinated chaotic mapping algorithm to map the update of the navigation endpoint coordinates to a uniformly distributed chaotic space specifically includes: firstly, mapping M... space ×M space The global space is divided into m×m local spaces. When iterating ut times, then u t M space ×u t M space The global space is divided into u t m×u t m local spaces Then, the search is performed outward from the starting point. Therefore, an improved chaotic mapping is needed to estimate the target position of each agent. The improved chaotic diffusion factor expression is as follows:

[0020] Where, μ t This is a control parameter, and its value range is (0,1). Given a chaotic mapping sequence with values ​​ranging from (0,1), in M... space ×M space A chaotic mapping of the agent's initial search endpoint within the search space, where the origin is located at M. space ×M space The midpoint of space, agent S i The expression for estimating the target location of the search is:

[0021] in, and These are the endpoint coordinates of the agent's search, u t Let u be the number of iterations in the chaotic mapping, and u be the number of iterations when generating the initial search endpoint. t The value is 1. and These are the chaos mapping factors for the x-axis and y-axis, respectively.

[0022] Further, step 3 includes: if an agent finds a target, i.e., a suspected target space exists, then a chaotic mapping is performed in the local space, mapping the chaotic factors of the x-axis and y-axis, both in the range [-1, 1], to the chaotic factor mapping range of the suspected target space. The mapping range of the chaotic contraction factor is:

[0023] The expression for the chaos contraction factor is:

[0024] Where w and h are the dimensions of the suspected target space, and the calculation formulas are as follows:

[0025] in, For the x-axis and y-axis intelligent agent S i The location of the target is then determined; at this point, the chaotic contraction factor is used to estimate the target location within the suspected target space. The target location estimation formula is as follows:

[0026] Furthermore, in step 4, the search space is generated by fusing the target detection confidence decision information and regional location information of all agents. Global search information Specifically, this includes the following: The decision information of the intelligent agent includes the following:

[0027] in, The target detection confidence of the intelligent agent. For agent S i Have we reached the navigation destination? For agent S i Regional location within the search space; global search information Specifically, it is expressed as follows:

[0028] Among them, C wh Representing local space The number of agents that detected the target, B wh Indicates that the local space has been searched. The number of intelligent agents; For local space The confidence level of target detection within the local space is determined by the local space. The target confidence scores detected by all agents within the system are fused to generate the target. Specifically, it is expressed as follows:

[0029] Furthermore, in step 4, the dynamic decision-making agent performs diffusion search and contraction search tasks based on global search information, specifically including: if All are 0, but there are intelligent agents. If True, meaning an agent has reached the search endpoint, then a spatial diffusion search is performed; if... All values ​​are 0, indicating no agent detected the target, but agents are present. If the value is True, it means an agent has reached the search endpoint. In this case, the search endpoints of all agents need to be updated, and the search range needs to be expanded, i.e., the number of iterations u. t Increase, at this time because the search range expands, local areas As this changes, it becomes necessary to inherit historical global information. The inheritance formula is as follows:

[0030] If it exists If the value is not 0, it indicates that there is an agent S. i Once the target is detected, a local agent space shrinkage search is performed. Corresponding local space As a suspected target space For local space The agents within eight local spaces around the target estimate the target location, thus achieving a shrinking search.

[0031] Furthermore, in step 5, the estimation of target location information based on the global search information of the intelligent agent, by fusing various sensor data from the multi-agent system and combining the Gray Wolf optimization algorithm, specifically includes the following: the perception information of the intelligent agent is as follows:

[0032] This includes lidar data. Real-time linear velocity and angular velocity Real-time deflection angle Combined with the state information of target detection Able to combine into intelligent agent S i Fusion data

[0033] Simultaneously receive status information from other intelligent agents. This is combined with its own state information to form a global state information array, specifically represented as follows:

[0034] Where N is the number of agents; based on the agent's own state vector The target detection probability is used to calculate the individual fitness value. Role assignment is completed based on the ranking of fitness values. The higher the target detection probability of an agent, the higher its fitness, indicating the emergence of a leader. The fitness value calculation formula is as follows:

[0035] Sort by fitness value to obtain Leader Wolf S α S β S δ And estimate the position of the target point directly in front:

[0036] in, To select the angle directly in front of the lidar, the coordinates of the nearest obstacle within that angle are used as the coordinates of the target object. The distance to the obstacle in the frontal angle is used to obtain the leader wolf's S. α S β S δ Each person's target point directly in front and This data is then merged into a precise target point location, as shown below:

[0037] S is calculated based on the angle with the minimum distance among the remaining angles not directly in front of the lidar. α S β S δ Coordinates of the nearest obstacle to each

[0038] in, and They are the intelligent agents S i The angle and distance with the smallest distance among the remaining angles not directly in front of the lidar data.

[0039] Furthermore, in step 5, the encirclement area is divided based on a dynamic Voronoi diagram, and the precise encirclement position of the agent specifically includes: the estimated target object coordinates (x, y, y). goal ,y goal ) and coordinates of nearby obstacles Including the coordinates of the alpha wolf, there are a total of seven coordinates. Initializing the Voronoi diagram yields seven discrete regions. and and the vertex set of each region and Calculate the target object V goal The center of gravity in the diagram The formula for calculating the center of gravity is as follows:

[0040] in, For discrete Voronoi diagram region V goalThe area is calculated, and then the wolf detection position update formula in the gray wolf optimization algorithm is improved using the centroid position of the target region to generate the updated wolf detection position. The improved formula is as follows:

[0041] Where A1 and C1 are the synergy coefficients: C1 = 2r2

[0042] Here, r1 and r2 are random vectors with values ​​between [0,1], and a is a control parameter that decreases linearly from 2 to 0 during the iteration process.

[0043] Furthermore, step 6 includes: based on the real-time position of the agent and the position of the obstacle, if an agent detects a target, updating the agent's role using the Gray Wolf Optimization Algorithm, and updating the coordinates of obstacles near the leader wolf. and and target coordinates Generate a new Voronoi diagram; if no agent detects the target, but the target is moving, update the agent's decision and the coordinates of the obstacle after a time T of movement. and and target coordinates Generate a new Voronoi diagram; simultaneously, detect the target object V. goal The area and boundary of the graph, if the area Less than the threshold If all boundaries are adjacent to a Voronoi diagram, then the encirclement ends. Beneficial effects:

[0044] 1. This invention is based on YOLOv5s target detection technology. To address the problem of real-time target motion, it uses predicted trajectories to complete the tracking task of dynamic targets.

[0045] 2. This invention addresses the problem of large search areas in unmanned combat scenarios by improving the chaotic mapping algorithm based on the principle of regional diffusion and contraction. It maps the search endpoint of unmanned intelligent agents in a coordinated diffusion and contraction manner and utilizes a decision-level brain-like fusion mechanism to complete the decision fusion of multiple intelligent agents, thereby reducing repeated regional searches.

[0046] 3. To address the problem of low efficiency in encircling moving targets, this invention proposes a multi-agent information-level brain-like fusion encirclement mechanism based on the role update mechanism of the gray wolf optimization algorithm and combined with dynamic Voronoi region partitioning. This mechanism accurately updates the encirclement endpoints of the agents, thereby completing the task of encircling the moving target. Attached Figure Description

[0047] Figure 1 is a flowchart of the distributed-inspired unmanned cluster brain-like swarm intelligence fusion search and capture method of the present invention.

[0048] Figure 2 is a layout diagram of the target detection box and the image of the present invention.

[0049] Figure 3 shows the search range as the number of iterations increases in this invention.

[0050] Figure 4 shows the end-of-encirclement marker of the present invention. Detailed Implementation

[0051] The invention will now be described in further detail with reference to the accompanying drawings.

[0052] As shown in Figure 1, this invention proposes a distributed-inspired unmanned cluster brain-like swarm intelligence fusion search and capture method, which includes the following steps:

[0053] Step 1: Perform single target detection based on the Yolov5 network, and combine it with the DeepSort algorithm to predict the position of the target in the next frame, and calculate the normalized distance difference diff between the points in the predicted bounding box and the central axis. det The normalized area difference (diff) between the predicted bounding box area and the defined imaging area. area ;

[0054] Step 2: Map the coordinates of the agent's navigation endpoint in physical space based on the distance difference and area difference.

[0055] Step 3: Construct a chaotic mapping algorithm that coordinates diffusion and contraction to map the update of the navigation endpoint coordinates to a uniformly distributed chaotic space;

[0056] Step 4: Based on the target detection confidence decision information and region location information of all agents, fuse them to generate the search space. Global search information Based on global search information, the dynamic decision-making agent performs both expansion search and contraction search tasks.

[0057] Step 5: Based on the global search information of the agents, integrate various sensor data from the multi-agent system, estimate the target location information, divide the encirclement area based on the dynamic Voronoi diagram, and accurately determine the encirclement position of the agents.

[0058] Step 6: Based on the agent's real-time position and the obstacle's position, and simultaneously detect the target object V based on the Voronoi diagram. goal Area of ​​the figure and boundaries, if area Less than the threshold If all boundaries are adjacent to a Voronoi diagram, then the encirclement ends.

[0059] The specific operation steps of the method of the present invention are as follows:

[0060] Step 1 of the present invention includes: performing single target detection based on the Yolov5 network, and combining it with the DeepSort algorithm to predict the possible location of the target in the next frame, wherein the predicted target location (x') is... det ,y' det ,width' det ,height' det The following is an example: (x') det ,y' det ,width' det ,height' det = DeepSort(x) det ,y det ,width det height det )

[0061] Among them, (x det ,y det ,width det height det The YOLOv5 object detection bounding box is defined by its center point and dimensions. Then, the normalized distance difference (diff) between the center point of the bounding box and the central axis is calculated based on the image width. det :

[0062] Simultaneously calculate the normalized area difference between the area of ​​the detection box and the defined imaging area:

[0063] Among them, width img and height img Figure 2 shows the width and height of the image, and the layout of the detection box and the image.

[0064] Step 2 of the present invention includes: performing single target detection based on a Yolov5 network, and mapping the navigation endpoint coordinates of the agent in physical space based on distance difference and area difference. First, based on the normalized distance difference and area difference, estimate the distance the agent needs to move forward and the distance it needs to travel to the left and right:

[0065] in, and Δ represents the maximum forward and left / right movement distances, respectively; Δ is the minimum real number, indicating that the agent stops moving forward and backward when the detection box area is greater than the optimal area, while ensuring the feasibility of subsequent angle calculations; and the mapped navigation endpoint coordinates in the agent's global coordinate system. The specific expression is:

[0066] in, and The real-time coordinates of the intelligent agent. The deflection angle of the agent in the global coordinate system.

[0067] Step 3 of the present invention includes: constructing a diffusion-contraction coordinated chaotic mapping algorithm to map the update of the navigation endpoint coordinates to a uniformly distributed chaotic space, firstly by M... space ×M space The global space is divided into m×m local spaces. When iteration u t Then, u t M space ×u t M space The global space is divided into u t m×u t m local spaces The iterative update space is shown in Figure 3. Then, the search is performed outward from the starting point. Therefore, an improved chaotic mapping is needed to estimate the target position of each agent. The expression for the improved chaotic diffusion factor is as follows:

[0068] Where, μ t This is a control parameter, and its value range is (0,1). Given a chaotic mapping sequence with values ​​ranging from (0,1), in M... space ×M space A chaotic mapping of the agent's initial search endpoint within the search space, where the origin is located at M. space ×M space The midpoint of space, agent S i The expression for estimating the target location of the search is:

[0069] in, and These are the endpoint coordinates of the agent's search along the x and y axes, respectively. t Let u be the number of iterations in the chaotic mapping, and u be the number of iterations when generating the initial search endpoint. t The value is 1. and These are the chaos mapping factors for the x-axis and y-axis, respectively.

[0070] If an agent finds a target in step 3 of this invention, i.e., a suspected target space exists, then a chaotic mapping is performed in the local space. The chaotic factors of the x-axis and y-axis, both ranging from [-1, 1], are mapped to the chaotic factor mapping range of the suspected target space. The mapping range of the chaotic contraction factor is:

[0071] The expression for the chaos contraction factor is:

[0072] Where w and h are the dimensions of the suspected target space, and the calculation formulas are as follows:

[0073] in, For the x-axis and y-axis intelligent agent S i The location of the target is then determined; at this point, the chaotic contraction factor is used to estimate the target location within the suspected target space. The target location estimation formula is as follows:

[0074] Step 4 of the present invention includes: fusing the target detection confidence decision information and regional location information of all agents to generate a search space. Global search information The decision information of the intelligent agent includes the following:

[0075] in, The target detection confidence of the intelligent agent. For agent S i Have we reached the navigation destination? For agent S i Regional location within the search space; global search information Specifically, it is expressed as follows:

[0076] Among them, C wh Representing local space The number of agents that detected the target, B wh Indicates that the local space has been searched. The number of intelligent agents; For local space The confidence level of target detection within the local space is determined by the local space. The target confidence scores detected by all agents within the system are fused to generate the target. Specifically, it is expressed as follows:

[0077] In step 4 of the present invention, the dynamic decision-making agent performs diffusion search and contraction search tasks based on global search information, specifically including: if All are 0, but there are intelligent agents. If True, meaning an agent has reached the search endpoint, then a spatial diffusion search is performed; if All values ​​are 0, indicating no agent detected the target, but agents are present. If the value is True, it means an agent has reached the search endpoint. In this case, the search endpoints of all agents need to be updated, and the search range needs to be expanded, i.e., the number of iterations u. t Increase, at this time because the search range expands, local areas As this changes, it becomes necessary to inherit historical global information. The inheritance formula is as follows:

[0078] If it exists If the value is not 0, it indicates that there is an agent S. i Once the target is detected, a local agent space shrinkage search is performed. Corresponding local space As a suspected target space For local space The agents within eight local spaces around the target estimate the target location, thus achieving a shrinking search.

[0079] Step 5 of the present invention includes: based on the global search information of the intelligent agent, fusing various types of sensor data in the multi-agent system, and combining the gray wolf optimization algorithm to estimate the target location information, the perception information of the intelligent agent is as follows:

[0080] This includes lidar data. Real-time linear velocity and angular velocity Real-time deflection angle Combined with the state information of target detection Able to combine into intelligent agent S i Fusion data

[0081] Simultaneously receive status information from other intelligent agents. This is combined with its own state information to form a global state information array, specifically represented as follows:

[0082] Where N is the number of agents; based on the agent's own state vector The target detection probability is used to calculate the individual fitness value. Role assignment is completed based on the ranking of fitness values. The higher the target detection probability of an agent, the higher its fitness, indicating the emergence of a leader. The fitness value calculation formula is as follows:

[0083] Sort by fitness value to obtain Leader Wolf S ɑ S β S δ And estimate the position of the target point directly in front:

[0084] in, To select the angle directly in front of the lidar, the coordinates of the nearest obstacle within that angle are used as the coordinates of the target object. The distance to the obstacle in the frontal angle is used to obtain the leader wolf's S. ɑ S β S δ Each person's target point directly in front and This data is then merged into a precise target point location, as shown below:

[0085] S is calculated based on the angle with the minimum distance among the remaining angles not directly in front of the lidar. ɑ S β S δ Coordinates of the nearest obstacle to each

[0086] in, and They are the intelligent agents S i The angle and distance with the smallest distance among the remaining angles not directly in front of the lidar data.

[0087] In step 5 of the present invention, the capture area is divided based on a dynamic Voronoi diagram, and the capture position of the agent is accurately determined by the estimated target object coordinates (x, y). goal ,y goal ) and coordinates of nearby obstacles Including the coordinates of the alpha wolf, there are a total of seven coordinates. Initializing the Voronoi diagram yields seven discrete regions. and and the vertex set of each region and

[0088] Calculate the target object V goal The center of gravity in the diagram The formula for calculating the center of gravity is as follows:

[0089] in, For discrete Voronoi diagram region V goal The area is then determined. The centroid position of the target region is then used to improve the wolf detection position update formula in the gray wolf optimization algorithm, generating the updated wolf detection position. The improved formula is as follows:

[0090] Where A1 and C1 are the synergy coefficients: C1 = 2r2

[0091] Here, r1 and r2 are random vectors with values ​​between [0,1], and a is a control parameter that decreases linearly from 2 to 0 during the iteration process.

[0092] Step 6 of the present invention includes: based on the real-time position of the agent and the position of the obstacle, if an agent detects a target, updating the agent's role using the Gray Wolf Optimization Algorithm, and updating the coordinates of obstacles near the leader wolf. and and target coordinates Generate a new Voronoi diagram; if no agent detects the target, but the target is moving, update the agent's decision and the coordinates of the obstacle after a time T of movement. and and target coordinates Generate a new Voronoi diagram; simultaneously, detect the target object V. goal The area and boundary of the graph, if the area Less than the threshold If all boundaries are adjacent to Voronoi diagrams, then the encirclement ends. A Voronoi diagram of the end of the encirclement is shown in Figure 4.

[0093] It should be noted that the above description of the embodiments is only for the purpose of helping to understand the method and core idea of ​​this application. For those skilled in the art, several improvements and modifications can be made to this application without departing from the principle of this application, and these improvements and modifications are also within the protection scope of the claims of this application.

Claims

1. A distributed-inspired unmanned swarm brain-like crowd intelligence fusion search and capture method, characterized in that, The method includes the following steps: Step 1: Perform single target detection based on the Yolov5 network, and combine it with the DeepSort algorithm to predict the position of the target in the next frame, and calculate the normalized distance difference diff between the points in the predicted bounding box and the central axis. det The normalized area difference (diff) between the predicted bounding box area and the defined imaging area. area ; Step 2: Map the coordinates of the agent's navigation endpoint in physical space based on the distance difference and area difference. Step 3: Construct a chaotic mapping algorithm that coordinates diffusion and contraction to map the update of the navigation endpoint coordinates to a uniformly distributed chaotic space; Step 4: Based on the target detection confidence decision information and region location information of all agents, fuse them to generate the search space. Global search information Based on global search information, the dynamic decision-making agent performs both expansion search and contraction search tasks. Step 5: Based on the global search information of the agents, integrate various sensor data from the multi-agent system, estimate the target location information, divide the encirclement area based on the dynamic Voronoi diagram, and accurately determine the encirclement position of the agents. Step 6: Based on the agent's real-time position and the obstacle's position, and simultaneously detect the target object V based on the Voronoi diagram. goal Area of ​​the figure and boundaries, if area Less than the threshold If all boundaries are adjacent to a Voronoi diagram, then the encirclement ends.

2. The distributed-inspired unmanned cluster brain-like swarm intelligence fusion search and capture method according to claim 1, characterized in that, In step 1, single target detection is performed based on the Yolov5 network, and the DeepSort algorithm is used to predict the possible location of the target in the next frame. Specifically, this includes: the predicted target position (x' det ,y' det ,width' det ,height' det The following is an example: (x') det ,y' det ,width' det ,height' det = DeepSort(x) det ,y det ,width det height det ) Among them, (x det ,y det ,width det height det The YOLOv5 object detection bounding box is defined by its center point and dimensions. Then, the normalized distance difference (diff) between the center point of the bounding box and the central axis is calculated based on the image width. det : Simultaneously calculate the normalized area difference between the area of ​​the detection box and the defined imaging area: Among them, width img and height img These are the width and height of the image, respectively.

3. The distributed-inspired unmanned swarm brain-like crowd intelligence fusion search and capture method according to claim 1, characterized in that, In step 2, the navigation endpoint coordinates of the intelligent agent in physical space are mapped based on the distance difference and area difference. Specifically, this includes: first, estimating the distance the agent needs to move forward and the distance it needs to travel to the left and right based on the normalized distance difference and area difference; in, and Δ represents the maximum forward and left / right movement distances, respectively; Δ is the minimum real number, indicating that the agent stops moving forward and backward when the detection box area is greater than the optimal area, while ensuring the feasibility of subsequent angle calculations; and the mapped navigation endpoint coordinates in the agent's global coordinate system. The specific expression is: in, and For the real-time coordinates of the intelligent agent. The deflection angle of the agent in the global coordinate system.

4. The distributed-inspired unmanned cluster brain-like swarm intelligence fusion search and capture method according to claim 1, characterized in that, In step 3, constructing a diffusion-contraction coordinated chaotic mapping algorithm to map the updated navigation endpoint coordinates to a uniformly distributed chaotic space specifically includes: firstly, mapping M... space ×M space The global space is divided into m×m local spaces. When iteration u t Then, u t M space ×u t M space The global space is divided into u t m×u t m local spaces Then, the search is performed outward from the starting point. Therefore, the target position of each agent needs to be estimated using an improved chaotic mapping. The improved chaotic diffusion factor expression is as follows: Where, μ t This is a control parameter, and its value range is (0,1). Given a chaotic mapping sequence with values ​​ranging from (0,1), in M... space ×M space A chaotic mapping of the agent's initial search endpoint within the search space, where the origin is located at M. space ×M space The midpoint of space, agent S i The expression for estimating the target location of the search is: in, and These are the endpoint coordinates of the agent's search along the x and y axes, respectively. t Let u be the number of iterations in the chaotic mapping, and u be the number of iterations when generating the initial search endpoint. t The value is 1. and These are the chaos mapping factors for the x-axis and y-axis, respectively.

5. The distributed-inspired unmanned cluster brain-like swarm intelligence fusion search and capture method according to claim 1, characterized in that, Step 3 includes: if an agent finds a target, i.e., a suspected target space exists, then a chaotic mapping is performed in the local space, mapping the chaotic factors of the x-axis and y-axis, both in the range of [-1, 1], to the chaotic factor mapping range of the suspected target space. The mapping range of the chaotic contraction factor is: The expression for the chaos contraction factor is: Where w and h are the dimensions of the suspected target space, and the calculation formulas are as follows: in, For the x-axis and y-axis intelligent agent S i The location of the target is then determined; at this point, the chaotic contraction factor is used to estimate the target location within the suspected target space. The target location estimation formula is as follows:

6. The distributed-inspired unmanned cluster brain-like swarm intelligence fusion search and capture method according to claim 1, characterized in that, In step 4, the search space is generated by fusing the target detection confidence decision information and regional location information of all agents. Global search information specific The decision information of the intelligent agent includes the following: in, The target detection confidence of the intelligent agent. For agent S i Have we reached the navigation destination? For agent S i Regional location within the search space; global search information Specifically, it is expressed as follows: Among them, C wh Representing local space The number of agents that detected the target, B wh Indicates that the local space has been searched. The number of intelligent agents; For local space The confidence level of target detection within the local space is determined by the local space. The target confidence scores detected by all agents within the system are fused to generate the target. Specifically, it is expressed as follows:

7. The distributed-inspired unmanned cluster brain-like swarm intelligence fusion search and capture method according to claim 1, characterized in that, In step 4, based on the global search information, the dynamic decision-making agent performs the diffusion search and contraction search tasks, specifically including: if All are 0, but there are intelligent agents. If True, meaning an agent has reached the search endpoint, then a spatial diffusion search is performed; if All values ​​are 0, indicating no agent detected the target, but agents are present. If the value is True, it means an agent has reached the search endpoint. In this case, the search endpoints of all agents need to be updated, and the search range needs to be expanded, i.e., the number of iterations u. t Increase, at this time because the search range expands, local areas As this changes, it becomes necessary to inherit historical global information. The inheritance formula is as follows: If it exists If the value is not 0, it indicates that there is an agent S. i Once the target is detected, a local agent space shrinkage search is performed. Corresponding local space As a suspected target space For local space The agents within eight local spaces around the target estimate the target location, thus achieving a shrinking search.

8. The distributed-inspired unmanned cluster brain-like swarm intelligence fusion search and capture method according to claim 1, characterized in that, In step 5, based on the global search information of the intelligent agent, various sensor data from the multi-agent system are fused, and the target location information is estimated using the Gray Wolf optimization algorithm. Specifically, this includes the following: the perception information of the intelligent agent is as follows: This includes lidar data. Real-time linear velocity and angular velocity Real-time deflection angle Combined with the state information of target detection Able to combine into intelligent agent S i Fusion data Simultaneously receive status information from other intelligent agents. This is combined with its own state information to form a global state information array, specifically represented as follows: Where N is the number of agents; based on the agent's own state vector The target detection probability is used to calculate the individual fitness value. Role assignment is completed based on the ranking of fitness values. The higher the target detection probability of an agent, the higher its fitness, indicating the emergence of a leader. The fitness value calculation formula is as follows: Sort by fitness value to obtain Leader Wolf S α S β S δ And estimate the position of the target point directly in front: in, To select the angle directly in front of the lidar, the coordinates of the nearest obstacle within that angle are used as the coordinates of the target object. The distance to the obstacle in the frontal angle is used to obtain the leader wolf's S. α S β S δ Each person's target point directly in front and And merged into a precise target point location X goal Specifically, it is expressed as follows: S is calculated based on the angle with the minimum distance among the remaining angles not directly in front of the lidar. α S β S β Coordinates of the nearest obstacle to each in, and They are the intelligent agents S i The angle and distance with the smallest distance among the remaining angles not directly in front of the lidar data.

9. The distributed-inspired unmanned cluster brain-like swarm intelligence fusion search and capture method according to claim 8, characterized in that, In step 5, the encirclement area is divided based on a dynamic Voronoi diagram. The precise encirclement position of the agent specifically includes: the estimated target coordinates (x, y, y). goal ,y goal ) and coordinates of nearby obstacles Including the coordinates of the alpha wolf, there are a total of seven coordinates. After initializing the Voronoi diagram, seven discrete regions can be obtained. and and the vertex set of each region and Calculate the target object V goal The center of gravity in the diagram The formula for calculating the center of gravity is as follows: in, For discrete Voronoi diagram region V goal The area is calculated, and then the wolf detection position update formula in the gray wolf optimization algorithm is improved using the centroid position of the target region to generate the updated wolf detection position. The improved formula is as follows: Where A1 and C1 are the synergy coefficients: C1＝2r2 Here, r1 and r2 are random vectors with values ​​between [0,1], and a is a control parameter that decreases linearly from 2 to 0 during the iteration process.

10. The distributed-inspired unmanned swarm brain-like crowd intelligence fusion search and capture method according to claim 10, characterized in that, Step 6 includes: based on the real-time position of the agent and the position of the obstacle, if an agent detects a target, updating the agent's role using the Gray Wolf Optimization Algorithm, and updating the coordinates of obstacles near the leader wolf. and and target object coordinates Generate a new Voronoi diagram; if no agent detects the target, but the target is moving, update the agent's decision and the coordinates of the obstacle after a time T of movement. and and target object coordinates Generate a new Voronoi diagram; simultaneously, detect the target object V. goal The area and boundary of the graph, if the area Less than the threshold If all boundaries are adjacent to a Voronoi diagram, then the encirclement ends.

Images