Unmanned cluster target search control method and device based on local information and medium

Abstract

The invention provides an unmanned cluster target search control method and device based on local information and a medium, and belongs to the technical field of swarm robot control. The method comprises the following steps: when a swarm robot explores a target, forming a tracking cluster by distributing a set tracking number of intelligent agents so as to track the target, so that a plurality of targets can be tracked at the same time; the method comprises the following steps of: tracking a cluster, establishing a local coordinate system, acquiring local information through the local coordinate system, generating a comprehensive concentration field about an obstacle and a target through a GRN upper network model, and simultaneously calculating the resultant velocity of potential field force of each agent in the tracking cluster and gradient information of each point in the comprehensive concentration field. Therefore, the operation speed of each agent in the tracking cluster is determined, the agents are enabled to realize rapid tracking of a target and avoid collision with an obstacle so as to complete group control, calculation is carried out through local information, the calculation amount is greatly reduced, and the tracking efficiency and the tracking precision are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of group robot control, in particular to an unmanned cluster target search control method, device and medium based on local information. Background Art

[0002] Swarm robot system tracking technology involves multiple cutting-edge fields, including artificial intelligence, machine learning, deep learning, computer vision, and sensor technology. Through these technologies, intelligent agents can achieve efficient target tracking in applications such as autonomous driving, drone tracking, and robot navigation.

[0003] In the existing technology, the complex tracking task is converted into a partially observable Markov decision process (POMDP) ​​problem, and an online solver is used to handle decisions in uncertain and dynamic environments, so as to achieve effective tracking of moving targets by the intelligent agent. However, the intelligent agent of this method mainly relies on the camera for visual odometer and landmark detection, which is easily disturbed by environmental factors. In addition, when applied in real time, the algorithm needs to process a large number of possible states and observations, resulting in high computational complexity, and the response of the intelligent agent may be delayed to a certain extent. Summary of the invention

[0004] The main purpose of the embodiments of the present application is to propose an unmanned cluster target search control method, device, and medium based on local information, so as to significantly reduce the amount of calculation and improve tracking efficiency and tracking accuracy.

[0005] To achieve the above-mentioned purpose, an embodiment of the present application proposes an unmanned cluster target search control method based on local information, the method comprising:

[0006] When the group robot explores the target, a set number of tracking agents are allocated from the group robots to form a tracking cluster to track the target;

[0007] A local coordinate system is established through the tracking cluster, and a first position information of the target, a second position information of the explored obstacle, and a third position information of each intelligent agent in the tracking cluster are determined according to the local coordinate system;

[0008] Generate a comprehensive concentration field about the obstacle and the target through a GRN upper network model according to the first position information and the second position information;

[0009] Determine, according to the third position information, a first set of agents having an attractive relationship with the current agent and a second set of agents having a repulsive relationship with the current agent, and calculate a resultant velocity of the potential field force of each agent in the tracking cluster according to the first set, the second set and the third position information;

[0010] The gradient information of each point in the comprehensive concentration field is determined by differential calculation, the running speed of each intelligent agent in the tracking cluster is determined according to the gradient information and the combined speed, and each intelligent agent in the tracking cluster is controlled to track the target according to the running speed.

[0011] Further, establishing a local coordinate system through the tracking cluster specifically includes:

[0012] In the tracking cluster, three agents that are not on the same straight line are selected, and the three agents are respectively used as a first coordinate agent, a second coordinate agent, and a third coordinate agent;

[0013] Taking the first coordinate intelligent body as the coordinate center o, the straight line from the first coordinate intelligent body to the second coordinate intelligent body is set as the x-axis, and the straight line perpendicular to the x-axis and pointing to one side of the third coordinate intelligent body is set as the y-axis to establish the local coordinate system xoy.

[0014] Further, determining the third position information of each agent in the tracking cluster according to the local coordinate system specifically includes:

[0015] Obtain a first distance from the first coordinate intelligent body to the second coordinate intelligent body, a second distance from the remaining intelligent bodies to the first coordinate intelligent body, and a third distance from the remaining intelligent bodies to the second coordinate intelligent body;

[0016] The third position information is determined based on the local coordinate system according to the first distance, the second distance and the third distance.

[0017] Furthermore, for the current agent R i , the calculating the combined velocity of the potential field force of each intelligent body in the tracking cluster according to the first set, the second set and the third position information specifically includes:

[0018] According to the first set G and the third position information, the current agent R is calculated by the gravitational field calculation formula. i The gravitational force U att (d ik ), wherein the first set G is d ik is the current agent R i With G k , k = 1, 2, ..., h, h is the number of agents in the first set G;

[0019] According to the second set F and the third position information, the current agent R is calculated by the repulsive field calculation formula. i The repulsive force Ureq (d ij ), wherein the second set F is d ij is the current agent R i With F j , j = 1, 2, ..., l, l is the number of agents in the second set F;

[0020] According to the gravitational force U att (d ik ) and the repulsive force U req (d ij ), calculate the current agent R i The combined velocity V of the potential force ip (t), where i = 1, 2, ..., m, where m is the number of agents in the tracking cluster, to calculate the combined speed of each agent in the tracking cluster.

[0021] Furthermore, determining the gradient information of each point in the comprehensive concentration field by differential calculation specifically includes:

[0022] Performing central difference on non-boundary points in the comprehensive concentration field to determine gradient information of the non-boundary points;

[0023] Forward difference and backward difference are performed on the boundary points in the comprehensive concentration field to determine the gradient information of the boundary points.

[0024] Furthermore, for the current agent R i , the step of determining the running speed of each agent in the tracking cluster according to the gradient information and the combined speed specifically includes:

[0025] According to the current agent R i The third position information of the integrated concentration field is used to determine the gradient information of the corresponding point;

[0026] According to the gradient information of the corresponding point and the current agent R i The corresponding combined speed is used to determine the running speed of the current agent, wherein i=1, 2, ..., m, and m is the number of agents in the tracking cluster, so as to calculate the running speed of each agent in the tracking cluster.

[0027] Furthermore, the unmanned cluster target search control method based on local information also includes:

[0028] Dividing the group of robots into a plurality of search groups, each of the search groups including a plurality of agents, and each of the search groups starting from a corresponding set initial position;

[0029] Each agent in the search group performs decentralized exploration in an unknown environment according to a Gaussian random walk strategy to explore the target.

[0030] Furthermore, when the group robot explores the target, allocating a set number of tracking agents from the group robot to form a tracking cluster specifically includes:

[0031] When the target is explored, for the current search group, the current search group communicates with the search group that has explored the target to determine the number of agents currently tracking the target;

[0032] Determine whether the tracking number of the intelligent agent is less than the set tracking number;

[0033] If so, the allocation number is determined based on the tracking number of the agents and the set tracking number, and the allocated number of agents are separated from the current search group to track the target to form the tracking cluster, and the remaining agents in the current search group perform decentralized exploration according to the Gaussian random walk strategy.

[0034] To achieve the above-mentioned purpose, another aspect of an embodiment of the present application proposes an electronic device, which includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, it implements the unmanned cluster target search control method based on local information as described above.

[0035] To achieve the above-mentioned purpose, another aspect of an embodiment of the present application proposes a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the unmanned cluster target search control method based on local information as described above.

[0036] The embodiments of the present application include at least the following beneficial effects: The present application provides an unmanned cluster target search control method, device, and medium based on local information. The scheme forms a tracking cluster by allocating a set number of tracking agents to track the target, so that multiple targets can be tracked at the same time; a local coordinate system is established, local information is obtained through the local coordinate system, and a GRN upper network model is introduced to construct a comprehensive concentration field, and the combined velocity of the potential field force of each agent and the gradient information of each point in the comprehensive concentration field are calculated at the same time, so as to determine the running speed of each agent in the tracking cluster, so that the agent can achieve rapid tracking of the target while avoiding collision with obstacles to complete group control, and calculate through local information, thereby greatly reducing the amount of calculation and improving tracking efficiency and tracking accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1It is a flowchart of an unmanned cluster target search control method based on local information provided by an embodiment of the present application;

[0038] Figure 2 It is a schematic diagram of a comprehensive concentration field of an unmanned cluster target search control method based on local information provided by an embodiment of the present application;

[0039] Figure 3 It is a schematic diagram of a tracking cluster tracking multiple target exploration phase of an unmanned cluster target search control method based on local information provided by an embodiment of the present application;

[0040] Figure 4 It is a schematic diagram of multiple target tracking stages of a tracking cluster tracking of an unmanned cluster target search control method based on local information provided by an embodiment of the present application. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is further described in detail below in conjunction with the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the embodiments of the present application. They are only examples of devices and methods consistent with some aspects of the embodiments of the present application as detailed in the attached claims.

[0042] It is understood that the terms "first", "second", etc. used in this application can be used to describe various concepts in this article, but unless otherwise specified, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another concept. For example, without departing from the scope of the embodiment of the present application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the words "if" and "if" as used herein can be interpreted as "at the time of" or "when" or "in response to determination".

[0043] The terms "at least one", "multiple", "each", "any", etc. used in this application, at least one includes one, two or more, multiple includes two or more, each refers to each of the corresponding multiple, and any refers to any one of the multiple.

[0044] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0045] In order to make the inventive concept of the present application easy to understand, before describing the embodiments of the present application in detail, the English abbreviations (terms) / related concepts involved in the embodiments of the present application are first described. The English abbreviations (terms) / related concepts involved in the embodiments of the present application are subject to the following explanations.

[0046] Gene Regulatory Network (GRN) is a network that describes the interaction between genes and their regulatory factors in an organism. It can apply the gene expression mechanism in biological morphogenesis to swarm robot control.

[0047] As described in the background technology, in the prior art, the intelligent agent mainly relies on cameras for visual odometer and landmark detection, which is easily interfered by environmental factors. In real-time applications, the algorithm needs to process a large number of possible states and observations, resulting in high computational complexity, and the response of the intelligent agent may be delayed to a certain extent.

[0048] Based on this, an embodiment of the present invention proposes an unmanned cluster target search control method, device, and medium based on local information, so as to significantly reduce the amount of calculation and improve tracking efficiency and tracking accuracy.

[0049] Figure 1 is an optional flow chart of an unmanned cluster target search control method based on local information provided in an embodiment of the present application. Figure 1 The method may include but is not limited to including S100 to S500.

[0050] S100, when the swarm robot explores the target, a set number of tracking agents are allocated from the swarm robots to form a tracking cluster to track the target;

[0051] S200, establishing a local coordinate system by tracking the cluster, and determining first position information of the target, second position information of the explored obstacle, and third position information of each intelligent agent in the tracking cluster according to the local coordinate system;

[0052] S300, generating a comprehensive concentration field of obstacles and targets through a GRN upper network model according to the first position information and the second position information;

[0053] S400, determining, based on the third position information, a first set of agents having an attractive relationship with the current agent and a second set of agents having a repulsive relationship with the current agent, and calculating a combined velocity of the potential field force of each agent in the tracking cluster based on the first set, the second set and the third position information;

[0054] S500, through differential calculation, determine the gradient information of each point in the comprehensive concentration field, determine the running speed of each intelligent agent in the tracking cluster according to the gradient information and the combined speed, and control each intelligent agent in the tracking cluster to track the target according to the running speed.

[0055] S100 to S500 shown in the embodiment of the present application, by allocating a set number of tracking agents to form a tracking cluster to track the target, can track multiple targets at the same time; establish a local coordinate system, obtain local information through the local coordinate system, and introduce the GRN upper network model to construct a comprehensive concentration field, and at the same time calculate the combined velocity of the potential field force of each agent and the gradient information of each point in the comprehensive concentration field, so as to determine the running speed of each agent in the tracking cluster, so that the agent can achieve rapid tracking of the target while avoiding collision with obstacles to complete group control.

[0056] In S100, when a target is discovered, a set number of tracking agents are allocated from the group of robots, and the set number of tracking agents form a tracking cluster, which tracks the target.

[0057] In one embodiment, the group of robots is equally divided into several search groups, and each search group starts from its own set initial position and disperses to explore in different directions. During the search process, after a search group explores the target, the search group that has explored the target communicates with the intelligent agents near the target to determine the number of intelligent agents currently involved in tracking the target M. It is judged whether the number of intelligent agent tracking M meets the set tracking number N. When the number of intelligent agent tracking M is less than the set tracking number N, the allocation number (NM) is determined, and the intelligent agents (NM) are divided to the search group that has explored the target to form a tracking cluster, so that the intelligent agents in the tracking cluster meet the set tracking number to ensure that the target is fully tracked.

[0058] By fixing the number of agents tracking each target and forming a tracking cluster, when the agent resources are sufficient, multiple targets can be tracked at the same time to ensure that each target is fully monitored.

[0059] In S200, when the tracking cluster is tracking, in order to determine the coordinates of the target and obstacles as inputs of the gene regulatory network, a local coordinate system covering the target and all obstacles needs to be established.

[0060] In the tracking cluster, a local coordinate system is established, and the coordinates of the target, the coordinates of the explored obstacles and the coordinates of each intelligent agent in the tracking cluster are determined through the local coordinate system, that is, the first position information, the second position information and the third position information are determined, so as to obtain local information and reduce dependence on global information.

[0061] In S300, for the tracking cluster, the first position information of the target and the second position information of the obstacle are used as inputs of the GRN upper network model through the local coordinate system established in S200, and a comprehensive concentration field based on the target position information and the obstacle position information is generated to achieve obstacle avoidance and target tracking of the tracking cluster in the tracking stage.

[0062] In S400, when the intelligent agent is tracking the target, since the information of the obstacle has been integrated in the comprehensive concentration field, the obstacle avoidance in the tracking process can be solved by the collision avoidance of the intelligent agent.

[0063] By tracking the third position information of each agent in the cluster, for the current agent in the tracking cluster, all agents that have an attractive relationship and a repulsive relationship with the current agent are counted, thereby determining a first set and a second set. The first set is the agents that have an attractive relationship with the current agent, and the second set is the agents that have a repulsive relationship with the current agent.

[0064] Through the first set, the second set and the third position information, the combined speed of the potential field force of each intelligent body in the tracking cluster is determined, so as to introduce the artificial potential field method for obstacle avoidance based on the comprehensive concentration field, and further improve the efficiency of tracking the target.

[0065] In S500, since the integrated concentration field is continuous, but the integrated concentration field is discrete when stored in the terminal, the gradient is replaced by difference to calculate the gradient information of each point in the integrated concentration field.

[0066] The combined speed of the potential field force superposition and the gradient information of each point are obtained through S400, the running speed of each intelligent agent in the tracking cluster is determined, and each intelligent agent in the tracking cluster is controlled to track the target according to the running speed, so that based on the gradient method of the environment, each intelligent agent can move to the place with high concentration gradient, that is, move to the first position information of the target, so as to achieve obstacle avoidance while tracking the target.

[0067] In some embodiments of S200 of the present invention, establishing a local coordinate system by tracking the cluster and determining the third position information of each agent in the tracking cluster according to the local coordinate system includes the following steps:

[0068] S210, in the tracking cluster, select three intelligent agents that are not on the same straight line, and use the three intelligent agents as the first coordinate intelligent agent, the second coordinate intelligent agent, and the third coordinate intelligent agent, respectively.

[0069] S220, taking the first coordinate intelligent body as the coordinate center o, setting the straight line from the first coordinate intelligent body to the second coordinate intelligent body as the x-axis, and setting the straight line perpendicular to the x-axis and pointing to one side of the third coordinate intelligent body as the y-axis, to establish a local coordinate system xoy.

[0070] S230, obtaining a first distance from the first coordinate intelligent body to the second coordinate intelligent body, a second distance from the remaining intelligent bodies to the first coordinate intelligent body, and a third distance from the remaining intelligent bodies to the second coordinate intelligent body.

[0071] S240, determining third position information of the remaining intelligent agents according to the first distance, the second distance and the third distance.

[0072] In S210, three agents that are not on the same straight line are selected from the tracking cluster and are respectively used as the first coordinate agent R1, the second coordinate agent R2 and the third coordinate agent R3.

[0073] In S220, the first coordinate intelligent body R1 is taken as the coordinate center o, the straight line connecting the first coordinate intelligent body R1 and the second coordinate intelligent body R2 is taken as the x-axis, and the straight line perpendicular to the x-axis and pointing to one side of the third coordinate intelligent body R3 is set as the y-axis to establish the local coordinate system xoy.

[0074] In S230, a first distance d between the first coordinate agent R1 and the second coordinate agent R2 is obtained. 12 , get the second distance d from the remaining agents to the first coordinate agent 1i and the third distance d from the rest of the agents to the second coordinate agent 2i .

[0075] In S240, the local coordinate system xoy is used according to the first distance d 12 , the second distance d 1i and the third distance d 2i , determine the third position information of the agent, the third position information of the agent is as follows:

[0076]

[0077] Wherein, i=1, 2, ..., m, m is the number of agents in the tracking cluster, and θ is the angle between the agent and the x-axis. Through S220, the third position information (0, 0) of the first coordinate agent R1 and the third position information (d 12 , 0).

[0078] Through S210 to S240, by tracking the intelligent agents in the cluster, a local coordinate system xoy is established between the intelligent agents, and the local coordinate system xoy covers the target and all obstacles, which is convenient for obtaining local information to achieve group control based on local information.

[0079] In some embodiments of S300 of the present invention, generating a comprehensive concentration field about obstacles and targets through a GRN upper layer network model according to the first position information and the second position information includes the following steps:

[0080] S310, based on the first position information p1 of the target and the second position information p2 of the detected obstacle n , a first concentration field g1 regarding the first position information p1 and the second position information p2 and a second concentration field g2 regarding the second position information p2 are generated.

[0081] S320, generating a comprehensive concentration field through the first concentration field g1 and the second concentration field g2.

[0082] In the above embodiment, the tracking cluster determines the first position information p1 of the target and the second position information p2 of the explored obstacle through the local coordinate system xoy, which is expressed as follows:

[0083]

[0084] Wherein, p is the total matrix including the first position information p1 and the second position information p2; (T x ,T y ) represents the first position information p1 of the assigned target, (O s,x ,O s,y )(s=1, 2, ..., z) represents the second position information p2 of the explored obstacles, and z is the number of explored obstacles.

[0085] The upper concentration field of GRN is calculated by the formula:

[0086]

[0087]

[0088] Among them, g1 is the concentration field formed by combining the target position information and the obstacle position information, that is, the first concentration field about the first position information p1 and the second position information p2; g2 is the concentration field formed by the obstacle position, that is, the second concentration field about the second position information p2; g3 represents the comprehensive concentration field that combines g1 and g2, corresponding to the generation of the tracking mode. θ1 and θ2 are the thresholds of the S-type function. k is the concentration field influencing factor.

[0089] Through S310 and S320, for the tracking cluster that is tracking the target, the position information of the explored obstacles and targets in the local coordinate system xoy is used as the input of the GRN upper network model to generate a comprehensive concentration field based on the target and obstacle position information. The gene regulatory network model is introduced as the selection basis for the intelligent agent tracking. The comprehensive concentration field is generated through the detected target information and obstacle information as a reference for the intelligent agent to track the target, so as to achieve flexible collision avoidance of obstacles and rapid tracking of targets.

[0090] In some embodiments of S400 of the present invention, the calculation process of the resultant velocity of the potential field force specifically includes:

[0091] S410, based on the third position information, counting the agents that have a gravitational field with the current agent to form a first set, and counting the agents that have a repulsive field with the current agent to form a second set.

[0092] In this embodiment, when tracking a target, since the information of obstacles has been integrated into the comprehensive concentration field, in the tracking cluster, only the intelligent agents in the tracking cluster and the collision avoidance of intelligent agents inside and outside the cluster need to be paid attention to. i , count all the agents with which they have gravitational and repulsive fields, the expression is as follows:

[0093]

[0094] Among them, G is all the agents related to the current agent R i The set of agents with gravitational fields between them, that is, the first set, F is all agents with gravitational fields between them and the current agent R i The set of agents among which there is a repulsive field is the second set.

[0095] S420, for the current agent R i , according to the first set G and the third position information, calculate the current agent R through the gravitational field calculation formula i The gravitational force U att (d ik ).

[0096] In this embodiment, for the current agent R i , through the third position information, determine the current agent R i With G k The distance d ik , d ik is a vector, and its direction is the agent G in the first set G k Point to the current agent R i , G kis an element in the first set G, that is, an agent in the first set G, k = 1, 2, ..., h, h is the number of agents in the first set G, and the gravitational field is calculated by the formula:

[0097]

[0098] Calculate the current agent R i The gravitational force U att (d ik ), where d att To maintain the maximum distance, the current agent R i The distance d from other agents ik Greater than d att When , gravity will appear, and η is the proportional gain coefficient.

[0099] S430, for the current agent R i According to the second set F and the third position information, the current agent R is calculated by the repulsive field calculation formula. i The repulsive force U req (d ij ).

[0100] In this embodiment, for the current agent R i , through the third position information, determine the current agent R i With F j The distance d ij , d ij is a vector, and its direction is the agent F in the second set F j Point to the current agent R i , F j is an element in the second set F, that is, an agent in the second set F, j = 1, 2, ..., l, l is the number of agents in the first set F, and the repulsive field is calculated by the formula:

[0101]

[0102] Calculate the current agent R i The repulsive force U req (d ij ), where d req is the minimum safe distance, the current agent R i The distance d from other agents ij Less than d att When , repulsion will appear, and λ is the proportional gain coefficient.

[0103] S440, for the current agent R i , according to the gravitational force U att (d ik ) and repulsive force dUreq (d ij ), calculate the current agent R i The combined velocity V of the potential force ip (t).

[0104] In this embodiment, for the current agent R i , calculated through S420 and S430, the gravitational force U is obtained att (d ik ) and repulsive force U req (d jj ), the resultant force field can be calculated as:

[0105]

[0106] Where i = 1, 2, ..., m, m is the number of agents in the tracking cluster, G k is an element in the first set G, that is, an agent in the first set G, F j is an element in the second set F, that is, an agent in the second set F.

[0107] Through the resultant force field, the potential field force and velocity are solved, and the formula is as follows:

[0108]

[0109] V ip (t) = E(i)

[0110] Where F(i) is the resultant force of the resultant force field, V ip (t) is the current agent R i The combined velocity of the potential field forces, and so on, can determine the combined velocity of the potential field forces of each intelligent agent.

[0111] Through S410 to S440, the gene regulatory network model is introduced to track the target, and the artificial potential field method is introduced for collision avoidance between intelligent agents to further improve the efficiency of exploration and tracking. On the basis of the comprehensive concentration field, the artificial potential field method is used for obstacle avoidance to reduce the amount of obstacle avoidance calculation.

[0112] Reference Figure 2 In some embodiments of S500 of the present invention, the process of determining the running speed of each agent in the tracking cluster specifically includes:

[0113] S510, performing central difference on non-boundary points in the integrated concentration field to determine gradient information of the non-boundary points.

[0114] In this embodiment, refer to Figure 2, the gradient corresponding to the target in the concentration field is positive, while the gradient corresponding to the obstacle in the concentration field is negative, then the gradient method based on the environment can not only track the target, but also avoid obstacles.

[0115] Since the integrated concentration field is continuous, but the integrated concentration field is discrete when stored on the terminal, the difference is used instead of the gradient, and the central difference method is used for the non-boundary points in the integrated concentration field. For a non-boundary point (x p ,y q ), the concentration gradient formula at this point is as follows:

[0116]

[0117] Among them, g(x,y) is the concentration field generated by the upper layer, For the point (x p ,y q ), that is, the concentration gradient at point (x p ,y q )’s gradient information; g(x p+1 ,y q )、g(x p-1 ,y q ) is g(x p ,y q ) the next point and the previous point along the positive direction of the X-axis of the integrated concentration field;

[0118] g(x p ,y q+1 )、g(x p ,y q-1 ) is g(x p ,y qw ) refers to the next and previous points along the positive Y-axis of the integrated concentration field.

[0119] S520, performing forward difference and backward difference on the boundary points in the integrated concentration field to determine the gradient information of the boundary points.

[0120] In this embodiment, when (x p ,y q ) is located at the boundary of the integrated concentration field, forward difference and backward difference are required. For the X-axis of the integrated concentration field, the gradient formula at the left and right boundaries is as follows:

[0121]

[0122] Among them, when x p When it is located at the left boundary of the positive direction of the X axis of the integrated concentration field, the forward difference method is used. pWhen the point is located at the right boundary of the positive direction of the X axis of the integrated concentration field, the backward difference method is used. The y value of the boundary point on the left boundary of the X axis of the integrated concentration field is q Greater than the y of the boundary point on the right boundary of the X-axis of the integrated concentration field q .

[0123] For the Y-axis of the integrated concentration field, the gradient formula at the left and right boundaries is as follows:

[0124]

[0125] Among them, when y q When it is located at the left boundary of the positive direction of the Y axis of the integrated concentration field, the forward difference method is used. q When the point is on the right boundary of the positive direction of the Y axis of the integrated concentration field, the backward difference method is used. The x value of the boundary point on the left boundary of the Y axis of the integrated concentration field is p is less than the x of the boundary point on the right boundary of the Y axis of the integrated concentration field p .

[0126] Through S510 and S520, the gradient information of the entire integrated concentration field can be obtained by differentiating the boundary points and non-boundary points of the integrated concentration field.

[0127] S530, for the current agent, according to the current agent R i The third position information is used to determine the gradient information of the corresponding point in the integrated concentration field.

[0128] In this embodiment, for the current agent R i , according to the current agent R i The third position information of the current agent R i The coordinates of the local coordinate system xoy correspond to the gradient of the point of the integrated concentration field That is, the gradient information of the corresponding point

[0129] S540, determining the running speed of the current agent according to the gradient information of the corresponding point and the combined speed corresponding to the current agent.

[0130] In this embodiment, for the current agent R i , according to the gradient information of the corresponding point determined in S540 and S400 to obtain the current agent R i The corresponding total speed V ip (t), determine the running speed of the current agent, the speed formula is as follows:

[0131]

[0132] Among them, α is the speed weight of the agent tracking the target, and so on, so as to determine the running speed of each agent in the tracking cluster, so that each agent can move to the place with high concentration gradient, that is, move to the target position, and achieve obstacle avoidance while tracking the target.

[0133] Through S510 to S540, the gene regulatory network is introduced to construct the concentration field. At the same time, the running speed of each intelligent agent is determined through the gradient method of the environment, so as to avoid collision with obstacles while quickly approaching the target.

[0134] Reference Figure 3 and Figure 4 In some embodiments of the present invention, the unmanned cluster target search control method further includes:

[0135] S101, dividing the group of robots into a number of search groups, each search group includes a number of intelligent agents, and each search group starts from a corresponding set initial position.

[0136] In this embodiment, in the initial stage, the group of robots is equally divided into several search groups. Each search group starts from its own set initial position and disperses to search in different directions to search for targets in an unknown environment. There are several intelligent agents in a single search group.

[0137] S102, the agents in each search group conduct decentralized exploration in the unknown environment according to the Gaussian random walk strategy to explore the target.

[0138] In this embodiment, each search group uses a Gaussian random walk strategy to search when exploring an unknown environment.

[0139] In order to maintain the formation, the step lengths extracted by agents in the same group follow the Gaussian distribution and remain consistent. i , the speed of the random walk at time t is:

[0140] V ig (t)=(∈ 1,t ,∈ 2,t )

[0141] Among them, V ig (t) represents the current agent R in the group i The velocity vector, ∈ 1,t and ∈ 2,t is derived from a Gaussian distribution N(μ,σ 2), μ is the mean of each move, which determines the distance from the initial position after repeated moves, and σ is the variance of each move, which determines the deviation between this move and the previous move. The larger σ is, the wider the search breadth is, and vice versa, the more refined the search is. By adjusting these two parameters, the speed and coverage of the agent's search for targets in unknown environments can be effectively improved, while enhancing its adaptability to diverse terrains.

[0142] For example, refer to Figure 3 and Figure 4 ,UAV clusters are used as group robots. The drones start from their respective initial positions and disperse in different directions to search for targets in unknown environments according to the Gaussian random walk strategy and track two targets.

[0143] Through S101 and S102, Gaussian random walk is introduced to enhance the agent's search ability in unknown environments, improve the search efficiency for targets, and enable the agent to quickly explore targets in complex and changing unknown environments.

[0144] S103, obtain the distance information between the current intelligent agent and other intelligent agents, and determine the third set of intelligent agents that have a gravitational relationship with the current intelligent agent and the fourth set of intelligent agents and obstacles that have a repulsive relationship with the current intelligent agent based on the distance information, and calculate the combined velocity of the potential field force of each intelligent agent in the several search groups based on the third set, the fourth set and the distance information.

[0145] In this embodiment, when the intelligent agents are performing the exploration task, the repulsive field is used to avoid collisions between different groups, between agents in a group, and between agents and obstacles. In order to simplify the calculation process, all obstacles are equivalent to agents in other groups, so only the interactions between agents in the same group and agents in different groups need to be considered. In order to maintain the formation, the intelligent agents in the same group need an additional gravitational field to achieve this purpose. i , count all individuals with which there is a gravitational field and a repulsive field, the expression is as follows:

[0146]

[0147] Among them, G' is all the agents related to the current agent R i There is a set of agents with a gravitational field between them, that is, the third set, t is the number of agents in the third set G', F' is the number of all agents with the current agent R i There is a set of agents and obstacles between which there is a repulsive field, that is, the fourth set, n is the number of agents in the fourth set F', and e is the number of obstacles in the fourth set F'.

[0148] Specifically, the current agent R iIf the distance to other agents or obstacles is less than the minimum safe distance d req , repulsion will occur; the current agent R i If the distance to other agents in the group is greater than the maximum distance d att , gravity will appear.

[0149] For the current agent R i , through the gravitational field calculation formula and the repulsive field calculation formula, according to the third set, the fourth set and the distance information, the gravitational force and the repulsive force are calculated, so as to calculate the combined speed of the potential field force of each intelligent agent in several search groups during the exploration phase.

[0150] In this embodiment, the calculation process of the combined velocity of the potential field force of each agent in the exploration phase is similar to the calculation process of the combined velocity of the potential field force of each agent in S420 to S440 of the tracking phase, and will not be described in detail in this embodiment. Among them, the second set in the tracking phase can be updated using the fourth set in the exploration phase to reduce the amount of calculation.

[0151] S104, according to the combined speed of the potential field force of each intelligent agent in the several search groups during the exploration phase, and in combination with the Gaussian random walk strategy, the total speed of each intelligent agent in the several search groups when exploring the unknown environment is obtained.

[0152] In this embodiment, for the current agent R i , the total speed formula is:

[0153] V i (t) = V ip (t)+V ig (t).

[0154] Through S101 to S104, when exploring the environment, the group robots are grouped, a Gaussian random walk strategy is introduced for walking, and an artificial potential field method is combined for obstacle avoidance to maintain the formation while avoiding obstacles, so as to improve the coverage and randomness of the search and improve the search efficiency.

[0155] In some embodiments of S100 of the present invention, when the swarm robot explores the target, allocating a set number of tracking agents from the swarm robot to form a tracking cluster specifically includes:

[0156] S110, when the target is discovered, for the current search group, the current search group communicates with the search group that discovered the target to determine the number of intelligent agents currently tracking the target.

[0157] In this embodiment, during the search process, after a search group has discovered a target, the search group that has discovered the target communicates with the current search group to determine the number M of intelligent agents currently involved in tracking the target.

[0158] That is, after the target is discovered, the agents in the search group that discovered the target communicate with the agents near the target to determine the number of agents M currently involved in tracking the target.

[0159] Among them, the current search group can be a group separated by a set distance from the search group that has explored the target, that is, the intelligent agents near the target can be intelligent agents separated by a set distance from the intelligent agents in the search group that has explored the target.

[0160] S120, determining whether the number of intelligent body tracking is less than the set tracking number.

[0161] In this embodiment, it is determined whether the agent tracking number M is less than the set tracking number N.

[0162] S130, if yes, determine the allocation number based on the number of intelligent agents to be tracked and the set number of tracking, separate the allocated number of intelligent agents from the current search group to track the target to form a tracking cluster, and the remaining intelligent agents in the current search group perform decentralized exploration according to the Gaussian random walk strategy.

[0163] In this embodiment, when the number of agents being tracked M is less than the set number of agents being tracked N, the allocation number (NM) is determined, and (NM) agents are separated from the current search group to the search group that has explored the target, forming a tracking cluster so that the agents in the tracking cluster meet the set number of agents being tracked, thereby ensuring that the target is fully tracked.

[0164] The remaining agents in the current search group will conduct decentralized exploration according to the Gaussian random walk strategy, that is, the agents that are not assigned tracking tasks will continue to explore decentralizedly.

[0165] If the number of agents in the current search group does not meet the allocated number (NM), all available agents in the current search group will be assigned to track the target, and continue to communicate with other search groups within the set communication range to supplement the agents until the number of agents participating in tracking the target reaches the set tracking number N.

[0166] Through S110 and S130, before tracking, by fixing the number of intelligent agents tracking each target, a tracking cluster is formed. Then, when the intelligent agent resources are sufficient, multiple targets can be tracked at the same time to ensure that each target is fully monitored.

[0167] The embodiment of the present application also provides an electronic device, the electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the above-mentioned unmanned cluster target search control method based on local information when executing the computer program. The electronic device can be any intelligent terminal including a tablet computer, a car computer, etc.

[0168] It can be understood that the contents of the above method embodiments are all applicable to the present device embodiments, the functions specifically implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0169] Electronic equipment includes:

[0170] The processor may be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;

[0171] The memory can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory can store an operating system and other applications. When the technical solution provided in the embodiments of this specification is implemented by software or firmware, the relevant program code is stored in the memory, and the processor is called to execute the unmanned cluster target search control method based on local information in the embodiments of this application;

[0172] Input / output interface, used to realize information input and output;

[0173] Communication interface, used to realize communication interaction between this device and other devices, which can be realized through wired mode (such as USB, network cable, etc.) or wireless mode (such as mobile network, WIFI, Bluetooth, etc.);

[0174] The bus, which transfers information between the various components of the device (such as the processor, memory, input / output interfaces, and communication interfaces);

[0175] The processor, memory, input / output interface and communication interface are connected to each other through a bus inside the device.

[0176] An embodiment of the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the unmanned cluster target search control method based on local information is implemented.

[0177] It can be understood that the contents of the above method embodiments are all applicable to the present storage medium embodiments, the functions specifically implemented by the present storage medium embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0178] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0179] Those skilled in the art will appreciate that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices may be implemented as software, firmware, hardware, or a suitable combination thereof.

[0180] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0181] The preferred embodiments of the present invention are described in detail above, but the invention is not limited to the embodiments. Those skilled in the art may make various equivalent modifications or substitutions without violating the spirit of the invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.

Claims

1. An unmanned cluster target search control method based on local information, characterized in that: The method comprises: When the group robot explores the target, a set number of tracking agents are allocated from the group robots to form a tracking cluster to track the target; A local coordinate system is established through the tracking cluster, and a first position information of the target, a second position information of the explored obstacle, and a third position information of each intelligent agent in the tracking cluster are determined according to the local coordinate system; Generate a comprehensive concentration field about the obstacle and the target through a GRN upper network model according to the first position information and the second position information; Determine, according to the third position information, a first set of agents having an attractive relationship with the current agent and a second set of agents having a repulsive relationship with the current agent, and calculate a resultant velocity of the potential field force of each agent in the tracking cluster according to the first set, the second set and the third position information; The gradient information of each point in the comprehensive concentration field is determined by differential calculation, the running speed of each intelligent agent in the tracking cluster is determined according to the gradient information and the combined speed, and each intelligent agent in the tracking cluster is controlled to track the target according to the running speed.

2. The unmanned cluster target search control method based on local information according to claim 1 is characterized in that: The establishing of the local coordinate system through the tracking cluster specifically includes: In the tracking cluster, three agents that are not on the same straight line are selected, and the three agents are respectively used as a first coordinate agent, a second coordinate agent, and a third coordinate agent; Taking the first coordinate intelligent body as the coordinate center o, the straight line from the first coordinate intelligent body to the second coordinate intelligent body is set as the x-axis, and the straight line perpendicular to the x-axis and pointing to one side of the third coordinate intelligent body is set as the y-axis to establish the local coordinate system xoy.

3. The unmanned cluster target search control method based on local information according to claim 2 is characterized in that: Determining the third position information of each agent in the tracking cluster according to the local coordinate system specifically includes: Obtain a first distance from the first coordinate intelligent body to the second coordinate intelligent body, a second distance from the remaining intelligent bodies to the first coordinate intelligent body, and a third distance from the remaining intelligent bodies to the second coordinate intelligent body; The third position information is determined based on the local coordinate system according to the first distance, the second distance and the third distance.

4. The unmanned cluster target search control method based on local information according to claim 1 is characterized in that: For the current agent R i , the calculating the combined velocity of the potential field force of each intelligent body in the tracking cluster according to the first set, the second set and the third position information specifically includes: According to the first set G and the third position information, the current agent R is calculated by the gravitational field calculation formula. i The gravitational force U att (d ik ), wherein the first set G is d ik is the current agent R i With G k , k = 1, 2, ..., h, h is the number of agents in the first set G; According to the second set F and the third position information, the current agent R is calculated by the repulsive field calculation formula. i The repulsive force U req (d ij ), wherein the second set F is d ij is the current agent R i With F j , j = 1, 2, ..., l, l is the number of agents in the second set F; According to the gravitational force U att (d ik ) and the repulsive force U req (d ij ), calculate the current agent R i The combined velocity V of the potential force ip (t), where i = 1, 2, ..., m, where m is the number of agents in the tracking cluster, to calculate the combined speed of each agent in the tracking cluster.

5. The unmanned cluster target search control method based on local information according to claim 1 is characterized in that: Determining the gradient information of each point in the comprehensive concentration field by differential calculation specifically includes: Performing central difference on non-boundary points in the comprehensive concentration field to determine gradient information of the non-boundary points; Forward difference and backward difference are performed on the boundary points in the comprehensive concentration field to determine the gradient information of the boundary points.

6. The unmanned cluster target search control method based on local information according to claim 1 is characterized in that: For the current agent R i , the step of determining the running speed of each agent in the tracking cluster according to the gradient information and the combined speed specifically includes: According to the current agent R i The third position information of the integrated concentration field is used to determine the gradient information of the corresponding point; According to the gradient information of the corresponding point and the current agent R i The corresponding combined speed is used to determine the running speed of the current agent, wherein i=1, 2, ..., m, where m is the number of agents in the tracking cluster, to calculate the running speed of each agent in the tracking cluster.

7. The unmanned cluster target search control method based on local information according to claim 1 is characterized in that: Also includes: Dividing the group of robots into a plurality of search groups, each of the search groups including a plurality of agents, and each of the search groups starting from a corresponding set initial position; Each agent in the search group performs decentralized exploration in an unknown environment according to a Gaussian random walk strategy to explore the target.

8. The unmanned cluster target search control method based on local information according to claim 7 is characterized in that: When the group robot explores the target, allocating a set number of tracking agents from the group robot to form a tracking cluster specifically includes: When the target is explored, for the current search group, the current search group communicates with the search group that has explored the target to determine the number of agents currently tracking the target; Determine whether the tracking number of the intelligent agent is less than the set tracking number; If so, the allocation number is determined based on the tracking number of the agents and the set tracking number, and the allocated number of agents are separated from the current search group to track the target to form the tracking cluster, and the remaining agents in the current search group perform decentralized exploration according to the Gaussian random walk strategy.

9. An electronic device, characterized in that: The electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the unmanned cluster target search control method based on local information as described in any one of claims 1 to 8 when executing the computer program.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the unmanned cluster target search control method based on local information described in any one of claims 1 to 8 is implemented.

Images