Heterogeneous swarm robot dynamic target tracking method and device, medium and product

By determining individual states and interactive neighbor sets in heterogeneous group robots, calculating speed parameters to control movement, the problems of poor information transmission and unoptimized resource utilization in dynamic target tracking of heterogeneous group robots are solved, and efficient collective response is achieved and decision-making efficiency.

CN118963348APending Publication Date: 2024-11-15FUDAN UNIVERSITY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411025502.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-30
Publication Date
2024-11-15

AI Technical Summary

Technical Problem

Heterogeneous group robots face problems such as poor information transmission, unoptimized resource utilization and low decision-making efficiency in dynamic target tracking, especially when there are different perception capabilities and large communication delays.

Method used

By obtaining the positions and velocities of individuals of heterogeneous group robots, determining individual states and interactive neighbor sets, calculating speed parameters to control individual movements, and achieving target tracking using repulsion, cohesion, alignment, self-drive and navigation speed parameters to promote rapid information transmission, and collective response.

Benefits of technology

It improves the response consistency and decision-making efficiency of heterogeneous group robots to external stimuli, optimizes resource utilization, and enhances the ability to respond quickly to environmental changes.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118963348A_ABST
    Figure CN118963348A_ABST
Patent Text Reader

Abstract

The invention discloses a heterogeneous swarm robot dynamic target tracking method and device, a medium and a product, and relates to the technical field of swarm robots, and the method comprises the steps: obtaining the positions and actual speeds of all robots in heterogeneous swarm robots, and the position and actual speed of a tracking target; determining an interactive neighbor set of the current individual based on the state of the current individual; calculating a speed parameter of the current individual based on the actual position and the actual speed of the current individual and the actual position and the actual speed of each neighbor individual in the interactive neighbor set of the current individual; determining a predicted speed of the current individual based on the state of the current individual and the speed parameter of the current individual; and determining the predicted position of the current individual based on the actual position of the current individual and the predicted speed of the current individual, thereby controlling the current individual to track the tracking target. Information in a robot group can be quickly transmitted, consistent collective response to external stimulation is ensured, and decision-making efficiency is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of swarm robots, and in particular to a method, device, medium and product for dynamic target tracking of heterogeneous swarm robots. Background Art

[0002] Swarm robot collaborative target tracking refers to a group of intelligent autonomous mobile robots that perceive the surrounding dynamic environment, interact and coordinate with neighboring robots within the perception range, make autonomous decisions and plan movements to adjust the relative position with the target object, thereby completing the tracking of a specific target. In the entire system, each robot is autonomously perceived, autonomously decided, and adopts a distributed control strategy. There is no centralized master control system.

[0003] Dynamic target tracking control of swarm robots can be applied to the escort of mobile targets. For example, a cluster of multiple ships on the sea can dynamically track an aircraft carrier in the middle and follow the battle, thus providing coordinated escort and protection for important targets in the middle. It can also be applied to the autonomous tracking of specific targets in a region. For example, through the collaborative interaction between multiple robots in the combat area, dynamic tracking of enemy mobile targets can be achieved.

[0004] Although swarm robots have made significant progress in collaborative tasks, heterogeneous swarm robots still face great challenges in dynamic target tracking. Traditional models usually assume that all agents have the same perception capabilities, which is an unrealistic expectation in complex practical applications. For example, in intelligent logistics systems, it is neither economical nor efficient to equip each transport vehicle with advanced remote perception sensors (such as lidar). Such an arrangement will lead to computational redundancy and unnecessary financial expenditure. In addition, for an area where a large number of intelligent mobile robots are deployed, if no target object enters or the target object is beyond the perception range of the swarm robots, these robots are information-free individuals; when a target enters, some robots closer to the target begin to perceive the target, become information individuals, and take actions to start tracking the target. Other robots remain stationary because they do not perceive the target, which leads to a decrease in the consistency of the swarm robot's motion. In addition, during the tracking process, the swarm robot also needs to perform obstacle avoidance control, first to avoid internal collisions caused by uncoordinated motion between individuals, and second to avoid obstacles in time when encountering obstacles to ensure the safety of the robots.

[0005] Some swarm intelligence algorithms have been applied to target tracking in actual multi-robot systems. (1) In the paper “Flocking for multi-agent dynamic systems: Algorithms and theory”, Olfati-Saber et al. introduced a method that integrates feedback items to monitor virtual navigators, effectively reducing the dispersion of group motion; (2) In the paper “Multi-robot cooperative target encirclement through learning distributed transferable policy”, Zhang et al. introduced a distributed policy network framework based on deep reinforcement learning, using graph communication mechanism for robot interaction, aiming to solve the problem of single target encirclement and collision avoidance. (3) The invention patent with application number 201911310097.5 involves a stability analysis method for a time-varying number of swarm robot relay target tracking system. The method includes: using differential equations to establish a tracking control model of a group robot relay target tracking system, and establishing a dynamic equation of the target: based on the tracking control model and the dynamic equation of the target, establishing an overall tracking error equation of a time-varying number of group robots relay target tracking system: using topological theory, Lyapunov energy function method and switching theory to establish a judgment condition for the asymptotic stability of the overall tracking error equation, and analyzing the stability of the group robot relay target tracking system according to the judgment condition. The implementation of the invention can perform stability analysis and controller design on the error jump of the relay system caused by the time-varying number of tracking robots, and can be applied to different scenarios. (4) The invention patent with application number 201811446461.6 relates to a distributed multi-target tracking method for swarm robots based on PHD filtering, including S1: in the initial state, the swarm robots are randomly distributed within a given constraint boundary, and a Power graph is constructed and the search area is divided according to the position coordinates and weights of the multiple robots; S2: the weighted center of mass of the Power unit corresponding to each robot is solved; S3: all robots start to move from the current position X to the corresponding h_i(0), collect relevant measurement information, calculate the normalization constant, and update the Power graph; S4: according to the measurement data, the PHD is updated and the new PHD is used as the weight, and S2 is executed to obtain the new center of mass position coordinates; S5: S3 and S4 are repeated until the target appears in the robot's field of view; S6: after the robot finds the target, it observes the motion state of the target, uses the functional relationship between the current radar measurement information and the acceleration, estimates the acceleration disturbance in real time, and performs adaptive adjustment of the acceleration variance to maintain synchronous tracking of the target.

[0006] However, these methods often face limitations in practical applications due to communication delays and limited onboard computing capacity, and their real-time performance is not strong. Summary of the invention

[0007] The purpose of this application is to provide a method, device, medium and product for dynamic target tracking of heterogeneous swarm robots, so that information within the robot group can be quickly transmitted, ensuring a consistent collective response to external stimuli, thereby optimizing resource utilization and improving decision-making efficiency.

[0008] To achieve the above objectives, this application provides the following solutions:

[0009] In a first aspect, the present application provides a method for dynamic target tracking of heterogeneous swarm robots, comprising:

[0010] At the current moment, the positions and actual speeds of all robots in the heterogeneous swarm robot and the position and actual speed of the tracking target are obtained; the heterogeneous swarm robot includes a plurality of individuals, one individual corresponding to one robot;

[0011] At the current moment, any individual in the heterogeneous group of robots is determined as the current individual;

[0012] Determine the state of the current individual at the current moment; the state includes: an environment activation state and a cluster activation state; the environment activation state is 1 or 0, and the cluster activation state is 1 or 0;

[0013] Based on the state of the current individual at the current moment, determine the interactive neighbor set of the current individual at the current moment; the interactive neighbor set includes: an exclusion neighbor set, a cohesive neighbor set and an alignment neighbor set, and each interactive neighbor set includes one or more neighbor individuals;

[0014] Based on the actual position and actual speed of the current individual at the current moment and the actual position and actual speed of each neighbor individual in the interactive neighbor set of the current individual at the current moment, the speed parameters of the current individual at the next moment are calculated; the speed parameters include: repulsion speed, cohesion speed, alignment speed, self-driving speed, navigation speed and attraction speed;

[0015] Based on the state of the current individual at the current moment and the speed parameter of the current individual at the next moment, determine the predicted speed of the current individual at the next moment;

[0016] Determine the predicted position of the current individual at the next moment based on the actual position of the current individual at the current moment and the predicted speed of the current individual at the next moment;

[0017] At the next moment, based on the predicted speed and predicted position of the current individual at the next moment, the current individual is controlled to track the tracking target.

[0018] Optionally, determining the state of the current individual at the current moment includes:

[0019] Determine whether the current individual satisfies a first judgment condition at the current moment, and obtain a first judgment result; the first judgment condition is that in the heterogeneous group of robots, only the current individual encounters an obstacle or detects the tracking target;

[0020] If the first determination result is yes, determining the environment activation state of the current individual at the current moment to be 1;

[0021] If the first determination result is no, determining the environmental activation state of the current individual at the current moment to be 0;

[0022] Determine whether the current individual has a high entropy neighbor individual at the current moment, and obtain a second determination result; the high entropy neighbor individual of the current individual is an individual whose information entropy among the neighbor individuals of the current individual is greater than a preset entropy value;

[0023] If the second determination result is no, determining the cluster activation state of the current individual at the current moment to be 0;

[0024] If the second determination result is yes, then determine whether the difference between the actual speed of the current individual at the current moment and the actual speed of the high entropy neighbor individual at the current moment is greater than a preset speed difference, and obtain a third determination result;

[0025] If the third determination result is yes, the cluster activation state of the current individual at the current moment is determined to be 1;

[0026] If the third determination result is no, the cluster activation state of the current individual at the current moment is determined to be 0.

[0027] Optionally, based on the state of the current individual at the current moment, determining the interactive neighbor set of the current individual at the current moment includes:

[0028] At the current moment, all connected neighbor individuals within the perception range of the current individual are determined as the first neighbor set of the current individual at the current moment;

[0029] Determine whether each neighbor individual in the first neighbor set of the current individual at the current moment meets a screening condition; the screening condition is: the neighbor individual is a high entropy neighbor individual and the difference between the actual speed of the current individual at the current moment and the actual speed of the high entropy neighbor individual at the current moment is greater than a preset speed difference;

[0030] If yes, then based on each neighbor individual that meets the screening condition, determine the second neighbor set of the current individual at the current moment;

[0031] The interactive neighbor set of the current individual at the current moment is determined according to the state of the current individual at the current moment, the first neighbor set and the second neighbor set.

[0032] Optionally, based on the actual position and actual speed of the current individual at the current moment and the actual position and actual speed of each neighbor individual in the interactive neighbor set of the current individual at the current moment, the speed parameter of the current individual at the next moment is calculated, including:

[0033] The speed parameter calculation formula is used to calculate the speed parameter of the current individual at the next moment according to the actual position and actual speed of the current individual at the current moment and the actual position and actual speed of each neighbor individual in the interactive neighbor set of the current individual at the current moment.

[0034] Optionally, the speed parameter calculation formula includes: a repulsion speed calculation formula, a cohesion speed calculation formula, an alignment speed calculation formula, a self-driving speed calculation formula, a navigation speed calculation formula and an attraction speed calculation formula;

[0035] The repulsion velocity calculation formula includes:

[0036]

[0037]

[0038] in, is the size of the exclusion term between the ith individual and the j1th neighbor individual in the exclusion neighbor set at the tth moment; C rep is the rejection term coefficient; is the maximum exclusion range; is the distance between the ith individual and the j1th neighbor individual in the exclusionary neighbor set at the tth moment; p i (t) is the actual position of the i-th individual at time t; is the actual position of the j1th neighbor individual in the excluded neighbor set at time t; is the repulsion velocity of the i-th individual at the t+1th moment; is the set of excluded neighbors of the i-th individual at time t;

[0039] The formula for calculating the condensation speed includes:

[0040]

[0041]

[0042] in, is the size of the cohesion term between the ith individual and the j2th neighbor individual in the cohesive neighbor set at the tth moment; C cohis the cohesion term coefficient; is the distance between the ith individual and the j2th neighbor individual in the cohesive neighbor set at the tth moment; is the minimum cohesion range; is the actual position of the j2th neighbor individual in the condensed neighbor set at time t; r i is the perception radius of the i-th individual; is the aggregation speed of the i-th individual at the t+1th moment; is the cohesive neighbor set of the i-th individual at time t;

[0043] The alignment speed calculation formula includes:

[0044]

[0045] in, is the alignment speed of the i-th individual at the t+1th moment; is the actual speed of the j3rd neighbor individual in the aligned neighbor set at time t; v i (t) is the actual speed of the i-th individual at the t-th moment; is the set of aligned neighbors of the i-th individual at time t; is the distance between the i-th individual and the j3-th neighbor individual in the aligned neighbor set at the t-th moment;

[0046] The self-driving speed calculation formula includes:

[0047]

[0048] in, is the self-driving speed of the i-th individual at the t+1th moment; C pro is the self-driving term coefficient; |·| is the Euclidean norm; v tar (t) is the actual speed of the tracking target at time t;

[0049] The navigation speed calculation formula includes:

[0050]

[0051] in, is the navigation speed of the ith individual at the t+1th moment; C nav is the navigation item coefficient; p tar (t) is the actual position of the tracking target at time t;

[0052] The suction speed calculation formula includes:

[0053]

[0054] in, is the attraction speed of the i-th individual at the t+1th moment; C tar is the attraction coefficient; d itar (t) is the distance between the i-th individual and the tracking target at the t-th moment.

[0055] Optionally, based on the state of the current individual at the current moment and the speed parameter of the current individual at the next moment, determining the predicted speed of the current individual at the next moment includes:

[0056] The speed prediction formula is used to determine the predicted speed of the current individual at the next moment according to the state of the current individual at the current moment and the speed parameter of the current individual at the next moment; the speed prediction formula includes:

[0057]

[0058] Among them, v′ i (t+1) is the predicted speed of the i-th individual at the t+1th time; is the alignment coefficient of individuals whose cluster activation state is 0; is the cluster activation state of the i-th individual at the t-th moment; is the environmental activation state of the i-th individual at time t; is the alignment coefficient for individuals whose cluster activation state is 1.

[0059] Optionally, based on the actual position of the current individual at the current moment and the predicted speed of the current individual at the next moment, determining the predicted position of the current individual at the next moment includes:

[0060] The predicted position of the current individual at the next moment is determined by using a position prediction formula according to the actual position of the current individual at the current moment and the predicted speed of the current individual at the next moment; the position prediction formula includes:

[0061] p′ i (t+1)=p i (t)+v′ i (t+1)×Δt;

[0062] Among them, p′ i (t+1) is the predicted position of the i-th individual at the t+1th time; Δt is the unit time step.

[0063] In a second aspect, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any of the heterogeneous swarm robot dynamic target tracking methods described above.

[0064] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the heterogeneous swarm robot dynamic target tracking methods described above.

[0065] In a fourth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements any of the heterogeneous swarm robot dynamic target tracking methods described above.

[0066] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0067] The present application discloses a method, device, medium and product for dynamic target tracking of heterogeneous swarm robots. First, at the current moment, the positions and actual speeds of all robots in the heterogeneous swarm robots and the position and actual speed of the tracking target are obtained; at the current moment, any individual in the heterogeneous swarm robots is determined as the current individual; the state of the current individual at the current moment is determined; secondly, based on the state of the current individual at the current moment, the interactive neighbor set of the current individual at the current moment is determined; the interactive neighbor set includes: an exclusive neighbor set, a cohesive neighbor set and an aligned neighbor set, and each interactive neighbor set includes one or more neighbor individuals; based on the actual position and actual speed of the current individual at the current moment and The actual position and actual speed of each neighbor individual in the interactive neighbor set of the current individual at the current moment are used to calculate the speed parameter of the current individual at the next moment; the speed parameters include: repulsion speed, cohesion speed, alignment speed, self-driving speed, navigation speed and attraction speed; again, based on the state of the current individual at the current moment and the speed parameter of the current individual at the next moment, the predicted speed of the current individual at the next moment is determined; based on the actual position of the current individual at the current moment and the predicted speed of the current individual at the next moment, the predicted position of the current individual at the next moment is determined; finally, at the next moment, based on the predicted speed and predicted position of the current individual at the next moment, the current individual is controlled to track the tracking target. This application can promote efficient information transmission between robots through each interactive neighbor set. Robots that perceive environmental changes will quickly spread information hierarchically within the group. Even if there is no robot that directly perceives the tracking target, it will keep speed alignment with the neighbor individuals that perceive the robot changes, which enables the group to make a rapid collective response to environmental changes and maximize the use of the characteristics of different perception ranges of heterogeneous groups, optimize resource utilization and improve decision-making efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0069] Figure 1 A schematic diagram of a flow chart of a dynamic target tracking method for heterogeneous swarm robots provided in one embodiment of the present application;

[0070] Figure 2 It is the perception radius diagram of ordinary individuals;

[0071] Figure 3 It is the perception radius diagram of sensitive individuals;

[0072] Figure 4 is a random distribution graph of heterogeneous group robots;

[0073] Figure 5 When there is no obstacle, the navigation speed direction points to the tracking target schematic diagram;

[0074] Figure 6 When approaching and sensing the existence of an obstacle, the navigation direction points to the direction of the largest obstacle clearance;

[0075] Figure 7 This is the PlusBot cluster control system structure diagram;

[0076] Figure 8 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0077] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0078] The purpose of this application is to provide a method, device, medium and product for dynamic target tracking of heterogeneous swarm robots, aiming to enable rapid transmission of information within the robot group and ensure a consistent collective response to external stimuli, thereby optimizing resource utilization and improving decision-making efficiency.

[0079] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0080] In an exemplary embodiment, Figure 1 As shown, the heterogeneous group robot dynamic target tracking method in this embodiment includes:

[0081] Step 1: At the current moment, obtain the position and actual speed of all robots in the heterogeneous swarm robot and the position and actual speed of the tracking target.

[0082] Among them, the heterogeneous group robot includes multiple individuals, and one individual corresponds to one robot.

[0083] Specifically, individuals in a heterogeneous swarm of robots can be divided into two categories according to their sensing range. Figure 2 The individuals with limited perception radius (perception radius r1) are regarded as ordinary individuals, while Figure 3 The individuals with a larger perception radius (perception radius is r1, r2>r1) are defined as sensitive individuals. Figure 4 As shown, these two types of individuals are randomly distributed in the initial area, and their numbers are defined as n1 and n2 respectively. The total number of individuals in the heterogeneous swarm robot is n=n1+n2.

[0084] Step 2: At the current moment, any individual in the heterogeneous swarm robot is determined as the current individual.

[0085] Step 3: Determine the state of the current individual at the current moment.

[0086] The status includes: environment activation status and cluster activation status; the environment activation status is 1 or 0, and the cluster activation status is 1 or 0.

[0087] As an optional implementation, step 3 includes:

[0088] Step 31: Determine whether the current individual satisfies the first judgment condition at the current moment, and obtain a first judgment result; the first judgment condition is that in the heterogeneous group robot, only the current individual encounters an obstacle or detects a tracking target.

[0089] Step 32: If the first determination result is yes, the environment activation state of the current individual at the current moment is determined to be 1.

[0090] Step 33: If the first determination result is no, the environment activation state of the current individual at the current moment is determined to be 0.

[0091] Step 34: Determine whether the current individual has a high entropy neighbor individual at the current moment, and obtain a second determination result; the high entropy neighbor individual of the current individual is an individual whose information entropy among the neighbor individuals of the current individual is greater than a preset entropy value.

[0092] Step 35: If the second determination result is no, the cluster activation state of the current individual at the current moment is determined to be 0.

[0093] Step 36: If the second judgment result is yes, determine whether the difference between the actual speed of the current individual at the current moment and the actual speed of the high entropy neighbor individual at the current moment is greater than the preset speed difference, and obtain a third judgment result.

[0094] Step 37: If the third determination result is yes, the cluster activation state of the current individual at the current moment is determined to be 1.

[0095] Step 38: If the third determination result is no, the cluster activation state of the current individual at the current moment is determined to be 0.

[0096] Step 4: Based on the state of the current individual at the current moment, determine the set of interactive neighbors of the current individual at the current moment.

[0097] The interactive neighbor sets include: an exclusive neighbor set, a cohesive neighbor set, and an aligned neighbor set, and each interactive neighbor set includes one or more neighbor individuals.

[0098] As an optional implementation, step 4 includes:

[0099] Step 41: At the current moment, based on all connected neighbor individuals within the perception range of the current individual, determine the first neighbor set of the current individual at the current moment.

[0100] Step 42: Determine whether each neighbor individual in the first neighbor set of the current individual at the current moment meets the screening condition; the screening condition is: the neighbor individual is a high entropy neighbor individual and the difference between the actual speed of the current individual at the current moment and the actual speed of the high entropy neighbor individual at the current moment is greater than a preset speed difference.

[0101] Step 43: If yes, then based on each neighbor individual that meets the screening condition, determine the second neighbor set of the current individual at the current moment.

[0102] Step 44: Determine the interactive neighbor set of the current individual at the current moment according to the state of the current individual at the current moment, the first neighbor set and the second neighbor set.

[0103] Specifically, the unique feature of heterogeneous swarm robots lies in the difference in the perception radius of individuals. Sensitive individuals have a larger perception radius and can detect a wider range of environmental changes. That is, sensitive individuals are more likely to perceive environmental stimuli during movement. Individuals that detect new environmental changes (such as the presence of obstacles or tracking targets) are classified as informed individuals. In contrast, individuals that do not detect such changes are called uninformed individuals. Inspired by the attention mechanism of fish schools, a small number of informed individuals can be fully utilized and the detected environmental information can be propagated to the entire group. By combining an adjustable self-driven swarm model with an attention mechanism, the group's responsiveness to early sensory stimuli can be enhanced. Heterogeneous swarm robots will exhibit stronger self-organizing characteristics. For group movement, individuals can respond to external stimuli more quickly and accurately while maintaining coordination.

[0104] In order to introduce the attention mechanism into cluster control, two activation states are defined for individuals: the environment activation state and the cluster activation state. The values ​​of these two states are not constant and change over time. If an individual encounters an obstacle or detects a tracking target, it will become an information individual and its environment activation state is 1. The information individual will not align with any neighbor individuals, and it will respond to environmental stimuli immediately.

[0105] Therefore, the information individual is the first individual to change its speed due to detecting an obstacle or tracking a target. Individuals with higher information entropy values ​​are defined as disordered individuals in an ordered group. These individuals make sudden changes in their motion state in response to environmental stimuli. These sudden changes in motion attract the attention of other members in the group. Although other individuals may not detect environmental changes and cannot react directly, they observe those individuals with higher entropy and align their speed with them. In this way, they indirectly achieve a fast response to environmental changes. When an individual has a high entropy neighbor individual, its cluster activation state is set to 1; if not, the individual remains in a normal state and the cluster activation state is set to 0.

[0106] However, in order to prevent individuals from frequently changing their cluster activation states, a threshold mechanism is further designed. Specifically, when at time t, the speed difference between the i-th individual and its high entropy neighbor exceeds the set threshold, that is, the preset speed difference θ max When , the i-th individual will change its cluster activation state from 0 to 1, and the high entropy neighbor individual belongs to the second neighbor set of the i-th individual at time t If the i-th individual has been activated, that is, the cluster activation state is 1, it will be The individuals in keep their speeds aligned and no longer respond to new stimuli. When the speed difference between the i-th individual and all its interacting neighbors is less than the threshold θ maxWhen , the cluster activation state will change from the activated state (1) to the normal state (0). This threshold mechanism is mathematically expressed as:

[0107]

[0108] in, is the cluster activation state of the i-th individual at the t-th moment; For existence; N i (t) is the first neighbor set of the i-th individual at time t; v j (t) is the actual speed of the jth neighbor individual in the first neighbor set of the i-th individual at time t; is the cluster activation state of the i-th individual at the t-1th moment; is any; otherwise is other.

[0109] The sets of interactive neighbors with different effects are summarized as follows.

[0110]

[0111] in, is the set of excluded neighbors of the i-th individual at time t; is the cohesive neighbor set of the i-th individual at time t; is the set of aligned neighbors of the i-th individual at time t; is the environmental activation state of the i-th individual at time t; is the second neighbor set of the i-th individual at time t; Is an empty set.

[0112] In summary, information individuals do not align and cohere with neighbor individuals. They tend to respond to external environmental stimuli rather than interactions within the group. Individuals with cluster activation status 1 and individuals with cluster activation status 0 interact with different neighbor individuals in terms of cohesion and alignment effects.

[0113] Step 5: Calculate the speed parameter of the current individual at the next moment based on the actual position and actual speed of the current individual at the current moment and the actual position and actual speed of each neighbor individual in the interactive neighbor set of the current individual at the current moment.

[0114] Among them, the speed parameters include: repulsion speed, cohesion speed, alignment speed, self-driving speed, navigation speed and attraction speed.

[0115] As an optional implementation, step 5 includes:

[0116] The speed parameter calculation formula is used to calculate the speed parameter of the current individual at the next moment according to the actual position and actual speed of the current individual at the current moment and the actual position and actual speed of each neighbor individual in the interactive neighbor set of the current individual at the current moment.

[0117] As an optional implementation, the speed parameter calculation formula includes: a repulsion speed calculation formula, a cohesion speed calculation formula, an alignment speed calculation formula, a self-driving speed calculation formula, a navigation speed calculation formula and an attraction speed calculation formula.

[0118] In order to maintain a safe distance between two neighboring individuals, a half-spring model is used, which calculates a distance-dependent central velocity term. For the repulsion effect, a maximum range is set As a cutoff point, within this range individuals begin to repel each other. In order to control the relative strength of the repulsive control, the repulsive term coefficient C is added rep As a linear gain. If the distance between two individuals is less than Then the size of the repulsion term between the two individuals is equal to 0. The total repulsion effect (i.e., repulsion speed) of the i-th individual is obtained by accumulating the neighbor individuals belonging to the repulsion neighbor set, which is expressed as the repulsion speed calculation formula. The repulsion speed calculation formula includes:

[0119]

[0120]

[0121] in, is the size of the exclusion term between the ith individual and the j1th neighbor individual in the exclusion neighbor set at the tth moment; C rep is the rejection term coefficient; is the maximum exclusion range; is the distance between the ith individual and the j1th neighbor individual in the exclusionary neighbor set at the tth moment; p i (t) is the actual position of the i-th individual at time t; is the actual position of the j1th neighbor individual in the excluded neighbor set at time t; is the repulsion velocity of the i-th individual at the t+1th moment; is the set of excluded neighbors of the i-th individual at time t.

[0122] The formula for cohesion control is defined as follows: Set the cohesion term coefficient C coh Characterizes the linear gain of the pairwise cohesive term, is the minimum condensation range, exceeding The individuals begin to cohere with each other. The total cohesion term (i.e., cohesion speed) of the i-th individual is calculated by summing the neighbor individuals belonging to the cohesive neighbor set, which is expressed as the cohesion speed calculation formula. The cohesion speed calculation formula includes:

[0123]

[0124]

[0125] in, is the size of the cohesion term between the ith individual and the j2th neighbor individual in the cohesive neighbor set at the tth moment; C coh is the cohesion term coefficient; is the distance between the ith individual and the j2th neighbor individual in the cohesive neighbor set at the tth moment; is the minimum cohesion range; is the actual position of the j2th neighbor individual in the condensed neighbor set at time t; r i is the perception radius of the i-th individual; is the aggregation speed of the i-th individual at the t+1th moment; is the cohesive neighbor set of the i-th individual at time t.

[0126] In addition to preventing collisions between individuals through distance control, the group must also achieve collective movement through alignment control, which is expressed as the alignment speed calculation formula. The alignment speed calculation formula includes:

[0127]

[0128] in, is the alignment speed of the i-th individual at the t+1th moment; is the actual speed of the j3rd neighbor individual in the aligned neighbor set at time t; v i (t) is the actual speed of the i-th individual at the t-th moment; is the set of aligned neighbors of the i-th individual at time t; is the distance between the i-th individual and the j3-th neighbor individual in the aligned neighbor set at the t-th moment.

[0129] The self-driving speed is designed to keep the individual moving at a certain speed. The self-driving speed calculation formula includes:

[0130]

[0131] in, is the self-driving speed of the i-th individual at the t+1th moment; C pro is the self-driven term coefficient, representing the linear gain of the self-driven term; |·| is the Euclidean norm; vtar (t) is the actual speed of the tracking target at the tth moment.

[0132] The self-driving swarm model with adjustable parameters uses repulsion velocity to slow down the speed of individuals in the swarm when approaching obstacles in the environment. However, when tracking dynamic targets and avoiding obstacles at the same time, the swarm can easily fall into a local minimum and oscillate in place. Therefore, the navigation velocity is used instead of the repulsion velocity to avoid oscillation.

[0133] Inspired by the vision-guided flight of pigeons around obstacles, swarm robots make steering decisions based on the relative information between obstacles and tracking targets. Figure 5 As shown in the figure, when there is no obstacle within the individual's perception range, the navigation speed direction points to the tracking target. The navigation speed calculation formula includes:

[0134]

[0135] in, is the navigation speed of the ith individual at the t+1th moment; C nav is the navigation item coefficient, which represents the linear gain of the navigation item; tar (t) is the actual position of the tracking target at time t.

[0136] In fact, if Figure 6 As shown in Figure 1, when the i-th individual approaches and senses the presence of an obstacle, it tends to navigate to the direction of the largest obstacle clearance to enhance the safety of the entire group. This steering decision can be obtained based on the local information of the robot. obs The angle θ of the obstacle k in ik is: ik =arctan(y k -y i ) / (x k -x i ), (x k ,y k ) is the center position of obstacle k; (x i ,y i ) is the center position of the ith individual.

[0137] The field of view where half of the obstacle k is blocked by the i-th individual is defined as φ ik Therefore, the prohibited angular range D of the i-th individual to the obstacle k is ik It can be expressed as D ik =[θ ik -φ ik ,θ ik +φ ik ]. Similarly, the detection field of the ith individual is defined as E iIn addition, the available gap A for the i-th individual to cross multiple obstacles i It can be calculated as Among them, A i is the set of available moving gaps; ∪ is the union; ∩ is the intersection; A i The mth available moving gap in N gap A i The total number of available moving gaps in . The occupied angle of the mth available moving gap is defined as α m , index ind1 is the largest visual gap, ind1 = argmax(α m ). The selected visual gap Determined by the steering decision, its boundary line is defined as β 1,i and β 2,i Therefore, the calculation formula of navigation speed can be as follows:

[0138]

[0139] in, is the navigation speed of the ith individual at the t+1th moment; C nav is the attraction coefficient; is the angle between the first boundary line of the visual gap of the ith individual at the tth moment and the positive direction of the x-axis; is the angle between the second boundary line of the visual gap of the ith individual at the tth moment and the positive direction of the x-axis.

[0140] When there are no obstacles within the individual's perception range, it moves toward the tracking target. However, when an obstacle is detected, it prioritizes safety and navigates through the largest gap between obstacles.

[0141] In order to ensure that all individuals can track moving targets and complete goal-oriented movements in a limited space, the attraction term is defined as follows. This effect only occurs when the tracking target is within the individual's perception range. The attraction speed calculation formula includes:

[0142]

[0143] in, is the attraction speed of the i-th individual at the t+1th moment; C tar is the attraction term coefficient, which represents the linear gain of the target attraction term; d itar (t) is the distance between the i-th individual and the tracking target at the t-th moment.

[0144] In fact, the repulsion speed, cohesion speed, alignment speed and self-driving speed are all calculated by considering the behavior within the cluster, and the navigation speed and attraction speed are all calculated by considering the behavior between the cluster and the environment.

[0145] Step 6: Based on the state of the current individual at the current moment and the speed parameter of the current individual at the next moment, determine the predicted speed of the current individual at the next moment.

[0146] As an optional implementation, step 6 includes:

[0147] The speed prediction formula is used to determine the predicted speed of the current individual at the next moment according to the state of the current individual at the current moment and the speed parameter of the current individual at the next moment; the speed prediction formula includes:

[0148]

[0149] Among them, v′ i (t+1) is the predicted speed of the i-th individual at the t+1th time; is the alignment coefficient of individuals whose cluster activation state is 0; is the cluster activation state of the i-th individual at the t-th moment; is the environmental activation state of the i-th individual at time t; is the alignment coefficient for individuals whose cluster activation state is 1.

[0150] Step 7: Based on the actual position of the current individual at the current moment and the predicted speed of the current individual at the next moment, determine the predicted position of the current individual at the next moment.

[0151] As an optional implementation, step 7 includes:

[0152] The position prediction formula is used to determine the predicted position of the current individual at the next moment according to the actual position of the current individual at the current moment and the predicted speed of the current individual at the next moment; the position prediction formula includes:

[0153] p′ i (t+1)=p i (t)+v′ i (t+1)×Δt.

[0154] Among them, p′ i (t+1) is the predicted position of the i-th individual at the t+1th time; Δt is the unit time step.

[0155] Step 8: At the next moment, based on the predicted speed and predicted position of the current individual at the next moment, control the current individual to track the tracking target. Specific embodiments

[0157] In such Figure 7The dynamic target tracking method of heterogeneous group robots of this application is implemented on the PlusBot cluster control system shown. Each robot is equipped with a WiFi module to enable it to communicate with the host computer. Control commands can be sent to the robot from the host computer, and it can also display various motion performance indicators returned by the robot. All cluster experiments were carried out in an indoor environment, and an ultra-wide-angle camera was installed on the ceiling of the room to record the experiment at 1080P and 60fps. In other words, this camera provides a global perspective to control the collective behavior of the entire robot. In order to implement the proposed dynamic target tracking method of heterogeneous group robots, it is necessary to ensure that each robot can only perceive its surroundings (neighboring individuals and obstacles) within a given perception radius. Although the global perspective can provide overall information, the control command of each robot only uses the local information it should have. These functions are implemented through a distributed parallel multi-threaded system. More specifically, each robot is assigned an independent control thread, all of which run on the host computer and communicate with the robot via WiFi. In this way, each robot has two controllers. One runs the proposed algorithm in the workstation, and the other runs in the onboard system and tracks the control commands sent from the workstation.

[0158] In an exemplary embodiment, a computer device is provided, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement a dynamic target tracking method for heterogeneous swarm robots.

[0159] In an exemplary embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, a dynamic target tracking method for a heterogeneous swarm robot is implemented.

[0160] In an exemplary embodiment, a computer program product is provided, including a computer program, which, when executed by a processor, implements a dynamic target tracking method for heterogeneous swarm robots.

[0161] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 8As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a dynamic target tracking method for a heterogeneous group robot is implemented.

[0162] Those skilled in the art will understand that Figure 8 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0163] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0164] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0165] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0166] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0167] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A dynamic target tracking method for heterogeneous swarm robots, characterized in that: The heterogeneous group robot dynamic target tracking method comprises: At the current moment, the positions and actual speeds of all robots in the heterogeneous swarm robot and the position and actual speed of the tracking target are obtained; the heterogeneous swarm robot includes a plurality of individuals, one individual corresponding to one robot; At the current moment, any individual in the heterogeneous group of robots is determined as the current individual; Determine the state of the current individual at the current moment; the state includes: an environment activation state and a cluster activation state; the environment activation state is 1 or 0, and the cluster activation state is 1 or 0; Based on the state of the current individual at the current moment, determine the interactive neighbor set of the current individual at the current moment; the interactive neighbor set includes: an exclusion neighbor set, a cohesive neighbor set and an alignment neighbor set, and each interactive neighbor set includes one or more neighbor individuals; Based on the actual position and actual speed of the current individual at the current moment and the actual position and actual speed of each neighbor individual in the interactive neighbor set of the current individual at the current moment, the speed parameters of the current individual at the next moment are calculated; the speed parameters include: repulsion speed, cohesion speed, alignment speed, self-driving speed, navigation speed and attraction speed; Based on the state of the current individual at the current moment and the speed parameter of the current individual at the next moment, determine the predicted speed of the current individual at the next moment; Determine the predicted position of the current individual at the next moment based on the actual position of the current individual at the current moment and the predicted speed of the current individual at the next moment; At the next moment, based on the predicted speed and predicted position of the current individual at the next moment, the current individual is controlled to track the tracking target.

2. The method for dynamic target tracking of heterogeneous swarm robots according to claim 1, characterized in that: Determine the current state of the individual at the current moment, including: Determine whether the current individual satisfies a first judgment condition at the current moment, and obtain a first judgment result; the first judgment condition is that in the heterogeneous group of robots, only the current individual encounters an obstacle or detects the tracking target; If the first determination result is yes, determining the environment activation state of the current individual at the current moment to be 1; If the first determination result is no, determining the environmental activation state of the current individual at the current moment to be 0; Determine whether the current individual has a high entropy neighbor individual at the current moment, and obtain a second determination result; the high entropy neighbor individual of the current individual is an individual whose information entropy among the neighbor individuals of the current individual is greater than a preset entropy value; If the second determination result is no, determining the cluster activation state of the current individual at the current moment to be 0; If the second determination result is yes, then determine whether the difference between the actual speed of the current individual at the current moment and the actual speed of the high entropy neighbor individual at the current moment is greater than a preset speed difference, and obtain a third determination result; If the third determination result is yes, the cluster activation state of the current individual at the current moment is determined to be 1; If the third determination result is no, the cluster activation state of the current individual at the current moment is determined to be 0.

3. The method for dynamic target tracking of heterogeneous swarm robots according to claim 2, characterized in that: Based on the state of the current individual at the current moment, determine the interactive neighbor set of the current individual at the current moment, including: At the current moment, all connected neighbor individuals within the perception range of the current individual are determined as the first neighbor set of the current individual at the current moment; Determine whether each neighbor individual in the first neighbor set of the current individual at the current moment meets a screening condition; the screening condition is: the neighbor individual is a high entropy neighbor individual and the difference between the actual speed of the current individual at the current moment and the actual speed of the high entropy neighbor individual at the current moment is greater than a preset speed difference; If yes, then based on each neighbor individual that meets the screening condition, determine the second neighbor set of the current individual at the current moment; The interactive neighbor set of the current individual at the current moment is determined according to the state of the current individual at the current moment, the first neighbor set and the second neighbor set.

4. The method for dynamic target tracking of heterogeneous swarm robots according to claim 1, characterized in that: Based on the actual position and actual speed of the current individual at the current moment and the actual position and actual speed of each neighbor individual in the interactive neighbor set of the current individual at the current moment, the speed parameter of the current individual at the next moment is calculated, including: The speed parameter calculation formula is used to calculate the speed parameter of the current individual at the next moment according to the actual position and actual speed of the current individual at the current moment and the actual position and actual speed of each neighbor individual in the interactive neighbor set of the current individual at the current moment.

5. The method for dynamic target tracking of heterogeneous swarm robots according to claim 4, characterized in that: The speed parameter calculation formula includes: a repulsion speed calculation formula, a cohesion speed calculation formula, an alignment speed calculation formula, a self-driving speed calculation formula, a navigation speed calculation formula and an attraction speed calculation formula; The repulsion velocity calculation formula includes: in, is the size of the exclusion term between the ith individual and the j1th neighbor individual in the exclusion neighbor set at the tth moment; C rep is the rejection term coefficient; is the maximum exclusion range; is the distance between the ith individual and the j1th neighbor individual in the exclusionary neighbor set at the tth moment; p i (t) is the actual position of the i-th individual at time t; is the actual position of the j1th neighbor individual in the excluded neighbor set at time t; is the repulsion velocity of the i-th individual at the t+1th moment; is the set of excluded neighbors of the i-th individual at time t; The formula for calculating the condensation speed includes: in, is the size of the cohesion term between the ith individual and the j2th neighbor individual in the cohesive neighbor set at the tth moment; C coh is the cohesion term coefficient; is the distance between the ith individual and the j2th neighbor individual in the cohesive neighbor set at the tth moment; is the minimum cohesion range; is the actual position of the j2th neighbor individual in the condensed neighbor set at time t; r i is the perception radius of the i-th individual; is the aggregation speed of the i-th individual at the t+1th moment; is the cohesive neighbor set of the i-th individual at time t; The alignment speed calculation formula includes: in, is the alignment speed of the i-th individual at the t+1th moment; is the actual speed of the j3rd neighbor individual in the aligned neighbor set at time t; v i (t) is the actual speed of the i-th individual at the t-th moment; is the set of aligned neighbors of the i-th individual at time t; is the distance between the i-th individual and the j3-th neighbor individual in the aligned neighbor set at the t-th moment; The self-driving speed calculation formula includes: in, is the self-driving speed of the i-th individual at the t+1th moment; C pro is the self-driving term coefficient; |·| is the Euclidean norm; v tar (t) is the actual speed of the tracking target at time t; The navigation speed calculation formula includes: in, is the navigation speed of the i-th individual at the t+1th moment; C nav is the navigation item coefficient; p tar (t) is the actual position of the tracking target at time t; The suction speed calculation formula includes: in, is the attraction speed of the i-th individual at the t+1th moment; C tar is the attraction coefficient; d itar (t) is the distance between the i-th individual and the tracking target at the t-th moment.

6. The method for dynamic target tracking of heterogeneous swarm robots according to claim 5, characterized in that: Based on the state of the current individual at the current moment and the speed parameter of the current individual at the next moment, the predicted speed of the current individual at the next moment is determined, including: The speed prediction formula is used to determine the predicted speed of the current individual at the next moment according to the state of the current individual at the current moment and the speed parameter of the current individual at the next moment; the speed prediction formula includes: Among them, v' i (t+1) is the predicted speed of the i-th individual at the t+1th time; is the alignment coefficient of individuals whose cluster activation state is 0; is the cluster activation state of the i-th individual at the t-th moment; is the environmental activation state of the i-th individual at time t; is the alignment coefficient for individuals whose cluster activation state is 1.

7. The method for dynamic target tracking of heterogeneous swarm robots according to claim 6, characterized in that: Based on the actual position of the current individual at the current moment and the predicted speed of the current individual at the next moment, the predicted position of the current individual at the next moment is determined, including: The predicted position of the current individual at the next moment is determined by using a position prediction formula according to the actual position of the current individual at the current moment and the predicted speed of the current individual at the next moment; the position prediction formula includes: p’ i (t+1)=p i (t)+v’ i (t+1)×Δt; Among them, p' i (t+1) is the predicted position of the i-th individual at the t+1th time; Δt is the unit time step.

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the heterogeneous swarm robot dynamic target tracking method as described in any one of claims 1-7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the dynamic target tracking method of heterogeneous swarm robots described in any one of claims 1 to 7 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the dynamic target tracking method of heterogeneous swarm robots described in any one of claims 1 to 7 is implemented.