Collaboration and grouping method of swarm robots inspired by bacterial tropism

Abstract

The invention relates to a collaboration and grouping method for swarm robots inspired by bacterial tropism, which comprises the following steps of: establishing a fitness function of a robot according to position information of neighbors in a local sensing range of an individual robot on the basis of a bacterial food source sensing mechanism; based on the fitness function and the bacterial tropism mechanism, determining the basic behavior state of the robot, and obtaining a motion switching result about whether behavior state switching occurs or not; correcting the motion state of the robot and determining the rotation angle of the robot based on the switching result; wherein the rotation angle is used for controlling the swarm robots to keep the same kind of subgroups agglomerated, keep away from different kinds of groups and search for other same kind of individuals, gathering in a certain area of the environment, and grouping the swarm robots. According to the method, the robot behavior state is determined by analyzing the bacterial tropism mechanism, the behavior state switching mechanism is constructed, the rotation angle is corrected, the swarm robot grouping effect is achieved, swarm robot distributed cooperative grouping is achieved, and the grouping efficiency is improved.

Description

Technical Field

[0001] This invention relates to the field of swarm robotics, and in particular to a method for collaboration and grouping of swarm robots inspired by bacterial tropism. Background Technology

[0002] The swarm behavior of swarm robots is often inspired by the social behaviors of organisms in nature, such as the division of labor in nest building by ants and bees. Swarming is a typical swarm behavior of swarm robots, primarily achieved through cooperation among individuals within the group, breaking down the swarm into subgroups with different functions and aggregating them in different locations. Current swarming methods generally suffer from problems such as a central control unit, poor swarming ability in multi-dissimilar swarm robot systems, complex individual behaviors, and the need for high computational power in individual units. These issues not only lead to high system power consumption but also affect the efficiency, robustness, and success rate of swarming. Therefore, to overcome these problems, it is necessary to deeply explore the underlying mechanisms through behavioral observations of typical social organisms, build models, and design new swarming methods. Currently, there is a lack of a fully distributed swarming method with fewer individual behavioral states, which could generate simple individual behaviors similar to bacteria through simple interactions between individuals, thereby achieving swarming and applying it to fields such as multi-regional reconnaissance by swarm robots. Summary of the Invention

[0003] The purpose of this invention is to provide a collaborative and grouping method for swarm robots inspired by bacterial tropism, applicable to reconnaissance tasks in different areas of swarm robots. The method involves initially determining the robot's sensory input through bacterial food source perception mechanisms; based on bacterial tropism analysis, using bacterial foraging behavior models or data analysis, determining the robot's basic behavioral state; inputting sensory information into individual robots for analysis and decision-making, designing a grouping method, and making preliminary predictions to determine behavioral states; and optimizing model parameters through experimental efficiency and success rate analysis.

[0004] To achieve the above objectives, the present invention provides the following solution:

[0005] A collaborative and grouping method for swarm robots inspired by bacterial tropism, comprising:

[0006] Based on the bacterial food source perception mechanism, a fitness function for the robot is established by taking into account the location information of its neighbors within the local perception range of an individual robot.

[0007] Based on the fitness function and bacterial tropism mechanism, the basic behavioral state of the robot is determined, and the motion switching result of whether a behavioral state switching has occurred is obtained.

[0008] Based on the switching result, the robot's motion state is corrected and the robot's rotation angle is determined; wherein, the rotation angle is used to control the group of robots to maintain the cohesion of similar subgroups, stay away from different groups, and search for other similar individuals, and to gather in a certain area of ​​the environment, thus performing grouping of the robots.

[0009] Optionally, establishing the robot fitness function includes:

[0010] Based on the robot's perception radius, obtain the same-type and different-type neighbors within the robot's local range, and determine the location information of the same-type and different-type neighbors;

[0011] The robot's fitness function is obtained based on the location information of its similar and dissimilar neighbors.

[0012] Based on the fitness function, a target function for the robot is established to determine whether the robot group has completed aggregation.

[0013] The fitness function value is used as the robot's perception input information at the current moment.

[0014] Optionally, the robot's fitness function is:

[0015]

[0016] Among them, F i Let x represent the fitness function of robot i, n represent the number of robots of the same type as robot i within its perception range, j represent robot j of the same type as robot i, and x represent the fitness function of robot i. i (t) represents the position information of robot i, x j (t) represents the position information of robot j at time t, m represents the number of alien robots within the perception range of robot i, k represents alien robot k, and y k (t) represents the position information of the alien robot k at time t;

[0017] The objective function is:

[0018]

[0019] Among them, F i (t) represents the fitness function of robot i of the same class at time t, where i represents the robot index and N represents the set of robots of the same class as robot i.

[0020] Optionally, determining the robot's basic behavioral state based on the fitness function includes:

[0021] Based on the fitness function, the difference between the fitness function values ​​of the robot at the current time and at historical time is obtained;

[0022] Based on the difference between the robot's fitness function values ​​at the current moment and at historical moments, a robot motion state switching strategy model is constructed.

[0023] Based on the robot motion state switching strategy model, the robot's basic behavioral states are determined, and it is determined whether the robot should switch behavioral states.

[0024] Optionally, the robot motion state switching strategy model is as follows:

[0025]

[0026] Among them, v i (t) represents the robot's velocity at the current moment, v i (t-1) represents the robot's velocity at the previous moment, ΔF i (t) represents the difference between the robot's fitness function values ​​at the current and historical moments, θ i (t) represents the robot's rotation angle, P i (t) represents the motion state of robot i, and Δ represents the behavior state.

[0027] Optionally, determining the robot's basic behavioral states includes:

[0028] Inspired by bacterial tropism, when the difference between the robot's current fitness function value and its historical fitness function value is greater than zero, the robot maintains its current favorable direction of motion; otherwise, the robot rotates to move and regains the fitness function value of its current position.

[0029] Optionally, based on the switching result, when the robot's fitness function value is less than zero at the previous and next time moments, correcting the robot's straight-line motion state to a rotational state and determining the robot's rotation angle includes:

[0030] When the robot changes its behavior state, the rotation angle is θ. i (t):

[0031]

[0032] In the formula, v i (t) represents the velocity of robot i at time t, v i (t-1) represents the velocity of robot i at time t-1;

[0033] Within the robot's perception range, individuals of the same type exhibit attraction F. i r (t), the alien robot exhibits a repulsive force F i g (t), taking into account both attractive and repulsive forces, the formula for the robot's velocity after rotation is as follows:

[0034]

[0035] In the formula, α1, α2, and α3 represent contribution factors, and w 1i (t), w 2i (t) and w 3i (t) represents the gain factors of the attraction term, repulsion term, and random search term, respectively, v i (t) represents the velocity of robot i after its rotation, rand(a,b) represents the random search term for other similar individuals outside robot i's perception range, k represents an indirect similar individual k within robot i's neighbor perception range but outside robot i's perception range, C represents the set of similar robots outside robot i's perception range, and γ ik (t) represents the effect of robot k on robot i.

[0036] Optionally, the gain factor is obtained based on the changing density information of similar subgroups, while balancing the repulsion of different subgroups, the cohesion of similar subgroups, and random search.

[0037] The formula for calculating the gain factor is:

[0038]

[0039] w 3i (t)=cw 1i (t),w 2i (t)=1-w 1i (t)-w 3i (t)

[0040] Among them, Num i (t) represents the number of robots of the same class i, where 0 ≤ c ≤ 1.

[0041] The beneficial effects of this invention are as follows:

[0042] This invention provides a collaborative and grouping method for swarm robots inspired by bacterial tropism. Based on the neighboring position information of the robot's current position, a fitness function is constructed for the system, and the fitness function value is used as the robot's sensory input. Behavioral state switching is determined based on existing bacterial tropism mechanisms and the robot's fitness function value. The rotation angle of different robot classes is determined by the distribution of similar and indirectly similar classes within a local range and random search terms, and the rotation angle is calculated based on the robot's position information within the sensing range. Gain and contribution factors are determined based on the density information of similar subgroups within the sensing range and an empirical value 'c' obtained from experimental data analysis. This invention achieves swarm robot grouping by analyzing bacterial tropism mechanisms to determine robot behavioral states, constructing behavioral state switching mechanisms, and correcting rotation angles, thus realizing distributed collaboration among swarm robots and improving grouping efficiency. Attached Figure Description

[0043] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0044] Figure 1 This is a schematic diagram of the mapping between bacteria and robots according to an embodiment of the present invention;

[0045] Figure 2 This is a schematic diagram of swarm robot grouping according to an embodiment of the present invention;

[0046] Figure 3 This invention provides a method for the collaboration and grouping of swarm robots inspired by bacterial tropism. Detailed Implementation

[0047] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0048] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0049] In this embodiment, the robot is considered as a bacterium, such as Figure 1As shown, the robot uses locally perceived neighbor location information as perceptual input to determine its current fitness function value, compares it with the fitness function value from the previous moment, and makes behavioral decisions, thereby generating corresponding body actions. This process is similar to bacterial tropism during bacterial foraging.

[0050] like Figure 3 As shown, this embodiment proposes a collaborative and grouping method for swarm robots inspired by bacterial tropism, including:

[0051] Based on the bacterial food source perception mechanism, a fitness function for the robot is established by taking into account the location information of its neighbors within the local perception range of an individual robot.

[0052] Based on the fitness function and bacterial tropism mechanism, the basic behavioral states of the robot are determined, and the motion switching results of whether a behavioral state switch has occurred are obtained.

[0053] Based on the switching results, the robot's motion state is corrected and the robot's rotation angle is determined. The rotation angle is used to control the group of robots to maintain cohesion of similar subgroups, stay away from dissimilar groups, and search for other similar individuals, so as to gather in a certain area of ​​the environment and perform grouping of the robots.

[0054] Furthermore, establishing the robot's fitness function includes:

[0055] Based on the robot's perception radius, obtain the same-type and different-type neighbors within the robot's local range, and determine the location information of the same-type and different-type neighbors;

[0056] The robot's fitness function is obtained based on the location information of its similar and dissimilar neighbors.

[0057] Based on the fitness function, the objective function of the robot is established to determine whether the robot group has completed aggregation;

[0058] The fitness function value is used as the robot's perception input information at the current moment.

[0059] Specifically, in this embodiment, the fitness function for the initial determination of the swarm robot grouping method includes:

[0060] Based on the existing bacterial food source perception mechanism, a fitness function is established to determine the robot's perception input information by taking into account the location information of neighbors within the local perception range of an individual robot.

[0061] For example, given an environment of size H×H, and robot i's perception radius is r... i At that time, the robot's local neighbors of the same type are obtained as a1, a2, ..., a n and different neighbors b1, b2, ..., bm The location information of similar and dissimilar neighbors are (x1, x2, ..., x...). n ) and (y1,y2,...,y n Based on the location information of its similar and dissimilar neighbors, the fitness function of robot i can be obtained as follows: Therefore, the objective function for the same subgroup is designed as follows: The subgroup system aims to maximize the objective function, max(F(t)). The fitness function is known as the robot's sensory input information at the current moment.

[0062] Furthermore, based on the fitness function, the basic behavioral states of the robot are determined as follows:

[0063] Based on the fitness function, the difference between the fitness function values ​​of the robot at the current time and at historical time is obtained;

[0064] Based on the difference between the robot's fitness function values ​​at the current moment and at historical moments, a robot motion state switching strategy model is constructed.

[0065] The robot motion state switching strategy model is used to determine whether the robot should switch its behavior state.

[0066] Furthermore, the robot's state switching strategy is as follows:

[0067] Inspired by bacterial tropism, when the difference between the robot's fitness function value at the current moment and at a previous moment is greater than zero, the robot maintains its current advantageous direction of movement. This process is similar to the "swimming" behavior in bacterial tropism. Conversely, the robot rotates to move and regains the fitness function value for its new position. This process is similar to the "rolling" behavior in bacterial tropism.

[0068] Specifically, in this embodiment, the basic behavioral state of the robot is determined based on the observation of bacterial foraging behavior and the analysis of bacterial tropism mechanism, and it is determined whether a behavioral state switch has occurred.

[0069] Among them, determining the basic behavioral states of the robot is to determine the basic behavioral states of the robot, such as the "swimming" and "rolling" of bacteria.

[0070] By using the robot's sensory input information from adjacent time intervals, the method for switching robot behavior states is determined, and it is judged whether the robot should switch behavior states.

[0071] Based on the obtained robot motion switching results, the robot motion state is corrected and the robot rotation angle is determined. The rotation angle is used to determine the motion direction of the group of robots in different areas, thus completing the group robot grouping method.

[0072] Based on observations of bacterial foraging behavior and analysis of bacterial tropism mechanisms, the robot's basic behavioral states were determined, and it was determined whether a behavioral state transition occurred. This included the following:

[0073] Bacteria capable of swimming, such as *E. coli*, typically possess flagella and receptors. These receptors allow bacteria to sense food source information and chemical molecules released by other individuals, switching between "swimming" and "tumbling" behavioral states. Inspired by this mechanism, robot behavioral states are defined. Based on the fitness function defined above, the robot obtains the difference ΔF between the fitness function values ​​at the current and previous times. i (t). Construct a robot motion state switching strategy model:

[0074]

[0075] In the formula, v i (t) represents the robot's current velocity, v i (t-1) represents the robot's velocity at the previous moment. When ΔF i When (t)>0, the robot maintains its current favorable direction of motion; otherwise, the robot rotates to move and regains the fitness function value F of its new position. i '(t+1).

[0076] Based on the fitness function, an objective function F is established for the same type of subgroup to determine whether the same type of robots have completed aggregation. If they have, they exit the system.

[0077] Specifically, in this embodiment, the robot motion state is corrected and the robot rotation angle is determined based on the obtained robot motion switching results.

[0078] The method for rotating the angle is as follows:

[0079] Based on the location information of neighbors within the individual's range, and combined with random search terms and indirect similar terms, the expression for the rotation angle is obtained.

[0080] Based on the changing density information of similar subgroups, while balancing subgroup cohesion and random search, the gain factors of each part of the velocity are obtained, and the contribution factor is obtained based on empirical values.

[0081] When the robot changes its behavior state, the rotation angle is θ. i (t):

[0082]

[0083] In the formula, v i (t) represents the direction of motion of robot i at time t. Within robot i's perception range, similar individuals exhibit attraction F. ir (t), the alien robot exhibits a repulsive force F i g (t), taking into account both attractive and repulsive forces, the formula for the velocity after rotation is as follows:

[0084]

[0085] In the formula, the first term on the right helps the robot attract similar individuals within its perception range, maintaining the compactness of the subgroup; the second term helps the robot repel different individuals; the third term can drive the robot to search randomly while maintaining the compactness of the similar subgroup; and the fourth term can indirectly expand the robot's perception range, helping the robot to gather towards other similar individuals that its neighbors can perceive.

[0086] Specifically, this embodiment also obtains a gain factor based on the changing density information of similar subgroups, while balancing the repulsion of different subgroups, the aggregation of similar subgroups, and random search.

[0087] Methods for determining gain factors and contribution factors include:

[0088] As the number of similar individuals within the robot's perception range increases, the gain factor of the attraction component will decrease, while the gain factor of the random search component will increase. Combined with the gain factor of the repulsion component, a balance is achieved among the three factors.

[0089] Based on the experimental data, the efficiency and success rate of system clustering were analyzed, and empirical values ​​for each contribution factor were obtained.

[0090] The robot's speed is mainly affected by the attraction of similar objects, the repulsion of dissimilar objects, and random search, i.e., the first three parts of formula (3). 1i (t), w 2i (t) and w 3i (t) can be used to amplify or reduce the effect of each force. This invention can take w 1i The range of (t) is [0c], 0≤c≤1, and the calculation formula is as follows:

[0091]

[0092] In the formula, Nu m i (t) represents the number of robots of the same class as robot i. i (t) and w3 i (t) satisfies w 3i (t)=cw 1i (t),w 2i (t)=1-w 1i (t)-w 3i (t). α1, α2 and α3 are contribution factors, which can be obtained empirically based on simulation data analysis.

[0093] Furthermore, w 1i (t), w 2i (t) and w 3i Substituting (t) into formula (3) yields the robot's velocity v. i (t), and then the rotation angle θ can be obtained using formula (2). i (t).

[0094] There are several types of robot groups distributed in the environment, such as Figure 2 As shown, different numbers represent different types of robots. The purpose of grouping robots is to separate heterogeneous robot groups based on the characteristics of each robot, so that robots of the same type can form a cohesive team, while different robots are separated from each other.

[0095] This embodiment proposes a method to determine whether a robot has switched behavioral states by observing bacterial foraging behavior and analyzing bacterial tropism mechanisms. Based on the obtained robot motion switching results, a method is used to correct the robot's motion state and determine the robot's rotation angle. Based on the changing density information of similar subgroups, and balancing heterogeneous repulsion, subgroup cohesion, and random search, a gain factor method is obtained.

[0096] This embodiment provides a collaborative and grouping method for swarm robots inspired by bacterial tropism. It constructs a system fitness function based on the neighboring positions of the robot's current position, using the fitness function value as the robot's sensory input. It judges robot behavior state transitions based on existing bacterial perception mechanisms and robot fitness function values. It calculates rotation angles based on robot position information within the perception range and random search terms. It determines gain and contribution factors based on density information of similar subgroups within the perception range and empirical values ​​obtained from experimental data analysis. This application achieves swarm robot grouping by constructing a behavior state transition mechanism and correcting rotation angles through bacterial tropism analysis, realizing distributed collaboration among swarm robots and improving grouping efficiency.

[0097] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for collaboration and grouping of swarm robots inspired by bacterial tropism, characterized in that, include: Based on the bacterial food source perception mechanism, a fitness function for the robot is established by taking into account the location information of its neighbors within the local perception range of an individual robot. Based on the fitness function and bacterial tropism mechanism, the basic behavioral state of the robot is determined, and the motion switching result of whether a behavioral state switching has occurred is obtained. Based on the switching result, the robot's motion state is corrected and the robot's rotation angle is determined; wherein, the rotation angle is used to control the group of robots to maintain the cohesion of similar subgroups, stay away from different groups, and search for other similar individuals, and to gather in a certain area of ​​the environment, thus performing grouping of the robots.

2. The method for cooperation and grouping of swarm robots inspired by bacterial tropism according to claim 1, characterized in that, Establishing the robot fitness function includes: Based on the robot's perception radius, obtain the same-type and different-type neighbors within the robot's local range, and determine the location information of the same-type and different-type neighbors; The robot's fitness function is obtained based on the location information of its similar and dissimilar neighbors. Based on the fitness function, a target function for the robot is established to determine whether the robot group has completed aggregation. The fitness function value is used as the robot's perception input information at the current moment.

3. The method for cooperation and grouping of swarm robots inspired by bacterial tropism according to claim 2, characterized in that, The robot's fitness function is: Among them, F i Let x represent the fitness function of robot i, n represent the number of robots of the same type as robot i within its perception range, j represent robot j of the same type as robot i, and x represent the fitness function of robot i. i (t) represents the position information of robot i, x j (t) represents the position information of robot j at time t, m represents the number of alien robots within the perception range of robot i, k represents alien robot k, and y k (t) represents the position information of the alien robot k at time t; The objective function is: Among them, F i (t) represents the fitness function of robot i of the same class at time t, where i represents the robot index and N represents the set of robots of the same class as robot i.

4. The method for cooperation and grouping of swarm robots inspired by bacterial tropism according to claim 1, characterized in that, Based on the fitness function, the basic behavioral states of the robot are determined as follows: Based on the fitness function, the difference between the fitness function values ​​of the robot at the current time and at historical time is obtained; Based on the difference between the robot's fitness function values ​​at the current moment and at historical moments, a robot motion state switching strategy model is constructed. Based on the robot motion state switching strategy model, the robot's basic behavioral states are determined, and it is determined whether the robot should switch behavioral states.

5. The method for cooperation and grouping of swarm robots inspired by bacterial tropism according to claim 4, characterized in that, The robot motion state switching strategy model is as follows: Among them, v i (t) represents the robot's velocity at the current moment, v i (t-1) represents the robot's velocity at the previous moment, ΔF i (t) represents the difference between the robot's fitness function values ​​at the current and historical moments, θ i (t) represents the robot's rotation angle, P i (t) represents the motion state of robot i, and Δ represents the behavior state.

6. The method for cooperation and grouping of swarm robots inspired by bacterial tropism according to claim 5, characterized in that, Determining the basic behavioral states of a robot includes: Inspired by bacterial tropism, when the difference between the robot's current fitness function value and its historical fitness function value is greater than zero, the robot maintains its current favorable direction of motion; otherwise, the robot rotates to move and regains the fitness function value of its current position.

7. The method for cooperation and grouping of swarm robots inspired by bacterial tropism according to claim 1, characterized in that, Based on the switching result, correcting the robot's motion state and determining the robot's rotation angle includes: When the robot changes its behavior state, the rotation angle is θ. i (t): In the formula, v i (t) represents the velocity of robot i at time t, v i (t-1) represents the velocity of robot i at time t-1; Within the robot's perception range, individuals of the same type exhibit attraction F. i r (t), the alien robot exhibits a repulsive force F i g (t), taking into account both attractive and repulsive forces, the formula for the robot's velocity after rotation is as follows: In the formula, α1, α2, and α3 represent the contribution factors of different terms, and w 1i (t), w 2i (t) and w 3i (t) represents the gain factors of the attraction term, repulsion term, and random search term, respectively, v i (t) represents the velocity of robot i after its rotation, rand(a,b) represents the random search term for other similar individuals outside robot i's perception range, k represents an indirect similar individual k within robot i's neighbor perception range but outside robot i's perception range, C represents the set of similar robots outside robot i's perception range, and γ ik (t) represents the effect of robot k on robot i.

8. The method for cooperation and grouping of swarm robots inspired by bacterial tropism according to claim 7, characterized in that, The gain factor is obtained based on the changing density information of similar subgroups, while balancing the repulsion of dissimilar subgroups, the cohesion of similar subgroups, and random search. The formula for calculating the gain factor is: w 3i (t)=c-w 1i (t),w 2i (t)=1-w 1i (t)-w 3i (t) Among them, Num i (t) represents the number of robots of the same class as robot i.

Images