A formation generation method based on self-organizing incremental topology

Abstract

This invention presents a formation generation method based on self-organizing incremental topology. It can self-organize and generate rigid and relatively excellent incremental communication topologies for multi-agent formations containing increments, significantly improving the stability and flexibility of the formations. Specifically, this invention combines the concept of incremental learning with the self-organizing topology generation method, proposing a self-organizing incremental topology generation method. The original topology can be regarded as a generalized topology node, and together with newly added nodes, it self-organizes to generate incremental topologies. This method maintains the original optimal topology while ensuring the rigidity of the newly generated incremental topology, greatly simplifying the computation of multi-agent formations and achieving the research goal of improving flexibility. Furthermore, during the formation generation process, each agent can obtain optimal input by utilizing local information from its neighbors and itself, avoiding the limitation of communication between agents due to limited communication bandwidth.

Description

Technical Field

[0001] This invention belongs to the field of intelligent control and artificial intelligence technology, and in particular relates to a formation generation method based on self-organizing incremental topology. Background Technology

[0002] In recent years, multi-agent swarm control has become increasingly popular, with many scholars dedicated to its research. Multi-agent swarm control refers to coordinated control achieved through the exchange of information among multiple agents. It possesses excellent adaptability and flexibility, enabling it to solve many tasks and jobs that a single agent cannot handle.

[0003] Multi-agent formation control can be divided into communication topology control and motion control corresponding to this communication topology. For multi-agent formations, the optimal communication topology is often difficult to obtain easily. Currently, the communication topology structure of most formations is pre-set manually. This method of topology generation has significant limitations and cannot cope with unexpected situations. Therefore, self-organizing topology generation methods are particularly important for multi-agent formations. They can greatly improve the autonomy and flexibility of the formation system, allowing it to better respond to unexpected situations. Relevant patents and papers have already integrated self-organizing topology generation methods into multi-agent formations, specifically utilizing position information to self-organize and generate excellent communication topologies for the formation.

[0004] However, a multi-agent formation system may have a large number of agents. If additional agents are needed during movement, directly adopting a self-organizing topology would require a large amount of position coordinate data. Such an operation would lead to a very complex computational load, which is not conducive to real-time communication between agents. Therefore, it is necessary to extend research in this direction to improve the stability, flexibility, and reduce the computational complexity of the formation. Summary of the Invention

[0005] To address the problems existing in the prior art, the purpose of this invention is to provide a formation generation method based on self-organizing incremental topology, which can self-organize and generate rigid and relatively excellent incremental communication topology structures for multi-agent formations containing increments, thereby improving the stability and flexibility of the formations.

[0006] The present invention adopts the following technical solution:

[0007] A formation generation method based on self-organizing incremental topology, comprising a multi-agent formation, characterized by the following steps:

[0008] Step 1: Analyze the communication topology of multi-agent formation based on graph theory;

[0009] Step 2: Based on graph theory analysis and the conditions required for communication topology, design a formation generation method based on self-organizing incremental topology;

[0010] B1. Obtain the position information of N initial agents at time t, and define the position of the agent at time t as: p ixy (t)=[x i (t),y i (t)] T (i∈V), where x i (t),y i (t) represents the position coordinates of agent i at time t;

[0011] B2. Treat the agents in the formation as points and draw a scatter plot;

[0012] B3. Use the Delaunay Triangulation function to triangulate the scatter plot to obtain the triangulated plot at time t. A triangulated plot is a rigid plot because it divides the scatter plot using many triangles.

[0013] B4. Calculate the distance of each edge in the rigid graph in step B3 using the distance function, and use the distance as the weight of each edge;

[0014] B5. Remove edges from the weighted rigid graph in B4 to obtain the optimal rigid graph, such that the number of edges in the graph satisfies: Num(E) = 2Num(V) - 3, and obtain the undirected graph G = (V, E, A), which is the optimal rigid graph.

[0015] B6. Select an agent R from the formation. L As a leader, R L If it can only receive commands from the host and cannot receive commands from other intelligent agents, then d in (Leader) = 0; other agents must also satisfy this condition. The condition is to add directions to the edges of the optimal rigid graph to obtain the optimal persistent graph, i.e., the optimal topology.

[0016] B7. At time T, M new agents are added to the multi-agent formation;

[0017] B8. Obtain the position information of the M newly added agents at time T, and define it as: p ixy (T)=[x i (T),y i (T)] T (i = N+1, ..., N+M);

[0018] B9. Obtain the optimal topology at time T. And construct the topological boundary edge set SE and the topological boundary point set SP;

[0019] B10. Use the convexHull function to obtain the convex hull (outer boundary of the scatter plot), construct the convex hull edge set DE and the convex hull boundary point set DP, and mark the geometric center of the convex hull as point C.

[0020] B11. Establish a rectangular coordinate system with point C as the origin, calculate the angle (0~2π) between the line connecting each boundary point and the center point C and the positive direction of the X-axis, and calculate the angle range between two adjacent convex hull boundary points.

[0021] B12. Divide the region outside the convex hull in step B11 into ranges and number each range.

[0022] B13. Add nodes to the region outside the convex hull, and determine the region to which the node belongs by the angle between the line connecting the new node and the center point C and the positive direction of the X-axis.

[0023] B14. If a region adds only one new point, then the two boundary points of the region are directly connected in the form of in-degree.

[0024] B15. If more than one new point is added to a region, the distance from each new point to the edge belonging to the convex hull of that region needs to be calculated. The point with the closest distance is considered the new point in this round, and the remaining points are considered subsequent new points. Step B13 is then performed on the new points in this round. Then, the region is partitioned again, and steps B13, B14, and B15 are repeated until topological links are completed for M new points. The final directed graph obtained is the incremental topology structure.

[0025] Preferably, the directed graph Represents the topology of an incremental multi-agent formation, where · + This only indicates that the data has changed after a new agent was added to the formation; its meaning itself remains unchanged.

[0026] Preferably, the communication topology of the multi-agent formation needs to meet the following conditions:

[0027] Rigid diagram condition: p ixy ≠p jxy ,||p ixy (t)-p jxy (t)||=l,p ixy =(x i ,y i ) is the position sequence of agent i, and l is a constant;

[0028] Minimum rigidity diagram condition: Num(E) = 2Num(V) - 3, where Num(g) represents the quantity;

[0029] Persistent graph conditions: din (i) represents the in-degree of agent i;

[0030] Navigator In-degree Restriction Conditions: d in (Leader) = 0 indicates that the leader's in-degree is 0. The leader can only receive information from the host and cannot receive information from other intelligent agents, hence its in-degree is 0.

[0031] Preferably, the topology has a topological boundary edge set SE, a topological boundary point set SP, a convex hull edge set DE, and a convex hull point set DP.

[0032] Preferably, step one can define G = (V, E, A) as an undirected graph with N nodes, where V = 1, 2, ..., N is a set of N nodes, E = {(i, j), i ≠ j, i, j ∈ V} is a set of undirected edges, and A = (a ij )∈R N×N Let R be the adjacency matrix. N×N Representing the real number field, such as If (i,j)∈E, then a ij =1, otherwise a ij =0.

[0033] Preferably, step one can further define a directed graph to represent the topology of a formation containing N agents, where V = 1, 2, ..., N represents a set of N agents. Let be the set of directed edges. This indicates that agent i in the formation sends information to agent j. Represents the adjacency matrix in the formation, if but on the contrary

[0034] Compared with the prior art, the advantages of this invention are:

[0035] 1) This invention can self-organize and generate rigid and relatively excellent incremental communication topologies for multi-agent formations with increments, thereby significantly improving the stability and flexibility of the formations. Specifically, this invention proposes a self-organizing incremental topology generation method based on the self-organizing topology generation method and combining it with the idea of ​​incremental learning. The original topology can be regarded as a generalized topology node, and the incremental topology can be generated together with the newly added node. This method can maintain the original optimal topology while making the newly generated incremental topology rigid, thereby greatly simplifying the computation of multi-agent formations and achieving the research and development goal of improving flexibility.

[0036] 2) This invention can automatically generate the optimal communication topology using the position information of each agent during the initial formation generation process, avoiding the situation where the manually preset topology becomes inapplicable due to formation movement;

[0037] 3) This invention can generate an incremental topology structure by utilizing the position information of the newly added agent and some agents in the original formation during the incremental formation generation process. It can maintain the original topology structure without change and form a new incremental topology structure on its basis, thus reducing the computational complexity.

[0038] 4) In the formation generation process, each agent can obtain the optimal input by using local information from its neighbors and itself, thus avoiding the situation where communication between agents is limited due to limited communication bandwidth. Attached Figure Description

[0039] Figure 1 This is an illustration of the topological boundary edges, convex hull edges, topological boundary points, and convex hull points of this invention.

[0040] Figure 2 This is a flowchart of the formation generation method of the present invention.

[0041] Figure 3 This is a simulation diagram of the present invention. Detailed Implementation

[0042] 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.

[0043] like Figures 1 to 3 As shown, this invention is a formation generation method based on self-organizing incremental topology, applicable to the formation generation of planar motion intelligent agents, such as differential motion wheeled robots, omnidirectional motion wheeled robots, four-wheel drive robots, etc. This invention utilizes the position information of the intelligent agents to self-organize and generate a rigid communication topology structure with low loss, resulting in good stability for multi-agent formations. Furthermore, the communication topology structure designed in this invention can be combined with some formation motion controllers, such as consensus formation, to obtain the ALB matrix information of the communication topology (A: adjacency matrix, L: Laplace matrix, B: traction matrix), construct the consensus deviation, and converge it to 0, thus completing the consensus formation control of multi-agents.

[0044] The specific operation can be divided into two steps:

[0045] Step 1: Analyze the communication topology of multi-agent formations based on graph theory:

[0046] We can define G = (V, E, A) as an undirected graph with N nodes, where V = 1, 2, ..., N is the set of N nodes, E = {(i, j), i ≠ j, i, j ∈ V} is the set of undirected edges, and A = (a ij )∈R N×N Let R be the adjacency matrix. N×N Representing the real number field, such as If (i,j)∈E, then a ij =1, otherwise a ij =0;

[0047] A directed graph can be defined to represent the topology of a group of N agents, where V = 1, 2, ..., N represents the set of N agents. Let be the set of directed edges. This indicates that agent i in the formation sends information to agent j. Represents the adjacency matrix in the formation, if but on the contrary

[0048] Step 2: Based on graph theory analysis and the conditions required for communication topology, design a formation generation method based on self-organizing incremental topology;

[0049] B1. Obtain the position information of N initial agents at time t, and define the position of the agent at time t as: p ixy (t)=[x i (t),y i (t)] T (i∈V), where x i (t),y i (t) represents the position coordinates of agent i at time t;

[0050] B2. Treat the agents in the formation as points and draw a scatter plot;

[0051] B3. Use the Delaunay Triangulation function to triangulate the scatter plot to obtain the triangulated plot at time t. A triangulated plot is a rigid plot because it divides the scatter plot using many triangles.

[0052] B4. Calculate the distance of each edge in the rigid graph in step B3 using the distance function, and use the distance as the weight of each edge;

[0053] B5. Remove edges from the weighted rigid graph in B4 to obtain the optimal rigid graph, such that the number of edges in the graph satisfies: Num(E) = 2Num(V) - 3, and obtain the undirected graph G = (V, E, A), which is the optimal rigid graph.

[0054] B6. Select an agent R from the formation. L As a leader, R L If it can only receive commands from the host and cannot receive commands from other intelligent agents, then d in (Leader) = 0; other agents must also satisfy this condition. The condition is to add directions to the edges of the optimal rigid graph to obtain the optimal persistent graph, i.e., the optimal topology.

[0055] B7. At time T, M new agents are added to the multi-agent formation;

[0056] B8. Obtain the position information of the M newly added agents at time T, and define it as: p ixy (T)=[x i (T),y i (T)] T (i = N+1, ..., N+M);

[0057] B9. Obtain the optimal topology at time T. And construct the topological boundary edge set SE and the topological boundary point set SP;

[0058] B10. Use the convexHull function to obtain the convex hull (outer boundary of the scatter plot), construct the convex hull edge set DE and the convex hull boundary point set DP, and mark the geometric center of the convex hull as point C.

[0059] B11. Establish a rectangular coordinate system with point C as the origin, calculate the angle (0~2π) between the line connecting each boundary point and the center point C and the positive direction of the X-axis, and calculate the angle range between two adjacent convex hull boundary points.

[0060] B12. Divide the region outside the convex hull in step B11 into ranges and number each range.

[0061] B13. Add nodes to the region outside the convex hull, and determine the region to which the node belongs by the angle between the line connecting the new node and the center point C and the positive direction of the X-axis.

[0062] B14. If a region adds only one new point, then the two boundary points of the region are directly connected in the form of in-degree.

[0063] B15. If more than one new point is added to a region, the distance from each new point to the edge belonging to the convex hull of that region needs to be calculated. The point with the closest distance is considered the new point in this round, and the remaining points are considered subsequent new points. Step B13 is then performed on the new points in this round. Then, the region is partitioned again, and steps B13, B14, and B15 are repeated until topological links are completed for M new points. The final directed graph obtained is the incremental topology structure.

[0064] Step two requires the following constraints to be met in order to self-organize and generate the optimal topology using the initial agent's position information, and to generate an incremental topology for the formation of mobile agents with increments. The specific conditions are:

[0065] Rigid diagram condition: p ixy ≠p jxy ,||p ixy (t)-p jxy (t)||=l,p ixy =(x i ,y i ) is the position sequence of agent i, and l is a constant;

[0066] Minimum rigidity diagram condition: Num(E) = 2Num(V) - 3, where Num(g) represents the quantity;

[0067] Persistent graph conditions: d in (i) represents the in-degree of agent i;

[0068] Navigator In-degree Restriction Conditions: d in (Leader) = 0 indicates that the leader's in-degree is 0. The leader can only receive information from the host and cannot receive information from other intelligent agents, hence its in-degree is 0.

[0069] Among them, the combination Figure 1 To distinguish between the following set definitions: topological boundary edge set SE, topological boundary vertex set SP, convex hull edge set DE, and convex hull vertex set DP:

[0070] The topological boundary edge set consists of the first directed line (a1), the second directed line (a2), the third directed line (a3), the fourth directed line (a4), the fifth directed line (a5), the sixth directed line (a6), and the seventh directed line (a7), i.e., the topological boundary edge set.

[0071] The topological boundary point set is the first point (1), the second point (2), the third point (3), the fourth point (4), the fifth point (5), the sixth point (6), and the seventh point (7), that is, the topological boundary point set SP = (1 2 3 4 5 6 7);

[0072] The convex hull set is the outermost undirected line (b1), i.e., the convex hull set.

[0073] Convex hull point set: DP=(2 3 4 5 6 7).

[0074] Based on the above combination Figure 3 The simulation diagram showing the step-by-step generation from A to F demonstrates that this invention significantly simplifies the computational workload of multi-agent formations, improves the flexibility of formations, and can be combined with various formation techniques. It is applicable to a wide range of practical scenarios, such as mobile robot formations exploring unknown terrains and reconstructing SLAM maps, and mobile robot formations identifying and tracking target points. This invention is an excellent choice for accomplishing these tasks because it treats the original topology as a generalized topological node and then self-organizes it with newly added nodes to generate a new topology. This maintains the original topology as the optimal topology, ensures the rigidity of the new topology, and preserves the stability of the formation.

[0075] In the description of this invention, it should be understood that the terms "center," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "axial," "radial," and "circumferential" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0076] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0077] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection, an electrical connection, or a communication connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0078] In this invention, unless otherwise expressly specified and limited, the first feature "on" or "below" the second feature may be in direct contact with the first and second features, or indirect contact through an intermediate medium. In the description of this specification, references to terms such as "an embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

Claims

1. A method for formation generation based on self-organizing incremental topology, comprising a multi-agent formation, characterized in that, The steps are as follows: Step one, analyzing the communication topology structure of multi-agent formation based on graph theory; Step two, designing the formation generation method based on self-organizing incremental topology according to the graph theory analysis and the required conditions of communication topology; B1, obtain the position information of N initial agents at time t, define the position of the agent at time t as: p ixy (t) = [x i (t), y i (t)] T (i∈V), wherein x i (t), y i (t) represents the position coordinates of agent i at time t, wherein V = 1, 2, …, N, and V represents a set of N agents; B2, regarding the agents in the formation as points, and drawing a scatter plot; B3, using the Delaunay Triangulation function to triangulate the scatter plot to obtain the triangulation graph at time t. The triangulation graph is a rigid graph obtained by dividing the scatter plot with many triangles; B4, using the distance function to calculate the distance of each edge in the rigid graph in step B3, and taking the distance as the weight of each edge; B5, the edge-deletion processing is performed on the rigid graph with weight in B4 to obtain an optimal rigid graph, so that the edge number of the graph satisfies: Num(E) = 2Num(V) - 3, and an undirected graph G = (V, E, A) is obtained, wherein E = {(i, j), i≠j, i, j∈V} is a set of undirected edges, and A = (a ij )∈R N×N is an adjacency matrix; B6, select an agent R from the formation L As the leader, the R L Can only receive host commands, cannot receive commands from other agents, then d in (Leader) = 0, where d in (Leader) = 0 means that the leader has an in-degree of 0, and the leader can only receive host information and cannot receive information from other agents, so its in-degree is 0; other agents also satisfy d in (i)≤2 The condition is to add direction to the edges of the optimal rigid graph to obtain the optimal persistent graph, that is, the optimal topology Where d in (i)≤2 d in (i) represents the in-degree of agent i, and is the set of directed edges, represents the adjacency matrix in the formation; B7, at time T, adding M new agents to the multi-agent formation; B8, obtain the position information of the M new intelligent agents at time T, and define it as: p ixy (T) = [x i (T),y i (T)] T i = N + 1, …, N + M; B9、obtain the optimal topology structure at T moment And construct the topological boundary edge set SE and the topological boundary point set SP; B10, using the convexHull function to obtain the convex hull of the above scatter plot, constructing the convex hull edge set DE and the convex hull boundary point set DP, and marking the geometric center of the convex hull as the C point; B11, establishing a rectangular coordinate system with the C point as the origin, calculating the angle between the connection line of each boundary point and the center point C and the positive direction of the X axis, and the range of the angle is 0~2π, and calculating the angle range of the adjacent two convex hull boundary points; B12, dividing the area outside the convex hull into ranges, and numbering each range; B13, adding nodes outside the convex hull, and judging the belonging area of the new point through the angle between the connection line of the new point and the center point C and the positive direction of the X axis; B14, if there is only one new point in a certain area, directly connecting the two boundary points of the area in the form of in-degree; B15、If the number of new points in a region is greater than 1, the distance from each new point to the edge of the convex hull belonging to the range needs to be calculated, the nearest point is regarded as the new point of this round, and the remaining points are regarded as subsequent new points, step B13 is performed on the new point of this round, then the region is divided again, and then steps B13, B14 and B15 are repeated until M new points are completed for topology connection, and finally, a directed graph is obtained, which is an incremental topology structure Directed graph represents the topology structure containing the incremental multi-agent formation, wherein + only indicates that the data has changed after the new agent of the formation, and the meaning itself does not change.

2. The method of claim 1, wherein, The communication topology structure of the multi-agent formation needs to meet the following conditions: Rigid graph condition: p ixy ≠ p jxy ,||p ixy (t)-p jxy (t)||=l, p ixy =(x i ,y i ) is the position sequence of the i-th agent, p jxy =(x j ,y j ) is the position sequence of the j-th agent, p ixy (t), p jxy (t) respectively represent the position sequence of the i-th, j-th agent at time t, and l is a constant; Minimum rigidity graph condition: Num(E) = 2Num(V) - 3, Num(·) represents the number; Persistent graph condition: d in (i) < 2 d in (i) denotes the in-degree of agent i; Leader in-degree restriction condition: d in (Leader) = 0 means that the in-degree of the leader is 0, the leader can only receive host information and cannot receive information from other agents, so its in-degree is 0.

3. The method of claim 1, wherein, The topology has a topological boundary edge set SE, a topological boundary point set SP, a convex hull edge set DE, and a convex hull point set DP.

4. The platoon generation method based on self-organizing incremental topology according to claim 1, characterized in that, The step one can define G = (V, E, A) as an undirected graph with N nodes, where V = 1, 2,..., N is a set of N nodes, R N×N represents the real number field, if then a ij = 1, otherwise a ij = 0.

5. The platoon generation method based on self-organizing incremental topology according to claim 1, characterized in that, The step one can also define a directed graph representing the topology of the formation containing N agents, represents that agent i in the formation sends information to agent j, if then otherwise

Images