A bio-inspired large-scale swarm networking and formation control method

Abstract

The application relates to a bio-inspired large-scale bee colony networking and formation control method, which uses a maximum-connectivity-based method to construct a hierarchical clustering network, constructs a route in and between clusters, fuses the hierarchical clustering network with grouping formation control, designs a message format, constructs formation groups by network clustering, realizes pigeon flock formation control in the groups, updates pigeon flock social grades by cluster heads, realizes European starling formation control between the groups, constructs a closed circle model, controls group distance by cluster heads, coordinates group speed by a leader, realizes European starling flock mode, obstacle avoidance mode and following mode at the group level. Compared with the prior art, the hierarchical clustering network architecture effectively reduces the routing overhead of a large-scale unmanned aerial vehicle cluster, and the pigeon flock formation control method and the European starling formation control method effectively reduce the control complexity and control communication overhead of the large-scale unmanned aerial vehicle cluster due to the distributed architecture.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of unmanned aerial vehicle cluster control, and particularly to a bio-inspired large-scale bee swarm networking and formation control method. BACKGROUND

[0002] OLSR (Optimized link state routing) as an active routing protocol, effectively improves the battlefield adaptability of unmanned aerial vehicle ad hoc network by actively maintaining the network topology. However, when the scale of the unmanned aerial vehicle cluster is further increased, the flooding overhead of TC (Topology control) messages will increase exponentially with the increase of the cluster scale, thereby exhausting the available bandwidth. The existing related research can reduce the topology flooding overhead of OLSR to different degrees, but most of the research only optimizes the routing topology flooding process of OLSR in a specific scenario. When the scale of the unmanned aerial vehicle cluster rises to a super large scale, the routing overhead is still huge, and the problems of computing overhead, storage overhead and communication hop count have not been solved, so the optimization of the expansion performance is relatively limited. HOLSR (Hierarchical optimized link state routing) changes the defect of the insufficient expansion performance of the flat routing by limiting the transmission of TC messages through the hierarchical clustering method, and can truly solve the scalability problem of the OLSR routing protocol, and has obvious potential in the design of large-scale unmanned aerial vehicle cluster network architecture.

[0003] The distributed control method points out the direction for the control of large-scale UAV cluster, and the bionic method provides the design idea for the distributed control. Since the bird flight has many similarities with the UAV formation control, the bird flight provides many bionic algorithms for the UAV formation control, among which, the formation control of bionic pigeon and the formation control of bionic European starling have attracted the attention of researchers. The control algorithm of bionic pigeon divides the UAV cluster into social grades, adopts the method of navigation, and realizes the generation and maintenance of close formation. However, when the scale of the formation increases, the following error will be accumulated layer by layer, and the accumulated error will greatly reduce the effect of topology maintenance and following effect of the formation. At the same time, the formation structure of the pigeon formation is relatively fixed, which is too single for the large-scale UAV cluster formation and is not suitable for the complex battlefield environment. The formation control algorithm of bionic European starling increases the phase change control of the UAV, simulates the speed following mechanism of European starling, and improves the maneuvering ability of the cluster such as obstacle avoidance and turning. The formation control algorithm of bionic European starling only focuses on the change of individual motion, so the connection relationship of the individual and the topology structure of the cluster will change with the change of the relative position between the individuals, which cannot form a relatively stable topology connection relationship, and thus leads to the generation of invalid routing and increases the routing maintenance overhead. At the same time, the completely distributed control mode of European starling formation is difficult to adapt to complex task planning.

[0004] Therefore, the existing method has the problems of insufficient adaptability of the HOLSR protocol and the formation control method, and insufficient expansibility of the pigeon formation control method and the European starling formation control method. SUMMARY

[0005] The purpose of the present application is to provide a biological inspired large-scale bee group networking and formation control method, which reduces the routing overhead and network maintenance overhead through the grouping and clustering architecture; avoids the conflict between the network grouping architecture and the UAV grouping formation control by combining the network clustering architecture with the UAV grouping formation, and effectively improves the stability of the network architecture; effectively reduces the communication overhead of the formation control through the distributed control architecture; and at the same time, considering the influence of the network communication architecture on the formation control, realizing the bionic pigeon formation control in the group and the bionic European starling formation control between groups, effectively improving the stability and maneuverability of the large-scale UAV cluster formation control.

[0006] The purpose of the present application can be realized by the following technical solutions:

[0007] A bio-inspired large-scale swarm networking and formation control method, which uses a maximum-connectivity-based method to construct a hierarchical clustering network and provides a network maintenance method to construct a route within and between clusters; the hierarchical clustering network is integrated with grouping formation control, and a message format is comprehensively designed; the hierarchical clustering network is used to construct a formation group, a pigeon-like formation control is realized within the group, the cluster head is used to update the pigeon social hierarchy, a distributed interaction rule is constructed, a European starling-like formation control is realized between groups, a closed circle model is constructed, the cluster head is used to control the distance between groups, and the leader is used to coordinate the speed between groups, realizing the European starling swarm mode, obstacle avoidance mode and following mode at the group level.

[0008] The method comprises the following steps:

[0009] Step 1) jointly designing the frame structure of the message according to the information required by the network architecture, route maintenance and formation control;

[0010] Step 2) obtaining the connectivity information, position information and cluster information of the neighbors within the two-hop range of the unmanned aerial vehicle through the interaction of neighbor messages, and electing the unmanned aerial vehicle with the maximum connectivity as the cluster head, the cluster head flooding the cluster management message to form a cluster, and completing the clustering of all unmanned aerial vehicles through multiple rounds of election;

[0011] Step 3) constructing a network maintenance mechanism for different network architecture failure conditions to adjust the network architecture to restore to a stable hierarchical clustering state;

[0012] Step 4) constructing an active route within and between clusters, limiting the route topology control message within the cluster, and the cluster head unmanned aerial vehicle flooding the route topology control message in the second layer of the network and broadcasting the member information of the cluster;

[0013] Step 5) using network clustering as the basis of formation grouping to construct a multi-formation group control architecture to realize the formation control of unmanned aerial vehicles, wherein the network clustering and the formation grouping are consistent:

[0014] Step 51) constructing a pigeon-like formation control mechanism within the group, using the cluster head to periodically manage the social hierarchy, establishing a pigeon-like neighbor discovery model and neighbor interaction model, and realizing close following flight;

[0015] Step 52) is equivalent to each UAV group using a closed circle model, and defines the equivalent position, equivalent radius and equivalent speed; based on the closed circle model, a European starling formation control mechanism is constructed between groups, the position information of neighbor groups is obtained through information interaction between cluster heads, the relative position control of the whole group and other groups is calculated and flooded by the cluster head, and the speed of the neighbor group is adjusted by the leader through listening; the cluster mode, obstacle avoidance mode and following mode of the European starling are mapped into the group formation control, and the collision avoidance and obstacle avoidance between UAV groups are realized.

[0016] The neighbor message includes the position, speed and cluster head sequence number information of the neighbor, each UAV interacts with the neighbor UAV the position, speed, cluster head sequence number and connectivity information of the local machine, and interacts the position, cluster head sequence number and connectivity information of one-hop neighbor.

[0017] The step 3) is specifically: for member UAV failure, cluster head UAV failure, new UAV joining, partial cluster overlap, cluster merging, respectively design network maintenance mechanism, determine the behavior of cluster head and member in different failure scenarios, wherein,

[0018] For the member UAV failure scenario, the cluster head queries the two-hop neighbor table, records the failed member UAV, and the member has no action;

[0019] For the cluster head UAV failure scenario, the cluster head has no action, and the member enters the unclustered state after waiting for a preset time;

[0020] For the new UAV joining scenario, the cluster head queries the two-hop neighbor table, records the new member UAV, and the member replaces the cluster head ID of the local UAV with the ID of the new cluster head;

[0021] For the partial cluster overlap scenario, the cluster head has no action, and the member joins the new cluster with fewer hops, and if there is no new cluster with fewer hops, it remains unchanged;

[0022] For the cluster merging scenario, the cluster head with smaller connectivity or the same connectivity but smaller ID gives up the cluster head position, and the member has no action.

[0023] In the step 51), the pigeon group neighbor discovery model is specifically: the UAV distinguishes the position of the neighbor according to the repulsion radius R Rep , the alignment radius R Al and the attraction radius R Att , determines the neighbor within the preset radius range with the UAV as the center, wherein the UAV in the blind area of the vision of the UAV is not selected as the neighbor, and the neighbor discovery model has a maximum number limit.

[0024] In the step 51), the pigeon group neighbor interaction model is specifically:

[0025] If node i and node j are in the same cluster, and node j is in the perception set N i of node i, the number of nodes in N i has a maximum limit, and there is an attraction or repulsion between nodes, as follows:

[0026]

[0027] Wherein, the node is a UAV, r ij is the distance between node i and node j, is the unit vector from node i to node j, R Rep is the repulsion radius, R Al is the alignment radius, R Att is the attraction radius, is the force between nodes, the superscript indicates that the force is used to control the relative position between UAVs, h i is the social rank of node i, A is a constant, A > 0; when 0 ≤ r ij < R Rep , the nodes produce repulsion, when R Al ≤ r ij < R Att , the nodes produce attraction;

[0028] If node j in N i and node i do not belong to the same cluster, in order to avoid mutual collision between nodes, the neighbor nodes of different clusters do not produce attraction, only produce repulsion, and the force function is as follows:

[0029]

[0030] In order to ensure the tightness of the formation, the nearest superior pigeon produces an additional following force to make up for the problem of insufficient attraction when the number of high-level individuals is small, and the following force is as follows:

[0031]

[0032] Wherein, represents the position vector of the node i, is the position vector of the neighbor node j, h i is the social rank of node i, and node j is the node closest to node i in the set N i

[0033] The UAV sets its own speed to the average value of the higher social rank individuals, as follows:

[0034]

[0035] ​wherein, is the velocity vector of the UAV i, H i is the set of neighbor UAVs in the perception range of the UAV i whose social rank is higher than the UAV, |H i | is the number of UAVs in the set.

[0036] In the step 52), the closed circle model groups the UAVs into a whole, so as to control the UAV groups as a single European starling, wherein the center of the closed circle model is the average position of the UAVs in the group, the radius of the closed circle is the distance from the farthest UAV to the center, and the velocity of the closed circle is the velocity of the leader.

[0037] In the step 52), the European starling-like neighbor discovery model is established between the groups, and the center, radius and velocity information of the equivalent closed circle of the neighbor group are exchanged between the cluster heads, and the European starling-like neighbor discovery model is specifically as follows: each UAV group has 5 fan-shaped perception areas, each perception area corresponds to a perception angle of 45°, and the remaining area is a blind area, and there is only one neighbor group in each perception area, and the radial difference between the UAV group i and the UAV group j is D ij , different areas are divided according to the size of the radial difference, if the radial difference is less than the repulsion distance D Rep , it is located in the repulsion area, if the radial difference is less than the alignment distance D Al and greater than the repulsion distance D Rep , it is located in the alignment area, if the radial difference is less than the attraction distance D Att and greater than the alignment distance D Al , it is located in the attraction area.

[0038] In the step 52), the action following sequence number information is added in the routing message and the cluster management message, when each UAV group informs the neighbor of its own velocity, the action following sequence number identifies the change of itself, the action following sequence number is increased with the increase of the interaction message, and the leader saves the action following sequence number of the UAV group that is followed, and whether the action of the UAV group is followed is judged according to the size of the action following sequence number.

[0039] In the step 52), the cluster mode, obstacle avoidance mode and following mode of the European starling are mapped into the group formation control, and specifically:

[0040] For the cluster mode, the action modes in the repulsion area and the attraction area are as follows:

[0041]

[0042] wherein, the group number of the present group is i, the group number of the neighbor group is j, is the unit vector from the group i to the group j, and Dij N represents the radial difference between group i and group j. i Let D be the set of neighboring groups of group i, when D ij <D Rep At that time, the groups are subjected to a repulsive force, when D Al ≤D ij <D Att At that time, an attractive force exists between the groups, and the radial difference will eventually converge at D. Rep With D Al between;

[0043] For neighbors located in the alignment region, the navigator adjusts its own speed based on the speed of neighbor grouping, as shown in the following formula:

[0044]

[0045] in, For the speed of group i, Let N be the velocity of group j. Al This indicates that group i is a neighbor of the alignment region, and v0 is the cruising speed, that is, the speed of group i is the average speed of the neighbors in the alignment region;

[0046] In obstacle avoidance mode, the navigator automatically changes its flight direction to avoid obstacles.

[0047] In follow mode, the navigator follows the neighbor with the largest speed change to avoid obstacles in advance. Specifically, it first determines whether to follow the group based on the action follow sequence number of the neighbor group. If it does follow, it makes the following judgment:

[0048]

[0049] Among them, A j (n) represents the velocity change index of neighbor group j. If the velocity of the neighbor group is a velocity that has not been imitated before, that is, the current velocity of the neighbor group. Not equal to the velocity of this group i a number of times m before. The velocity change of the neighbor is measured by comparing it with the velocity change at the previous time step; otherwise, it is considered that the neighbor cannot follow, and A is set to [a different value]. j (n) is set to 0;

[0050] In follow mode, only one neighbor group is followed by this group. Therefore, the node with the largest change in velocity among the neighbor groups is selected. Assuming the index of this group is I(n), its expression is as follows:

[0051]

[0052] Establish a speed switching threshold to evaluate the magnitude of group speed changes, wherein the speed switching threshold Ti is expressed as:

[0053]

[0054] wherein, delta is a coefficient, N i is a neighbor set, is the number of followable neighbors in the neighbor set, gamma i reacts to the average speed change degree of the current followable neighbors;

[0055] if I(n) corresponds to the group number k, if I(n)>T i then the leader enters the following mode, and the group i imitates the speed of the group k; if I(n)≤T i then the leader enters the cluster mode, and the group i does not imitate the speed of the group k, and maintains the speed consistency of the group and the neighbor group, as shown in the following formula:

[0056]

[0057] wherein, v0 is the cruise speed of the unmanned aerial vehicle.

[0058] Compared with the prior art, the present application has the following beneficial effects:

[0059] (1) The present application unifies the clustering of the network and the grouping of the formation, designs a fusion message structure, avoids the conflict between the network structure and the formation structure, and reduces the additional communication overhead.

[0060] (2) The present application designs a formation control method based on the existing network architecture and routing information, realizes part of the control function of the leader and the cluster head respectively, avoids the generation of the deadlock phenomenon, and improves the control effect.

[0061] (3) The present application realizes the pigeon group leader flight in the group, designs a pigeon group social class dynamic maintenance mechanism, and realizes the leader dense flight in the high-mobility environment.

[0062] (4) The present application combines the European starling formation control method and the large-scale formation control, effectively improves the obstacle avoidance and collision avoidance of the large-scale unmanned aerial vehicle cluster.

[0063] (5) The present application is designed for the large-scale unmanned aerial vehicle cluster, considers the mutual influence of the network communication and the formation control, and has strong engineering value. BRIEF DESCRIPTION OF DRAWINGS

[0064] Figure 1 is the flow chart of the method of the present application;

[0065] Figure 2 is the schematic diagram of the design points of the present application;

[0066] Figure 3 A layered clustering network architecture diagram in an embodiment of the present application;

[0067] Figure 4 A pigeon social hierarchy diagram in an embodiment of the present application;

[0068] Figure 5 A pigeon flock neighbor discovery model diagram in an embodiment of the present application;

[0069] Figure 6 An equivalent circle model diagram in an embodiment of the present application;

[0070] Figure 7 A European starling neighbor discovery model diagram in an embodiment of the present application;

[0071] Figure 8 A routing overhead comparison diagram in an embodiment of the present application;

[0072] Figure 9 A communication overhead comparison diagram in an embodiment of the present application;

[0073] Figure 10 A BINC obstacle avoidance effect diagram without following mechanism in an embodiment of the present application;

[0074] Figure 11 A BINC obstacle avoidance effect diagram with following mechanism in an embodiment of the present application;

[0075] Figure 12 A turning radius comparison diagram in an embodiment of the present application;

[0076] Figure 13 A clustering switching times comparison diagram of different models in an embodiment of the present application;

[0077] Figure 14 A clustering switching times comparison diagram in a straight-line navigation scenario in an embodiment of the present application;

[0078] Figure 15 A clustering switching times comparison diagram in a turning obstacle avoidance scenario in an embodiment of the present application; DETAILED DESCRIPTION

[0079] The present application will be described in detail below with reference to the drawings and specific embodiments. The present embodiment is implemented on the premise of the technical solution of the present application, and gives a detailed implementation and specific operation process, but the protection scope of the present application is not limited to the following embodiments.

[0080] Embodiment 1

[0081] The embodiment provides a bio-inspired large-scale bee colony networking and formation control method, which has low routing overhead, formation control overhead and network maintenance overhead, uses a maximum connectivity-based method to construct a hierarchical clustering network, and provides a network maintenance method to construct a route within a cluster and between clusters; the hierarchical clustering network is fused with group formation control, and a message format is comprehensively designed; the hierarchical clustering network is used to construct a formation group, pigeon flock formation control is realized in the group, a cluster head is used to update a pigeon flock social rank, a distributed interaction rule is constructed, European starling formation control is realized between groups, a closed circle model is constructed, a cluster head is used to control a distance between groups, and a leader is used to coordinate a speed between groups, and a European starling swarm mode, an obstacle avoidance mode and a following mode are realized at a group level.

[0082] Specifically, as shown in Figure 1 and Figure 2 , the method comprises the following steps:

[0083] Step 1) According to the information required by the network architecture, route maintenance and formation control, the frame structure of the message is jointly designed.

[0084] Step 2) Through the interaction of the neighbor message, the connectivity information, position information and cluster information of the neighbors within the two-hop range of the unmanned aerial vehicle are obtained, and the unmanned aerial vehicle with the maximum connectivity is elected as the cluster head; the cluster head floods the cluster management message, forms a cluster, and completes the clustering of all unmanned aerial vehicles through multiple rounds of election.

[0085] The neighbor message comprises the position, speed and cluster head serial number information of the neighbor, and the unmanned aerial vehicle interacts with the neighbor unmanned aerial vehicle through the HELLO message to interact the position, speed, cluster head serial number and connectivity information of the local machine, and the position, cluster head serial number and connectivity information of the one-hop neighbor, so that each unmanned aerial vehicle obtains the connectivity, cluster head serial number and position information of the neighbors within the two-hop range. In the embodiment, the connectivity of a single unmanned aerial vehicle node in the swarm refers to the number of neighbor nodes directly connected to the node. That is, how many other unmanned aerial vehicle nodes can be directly communicated with by the unmanned aerial vehicle node through the wireless communication link, and the number of the neighbor nodes capable of direct communication is the connectivity of the node.

[0086] According to the connectivity information of the neighbors within the two-hop range, a clustering structure is constructed using the maximum connectivity criterion, and the clustering network structure is as shown in Figure 3 .

[0087] In the embodiment, the unmanned aerial vehicle has two communication channels, one responsible for long-distance inter-cluster communication and one responsible for short-distance intra-cluster communication. The network management message of the cluster head fuses the position transformation information, social rank and cluster head serial number information required by the formation movement control.

[0088] Step 3) Construct network maintenance mechanism for different network architecture failure cases, adjust network architecture to restore to stable hierarchical clustering state.

[0089] Design network maintenance mechanism for member UAV failure, cluster head UAV failure, new UAV joining, partial clustering overlap, clustering merging, respectively, determine the behavior of cluster head and member in different failure scenarios, as shown in Table 1.

[0090] Table 1

[0091]

[0092] Data packets in the same cluster are transmitted by direct forwarding, and data packets in different clusters are transmitted by default through the cluster head.

[0093] Step 4) Construct active routing within and between clusters, limit routing topology control messages within the cluster, and broadcast member information of the cluster.

[0094] In the network architecture of Figure 3 , limit routing topology messages within the cluster, and flood member UAV addresses between cluster heads. When member UAV 1 sends a data packet to member UAV 4, the addressing process of the data packet is 1→A1→3→4. When UAV 1 sends a data packet to UAV 14, the addressing process of the data packet is 1→A1→A2→B2→C2→D2→D1→14.

[0095] Step 5) Construct multi-formation grouping control architecture based on network clustering as the basis for formation grouping, to achieve formation control of UAVs, wherein network clustering and formation grouping have consistency.

[0096] Formation control grouping is based on network clustering, and member nodes generate attraction according to the position of the cluster head to aggregate and form a group.

[0097] Step 51) Construct pigeon flock formation control mechanism within the group, use cluster head to periodically manage social rank, establish pigeon flock neighbor discovery model and neighbor interaction model, and achieve close following flight.

[0098] Construct pigeon flock formation control mechanism within the group, including periodic distribution of social rank, neighbor discovery model, and pigeon flock interaction rules, the specific design is as follows:

[0099] (1) The cluster head periodically calculates the social rank according to the position of the member node and distributes it.

[0100] Specifically, the cluster head determines the social rank according to the advancing direction of the unmanned aerial vehicle group, and the higher the social rank, the lower the rank number, and the number of unmanned aerial vehicles in each level is shown in the following formula:

[0101]

[0102] Wherein n is the social rank number, K(n) is the number of unmanned aerial vehicles corresponding to the social rank number, and k is a constant. When n=0, the unmanned aerial vehicle is selected as the leader. The pigeon social rank structure is shown in Figure 4 .

[0103] (2) The pigeon-like neighbor discovery model is shown in Figure 5 . Each unmanned aerial vehicle has a limited number of neighbors, and there is a blind area. The relative distance between the neighbor unmanned aerial vehicle and other unmanned aerial vehicles is r, and different interaction behaviors are generated around the unmanned aerial vehicle according to the repulsion radius R Rep , the alignment radius R Al , and the attraction radius R Att .

[0104] The unmanned aerial vehicle distinguishes the position of the neighbor according to the repulsion radius R Rep , the alignment radius R Al , and the attraction radius R Att , and determines the neighbor within the preset radius range with the unmanned aerial vehicle as the center. The unmanned aerial vehicle in the blind area of the vision of the unmanned aerial vehicle is not selected as the neighbor, and the neighbor discovery model has a maximum number limit. In this embodiment, the maximum number of neighbors of the unmanned aerial vehicle is 7.

[0105] (3) The pigeon neighbor interaction rule includes repulsion behavior, speed alignment behavior, attraction behavior, and close following behavior, as shown in the following table:

[0106] Table 2

[0107] Relative distance of drone to neighbours Drone behaviour 0 < r < R Rep ]] Drone subject to repulsion [R Rep ≤r<R Al ]] Drone alignment speed [R Al ≤r<R Att ]] Drone subject to attraction No Drone subject to attraction by nearest high ranking drone

[0108] If node i and node j are in the same cluster, and node j is in the perception set N i of node i, the number of nodes in N i is not more than 7, and there is an attraction or repulsion between the nodes, as shown below:

[0109]

[0110] Wherein the node is the unmanned aerial vehicle, r ij is the distance between node i and node j, is the unit vector from node i to node j, R Rep is the repulsion radius, R Al is the alignment radius, and R Attis the attractive radius, is the force between nodes, the superscript indicates that the force is used to control the relative position between UAVs, h i is the social rank of node i, A is a constant, A>0; when 0≤r ij <R Rep , the nodes produce repulsive force, when R Al ≤r ij <R Att , the nodes produce attractive force;

[0111] If node j in N i and node i do not belong to the same cluster, in order to avoid mutual collision between nodes, the neighbor nodes of different clusters do not produce attractive action, only produce repulsive action, the force function is shown as follows:

[0112]

[0113] In order to ensure the tightness of the formation, the nearest superior pigeon produces additional following force to make up for the problem of insufficient attractive force when the number of high-level individuals is small, and the following force is shown as follows:

[0114]

[0115] wherein, represents the position vector of the node i, is the position vector of the neighbor node j, h i is the social rank of node i, and node j is the node closest to node i in the set N i .

[0116] The UAV sets its own speed to the average value of the higher individual, so as to achieve the consistency of the entire cluster in the path planning, which is shown as follows:

[0117]

[0118] wherein, is the speed vector of UAV i, H i is the set of neighbor UAVs whose social rank is higher than the UAV in the perception range of the UAV, |H i | is the number of UAVs in the set.

[0119] The UAV follows the motion information of the nearest high-level UAV.

[0120] Step 52) using a closed circle model to equivalently each UAV group, and defining the equivalent position, equivalent radius and equivalent speed; based on the closed circle model, building a mimic European starling formation control mechanism between groups, obtaining the position information of neighbor groups through information interaction between cluster heads, controlling the relative position of the whole group and other groups through cluster head calculation and message flooding, and adjusting the speed of the leader through listening to the speed information of neighbor groups; mapping the cluster mode, obstacle avoidance mode and following mode of European starling to the group formation control, realizing collision avoidance and obstacle avoidance between UAV groups.

[0121] Specifically, the mimic European starling formation flight mechanism is built in the cluster, including a closed circle equivalent model, a mimic European starling neighbor discovery model, and an interaction rule in different modes.

[0122] (1) The center of the closed circle is the geometric center of the UAV group, the radius is the distance from the geometric center of the farthest UAV, and the speed is the speed of the leader.

[0123] As shown in Figure 6 , the closed circle model equivalently regards the UAV group as a whole, so that the UAV group is controlled as a single European starling, wherein the center of the closed circle model is the average position of the UAVs in the group, the radius R of the closed circle is the distance from the farthest UAV to the center, and the speed of the closed circle is the speed of the leader.

[0124] (2) A mimic European starling neighbor discovery model is established between groups, as shown in Figure 7 , the equivalent closed circle center, radius and speed information of neighbor groups are interacted between cluster heads. Wherein, each UAV group has 5 fan-shaped sensing areas, each sensing area corresponds to a sensing angle of 45°, and the remaining area is a blind area. There is only one neighbor group in each sensing area. The radial difference between UAV group i and UAV group j is D ij . According to the size of the radial difference, different areas are divided. If the radial difference is less than the repulsion distance D Rep , it is located in the repulsion area. If the radial difference is less than the alignment distance D Al and greater than the repulsion distance D Rep , it is located in the alignment area. If the radial difference is less than the attraction distance D Att and greater than the alignment distance D Al , it is located in the attraction area. Only one group can be selected in each sensing area, as shown in Figure 7 , in group m and group n, the group m with shorter distance is selected as the neighbor of this area.

[0125] This step realizes the separation of formation control, the leader controls the speed and direction of the flight path planning, and the cluster head controls the relative position of the neighbor group. The inter-group formation control can maintain the relative stable group distance, thereby avoiding the network structure conflict between the clusters. The leader can autonomously plan the flight path, thereby realizing obstacle avoidance.

[0126] (3) The neighbor interaction rules of the European starling imitation are determined by the cluster mode, obstacle avoidance mode and following mode, as shown in Table 3.

[0127] Table 3

[0128]

[0129] In the routing message and the cluster management message, the action following sequence number information is added, each UAV group informs the neighbor of its own speed, the action following sequence number is added to identify the change of itself, the action following sequence number is increased with the increase of the interaction message, the leader saves the action following sequence number of the UAV group received, and whether the action of the UAV group is followed is judged according to the size of the action following sequence number.

[0130] For the cluster mode, the action mode of the repulsion region and the attraction region is as follows:

[0131]

[0132] wherein, the group number of the present group is i, the group number of the neighbor group is j, is the unit vector from group i to group j, D ij represents the radial difference between group i and group j, N i is the neighbor group set of group i, when D ij <D Rep , the groups are subjected to repulsive force, when D Al ≤D ij <D Att , the groups generate attractive force, and the radial difference will finally converge between D Rep and D Al ;

[0133] For the neighbor located in the alignment region, the leader adjusts its own speed according to the speed of the neighbor group, as shown in the following formula:

[0134]

[0135] wherein, is the speed of group i, is the speed of group j, N Al represents that group i is in the alignment region neighbor, v0 is the cruise speed, that is, the speed of group i is the average speed of the alignment region neighbor.

[0136] In obstacle avoidance mode, the navigator automatically changes its flight direction to avoid obstacles.

[0137] In follow mode, the navigator follows the neighbor with the largest speed change to avoid obstacles in advance. First, it determines whether to follow the group based on the action follow sequence number of the neighbor group. If following is necessary, the following judgment is made:

[0138]

[0139] Among them, A j (n) represents the velocity change index of neighbor group j. If the velocity of the neighbor group is a velocity that has not been imitated before, that is, the current velocity of the neighbor group. Not equal to the velocity of this group i a number of times m before. The velocity change of the neighbor is measured by comparing it with the velocity change at the previous time step; otherwise, it is considered that the neighbor cannot follow, and A is set to [a different value]. j (n) is set to 0. Where D ij The value of the radial difference between neighboring groups is used to normalize the index size.

[0140] In follow mode, only one neighbor group is followed by this group. Therefore, the node with the largest change in velocity among the neighbor groups is selected. Assuming the index of this group is I(n), its expression is as follows:

[0141]

[0142] A speed switching threshold is established to evaluate the magnitude of group speed changes. When speed consistency among multiple clusters is poor, arbitrary speed following will lead to difficulty in cluster speed convergence. A higher threshold should be set to keep clusters in replication mode, resulting in consistent speeds. When speed consistency among multiple clusters is good, a lower threshold should be set. In this embodiment, the speed switching threshold T... i Represented as:

[0143]

[0144] Where δ is the coefficient, N i It is a collection of neighbors. γ is the number of followable neighbors in the neighbor set. i The reaction reflects the degree to which the average speed of the current neighbor changes.

[0145] Let k be the group number corresponding to I(n). If I(n) > T i Then the navigator enters follow mode, and group i imitates the speed of group k; if I(n)≤T iThen the leader enters cluster mode, and group i does not imitate the speed of group k, but maintains the speed consistency between the group and its neighboring groups, as shown in the following formula:

[0146]

[0147] Where v0 is the drone's cruising speed.

[0148] Example 2

[0149] This embodiment, based on the method of Embodiment 1, specifically provides a bio-inspired large-scale UAV swarm network architecture and formation control integrated design method, abbreviated as BINC, which includes the following steps:

[0150] S1, at 30000×30000m 2 Within the range, at 40 aircraft / km 2 Simulations were performed on the density of drones, with simulated node numbers of 520, 640, 760, 880, and 1000. Each drone had two channels: one for long-range inter-packet communication and one for short-range intra-packet communication. The large-scale drone swarm was initially randomly distributed within a certain area.

[0151] S2, add the node's connectivity, location, cluster head number, speed, and follower number information to the HELLO message to obtain the connectivity, location, cluster head number, speed, and follower number information of the one-hop neighbor; add connectivity, location, and cluster head number to the one-hop neighbor table to obtain the connectivity, location, and cluster head number information of the two-hop neighbor.

[0152] In S3, large-scale drone swarms cluster together through multiple rounds of elections, using connectivity information exchanged within two hops. Each cluster's TC (Traffic Control) messages are confined within its own cluster, significantly reducing routing overhead. As the drone swarm size increases, the routing overhead consistently decreases. The communication overhead required for distributed formation control depends only on local conditions. With increasing drone swarm size, both intra-group and inter-group formation control overhead remain small and stable, resulting in superior scalability. Figure 8 By comparing the routing overhead at different drone swarm sizes, it can be seen that this solution significantly reduces routing overhead compared to solutions such as OLSR and CLSR. Figure 9 This demonstrates how routing overhead and motion control overhead change as the size of the drone swarm varies.

[0153] S4, add social class field, group motion field, cluster head serial number field in the cluster management message, so as to realize network management and group motion control. In the HELLO message of the second layer of the network, add group position, group speed, group radius, cluster head serial number field, and realize information interaction of neighbor groups within one hop.

[0154] S5, large-scale UAV cluster flies in straight-line navigation scene and turning obstacle avoidance scene respectively. The size of the fixed UAV cluster is compared with the obstacle avoidance effect of the UAV group with and without the European starling following mechanism in the turning obstacle avoidance scene. The leader of the group with the European starling following mechanism follows according to the following serial number and the speed of the neighbor group, so as to realize early obstacle avoidance. When the size of the fixed UAV cluster is fixed, the distance between the UAV groups is maintained by the formation control between the European starling groups, and the density of the UAV groups is maintained by the pigeon group formation control within the group. Figure 10 、 Figure 11 It is shown that the turning radius of the cluster without the European starling following mode is small, and there is a phenomenon of hitting the obstacle. The cluster with the European starling following mode does not have the above problems. Figure 12 It is shown whether the obstacle avoidance distance with the European starling following mechanism.

[0155] S6, large-scale UAV cluster flies in straight-line navigation scene and turning obstacle avoidance scene respectively. The UAVs may switch between different groups, but the inter-group distance control of BINC will limit the distance between the network clusters, and reduce the occurrence of cluster switching. For different inter-group distance constraints, gradually increase the size of the UAV cluster, and the smaller inter-group distance constraint will lead to the occurrence of UAV group switching, and the larger inter-group distance constraint can reduce the occurrence of cluster switching. Figure 13 The number of cluster switching times of BINC and Boids model in different scenes and different cluster sizes is counted in Figs. 6 and 7. It can be seen that BINC can effectively reduce the number of cluster switching times in straight-line navigation scene and turning obstacle avoidance scene, and maintain the stability of the network. Figure 14 and Figure 15 The number of cluster switching times in the straight-line navigation scene and the turning obstacle avoidance scene is compared in Figs. 6 and 7 respectively. It can be seen that the larger the inter-group distance constraint, the better the network maintenance effect.

[0156] The above detailed description of the preferred embodiments of the present application. It should be understood that those skilled in the art can make many modifications and changes without creative labor according to the concept of the present application. Therefore, any technical solution obtained by logical analysis, reasoning or limited experiment on the basis of the prior art according to the concept of the present application shall be within the protection scope determined by the claims.

Claims

1. A bio-inspired method for large-scale bee colony networking and formation control, characterized in that, This method constructs a hierarchical clustered network using a maximum connectivity-based approach and provides network maintenance methods, establishing routes within and between clusters. It integrates the hierarchical clustered network with group formation control and comprehensively designs message formats. It constructs group formation based on network clusters, implementing pigeon flock-like group formation control within groups, using cluster heads to update the pigeon flock's social hierarchy, and constructing distributed interaction rules. It implements starling-like group formation control between groups, constructing a closed circle model, using cluster heads to control the distance between groups, and using navigators to coordinate the speed between groups. At the group level, it implements starling flock mode, obstacle avoidance mode, and following mode. The method includes the following steps: Step 1) Based on the information required for network architecture, routing maintenance, and formation control, jointly design the message frame structure; Step 2) By interacting with neighbor messages, obtain the connectivity information, location information and cluster information of the neighbors within two hops of the drone, and elect the drone with the highest connectivity as the cluster head. The cluster head floods the cluster management message to form a cluster. After multiple rounds of election, the clustering of all drones is completed. Step 3) Build a network maintenance mechanism for different network architecture failure scenarios, and adjust the network architecture to restore it to a stable hierarchical and clustered state; Step 4) Construct active routing within and between clusters, restricting the routing topology control messages within the cluster to the cluster itself. The cluster head UAV floods the routing topology control messages in the second layer of the network and broadcasts the cluster member information. Step 5) Using network clustering as the basis for formation grouping, construct a multi-formation grouping control architecture to realize UAV formation control. Network clustering and formation grouping must be consistent. Step 51) Construct a pigeon flock formation control mechanism within the group, use the flock leader to periodically manage the social hierarchy, establish a pigeon flock neighbor discovery model and a neighbor interaction model, and achieve close following flight; Step 52) Use a closed circle model to perform equivalent analysis on each UAV group, and define the equivalent position, equivalent radius, and equivalent speed. Based on the closed circle model, construct a starling-like formation control mechanism between groups. Obtain the position information of neighboring groups through information interaction between cluster heads. Control the relative position of the entire group with other groups by calculating and flooding messages through the cluster heads. The navigator adjusts its own speed by listening to the speed information of neighboring groups. Map the starling's swarm mode, obstacle avoidance mode, and follow mode to the group formation control to achieve collision avoidance and obstacle avoidance between UAV groups.

2. The bio-inspired large-scale bee colony networking and formation control method according to claim 1, characterized in that, The neighbor message includes the neighbor's location, speed, and cluster head number information. Each UAV exchanges its own location, speed, cluster head number, and connectivity information with neighboring UAVs, and also exchanges the location, cluster head number, and connectivity information of one-hop neighbors.

3. The bio-inspired large-scale bee colony networking and formation control method according to claim 1, characterized in that, Step 3) specifically involves designing network maintenance mechanisms for member drone failures, cluster head drone failures, new drone additions, partial cluster overlap, and cluster merging, respectively, and determining the behavior of the cluster head and members under different failure scenarios. In the case of a member drone failure, the cluster head queries the two-hop neighbor table to record the failed member drone, and the member has no action. In the scenario where the cluster leader drone fails, the cluster leader does not move, and the members wait for a preset time before entering the un-clustered state. When a new drone joins the scene, the cluster head queries the two-hop neighbor table to record the newly joined member drone, and the member replaces its own cluster head ID with the new cluster head ID; In scenarios where some clusters overlap, the cluster head remains inactive, and members join a new cluster with fewer hops. If no new cluster with fewer hops exists, the process remains unchanged. In cluster merging scenarios, the cluster head with lower connectivity or the same connectivity but a smaller ID relinquishes its cluster head status, and its members take no action.

4. The bio-inspired large-scale bee colony networking and formation control method according to claim 1, characterized in that, In step 51), the simulated pigeon flock neighbor detection model specifically involves the drone determining the repulsion radius. Alignment radius and attraction radius The location of neighbors is distinguished, and neighbors within a preset radius centered on the drone are identified. Drones located in the drone's blind spot are not selected as neighbors, and there is a maximum limit to the number of neighbors that can be detected.

5. The bio-inspired large-scale bee colony networking and formation control method according to claim 1, characterized in that, In step 51), the simulated pigeon flock neighbor interaction model is specifically as follows: If node With nodes Within the same cluster, and nodes At the node Perceptual set middle, There is a maximum limit to the number of nodes in the system, and nodes can attract or repel each other, as shown below: The node is the drone. For nodes With nodes The distance between them It is a node To the node unit vector, For the repulsion radius, To align the radius, For the radius of attraction, This represents the force between nodes; the superscript indicates that the force is used to control the relative positions of the drones. For nodes social hierarchy, It is a constant. ;when At this time, a repulsive force is generated between the nodes. At that time, attraction is generated between nodes; like Nodes in With nodes If nodes do not belong to the same cluster, then to avoid collisions between nodes, neighboring nodes from different clusters do not attract each other, but only repel each other. The force function is shown in the following equation: To ensure the tightness of the formation, the closest senior pigeons generate an additional following force to compensate for the insufficient attraction when there are few senior individuals. The following force is shown in the following formula: in, Indicates this node The position vector, Neighboring nodes The position vector, For nodes Social hierarchy, nodes For set Mid-distance nodes The nearest node; The drone sets its speed to the average speed of individuals with higher social ranks, as shown in the following formula: in, For drones velocity vector For drones The set of neighboring drones within the perception range that have a higher social status than the drone in question. This represents the number of drones in the set.

6. The bio-inspired large-scale bee colony networking and formation control method according to claim 1, characterized in that, In step 52), the closed circle model treats the drone group as an equivalent whole, thereby controlling the drone group as a single starling. The center of the closed circle model is the average position of the drones in the group, the radius of the closed circle is the distance from the center of the drone farthest from the center, and the speed of the closed circle is the speed of the navigator.

7. The bio-inspired large-scale bee colony networking and formation control method according to claim 1, characterized in that, In step 52), a simulated starling neighbor detection model is established between groups. This model uses the center, radius, and velocity information of the equivalent closed circle of neighbor groups interacting between cluster heads. Specifically, each UAV group has 5 fan-shaped sensing regions, and the sensing angle corresponding to each sensing region is... The remaining area is a blind spot; only one neighbor group exists in each sensing area, and the drone group... Grouping with drones The radial difference is Different regions are divided based on the magnitude of the radial difference; if the radial difference is less than the repulsion distance... If the radial difference is less than the alignment distance, then it is located in the repulsion region. And greater than the repulsion distance If the radial difference is less than the attraction distance, then it is located in the alignment region. And greater than the alignment distance Then it is located in the attraction area.

8. The bio-inspired large-scale bee colony networking and formation control method according to claim 1, characterized in that, In step 52), action follow sequence number information is added to the routing message and cluster management message. When each drone group informs its neighbor of its speed, it adds an action follow sequence number to identify its own changes. The action follow sequence number increases with the increase of interaction messages. The navigator saves the action follow sequence number of the drone group it receives and determines whether it has followed the action of the drone group based on the size of the action follow sequence number.

9. A bio-inspired large-scale bee colony networking and formation control method according to claim 7, characterized in that, In step 52), the flocking mode, obstacle avoidance mode, and follow mode of the starling are mapped to the group formation control, specifically as follows: For cluster mode, the interaction pattern between the repulsive and attractive regions is shown in the following equation: The group number is as follows: The neighbor group number is , Grouping To Group unit vector, Indicates grouping Grouping The radial difference between them Grouping The neighbor group set, when At that time, the groups are subjected to a repulsive force. At that time, an attractive force exists between the groups, and the radial difference will eventually converge at... and between; For neighbors located in the alignment region, the navigator adjusts its own speed based on the speed of neighbor grouping, as shown in the following formula: in, Grouping speed, Grouping speed, Indicates grouping Neighbors in the alignment region, For cruising speed, i.e., grouping The speed is the average speed of the neighbors in the aligned region; In obstacle avoidance mode, the navigator automatically changes its flight direction to avoid obstacles. In follow mode, the navigator follows the neighbor with the largest speed change to avoid obstacles in advance. Specifically, it first determines whether to follow the group based on the action follow sequence number of the neighbor group. If it does follow, it makes the following judgment: in, Indicates neighbor grouping The speed change indicator, if the speed of the neighbor group is a speed that has not been imitated before, that is, the current speed of the neighbor group. Not equal to a number of moments The previous group speed The velocity change of the neighbor is measured by comparing it with the velocity change at the previous time step; otherwise, it is considered that the neighbor cannot follow. Set to 0; In follow mode, only one neighbor group is followed by this group, so the node with the largest speed change among the neighbor groups is selected. Let's assume the metric for this group is... Its expression is as follows: Establish a speed switching threshold to evaluate the magnitude of group speed changes, wherein the speed switching threshold... Represented as: in, It is a coefficient. It is a collection of neighbors. It is the number of followable neighbors in the neighbor set. The degree to which the average velocity of the current neighbor can be followed changes; set up Corresponding group number ,like Then the navigator enters follow mode and groups up. Imitation grouping The speed; if Then the navigator enters cluster mode and is divided into groups. Do not imitate grouping The speed of the packet is maintained to ensure consistency between the speed of the packet and the speed of the neighboring packets, as shown in the following formula: in, This refers to the drone's cruising speed.

Images