Self-adaptive scaling formation cooperation method for unmanned bee colony

Abstract

The invention relates to a self-adaptive scaling formation cooperation method for unmanned swarms, the unmanned swarms can realize self-adaptive formation scaling according to a complex environment, the flight safety of the unmanned swarms is ensured, and the method can be widely applied to multiple fields of public safety monitoring, environmental protection monitoring, agricultural management, traffic dispersion and the like. Through cooperative obstacle avoidance and formation, tasks can be efficiently and accurately completed in a complex environment, real-time data support is provided, and powerful guarantee is provided for development of various fields. Along with continuous maturation of the technology, the application prospect of unmanned bee colony formation and obstacle avoidance collaboration is wider.

Description

Technical Field

[0001] The present application relates to the technical field of multi-agent system collaboration, and in particular, to an unmanned bee colony adaptive scaling formation collaboration method. Background Art

[0002] Unmanned swarms are composed of a large number of patrol aircraft. The safe formation flight technology of unmanned swarms in obstacle environments has broad application prospects. With the rapid development of industries such as logistics, environmental monitoring, and disaster relief, higher requirements are placed on efficient and flexible aerial operations. Collaborative formation obstacle avoidance technology has become an important means to meet these needs. Collaborative formation obstacle avoidance technology enables unmanned swarms to autonomously perceive the surrounding environment during flight, identify and avoid obstacles in real time. This greatly reduces the risk of collisions that may occur during flight and improves the flight safety of unmanned swarms. In formation flight, each unmanned swarm needs to maintain a certain distance and formation with other members. Collaborative formation obstacle avoidance technology can ensure that each member can make coordinated adjustments when encountering obstacles, maintaining the stability of the formation and the consistency of the formation. Through collaborative obstacle avoidance, unmanned swarm formations can complete flight missions more efficiently in complex environments. For example, in logistics distribution, formation flight can reach the destination faster and improve distribution efficiency; in environmental monitoring, formation flight can complete data collection and analysis of designated areas faster. Collaborative formation obstacle avoidance technology enables unmanned swarms to work in more complex and challenging environments, such as mountainous areas, forests, and cities. This greatly expands the application areas of unmanned swarms and provides strong support for the widespread application and promotion of drone technology. However, the adaptive ability of the current unmanned swarm collaborative formation obstacle avoidance technology still needs to be further improved. Summary of the invention

[0003] In order to overcome at least one shortcoming in the prior art, the present application provides an unmanned swarm adaptive scaling formation collaboration method.

[0004] In a first aspect, a method for adaptively scaling formation collaboration of an unmanned bee colony is provided, comprising:

[0005] Construct an unmanned swarm model; the unmanned swarm model is used to describe the relationship between the state information of the UAV and the control input of the UAV;

[0006] Determine the safe distance between drones and the safe distance between drones and obstacles;

[0007] Based on the unmanned bee swarm model, a cost function for the adaptive scaling formation collaboration of unmanned bee swarms is constructed. The cost function includes formation cost term, aggregation cost term, rate cost term, direction cost term and control cost term. The constraints of the cost function are set according to the safe distance between UAVs and the safe distance between UAVs and obstacles.

[0008] When the state information of the UAV is obtained, the control input of the UAV is solved when the cost function reaches the minimum value.

[0009] In one embodiment, the unmanned bee colony model is:

[0010]

[0011]

[0012] in, is the state information of UAV i at time k, T represents transposition, and p i (k) is the position of UAV i at time k, v i (k) is the speed of UAV i at time k, x i (k+1) is the state information of drone i at time k+1, i is the number of the drone in the drone swarm, is the label set of all drones; u i (k) is the control input of UAV i at time k; A i is the system matrix, B i is the input matrix; I n is the n-order unit matrix, 0 n is an n-order matrix with 0 elements, n is the system dimension, and ΔT is the sampling period.

[0013] In one embodiment, the safe distance between drones is:

[0014] d safety-agent =2(r agent +r margin-agent )

[0015] Among them, d safety-agent is the safe distance between drones, r agent is the radius of the drone, r margin-agent Provides a safety buffer radius for drones;

[0016] The safe distance between the drone and obstacles is:

[0017] d safety-obs =r agent +r margin-agent +r obs +r margin-obs

[0018] Among them, d safety-obs is the safe distance between the drone and the obstacle, r obs is the radius of the obstacle, r margin-obs It is the safety buffer radius of obstacles.

[0019] In one embodiment, the cost function is:

[0020] J ∑ (k) = J f (k)+J coh (k)+J s (k)+J d (k)+J c (k)

[0021] The constraints are:

[0022] x i (k+l+1|j)=A i x i (k+l|k)+B i u i (k+l|k)

[0023]

[0024]

[0025]

[0026] ‖p m -p i (k+l|k)‖≥d safety-obs , m∈M obs

[0027] Among them, J ∑ (k) is the cost function value, J f (k) is the formation cost, J coh (k) is the aggregation cost term, J s (k) is the rate cost term, J d (k) is the direction cost term, J c (k) is the control consideration item;

[0028] x i (k+l+1|k) is the state information of UAV i at the k+l+1th moment predicted when the current moment is k, l is the time step, l=0,1,2,…,P-1, P is the prediction time domain, A i is the system matrix, B i is the input matrix, x i (k+l|k) is the state information of UAV i predicted at the k+lth moment when the current moment is k, u i (k+l|k) is the control input of UAV i predicted at the k+lth moment when the current moment is k, is the state information component set of UAV i, is the set of control input components of UAV i, p j(k+l|k) is the predicted position of drone j at the k+lth moment when the current moment is k, p i (k+l|k) is the predicted position of UAV i at the k+lth moment when the current moment is k, N i is the set of neighboring drones of drone i, d safety-agent is the safe distance between drones; p m is the position of obstacle m, M obs is the obstacle set, d safety-obs is the safe distance between the drone and obstacles.

[0029] In one embodiment, the formation cost term is:

[0030]

[0031] Among them, J f (k) is the formation cost term, l is the time step, P is the prediction time domain, i is the number of drones in the unmanned bee swarm, N is the number of drones in the unmanned bee swarm, and J f,i (k+l|k) is the formation cost item predicted by UAV i at the k+lth moment when the current moment is k, λ f is the formation cost item weight, ‖·‖ F is the F-norm of the matrix, is the simple normalized Laplace matrix of the graph corresponding to the formation formed by the unmanned swarm, is the simple normalized Laplacian matrix of the graph corresponding to the desired formation.

[0032] In one embodiment, the aggregate cost term is:

[0033]

[0034]

[0035] Among them, J coh (k) is the aggregation cost term, l is the time step, P is the prediction time domain, i is the number of drones in the unmanned bee swarm, N is the number of drones in the unmanned bee swarm, and J coh,i (k+l|k) is the aggregate cost term predicted by drone i at the k+lth moment when the current moment is k, λ coh is the weight of the aggregate cost term, N i is the set of neighboring drones of drone i, j represents N i UAV in the j (k+l|k) is the predicted position of drone j at the k+kth moment when the current moment is k, p i (k+k|k) is the predicted position of drone i at the k+lth moment when the current moment is k, w ij, ref is the element in the adjacency matrix of the graph corresponding to the desired formation, ‖·‖ 2 represents the 2-norm, and |·| represents the absolute value.

[0036] In one embodiment, the rate cost term is:

[0037]

[0038] J s，i (k+l||k)=λ s (||v i (k+l||k)|||-v ref ) 2

[0039] Among them, J S (k) is the rate cost term, l is the time step, P is the prediction time domain, i is the number of drones in the unmanned bee swarm, N is the number of drones in the unmanned bee swarm, and J S,i (k+l|k) is the rate cost item predicted by UAV i at the k+lth moment when the current moment is k, λ S is the rate cost item weight, v i (k+l|k) is the predicted speed of UAV i at the k+lth moment when the current moment is k, ‖·‖ represents the norm, and v ref is the expected speed of the formation;

[0040] The direction cost term is:

[0041]

[0042]

[0043] Among them, J d (k) is the direction cost term, J d,i (k+l|k) is the direction cost item predicted by UAV i at the k+lth moment when the current moment is k, λ d is the direction cost item weight, u ref is the expected direction of formation motion, ‖·‖ represents the norm.

[0044] In one embodiment, the control cost term is:

[0045]

[0046] J c，i (k+l|k)=λ c ||u i (k+l|k)|| 2

[0047] Among them, J C(k) is the control cost term, l is the time step, P is the prediction time domain, i is the number of drones in the unmanned bee swarm, N is the number of drones in the unmanned bee swarm, and J C,i (k+l|k) is the control cost item predicted by UAV i at the k+lth moment when the current moment is k, λ C To control the weight of the cost term, u i (k+l|k) is the control input of UAV i at the k+lth moment predicted when the current moment is k, ‖·‖ 2 represents the 2-norm.

[0048] In one embodiment, when the state information of the drone is obtained, solving the control input of the drone when the cost function reaches the minimum value includes:

[0049] Step S41, for UAV i in the UAV swarm, set the formation cost item weight λ f , weight of aggregate cost term λ coh , rate cost term weight λ S , direction cost term weight λ d , control cost item weight λ C ; Set the sampling period ΔT, prediction time domain P, and maximum control step K max ; Set the initial state information x of drone i i (0);

[0050] Step S42: at the current time k=0, x i (0) Substitute the cost function and solve the control input u of the drone when the cost function reaches the minimum value i (0+l|0), l is the time step, l=0,1,2,…,P-1; take the control input u corresponding to l=0 i (0|0) is used as the control input u executed by drone i i (0), that is u i (0) = u i (0|0), drone i executes control input u i (0);

[0051] Step S43, when the current time k>0, perform the following steps:

[0052] Step S431, obtain the status information x of drone i at the current moment k i (k);

[0053] Step S432: x i (k) Substitute the cost function and solve for the control input u of the drone when the cost function reaches the minimum value i (k+l|k), l is the time step, k=0,1,2,…,P-1;

[0054] Step S433, take the control input u corresponding to l=0 i (k|k) is the control input u executed by UAV i i (k), that is, u i (k) = u i (k|k), drone i executes control input u i (k);

[0055] Step S434, if at the next moment k+1>K max , then end, otherwise, return to step S431.

[0056] In a second aspect, an unmanned bee colony adaptive scaling formation coordination device is provided, comprising:

[0057] The unmanned bee swarm model building module is used to build an unmanned bee swarm model; the unmanned bee swarm model is used to describe the relationship between the state information of the UAV and the control input of the UAV;

[0058] A safety distance determination module is used to determine the safety distance between drones and the safety distance between drones and obstacles;

[0059] The cost function construction module is used to construct the cost function of the unmanned swarm adaptive scaling formation collaboration based on the unmanned swarm model. The cost function includes formation cost term, aggregation cost term, rate cost term, direction cost term and control cost term. The constraint conditions of the cost function are set according to the safe distance between UAVs and the safe distance between UAVs and obstacles.

[0060] The solving module is used to solve the control input of the UAV when the cost function reaches the minimum value while obtaining the state information of the UAV.

[0061] Compared with the prior art, this application has the following beneficial effects:

[0062] 1. This application uses a formation cost term designed based on a simple normalized Laplace matrix, so that the formed formation has scaling and rotation characteristics. Therefore, the formation has more flexibility when passing through an obstacle area and can better coordinate the formation and obstacle avoidance.

[0063] 2. This application constructs a speed cost item to achieve controllable speed of the unmanned swarm formation. The speed cost item and the formation cost item cooperate with each other to make the unmanned swarm formation have a certain degree of adaptability, which can give full play to the scaling characteristics of the formation. Once the formation is blocked in an obstacle environment, its rate will drop, and the rate cost item will increase rapidly, further causing the total cost of the system to increase rapidly. At this time, the various cost items can be re-optimized to try to get out of the predicament. In particular, enlarging or reducing the formation does not cause changes in the formation cost item. This attempt may enable the formation to get out of the predicament, which enhances its obstacle avoidance ability.

[0064] 3. This application uses nonlinear model prediction, taking into account the safety issues of unmanned swarm formation flight, and also taking into account the dynamic constraints of the unmanned swarm itself, such as actuator saturation constraints, speed upper limits, acceleration upper limits and other practical constraints, making this application very practical. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] The present application may be better understood by referring to the following description given in conjunction with the accompanying drawings, which together with the following detailed description are included in this specification and form a part of this specification. In the drawings:

[0066] Figure 1 A flowchart of an unmanned bee colony adaptive scaling formation collaboration method according to an embodiment of the present application is shown;

[0067] Figure 2 A schematic diagram of setting a safe distance is shown;

[0068] Figure 3 A structural block diagram of an unmanned bee colony adaptive scaling formation collaboration device according to an embodiment of the present application is shown. DETAILED DESCRIPTION

[0069] The exemplary embodiments of the present application will be described below in conjunction with the accompanying drawings. For the sake of clarity and conciseness, not all features of the actual embodiments are described in the specification. However, it should be understood that many implementation-specific decisions can be made in the process of developing any such actual embodiments in order to achieve the specific goals of the developer, and these decisions may vary from embodiment to embodiment.

[0070] It is also necessary to explain here that, in order to avoid obscuring the present application due to unnecessary details, only the device structure closely related to the scheme according to the present application is shown in the drawings, while other details that are not closely related to the present application are omitted.

[0071] It should be understood that the present application is not limited to the described implementation forms due to the following description with reference to the accompanying drawings. In this article, where feasible, the embodiments can be combined with each other, features between different embodiments can be replaced or borrowed, and one or more features can be omitted in one embodiment.

[0072] In response to the current lack of unmanned bee swarms' adaptive formation capabilities in obstacle environments, the present application provides an unmanned bee swarm adaptive scaling formation collaboration method, which aims to establish a mapping relationship between adaptive scaling formations and key cost functions, and to design an unmanned bee swarm formation that can adapt to the current environment based on nonlinear model prediction, thereby achieving autonomous and safe formation collaboration of unmanned bee swarms in complex environments.

[0073] Figure 1 A flowchart of the unmanned bee colony adaptive scaling formation collaboration method according to an embodiment of the present application is shown, see Figure 1 , methods include:

[0074] Step S1, constructing an unmanned swarm model; the unmanned swarm model is used to describe the relationship between the state information of the unmanned aerial vehicle and the control input of the unmanned aerial vehicle.

[0075] The unmanned bee swarm is a formation of N drones, each of which obeys the following discrete second-order linear system, namely the unmanned bee swarm model:

[0076]

[0077]

[0078] in, is the state information of UAV i at time k, T represents transposition, and p i (k) is the position of UAV i at time k, v i (k) is the speed of UAV i at time k, x i (k+1) is the state information of drone i at time k+1, i is the number of the drone in the drone swarm, is the label set of all drones, u i (k) is the control input of UAV i at time k. Assuming that the UAV swarm has an acceleration controller, the control input here is acceleration; A i is the system matrix, B i is the input matrix; I n is the n-order unit matrix, 0 n It is an n-order matrix with 0 elements, n is the system dimension, which can be a 2D plane or a 3D space, and ΔT is the sampling period.

[0079] Consider the state and control input constraints of the unmanned swarm: Indicates the definition, is the state information component set of UAV i, is the set of control input components of UAV i, x ih For x i The hth component of ih and x i The lower and upper bounds of the hth component of ih for u i The hth component of ih and u i The lower and upper bounds of the h-th component of .

[0080] Step S2, determining the safe distance between drones and the safe distance between drones and obstacles.

[0081] Figure 2 The schematic diagram of the safety distance setting is shown. To improve the robustness of the method, the safety buffer radius r of the drone is set margin-agent , the safety buffer radius r of the obstacle margin-obs ; Figure 2 In the figure, the radius of the unmanned bee colony is r agent , the minimum distance d that needs to be maintained between unmanned bee colonies without collision coll-agent =2r agent The black area is an obstacle with a radius of r. obs , the minimum distance d between the unmanned bee swarm and obstacles without collision coll-obs =r agent +r obs . Figure 2 The medium and long strip area is the forward prediction trajectory predicted by the model, and its length is determined by the prediction time domain P.

[0082] The safe distance between drones is:

[0083] d safety-agent =2(r agent +r margin-agent )

[0084] Among them, d safety-agent is the safe distance between drones, r agent is the radius of the drone, r margin-agent Provides a safety buffer radius for drones;

[0085] The safe distance between the drone and obstacles is:

[0086] d safety-obs =r agent +rmargin-agent +r obs +r margin-obs

[0087] Among them, d safety-obs is the safe distance between the drone and the obstacle, r obs is the radius of the obstacle, r margin-obs It is the safety buffer radius of obstacles.

[0088] Step S3, constructing a cost function for the adaptive scaling formation collaboration of the unmanned swarm based on the unmanned swarm model, the cost function includes a formation cost term, an aggregation cost term, a rate cost term, a direction cost term and a control cost term; the constraints of the cost function are set according to the safe distance between the drones and the safe distance between the drones and obstacles.

[0089] Step S4, when the state information of the UAV is obtained, the control input of the UAV is solved when the cost function reaches the minimum value.

[0090] The unmanned bee swarm of this embodiment can realize adaptive formation scaling according to complex environments, ensuring the flight safety of the unmanned bee swarm, and can be widely used in public safety monitoring, environmental protection monitoring, agricultural management, traffic diversion and other fields. Through collaborative obstacle avoidance, the formation can complete tasks efficiently and accurately in complex environments, provide real-time data support, and provide strong guarantees for the development of various fields. As the technology continues to mature, the application prospects of unmanned bee swarm formation obstacle avoidance collaboration will be broader.

[0091] In one embodiment, the cost function is:

[0092] J ∑ (k) = J f (k)+J coh (k)+J s (k)+J d (k)+J c (d)

[0093] The constraints are:

[0094] x i (k+l+1|k)=A i x i (k+l|k)+B i u i (k+l|k)

[0095]

[0096]

[0097]

[0098] ‖p m -p i (k+l|k)‖≥d safety-obs , m∈M obs

[0099] Among them, J ∑ (k) is the cost function value, J f (k) is the formation cost, J coh (k) is the aggregation cost term, J s (k) is the rate cost term, J d (k) is the direction cost term, J c (k) is the control consideration item;

[0100] x i (k+l+1|k) is the state information of UAV i at the k+l+1th moment predicted when the current moment is k, l is the time step, l=0,1,2,…,P-1, P is the prediction time domain, A i is the system matrix, B i is the input matrix, x i (k+l|k) is the state information of UAV i predicted at the k+k moment when the current moment is k, u i (k+l|k) is the control input of UAV i predicted at the k+lth moment when the current moment is k, is the state information set of UAV i, The control input set of drone i, p j (k+l|k) is the predicted position of drone j at the k+lth moment when the current moment is k, p i (k+l|k) is the predicted position of UAV i at the k+lth moment when the current moment is k, N i is the set of neighboring drones of drone i, d safety-agent is the safe distance between drones; p m is the position of obstacle m, M obs is the obstacle set, d safety-obs is the safe distance between the drone and obstacles.

[0101] Specifically, the unmanned swarm is controlled to form a desired formation based on the distance constraint. The desired formation can be represented by a graph, where the nodes in the graph are the drones in the unmanned swarm, and the node information is the status information of the drones.

[0102] The adjacency matrix of the graph corresponding to the expected formation is:

[0103]

[0104] w ij,ref is the element in the adjacency matrix of the graph corresponding to the desired formation, represents the distance between UAVs i and j, R N×N Represents an N×N real number matrix. According to the adjacency matrix, the simple normalized Laplace matrix of the desired formation can be obtained I is the unit matrix, is the in-degree matrix D ref The inverse matrix of for The elements in .

[0105] The formation cost is:

[0106]

[0107]

[0108] Among them, J f (k) is the formation cost term, l is the time step, P is the prediction time domain, i is the number of drones in the unmanned bee swarm, N is the number of drones in the unmanned bee swarm, and J f,i (k+l|k) is the formation cost item predicted by UAV i at the k+lth moment when the current moment is k, λ f is the formation cost item weight, ‖·‖ F is the F-norm of the matrix, is the simple normalized Laplace matrix of the graph corresponding to the formation formed by the unmanned swarm, is the simple normalized Laplacian matrix of the graph corresponding to the desired formation.

[0109] N i is the set of neighboring drones of drone i, j represents N i UAV in the j (k+l|k) is the predicted position of drone j at the k+lth moment when the current moment is k, p i (k+l|k) is the predicted position of UAV i at the k+lth moment when the current moment is k, ‖·‖ 2 represents the 2-norm, for The elements in .

[0110] Specifically, if we only rely on the formation cost, the formation shape will scale without limit, and when there is no obstacle, the formation shape will be too large or too small. Therefore, the aggregation cost is designed to ensure that the formation is limited in scale. The aggregation cost is:

[0111]

[0112]

[0113] Among them, J coh(k) is the aggregation cost term, l is the time step, P is the prediction time domain, i is the number of drones in the unmanned bee swarm, N is the number of drones in the unmanned bee swarm, and J coh,i (k+l|k) is the aggregate cost term predicted by drone i at the k+lth moment when the current moment is k, λ coh is the weight of the aggregate cost term, N i is the set of neighboring drones of drone i, j represents N i UAV in the j (k+l|k) is the predicted position of drone j at the k+lth moment when the current moment is k, p i (k+l|k) is the predicted position of drone i at the k+lth moment when the current moment is k, w ij,ref is the element in the adjacency matrix of the graph corresponding to the desired formation, ‖·‖ 2 represents the 2-norm, and |·| represents the absolute value.

[0114] Specifically, the speed cost term drives the formation to move forward at a desired speed, and the direction cost term ensures that the formation moves in a desired direction until it reaches a specified area.

[0115] The rate cost term is:

[0116]

[0117] J si (k+l||k)=λ s (|||v i (k+l||k)||-v ref ) 2

[0118] Among them, J S (k) is the rate cost term, l is the time step, P is the prediction time domain, i is the number of drones in the unmanned bee swarm, N is the number of drones in the unmanned bee swarm, and J S,i (k+l|k) is the rate cost item predicted by UAV i at the k+lth moment when the current moment is k, λ S is the rate cost item weight, v i (k+l|k) is the predicted speed of UAV i at the k+lth moment when the current moment is k, ‖·‖ represents the norm, and v ref is the expected velocity of the formation. ref =||v ref ||;v ref is the expected speed of the formation.

[0119] The direction cost term is:

[0120]

[0121]

[0122] Among them, J d (k) is the direction cost term, J d,i (k+l|k) is the direction cost item predicted by UAV i at the k+lth moment when the current moment is k, λ d is the weight of the direction cost term. ref is the expected direction of the formation movement, which is a unit vector, u ref =v ref / ||v ref ||.

[0123] Specifically, in the process of formation movement, the ideal state is that each drone moves in coordination at the same speed. That is, the control input (acceleration) of each drone approaches 0 when the formation moves stably. At the same time, the control input can physically characterize the energy consumption of the unmanned bee swarm. From this perspective, taking 0 as the expected value of the control input can make the formation take into account the energy optimization during the obstacle avoidance process. The control cost term is:

[0124]

[0125] J c，i (k+l|k)=λ c ||u i (k+l|k)|| 2

[0126] Among them, J C (k) is the control cost term, l is the time step, P is the prediction time domain, i is the number of drones in the unmanned bee swarm, N is the number of drones in the unmanned bee swarm, and J C,i (k+l|k) is the control cost item predicted by UAV i at the k+lth moment when the current moment is k, λ C To control the weight of the cost term, u i (k+l|k) is the control input of UAV i at the k+lth moment predicted when the current moment is k, ‖·‖ 2 represents the 2-norm.

[0127] In one embodiment, in step S4, when the state information of the drone is obtained, solving the control input of the drone when the cost function reaches the minimum value includes:

[0128] Step S41, for UAV i in the UAV swarm, set the formation cost item weight λ f , weight of aggregate cost term λ coh , rate cost term weight λ S , direction cost term weight λ d , control cost item weight λC ; Set the sampling period ΔT, prediction time domain P, and maximum control step K max ; Set the initial state information x of drone i i (0); K max =T max / ΔT,T max The task end time.

[0129] Step S42: at the current time k=0, x i (0) Substitute the cost function and solve the control input u of the drone when the cost function reaches the minimum value i (0+l|0), l is the time step, l=0,1,2,…,P-1; take the control input u corresponding to l=0 i (0|0) is used as the control input u executed by drone i i (0), that is u i (0) = u i (0|0), drone i executes control input u i (0).

[0130] Step S43, when the current time k>0, perform the following steps:

[0131] Step S431, obtain the status information x of drone i at the current moment k i (k);

[0132] Step S432: x i (k) Substitute the cost function and solve for the control input u of the drone when the cost function reaches the minimum value i (k+l|k), l is the time step, l=0,1,2,…,P-1;

[0133] Step S433, take the control input u corresponding to l=0 i (k|k) is the control input u executed by UAV i i (k), that is, u i (k) = u i (k|k), drone i executes control input u i (k);

[0134] Step S434, if at the next moment k+1>K max , then it ends, otherwise, return to step S431 and execute the solution process at the next moment.

[0135] Based on the same inventive concept as the unmanned bee colony adaptive scaling formation collaboration method, this embodiment also provides a corresponding unmanned bee colony adaptive scaling formation collaboration device. Figure 3A structural block diagram of an unmanned bee colony adaptive scaling formation collaboration device according to an embodiment of the present application is shown, and the device includes:

[0136] The unmanned bee swarm model building module 31 is used to build an unmanned bee swarm model; the unmanned bee swarm model is used to describe the relationship between the state information of the unmanned aerial vehicle and the control input of the unmanned aerial vehicle;

[0137] A safety distance determination module 32, used to determine the safety distance between drones and the safety distance between drones and obstacles;

[0138] The cost function construction module 33 is used to construct a cost function of the unmanned bee swarm adaptive scaling formation collaboration based on the unmanned bee swarm model, wherein the cost function includes a formation cost term, an aggregation cost term, a rate cost term, a direction cost term, and a control cost term; the constraint conditions of the cost function are set according to the safe distance between the unmanned aerial vehicles and the safe distance between the unmanned aerial vehicles and obstacles;

[0139] The solving module 34 is used to solve the control input of the UAV when the cost function reaches the minimum value when the state information of the UAV is obtained.

[0140] The unmanned bee swarm adaptive scaling formation collaboration device of this embodiment has the same inventive concept as the unmanned bee swarm adaptive scaling formation collaboration method described above, so the specific implementation method of the device can be seen in the implementation example part of the unmanned bee swarm adaptive scaling formation collaboration method in the previous text, and its technical effect corresponds to the technical effect of the above-mentioned method, which will not be repeated here.

[0141] In summary, this application has the following technical effects:

[0142] 1. This application uses a formation cost term designed based on a simple normalized Laplace matrix to make the formed formation have scaling and rotation characteristics, so the formation has more flexibility when passing through an obstacle area and can better coordinate the formation and obstacle avoidance.

[0143] 2. This application constructs a speed cost item to achieve controllable speed of the unmanned swarm formation. The speed cost item and the formation cost item cooperate with each other to make the unmanned swarm formation have a certain degree of adaptability, which can give full play to the scaling characteristics of the formation. Once the formation is blocked in an obstacle environment, its rate will drop, and the rate cost item will increase rapidly, further causing the total cost of the system to increase rapidly. At this time, the various cost items can be re-optimized to try to get out of the predicament. In particular, enlarging or reducing the formation does not cause changes in the formation cost item. This attempt may enable the formation to get out of the predicament, which enhances its obstacle avoidance ability.

[0144] 3. This application uses nonlinear model prediction, taking into account the safety issues of unmanned swarm formation flight, and also taking into account the dynamic constraints of the unmanned swarm itself, such as actuator saturation constraints, speed upper limits, acceleration upper limits and other practical constraints, making this application very practical.

[0145] The above are only various implementations of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. An unmanned bee colony adaptive scaling formation collaboration method, characterized in that: include: Constructing an unmanned bee swarm model; the unmanned bee swarm model is used to describe the relationship between the state information of the unmanned aerial vehicle and the control input of the unmanned aerial vehicle; Determine the safe distance between drones and the safe distance between drones and obstacles; Based on the unmanned bee swarm model, a cost function for the adaptive scaling formation collaboration of the unmanned bee swarm is constructed, wherein the cost function includes a formation cost term, an aggregation cost term, a rate cost term, a direction cost term, and a control cost term; and the constraint conditions of the cost function are set according to the safe distance between the unmanned aerial vehicles and the safe distance between the unmanned aerial vehicles and obstacles; When the state information of the UAV is obtained, the control input of the UAV is solved when the cost function reaches the minimum value.

2. The method according to claim 1, characterized in that The unmanned bee colony model is: in, is the state information of UAV i at time k, T represents transposition, and p i (k) is the position of UAV i at time k, v i (k) is the speed of UAV i at time k, x i (k+1) is the state information of drone i at time k+1, i is the number of the drone in the drone swarm, is the label set of all drones; u i (k) is the control input of UAV i at time k; A i is the system matrix, B i is the input matrix; I n is the n-order unit matrix, 0 n is an n-order matrix with 0 elements, n is the system dimension, and ΔT is the sampling period.

3. The method according to claim 1, characterized in that The safe distance between the drones is: d safety-agent ＝2(r agent +r margin-agent ) Among them, d safety-agent is the safe distance between drones, r agent is the radius of the drone, r margin-agent Provides a safety buffer radius for drones; The safe distance between the drone and the obstacle is: d safety-obs ＝r agent +r margin-agent +r obs +r margin-obs Among them, d safety-obs is the safe distance between the drone and the obstacle, r obs is the radius of the obstacle, r margin-obs It is the safety buffer radius of obstacles.

4. The method according to claim 1, characterized in that The cost function is: J ∑ (k)=J f (k)+J coh (k)+J s (k)+J d (k)+J c (k) The constraints are: x i (k+l+1|k)=A i x i (k+l|k)+B i u i (k+l|k) ||p j (k+l|k)-p i (k+l|k)||≥d safety-agent ,j∈N i ‖p m -p i (k+l|k)‖≥d safety-obs ,m∈M obs Among them, J ∑ (k) is the cost function value, J f (k) is the formation cost, J coh (k) is the aggregation cost term, J s (k) is the rate cost term, J d (k) is the direction cost term, J c (k) is the control consideration item; x i (k+l+1|k) is the state information of UAV i at the k+l+1th moment predicted when the current moment is k, l is the time step, l=0,1,2,…,P-1, P is the prediction time domain, A i is the system matrix, B i is the input matrix, x i (k+l|k) is the state information of UAV i predicted at the k+lth moment when the current moment is k, u i (k+l|k) is the control input of UAV i predicted at the k+lth moment when the current moment is k, is the state information component set of UAV i, is the set of control input components of UAV i, p j (k+l|k) is the predicted position of drone j at the k+lth moment when the current moment is k, p i (k+l|k) is the predicted position of UAV i at the k+lth moment when the current moment is k, N i is the set of neighboring drones of drone i, d safety-agent is the safe distance between drones; p m is the position of obstacle m, M obs is the obstacle set, d safety-obs is the safe distance between the drone and obstacles.

5. The method according to claim 1, characterized in that The formation cost term is: Among them, J f (k) is the formation cost term, l is the time step, P is the prediction time domain, i is the number of drones in the unmanned bee swarm, N is the number of drones in the unmanned bee swarm, and J f,i (k+l|k) is the formation cost item predicted by UAV i at the k+lth moment when the current moment is k, λ f is the formation cost item weight, ‖·‖ F is the F-norm of the matrix, is the simple normalized Laplace matrix of the graph corresponding to the formation formed by the unmanned swarm, is the simple normalized Laplacian matrix of the graph corresponding to the desired formation.

6. The method according to claim 1, characterized in that The aggregation cost term is: Among them, J coh (k) is the aggregation cost term, l is the time step, P is the prediction time domain, i is the number of drones in the unmanned bee swarm, N is the number of drones in the unmanned bee swarm, and J coh,i (k+l|k) is the aggregate cost term predicted by drone i at the k+lth moment when the current moment is k, λ coh is the weight of the aggregate cost term, N i is the set of neighboring drones of drone i, j represents N i UAV in the j (k+l|k) is the predicted position of drone j at the k+lth moment when the current moment is k, p i (k+l|k) is the predicted position of drone i at the k+lth moment when the current moment is k, w ij , ref is the element in the adjacency matrix of the graph corresponding to the desired formation, ‖·‖ 2 represents the 2-norm, and |·| represents the absolute value.

7. The method according to claim 1, characterized in that The rate cost term is: J,(k+7k)=V(llv,(k+7[k)Ⅲ-Vrer) 2 Among them, J S (k) is the rate cost term, l is the time step, P is the prediction time domain, i is the number of drones in the unmanned bee swarm, N is the number of drones in the unmanned bee swarm, and J S,i (k+l|k) is the rate cost item predicted by UAV i at the k+lth moment when the current moment is k, λ S is the rate cost item weight, v i (k+l|k) is the predicted speed of UAV i at the k+lth moment when the current moment is k, ‖·‖ represents the norm, and v ref is the expected speed of the formation; The direction cost term is: Among them, J d (k) is the direction cost term, J d,i (k+l|k) is the direction cost item predicted by UAV i at the k+lth moment when the current moment is k, λ d is the direction cost item weight, u ref is the expected direction of formation motion, ‖·‖ represents the norm.

8. The method according to claim 1, characterized in that The control cost term is: J.(k+lk)=UI(k+ / ]k)2 Among them, J C (k) is the control cost term, l is the time step, P is the prediction time domain, i is the number of drones in the unmanned bee swarm, N is the number of drones in the unmanned bee swarm, and J C,i (k+l|k) is the control cost item predicted by UAV i at the k+lth moment when the current moment is k, λ C To control the weight of the cost term, u i (k+l|k) is the control input of UAV i predicted at the k+lth moment when the current moment is k, ‖·‖ 2 represents the 2-norm.

9. The method according to claim 1, characterized in that in, In the case of obtaining the state information of the drone, solving the control input of the drone when the cost function reaches the minimum value includes: Step S41, for UAV i in the UAV swarm, set the formation cost item weight λ f , weight of aggregate cost term λ coh , rate cost term weight λ S , direction cost term weight λ d , control cost item weight λ C ; Set the sampling period ΔT, prediction time domain P, and maximum control step K max ; Set the initial state information x of drone i i (0); Step S42: at the current time k=0, x i (0) Substitute the cost function and solve the control input u of the UAV when the cost function reaches the minimum value i (0+l|0), l is the time step, l=0,1,2,…,P-1; take the control input u corresponding to l=0 i (0|0) is used as the control input u executed by drone i i (0), that is u i (0) = u i (0|0), drone i executes control input u i (0); Step S43, when the current time k>0, perform the following steps: Step S431, obtain the status information x of drone i at the current moment k i (k); Step S432: x i (j) Substitute the cost function and solve the control input u of the drone when the cost function reaches the minimum value i (k+l|k), l is the time step, l=0,1,2,…,P-1; Step S433, take the control input u corresponding to l=0 i (k|k) is the control input u executed by UAV i i (k), that is, u i (k) = u i (k|k), drone i executes control input u i (k); Step S434, if at the next moment k+1>K max , then end, otherwise, return to step S431.

10. An unmanned bee colony adaptive scaling formation coordination device, characterized in that: include: An unmanned bee swarm model building module, used to build an unmanned bee swarm model; the unmanned bee swarm model is used to describe the relationship between the state information of the unmanned aerial vehicle and the control input of the unmanned aerial vehicle; A safety distance determination module is used to determine the safety distance between drones and the safety distance between drones and obstacles; A cost function construction module is used to construct a cost function for the adaptive scaling formation collaboration of the unmanned bee swarm based on the unmanned bee swarm model, wherein the cost function includes a formation cost term, an aggregation cost term, a rate cost term, a direction cost term and a control cost term; and the constraint conditions of the cost function are set according to the safe distance between the unmanned aerial vehicles and the safe distance between the unmanned aerial vehicles and obstacles; The solution module is used to solve the control input of the drone when the cost function reaches a minimum value when the state information of the drone is obtained.

Images