Unmanned aerial vehicle group distributed resource scheduling method based on dynamic event triggering

Abstract

The invention discloses an unmanned aerial vehicle group distributed resource scheduling method based on dynamic event triggering, and belongs to the technical field of distributed optimization, and the method comprises the steps: firstly constructing an unmanned aerial vehicle group resource scheduling model with global equality constraints and a communication topology model according to a multi-agent system; secondly, designing a distributed resource scheduling algorithm for each unmanned aerial vehicle group based on a dynamic event triggering method; and finally, designing a distributed event trigger controller by utilizing the dynamic error and giving controller parameters to realize distributed resource scheduling. According to the unmanned aerial vehicle group distributed resource scheduling method based on dynamic event triggering provided by the invention, for the unmanned aerial vehicle group distributed resource scheduling problem with global equality constraints, the optimal solution of the problem can be obtained through exponential convergence under discrete communication, the bandwidth occupation of communication is greatly reduced, the probability of time delay and packet loss is reduced, and the communication efficiency is improved. The efficiency of the system is improved.

Description

Technical Field

[0001] The present invention belongs to the field of distributed optimization technology, and in particular relates to a distributed resource scheduling method for a drone swarm based on dynamic event triggering. Background Art

[0002] Large-scale drone swarms feature distributed communication. The goal of distributed optimization for drone swarms is to design distributed algorithms to find the optimal solution to a global optimization problem through communication and collaboration between drones. Compared to traditional centralized optimization algorithms, distributed optimization algorithms offer the following significant advantages: each drone in a swarm only needs to communicate with its neighbors, which not only improves network security but also reduces communication costs. Furthermore, the information and objective function of each drone in the swarm are effectively protected. Therefore, studying distributed optimization problems in drone swarm systems is of great significance.

[0003] However, previous research relied on continuous information exchange between drone swarms. However, in practical applications, system bandwidth and energy resources are limited, and continuous communication is often unavailable. Furthermore, excessive communication frequency can increase system operational burden and energy consumption. Event-triggered mechanisms, by designing trigger functions that trigger only when the trigger conditions are met, perform data sampling, information transmission, and controller updates upon triggering, effectively reducing communication overhead. Therefore, it is necessary to conduct research on event-triggered mechanisms for distributed resource scheduling.

[0004] Currently, research on distributed event triggering mechanisms for network systems can be divided into static event triggering methods and dynamic event triggering methods. One paper pioneered the application of static event triggering to distributed control. While traditional static event triggering can utilize dynamic errors to design trigger conditions, this method cannot eliminate the Zeno phenomenon (see Dimarogonas, Dimos V., Emilio Frazzoli, and Karl H. Johansson, "Distributed event-triggered control for multi-agent systems," IEEE Transactions on automatic control 57.5 (2011): 1291-1297). Building on the research of the former, a dynamic event triggering mechanism was applied to distributed control instead of the static event triggering mechanism, successfully eliminating the Zeno phenomenon (see Yi X, Liu K, Dimarogonas DV, et al., "Dynamic event-triggered and self-triggered control for multi-agent systems," IEEE Transactions on Automatic Control 64.8 (2018): 3300-3307). However, existing research has not applied the dynamic event triggering mechanism to distributed resource scheduling with equality constraints in directed graphs. Summary of the Invention

[0005] The purpose of the present invention is to provide a distributed resource scheduling method for drone swarms based on dynamic event triggering, to solve the problem of distributed resource scheduling of drone swarms under equality constraints existing in the prior art, and to reduce the communication pressure of the system.

[0006] To achieve the above objectives, the present invention provides a method for distributed resource scheduling of drone swarms based on dynamic event triggering, comprising the following steps:

[0007] Step 1: Based on the multi-agent system, a UAV swarm resource scheduling model and communication topology model with global equality constraints are constructed;

[0008] Step 2: Design a distributed resource scheduling algorithm for each drone swarm based on the dynamic event triggering method;

[0009] Step 3: Use dynamic error to design a distributed event trigger controller and give controller parameters to achieve distributed resource scheduling.

[0010] Preferably, the expression of the drone swarm resource scheduling model with global equality constraints constructed in step 1 is as follows:

[0011]

[0012] Where N is the number of drones in the swarm, x i Indicates the status of drone group i, representing the proportion of the number of drones assigned to drone group i to the total number of drones, x=(x1,x2,…,x N ) T Represents the resource allocation of each drone group, x satisfies the equality constraint The sum of all drone swarm resource allocations is 1, f i (x i ) is the cost function of the drone group i, which satisfies the strong convexity condition And f i (x i ) derivative is Lipschitz continuous: l0 and l are both positive numbers. represents the set of real numbers.

[0013] Preferably, the drone swarm communication topology model in step 1 is specifically described as follows:

[0014] The UAV swarm communication topology model is regarded as a multi-agent network system consisting of N nodes, and the communication topology is recorded as in Represents a node set, an edge set (i,j)∈ε means that node j receives information from node i, and the information from node i1 to node i l The path is an ordered sequence of edges (i k ,i k+1 ), k=1,···,l-1; the set of neighbors within node i is picture The adjacency matrix Defined as a ii =0, if (j,i)∈ε, a ij =1, otherwise 0; The Laplace matrix of Defined as When i≠j, l ij =-a ij ;ε represents edge, a ij Representation matrix The i-th row and j-th column element, l ij Representation matrix The element in row i and column j, l ii Representation matrix The element in row i and column i.

[0015] Preferably, the drone swarm communication topology model needs to meet the following requirements: Communication topology is a strongly connected balanced graph. And satisfy Represents the communication topology The Laplace matrix of express Multiply left by 1, 1 T Represents the transpose of a vector whose elements are all 1.

[0016] Preferably, the specific process of designing a distributed resource scheduling algorithm for each drone group based on the dynamic event triggering method in step 2 is as follows:

[0017] Step 201: Update the status of the drone group using the derivative information of each drone group;

[0018] Step 202: Design an updating method for auxiliary variables based on dynamic errors.

[0019] Preferably, in step 201, the expression for updating the drone swarm state using the derivative information of each drone swarm is as follows:

[0020]

[0021] in, represents the derivative of the state of the drone group i, β is the feedback gain and satisfies β>0, yes Time to The state prediction of drone group j at any moment, is the moment when drone swarm i is triggered for the kth time, yes Time to The state prediction of drone swarm i at any moment.

[0022] Preferably, the expression for updating the auxiliary variable based on the dynamic error design in step 202 is as follows:

[0023]

[0024] in, Derivative of the designed auxiliary variable, η i (t) represents the auxiliary variable of the design, represents the set of neighbors in drone group i, k i is the feedback gain and satisfies k i >0,δ i ∈[0,1] is an arbitrarily selected parameter, and m is the parameter to be designed.

[0025] Preferably, the specific expression for designing a distributed event trigger controller using dynamic error in step 3 is as follows:

[0026]

[0027] in, represents the time when the drone group i is triggered for the k+1th time, inf represents the maximum infimum, m and θ i are the controller parameters to be designed.

[0028] Preferably, the controller parameters in step 3 are as follows:

[0029]

[0030] in yes The second norm of the matrix, yes The smallest non-zero eigenvalue of , σ∈(0,1).

[0031] Therefore, the present invention adopts the above-mentioned method for distributed resource scheduling of drone swarms based on dynamic event triggering, which has the following beneficial effects:

[0032] (1) The present invention utilizes dynamic errors to design distributed algorithms and auxiliary variables, and only the information of the algorithm itself and its neighbors needs to be used in the algorithm iteration;

[0033] (2) The present invention designs an event trigger function, based on which trigger time is generated instead of continuous time, thereby reducing system communication costs;

[0034] (3) The present invention proposes a distributed resource scheduling method for drone swarms based on dynamic event triggering, and provides parameter conditions that make the algorithm converge exponentially.

[0035] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 This is an overall flow chart of a distributed resource scheduling method for drone swarms based on dynamic event triggering according to the present invention;

[0037] Figure 2 Schematic diagram of the UAV swarm communication topology model for the simulation experiment of the present invention;

[0038] Figure 3 is a schematic diagram of the resource ratio allocated to each drone group in the simulation experiment of the present invention;

[0039] Figure 4 Schematic diagram of the triggering moments of each drone group in the simulation experiment of the present invention;

[0040] Figure 5 is a schematic diagram of the change of total UAV resources in the simulation experiment of the present invention;

[0041] Figure 6 It is a schematic diagram of the change of the total system cost of the simulation experiment of the present invention. DETAILED DESCRIPTION

[0042] The following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but rather merely represents selected embodiments of the present invention. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without inventive effort are intended to fall within the scope of protection of the present invention.

[0043] See also Figure 1-6 ,A distributed resource scheduling method for a UAV swarm based on dynamic event triggering, includes the following steps:

[0044] Step 1: Construct a UAV swarm resource scheduling model and a communication topology model with global equality constraints based on the multi-agent system. The expression of the constructed UAV swarm resource scheduling model with global equality constraints is as follows:

[0045]

[0046] Where N is the number of drones in the swarm, x i Indicates the status of drone group i, representing the proportion of the number of drones assigned to drone group i to the total number of drones, x=(x1,x2,…,x N ) T Represents the resource allocation of each drone group, x satisfies the equality constraint The sum of all drone swarm resource allocations is 1, f i (x i ) is the cost function of the drone group i, which satisfies the strong convexity condition And f i (x i ) derivative is Lipschitz continuous: l0 and l are both positive numbers. The significance of this model is to minimize the sum of the cost functions of each drone group while keeping the total amount of resources allocated to each drone group unchanged.

[0047] The UAV swarm communication topology model is described as follows: The UAV swarm communication topology model is regarded as a multi-agent network system consisting of N nodes, and the communication topology structure is recorded as in Represents a node set, an edge set (i,j)∈ε means that node j receives information from node i, and the information from node i1 to node i l The path is an ordered sequence of edges (i k ,i k+1 ), k=1,···,l-1; the set of neighbors within node i is picture The adjacency matrix Defined as a ii =0, if (j,i)∈ε, a ij =1, otherwise 0; The Laplace matrix of Defined as When i≠j, l ij =-a ij ;ε represents edge, a ij Representation matrix The i-th row and j-th column element, l ij Representation matrix The element in row i and column j, l ii Representation matrix The element in row i and column i. In addition, the drone swarm communication topology model needs to meet the following requirements: Communication topology is a strongly connected balanced graph. And satisfy Represents the communication topology The Laplace matrix of express Multiply left by 1, 1 T Represents the transpose of a vector whose elements are all 1.

[0048] Step 2: Design a distributed resource scheduling algorithm for each drone swarm based on the dynamic event triggering method. The specific process is as follows:

[0049] Step 201: Update the state of the drone group using the derivative information of each drone group. The specific expression is as follows:

[0050]

[0051] in, represents the derivative of the state of the drone group i, β is the feedback gain and satisfies β>0, yes Time to The state prediction of drone group j at any moment, is the moment when drone swarm i is triggered for the kth time, yes Time to The state prediction of drone swarm i at any moment.

[0052] Step 202: Design an updating method for auxiliary variables based on dynamic errors; the specific expression is as follows:

[0053]

[0054] in, Derivative of the designed auxiliary variable, η i (t) represents the auxiliary variable of the design, represents the set of neighbors in drone group i, k i is the feedback gain and satisfies k i >0,δ i ∈[0,1] is an arbitrarily selected parameter, and m is the parameter to be designed.

[0055] Step 3: Use dynamic error to design a distributed event trigger controller and give controller parameters to achieve distributed resource scheduling. The specific expression for using dynamic error to design a distributed event trigger controller is as follows:

[0056]

[0057] in, represents the time when the drone group i is triggered for the k+1th time, inf represents the maximum infimum, m and θ i are the controller parameters to be designed.

[0058] The controller parameters are as follows:

[0059]

[0060] in yes The second norm of the matrix, yes The smallest non-zero eigenvalue of , σ∈(0,1).

[0061] Example:

[0062] The method of the present invention is verified by simulation experiments as follows: Considering that three drone groups need to perform their respective target tasks, the cost function of drone group 1 is The cost function of drone swarm 2 is The cost function of drone swarm 3 is The communication topology of the simulated drone swarm is as follows: Figure 2 As shown in Figure 1, the topology satisfies the equilibrium graph and is strongly connected. The initial values ​​of each drone group are set to x1(0) = 0.33, x2(0) = 0.33 and x3(0) = 0.34. Each drone group is updated according to the algorithm given in the present invention, and the results are as follows: Figure 3 、 Figure 4 、 Figure 5 、 Figure 6 As shown, Figure 3 It shows that the state index of each drone group converges to the optimal solution of the distributed resource scheduling model under equality constraints; Figure 4 This shows that the triggering moments of each drone group are discrete, which reduces the communication frequency of the system; Figure 5 This shows that during the algorithm iteration, the total UAV resources remain unchanged, that is, the equality constraint is established in real time; Figure 6 This indicates that the total system cost exponentially converges to the optimal value under the equality constraints. The overall experimental results demonstrate that each drone swarm achieves resource allocation under the conditions of constant total resources and discrete communication, and the algorithm converges exponentially. This demonstrates that the proposed distributed resource scheduling method for drone swarms based on dynamic event triggering can reduce communication bandwidth usage and improve system performance.

[0063] Therefore, the present invention adopts the above-mentioned distributed resource scheduling method of drone swarms based on dynamic event triggering. For the distributed resource scheduling problem of drone swarms with global equality constraints, it can exponentially converge to the optimal solution of the problem under discrete communication, greatly reducing the bandwidth occupancy of communication, reducing the probability of time delay and packet loss, and improving the performance of the system.

[0064] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit the same. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solutions of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A distributed resource scheduling method for drone swarms based on dynamic event triggering, characterized in that: The following steps are involved: Step 1: Based on the multi-agent system, a UAV swarm resource scheduling model and communication topology model with global equality constraints are constructed; Step 2: Design a distributed resource scheduling algorithm for each drone swarm based on the dynamic event triggering method; Step 3: Use dynamic error to design a distributed event trigger controller and give controller parameters to achieve distributed resource scheduling.

2. The method for distributed resource scheduling of drone swarms based on dynamic event triggering according to claim 1, characterized in that: The expression of the drone swarm resource scheduling model with global equality constraints constructed in step 1 is as follows: Where N is the number of drones in the swarm, x i Indicates the status of drone group i, representing the proportion of the number of drones assigned to drone group i to the total number of drones, x=(x1,x2,…,x N ) T Represents the resource allocation of each drone group, x satisfies the equality constraint The sum of all drone swarm resource allocations is 1, f i (x i ) is the cost function of the drone group i, which satisfies the strong convexity condition And f i (x i ) derivative is Lipschitz continuous: l0 and l are both positive numbers. represents the set of real numbers.

3. The method for distributed resource scheduling of drone swarms based on dynamic event triggering according to claim 2 is characterized in that: The drone swarm communication topology model in step 1 is described as follows: The UAV swarm communication topology model is regarded as a multi-agent network system consisting of N nodes, and the communication topology is recorded as in Represents a node set, an edge set (i,j)∈ε means that node j receives information from node i, and the information from node i1 to node i l The path is an ordered sequence of edges (i k ,i k+1 ), k=1,···,l-1; the set of neighbors within node i is picture The adjacency matrix Defined as a ii =0, if (j,i)∈ε, a ij =1, otherwise 0; The Laplace matrix of Defined as When i≠j, l ij =-a ij ; ε represents the edge, a ij Representation matrix The i-th row and j-th column element, l ij Representation matrix The element in row i and column j, l ii Representation matrix The element in row i and column i.

4. The method for distributed resource scheduling of drone swarms based on dynamic event triggering according to claim 3 is characterized in that: The UAV swarm communication topology model needs to meet the following requirements: Communication topology is a strongly connected balanced graph. And satisfy Represents the communication topology The Laplace matrix of express Multiply left by 1, 1 T Represents the transpose of a vector whose elements are all 1.

5. The method for distributed resource scheduling of drone swarms based on dynamic event triggering according to claim 4 is characterized in that: The specific process of designing a distributed resource scheduling algorithm for each drone swarm based on the dynamic event triggering method in step 2 is as follows: Step 201: Update the status of the drone group using the derivative information of each drone group; Step 202: Design an updating method for auxiliary variables based on dynamic errors.

6. The method for distributed resource scheduling of drone swarms based on dynamic event triggering according to claim 5, characterized in that: In step 201, the expression for updating the state of the drone group using the derivative information of each drone group is as follows: in, represents the derivative of the state of the drone group i, β is the feedback gain and satisfies β>0, yes Time to The state prediction of drone group j at any moment, is the moment when drone swarm i is triggered for the kth time, yes Time to The state prediction of drone swarm i at any moment.

7. The method for distributed resource scheduling of drone swarms based on dynamic event triggering according to claim 6, characterized in that: The expression for the updating method of the auxiliary variable based on the dynamic error design in step 202 is as follows: in, Denotes the derivative of the designed auxiliary variable, η i (t) represents the auxiliary variable of the design, represents the set of neighbors in the drone group i, k i is the feedback gain and satisfies k i >0,δ i ∈[0,1] is an arbitrarily selected parameter, and m is the parameter to be designed.

8. The method for distributed resource scheduling of drone swarms based on dynamic event triggering according to claim 7 is characterized in that: The specific expression for designing a distributed event-triggered controller using dynamic error in step 3 is as follows: in, represents the time when the drone group i is triggered for the k+1th time, inf represents the maximum infimum, m and θ i are the controller parameters to be designed.

9. The method for distributed resource scheduling of drone swarms based on dynamic event triggering according to claim 8, characterized in that: The controller parameters in step 3 are as follows: in yes The second norm of the matrix, yes The smallest non-zero eigenvalue of , σ∈(0,1).

Images