Cluster system formation serialization switching method

Abstract

The invention discloses a cluster system formation serialization switching method. The method comprises the following steps: defining a formation sequence which is composed of N formation stages and comprises a time domain and a state domain; wherein the time domain is composed of N continuous formation time domains, and each time domain has a convergence moment; a progressive controller prototype is designed based on the defined state domain, the controller prototype can achieve a progressive convergence effect of an expected formation, and the prototype provides a basis for subsequent design; synchronously designing a time mapping function kappa based on a defined time domain, constructing a predetermined time controller called as a composite formation protocol, and executing the composite formation protocol by cluster individuals in a formation forming section and a formation maintaining section; according to the method, scenes related to multiple formation types are used, formation can be completed according to specified convergence time accurately in each formation stage, and the composite formation protocol has robustness for the target real-time acceleration and environment uncertainty.

Description

Technical Field

[0001] The present invention relates to the field of computer science and control technology, and in particular to a cluster system formation serialization switching method. Background Art

[0002] Swarm system formation technology is being widely used in scenarios such as exploration, environmental monitoring, surveillance and interception because it can improve performance in aspects such as the quality of collected data and robustness in the face of random environments. In many occasions, the formation system needs to switch quickly and accurately in the set formation sequence according to external reactions or its own needs, that is, in a certain formation sequence, switch from one steady-state formation to another steady-state formation.

[0003] At present, formation technology regards each formation as a non-dynamic isolated task, and lacks a flexible transition method for switching formation sequences. In addition, existing work usually achieves gradual convergence and finite time convergence, which lacks controllability in the time dimension. For example, since the convergence time cannot be accurately set, the convergence time may be greater than the total duration of the formation. Summary of the invention

[0004] The purpose of the present invention is to provide a cluster system formation serialization switching method, by defining a formation sequence including a time domain and a state domain, designing a progressive controller prototype and a time mapping function, and using the progressive controller prototype and the time mapping function to compositely construct a predetermined time controller, so that cluster individuals execute the constructed composite formation protocol in the formation formation segment and the formation maintenance segment respectively, so as to be applicable to scenarios involving various formation types, so that each formation stage can complete the formation accurately according to the specified convergence time.

[0005] The present invention provides a specific technical solution as follows: a cluster system formation serialization switching method, comprising the following steps:

[0006] S1. Define a formation sequence consisting of N formation stages, wherein the formation sequence includes a time domain and a state domain;

[0007] Preferably, as a definition of the state domain, the model of the cluster system is described by second-order integrator dynamics, where ρ, η, ω is the state variable in the state domain model. The total number of individuals in the cluster system is n, and they are marked as 1, 2, ..., n in counterclockwise spatial order. The subscript i of the variable indicates that the state variable belongs to the cluster individual. The variable with the subscript i omitted indicates that the state variable belongs to the cluster system, that is, In addition, the target individual of the marked formation is 0, and the cluster system is connected internally through a ring topology graph G;

[0008] The state domain represents the convergence of the system state, and the time domain represents a clear limit on the formation convergence time of the current stage. For the formation situation in the two-dimensional plane space, the formation state satisfies the limiting conditions represented by the following four variables: Among them, ρ * =R and η * =0 n Respectively represent the relative distance and speed between the cluster system and the target 0, which is the radial restriction condition of the state domain, ω * =Ω1 n and They represent the angular distribution and angular velocity between the cluster system and the target 0, respectively, and are the angular limitation conditions of the state domain.

[0009] S2. The time domain of the formation sequence includes N consecutive formation time domains [t0,t s ), each time domain has a convergence time t0+τ, the formation time domain [t0,t s ) includes the formation formation phase [t0, t0+τ) and the formation maintenance phase [t0+τ, t s );

[0010] Preferably, the formation time domain [t0,t s ) is a complete formation process, which is divided into two sections according to the convergence time t0+τ, namely the formation section [t0, t0+τ) and the formation maintenance section [t0+τ, t s ).

[0011] S3. Based on the state domain in S1, a progressive controller prototype is designed, wherein the progressive controller prototype can achieve a progressive convergence effect to a desired formation;

[0012] Preferably, the designed progressive controller prototype is specifically as follows:

[0013]

[0014] in, and are the controller inputs in the local radial and angular dimensions, ω i is the angular velocity of individual i relative to the target, ρ i With η i is the relative distance and speed between individual i and target 0, is the radial real-time state of the state domain, and f i for The negative feedback regulation term, R i ,d i Respectively represent ρ i and The expected value of Ω represents the expected angular velocity of the cluster individuals around the target individual. and are the equivalent values ​​of the target acceleration on the x-axis and y-axis of the local coordinate system of individual i, respectively. d=[d1,d2,…,d n ] T , R=[R1,R2,…,R n ] T , control parameter k i >0(i=1,2,3,4).

[0015] S4, based on the time domain in S2, design a time mapping function k, use the progressive controller prototype and the time mapping function k to construct a predetermined time controller, and the predetermined time controller is a composite formation protocol;

[0016] Preferably, the designed progressive controller prototype is combined with the time mapping function κ, that is, based on the progressive controller prototype, the time mapping function is introduced Construct the following composite formation protocol:

[0017]

[0018] Among them, the formation time domain [t0,t s ) is divided into the formation segment [t0, t0+τ) and the formation maintenance segment [t0+τ, t s ), Δt is the total formation duration in the formation segment [t0, t0+τ), Δt=t-t0, express The square of represents the first-order derivative of κ with respect to time, represents the second-order derivative of κ with respect to time. The two-stage control protocol has different forms, corresponding to different functions;

[0019] At the convergence time t0+τ, there is a control protocol switch in the composite formation protocol, and the control parameter k i >0(i=1,2,3,4) is selected around the controller switching time t=t0+τ;

[0020] In the formation stage [t0, t0+τ), the composite formation protocol is applied to the dynamic system of the cluster individuals, and the compact closed-loop form of the cluster system is expressed by variables without subscripts, and the final system dynamic equation is obtained:

[0021]

[0022] Among them, ω + =[ω2,ω3,...,ωn ,ω1] T is ω=[ω1,ω2,ω3,...,ω n ] T The circular dislocation term, diag(·) represents the diagonal matrix of the vector, that is, the elements in the vector are sequentially used as the diagonal elements of the diagonal matrix, 1 n is an n-dimensional column vector whose elements are all 1, R = [R1, R1, ..., R n ] T and d=[d1,d1,...,d n ] T Represent the vector form of each expected value respectively.

[0023] S5, in the formation forming section [t0, t0+τ) and the formation maintaining section [t0+τ, t s ), the cluster individuals execute the composite formation protocol in S4.

[0024] Preferably, due to the characteristics of the time mapping function κ, the composite formation protocol does not involve the real-time acceleration a0(t) of the formation target, so by defining the acceleration a0(t) of the target individual as not directly measurable but bounded, that is:

[0025]

[0026] in, is a certain unknown upper bound. When the real-time acceleration a0(t) of the formation target satisfies the continuity condition, a0(t) can be changed arbitrarily.

[0027] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: the present invention defines a formation sequence including a time domain and a state domain, designs a progressive controller prototype based on the defined state domain, designs a time mapping function based on the defined time domain, constructs a predetermined time controller called a composite formation protocol, and the cluster individuals execute the composite formation protocol in the formation formation segment and the formation maintenance segment; the method is applicable to scenarios involving various formation types, and can enable each formation stage to complete the formation accurately according to the specified convergence time, and the composite formation protocol is robust to the target real-time acceleration and environmental uncertainty, the composite formation protocol can effectively and smoothly drive the cluster system to switch in the formation subsequence, the cluster system achieves various types of formation forms, and can perform dynamic formation switching in a given formation sequence as expected, the technical solution of the present invention is applicable to more practical engineering scenarios, such as the encirclement of high-maneuverable targets and the presence of random interference, and effectively fills the research gap in the formation technology of related cluster systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0029] Figure 1 A principle flow chart of a cluster system formation serialization switching method provided by an embodiment of the present invention;

[0030] Figure 2 A schematic diagram of a time mapping mechanism provided in an embodiment of the present invention;

[0031] Figure 3 A schematic diagram of a trajectory of a medium formation provided in an embodiment of the present invention;

[0032] Figure 4 A rendering of a surround experiment for a dynamic target provided by an embodiment of the present invention;

[0033] Figure 5 A result schematic diagram of the real-time value of the cluster system status provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0034] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0035] The present invention combines Figures 1 to 5 , providing the following technical solution: a cluster system formation serialization switching method, comprising:

[0036] S1. Define a formation sequence consisting of N formation stages, including the time domain and the state domain;

[0037] In this embodiment, combined with Figure 2 , Figure 2 A schematic diagram of a time mapping mechanism provided by an embodiment of the present invention, specifically, as follows Figure 2 As shown in the figure, as the definition of the state domain, the model of the cluster system is described by the second-order integrator dynamics, where ρ, η, ω is the state variable in the state domain model. The total number of cluster individuals is n, and they are marked as 1, 2, ..., n in counterclockwise spatial order. The surrounded target individual is 0. The cluster system is connected internally through a ring topology graph G(V, E), where V = {v1, ...v n} and E=V×V represent the point set and edge set respectively, and the edge e ij∈E means that cluster individuals i and j can transmit information to each other. In other words, G(V,E) is an undirected graph;

[0038] Furthermore, N i ={j|e ij ∈E,j≠i} represents the communication neighbor set of individual i, and stipulates that each individual i has only two neighbors N i ={i - ,i +},in:

[0039]

[0040] For example, the variable subscript i indicates that the state variable belongs to a cluster individual, and the variable without a subscript indicates that the state variable belongs to a cluster system, for example:

[0041] In this embodiment, the state domain represents the convergence of the system state, and the time domain clearly limits the formation convergence time at the current stage. For the formation situation in a two-dimensional plane space, a stable formation state must satisfy the requirements of stable distance and adjustable layout. This limiting condition can be represented by the following variables:

[0042]

[0043] Among them, ρ * =R and η * =0 n Respectively represent the relative distance and speed between the cluster system and the target 0, which is the radial restriction condition of the state domain, ω * =Ω1 n and They represent the angle distribution and angular velocity between the cluster system and the target 0, respectively, and are the angular constraints of the state domain;

[0044] Exemplarily, the problem of determining whether the formation restriction condition is satisfied is modeled as a control problem of the following cluster individual dynamic system:

[0045]

[0046] Among them, ρ i With η i is the relative distance and speed between individual i and target 0, is the radial real-time state of the state domain, Represents individual i and its forward neighbor i + The relative angle between i ,ω i+ With ω i- Then, individuals i and i + 、i -The angular velocity of the target rotation, and ω i Together they define the angular real-time state of the state domain. and are the equivalent values ​​of the target acceleration on the x-axis and y-axis of the local coordinate system of individual i, and is the controller input in the local radial and angular dimensions;

[0047] Preferably, Figure 2 The meaning of each state is further explained in . It should be pointed out that the equivalent system model combines the second-order integral dynamic system of the cluster individuals and the target individuals with the coordinate rotation characteristics of the cluster individuals, and has a beneficial completely local characteristic. The individual can obtain all state information only through measurement and calculation between neighbors in the moving local coordinate system, that is, the radial state ρ i , η i and angular state ω i , further communication with neighbors can obtain ω i+ and i- .

[0048] S2. The time domain of the formation sequence includes N consecutive formation time domains [t0,t s ), each time domain has a convergence moment t0+τ, which is the moment when the formation is completed;

[0049] In this embodiment, the formation time domain [t0,t s ) is a complete formation process, which is divided into two sections according to the convergence time t0+τ, namely the formation section [t0, t0+τ) and the formation maintenance section [t0+τ, t s ).

[0050] S3. Based on the state domain in S1, design a prototype of a progressive controller that can achieve a progressive convergence effect to the desired formation;

[0051] In this embodiment, a controller prototype with asymptotic convergence capability is designed to surround a target entity with dynamic acceleration a0(t):

[0052]

[0053] Among them, f i for The negative feedback regulation term, R i ,d i Respectively represent ρ i and , Ω represents the expected angular velocity of the cluster individuals around the target individual. For the convenience of expression, let d = [d1, d2, …, d n ] T , R=[R1,R2,…,R n ] T , control parameter k i >0(i=1,2,3,4).

[0054] S4, based on the time domain in S2, design the time mapping function κ;

[0055] In this embodiment, combined with Figure 3 , Figure 3 The schematic diagram of the mapping mechanism provided in this embodiment is as follows: Time-like mapping functions It has the following characteristics: κ(·):[0,τ)→[0,∞), monotonically increasing, and and

[0056] Exemplarily, the progressive controller prototype in S3 is compounded with the time mapping function κ to construct a scheduled time controller, i.e., a composite formation protocol, which allows formation time reservation in the time domain and flexible switching of formation states in the state domain, thereby realizing serialized switching of cluster system formations in a composite manner.

[0057] Specifically, based on the prototype of the progressive controller, the time mapping function is introduced Construct the following composite formation protocol:

[0058]

[0059] Among them, the formation time domain [t0,t s ) is divided into the formation segment [t0, t0+τ) and the formation maintenance segment [t0+τ, t s ), Δt is the total formation duration in the formation segment [t0, t0+τ), Δt=t-t0, express The square of represents the first-order derivative of κ with respect to time, represents the second-order derivative of κ with respect to time. The two-stage control protocols have different forms and correspond to different functions. At the convergence time t0+τ, there is a control protocol switch in the composite formation protocol. The control parameter k i The selection of >0(i=1,2,3,4) should ensure that the amplitude change of the control input should be as small as possible before and after the controller switching time t=t0+τ.

[0060] Furthermore, in the formation segment [t0, t0+τ), the composite formation protocol is applied to the dynamic system of the cluster individuals. If the compact closed-loop form of the cluster system is represented by variables without subscripts, the final system dynamic equation is obtained:

[0061]

[0062] Among them, ω + =[ω2,ω3,...,ω n ,ω1] T is ω=[ω1,ω2,ω3,...,ω n ] T The circular dislocation term; diag(·) represents the diagonal matrix of the vector, that is, the elements in the vector are taken as the diagonal elements of the diagonal matrix in turn, 1 n is an n-dimensional column vector whose elements are all 1, R = [R1, R1, ..., R n ] T and d=[d1,d1,...,d n ] T Represent the vector form of each expected value respectively.

[0063] S5, in the formation formation stage [t0, t0+τ) and the formation maintenance stage [t0+τ, t s ), the cluster individuals execute the composite formation protocol in S4;

[0064] In this embodiment, for the compact closed-loop system in step S4, if the control parameter k i >0(i=1,2,3,4) meets the conditions in step S4, then it is guaranteed that the cluster system can accurately achieve the desired formation within the predetermined time τ, including two aspects:

[0065] Exemplarily, in the formation segment [t0, t0+τ), the state of the cluster system converges to a state domain that satisfies the limited conditions:

[0066] When t→t0+τ

[0067] For example, in the formation maintenance phase [t0+τ,t s ), because there is a control protocol switch in the composite formation protocol, the cluster system can maintain the desired formation state and achieve a stable formation state.

[0068] Preferably, due to the characteristics of the time mapping function κ, the composite formation protocol does not involve the real-time acceleration a0(t) of the formation target. Therefore, the present invention only assumes that the acceleration a0(t) of the target individual cannot be directly measured but is bounded, that is:

[0069]

[0070] in, is a certain unknown upper bound. When the real-time acceleration a0(t) of the formation target satisfies the continuity condition, a0(t) can be changed arbitrarily, especially, it can be changed randomly. In other words, the composite formation protocol is robust to the target real-time acceleration and environmental uncertainty, and has beneficial generalization characteristics.

[0071] In this embodiment, combined Figure 4 , Figure 4 The effect of the surround experiment on dynamic targets is given. In this experiment, a formation target in uniform circular motion in the counterclockwise direction is simulated. The real-time acceleration of the formation target is unknown, and the system is commanded to perform dynamic formation switching after the formation converges. It is set that there are n=5 cluster individuals in the cluster system, and the individuals also move around the formation target in the counterclockwise direction. For the formation sequence, the cluster system is required to execute the following in sequence: 1) regular pentagon formation; 2) cross formation; 3) irregular pentagon formation. The total duration of each sequence stage is 20s. It is required that the formation transition process should be completed within τ=10s. For the time mapping function in the composite formation protocol, κ(t)=-ln(1-t / τ) is selected.

[0072] In this embodiment, combined Figure 5 , Figure 5 The real-time value of the cluster system status is given. The results show that the composite formation protocol can effectively and smoothly drive the cluster system to switch in the formation subsequence. The cluster system has achieved various types of formation forms, and subsequently performs dynamic formation switching in a given formation sequence as expected. In summary, in view of the research gap in the relevant cluster system formation technology, a flexible cluster system formation serialization switching method is proposed, which is suitable for more practical engineering scenarios, such as the encirclement of highly maneuverable targets and the presence of random interference.

[0073] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0074] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A cluster system formation serialization switching method, characterized by: The following steps are involved: S1. Define a formation sequence consisting of N formation stages, wherein the formation sequence includes a time domain and a state domain; S2. The time domain of the formation sequence includes N consecutive formation time domains [t0,t s ), each time domain has a convergence time t0+τ, the formation time domain [t0,t s ) includes the formation formation phase [t0, t0+τ) and the formation maintenance phase [t0+τ, t s ); S3. Based on the state domain in S1, a progressive controller prototype is designed, wherein the progressive controller prototype can achieve a progressive convergence effect to a desired formation; S4, based on the time domain in S2, design a time mapping function κ, use the progressive controller prototype and the time mapping function κ to construct a predetermined time controller, and the predetermined time controller is a composite formation protocol; S5, in the formation forming section [t0, t0+τ) and the formation maintaining section [t0+τ, t s ), the cluster individuals execute the composite formation protocol in S4.

2. A cluster system formation serialized switching method according to claim 1, characterized in that: The definition of the state domain includes: the model of the cluster system is described by the second-order integrator dynamics, where ρ, η, ω is the state variable in the state domain model. The total number of cluster individuals is n, and they are marked as 1, 2, ..., n in counterclockwise spatial order. The subscript i of the variable indicates that the state variable belongs to the cluster individual. The variable with the subscript i omitted indicates that the state variable belongs to the cluster system, that is, The state domain represents the convergence of the system state, and the time domain represents a clear limit on the formation convergence time of the current stage. For the formation situation in the two-dimensional plane space, the formation state satisfies the limiting conditions represented by the following four variables: Among them, ρ * =R and η * =0 n Respectively represent the relative distance and speed between the cluster system and the target 0, which is the radial restriction condition of the state domain, ω * =Ω1 n and They represent the angular distribution and angular velocity between the cluster system and the target 0, respectively, and are the angular limitation conditions of the state domain.

3. A cluster system formation serialized switching method according to claim 2, characterized in that: The convergence time t0+τ is the time when the formation is completed, and the formation formation section [t0, t0+τ) and the formation maintenance section [t0+τ, t s ) is divided according to the convergence time t0+τ, and the formation time domain [t0,t s ) is a complete formation process.

4. A cluster system formation serialized switching method according to claim 3, characterized in that: The prototype of the progressive controller designed is as follows: in, and are the controller inputs in the local radial and angular dimensions, ω i is the angular velocity of individual i relative to the target, ρ i With η i is the relative distance and speed between individual i and target 0, is the radial real-time state of the state domain, and f i for The negative feedback regulation term, R i ,d i Respectively represent ρ i and The expected value of Ω represents the expected angular velocity of the cluster individuals around the target individual. and are the equivalent values ​​of the target acceleration on the x-axis and y-axis of the local coordinate system of individual i, respectively. d=[d1,d2,…,d n ] T , R=[R1,R2,…,R n ] T , control parameter k i >0(i=1,2,3,4).

5. A cluster system formation serialized switching method according to claim 4, characterized in that: The design time mapping function κ includes: selecting Time-like mapping functions It has the following characteristics: κ(·):[0,τ)→[0,∞), monotonically increasing, and and 6. A cluster system formation serialized switching method according to claim 5, characterized in that: The construction of the predetermined time controller specifically includes: combining the designed progressive controller prototype with the time mapping function κ, that is, introducing the time mapping function on the basis of the progressive controller prototype Construct the following composite formation protocol: Among them, the formation time domain [t0,t s ) is divided into the formation segment [t0, t0+τ) and the formation maintenance segment [t0+τ, t s ), Δt is the total formation duration in the formation segment [t0, t0+τ), Δt=t-t0, express The square of represents the first-order derivative of κ with respect to time, represents the second-order derivative of κ with respect to time. The two-stage control protocol has different forms, corresponding to different functions; At the convergence time t0+τ, there is a control protocol switch in the composite formation protocol, and the control parameter k i >0(i=1,2,3,4) is selected around the controller switching time t=t0+τ; In the formation stage [t0, t0+τ), the composite formation protocol is applied to the dynamic system of the cluster individuals, and the compact closed-loop form of the cluster system is expressed by variables without subscripts, and the final system dynamic equation is obtained: Among them, ω + =[ω2,ω3,...,ω n ,ω1] T is ω=[ω1,ω2,ω3,...,ω n ] T The circular dislocation term, diag(·) represents the diagonal matrix of the vector, that is, the elements in the vector are sequentially used as the diagonal elements of the diagonal matrix, 1 n is an n-dimensional column vector whose elements are all 1, R = [R1, R1, ..., R n ] T and d=[d1,d1,...,d n ] T Represent the vector form of each expected value respectively.

7. A cluster system formation serialized switching method according to claim 6, characterized in that: The cluster individuals execute the composite formation protocol specifically including: when the control parameter k i >0(i=1,2,3,4), the cluster system can achieve the desired formation within the predetermined time τ, including two aspects: In the formation stage [t0, t0+τ), the state of the cluster system converges to the state domain that satisfies the restricted conditions: When t→t0+τ in the formation maintenance phase [t0+τ,t s ), because there is a control protocol switch in the composite formation protocol, the cluster system maintains the expected formation state and reaches a stable formation state.

8. A cluster system formation serialized switching method according to claim 7, characterized in that: The cluster individuals execute the composite formation protocol in S4 and also include the following conditions: Due to the characteristics of the time mapping function κ, the composite formation protocol does not involve the real-time acceleration a0(t) of the formation target, so by defining the acceleration a0(t) of the target individual, it cannot be directly measured but is bounded, that is: in, is a certain unknown upper bound. When the real-time acceleration a0(t) of the formation target satisfies the continuity condition, a0(t) can be changed arbitrarily.

Images