Iterative learning based heterogeneous unmanned cluster formation control method, system and device

Abstract

The application provides a heterogeneous unmanned cluster formation control method, system and device based on iterative learning, wherein the heterogeneous unmanned cluster formation control method based on iterative learning comprises the following steps: constructing a communication topology directed graph of a leading unmanned vehicle and a communication topology directed graph of a following unmanned aerial vehicle; constructing a reference trajectory discrete iterative model of the leading unmanned vehicle and the following unmanned aerial vehicle respectively, and determining a distributed control strategy of the reference trajectory of the leading unmanned vehicle and the following unmanned aerial vehicle; constructing an unmanned vehicle discrete kinematics model, and determining the control input of the linear velocity and the angular velocity of each leading unmanned vehicle according to the formation error of the reference trajectory of the leading unmanned vehicle. Through the application, the different flexibility of the UGV and the UAV and the different accuracy requirements of the reference trajectory are considered, and the problem of poor control stability existing in the prior art is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of unmanned cluster formation, and particularly relates to a heterogeneous unmanned cluster formation control method and device based on iterative learning and electronic equipment. BACKGROUND

[0002] Heterogeneous unmanned aerial vehicles (UAVs) and unmanned ground vehicles (UGVs) have been widely applied in military and civilian fields such as cooperative mapping, cooperative search or tracking, and surveillance tasks due to their complementary advantages such as wider field of view and cooperative endurance compared with homogeneous UAVs or UGVs. However, current researches on heterogeneous unmanned systems mostly focus on the cooperation between a single UAV and a single UGV. The interaction between heterogeneous multi-UAVs and multi-UGVs makes the function and control of the heterogeneous unmanned cluster system more complex, which brings many challenges to the formation control of the heterogeneous unmanned cluster. Due to the size and weight limitations of unmanned systems, the power of the communication module carried is limited and difficult to support large-scale formation. In order to make up for the lack of communication capability of unmanned equipment, it is necessary to study the distributed heterogeneous formation method.

[0003] The formation of heterogeneous UAVs and UGVs needs to consider two key problems, one is the heterogeneous dimension of UAVs and UGVs, and the other is the nonlinear heterogeneous dynamics. For the formation of UAVs and UGVs, a common solution is a hierarchical distributed control framework, in which the upper layer generates a reference trajectory for each UAV and UGV, and the lower layer designs a formation controller to make the UAVs and UGVs track the reference trajectory. The existing method does not consider the different sampling periods of UAVs and UGVs when generating the reference trajectory, and does not further update the reference trajectory within the sampling interval. At the same time, the existing heterogeneous formation controller design mostly considers heterogeneous multi-agent systems, ignoring the nonlinear dynamics of each UAV and UGV. The update lag of the reference trajectory and the neglect of the dynamics model will affect the formation control accuracy of UAVs and UGVs to some extent. Therefore, when studying the formation control of heterogeneous unmanned clusters, it is necessary to use an iterative learning method with different time scales in the upper layer to generate a reference trajectory, and to consider the flexibility and maneuverability of UAVs and UGVs in the lower layer formation control and model the specific physical characteristics.

[0004] There is no effective solution to the problem of poor control stability in the prior art. SUMMARY

[0005] The present application provides a heterogeneous unmanned cluster formation control method, system and device based on iterative learning, to solve the defect of poor control stability in the prior art.

[0006] The first aspect provides a heterogeneous unmanned cluster formation control method based on iterative learning, comprising:

[0007] constructing a communication topology directed graph of the leading unmanned vehicle and a communication topology directed graph of the follower unmanned aerial vehicle;

[0008] constructing a reference trajectory discrete iterative model of the leading unmanned vehicle and the follower unmanned aerial vehicle respectively, and determining a distributed control strategy of the reference trajectory of the leading unmanned vehicle and the follower unmanned aerial vehicle;

[0009] constructing a discrete kinematics model of the unmanned vehicle, and determining a control input of the linear velocity and the angular velocity of each leading unmanned vehicle according to a formation error of the reference trajectory of the leading unmanned vehicle;

[0010] constructing a discrete kinematics model of the unmanned aerial vehicle, and generating a formation error discrete sliding mode function of the leading unmanned vehicle and the follower unmanned aerial vehicle at different time scales to determine a control input of the reference velocity of the follower unmanned aerial vehicle;

[0011] constructing a small time scale unmanned aerial vehicle velocity discrete model, and determining a control input of the small time scale acceleration of the follower unmanned aerial vehicle.

[0012] According to the heterogeneous unmanned cluster formation control method based on iterative learning, the communication topology directed graph of the leading unmanned vehicle is constructed, comprising:

[0013] defining a set of unmanned vehicles of the leading unmanned vehicle and a virtual leading unmanned vehicle; the virtual leading unmanned vehicle is a leader of the leading unmanned vehicle;

[0014] generating the communication topology directed graph of the leading unmanned vehicle based on a communication relationship between the leading unmanned vehicles.

[0015] According to the heterogeneous unmanned cluster formation control method based on iterative learning, the communication topology directed graph of the follower unmanned aerial vehicle is constructed, comprising:

[0016] for each leading unmanned vehicle, defining a set of unmanned aerial vehicles in the unmanned cluster corresponding to the leading unmanned vehicle, and setting the leading unmanned vehicle as a leader of the unmanned cluster;

[0017] generating the communication topology directed graph of the follower unmanned aerial vehicle based on a communication relationship between the follower unmanned aerial vehicles.

[0018] According to the heterogeneous unmanned cluster formation control method based on iterative learning, the reference trajectory discrete iterative model of the leading unmanned vehicle and the follower unmanned aerial vehicle is constructed respectively, and the distributed control strategy of the reference trajectory of the leading unmanned vehicle and the follower unmanned aerial vehicle is determined, comprising:

[0019] determining the iteration time interval and the iteration number of the reference trajectories of the leading unmanned vehicle and the follower unmanned aerial vehicle in the same time period;

[0020] determining the iteration number of the reference trajectories of the leading unmanned vehicle and the follower unmanned aerial vehicle in the same time period according to the event point of the alignment reference trajectory control input update;

[0021] constructing the discrete iteration model of the reference trajectories of the leading unmanned vehicle and the follower unmanned aerial vehicle, and determining the distributed control strategy of the reference trajectories of the leading unmanned vehicle and the follower unmanned aerial vehicle;

[0022] obtaining the value range of the iteration number ratio of the reference trajectories of the leading unmanned vehicle and the follower unmanned aerial vehicle in the same time interval according to Lyapunov stability theory.

[0023] According to the present application, a heterogeneous unmanned cluster formation control method based on iteration learning is provided, a discrete kinematic model of an unmanned vehicle is constructed, and the control input of the linear velocity and the angular velocity of each leading unmanned vehicle is determined according to the formation error of the reference trajectory of the leading unmanned vehicle, including:

[0024] determining the discrete kinematic model of the unmanned vehicle based on the displacement and the yaw angle of the leading unmanned vehicle;

[0025] determining the formation error of the reference trajectory of the leading unmanned vehicle based on the expected displacement, the expected yaw angle and the expected formation distance of the leading unmanned vehicle;

[0026] combining the formation error of the reference trajectory of the leading unmanned vehicle, and designing the control input of the linear velocity and the angular velocity of the formation controller of the leading unmanned vehicle.

[0027] According to the present application, a heterogeneous unmanned cluster formation control method based on iteration learning is provided, a discrete kinematic model of an unmanned vehicle is constructed, and the formation error discrete sliding mode function of the leading unmanned vehicle and the follower unmanned aerial vehicle in different time scales is generated, including:

[0028] constructing the discrete kinematic model of the unmanned vehicle based on the displacement of the follower unmanned aerial vehicle;

[0029] taking the iteration time interval of the upper reference trajectory as the time scale of the formation in which the follower unmanned aerial vehicle is located, and designing the formation error discrete sliding mode function of the leading unmanned vehicle and the follower unmanned aerial vehicle containing different time scales.

[0030] According to the present application, a heterogeneous unmanned cluster formation control method based on iteration learning is provided, the control input of the reference velocity of the follower unmanned aerial vehicle is determined, including:

[0031] determining the control input of the follower unmanned aerial vehicle based on the stability requirement of the discrete sliding mode function.

[0032] The range of values of the sliding mode gain parameter and the ratio of the formation time scale is obtained according to Lyapunov stability analysis.

[0033] The application provides a heterogeneous unmanned cluster formation control method based on iterative learning, a small-time-scale unmanned vehicle speed discrete model is constructed, and control input of small-time-scale acceleration of a follower unmanned vehicle is determined.

[0034] A small-time-scale unmanned vehicle speed discrete model is constructed according to actual speed of the follower unmanned vehicle.

[0035] A sliding mode function of a speed loop is constructed according to an error between the actual speed of the follower unmanned vehicle and a reference speed.

[0036] Control input of small-time-scale acceleration of the follower unmanned vehicle is determined according to convergence requirements of a sliding surface.

[0037] In a second aspect, the application further provides a heterogeneous unmanned cluster formation control system based on iterative learning, comprising:

[0038] A graph construction module is configured to construct a leading unmanned vehicle communication topology directed graph and a follower unmanned vehicle communication topology directed graph.

[0039] A strategy making module is configured to construct reference trajectory discrete iterative models of the leading unmanned vehicle and the follower unmanned vehicle respectively, and determine a distributed control strategy of the reference trajectories of the leading unmanned vehicle and the follower unmanned vehicle.

[0040] An unmanned vehicle control module is configured to construct an unmanned vehicle discrete kinematic model, and determine control input of linear speed and angular speed of each leading unmanned vehicle according to formation error of the reference trajectory of the leading unmanned vehicle.

[0041] An unmanned vehicle control module is configured to construct an unmanned vehicle discrete kinematic model, and generate formation error discrete sliding mode functions of the leading unmanned vehicle and the follower unmanned vehicle in different time scales, and determine control input of the reference speed of the follower unmanned vehicle.

[0042] A target control module is configured to construct a small-time-scale unmanned vehicle speed discrete model, and determine control input of small-time-scale acceleration of the follower unmanned vehicle.

[0043] In a third aspect, the application further provides a heterogeneous unmanned cluster formation control device based on iterative learning, comprising:

[0044] A construction module is configured to construct a leading unmanned vehicle communication topology directed graph and a follower unmanned vehicle communication topology directed graph.

[0045] The processing module is configured to construct a reference trajectory discrete iteration model of the leading unmanned vehicle and the following unmanned aerial vehicle respectively, and determine a distributed control strategy of the reference trajectory of the leading unmanned vehicle and the following unmanned aerial vehicle.

[0046] The first control module is configured to construct a discrete kinematics model of the unmanned vehicle, and determine a control input of linear velocity and angular velocity of each leading unmanned vehicle according to a formation error of the reference trajectory of the leading unmanned vehicle.

[0047] The second control module is configured to construct a discrete kinematics model of the unmanned aerial vehicle, and generate a formation error discrete sliding mode function of the leading unmanned vehicle and the following unmanned aerial vehicle at different time scales, and determine a control input of the reference velocity of the following unmanned aerial vehicle.

[0048] The third control module is configured to construct a small time scale discrete model of the unmanned aerial vehicle velocity, and determine a control input of small time scale acceleration of the following unmanned aerial vehicle.

[0049] In a fourth aspect, the present application further provides an electronic device, which comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the iteration learning based heterogeneous unmanned cluster formation control method according to the first aspect when executing the program.

[0050] In a fifth aspect, the present application further provides a non-transitory computer readable storage medium, which stores a computer program executable by a processor to implement the iteration learning based heterogeneous unmanned cluster formation control method according to the first aspect.

[0051] In a sixth aspect, the present application further provides a computer program product, which comprises a computer program executable by a processor to implement the iteration learning based heterogeneous unmanned cluster formation control method according to the first aspect.

[0052] Compared with the prior art, the present application has the following beneficial effects:

[0053] The application provides a heterogeneous unmanned cluster formation control method based on iterative learning, which generates a reference trajectory of itself through iterative learning in a small time scale, and only updates the distributed control input according to the iterative learning result at a fixed time interval. In the upper reference trajectory iterative learning tracking control and the lower heterogeneous formation control, the different time scale responses of the UGV and the UAV are considered. Considering the different flexibility of the UGV and the UAV and the different accuracy requirements of the reference trajectory, the reference trajectory iteration and the formation control input of the UGV have a large time scale, and the reference trajectory iteration and the formation control input of the UAV have a small time scale. Unlike existing methods, this time scale separation method is more in line with the actual characteristics of the UGV and the UAV in the heterogeneous formation, and solves the problem of poor control stability in the related art. BRIEF DESCRIPTION OF DRAWINGS

[0054] In order to more clearly illustrate the technical solutions of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings described below are some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0055] Figure 1 is a flowchart of the heterogeneous unmanned cluster formation control method based on iterative learning provided by the application;

[0056] Figure 2 is a schematic diagram of the communication topology structure of the heterogeneous unmanned cluster in the embodiment of the application;

[0057] Figure 3 is a schematic diagram of the process of unmanned cluster formation control in the embodiment of the application;

[0058] Figure 4 is a schematic diagram of the heterogeneous unmanned cluster built in the embodiment of the application;

[0059] Figure 5 is a schematic diagram of the upper reference trajectory of the heterogeneous unmanned cluster formation control method based on iterative learning in the embodiment of the application;

[0060] Figure 6 is a schematic diagram of the tracking error of the reference trajectory in the embodiment of the application;

[0061] Figure 7 is a schematic diagram of the formation trajectory of the lead unmanned vehicle and the follower unmanned aerial vehicle in the heterogeneous cluster in the embodiment of the application;

[0062] Figure 8 is a schematic diagram of the formation trajectory of three lead unmanned vehicles in the embodiment of the application;

[0063] Figure 9 is a curve diagram of the angle change of the unmanned vehicle formation in the embodiment of the application;

[0064] Figure 10 is a displacement error schematic diagram of the unmanned vehicle formation in the embodiment of the application;

[0065] Figure 11 is an angle error schematic diagram of the unmanned vehicle formation in the embodiment of the application;

[0066] Figure 12 is a displacement error of the unmanned aerial vehicle formation in the embodiment of the application;

[0067] Figure 13 is the reference speed and speed tracking error of the unmanned aerial vehicle formation in the X-axis direction in the embodiment of the application;

[0068] Figure 14 is the reference speed and speed tracking error of the unmanned aerial vehicle formation in the Y-axis direction in the embodiment of the application;

[0069] Figure 15 is a structural block diagram of the heterogeneous unmanned cluster formation control device based on iterative learning provided by the application;

[0070] Figure 16 is a structural schematic diagram of the electronic device provided by the application. DETAILED DESCRIPTION

[0071] To make the objectives, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be described clearly and completely below with reference to the drawings in the present application. Obviously, the described embodiments are some embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the present application.

[0072] The present application provides a heterogeneous unmanned cluster formation control method based on iterative learning, Figure 1 is a flowchart of the heterogeneous unmanned cluster formation control method based on iterative learning provided by the application, as Figure 1 shown, the method comprises the following steps:

[0073] Step S101, constructing a communication topology directed graph of the leading unmanned vehicle and a communication topology directed graph of the following unmanned vehicle;

[0074] Step S102, constructing a reference trajectory discrete iterative model of the leading unmanned vehicle and the following unmanned vehicle respectively, and determining a distributed control strategy of the reference trajectory of the leading unmanned vehicle and the following unmanned vehicle;

[0075] Step S103, a discrete kinematic model of the unmanned vehicle is constructed, and control input of linear velocity and angular velocity of each lead unmanned vehicle is determined according to the formation error of the reference trajectory of the lead unmanned vehicle;

[0076] Step S104, a discrete kinematic model of the unmanned aerial vehicle is constructed, and a formation error discrete sliding mode function of the lead unmanned vehicle and the follower unmanned aerial vehicle in different time scales is generated, and control input of the reference velocity of the follower unmanned aerial vehicle is determined;

[0077] Step S105, a small-time-scale discrete model of the unmanned aerial vehicle velocity is constructed, and control input of the small-time-scale acceleration of the follower unmanned aerial vehicle is determined.

[0078] In the method, first, a lead unmanned vehicle communication topology and a follower unmanned aerial vehicle cluster communication topology directed graph are constructed. In the lead unmanned vehicle communication topology directed graph, the lead unmanned vehicle is regarded as a graph node, and the communication relationship between different lead unmanned vehicles can be clearly represented. In the follower unmanned aerial vehicle communication topology directed graph, the follower unmanned aerial vehicle is regarded as a graph node, and the communication relationship between different follower unmanned aerial vehicles can be clearly represented. Then, a reference trajectory discrete iterative model of the lead unmanned vehicle and the follower unmanned aerial vehicle is constructed respectively, and a distributed control strategy of the reference trajectory of the lead unmanned vehicle and the follower unmanned aerial vehicle is determined. Then, a discrete kinematic model of the unmanned vehicle is constructed, and control input of linear velocity and angular velocity of each lead unmanned vehicle is determined according to the formation error of the reference trajectory of the lead unmanned vehicle. Then, a discrete kinematic model of the unmanned aerial vehicle is constructed, and a formation error discrete sliding mode function of the lead unmanned vehicle and the follower unmanned aerial vehicle in different time scales is generated, and control input of the reference velocity of the follower unmanned aerial vehicle is determined. Finally, a small-time-scale discrete model of the unmanned aerial vehicle velocity is constructed, and control input of the small-time-scale acceleration of the follower unmanned aerial vehicle is determined.

[0079] In the above process, the reference trajectory of itself is generated through iterative learning in a small time scale, and the distributed control input is updated only at a fixed time interval according to the iterative learning result. In the upper reference trajectory iterative learning tracking control and the lower heterogeneous formation control, the different time scale responses of the UGV and the UAV are considered. Considering the different flexibility of the UGV and the UAV and the different accuracy requirements of the reference trajectory, the reference trajectory iteration and the formation control input of the UGV have a large time scale, and the reference trajectory iteration and the formation control input of the UAV have a small time scale. Unlike existing methods, this time scale separation method is more in line with the actual characteristics of the UGV and the UAV in the heterogeneous formation, and solves the problem of poor control stability existing in the related art.

[0080] The method is described below with specific examples, Figure 2 is a schematic diagram of a heterogeneous unmanned cluster formation communication topology structure in an embodiment of the application, as Figure 2As shown, 3 unmanned cluster platforms are carried, each unmanned cluster contains a leading unmanned vehicle and several follower unmanned vehicles. According to the specific process shown in the figure, the method is actually verified, Figure 3 Figure 3 is a process diagram for unmanned cluster formation control in the embodiment of the present application. Only UGV1 in the unmanned vehicle formation can communicate with the virtual leading unmanned vehicle to obtain the motion trajectory parameters required to be tracked, UGV2 communicates with UGV1, and UGV3 communicates with UGV2. Only one follower unmanned vehicle in each unmanned vehicle cluster can communicate with the leading unmanned vehicle. The iteration time interval T of the leading unmanned vehicle is set to 0.05s, the iteration time interval τ of the follower unmanned vehicle is set to 0.01s, T * = 2, τ * = 10. The upper reference trajectory control gain = 0.05, = 0.01, the unmanned vehicle formation control gain k1 = 0.6, k2 = 8, k3 = 6, the unmanned vehicle formation control gain k4 = 0.001, k5 = 0.001, = 0.45, = 0.6, the speed loop control k6 = 0.0005, = 0.005, = 0.6. The expected formation distance of the unmanned vehicle is set to = 0, =0, =-3, = 4, =-3, =-4. The expected formation distance of the unmanned vehicle is set to = 0, = 0, = 2.5, =-1.5, = 0, = 2.5, =-1.5, =2.5, = 2.5, =-3, = 4, = 2.5, =-4.5, = 5.5, = 2.5, =-3, =-4, = 2.5, =-4.5, =-5.5, = 2.5,​ =-4.5, =-2.5, = 2.5.

[0081] In some embodiments, step S101, constructing a directed graph of the communication topology of the navigating autonomous vehicle, includes: defining a set of autonomous vehicles and a virtual navigating autonomous vehicle; the virtual navigating autonomous vehicle is the navigator of the navigating autonomous vehicles; and generating a directed graph of the communication topology of the navigating autonomous vehicle based on the communication relationships between the navigating autonomous vehicles.

[0082] For example, a set of autonomous vehicles is defined, with the following formula:

[0083]

[0084] in, Indicates a collection of driverless cars. m This represents the number of the lead autonomous vehicles. The lead autonomous vehicle is assigned the number 0. If each lead autonomous vehicle in the set is considered a node in a directed graph of the lead autonomous vehicle communication topology, then the directed graph of the lead autonomous vehicle communication topology is defined as follows:

[0085]

[0086] in, This represents a directed graph representing the communication topology of the autonomous vehicle. Represents the set of vertices. Describe the set of edges. This represents the communication relationship between the k-th and j-th autonomous vehicles. Represents the adjacency matrix. This indicates that the matrix has both row and column dimensions of m. If the k-th navigator vehicle can receive the first... j Information about the first autonomous vehicle is referred to as the first... j The lead vehicle is the communication neighbor of the k-th lead vehicle and ,otherwise .definition This represents the set of communication neighbors of the k-th navigating autonomous vehicle. m This represents the number of lead autonomous vehicles. If the k-th lead autonomous vehicle can receive information from the virtual lead autonomous vehicle, then... ,otherwise .

[0087] Correspondingly, a directed graph of the communication topology of the following drones is constructed, including: for each unmanned cluster corresponding to the lead drone, defining the set of drones in the unmanned cluster and setting the lead drone as the leader of the unmanned cluster; and generating a directed graph of the communication topology of the following drones based on the communication relationship between the following drones.

[0088] For example, the set of follower UAVs in the heterogeneous cluster (i.e., the heterogeneous cluster k) where each lead UAV k is located is defined as:

[0089]

[0090] wherein, denotes the set of follower UAVs in the heterogeneous cluster k, n denotes the number of follower UAVs in the heterogeneous cluster k, and lead UAV k is the leader of the heterogeneous cluster. If each follower UAV in the set of follower UAVs is regarded as a node in the follower UAV communication topology directed graph, the follower UAV communication topology directed graph of the heterogeneous cluster k is defined as follows:

[0091]

[0092] wherein, denotes the follower UAV communication topology directed graph, denotes the set of vertices, denotes the set of edges, denotes the communication relationship between the i-th follower UAV and the j-th follower UAV, i denotes the adjacency matrix, j denotes the matrix with row and column dimensions both being n, and if the i-th follower UAV can receive information from the j-th follower UAV, it is said that the i-th follower UAV is a communication neighbor of the j-th follower UAV, and , otherwise . It is defined that i denotes the set of communication neighbors of the i-th follower UAV, and n denotes the number of follower UAVs, if the i-th follower UAV can receive information from the lead UAV k of the heterogeneous cluster, then j , otherwise i . i i

[0093] In this embodiment, both the lead UAV communication topology directed graph and the follower UAV communication topology directed graph are distributed topology directed graphs. The motion trajectory of the virtual lead UAV is generated by the system, only one lead UAV in the set of lead UAVs communicates with the virtual lead UAV, and the remaining lead UAVs only interact with their neighbor lead UAVs. Only one follower UAV in the set of follower UAVs communicates with the lead UAV of the heterogeneous cluster, and the remaining follower UAVs only interact with their neighbor follower UAVs.

[0094] ​​​​​​​In some embodiments, step S102, respectively constructing the reference trajectory discrete iteration model of the leading unmanned vehicle and the following unmanned vehicle, and determining the distributed control strategy of the reference trajectory of the leading unmanned vehicle and the following unmanned vehicle, comprising: respectively determining the discrete time sampling period and the iteration time interval of the reference trajectory of the leading unmanned vehicle and the following unmanned vehicle; aligning the event points of the reference trajectory control input update, determining the iteration times of the reference trajectory of the leading unmanned vehicle and the following unmanned vehicle in the same time period; constructing the reference trajectory discrete iteration model of the leading unmanned vehicle and the following unmanned vehicle, and determining the distributed control strategy of the reference trajectory of the leading unmanned vehicle and the following unmanned vehicle; obtaining the value range of the iteration times ratio of the reference trajectory of the leading unmanned vehicle and the following unmanned vehicle in the same time interval according to Lyapunov stability theory.

[0095] In the present embodiment, the discrete time sampling period is , the iteration times of each leading unmanned vehicle itself in the time interval is , the time interval of each iteration is , the iteration times of each following unmanned vehicle itself is , and the time interval of each iteration is .

[0096] The discrete iteration model update process of the reference trajectory of the leading unmanned vehicle is as follows:

[0097]

[0098] wherein, is the reference information of the displacement of the X-axis direction, the displacement of the Y-axis direction, the yaw angle, the linear velocity and the angular velocity of the kth leading unmanned vehicle. and respectively represent the values at the time points of and .

[0099] is the control input of the reference trajectory of the kth leading unmanned vehicle, and is specifically as follows:

[0100]

[0101] wherein, represents the control gain, represents the displacement of the X-axis direction, the displacement of the Y-axis direction, the yaw angle, the linear velocity and the angular velocity information of the virtual leading unmanned vehicle, represents the communication neighbor set of the kth leading unmanned vehicle, represents the reference information of the kth leading unmanned vehicle at the time point of j , and represents the reference information of the kth leading unmanned vehicle at the time point of represents the reference information of the kth leading unmanned vehicle at the time point of​i exist Reference information for the moment, Indicates that the virtual navigation driverless car is in Information about the time.

[0102] The discrete iterative model update process following the UAV reference trajectory is as follows:

[0103]

[0104] in, Indicates following the drone i Reference information for displacement in the X-axis direction, displacement in the Y-axis direction, yaw angle, linear velocity, and angular velocity, respectively; and They represent exist and The value at any given moment.

[0105] Indicates following the drone i The control inputs for the reference trajectory are as follows:

[0106]

[0107] in, Indicates control gain. Indicates the first i A collection of communication neighbors that follow the drone. Indicates following the drone j exist Reference information for the moment, Indicates following the drone i exist Reference information for the moment, Indicates that the autonomous vehicle K is in Reference information for the moment, Indicates following the drone i exist Reference information for the moment.

[0108] To ensure the convergence of the reference trajectory iterations, the number of iterations T for the navigating autonomous vehicle is [number]. * and the number of iterations following the drone itself The range of values ​​for is:

[0109]

[0110] in, L The Laplace matrix represents the communication topology of autonomous vehicle platoons. , The Laplace matrix represents the communication topology of the k-th UAV swarm.B represents a communication weight matrix of the unmanned vehicle formation and the virtual leader unmanned vehicle, , represents a communication weight matrix of the kth unmanned vehicle cluster and the cluster leader unmanned vehicle; and respectively represent the minimum eigenvalue and the maximum eigenvalue of the matrix, represents a Laplacian matrix of the unmanned vehicle formation communication topology, represents a communication weight matrix of the unmanned vehicle formation and the virtual leader unmanned vehicle.

[0111] In some embodiments, step S103 of constructing the unmanned vehicle discrete kinematics model, determining the linear velocity and angular velocity control input of each leader unmanned vehicle according to the formation error of the leader unmanned vehicle reference trajectory, includes: determining the unmanned vehicle discrete kinematics model based on the displacement and yaw angle of the leader unmanned vehicle; determining the formation error of the leader unmanned vehicle reference trajectory based on the expected displacement, expected yaw angle and expected formation distance of the leader unmanned vehicle; combining the formation error of the leader unmanned vehicle reference trajectory, designing the linear velocity and angular velocity control input of the formation controller of the leader unmanned vehicle.

[0112] For example, the specific formula of the unmanned vehicle discrete kinematics model is as follows:

[0113]

[0114] wherein, represents the X-axis displacement of the kth leader unmanned vehicle, represents the Y-axis displacement of the kth leader unmanned vehicle, represents the yaw angle of the kth leader unmanned vehicle, represents the linear velocity of the leader unmanned vehicle, represents the angular velocity of the leader unmanned vehicle, T represents the iteration time interval of the leader unmanned vehicle, represents the iteration number of the leader unmanned vehicle itself.

[0115] The specific formula of the unmanned vehicle formation error is as follows:

[0116]

[0117] wherein, , and respectively represent the formation error of the kth leader unmanned vehicle in the X-axis direction, the Y-axis direction and the yaw angle; , and respectively represent the expected displacement of the kth leader unmanned vehicle in the X-axis direction, the expected displacement of the kth leader unmanned vehicle in the Y-axis direction and the expected yaw angle of the kth leader unmanned vehicle; and respectively represent the desired formation distance in the X-axis direction and the Y-axis direction. The line speed and angular velocity inputs of the unmanned vehicle formation controller are designed as follows:

[0118]

[0119]

[0120] wherein, represents the line speed input, represents the angular velocity input, k 1, k 2 and k 3 are control gains, and for the convergence of the formation controller, the value range is:

[0121]

[0122] wherein, T represents the iteration time interval of the lead unmanned vehicle, represents the iteration number of the lead unmanned vehicle itself.

[0123] In some embodiments, the step S104 of constructing the unmanned vehicle discrete kinematics model and generating the formation error discrete sliding mode function of the lead unmanned vehicle and the follower unmanned vehicle in different time scales comprises: constructing the unmanned vehicle discrete kinematics model based on the displacement of the follower unmanned vehicle; taking the iteration time interval of the upper reference trajectory as the time scale of the formation in which the follower unmanned vehicle is located, and designing the formation error discrete sliding mode function of the lead unmanned vehicle and the follower unmanned vehicle containing different time scales.

[0124] On this basis, the control input of the reference speed of the follower unmanned vehicle is determined, comprising: determining the control input of the follower unmanned vehicle based on the stability requirement of the discrete sliding mode function; and obtaining the value range of the sliding mode gain parameter and the ratio of the formation time scale according to Lyapunov stability analysis.

[0125] For example, the unmanned vehicle discrete kinematics model is constructed as follows:

[0126]

[0127] wherein, , and respectively represent the displacement of the kth heterogeneous cluster in the X, Y and Z axes. i , , respectively represent the formation error of the follower unmanned vehicle i and its reference trajectory in the X, Y and Z axes. Let the formation error of the follower unmanned vehicle i ​a formation error vector of the follower UAV a formation error vector of the leader UAV k The formation error of the follower UAV and the leader UAV considering different response time scales is:

[0128]

[0129] wherein, denotes the formation error of the follower UAV considering different response time scales, denotes the formation error of the leader UAV considering different response time scales.

[0130] In the embodiment, the expression of the discrete sliding mode function is as follows:

[0131]

[0132] wherein, denotes the discrete sliding mode function, k 4 and k 5 are control gains. According to the stability requirement of the sliding mode, the control input of the UAV formation is designed as:

[0133]

[0134] wherein, , sgn is a sign function, and the exponent is a gain . In order to make the stability of the follower UAV in tracking the reference trajectory and the leader UAV, the control gain and the upper sampling time interval are given as follows:

[0135]

[0136] wherein, denotes the maximum value of .

[0137] In some of the embodiments, the step S105 of constructing the small-time-scale UAV speed discrete model and determining the control input of the small-time-scale acceleration of the follower UAV comprises: constructing the small-time-scale UAV speed discrete model according to the actual speed of the follower UAV; constructing the sliding mode function of the speed loop according to the error between the actual speed of the follower UAV and the reference speed; and determining the control input of the small-time-scale acceleration of the follower UAV according to the convergence requirement of the sliding surface.

[0138] For example, the small-time-scale UAV speed discrete model is constructed as follows:

[0139]

[0140] wherein, , and respectively follow drone i The actual velocity in the X, Y, and Z axis directions. , and These respectively represent following the drone i Acceleration control inputs in the X, Y, and Z axes. Let the speed tracking error of the drone be:

[0141]

[0142] in, , and respectively follow drone i Velocity tracking error in the X, Y, and Z axes. , and These respectively represent following the drone i The desired velocity in the X, Y, and Z axes. Let the velocity error set be... The discrete sliding mode function is constructed as follows:

[0143]

[0144] in, Represents the discrete sliding mode function. k 6 represents the control gain, discrete sliding mode function. .

[0145] In this embodiment, the acceleration control input on a small time scale is:

[0146]

[0147] in, To follow the drone i Acceleration control inputs in the X, Y, and Z axes; sgn is the sign function, exponent. Gain .

[0148] To verify the effectiveness of the above method, a heterogeneous unmanned cluster simulation environment was built on the Gazebo platform for simulation experiments, such as... Figure 4 As shown, Figure 4 This is a schematic diagram of the heterogeneous unmanned swarm constructed in an embodiment of the present invention. In this experiment, the upper-level reference trajectory of the heterogeneous unmanned swarm formation control method based on iterative learning is as follows: Figure 5 As shown, the tracking error of the reference trajectory is as follows: Figure 6 As shown, the iterative learning strategy employed in this method enables the reference trajectories of the navigating autonomous vehicle and the following drone to...Figure 2 The method can converge to track the corresponding leader under the distributed communication structure, the iterative learning scheme in the method reduces the communication interaction times between neighbors, and has greater application value in system task planning of large-scale unmanned cluster formation.

[0149] Figure 7 is a formation trajectory schematic diagram of the leader unmanned vehicle and the follower unmanned vehicle in the heterogeneous cluster in the embodiment of the application, Figure 8 is a formation trajectory schematic diagram of three leader unmanned vehicles, Figure 9 is an angle change curve diagram of the unmanned vehicle formation, Figure 10 is a displacement error schematic diagram of the unmanned vehicle formation, Figure 11 is an angle error schematic diagram of the unmanned vehicle formation. Figures 7-11 It can be seen that, under the formation control strategy of the method, the formation error of the leader unmanned vehicle can converge to within 0.03 m.

[0150] Figure 12 is a displacement error of the unmanned vehicle formation, Figure 13 is a reference speed and speed tracking error of the unmanned vehicle formation in the X-axis direction, Figure 14 is a reference speed and speed tracking error of the unmanned vehicle formation in the Y-axis direction. Figures 12-14 It can be seen that, the unmanned vehicle formation controller considering different time scales set in the method makes the unmanned vehicle formation track the reference trajectory while taking into account the cooperation with the leader unmanned vehicle. Moreover, the sliding mode control method designed in the method makes the speed of the follower unmanned vehicle quickly track the reference speed, and improves the robustness of the formation, and the formation error is kept at about 0.05 m. Therefore, the method has excellent formation keeping ability and robustness.

[0151] In summary, the method generates its own reference trajectory through iterative learning in a small time scale, and only updates the distributed control input at a fixed time interval according to the iterative learning result. In the upper layer reference trajectory iterative learning tracking control and the lower layer heterogeneous formation control, the different time scale responses of the UGV and the UAV are considered. Considering the different flexibility of the UGV and the UAV and the different accuracy requirements of the reference trajectory, the reference trajectory iteration and the formation control input of the UGV have a larger time scale, while the reference trajectory iteration and the formation control input of the UAV have a smaller time scale. Unlike existing methods, this time scale separation method is more in line with the actual characteristics of the UGV and the UAV in the heterogeneous formation. Through Lyapunov stability analysis, an upper limit of the ratio of the two time scales is given to ensure the convergence of the reference trajectory iteration and the stability of the heterogeneous formation controller under different time scales.

[0152] The application also provides a heterogeneous unmanned cluster formation control system based on iterative learning, comprising:

[0153] a graph construction module, configured to construct a leading unmanned vehicle communication topology directed graph and a follower unmanned aerial vehicle communication topology directed graph;

[0154] a strategy formulation module, configured to construct a reference trajectory discrete iterative model of the leading unmanned vehicle and the follower unmanned aerial vehicle respectively, and determine a distributed control strategy of the reference trajectory of the leading unmanned vehicle and the follower unmanned aerial vehicle;

[0155] an unmanned vehicle control module, configured to construct an unmanned vehicle discrete kinematics model, and determine a control input of linear velocity and angular velocity of each leading unmanned vehicle according to a formation error of the reference trajectory of the leading unmanned vehicle;

[0156] an unmanned aerial vehicle control module, configured to construct an unmanned aerial vehicle discrete kinematics model, and generate a formation error discrete sliding mode function of the leading unmanned vehicle and the follower unmanned aerial vehicle at different time scales, and determine a control input of a reference velocity of the follower unmanned aerial vehicle;

[0157] a target control module, configured to construct a small time scale unmanned aerial vehicle velocity discrete model, and determine a control input of a small time scale acceleration of the follower unmanned aerial vehicle.

[0158] The application also provides a heterogeneous unmanned cluster formation control device based on iterative learning, and the following description is made to the heterogeneous unmanned cluster formation control device based on iterative learning provided by the application. Figure 15 The heterogeneous unmanned cluster formation control device based on iterative learning provided by the application is a structural block diagram as shown in Figure 15 The device comprises:

[0159] a construction module 1501, configured to construct a leading unmanned vehicle communication topology directed graph and a follower unmanned aerial vehicle communication topology directed graph;

[0160] a processing module 1502, configured to construct a reference trajectory discrete iterative model of the leading unmanned vehicle and the follower unmanned aerial vehicle respectively, and determine a distributed control strategy of the reference trajectory of the leading unmanned vehicle and the follower unmanned aerial vehicle;

[0161] a first control module 1503, configured to construct an unmanned vehicle discrete kinematics model, and determine a control input of linear velocity and angular velocity of each leading unmanned vehicle according to a formation error of the reference trajectory of the leading unmanned vehicle;

[0162] The second control module 1504 is configured to construct a discrete kinematic model of the UAV, and generate a formation error discrete sliding mode function of the lead UGV and the following UAV at different time scales, and determine a control input of a reference speed of the following UAV.

[0163] The third control module 1505 is configured to construct a small time scale discrete model of the UAV speed, and determine a control input of a small time scale acceleration of the following UAV.

[0164] In use, first, the construction module 1501 constructs a directed graph of a communication topology of the lead UGV and a communication topology of a following UAV cluster. In the directed graph of the communication topology of the lead UGV, the lead UGV is regarded as a graph node, and the communication relationship between different lead UGVs can be clearly represented. In the directed graph of the communication topology of the following UAV, the following UAV is regarded as a graph node, and the communication relationship between different following UAVs can be clearly represented. Then, the processing module 1502 respectively constructs a reference trajectory discrete iterative model of the lead UGV and the following UAV, and determines a distributed control strategy of the reference trajectory of the lead UGV and the following UAV. The first control module 1503 further constructs a discrete kinematic model of the UGV, and determines a control input of a linear speed and an angular speed of each lead UGV according to a formation error of the reference trajectory of the lead UGV. Then, the second control module 1504 constructs a discrete kinematic model of the UAV, and generates a formation error discrete sliding mode function of the lead UGV and the following UAV at different time scales, and determines a control input of a reference speed of the following UAV. Finally, the third control module 1505 constructs a small time scale discrete model of the UAV speed, and determines a control input of a small time scale acceleration of the following UAV.

[0165] In the above process, the reference trajectory of the self is generated through iterative learning in a small time scale, and the distributed control input is updated according to the iterative learning result only at a fixed time interval. In the upper reference trajectory iterative learning tracking control and the lower heterogeneous formation control, the different time scale responses of the UGV and the UAV are considered. Considering the different flexibility of the UGV and the UAV and the different accuracy requirements of the reference trajectory, the reference trajectory iteration and the formation control input of the UGV have a large time scale, and the reference trajectory iteration and the formation control input of the UAV have a small time scale. Unlike existing methods, this time scale separation method is more in line with the actual characteristics of the UGV and the UAV in the heterogeneous formation, and solves the problem of poor control stability in the related art.

[0166] Figure 16 An example of an electronic device is shown in FIG. 1, which is a schematic diagram of a physical structure of an electronic device. Figure 16As shown, the electronic device can include a processor 1601, a communications interface 1602, a memory 1603, and a communications bus 1604, wherein the processor 1601, the communications interface 1602, and the memory 1603 complete mutual communication through the communications bus 1604. The processor 1601 can invoke the logical instructions in the memory 1603 to execute the heterogeneous unmanned cluster formation control method based on iterative learning, which includes:

[0167] Construct a leading unmanned vehicle communication topology directed graph and a following unmanned aerial vehicle communication topology directed graph;

[0168] Discrete iterative models of the reference trajectories of the leading unmanned vehicle and the following unmanned aerial vehicle are constructed respectively, and a distributed control strategy of the reference trajectories of the leading unmanned vehicle and the following unmanned aerial vehicle is determined;

[0169] A discrete kinematic model of the unmanned vehicle is constructed, and control inputs of linear velocity and angular velocity of each leading unmanned vehicle are determined according to formation errors of the reference trajectories of the leading unmanned vehicle;

[0170] A discrete kinematic model of the unmanned aerial vehicle is constructed, and discrete sliding mode functions of formation errors of the leading unmanned vehicle and the following unmanned aerial vehicle at different time scales are generated, and control inputs of reference velocities of the following unmanned aerial vehicle are determined;

[0171] A small-time-scale discrete model of the unmanned aerial vehicle velocity is constructed, and control inputs of small-time-scale accelerations of the following unmanned aerial vehicle are determined.

[0172] In addition, the logical instructions in the memory 1603 described above can be implemented in the form of a software function unit and sold or used as an independent product, which can be stored in a computer-readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0173] In another aspect, the present application also provides a computer program product, which comprises a computer program, the computer program being stored in a non-transitory computer-readable storage medium and executable by a processor to enable a computer to perform the iteration learning based heterogeneous unmanned cluster formation control method provided by the above method, which comprises:

[0174] constructing a communication topology directed graph of the leading unmanned vehicle and a communication topology directed graph of the follower unmanned aerial vehicle;

[0175] constructing a reference trajectory discrete iterative model of the leading unmanned vehicle and the follower unmanned aerial vehicle respectively, and determining a distributed control strategy of the reference trajectory of the leading unmanned vehicle and the follower unmanned aerial vehicle;

[0176] constructing a discrete kinematics model of the unmanned vehicle, and determining a control input of the linear velocity and the angular velocity of each leading unmanned vehicle according to a formation error of the reference trajectory of the leading unmanned vehicle;

[0177] constructing a discrete kinematics model of the unmanned aerial vehicle, and generating a formation error discrete sliding mode function of the leading unmanned vehicle and the follower unmanned aerial vehicle at different time scales to determine a control input of the reference velocity of the follower unmanned aerial vehicle;

[0178] constructing a small time scale discrete model of the unmanned aerial vehicle velocity, and determining a control input of the small time scale acceleration of the follower unmanned aerial vehicle.

[0179] In another aspect, the present application also provides a non-transitory computer-readable storage medium, which stores a computer program, the computer program being executable by a processor to implement the iteration learning based heterogeneous unmanned cluster formation control method provided by the above method, which comprises:

[0180] constructing a communication topology directed graph of the leading unmanned vehicle and a communication topology directed graph of the follower unmanned aerial vehicle;

[0181] constructing a reference trajectory discrete iterative model of the leading unmanned vehicle and the follower unmanned aerial vehicle respectively, and determining a distributed control strategy of the reference trajectory of the leading unmanned vehicle and the follower unmanned aerial vehicle;

[0182] constructing a discrete kinematics model of the unmanned vehicle, and determining a control input of the linear velocity and the angular velocity of each leading unmanned vehicle according to a formation error of the reference trajectory of the leading unmanned vehicle;

[0183] constructing a discrete kinematics model of the unmanned aerial vehicle, and generating a formation error discrete sliding mode function of the leading unmanned vehicle and the follower unmanned aerial vehicle at different time scales to determine a control input of the reference velocity of the follower unmanned aerial vehicle;

[0184] constructing a small time scale discrete model of the unmanned aerial vehicle velocity, and determining a control input of the small time scale acceleration of the follower unmanned aerial vehicle.

[0185] The device embodiments described above are merely illustrative, wherein the units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, i.e., can be located in one place, or can be distributed to multiple network units. Part or all of the modules can be selected to achieve the purposes of the embodiments according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0186] Through the description of the above embodiments, those skilled in the art can clearly understand that the embodiments can be realized by means of software and the necessary general hardware platform, and of course can also be realized by hardware. Based on such understanding, the above technical solutions can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods described in each embodiment or some parts of the embodiments.

[0187] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement to part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A heterogeneous unmanned swarm formation control method based on iterative learning, characterized in that, include: Construct a directed graph of the communication topology for the leading unmanned vehicle and a directed graph of the communication topology for the following unmanned aerial vehicle; Discrete iterative models of the reference trajectories of the leading unmanned vehicle and the following unmanned vehicle are constructed respectively, and the distributed control strategies of the reference trajectories of the leading unmanned vehicle and the following unmanned vehicle are determined. A discrete kinematic model of the unmanned vehicle is constructed, and the control inputs of the linear velocity and angular velocity of each of the unmanned vehicles are determined based on the formation error of the reference trajectory of the navigating unmanned vehicle. A discrete kinematic model of the UAV is constructed, and discrete sliding mode functions of the formation error between the leading UAV and the following UAV at different time scales are generated to determine the control input of the reference speed of the following UAV. Construct a discrete velocity model of the UAV with a small time scale, and determine the control input for the small time scale acceleration of the following UAV; Discrete iterative models of the reference trajectories for the leading unmanned vehicle and the following unmanned aerial vehicle are constructed respectively, and distributed control strategies for the reference trajectories of the leading unmanned vehicle and the following unmanned aerial vehicle are determined, including: The discrete-time sampling period and iteration time interval of the reference trajectories of the leading unmanned vehicle and the following unmanned vehicle are determined respectively; Align the event points of the reference trajectory control input update to determine the number of iterations of the reference trajectories of the navigating drone and the following drone within the same time period; Construct a discrete iterative model of the reference trajectory of the leading unmanned vehicle and the following unmanned vehicle, and determine the distributed control strategy of the reference trajectory of the leading unmanned vehicle and the following unmanned vehicle; Based on Lyapunov stability theory, the range of values ​​for the ratio of the number of iterations of the reference trajectories of the leading unmanned vehicle and the following unmanned vehicle within the same time interval is obtained; Construct a discrete kinematics model of the UAV and generate discrete sliding mode functions for the formation error of the leading UAV and the following UAV at different time scales, including: Based on the displacement of the following UAV, a discrete kinematic model of the UAV is constructed; Using the iteration time interval of the upper reference trajectory as the time scale of the formation of the following UAV, a discrete sliding mode function for the formation error of the leading UAV and the following UAV with different time scales is designed. The control input for determining the reference speed of the following drone includes: Based on the stability requirements of the discrete sliding mode function, the control input of the following UAV is determined; The range of values ​​for the ratio of sliding mode gain parameter to formation time scale is obtained based on Lyapunov stability analysis; Construct a discrete velocity model of the UAV with a small time scale, and determine the control input for the small time scale acceleration of the following UAV, including: Based on the actual speed of the following drone, a small-time-scale discrete model of drone speed is constructed; A sliding mode function for the velocity loop is constructed based on the error between the actual speed of the following UAV and the reference speed; Based on the convergence requirements of the sliding surface, the control input for the small-timescale acceleration of the following UAV is determined.

2. The heterogeneous unmanned swarm formation control method based on iterative learning according to claim 1, characterized in that, Constructing a directed graph of the communication topology for the autonomous vehicle, including: Define the set of autonomous vehicles for the navigation vehicle and the virtual navigation vehicle; the virtual navigation vehicle is the navigator of the navigation vehicle. Based on the communication relationships between the lead unmanned vehicles, a directed graph of the communication topology of the lead unmanned vehicles is generated.

3. The heterogeneous unmanned swarm formation control method based on iterative learning according to claim 1, characterized in that, Constructing a directed graph of the communication topology for following drones, including: For each of the aforementioned unmanned vehicles corresponding to an unmanned cluster, a set of unmanned drones in the unmanned cluster is defined, and the aforementioned unmanned vehicle is designated as the leader of the unmanned cluster. Based on the communication relationships between the following drones, a directed graph of the communication topology of the following drones is generated.

4. The heterogeneous unmanned swarm formation control method based on iterative learning according to claim 1, characterized in that, Construct a discrete kinematic model of the unmanned vehicle, and determine the control inputs for the linear velocity and angular velocity of each of the lead unmanned vehicles based on the formation error of the reference trajectory of the lead vehicle, including: Based on the displacement and yaw angle of the navigating unmanned vehicle, the discrete kinematic model of the unmanned vehicle is determined; Based on the expected displacement, expected yaw angle and expected formation distance of the lead vehicle, the formation error of the lead vehicle's reference trajectory is determined; Based on the formation error of the reference trajectory of the navigating unmanned vehicle, the control inputs of the linear velocity and angular velocity of the formation controller of the navigating unmanned vehicle are designed.

5. A heterogeneous unmanned swarm formation control system based on iterative learning, used to implement the heterogeneous unmanned swarm formation control method based on iterative learning as described in any one of claims 1-4, characterized in that, include: The graph construction module is used to construct directed graphs of the communication topology for the leading unmanned vehicle and the following unmanned aerial vehicle. The strategy formulation module is used to construct discrete iterative models of the reference trajectories of the leading unmanned vehicle and the following unmanned vehicle, and to determine the distributed control strategy of the reference trajectories of the leading unmanned vehicle and the following unmanned vehicle. The unmanned vehicle control module is used to construct a discrete kinematic model of the unmanned vehicle and determine the control inputs of the linear velocity and angular velocity of each of the lead unmanned vehicles based on the formation error of the reference trajectory of the lead unmanned vehicle. The UAV control module is used to construct a discrete kinematic model of the UAV and generate discrete sliding mode functions of the formation error between the leading UAV and the following UAV at different time scales, and to determine the control input of the reference speed of the following UAV. The target control module is used to construct a discrete model of the UAV velocity on a small time scale and determine the control input for the small time scale acceleration of the following UAV.

6. A heterogeneous unmanned swarm formation control device based on iterative learning, used to implement the heterogeneous unmanned swarm formation control method based on iterative learning as described in any one of claims 1-4, characterized in that, include: The module is used to construct the directed graph of the communication topology for the leading unmanned vehicle and the following unmanned vehicle. The processing module is used to construct discrete iterative models of the reference trajectories of the leading unmanned vehicle and the following unmanned vehicle, respectively, and to determine the distributed control strategy of the reference trajectories of the leading unmanned vehicle and the following unmanned vehicle. The first control module is used to construct a discrete kinematic model of the unmanned vehicle and determine the control inputs of the linear velocity and angular velocity of each of the unmanned vehicles based on the formation error of the reference trajectory of the unmanned vehicle. The second control module is used to construct a discrete kinematic model of the UAV and generate discrete sliding mode functions of the formation error between the leading UAV and the following UAV at different time scales, and to determine the control input of the reference speed of the following UAV. The third control module is used to construct a discrete model of the UAV velocity on a small time scale and determine the control input for the small time scale acceleration of the following UAV.

Images