A dynamic target tracking control method for distributed cooperative maneuvering of multiple unmanned vessels

Abstract

A dynamic target tracking control method for distributed cooperative maneuvering of multiple unmanned boats relates to the field of cooperative control of unmanned boats. In order to solve the defect in the existing technology that it is impossible to achieve high-precision maneuvering tracking of dynamic targets in irregular waters, the technical solution provided by the present invention includes: based on the original position and heading of the unmanned boat, a new position expression with clearly defined relative degrees is generated according to preset parameters; the speed of the dynamic target is estimated online, and the target speed estimate is output; the information of each virtual leader is aggregated to estimate the leader speed; the formation rotation, scaling and translation parameters are generated according to the affine transformation theory; the system uncertainty and external disturbance are compensated online, and the final control signal for achieving maneuvering tracking control is output. It is suitable for the control task of multiple under-actuated unmanned boat clusters to perform autonomous cooperative maneuvering tracking of dynamic moving targets in complex sea conditions.

Description

Technical Field

[0001] The present invention relates to the field of cooperative control of unmanned boats, and particularly designs an unmanned boat cooperative target tracking control method and system. Background Art

[0002] Collaborative target tracking, a core technology for swarm intelligent control, requires multiple unmanned vehicles (UAVs) to coordinate in real time in a dynamic environment to achieve continuous and accurate tracking of moving targets in a desired formation. However, this technology faces multiple challenges, including inter-vessel collaborative coupling constraints and marine environmental interference. Control algorithm design must ensure trajectory tracking accuracy and swarm motion stability in complex sea conditions. Therefore, developing a collaborative tracking control system with strong anti-interference capabilities and high adaptability has become a cutting-edge issue in improving the operational effectiveness of UAV swarms.

[0003] Existing fixed-formation target tracking control methods have significant deficiencies in formation description and overall maneuverability. Specifically, traditional approaches often construct rigid formation constraints based on a static environment, failing to account for the non-static and unpredictable nature of the environment. This makes it difficult to address the maneuverability requirements of swarm systems in irregular waters. Furthermore, multiple constraints inherent in unmanned vehicle swarm systems, such as model parameter uncertainty, the complexity of external interference, and underactuated characteristics, further complicate the design of dynamic target tracking control strategies.

[0004] In summary, the existing technology has the defects of relying on fixed rigid formations, lacking adaptive processing for under-actuated characteristics and complex sea conditions, and being unable to achieve high-precision maneuvering tracking of dynamic targets in irregular waters. Summary of the Invention

[0005] To address the shortcomings of existing technologies, such as reliance on fixed rigid formations, lack of adaptive processing for underactuated characteristics and complex sea conditions, and inability to achieve high-precision maneuverable tracking of dynamic targets in irregular waters, the present invention provides the following technical solutions:

[0006] A dynamic target tracking control method for distributed cooperative maneuvering of multiple unmanned vessels, comprising:

[0007] Step 1: Output redefinition processing: Based on the original position and heading of the unmanned boat, a new position expression with clearly defined relative degrees is generated according to preset parameters, and the position expression and corresponding speed expression are output;

[0008] Step 2: Target speed estimation: Receive the output of step 1, perform online estimation of the dynamic target speed by interacting with neighboring nodes’ distributed information, and output the target speed estimate;

[0009] Step 3: Leader speed estimation: Receive the target speed estimate from step 2, aggregate the information of each virtual leader, estimate the leader speed through the consensus algorithm, and output the leader speed estimate;

[0010] Step 4: Formation transformation and control: Receive the velocity estimates from steps 2 and 3, generate formation rotation, scaling, and translation parameters based on affine transformation theory, and calculate and output basic control instructions in combination with the dynamic surface control algorithm;

[0011] Step 5: Adaptive compensation: Receive the basic control instructions from step 4, use minimum parameter learning combined with neural network approximation to perform online compensation for system uncertainty and external disturbances, and output the final control signal for achieving maneuver tracking control.

[0012] Furthermore, a preferred embodiment is provided, in step 5, the adaptive compensation performs functional approximation on the unknown dynamics of the system and external disturbances through a radial basis function neural network, and combines the minimum parameter online learning strategy to update only the constant weights.

[0013] Furthermore, a preferred embodiment is provided. In step 4, the formation transformation calculates the position deviation between the current formation and the nominal formation in real time, constructs an affine mapping based on the rotation matrix and translation vector, and then combines the adaptive adjustment of the formation scaling factor to realize dynamic rotation, scaling and translation control of the unmanned boat cluster.

[0014] Furthermore, a preferred embodiment is provided, in step 1, the new position expression is obtained by multiplying the original position value and the heading value by the cosine weight and the sine weight respectively and then adding the results.

[0015] Furthermore, a preferred embodiment is provided, in step 2, the target speed estimation is based on a distributed second-order consensus protocol, and the speed information of neighboring nodes is weighted averaged and differentiated to achieve real-time tracking estimation of the target speed.

[0016] Furthermore, a preferred embodiment is provided, in step 3, the leader speed estimation adopts the Laplace matrix weighted consensus algorithm to perform information fusion on the speed estimations between the virtual leader nodes.

[0017] A dynamic target tracking control device for distributed cooperative maneuvering of multiple unmanned ships is also provided, comprising:

[0018] Module 1: Output redefinition processing: Based on the original position and heading of the unmanned boat, a new position expression with clearly defined relative degrees is generated according to preset parameters, and the position expression and corresponding speed expression are output;

[0019] Module 2: Target speed estimation: Receive the output of step 1, perform online estimation of the dynamic target speed by interacting with neighboring nodes’ distributed information, and output the target speed estimate;

[0020] Module 3: Leader speed estimation: Receives the target speed estimate from step 2, aggregates information from each virtual leader, estimates the leader speed using a consensus algorithm, and outputs the leader speed estimate.

[0021] Module 4: Formation Transformation and Control: Receives the velocity estimates from steps 2 and 3, generates formation rotation, scaling, and translation parameters based on affine transformation theory, and calculates and outputs basic control instructions in conjunction with the dynamic surface control algorithm.

[0022] Module 5: Adaptive Compensation: Receives the basic control instructions from step 4, uses minimum parameter learning combined with neural network approximation to perform online compensation for system uncertainties and external disturbances, and outputs the final control signal for achieving maneuver tracking control.

[0023] A computer storage medium is also provided for storing a computer program, and when the computer program is read by a computer, the computer executes the method.

[0024] A computer is also provided, comprising a processor and a storage medium, wherein when the processor reads a computer program stored in the storage medium, the computer executes the method.

[0025] A computer program product is also provided, which is a computer program that implements the method when the computer program is executed.

[0026] Compared with the prior art, the technical solution provided by the present invention is beneficial in that:

[0027] First, the dynamic conversion method based on output redefinition completely eliminates the control difficulties caused by the mutual coupling of position and attitude in under-actuated unmanned boats. This method constructs a new output with clearly defined relative degrees by combining the original position output and heading angle output according to preset parameters, eliminating the need to deal with coupling relationships in controller design. The present invention applies this method to the under-actuated unmanned boat model, simplifying the control law construction process and improving design efficiency. Compared with traditional methods that often require the construction of complex virtual controllers to solve the mutual coupling problem, this solution relies on only one output mapping to complete relative degree compensation, significantly reducing the design difficulty and engineering implementation costs.

[0028] Secondly, a control architecture combining a hierarchical distributed velocity observer with affine transformation significantly improves the formation description capability and maneuvering tracking performance of the unmanned watercraft swarm. The distributed target velocity observer and distributed virtual leader velocity observer can estimate target and leader velocities based solely on neighbor communication, reducing reliance on global information. The affine transformation-based tracking control enables the swarm to flexibly implement complex formation changes such as rotation, scaling, and translation when performing dynamic target maneuvering tracking. Compared with traditional fixed rigid formation control methods, this scheme exhibits greater anti-interference and safety in irregular waters, and can maintain stable tracking in complex sea conditions or environments with limited communication.

[0029] Finally, to address the model uncertainty and complex external interference commonly found in unmanned watercraft systems, this paper integrates radial basis function neural networks with minimum parameter learning to online approximate and adaptively compensate for unknown dynamics and disturbances. This strategy only requires estimating constant parameters, avoiding the extensive calculation of matrix parameters required in traditional pure RBF network online learning. This significantly reduces the computational burden and improves system real-time performance while ensuring control accuracy. Unlike existing research that relies heavily on high-dimensional online parameter updates, this scheme's minimum parameter learning significantly shortens convergence time and reduces the computational resource requirements of the unmanned watercraft controller.

[0030] It is suitable for the control task of multiple under-actuated unmanned boats clusters to perform autonomous coordinated maneuvering and tracking of dynamic moving targets in complex sea conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 A hierarchical diagram for dynamic target tracking;

[0032] Figure 2 A schematic diagram of a dynamic target tracking control method for distributed cooperative maneuvering of multiple unmanned vessels;

[0033] Figure 3 This is the tracking trajectory diagram of the maneuvering target of the unmanned boat swarm system;

[0034] Figure 4 It is the tracking distance error variation curve;

[0035] Figure 5 is the relative position error change curve;

[0036] Figure 6 It is the adaptive law change curve;

[0037] Figure 7 This is the control input change curve. DETAILED DESCRIPTION

[0038] In order to make the advantages and benefits of the technical solution provided by the present invention more clearly reflected, the technical solution provided by the present invention is now further described in detail with reference to the accompanying drawings, specifically:

[0039] Implementation 1: This implementation provides a dynamic target tracking control method for distributed coordinated maneuvering of multiple unmanned vessels, including:

[0040] Step 1: Output redefinition processing: Based on the original position and heading of the unmanned boat, a new position expression with clearly defined relative degrees is generated according to preset parameters, and the position expression and corresponding speed expression are output;

[0041] Step 2: Target speed estimation: Receive the output of step 1, perform online estimation of the dynamic target speed by interacting with neighboring nodes’ distributed information, and output the target speed estimate;

[0042] Step 3: Leader speed estimation: Receive the target speed estimate from step 2, aggregate the information of each virtual leader, estimate the leader speed through the consensus algorithm, and output the leader speed estimate;

[0043] Step 4: Formation transformation and control: Receive the velocity estimates from steps 2 and 3, generate formation rotation, scaling, and translation parameters based on affine transformation theory, and calculate and output basic control instructions in combination with the dynamic surface control algorithm;

[0044] Step 5: Adaptive compensation: Receive the basic control instructions from step 4, use minimum parameter learning combined with neural network approximation to perform online compensation for system uncertainty and external disturbances, and output the final control signal for achieving maneuver tracking control.

[0045] In step 5, the adaptive compensation performs functional approximation on the unknown dynamics of the system and external disturbances through the radial basis function neural network, and combines the minimum parameter online learning strategy to update only the constant weights.

[0046] In step 4, the formation transformation calculates the position deviation between the current formation and the nominal formation in real time, constructs an affine mapping based on the rotation matrix and translation vector, and then combines the adaptive adjustment of the formation scaling factor to achieve dynamic rotation, scaling and translation control of the unmanned boat cluster.

[0047] In step 1, the new position expression is obtained by multiplying the original position value and heading value by the cosine weight and sine weight respectively and then adding them together.

[0048] In step 2, the target speed estimation is based on the distributed second-order consensus protocol, and the speed information of neighboring nodes is weighted averaged and differentiated to achieve real-time tracking estimation of the target speed.

[0049] In step 3, the leader speed estimation uses the Laplace matrix weighted consensus algorithm to perform information fusion on the speed estimates between the virtual leader nodes.

[0050] Implementation Method 2: This implementation method further describes the technical solution provided in Implementation Method 1 in detail. Specifically:

[0051] Step 1:

[0052] The coupling problem between the position and heading of the unmanned boat is eliminated through the dynamic conversion method of output redefinition.

[0053] This step first generates a new position representation based on the original heading and position measurements of the UAV, using a preset parameter ratio. This representation has a clear definition of relative degrees and avoids control problems caused by insufficient input dimensionality. Once the conversion is complete, the new position representation and corresponding velocity representation are directly obtained for subsequent observer and controller design.

[0054] The inputs to the previous step were the original position and heading measurements; the outputs of this step are the new position representation and the corresponding velocity representation.

[0055] Step 2:

[0056] A target speed estimation module is designed based on a distributed structure to collaboratively obtain the speed information of dynamic targets among nodes.

[0057] In this step, in each virtual leader node, the real-time changes in the current position expression and the speed estimate at the previous moment are utilized. By exchanging estimated values ​​with adjacent nodes and combining them with the design parameters, a speed estimation algorithm is constructed so that the estimated results can consistently converge to the approximate value of the target true speed, providing a smooth and reliable target speed input for subsequent control.

[0058] The input of the previous step is the new position expression and the corresponding velocity expression; the output of this step is the distributed estimation result of the target velocity by each virtual leader.

[0059] Step 3:

[0060] A leader speed estimation module is designed based on a distributed structure to complete the required leader speed information for each follower boat.

[0061] In this step, each follower node collects velocity estimates from all virtual leaders. Through information transmission and a consensus convergence mechanism between neighboring nodes, an online estimate of the virtual leader's true velocity is performed. This estimation enables each follower node to obtain stable and reliable leader velocity data, using only neighboring communication, as input for the next control step.

[0062] The input of the previous step is the target speed estimation result of each virtual leader; the output of this step is the distributed estimation result of the virtual leader's speed by each follower boat.

[0063] Step 4:

[0064] Based on the transformation mapping theory, a distributed tracking control method is designed to realize the dynamic rotation, scaling and translation maneuvers of the cluster formation.

[0065] This step first calculates the position and speed deviations between each boat and the nominal formation based on the target speed estimate and the leader's speed estimate. Then, a transformation mapping technique is used to generate formation adjustment parameters, allowing the nominal formation to dynamically deform according to rotation, scaling, and translation requirements. Finally, combined with the error dynamic surface control algorithm, control instructions are generated to drive the coordinated maneuvers of each boat, enabling the entire cluster to maintain precise formation and continuous tracking under changing sea conditions and target maneuvers.

[0066] The input of the previous step is each boat's estimation of the target and leader's speed; the output of this step is the basic control command used to drive the maneuvering formation tracking.

[0067] Step 5:

[0068] A method combining minimum parameter learning and neural network approximation is used to online compensate for system model uncertainty and external disturbances to achieve adaptive control.

[0069] In this step, all unknown dynamic terms and sea state disturbances are considered as overall uncertainties, and a neural network structure is used to perform functional approximation on this uncertainty, with only a small number of constant parameters updated online. The parameters are adaptively adjusted based on the tracking error, so that compensation for uncertainties and disturbances can take effect immediately. While ensuring tracking accuracy, the amount of computation is significantly reduced, ensuring the real-time and robustness of the entire control system in complex sea conditions.

[0070] Implementation Method 3: Combination Figure 1-7 This embodiment further describes the above technical solution in detail through specific examples, specifically:

[0071] (1) Establish a mathematical model of unmanned boat motion. Define unmanned boat The mathematical model is:

[0072]

[0073] Where, , , , Unmanned boat quality, Unmanned boat The moment of inertia, , and Unmanned boat The hydrodynamic coefficient can be calculated through tank tests or its approximate value can be estimated through relevant empirical formulas. and Unmanned boat The control input, 、 and is the lumped term of model uncertainty and external interference.

[0074] Considering the problems that the underactuated characteristics of the unmanned boat bring to the control method design, a position output redefinition conversion method is designed. First, define 、 To achieve relativity, the control matrix needs to be non-singular, that is, the two system inputs must appear in the input-output dynamic system. , with only one input appears in an input-output dynamic system, and another input Appears in Therefore, by combining the original output and To construct a new output to solve the relative degree missing problem.

[0075] definition ,in, is the design parameter, the new position output vector can be expressed as Based on the above position conversion, the mathematical model of the unmanned boat can be expressed as:

[0076]

[0077] Where, , is the lumped disturbance term, and is a nonlinear coupling term, and its specific form is:

[0078]

[0079] It can be seen that the underactuated unmanned boat system selects a combination of primitive variables without any complex calculations. As the new output of the system, the new input-output system has a well-defined relative degree. However, even if the system output changes, the target tracking task does not change. Therefore, it is necessary to select a more appropriate new system output as an approximation of the original output so that when the new system output tracks the target, the tracking error of the original system output remains within a small range.

[0080] (2) Establish affine transformation relationship.

[0081] Include Diagram of information interaction topology of swarm system of unmanned boats Indicates that, Represents the unmanned boat set, Represents an edge set. , there is a directed edge , unmanned boat Can receive unmanned boats The relevant information sent, in addition, There is no loop in .definition Unmanned boat The set of neighbors of . Definition For the picture The non-negative adjacency matrix of , where if ,So ,otherwise .definition is a directed graph The Laplace matrix of , where:

[0082]

[0083] Define the configuration for A finite set of unmanned boats is represented by ,Right now The lumped vector consisting of the position vectors of the unmanned boats. Definition The unmanned boat is the leader, and the rest The unmanned boat is defined as a follower. Under this premise, and Used to represent sets of leaders and followers respectively; configuration It can be decomposed into leader configuration and follower configuration, namely and Denote the configurations of leaders and followers respectively. Definition and denote the speed sets of the leader and follower respectively.

[0084] Stress is a set of scalar quantities ,in With edge Related, when unmanned boat With unmanned boats When there is attraction between When the unmanned boat With unmanned boats When there is repulsive force between ; Other situations, If the stress satisfies the following equation, it is called equilibrium stress:

[0085]

[0086] It can be further expressed in matrix form as follows:

[0087]

[0088] Where, For the framework The stress matrix satisfies:

[0089]

[0090] Next, the definition of the target affine formation is given. The target frame of the time-varying target affine formation is:

[0091]

[0092] Where, represents an affine transformation. is a matrix used to achieve relative to the nominal configuration Rotation, scaling, and shearing; is a vector that realizes the nominal configuration Translation of unmanned boats. The expected position in the target affine formation is .

[0093] The communication topology of the entire unmanned boat system is defined as For the first layer, the directed topology between the target and the virtual leader is defined as For the second layer, the communication graph between the virtual leader and followers is defined as .

[0094] For the topology graph defined above, the following assumptions are made.

[0095] Assumption 1: In a directed communication graph In the , the root node of the directed spanning tree is the target.

[0096] Assumption 2: In a directed communication graph There is at least one communication path from the virtual leader to the follower.

[0097] Assumption 3: Consider A swarm system consisting of 20 unmanned boats has at least 3 virtual leaders.

[0098] Hypothesis 4: The leader unmanned boat is equipped with advanced sensing equipment, which can detect obstacles in the surrounding environment in real time and accurately control its navigation position through intelligent planning programs.

[0099] Assumption 5: There exists a positive constant , making Established, .

[0100] Assumption 6: Maximum speed of the unmanned boat satisfy ,in represents the maximum speed of the target, Represents the upper bound of the velocity deviation caused by environmental disturbances.

[0101] Then, the topology of the entire unmanned boat cluster system can be expressed by the Laplace matrix The specific form is:

[0102]

[0103] Where, represents the Laplace matrix between the target and the virtual leader. If A virtual leader can receive the target's information, then , if it is not satisfied, then . Represents the Laplace matrix between virtual leaders, which is also a symmetric matrix. Represents the Laplacian matrix between the virtual leader and the followers. represents the Laplacian matrix between followers.

[0104] If and only if Assumption 1 holds, the matrix is a non-singular In addition, if Assumption 2 holds, then Every term of is non-negative, and All rows of equal 1.

[0105] In practical applications, when the unmanned boat swarm system performs target tracking tasks in narrow waterways or complex sea areas, there are many obstacles and safety is of paramount importance. Assumption 4 ensures that the swarm system can flexibly respond to potential collision risks in dynamic and complex environments and can always track the desired position, that is, for any time ,have ,in represents the desired position of the leader. By taking advantage of the affine transformation method, by adjusting the parameters and Effective collision avoidance maneuver control can be achieved.

[0106] The stress matrix is ​​divided into leaders and followers:

[0107]

[0108] Where, 、 、 、 、 、 、 and .

[0109] Define the desired position of the follower as According to the properties of the stress matrix, we can get:

[0110]

[0111] Based on the above, the position tracking error can be defined as follows:

[0112]

[0113] The stack form of position tracking error is defined as . Combined ,have:

[0114]

[0115] The unmanned boat swarm system is divided into two parts: a virtual leader group and a follower group. By combining this structure with affine transformation theory, when performing tracking missions, only a small number of unmanned boats need to be sent command instructions to achieve complex maneuver planning for the entire swarm system. This improves the formation description and maneuverability of the unmanned boat swarm system, as well as its safety when performing tracking missions in irregular waters. This also expands the mission capabilities of the unmanned boat swarm system when traversing irregular waters during the tracking process.

[0116] (3) Design a distributed speed observer.

[0117] The speed observer design is divided into two modules: 1) a distributed target speed observer and 2) a distributed virtual leader speed observer. These two speed observers are described below.

[0118] 1) Distributed target speed observer

[0119] Followers The boat can observe the target information, namely . The target speed observer designed on the boat is as follows:

[0120]

[0121] Where, For the The target speed estimate of the virtual leader, For the The target position obtained by the virtual leader, is a dummy variable, is a positive design parameter, is the speed at the last sampling moment.

[0122] The target velocity observer can estimate the target velocity, and the estimation error converges to a small range near zero.

[0123] 2) Distributed Virtual Leader Speed ​​Observer

[0124] To estimate the speed of the virtual leader, the distributed virtual leader speed observer is designed as:

[0125]

[0126] Where, is the observed value of the virtual leader's velocity. Speed ​​for virtual leaders. is a positive design parameter and satisfies .

[0127] The distributed virtual leader speed observer can observe the speed of the virtual leader, that is, the system observation error converges uniformly and asymptotically to a small range near zero.

[0128] (4) Design a distributed cooperative maneuvering target tracking control method.

[0129] Based on dynamic surface control, a distributed cooperative target tracking control method for multiple underactuated unmanned vehicles is proposed. Below, the relative position error is first defined as , the speed error is , the specific form is:

[0130]

[0131] Before continuing, define the relative position error The stack form is , we can get:

[0132]

[0133] Combine , we can get:

[0134]

[0135] Therefore, once you have determined The stability of is stable. Taking the derivative with respect to time, we get:

[0136]

[0137] According to the above formula, the designed virtual control law is:

[0138]

[0139] Where, Indicates that the control parameters are being designed.

[0140] In actual engineering applications, model parameters Generally, it is an unknown term, which makes it impossible to accurately design the control method. Therefore, by using the online approximation characteristics of radial basis neural network for any continuous function, the unknown model parameters and external interference can be approximated as:

[0141]

[0142] Where, represents the ideal weight, represents the output based on the Gaussian function, Represents the estimation error.

[0143] Considering that the large number of parameters in the online approximation process of the radial basis neural network will increase the complexity of the calculation process, an unknown model online learning algorithm based on the radial basis neural network and combined with the minimum parameter learning method is proposed, which can effectively reduce the time-consuming and complex problems of online learning. It can be further expressed as the following formula:

[0144]

[0145] Where, , , .

[0146] In the formula, by combining the minimum parameter learning method with the radial basis function neural network, only constant variables need to be estimated instead of matrix variables, which effectively reduces the amount of calculation and the time-consuming online learning process of the radial basis function neural network, and can greatly improve the real-time performance of the control system.

[0147] Based on the above analysis, the controller designed for each unmanned boat is:

[0148]

[0149] Where, is the positive design control parameter matrix. and They are and The adaptive control law is designed as follows:

[0150]

[0151] Where, 、 、 and is a positive design parameter.

[0152] Consider a swarm of unmanned underwater vehicles with a leader-follower architecture. Using virtual control laws, control laws, and adaptive laws, all signals in the target tracking control system are uniformly asymptotically stable. The cascaded system of distributed target velocity observers, distributed virtual leader velocity observers, and distributed cooperative maneuvering target tracking control is globally uniformly asymptotically stable.

[0153] Technical Effects

[0154] (1) Based on the affine transformation theory, this paper designs a distributed cooperative maneuvering target tracking control method for unmanned boats, which improves the formation description capability and maneuverability of the unmanned boat swarm system. The design of the distributed speed observer reduces the demand for onboard equipment of the unmanned boat swarm system and improves its safety and reliability when performing tracking tasks in irregular waters.

[0155] (2) The present invention designs a dynamic conversion method based on output redefinition, which eliminates the problems caused by the mutual coupling of the position and attitude of the unmanned boat, overcomes the difficulties brought by the under-actuated characteristics to the controller design, and simplifies the design process.

[0156] (3) The present invention sets the unknown disturbances inside and outside the system as the total uncertainty term, adopts radial basis function neural network to approximate it online, and effectively reduces the computational burden through the minimum parameter learning method.

[0157] The above further describes the technical solution provided by the present invention in detail through several specific embodiments in order to highlight the advantages and benefits of the technical solution provided by the present invention. However, the several specific embodiments described above are not intended to limit the present invention. Any reasonable modification and improvement of the present invention, combination of embodiments and equivalent replacement based on the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A dynamic target tracking control method for distributed cooperative maneuvering of multiple unmanned vessels, characterized in that: include: Step 1: Output redefinition processing: Based on the original position and heading of the unmanned boat, a new position expression with clearly defined relative degrees is generated according to preset parameters, and the position expression and corresponding speed expression are output; Step 2: Target speed estimation: Receive the output of step 1, perform online estimation of the dynamic target speed by interacting with neighboring nodes’ distributed information, and output the target speed estimate; Step 3: Leader speed estimation: Receive the target speed estimate from step 2, aggregate the information of each virtual leader, estimate the leader speed through the consensus algorithm, and output the leader speed estimate; Step 4: Formation transformation and control: Receive the velocity estimates from steps 2 and 3, generate formation rotation, scaling, and translation parameters based on affine transformation theory, and calculate and output basic control instructions in combination with the dynamic surface control algorithm; In step 4, the formation transformation calculates the position deviation between the current formation and the nominal formation in real time, constructs an affine mapping based on the rotation matrix and translation vector, and then combines it with the adaptive adjustment of the formation scaling factor to achieve dynamic rotation, scaling and translation control of the unmanned boat cluster; Step 5: Adaptive compensation: Receive the basic control instructions from step 4, use minimum parameter learning combined with neural network approximation to perform online compensation for system uncertainty and external disturbances, and output the final control signal for achieving maneuver tracking control.

2. The dynamic target tracking control method for distributed cooperative maneuvering of multiple unmanned vessels according to claim 1 is characterized in that: In step 5, the adaptive compensation performs functional approximation on the unknown dynamics of the system and external disturbances through the radial basis function neural network, and combines the minimum parameter online learning strategy to update only the constant weights.

3. The dynamic target tracking control method for distributed coordinated maneuvering of multiple unmanned vessels according to claim 1 is characterized in that: In step 1, the new position expression is obtained by multiplying the original position value and heading value by the cosine weight and sine weight respectively and then adding them together.

4. The dynamic target tracking control method for distributed coordinated maneuvering of multiple unmanned vessels according to claim 1 is characterized in that: In step 2, the target speed estimation is based on the distributed second-order consensus protocol, and the speed information of neighboring nodes is weighted averaged and differentiated to achieve real-time tracking estimation of the target speed.

5. The dynamic target tracking control method for distributed cooperative maneuvering of multiple unmanned vessels according to claim 1 is characterized in that: In step 3, the leader speed estimation uses the Laplace matrix weighted consensus algorithm to perform information fusion on the speed estimates between the virtual leader nodes.

6. A dynamic target tracking control device for distributed cooperative maneuvering of multiple unmanned ships, characterized in that: include: Module 1: Output redefinition processing: Based on the original position and heading of the unmanned boat, a new position expression with clearly defined relative degrees is generated according to preset parameters, and the position expression and corresponding speed expression are output; Module 2: Target speed estimation: Receives the output of module 1, performs online estimation of the dynamic target speed by interacting with neighboring nodes’ distributed information, and outputs the target speed estimate; Module 3: Leader speed estimation: Receives the target speed estimate from module 2, aggregates information from each virtual leader, estimates the leader speed using a consensus algorithm, and outputs the leader speed estimate. Module 4: Formation Transformation and Control: Receives the velocity estimates from modules 2 and 3, generates formation rotation, scaling, and translation parameters based on affine transformation theory, and calculates and outputs basic control instructions in combination with the dynamic surface control algorithm. In module 4, the formation transformation calculates the position deviation between the current formation and the nominal formation in real time, constructs an affine mapping based on the rotation matrix and translation vector, and then combines it with the adaptive adjustment of the formation scaling factor to achieve dynamic rotation, scaling and translation control of the unmanned boat cluster; Module 5: Adaptive Compensation: Receives the basic control instructions from module 4, uses minimum parameter learning combined with neural network approximation to perform online compensation for system uncertainties and external disturbances, and outputs the final control signal for achieving maneuver tracking control.

7. A computer storage medium for storing a computer program, characterized in that When the computer program is read by a computer, the computer executes the method according to claim 1 .

8. A computer comprising a processor and a storage medium, characterized in that When the processor reads the computer program stored in the storage medium, the computer executes the method according to claim 1 .

9. Computer program product comprising a computer program, characterized in that When the computer program is executed, the method according to claim 1 is implemented.

Images