A robust adaptive consistency control method for multi-AUV distributed clusters

Abstract

This invention discloses a robust adaptive consensus control method for a multi-AUV distributed cluster, comprising: establishing a kinematic mathematical model of virtual leaders and followers in the multi-AUV cluster system and rewriting it into a standard second-order system suitable for consensus control protocol design; determining the communication topology of the multi-AUV cluster system; calculating the consensus error and filtered consensus error for each AUV; constructing a robust neural damping term; designing a distributed consensus control protocol based on the robust neural damping term and adaptive control gain; designing the adaptive law of the adaptive parameters of the robust damping term and the adaptive law of the time-varying control gain; adjusting the parameters involved in the consensus control protocol and the adaptive law, and applying them to the consensus control of the multi-AUV cluster. This invention can achieve asymptotic consensus control even when the AUV model has arbitrary nonlinear model uncertainties and unknown leader dynamics, and can be extended to formation control, inclusion control, and cooperative target tracking of multi-AUV clusters.

Description

Technical Field

[0001] This invention relates to distributed control technology for multiple autonomous underwater vehicles (AUVs), and more particularly to a robust adaptive consistency control method for a distributed cluster of multiple AUVs. Background Technology

[0002] Cooperative swarm control of multiple autonomous underwater vehicles (AUVs) has attracted widespread attention from experts and engineers in the field of marine cybernetics in recent years due to its high efficiency and low power consumption. Multi-AUV swarm modes can be divided into centralized and distributed types. The centralized mode relies on a specific central node that monitors the state of all AUVs in the swarm for decision-making and control, placing high demands on communication and computing resources and exhibiting poor scalability. Furthermore, the centralized model suffers from a "single point of failure" problem; if the central node is attacked or damaged, the entire swarm system struggles to function properly. In contrast, in the distributed mode, each individual agent obtains information from only a subset of its neighboring nodes through local information exchange, significantly reducing the demands on communication and computing capabilities, providing greater flexibility, and making it suitable for communication-constrained underwater swarm environments. The consistency problem is a fundamental issue in multi-agent systems, referring to the design of distributed consistency control protocols to achieve state convergence among all agents through local information exchange. Consistency control methods can be extended to problems such as cooperative formation, cooperative target tracking, and coordinated target inclusion. Researching distributed consensus control methods for multiple underwater vehicles is not only of significant theoretical importance, but also provides valuable reference for engineering problems such as collaborative exploration and combat in underwater swarm systems.

[0003] The motion models of underwater vehicles are highly nonlinear and uncertain, and are affected by disturbances from unknown marine environments, making it difficult to establish accurate mathematical models. Existing control methods cannot effectively solve the consensus problem of multiple underwater vehicles considering model uncertainties. Control schemes based on approximators such as neural networks and fuzzy systems require a sufficient number of weight parameters for online learning, resulting in a significant computational burden. Furthermore, to solve for the control gain parameters that satisfy stability conditions, certain topological information, such as the minimum eigenvalue of the Laplace matrix, must be assumed to be globally known, which has significant limitations and generally only guarantees that the consensus error will eventually be bounded. Therefore, it is necessary to study a multi-underwater vehicle asymptotic consensus control protocol that is independent of global information, has good robustness and adaptability, and has a lower computational burden under model uncertainty. Summary of the Invention

[0004] The technical problem to be solved by this invention is to provide a robust adaptive consistency control method for multi-AUV distributed clusters, which addresses the shortcomings of existing technologies. Compared with traditional control methods, this method can achieve accurate consistency control with a small computational burden under arbitrary model uncertainty and environmental disturbances, and does not depend on any global information.

[0005] The technical solution adopted by this invention to solve its technical problem is: This invention provides a robust adaptive consistency control method for multi-AUV distributed clusters, the method comprising the following steps: Step 1: Establish a motion mathematical model of a virtual leader and multiple followers in a multi-AUV cluster system, and rewrite it into a standard second-order system suitable for the design of a consensus control protocol; Step 2: Determine the communication topology of the multi-AUV cluster system; Step 3: Calculate the consistency error and filter consistency error for each AUV; Step 4: Construct robust neural damping terms, and design a distributed consensus control protocol based on robust neural damping terms and adaptive control gain; Step 5: Based on the consistency error of each AUV obtained in Step 3, design the adaptive law of the adaptive parameters of the robust damping term, and the adaptive law of the time-varying control gain. Step 6: Adjust the parameters involved in the consistency control protocol and adaptive law, and apply them to the consistency control of the multi-AUV cluster.

[0006] Furthermore, the method for establishing a motion mathematical model consisting of one virtual leader and multiple followers in a multi-AUV cluster system in step 1 of the present invention is as follows: A multi-AUV swarm system comprises one virtual leader and n follower AUVs; the motion model of the i-th AUV can be described as follows: In the formula, This indicates the vehicle's position and heading in the geodetic coordinate system. These are the longitudinal and lateral linear velocities and yaw angular velocities of the aircraft. This is the state transition matrix from the carrier coordinate system to the geodetic coordinate system; The matrix represents the Coriolis force and the centripetal force. The damping force matrix; This is to control the input vector, i.e., the control force and torque.

[0007] Furthermore, the method for rewriting the system into a standard second-order system in step 1 of this invention is as follows: The motion model of the AUV is converted into the following standard second-order nonlinear system: In the formula, , It is a nonlinear function of the vehicle's speed and heading; The transformed control input vector; similarly, the motion model of the virtual leader is described as follows: The virtual leader's course and speed are specified by the user based on the cluster task.

[0008] Furthermore, the method in step 2 of the present invention is as follows: The communication structure of a multi-AUV cluster system is as follows: only some of the aircraft need to know the status information of the virtual leader; each follower aircraft communicates bidirectionally only with other aircraft within its communication range, forming a connected undirected graph; there is a directed path from the virtual leader to any follower, i.e., there exists a spanning tree with the virtual leader as the root node.

[0009] Furthermore, the calculation method for step 3 of the present invention is as follows: Calculate the consistency error and filter consistency error for each AUV, specifically described as follows: in, This represents the communication weight between spacecraft i and spacecraft j; if they can communicate, then... ,on the contrary, ; This represents the communication weight between vehicle i and the virtual leader. If vehicle i can obtain information from the virtual leader, then... ,on the contrary, Based on this, the filtering consistency error of each follower aircraft is defined as: .

[0010] in, The adjustment parameter is greater than 0.

[0011] Furthermore, the method in step 4 of the present invention is as follows: The robust damping term is in the form of: in, For adaptive parameter vectors; , Let be a radial basis function vector, where each radial basis function... The formal description is as follows: In the formula, Let be the input vector of the radial basis functions. and The center vector and width of the radial basis functions; The integral-bounded function is used to ensure asymptotic consistency in multi-AUV systems. and The adjustment parameter is greater than 0; Based on the constructed robust adaptive damping term, the following robust adaptive consensus control protocol is designed: In the formula, The time-varying control gain vector is used; according to the robust adaptive consensus protocol, the actual control input vector of each vehicle before coordinate transformation is obtained: .

[0012] Furthermore, the method in step 5 of the present invention is as follows: The specific form of the adaptive law for controlling the gain is as follows: The adaptive law for the adaptive parameter in the robust adaptive damping term is as follows: In the formula, and The adaptive law adjustment parameter is greater than 0 and is specified by the user. and The adaptive law adjustment parameter is greater than 0 and is specified by the user.

[0013] Furthermore, the method for multi-AUV cluster consistency control in step 6 of the present invention is as follows: At each sampling time, each AUV collects its own state, exchanges information with neighboring AUVs according to a predetermined communication topology, calculates its own actual control input, and updates the control gain and adaptive parameters according to the parameter adaptive law, ultimately enabling the multi-AUV system to achieve asymptotic consistency.

[0014] The beneficial effects of this invention are: The method proposed in this invention can achieve asymptotic consistency control even when the AUV model has arbitrary nonlinear model uncertainties and unknown leader dynamics. This method can be extended to formation control, inclusion control, and cooperative target tracking of multiple AUV clusters. Compared with existing technologies, the advantages of this invention are: 1. This invention uses robust neural damping terms to handle the model uncertainties of the AUV itself and the unknown leader dynamics, requiring only three adaptive parameters to be updated online to stabilize the composite uncertainties, resulting in a smaller online computational burden and suitability for engineering practice; 2. It uses online-updated control gains for consistency protocol design and introduces an integral bounded function into the parameter adaptive law, enabling asymptotic consistency control without requiring global information, achieving higher accuracy compared to traditional control methods that only guarantee the eventual boundedness of the consistency error. Attached Figure Description

[0015] The present invention will be further described below with reference to the accompanying drawings and embodiments. In the accompanying drawings: Appendix Figure 1 This is a block diagram of the algorithm execution logic of the present invention; Appendix Figure 2 The communication topology of a multi-AUV cluster system; Appendix Figure 3 There are 3 horizontal trajectories for AUVs; Appendix Figure 4 The simulation results show the consistency of the three AUVs in the x-direction; Appendix Figure 5 The simulation results show the consistency of the three AUVs in the y-direction; Appendix Figure 6 The simulation results show the consistency of heading angles for the three AUVs. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0017] In this example, the robust adaptive consistency control method of the present invention is used to achieve consistency control of a cluster system consisting of three AUVs.

[0018] Step 1: Establish a motion mathematical model of the virtual leader and followers in a multi-AUV cluster system, and rewrite it into a standard second-order system suitable for the design of a consensus control protocol.

[0019] Specifically, the motion model of the three AUVs in this embodiment is described as follows: (1) In the formula, This indicates the vehicle's position and heading in the geodetic coordinate system. These are the longitudinal and lateral linear velocities and yaw angular velocities of the vehicle. This is the state transition matrix from the carrier coordinate system to the geodetic coordinate system. The matrix represents the Coriolis force and the centripetal force. This is the damping force matrix. To control the input vectors, i.e., control force and torque, without loss of generality, we assume the three vehicles have the same model parameters. The specific forms of the mass matrix and damping force matrix are: , Furthermore, the AUV motion model is rewritten into a standard second-order system form suitable for the design of consensus control protocols.

[0020] (2) In the formula, , It is a nonlinear function of the vehicle's speed and heading. This is the transformed control input vector. Similarly, the motion model of the virtual leader can be described as: (3) in, satisfy Specifically, in this example It serves as a periodic reference signal.

[0021] Step 2: Determine the communication topology of the multi-AUV cluster system.

[0022] Specifically, in this example, only AUV 1 can receive the virtual leader's status information. AUV 1 communicates bidirectionally with AUV 2, and AUV 2 communicates bidirectionally with AUV 3. Based on this, it can be determined that a spanning tree exists with the virtual leader as its root node. Furthermore, the Laplace matrix and leader information matrix of the cluster system can be written as follows: (4) In addition, matrices can be defined. Given a spanning tree with the leader as the root node, it can be proven that the H matrix is ​​positive definite.

[0023] Step 3: Calculate the consistency error and filter consistency error for each AUV.

[0024] Specifically, in this example, the consistency error of each aircraft can be calculated as follows: (5) in, , Furthermore, the filter consistency error for each aircraft can be calculated as follows: (6) in, The adjustment parameter is greater than 0.

[0025] Step 4: Construct a robust neural damping term, and design a distributed consensus control protocol based on the robust neural damping term and adaptive control gain.

[0026] Specifically, in this example, a robust neural damping term vector is constructed for each follower vehicle based on the filter consistency error, which yields the following results: (7) in, This is an adaptive parameter vector. , Let be a radial basis function vector, where each radial basis function... The form can be described as (8) In the formula, Let be the input vector of the radial basis functions. and Let be the center vector and width of the radial basis functions. The integral-bounded function is used to ensure asymptotic consistency in multi-AUV systems. and The adjustment parameter is greater than 0.

[0027] Based on the constructed robust adaptive damping term, the following robust adaptive consensus control protocol can be designed: (9) In the formula, This is the time-varying control gain vector.

[0028] According to the robust adaptive consensus protocol, the actual control input vector of each vehicle before coordinate transformation can be obtained: (10) Step 5: Based on the consistency error of each aircraft obtained in Step 3, design the adaptive law of the adaptive parameters of the robust damping term, and the adaptive law of the time-varying control gain.

[0029] Specifically, in this example, the adaptive law design for the adaptive parameters of the robust adaptive damping term is as follows: (11) The adaptive law design for the time-varying control gain in the consensus protocol is as follows: (12) In the formula, and The adaptive law adjustment parameter is greater than 0 and is specified by the user.

[0030] Step 6: Adjust the relevant parameters of the aforementioned consistency control protocols and adaptive laws, and apply them to the consistency control of the multi-AUV cluster.

[0031] Specifically, in this example, the control parameters are selected as follows: Initial value of control gain: Initial values ​​for adaptive parameters: Adjustment parameters of the adaptive law: Other adjustment parameters: At each sampling time, each AUV collects its own state, interacts with neighboring AUVs according to a predetermined communication topology, calculates its actual control input, and updates the control gain and adaptive parameters based on the parameter adaptive law, ultimately achieving asymptotic consistency in the multi-AUV system. Simulation results are attached. Figure 3-6 As shown.

[0032] Figure 3-6 The horizontal trajectories of three AUVs, the consistency control results in the X direction, and the consistency control results in the Y direction are presented respectively. Simulation results show that the method proposed in this invention can enable multi-AUV cluster systems to accurately and quickly achieve consistency across states even with model uncertainties and unknown leader dynamics, demonstrating good robustness and adaptability.

[0033] The robust adaptive consistency control method for a multi-AUV distributed cluster in this embodiment has the following two main characteristics: (1) A radial basis function neural network is used to handle the uncertainty and unknown leader dynamics of the AUV model, and a minimum parameterization method is introduced. Robust neural damping technology is used to compress the number of adaptive parameters, requiring only 3 adaptive parameters to be learned online to stabilize the composite uncertainty. Therefore, this invention further reduces the computational burden while maintaining good robustness and adaptability, making it more suitable for engineering practice.

[0034] (2) The consensus protocol is designed using online updated time-varying control gain, thereby realizing a fully distributed design that does not require global information. An integral bounded function is introduced into the adaptive law, so that the consensus error asymptotically converges to 0. Compared with the traditional method that can only guarantee the boundedness of the consensus error, it has better control accuracy.

[0035] It should be noted that, as described in step 2, the method proposed in this invention requires bidirectional communication between AUVs in the AUV cluster system, and the existence of a spanning tree with a virtual leader as the root node. Based on meeting this requirement, by designing according to the steps in the specification and reasonably adjusting the adjustment parameters in each step, the same or similar control effect can still be achieved for AUV cluster systems with different model parameters than in this example. Furthermore, this invention only requires the AUVs to exchange position and velocity information. For consistency control problems in specific degrees of freedom, such as heading consistency problems and forward velocity consistency problems, it is only necessary to reduce the order of the vehicle mathematical model in step 1 and design according to the remaining steps. In this case, the vehicles only need to exchange state information in specific degrees of freedom. With an underwater acoustic communication system rate of not less than 100 bps, the information transmission requirements of the technical solution of this invention can be met.

[0036] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0037] It should be understood that those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. A robust adaptive consistency control method for multi-AUV distributed clusters, characterized in that, The method includes the following steps: Step 1: Establish a motion mathematical model of a virtual leader and multiple followers in a multi-AUV cluster system, and rewrite it into a standard second-order system suitable for the design of a consensus control protocol; Step 2: Determine the communication topology of the multi-AUV cluster system; Step 3: Calculate the consistency error and filter consistency error for each AUV; Step 4: Construct robust neural damping terms, and design a distributed consensus control protocol based on robust neural damping terms and adaptive control gain; The method for step 4 is as follows: The robust damping term is in the form of: in, For adaptive parameter vectors; , Let be a radial basis function vector, where each radial basis function... The formal description is as follows: In the formula, Let be the input vector of the radial basis functions. and The center vector and width of the radial basis functions; The integral-bounded function is used to ensure asymptotic consistency in multi-AUV systems. and The adjustment parameter is greater than 0; Based on the constructed robust adaptive damping term, the following robust adaptive consensus control protocol is designed: In the formula, The time-varying control gain vector is used; according to the robust adaptive consensus protocol, the actual control input vector of each vehicle before coordinate transformation is obtained: in, M i Represents the mass matrix, This is the state transition matrix from the carrier coordinate system to the geodetic coordinate system; Step 5: Based on the consistency error of each AUV obtained in Step 3, design the adaptive law of the adaptive parameters of the robust damping term, and the adaptive law of the time-varying control gain. Step 6: Adjust the parameters involved in the consistency control protocol and adaptive law, and apply them to the consistency control of the multi-AUV cluster.

2. The robust adaptive consistency control method for multi-AUV distributed clusters according to claim 1, characterized in that, The method for establishing the motion mathematical model consisting of one virtual leader and multiple followers in the multi-AUV cluster system in step 1 is as follows: A multi-AUV swarm system comprises one virtual leader and n follower AUVs; the motion model of the i-th AUV can be described as follows: In the formula, M i Represents the mass matrix, This indicates the vehicle's position and heading in the geodetic coordinate system. These are the longitudinal and lateral linear velocities and yaw angular velocities of the aircraft. This is the state transition matrix from the carrier coordinate system to the geodetic coordinate system; The matrix represents the Coriolis force and the centripetal force. The damping force matrix; This is to control the input vector, i.e., the control force and torque.

3. The robust adaptive consistency control method for multi-AUV distributed clusters according to claim 2, characterized in that, The method for rewriting the system into a standard second-order system in step 1 is as follows: The motion model of the AUV is converted into the following standard second-order nonlinear system: In the formula, , It is a nonlinear function of the vehicle's speed and heading; The transformed control input vector; similarly, the motion model of the virtual leader is described as follows: The virtual leader's course and speed are specified by the user based on the cluster task.

4. The robust adaptive consistency control method for multi-AUV distributed clusters according to claim 1, characterized in that, The method for step 2 is as follows: The communication structure of a multi-AUV cluster system is as follows: only some of the aircraft need to know the status information of the virtual leader; each follower aircraft communicates bidirectionally only with other aircraft within its communication range, forming a connected undirected graph; there is a directed path from the virtual leader to any follower, i.e., there exists a spanning tree with the virtual leader as the root node.

5. The robust adaptive consistency control method for multi-AUV distributed clusters according to claim 1, characterized in that, The calculation method for step 3 is as follows: Calculate the consistency error and filter consistency error for each AUV, specifically described as follows: in, This represents the communication weight between spacecraft i and spacecraft j; if they can communicate, then... ,on the contrary, ; This represents the communication weight between vehicle i and the virtual leader. If vehicle i can obtain information from the virtual leader, then... ,on the contrary, Based on this, the filtering consistency error of each follower aircraft is defined as: , in, The adjustment parameter is greater than 0.

6. The robust adaptive consistency control method for multi-AUV distributed clusters according to claim 1, characterized in that, The method for step 5 is as follows: The specific form of the adaptive law for controlling the gain is as follows: The adaptive law for the adaptive parameter in the robust adaptive damping term is as follows: In the formula, and The adaptive law adjustment parameter is greater than 0 and is specified by the user. and The adaptive law adjustment parameter is greater than 0 and is specified by the user.

7. The robust adaptive consistency control method for multi-AUV distributed clusters according to claim 1, characterized in that, The method for multi-AUV cluster consistency control in step 6 is as follows: At each sampling time, each AUV collects its own state, exchanges information with neighboring AUVs according to a predetermined communication topology, calculates its own actual control input, and updates the control gain and adaptive parameters according to the parameter adaptive law, ultimately enabling the multi-AUV system to achieve asymptotic consistency.

Images