CN-121995962-A - Method and device for controlling formation of underwater vehicle and electronic equipment

CN121995962ACN 121995962 ACN121995962 ACN 121995962ACN-121995962-A

Abstract

The invention relates to an underwater vehicle formation control method, an underwater vehicle formation control device and electronic equipment, wherein the method comprises the steps of constructing a formation configuration retaining frame based on a rigid graph theory, and determining the relative position relationship among all vehicles through distance constraint; the method comprises the steps of establishing an obstacle avoidance control model based on an artificial potential field method, calculating potential field force according to the distance between an aircraft and an obstacle and the distance between the aircraft to avoid collision between the obstacle and the aircraft, carrying out on-line identification and compensation on unmodeled dynamics and external disturbance in a dynamic model of the aircraft based on a radial basis function neural network, adopting a Gaussian function as an activation function by the radial basis function neural network, carrying out vector fusion on formation maintenance control force, the potential field force and compensation output of the radial basis function neural network, generating control instructions of each aircraft, and realizing formation control. The invention can realize formation high-precision maintenance, effective obstacle avoidance and disturbance self-adaptive compensation capability in a complex underwater environment.

Inventors

LI HONGFEI
ZHU DAQI
CHEN MINGZHI
MA YUHONG
YANG DONG
LIU JIANYU
LI LIUFANG

Assignees

上海理工大学

Dates

Publication Date: 20260508
Application Date: 20260410

Claims (10)

1. An underwater vehicle formation control method, the method comprising: Constructing a formation configuration holding frame based on a rigid graph theory, and determining the relative position relation among all the aircrafts by constructing a rigid matrix and applying a distance constraint to generate formation holding control force; An obstacle avoidance control model is established based on an artificial potential field method, potential field force is calculated according to the distance between an aircraft and an obstacle and the distance between the aircraft by adopting a continuously-conductive partial function structure between the minimum safe distance and the maximum induction radius, and the obstacle avoidance and the collision prevention between the aircraft are realized, wherein the partial function structure comprises the following specific steps of And potential field function between aircraft and obstacle , wherein, Representing the euclidean distance between the aircraft, Indicating the euclidean distance between the vehicle and the obstacle, In order for the safety distance to be a minimum, At the point of maximum radius of induction, Is the first Position vectors of the individual obstacles; On-line identification and self-adaptive compensation are carried out on unmodeled dynamics and external disturbance in the vehicle dynamics model based on a radial basis function neural network, the radial basis function neural network adopts a Gaussian function as an activation function, and the network weight is adjusted on line through a weight self-adaptive update law; And vector fusion is carried out on the formation maintaining control force, the potential field force and the compensation output of the radial basis function neural network, so that control instructions of all aircrafts are generated, and formation control is realized.
2. The underwater vehicle formation control method according to claim 1, wherein the constructing a formation configuration holding frame based on the rigid graph theory includes: defining an undirected graph Representing information interactions between aircraft, wherein For a set of nodes, For the collection of edges, Is a connection matrix; Construction of rigid matrices Wherein For a set of all of the vertex coordinates, Is an edge function; by connecting matrices Describing the relationship between vertices, where 。
3. The underwater vehicle formation control method according to claim 1, wherein the vector fusion includes: When the distance between the aircraft and the obstacle is smaller than the maximum induction radius, dynamically reducing the weight of the formation maintaining control force and increasing the weight of the potential field force to ensure that the obstacle avoidance response is maintained in preference to the formation, and simultaneously, the compensation output of the radial basis function neural network is used as a feedforward term to form a composite control structure with a feedback term formed by the potential field force so as to jointly correct the track of the aircraft.
4. The underwater vehicle formation control method according to claim 1, wherein the on-line identification and adaptive compensation of unmodeled dynamics and external disturbances in the vehicle dynamics model based on the radial basis function neural network comprises: Constructing radial basis function neural network approximation model Wherein As a matrix of the weights to be used in the ideal, Is the first The input vector of the individual aircraft is set, Is an approximation error; using Gaussian functions as radial basis functions Wherein As a center vector of the two-dimensional image, Is a width parameter; And outputting disturbance compensation quantity according to the approximation model.
5. The underwater vehicle formation control method according to claim 1, wherein the vector-fusing the compensation outputs of the formation holding control force, the potential field force, and the radial basis function neural network to generate the control command for each vehicle includes: Establishing an aircraft dynamics equation Wherein 、 The generalized matrixes are respectively an inertia matrix and a Coriolis centripetal force matrix after coordinate transformation, In the form of a generalized velocity vector, For a generalized control input, Is generalized external disturbance; Design control law Wherein As an approximation of the weights of the neural network, In the form of a gain matrix, Is a known kinetic term.
6. The method of claim 1, further comprising modeling the kinematics of the vehicle Wherein Indicating the position information and heading angle of the aircraft, Information representative of the speed of the aircraft, Is a rotation matrix.
7. The underwater vehicle formation control method of claim 6, wherein the rotation matrix Wherein Is the heading angle.
8. The underwater vehicle formation control method of claim 1, wherein the method further comprises modeling the dynamics of the aircraft Wherein Is a matrix of inertia which is a matrix of inertia, In the form of a coriolis centripetal force matrix, Is a hydrodynamic damping matrix, which is a dynamic damping matrix, For the purpose of an external disturbance, Is a control input.
9. The underwater vehicle formation control method according to claim 1, wherein the vector fusion further comprises a dual-threshold-based intelligent switching mechanism, specifically: Monitoring formation errors and disturbance estimated values in real time, and when the formation errors are larger than a preset first threshold value and the disturbance estimated values are smaller than a preset second threshold value, increasing the weight of formation holding control force and reducing the weight of potential field force so as to preferentially correct formation; When the formation error is smaller than the first threshold value and the distance between the aircraft and the obstacle is smaller than the maximum induction radius, the weight of the potential field force and the weight of the radial basis function neural network compensation output are increased, and the weight of the formation maintaining control force is reduced, so that the navigation safety is guaranteed preferentially.
10. The underwater vehicle formation control method of claim 4, wherein the online adjustment of the network weights by the weight adaptive update law includes: The design weight updating law is a product term of tracking error and hidden layer activation function output, and a correction term for inhibiting parameter drift is overlapped, so that the network weight estimated value of the radial basis function neural network is always maintained within a preset bounded range under the action of continuous external disturbance.

Description

Method and device for controlling formation of underwater vehicle and electronic equipment Technical Field The invention relates to the technical field of underwater vehicle formation control, in particular to an underwater vehicle formation control method, an underwater vehicle formation control device and electronic equipment. Background Autonomous underwater vehicles (Autonomous Underwater Vehicle, AUV) play an irreplaceable role in tasks such as subsea resource exploration, underwater search and rescue, marine environment monitoring, and the like. In order to meet the increasingly complex high-timeliness and large-scale task demands, the multi-AUV cooperative operation mode can remarkably improve the task execution efficiency and the system robustness compared with a single AUV. The stable and reliable formation control is a foundation for realizing multi-AUV cooperative operation, and is characterized in that firstly, a preset geometric configuration is maintained in a complex underwater environment, the formation generation and maintenance capacity is ensured, and secondly, coordinated control of overall movement of the formation is realized, so that the cluster can stably navigate along a desired track and adapt to environmental constraints. The underwater environment has strong uncertainty characteristics, including factors such as ocean current disturbance, underwater obstacle distribution, limited communication and the like, and the environmental factors form a serious challenge for formation control precision and stability. The conventional formation control method has the defects that a stable formation structure is difficult to maintain when external disturbance exists in the conventional formation control method, formation errors are large, convergence speed is low, effective coordination is not available between an existing obstacle avoidance control strategy and the formation control strategy, instability or track deviation is easy to occur in formation under an obstacle-dense environment, an adaptive compensation mechanism aiming at uncertainty of an underwater environment is imperfect, and control performance degradation caused by time-varying disturbance is difficult to solve by a fixed gain control method. Disclosure of Invention In view of the above, it is necessary to provide an underwater vehicle formation control method, apparatus and electronic device capable of realizing high-precision maintenance of formation, effective obstacle avoidance and disturbance adaptive compensation capability in a complex underwater environment. The invention provides an underwater vehicle formation control method, which comprises the following steps: Constructing a formation configuration holding frame based on a rigid graph theory, and determining the relative position relation among all the aircrafts by constructing a rigid matrix and applying a distance constraint to generate formation holding control force; An obstacle avoidance control model is established based on an artificial potential field method, potential field force is calculated according to the distance between an aircraft and an obstacle and the distance between the aircraft by adopting a continuously-conductive partial function structure between the minimum safe distance and the maximum induction radius, and the obstacle avoidance and the collision prevention between the aircraft are realized, wherein the partial function structure comprises the following specific steps of And potential field function between aircraft and obstacle, wherein,Representing the euclidean distance between the aircraft,Indicating the euclidean distance between the vehicle and the obstacle,In order for the safety distance to be a minimum,At the point of maximum radius of induction,Is the firstPosition vectors of the individual obstacles; On-line identification and self-adaptive compensation are carried out on unmodeled dynamics and external disturbance in the vehicle dynamics model based on a radial basis function neural network, the radial basis function neural network adopts a Gaussian function as an activation function, and the network weight is adjusted on line through a weight self-adaptive update law; And vector fusion is carried out on the formation maintaining control force, the potential field force and the compensation output of the radial basis function neural network, so that control instructions of all aircrafts are generated, and formation control is realized. In one embodiment, the constructing a formation configuration retention frame based on rigid graph theory includes: defining an undirected graph Representing information interactions between aircraft, whereinFor a set of nodes,For the collection of edges,Is a connection matrix; Construction of rigid matrices WhereinFor a set of all of the vertex coordinates,Is an edge function; by connecting matrices Describing the relationship between vertices, where。 In one embodiment, the vector fusion c