An underwater AUV-mounted RIS optical communication cooperative deployment and reflection regulation method and system

Abstract

The application provides an underwater AUV-mounted RIS optical communication cooperative deployment and reflection regulation method and system, and relates to the technical fields of underwater wireless optical communication and intelligent reflecting surface control. The method comprises the following steps: collecting node coordinates to construct a three-dimensional geometric model, calculating the position offset of the transmitting node caused by water disturbance based on a Gaussian water flow model, and solving state parameters such as cascade link signal-to-noise ratio and shielding probability in combination with factors such as water optical loss, turbulence and shielding; unknown motion disturbance is predicted by using a Gaussian process, a mobile time domain optimization model with adaptive robust constraints is constructed, and the optimal propulsion control sequence of the AUV is solved by rolling; the RIS reference normal vector is calculated based on the geometric relationship of the light path, the DDPG algorithm is used to output a bounded residual correction amount, and the reflection control vector is obtained by superposition and normalization; the sailing and angle instructions are synchronously issued, and the new data is collected for closed-loop iteration. The application cooperatively optimizes the position of the AUV and the reflection direction of the RIS, effectively improves the stability of underwater optical communication, and is suitable for underwater unmanned networking scenes.

Description

Technical Field

[0001] This invention relates to the field of underwater wireless optical communication and intelligent reflector control technology, specifically to a method and system for coordinated deployment and reflection control of RIS optical communication on an underwater AUV. Background Technology

[0002] Underwater wireless optical communication (UWOC) boasts advantages such as high transmission rate, low latency, and strong resistance to electromagnetic interference. However, optical signals are susceptible to absorption, scattering, turbulence, and obstruction when propagating in water, leading to decreased received power and reduced link stability. Intelligent reflectors can construct auxiliary optical transmission links by adjusting the reflection direction; however, existing intelligent reflectors are mostly deployed in a fixed manner, making it difficult to adapt to the dynamic movement of underwater optical transmitting nodes. While mounting intelligent reflectors on autonomous underwater vehicles (AUVs) allows for position adjustment, their trajectory planning and reflection control must simultaneously satisfy constraints such as range of motion, speed, signal-to-noise ratio, and optical reflection laws. Existing methods do not adequately consider underwater random disturbances and the future motion state of nodes, easily resulting in deployment lag, reflection mismatch, and communication interruptions. Therefore, it is necessary to propose a dynamic deployment and reflection coordinated control method for underwater intelligent reflectors. Summary of the Invention

[0003] The purpose of this invention is to provide a method and system for coordinated deployment and reflection control of RIS optical communication on an underwater AUV, so as to solve at least one technical problem existing in the prior art.

[0004] On one hand, embodiments of the present invention provide a method for coordinated deployment and reflection control of RIS optical communication on an underwater AUV, the method comprising: step S1, receiving a preset position of an underwater optical transmitting node, a measured position of an autonomous underwater vehicle equipped with a smart reflector, and a fixed position of an optical receiving node as inputs to establish a three-dimensional geometric model of underwater intelligent reflector-assisted optical communication; step S2, calculating the position offset of the underwater optical transmitting node based on water disturbance information, and combining the position offset with the preset position of the underwater optical transmitting node to calculate the actual position of the underwater optical transmitting node; step S3, based on the measured position of the autonomous underwater vehicle... Step S4: Establish a cascaded optical link model between the underwater optical transmitter, the smart reflector, and the optical receiver at the fixed positions of the optical receiver and the actual positions of the underwater optical transmitter. Output the total cascaded propagation distance, the signal-to-noise ratio of the cascaded optical link, and the cascaded occlusion probability. Step S5: Construct the motion state matrix of the autonomous underwater vehicle based on the measured position of the autonomous underwater vehicle. Construct Gaussian process input features based on the motion state matrix, the total cascaded propagation distance, the signal-to-noise ratio of the cascaded optical link, and the cascaded occlusion probability. Step S6: Based on the Gaussian process input features, predict the unknown motion disturbances caused by the water body and the propulsion mechanism through a Gaussian process, and introduce a loop. A mobile time-domain multi-objective optimization model is constructed using adaptive weights and communication security constraints. The optimal motion sequence for AUV's dynamic global deployment, adapted to real-time link conditions, is solved using a rolling algorithm, outputting the optimal propulsion control input for the autonomous underwater vehicle at the current moment. Step S6: Based on the actual position of the underwater optical transmitter node, the measured position of the autonomous underwater vehicle equipped with the intelligent reflector, and the fixed position of the optical receiver node, the geometric reference normal vector of the intelligent reflector is calculated. A depth-deterministic gradient algorithm is used to solve for the bounded residual correction amount corresponding to the geometric reference normal vector. The geometric reference normal vector and the bounded residual correction amount are then superimposed and normalized to output the optimal propulsion control input for the intelligent reflector. Final normal vector; Step S7: Using the optimal propulsion control input of the autonomous underwater vehicle output in Step S5 and the final normal vector of the intelligent reflector output in Step S6 as inputs, send them to the propulsion device of the autonomous underwater vehicle and the angle adjustment mechanism of the intelligent reflector to perform dynamic deployment adjustment of the AUV and synchronous control of the reflection angle; Step S8: Obtain the motion state of the autonomous underwater vehicle, the position of the underwater optical emission node, the signal-to-noise ratio of the cascaded optical link, the intensity of water disturbance and the link blockage state after the control is executed, and re-execute steps S1 to S7 to achieve continuous coordinated adjustment of the position of the autonomous underwater vehicle and the reflection direction of the intelligent reflector.

[0005] Furthermore, step S1, receiving the preset position of the underwater optical transmitting node, the measured position of the autonomous underwater vehicle equipped with the intelligent reflector, and the fixed position of the optical receiving node as inputs, and establishing a three-dimensional geometric model of underwater intelligent reflector-assisted optical communication, includes: representing the actual position of the underwater optical transmitting node as the sum of the preset position and the position offset caused by water disturbance. ; in, This represents the actual position of the underwater optical emission node at time t. This represents the preset position of the underwater optical emission node at time t. The position shift caused by water disturbance at time t; The location of the autonomous underwater vehicle equipped with a smart reflector is represented as: ; in, These are the position coordinates of the autonomous underwater vehicle along the three coordinate axes at time t; The location of the optical receiving node is represented as follows: ; in, These are the position coordinates of the optical receiving node along the three coordinate axes.

[0006] Furthermore, step S2, calculating the position offset of the underwater optical emitting node based on the water disturbance information, and combining the position offset with the preset position of the underwater optical emitting node to calculate the actual position of the underwater optical emitting node, includes: The water velocity perturbation acting on the underwater optical emission node is represented as a zero-mean random vector: ; in, Let be the three-dimensional water velocity perturbation vector acting on the underwater optical emission node at time t. Indicates a Gaussian distribution. Let be the covariance matrix of the water velocity perturbation at time t; The covariance matrix is ​​expressed as: ; in, The intensity of water disturbance. This is the normalized directional covariance matrix; The perturbation force acting on the underwater optical emission node is calculated based on the water velocity disturbance: ; in, For water density, The drag coefficient, The effective projected area of ​​the underwater optical emission node and its onboard equipment; Converting water disturbance forces into position offsets for underwater optical emission nodes: ; in, is the conversion coefficient from disturbance force to position offset, and M is the equivalent mass of the underwater optical transmitting node and its onboard equipment.

[0007] Furthermore, step S3, establishing a cascaded optical link model between the underwater optical transmitter, the smart reflector, and the optical receiver based on the measured position of the autonomous underwater vehicle, the fixed position of the optical receiver node, and the actual position of the underwater optical transmitter node, and outputting the total cascaded propagation distance, the cascaded optical link signal-to-noise ratio, and the cascaded obstruction probability, includes: Calculate the first propagation distance from the underwater optical emitting node to the smart reflector. : ; Calculate the second propagation distance from the smart reflector to the light receiving node. : ; Cascade propagation distance of underwater intelligent reflector-assisted optical communication links for: ; According to the cascade propagation distance By combining water absorption, scattering, turbulent scintillation, obstacle occlusion, and the reflection efficiency of the smart reflector, the signal-to-noise ratio of the smart reflector-assisted cascaded optical link is calculated. and cascading occlusion probability .

[0008] Furthermore, step S4, constructing the motion state matrix of the autonomous underwater vehicle based on its measured position, and constructing Gaussian process input features based on the motion state matrix, cascaded total propagation distance, cascaded optical link signal-to-noise ratio, and cascaded occlusion probability, includes: Constructing the augmented motion state of an autonomous underwater vehicle: ; in, For autonomous underwater vehicles at all times speed, The residual state is caused by local water disturbance, propulsion error, and unmodeled dynamics. Gaussian process input features are constructed based on the autonomous underwater vehicle's motion state, control input, cascade propagation distance, link signal-to-noise ratio, turbulent scintillation intensity, cascade blockage probability, and reflection geometric error. : ; in, To construct a function by inputting features into a Gaussian process, For the control input of autonomous underwater vehicles, The turbulent scintillation intensity of the cascaded optical link. For reflection geometric error; The input features are then input into the Gaussian process perturbation prediction model to obtain the perturbation prediction mean and perturbation prediction covariance: ; in, The training dataset is for Gaussian processes. The mean of the disturbance prediction, To predict the covariance of the perturbation, This is a pre-trained Gaussian process perturbation prediction model.

[0009] Furthermore, in step S5, based on the Gaussian process input characteristics, the unknown motion disturbances caused by the water body and propulsion mechanism are predicted through the Gaussian process. An environmental adaptive weight and communication security constraints are introduced to build a mobile time-domain multi-objective optimization model. This model is then used to continuously solve for the optimal motion sequence for AUV global dynamic deployment, adapting to the real-time link status. The output of the optimal propulsion control input for the autonomous underwater vehicle at the current moment includes: A state model of the autonomous underwater vehicle in the prediction time domain is established based on the mean of the disturbance prediction: ; in, This is the nominal motion model of an autonomous underwater vehicle that does not consider residual disturbances from the Gaussian process. The perturbation input mapping matrix; An environment-adaptive weighting is constructed based on turbulent scintillation intensity, cascade occlusion probability, distance feedback error, and Gaussian process prediction uncertainty. ; in, , , and These are non-negative weighting coefficients. To provide feedback on the error between the reconstructed distance and the geometrically predicted distance; A multi-objective optimization function in the time domain is established, with the optimization objectives being cascade propagation distance, link feedback error, insufficient signal-to-noise ratio, cascade blockage risk, autonomous underwater vehicle control input, and trajectory smoothness: ; in, To predict the length of the time domain, The error between the measured link feedback and the predicted link feedback, For the minimum permissible signal-to-noise ratio, To maximize the probability of cascading occlusion, , , , , and These are the weighting coefficients for the corresponding target items; The multi-objective optimization function satisfies communication security constraints: ; ; ; ; ; in, The permitted movement area for autonomous underwater vehicles. For the set of allowed states, and This is a tightened set of uncertainties constructed based on the covariance predicted by the Gaussian process. For set difference operation, For the maximum permissible control input, and This refers to the safety margin factor. After solving the finite prediction time-domain optimization problem at each control time step, only the first control input in the current optimization sequence is executed: ; Based on the updated motion state of the autonomous underwater vehicle and the rolling update prediction time domain of the cascaded optical link feedback, the optimal motion trajectory of the autonomous underwater vehicle equipped with the intelligent reflector is obtained: ; in, for The optimal propulsion control input for the AUV is ultimately issued and executed at any given time. This is the optimal control sequence obtained by constrained optimization. This is the optimal deployment trajectory set for AUVs across the entire domain. The optimal three-dimensional position of the AUV is obtained after optimization at time t.

[0010] Furthermore, step S6, which involves calculating the geometric reference normal vector of the intelligent reflector based on the actual position of the underwater optical emitting node, the measured position of the autonomous underwater vehicle equipped with the intelligent reflector, and the fixed position of the optical receiving node, and then using a depth-deterministic gradient algorithm to solve for the bounded residual correction amount corresponding to the geometric reference normal vector, and finally outputting the final normal vector of the intelligent reflector after superimposing and normalizing the geometric reference normal vector and the bounded residual correction amount, includes: Based on the spatial relationship between the underwater light emitting node, the smart reflector, and the light receiving node, a geometric normal vector mapping is established: ; The geometric reference normal vector at the current moment is: ; The predicted geometric normal vector for the next time step is: ; The current geometric reference normal vector, the predicted geometric normal vector for the next time step, the cascaded optical link state, and the smart reflector normal vector executed in the previous time step are used to construct the state vector of the deep deterministic policy gradient algorithm: ; The state characteristics of the cascaded optical link are as follows: ; The deep deterministic policy gradient algorithm outputs the residual correction relative to the geometric reference normal vector: ; in, The control actions output by the deep deterministic policy gradient algorithm. To form a collection of actionable actions This is the residual correction amount relative to the geometric reference normal vector; The set of possible actions for the residual correction is as follows: ; in, The maximum permissible residual correction magnitude, This represents the maximum permissible angular deviation of the normal vector after residual correction relative to the geometric reference normal vector. The geometric reference normal vector is added to the residual correction and projected onto the unit sphere to obtain the final normal vector of the intelligent reflective surface. : ; The unit spherical projection function is: .

[0011] Furthermore, the instantaneous reward function of the deep deterministic policy gradient algorithm is: ; in, , , , and The non-negative weighting coefficients represent the reflection geometric error as follows: ; The instant reward function is used to improve the signal-to-noise ratio of the intelligent reflector-assisted cascaded optical link and suppress link occlusion, turbulent scintillation, reflection geometry mismatch, and abrupt changes in reflection direction between adjacent time moments.

[0012] Furthermore, step S7, using the optimal propulsion control input of the autonomous underwater vehicle output in step S5 and the final normal vector of the intelligent reflector output in step S6 as inputs, and sending them to the propulsion device of the autonomous underwater vehicle and the angle adjustment mechanism of the intelligent reflector to perform dynamic deployment adjustment and synchronous control of the reflection angle of the AUV, includes: sending the current optimal control input obtained in step S5 to the propulsion control device of the autonomous underwater vehicle, converting the final normal vector obtained in step S6 into azimuth and pitch angle control commands of the intelligent reflector, and sending them to the reflection angle adjustment device of the intelligent reflector.

[0013] Secondly, embodiments of the present invention provide an underwater AUV equipped with a RIS optical communication collaborative deployment and reflection control system. The system is implemented using the aforementioned underwater AUV equipped with a RIS optical communication collaborative deployment and reflection control method. The system includes: a geometric modeling module, adapted to receive a preset position of the underwater optical transmitting node, the measured position of the autonomous underwater vehicle equipped with a smart reflector, and the fixed position of the optical receiving node as inputs to establish a three-dimensional geometric model of underwater smart reflector-assisted optical communication; and an underwater optical transmitting node actual position calculation module, adapted to calculate the position offset of the underwater optical transmitting node based on water disturbance information, and combine the position offset with the preset position of the underwater optical transmitting node to calculate the underwater optical transmitting node. The system includes: a cascaded optical link modeling module, suitable for establishing a cascaded optical link model between the underwater optical transmitter, smart reflector, and optical receiver based on the measured position of the autonomous underwater vehicle, the fixed position of the optical receiver node, and the actual position of the underwater optical transmitter node; outputting the total cascaded propagation distance, the cascaded optical link signal-to-noise ratio, and the cascaded occlusion probability; a Gaussian process feature construction and disturbance prediction module, suitable for constructing the motion state matrix of the autonomous underwater vehicle based on its measured position, and constructing Gaussian process input features based on the motion state matrix, the total cascaded propagation distance, the cascaded optical link signal-to-noise ratio, and the cascaded occlusion probability; and a robust rolling time-domain optimization module, suitable for optimizing the Gaussian process input features using a Gaussian process. The system predicts unknown motion disturbances caused by water bodies and propulsion mechanisms, introduces environmental adaptive weights and communication security constraints to build a multi-objective optimization model in the mobile time domain, and continuously solves for the optimal motion sequence for AUV's dynamic global deployment, adapting to real-time link status, and outputs the optimal propulsion control input for the autonomous underwater vehicle at the current moment. The RIS reflection normal vector calculation module is suitable for calculating the geometric reference normal vector of the intelligent reflector based on the actual position of the underwater optical transmitter node, the measured position of the autonomous underwater vehicle equipped with the intelligent reflector, and the fixed position of the optical receiver node. It uses a depth-deterministic gradient algorithm to solve for the bounded residual correction amount corresponding to the geometric reference normal vector, and outputs the intelligent reflector after superimposing and normalizing the geometric reference normal vector and the bounded residual correction amount. The final normal vector; the execution drive module, which is suitable for using the optimal propulsion control input of the autonomous underwater vehicle (AUV) output by the robust rolling time-domain optimization module and the final normal vector of the intelligent reflector output by the RIS reflection normal vector solving module as inputs, and sending them to the AUV propulsion device and the intelligent reflector angle adjustment mechanism to perform dynamic deployment adjustment and synchronous control of reflection angle; the data acquisition closed-loop feedback module, which is suitable for acquiring the motion state of the AUV, the position of the underwater optical emission node, the signal-to-noise ratio of the cascaded optical link, the intensity of water disturbance and the link blockage state after the control is executed, and re-executing the geometric modeling module to the execution drive module to achieve continuous and coordinated adjustment of the AUV position and the reflection direction of the intelligent reflector.

[0014] Thirdly, embodiments of the present invention also provide an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the above-described underwater AUV-mounted RIS optical communication collaborative deployment and reflection control method.

[0015] Fourthly, embodiments of the present invention also provide a readable storage medium, which, when the instructions in the storage medium are executed by the processor of an electronic device, enables the electronic device to execute the above-described underwater AUV-mounted RIS optical communication collaborative deployment and reflection control method.

[0016] The beneficial effects of this invention are as follows: By establishing a three-dimensional geometric model between the underwater optical transmitting node, the autonomous underwater vehicle equipped with a smart reflector, and the optical receiving node, the preset motion position of the underwater optical transmitting node is combined with the position offset caused by water disturbance, which can characterize the actual motion state of the underwater optical transmitting node in a dynamic water environment; by predicting the unknown motion disturbances experienced by the autonomous underwater vehicle through a Gaussian process, and introducing the predicted mean and predicted covariance of the disturbance into the moving time domain optimization process, the trajectory of the autonomous underwater vehicle can be adjusted according to the future motion state of the node and the uncertainty of the disturbance; by combining the geometric reference normal vector with the bounded residual correction amount output by the depth deterministic strategy gradient algorithm, the reflection direction deviation caused by water disturbance, link obstruction, and position prediction error can be compensated while satisfying the optical reflection law, thereby achieving the coordinated optimization of the deployment position and reflection direction of the underwater smart reflector. Attached Figure Description

[0017] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0018] Figure 1 This is a schematic diagram of a method for coordinated deployment and reflection control of an underwater AUV equipped with RIS optical communication, provided in Embodiment 1 of the present invention.

[0019] Figure 2 This is a schematic diagram of the three-dimensional geometric optical path of the underwater transmitting node-AUV(RIS)-receiving node at time t1 provided in Embodiment 1 of the present invention.

[0020] Figure 3 This is a schematic diagram of the three-dimensional geometric optical path of the underwater transmitting node-AUV(RIS)-receiving node at time t2 provided in Embodiment 1 of the present invention.

[0021] Figure 4 This is a schematic diagram of the geometric modeling principle of a three-dimensional unified coordinate system for underwater optical communication provided in Embodiment 1 of the present invention.

[0022] Figure 5This is a schematic diagram of the structure of an underwater AUV equipped with a RIS optical communication collaborative deployment and reflection control system provided in Embodiment 2 of the present invention.

[0023] Figure 6 This is a partial block diagram of the electronic device provided in Embodiment 3 of the present invention. Detailed Implementation

[0024] Before discussing the exemplary embodiments in more detail, it should be noted that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe the operations as sequential processes, many of these operations can be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but may also have additional steps not included in the figures. The process can correspond to a method, function, procedure, subroutine, subroutine, etc.

[0025] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0026] The present invention will now be described in detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0027] For ease of understanding, the following technical terms are explained here: AUV: Autonomous Underwater Vehicle; RIS: Reconfigurable Intelligent Surface.

[0028] Example 1 The specific implementation method is as follows: like Figure 1 The diagram shown is a flowchart of a method for coordinated deployment and reflection control of RIS optical communication on an underwater AUV provided by the present invention.

[0029] As an example, the method includes: Step S1, receiving the preset position of the underwater optical transmitting node, the measured position of the autonomous underwater vehicle equipped with a smart reflector, and the fixed position of the optical receiving node as inputs, and establishing a three-dimensional geometric model of underwater smart reflector-assisted optical communication; Step S2, calculating the position offset of the underwater optical transmitting node based on water disturbance information, and combining the position offset with the preset position of the underwater optical transmitting node to calculate the actual position of the underwater optical transmitting node; Step S3, establishing a three-dimensional geometric model of underwater smart reflector-assisted optical communication based on the measured position of the autonomous underwater vehicle, the fixed position of the optical receiving node, and the actual position of the underwater optical transmitting node. A cascaded optical link model is established between the optical transmitting node, the intelligent reflector, and the optical receiving node, outputting the total cascaded propagation distance, the cascaded optical link signal-to-noise ratio, and the cascaded occlusion probability. Step S4: Based on the measured position of the autonomous underwater vehicle (AUV), a motion state matrix is ​​constructed. Based on the motion state matrix, the total cascaded propagation distance, the cascaded optical link signal-to-noise ratio, and the cascaded occlusion probability, Gaussian process input features are constructed. Step S5: Based on the Gaussian process input features, unknown motion disturbances caused by the water body and propulsion mechanism are predicted using a Gaussian process. Environmental adaptive weights and communication security constraints are introduced to build a mobile... A time-domain multi-objective optimization model is used to continuously solve for the optimal motion sequence for AUV's global dynamic deployment, adapting to the real-time link status, and outputting the optimal propulsion control input for the autonomous underwater vehicle at the current moment; Step S6: Based on the actual position of the underwater optical transmitting node, the measured position of the autonomous underwater vehicle equipped with the intelligent reflector, and the fixed position of the optical receiving node, the geometric reference normal vector of the intelligent reflector is calculated. The bounded residual correction amount corresponding to the geometric reference normal vector is solved using a depth deterministic strategy gradient algorithm. The geometric reference normal vector and the bounded residual correction amount are superimposed and normalized to output the final normal vector of the intelligent reflector; Step S 7. Using the optimal propulsion control input of the autonomous underwater vehicle (AUV) output in step S5 and the final normal vector of the intelligent reflector output in step S6 as inputs, the data is sent to the AUV propulsion device and the intelligent reflector angle adjustment mechanism to perform dynamic deployment adjustment and synchronous control of the reflection angle; Step S8: After the control is executed, the motion state of the AUV, the position of the underwater optical emission node, the signal-to-noise ratio of the cascaded optical link, the water disturbance intensity and the link blockage state are obtained, and steps S1 to S7 are executed again to achieve continuous and coordinated adjustment of the position of the AUV and the reflection direction of the intelligent reflector.

[0030] Preferred, such as Figure 2 and Figure 3 As shown, the underwater optical communication system in this embodiment includes an underwater optical transmitting node, an autonomous underwater vehicle equipped with a smart reflector, and an optical receiving node. The optical signal emitted by the underwater optical transmitting node first propagates to the smart reflector mounted on the autonomous underwater vehicle, and then the smart reflector adjusts the reflection direction and reflects the optical signal to the optical receiving node, thereby forming a cascaded optical link between the underwater optical transmitting node, the smart reflector, and the optical receiving node.

[0031] Preferred, such as Figure 4 As shown, in a unified three-dimensional Cartesian coordinate system, the actual position of the underwater optical emission node is represented as the sum of the preset motion position and the position offset caused by water disturbance: ; in, This represents the actual position of the underwater optical emission node at time t. This represents the preset position of the underwater optical emission node at time t. Let be the position shift caused by water disturbance at time t.

[0032] The positions of the autonomous underwater vehicle equipped with a smart reflector and the optical receiver node are represented as follows: ; ; in, , and These represent the position coordinates of the autonomous underwater vehicle along the three coordinate axes. , and These are the position coordinates of the optical receiving node. Through the above modeling, the spatial relationship between the three types of nodes in a unified coordinate system is determined, providing a geometric basis for subsequent position offset calculations, cascaded optical link modeling, autonomous underwater vehicle trajectory optimization, and intelligent reflector reflection direction control.

[0033] Preferably, step S2, calculating the position offset of the underwater optical transmitting node based on the water disturbance information, and combining the position offset with the preset position of the underwater optical transmitting node to calculate the actual position of the underwater optical transmitting node, includes: The water velocity perturbation acting on the underwater optical emission node is represented as a zero-mean random vector: ; in, Let be the three-dimensional water velocity perturbation vector acting on the underwater optical emission node at time t. Indicates a Gaussian distribution. Let be the covariance matrix of the water velocity perturbation at time t; The covariance matrix is ​​expressed as follows: ; in, The intensity of water disturbance. This is the normalized directional covariance matrix; The perturbation force acting on the underwater optical emission node is calculated based on the water velocity disturbance: ; in, For water density, The drag coefficient, The effective projected area of ​​the underwater optical emission node and its onboard equipment; Converting water disturbance forces into position offsets for underwater optical emission nodes: ; in, is the conversion coefficient from disturbance force to position offset, and M is the equivalent mass of the underwater optical transmitting node and its onboard equipment.

[0034] In some feasible implementations, step S3, establishing a cascaded optical link model between the underwater optical transmitter, the smart reflector, and the optical receiver based on the measured position of the autonomous underwater vehicle, the fixed position of the optical receiver node, and the actual position of the underwater optical transmitter node, and outputting the total cascaded propagation distance, the cascaded optical link signal-to-noise ratio, and the cascaded obstruction probability, includes: Calculate the first propagation distance from the underwater optical emitting node to the smart reflector. : ; Calculate the second propagation distance from the smart reflector to the light receiving node. : ; Cascade propagation distance of underwater intelligent reflector-assisted optical communication links for: ; According to the cascade propagation distance By combining water absorption, scattering, turbulent scintillation, obstacle occlusion, and the reflection efficiency of the smart reflector, the signal-to-noise ratio of the smart reflector-assisted cascaded optical link is calculated. and cascading occlusion probability .

[0035] Specifically, after obtaining the propagation distances of the two optical paths, the signal-to-noise ratio of the entire cascaded optical link is quantitatively calculated by considering five major optical interference conditions: seawater absorption loss, suspended particle scattering loss, light intensity scintillation loss caused by underwater turbulence, reflection loss from the intelligent reflective panel, and optical path obstruction by underwater obstacles. and cascading occlusion probability .

[0036] In some feasible implementations, step S4, constructing the motion state matrix of the autonomous underwater vehicle based on its measured position, and constructing Gaussian process input features based on the motion state matrix, cascaded total propagation distance, cascaded optical link signal-to-noise ratio, and cascaded occlusion probability, includes: Constructing the augmented motion state of an autonomous underwater vehicle: ; in, For autonomous underwater vehicles at all times speed, The residual state is caused by local water disturbance, propulsion error, and unmodeled dynamics. Gaussian process input features are constructed based on the autonomous underwater vehicle's motion state, control input, cascade propagation distance, link signal-to-noise ratio, turbulent scintillation intensity, cascade blockage probability, and reflection geometric error. : ; in, To construct a function by inputting features into a Gaussian process, For the control input of autonomous underwater vehicles, The turbulent scintillation intensity of the cascaded optical link. For reflection geometric error; The input features are then input into the Gaussian process perturbation prediction model to obtain the perturbation prediction mean and perturbation prediction covariance: ; in, The training dataset is for Gaussian processes. The mean of the disturbance prediction, To predict the covariance of the perturbation, This is a pre-trained Gaussian process perturbation prediction model. The predicted mean is used to compensate for the nominal motion model of the autonomous underwater vehicle, and the predicted covariance is used to characterize the uncertainty of the perturbation prediction.

[0037] In some feasible implementations, step S5, based on the Gaussian process input characteristics, predicts unknown motion disturbances caused by the water body and propulsion mechanism through the Gaussian process, introduces environmental adaptive weights and communication security constraints to build a mobile time-domain multi-objective optimization model, and continuously solves for the optimal motion sequence of AUV global dynamic deployment adapted to the real-time link state, outputting the optimal propulsion control input of the autonomous underwater vehicle at the current moment, including: A state model of the autonomous underwater vehicle in the prediction time domain is established based on the mean of the disturbance prediction: ; in, This is the nominal motion model of an autonomous underwater vehicle that does not consider residual disturbances from the Gaussian process. The perturbation input mapping matrix; An environment-adaptive weighting is constructed based on turbulent scintillation intensity, cascade occlusion probability, distance feedback error, and Gaussian process prediction uncertainty. ; in, , , and These are non-negative weighting coefficients. To provide feedback on the error between the reconstructed distance and the geometrically predicted distance; A multi-objective optimization function in the time domain is established, with the optimization objectives being cascade propagation distance, link feedback error, insufficient signal-to-noise ratio, cascade blockage risk, autonomous underwater vehicle control input, and trajectory smoothness. This function includes: at the current time... Establish a length of Finite prediction time-domain optimization problem: ; in, To predict the length of the time domain, The error between the measured link feedback and the predicted link feedback, For the minimum permissible signal-to-noise ratio, To maximize the probability of cascading occlusion, , , , , and These are the weighting coefficients for the corresponding objective terms; the optimization satisfies constraints on the autonomous underwater vehicle's motion area, maximum speed, maximum acceleration, control input, minimum signal-to-noise ratio, and maximum obstruction probability, including: ; ; ; ; ; in, The permitted movement area for autonomous underwater vehicles. For the set of allowed states, and This is a tightened set of uncertainties constructed based on the covariance predicted by the Gaussian process. For set difference operation, For the maximum permissible control input, and This refers to the safety margin factor. After solving the finite prediction time-domain optimization problem at each control time step, only the first control input in the current optimization sequence is executed: ; Based on the updated motion state of the autonomous underwater vehicle and the rolling update prediction time domain of the cascaded optical link feedback, the optimal motion trajectory of the autonomous underwater vehicle equipped with the intelligent reflector is obtained: ; in, for The optimal propulsion control input for the AUV is finally issued and executed at any given time. This is the optimal control sequence obtained by constrained optimization. This is the optimal deployment trajectory set for AUVs across the entire domain. The optimal three-dimensional position of the AUV is obtained after optimization at time t.

[0038] In some feasible implementations, step S6, calculating the geometric reference normal vector of the intelligent reflector based on the actual position of the underwater optical emitting node, the measured position of the autonomous underwater vehicle equipped with the intelligent reflector, and the fixed position of the optical receiving node, and using a depth-deterministic gradient algorithm to solve for the bounded residual correction amount corresponding to the geometric reference normal vector, and then outputting the final normal vector of the intelligent reflector after superimposing and normalizing the geometric reference normal vector and the bounded residual correction amount, includes: Based on the spatial relationship between the underwater light emitting node, the smart reflector, and the light receiving node, a geometric normal vector mapping is established: ; The geometric reference normal vector at the current moment is: ; The predicted geometric normal vector for the next time step is: ; The current geometric reference normal vector, the predicted geometric normal vector for the next time step, the cascaded optical link state, and the smart reflector normal vector executed in the previous time step are used to construct the state vector of the deep deterministic policy gradient algorithm: ; The state characteristics of the cascaded optical link are as follows: ; The deep deterministic policy gradient algorithm outputs the residual correction relative to the geometric reference normal vector: ; in, The control actions output by the deep deterministic policy gradient algorithm. To form a collection of actionable actions This is the residual correction amount relative to the geometric reference normal vector; The set of possible actions for the residual correction is as follows: ; in, The maximum permissible residual correction magnitude, This represents the maximum permissible angular deviation of the normal vector after residual correction relative to the geometric reference normal vector. The geometric reference normal vector is added to the residual correction and projected onto the unit sphere to obtain the final normal vector of the intelligent reflective surface. : ; The unit spherical projection function is: .

[0039] In some feasible implementations, the instantaneous reward function of the deep deterministic policy gradient algorithm is: ; in, , , , and The non-negative weighting coefficients represent the reflection geometric error as follows: ; The instant reward function is used to improve the signal-to-noise ratio of the intelligent reflector-assisted cascaded optical link and suppress link occlusion, turbulent scintillation, reflection geometry mismatch, and abrupt changes in reflection direction between adjacent time moments.

[0040] The algorithm's reward function comprehensively considers the signal-to-noise ratio of the cascaded optical link, occlusion probability, turbulent scintillation intensity, reflection geometric error, and the change in normal vector between adjacent time points. It is trained through an actor network, an evaluation network, an experience replay network, and a target network. After training is complete, the online control phase only calls the bounded residual correction amount output by the actor network, thereby compensating for reflection deviations caused by water disturbance, position prediction errors, and changes in link state while maintaining the feasibility of optical reflection.

[0041] In some feasible implementations, step S7, using the optimal propulsion control input of the autonomous underwater vehicle output in step S5 and the final normal vector of the intelligent reflector output in step S6 as inputs, and sending them to the propulsion device of the autonomous underwater vehicle and the angle adjustment mechanism of the intelligent reflector to perform dynamic deployment adjustment and synchronous control of the reflection angle of the AUV, includes: sending the current optimal control input obtained in step S5 to the propulsion control device of the autonomous underwater vehicle, converting the final normal vector obtained in step S6 into azimuth and pitch angle control commands of the intelligent reflector, and sending them to the reflection angle adjustment device of the intelligent reflector.

[0042] Specifically, the current optimal control input obtained in step S5 is sent to the propulsion control module of the autonomous underwater vehicle (AUV) to control the AUV to adjust its direction of motion, speed, and spatial position within the permitted activity area, so that its onboard intelligent reflector moves to the optimized position corresponding to the current control cycle. The temporal optimization executes only the first optimal control input in the obtained control sequence each time, thus enabling timely correction of the trajectory based on subsequent feedback.

[0043] Specifically, the final normal vector obtained in step S6 is sent to the control module of the smart reflector. The control module converts the final normal vector into a corresponding reflection angle control command according to the installation direction of the smart reflector and the calibration relationship of the actuator, and drives the micro actuator to adjust the spatial orientation of the activated reflection unit so that the light signal emitted by the underwater light emitting node is reflected by the smart reflector and points to the light receiving node.

[0044] In some feasible implementations, step S8 includes: after completing the current control cycle, reacquiring the actual position of the underwater optical transmitting node, the position and velocity of the autonomous underwater vehicle, the cascade propagation distance, the link signal-to-noise ratio, the turbulent scintillation intensity, the occlusion probability, and the reflection geometric error, and feeding back the updated node status and cascade optical link status to step S1. By cyclically executing steps S1 to S7, a closed-loop process of position acquisition, disturbance calculation, trajectory optimization, reflection correction, control execution, and status feedback is formed, realizing continuous and coordinated adjustment of the autonomous underwater vehicle deployment position and the reflection direction of the intelligent reflector.

[0045] In some feasible implementations, the working principle of the dynamic deployment and reflection control method of the autonomous underwater vehicle equipped with a smart reflector in this embodiment is as follows: First, the positional information of the underwater optical transmitting node, the autonomous underwater vehicle equipped with a smart reflector, and the optical receiving node are acquired, and their spatial relationship is established in a unified three-dimensional coordinate system. The optical signal emitted by the underwater optical transmitting node first propagates to the smart reflector, and then is reflected by the smart reflector to the optical receiving node, forming a two-stage cascaded optical link. The actual position of the underwater optical transmitting node is determined by a preset position and the positional offset caused by water disturbance.

[0046] Secondly, the positional offset of the underwater optical emission node is calculated based on water velocity disturbance, water density, drag coefficient, effective projected area, and equivalent mass. The two propagation distances and cascaded propagation distances are then calculated by combining the real-time positions of the three types of nodes. Simultaneously, water absorption and scattering losses, turbulent scintillation, intelligent reflector efficiency, and link occlusion status are incorporated into the link model to obtain state parameters such as signal-to-noise ratio, occlusion probability, and reflection geometric error.

[0047] Next, a motion state including the autonomous underwater vehicle's position, velocity, and residual disturbances is constructed, and combined with the control inputs and cascaded link states to form Gaussian process input features. The Gaussian process outputs the predicted mean and predicted covariance of the unknown motion disturbances, where the predicted mean is used to compensate for the nominal motion model, and the predicted covariance is used to characterize the prediction uncertainty. The temporal optimization comprehensively considers cascade propagation distance, link quality, occlusion risk, control consumption, and trajectory smoothness, and solves for the optimal control sequence under the constraints of motion area, velocity, acceleration, and link quality. Only the first optimal control input is executed in each control cycle, and updates are performed continuously based on the latest state.

[0048] Subsequently, the geometric reference normal vector of the intelligent reflector is calculated based on the incident and exit directions, and this normal vector, along with the predicted geometric normal vector, the link state, and the normal vector from the previous time step, is input into the depth-deterministic policy gradient controller. The controller outputs a bounded residual correction, which, after amplitude limiting, deflection angle limiting, and normalization, yields the final normal vector, thereby compensating for disturbances and prediction errors while satisfying the optical reflection laws.

[0049] Finally, the optimal control input is sent to the autonomous underwater vehicle propulsion control module, which converts the final normal vector into angle control commands for the intelligent reflector. After completing the current control cycle, the node status and cascade link status are reacquired, and steps S1 to S7 are repeated to form a closed-loop control process of "position acquisition - disturbance modeling - trajectory optimization - reflection correction - control execution - status feedback".

[0050] The core principle of this method is to coordinate the spatial deployment optimization of the autonomous underwater vehicle (AUV) with the reflection direction control of the intelligent reflector. First, a three-dimensional cascaded optical link model is established based on the positions of the underwater optical transmitting node, the intelligent reflector, and the optical receiving node. The positional offset of the transmitting node and the link status are calculated in conjunction with water disturbances. Then, a Gaussian process is used to predict unknown motion disturbances experienced by the AUV, and the optimal trajectory satisfying motion and link constraints is solved through time-domain optimization. Based on this, the geometric reference normal vector of the intelligent reflector is calculated according to the incident and exit directions. A depth-deterministic strategy gradient algorithm is then used to output a bounded residual correction to compensate for disturbances and prediction errors. Finally, the trajectory and reflection angle are updated cyclically through state feedback, achieving dynamic coordinated control of the position and orientation of the underwater intelligent reflector.

[0051] In summary, this invention constructs a complete closed-loop collaborative control scheme for AUV-equipped RIS underwater optical communication in time-varying turbulent waters. It accurately characterizes the spatial relationships of multiple underwater nodes through three-dimensional geometric modeling, quantifies the positional shift of the transmitting node caused by ocean currents and turbulence using a Gaussian perturbation model, establishes two-segment cascaded optical links, and calculates multi-dimensional link quality indicators by comprehensively considering water absorption and scattering, turbulent scintillation, optical path obstruction, and RIS reflection loss. It introduces Gaussian processes for offline learning of various unknown underwater motion perturbations, outputting the mean and covariance matrix of the perturbation online. Based on this, a robust moving time-domain optimization algorithm with adaptive tightening according to perturbation uncertainty is designed to solve for the optimal dynamic deployment trajectory of the AUV that balances communication performance, navigation energy consumption, and trajectory smoothness. Simultaneously, based on the geometric laws of optical reflection, the theoretical reference normal vector of the RIS is solved, and bounded local residual corrections are output using DDPG deep reinforcement learning to perform environmentally adaptive fine-tuning of the reference reflection direction. Finally, AUV propulsion commands and RIS angle control signals are simultaneously issued, achieving full-process iterative closed-loop control through continuous acquisition and updating of the underwater environment and communication status.

[0052] Compared to existing technologies that use fixed RIS, decoupled AUV and reflection direction control, and fixed safety margins, this invention, on the one hand, relies on Gaussian processes to quantitatively characterize disturbance uncertainties and dynamically balance communication transmission performance with link stability and safety margins; on the other hand, it employs a two-layer reflection control architecture of "geometric reference + reinforcement learning residual correction," significantly reducing the online computational overhead of reinforcement learning, improving real-time adaptability, and achieving global optimal coordination between AUV deployment location and RIS reflection direction. All physical parameters of the entire method can be collected in real time using conventional underwater positioning, flow velocity, and optical sensing equipment. Optimized model weights and hardware constraint parameters can be easily tuned offline according to different water depths and flow velocity conditions. It can be directly integrated into the AUV embedded controller for real-time computation, demonstrating high practical value and promising prospects in engineering scenarios such as underwater sensor networking, seabed exploration, and high-speed wireless data transmission for underwater unmanned platforms.

[0053] Example 2 Please see Figure 5 This embodiment provides a schematic diagram of the structure of an underwater AUV equipped with a RIS optical communication collaborative deployment and reflection control system.

[0054] As an example, the system is implemented using the underwater AUV equipped with RIS optical communication collaborative deployment and reflection control method described in Example 1. The system includes: The geometric modeling module 510 is suitable for receiving the preset position of the underwater optical transmitting node, the measured position of the autonomous underwater vehicle equipped with the intelligent reflector, and the fixed position of the optical receiving node as input to establish a three-dimensional geometric model of underwater intelligent reflector-assisted optical communication. The underwater optical transmitting node actual position calculation module 520 is suitable for calculating the position offset of the underwater optical transmitting node based on water disturbance information, and calculating the actual position of the underwater optical transmitting node by combining the position offset with the preset position of the underwater optical transmitting node. The cascaded optical link modeling module 530 is suitable for establishing a cascaded optical link model between the underwater optical transmitter node, the intelligent reflector and the optical receiver node based on the measured position of the autonomous underwater vehicle, the fixed position of the optical receiver node and the actual position of the underwater optical transmitter node, and outputting the total cascaded propagation distance, the cascaded optical link signal-to-noise ratio and the cascaded occlusion probability. The Gaussian process feature construction and disturbance prediction module 540 is suitable for constructing the motion state matrix of the autonomous underwater vehicle based on the measured position of the autonomous underwater vehicle, and constructing Gaussian process input features based on the motion state matrix, cascaded total propagation distance, cascaded optical link signal-to-noise ratio and cascaded occlusion probability. The robust rolling time-domain optimization module 550 is suitable for predicting unknown motion disturbances caused by water and propulsion mechanism through Gaussian process based on the input characteristics of the Gaussian process. It introduces environmental adaptive weights and communication security constraints to build a mobile time-domain multi-objective optimization model, rolls to solve the optimal motion sequence of AUV global dynamic deployment that adapts to the real-time link status, and outputs the optimal propulsion control input of the autonomous underwater vehicle at the current moment. The RIS reflection normal vector solving module 560 is suitable for calculating the geometric reference normal vector of the intelligent reflector based on the actual position of the underwater optical transmitting node, the measured position of the autonomous underwater vehicle equipped with the intelligent reflector, and the fixed position of the optical receiving node. It uses a depth deterministic strategy gradient algorithm to solve the bounded residual correction amount corresponding to the geometric reference normal vector, and outputs the final normal vector of the intelligent reflector after superimposing and normalizing the geometric reference normal vector and the bounded residual correction amount. The execution drive module 570 is suitable for using the optimal propulsion control input of the autonomous underwater vehicle output by the robust rolling time domain optimization module 550 and the final normal vector of the intelligent reflector output by the RIS reflection normal vector solving module 560 as inputs, and sending them to the autonomous underwater vehicle propulsion device and the intelligent reflector angle adjustment mechanism to perform dynamic deployment adjustment and synchronous control of reflection angle of the AUV. The data acquisition closed-loop feedback module 580 is suitable for acquiring the motion state of the autonomous underwater vehicle, the position of the underwater optical transmitting node, the signal-to-noise ratio of the cascaded optical link, the intensity of water disturbance, and the link blockage state after the control is executed. It re-executes the geometric modeling module 510 to the execution drive module 570 to achieve continuous and coordinated adjustment of the autonomous underwater vehicle's position and the reflection direction of the intelligent reflector.

[0055] It is not difficult to see that this embodiment is a system implementation corresponding to the first embodiment, and this embodiment can be implemented in conjunction with the first embodiment. The relevant technical details mentioned in the first embodiment are still valid in this embodiment, and will not be repeated here to reduce repetition. Accordingly, the relevant technical details mentioned in this embodiment can also be applied to the first embodiment.

[0056] It is worth mentioning that all modules involved in this embodiment are logical units. In practical applications, a logical unit can be a physical unit, a part of a physical unit, or a combination of multiple physical units. Furthermore, to highlight the innovative aspects of this invention, this embodiment does not introduce units that are not closely related to solving the technical problem proposed by this invention; however, this does not mean that other units are absent from this embodiment.

[0057] Example 3 Please see Figure 6 The present invention also provides an electronic device, including: a memory and a processor; the memory stores at least one program instruction; the processor loads and executes the at least one program instruction to implement the underwater AUV equipped with RIS optical communication collaborative deployment and reflection control method provided in Embodiment 1.

[0058] The memory 702 and processor 701 are connected via a bus, which may include any number of interconnecting buses and bridges, connecting various circuits of one or more processors 701 and memory 702 together. The bus may also connect various other circuits, such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. A bus interface provides an interface between the bus and the transceiver. The transceiver may be a single element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by processor 701 is transmitted over a wireless medium via an antenna, which further receives data and transmits it to processor 701.

[0059] Processor 701 is responsible for managing the bus and general processing, and can also provide various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. Memory 702 can be used to store data used by processor 701 during operation.

[0060] Example 4 This invention also proposes a storage medium storing a method for coordinated deployment and reflection control of RIS optical communication on an underwater AUV. When executed by a processor, the method implements the steps described above. Since this storage medium employs all the technical solutions of the above embodiments, it possesses at least all the beneficial effects brought about by the technical solutions of the above embodiments, which will not be elaborated further here.

[0061] The above descriptions are merely embodiments of the present invention. Commonly known structures and characteristics are not described in detail here. Those skilled in the art are aware of all common technical knowledge in the field prior to the application date or priority date, are aware of all existing technologies in that field, and have the ability to apply conventional experimental methods prior to that date. Those skilled in the art can, based on the guidance provided in this application, improve and implement this solution in combination with their own capabilities. Some typical known structures or methods should not be obstacles for those skilled in the art to implement this application. It should be noted that those skilled in the art can make several modifications and improvements without departing from the structure of the present invention. These should also be considered within the scope of protection of the present invention, and will not affect the effectiveness of the implementation of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A method for coordinated deployment and reflection control of RIS optical communication on an underwater AUV, characterized in that, The method includes: Step S1: Receive the preset position of the underwater optical transmitting node, the measured position of the autonomous underwater vehicle equipped with the intelligent reflector, and the fixed position of the optical receiving node as inputs to establish a three-dimensional geometric model of underwater intelligent reflector-assisted optical communication. Step S2: Calculate the position offset of the underwater optical transmitting node based on the water disturbance information, and calculate the actual position of the underwater optical transmitting node by combining the position offset with the preset position of the underwater optical transmitting node. Step S3: Based on the measured position of the autonomous underwater vehicle, the fixed position of the optical receiving node, and the actual position of the underwater optical transmitting node, establish a cascaded optical link model between the underwater optical transmitting node, the intelligent reflective surface, and the optical receiving node, and output the total cascaded propagation distance, the cascaded optical link signal-to-noise ratio, and the cascaded occlusion probability. Step S4: Construct the motion state matrix of the autonomous underwater vehicle based on the measured position of the autonomous underwater vehicle, and construct Gaussian process input features based on the motion state matrix, cascaded total propagation distance, cascaded optical link signal-to-noise ratio, and cascaded occlusion probability. Step S5: Based on the input characteristics of the Gaussian process, predict the unknown motion disturbances brought by the water body and propulsion mechanism through the Gaussian process, introduce environmental adaptive weights and communication security constraints to build a mobile time domain multi-objective optimization model, and solve the optimal motion sequence of AUV global dynamic deployment that adapts to the real-time link status in a rolling manner, and output the optimal propulsion control input of the autonomous underwater vehicle at the current moment. Step S6: Calculate the geometric reference normal vector of the intelligent reflector based on the actual position of the underwater optical transmitting node, the measured position of the autonomous underwater vehicle equipped with the intelligent reflector, and the fixed position of the optical receiving node. Use the depth deterministic strategy gradient algorithm to solve the bounded residual correction amount corresponding to the geometric reference normal vector. After superimposing and normalizing the geometric reference normal vector and the bounded residual correction amount, output the final normal vector of the intelligent reflector. Step S7: Using the optimal propulsion control input of the autonomous underwater vehicle output in step S5 and the final normal vector of the intelligent reflector output in step S6 as inputs, the data is sent to the propulsion device of the autonomous underwater vehicle and the angle adjustment mechanism of the intelligent reflector to perform dynamic deployment adjustment and synchronous control of the reflection angle of the AUV. Step S8: After the control is executed, obtain the motion state of the autonomous underwater vehicle, the position of the underwater optical transmitting node, the signal-to-noise ratio of the cascaded optical link, the intensity of water disturbance and the link blockage status, and re-execute steps S1 to S7 to achieve continuous and coordinated adjustment of the position of the autonomous underwater vehicle and the reflection direction of the intelligent reflector.

2. The underwater AUV equipped with RIS optical communication for coordinated deployment and reflection control according to claim 1, characterized in that, Step S1, which involves receiving the preset position of the underwater optical transmitting node, the measured position of the autonomous underwater vehicle equipped with the intelligent reflector, and the fixed position of the optical receiving node as inputs, and establishing a three-dimensional geometric model of underwater intelligent reflector-assisted optical communication, includes: The actual position of the underwater optical transmitting node is represented as the sum of the preset position and the position offset caused by water disturbance: ； in, This represents the actual position of the underwater optical emission node at time t. This represents the preset position of the underwater optical emission node at time t. The positional shift caused by water disturbance at time t; The location of the autonomous underwater vehicle equipped with a smart reflector is represented as: ； in, These are the position coordinates of the autonomous underwater vehicle along the three coordinate axes at time t; The location of the optical receiving node is represented as follows: ； in, These are the position coordinates of the optical receiving node along the three coordinate axes.

3. The underwater AUV equipped with RIS optical communication for coordinated deployment and reflection control according to claim 2, characterized in that, Step S2, calculating the position offset of the underwater optical transmitting node based on the water disturbance information, and combining the position offset with the preset position of the underwater optical transmitting node to calculate the actual position of the underwater optical transmitting node, includes: The water velocity perturbation acting on the underwater optical emission node is represented as a zero-mean random vector: ； in, Let be the three-dimensional water velocity perturbation vector acting on the underwater optical emission node at time t. Indicates a Gaussian distribution. Let be the covariance matrix of the water velocity perturbation at time t; The covariance matrix is ​​expressed as: ； in, The intensity of water disturbance. This is the normalized directional covariance matrix; The perturbation force acting on the underwater optical emission node is calculated based on the water velocity disturbance: ； in, For water density, The drag coefficient, The effective projected area of ​​the underwater optical emission node and its onboard equipment; Converting water disturbance forces into position offsets for underwater optical emission nodes: ； in, is the conversion coefficient from disturbance force to position offset, and M is the equivalent mass of the underwater optical transmitting node and its onboard equipment.

4. The underwater AUV equipped with RIS optical communication for coordinated deployment and reflection control according to claim 2, characterized in that, Step S3 involves establishing a cascaded optical link model between the underwater optical transmitter, the smart reflector, and the optical receiver based on the measured position of the autonomous underwater vehicle, the fixed position of the optical receiver node, and the actual position of the underwater optical transmitter node. The outputs include the total cascaded propagation distance, the cascaded optical link signal-to-noise ratio, and the cascaded obstruction probability. Calculate the first propagation distance from the underwater optical emitting node to the smart reflector. : ； Calculate the second propagation distance from the smart reflector to the light receiving node. : ； Cascade propagation distance of underwater intelligent reflector-assisted optical communication links for: ； According to the cascade propagation distance By combining water absorption, scattering, turbulent scintillation, obstacle occlusion, and the reflection efficiency of the smart reflector, the signal-to-noise ratio of the smart reflector-assisted cascaded optical link is calculated. and cascading occlusion probability .

5. The underwater AUV equipped with RIS optical communication for coordinated deployment and reflection control according to claim 4, characterized in that, Step S4 involves constructing the motion state matrix of the autonomous underwater vehicle (AUV) based on its measured position, and then constructing Gaussian process input features based on the motion state matrix, cascaded total propagation distance, cascaded optical link signal-to-noise ratio, and cascaded occlusion probability. This includes: Constructing the augmented motion state of an autonomous underwater vehicle: ； in, For autonomous underwater vehicles at all times speed, The residual state is caused by local water disturbance, propulsion error, and unmodeled dynamics. Gaussian process input features are constructed based on the autonomous underwater vehicle's motion state, control input, cascade propagation distance, link signal-to-noise ratio, turbulent scintillation intensity, cascade blockage probability, and reflection geometric error. : ； in, To construct a function by inputting features into a Gaussian process, For the control input of autonomous underwater vehicles, The turbulent scintillation intensity of the cascaded optical link. For reflection geometric error; The input features are then input into the Gaussian process perturbation prediction model to obtain the perturbation prediction mean and perturbation prediction covariance: ; in, The training dataset is for Gaussian processes. The mean of the disturbance prediction, To predict the covariance of the perturbation, This is a pre-trained Gaussian process perturbation prediction model.

6. The underwater AUV equipped with RIS optical communication for coordinated deployment and reflection control according to claim 5, characterized in that, Step S5 involves predicting unknown motion disturbances caused by the water body and propulsion mechanism based on the Gaussian process input characteristics, introducing environmental adaptive weights and communication security constraints to build a mobile time-domain multi-objective optimization model, and continuously solving for the optimal motion sequence for AUV global dynamic deployment that adapts to the real-time link status. The optimal propulsion control input for the autonomous underwater vehicle at the current moment includes: A state model of the autonomous underwater vehicle in the prediction time domain is established based on the mean of the disturbance prediction: ； in, This is the nominal motion model of an autonomous underwater vehicle that does not consider residual disturbances from the Gaussian process. The perturbation input mapping matrix; An environment-adaptive weighting is constructed based on turbulent scintillation intensity, cascade occlusion probability, distance feedback error, and Gaussian process prediction uncertainty. ； in, , , and These are non-negative weighting coefficients. To provide feedback on the error between the reconstructed distance and the geometrically predicted distance; A multi-objective optimization function in the time domain is established, with the optimization objectives being cascade propagation distance, link feedback error, insufficient signal-to-noise ratio, cascade blockage risk, autonomous underwater vehicle control input, and trajectory smoothness: ； in, To predict the length of the time domain, The error between the measured link feedback and the predicted link feedback, For the minimum permissible signal-to-noise ratio, To maximize the probability of cascading occlusion, , , , , and These are the weighting coefficients for the corresponding target items; The multi-objective optimization function satisfies communication security constraints: ； ； ； ； ； in, The permitted movement area for autonomous underwater vehicles. For the set of allowed states, and This is a tightened set of uncertainties constructed based on the covariance predicted by the Gaussian process. For set difference operation, For the maximum permissible control input, and This refers to the safety margin factor. After solving the finite prediction time-domain optimization problem at each control time step, only the first control input in the current optimization sequence is executed: ； Based on the updated motion state of the autonomous underwater vehicle and the rolling update prediction time domain of the cascaded optical link feedback, the optimal motion trajectory of the autonomous underwater vehicle equipped with the intelligent reflector is obtained: ； in, for The optimal propulsion control input for the AUV is finally issued and executed at any given time. This is the optimal control sequence obtained by constrained optimization. This is the optimal deployment trajectory set for AUVs across the entire domain. The optimal three-dimensional position of the AUV is obtained after optimization at time t.

7. The underwater AUV equipped with RIS optical communication for coordinated deployment and reflection control according to claim 6, characterized in that, Step S6 involves calculating the geometric reference normal vector of the intelligent reflector based on the actual position of the underwater optical emitting node, the measured position of the autonomous underwater vehicle equipped with the intelligent reflector, and the fixed position of the optical receiving node. A depth-deterministic gradient algorithm is then used to solve for the bounded residual correction amount corresponding to the geometric reference normal vector. The final normal vector of the intelligent reflector is then output after superimposing and normalizing the geometric reference normal vector and the bounded residual correction amount. Based on the spatial relationship between the underwater light emitting node, the smart reflector, and the light receiving node, a geometric normal vector mapping is established: ； The geometric reference normal vector at the current moment is: ； The predicted geometric normal vector for the next time step is: ； The current geometric reference normal vector, the predicted geometric normal vector for the next time step, the cascaded optical link state, and the smart reflector normal vector executed in the previous time step are used to construct the state vector of the deep deterministic policy gradient algorithm: ； The state characteristics of the cascaded optical link are as follows: ； The deep deterministic policy gradient algorithm outputs the residual correction relative to the geometric reference normal vector: ； in, The control actions output by the deep deterministic policy gradient algorithm. To form a collection of actionable actions This is the residual correction amount relative to the geometric reference normal vector; The set of possible actions for the residual correction is as follows: ； in, The maximum permissible residual correction magnitude, This represents the maximum permissible angular deviation of the normal vector after residual correction relative to the geometric reference normal vector. The geometric reference normal vector is added to the residual correction and projected onto the unit sphere to obtain the final normal vector of the intelligent reflective surface. : ； The unit spherical projection function is: 。 8. The underwater AUV equipped with RIS optical communication for coordinated deployment and reflection control according to claim 7, characterized in that, The instantaneous reward function of the deep deterministic policy gradient algorithm is: ； in, , , , and The non-negative weighting coefficients represent the reflection geometric error as follows: ； The instant reward function is used to improve the signal-to-noise ratio of the intelligent reflector-assisted cascaded optical link and suppress link occlusion, turbulent scintillation, reflection geometry mismatch, and abrupt changes in reflection direction between adjacent time moments.

9. The underwater AUV equipped with RIS optical communication for coordinated deployment and reflection control according to claim 1, characterized in that, Step S7, using the optimal propulsion control input of the autonomous underwater vehicle (AUV) output in step S5 and the final normal vector of the intelligent reflector output in step S6 as inputs, is respectively sent to the AUV propulsion device and the intelligent reflector angle adjustment mechanism to perform dynamic deployment adjustment and synchronous control of the reflection angle of the AUV, including: The current optimal control input obtained in step S5 is sent to the propulsion control device of the autonomous underwater vehicle. The final normal vector obtained in step S6 is converted into azimuth and pitch angle control commands for the intelligent reflector and sent to the reflection angle adjustment device of the intelligent reflector.

10. A system for coordinated deployment and reflection control of RIS optical communication on an underwater AUV, wherein the system is implemented using the method for coordinated deployment and reflection control of RIS optical communication on an underwater AUV as described in any one of claims 1-9, characterized in that, The system includes: The geometric modeling module is suitable for receiving the preset position of the underwater optical transmitting node, the measured position of the autonomous underwater vehicle equipped with the intelligent reflector, and the fixed position of the optical receiving node as inputs to establish a three-dimensional geometric model of underwater intelligent reflector-assisted optical communication. The underwater optical transmitting node actual position calculation module is suitable for calculating the position offset of the underwater optical transmitting node based on water disturbance information, and calculating the actual position of the underwater optical transmitting node by combining the position offset with the preset position of the underwater optical transmitting node. The cascaded optical link modeling module is suitable for establishing a cascaded optical link model between the underwater optical transmitter node, the smart reflector, and the optical receiver node based on the measured position of the autonomous underwater vehicle, the fixed position of the optical receiver node, and the actual position of the underwater optical transmitter node. It outputs the total cascaded propagation distance, the cascaded optical link signal-to-noise ratio, and the cascaded occlusion probability. The Gaussian process feature construction and disturbance prediction module is suitable for constructing the motion state matrix of the autonomous underwater vehicle based on the measured position of the autonomous underwater vehicle, and constructing Gaussian process input features based on the motion state matrix, cascaded total propagation distance, cascaded optical link signal-to-noise ratio and cascaded occlusion probability. The robust rolling time-domain optimization module is suitable for predicting unknown motion disturbances caused by water bodies and propulsion mechanisms based on the input characteristics of the Gaussian process through the Gaussian process. It introduces environmental adaptive weights and communication security constraints to build a mobile time-domain multi-objective optimization model, and rolls to solve for the optimal motion sequence of AUV global dynamic deployment that adapts to the real-time link status. It outputs the optimal propulsion control input of the autonomous underwater vehicle at the current moment. The RIS reflection normal vector solving module is suitable for calculating the geometric reference normal vector of the intelligent reflector based on the actual position of the underwater optical emitting node, the measured position of the autonomous underwater vehicle equipped with the intelligent reflector, and the fixed position of the optical receiving node. It uses a depth deterministic strategy gradient algorithm to solve the bounded residual correction amount corresponding to the geometric reference normal vector. After superimposing and normalizing the geometric reference normal vector and the bounded residual correction amount, the final normal vector of the intelligent reflector is output. The execution drive module is suitable for using the optimal propulsion control input of the autonomous underwater vehicle output by the robust rolling time domain optimization module and the final normal vector of the intelligent reflector output by the RIS reflection normal vector solving module as inputs, and sending them to the autonomous underwater vehicle propulsion device and the intelligent reflector angle adjustment mechanism to perform dynamic deployment adjustment and synchronous control of reflection angle of the AUV. The data acquisition closed-loop feedback module is suitable for acquiring the motion state of the autonomous underwater vehicle, the position of the underwater optical emission node, the signal-to-noise ratio of the cascaded optical link, the intensity of water disturbance, and the link blockage state after the control is executed. It re-executes the geometric modeling module to the execution drive module to achieve continuous and coordinated adjustment of the autonomous underwater vehicle's position and the reflection direction of the intelligent reflector.

Images