A marine stereo-synchronous observation system and method fixed with motor coordination

Abstract

The present application relates to a kind of fixed and motorized cooperative marine three-dimensional synchronous observation system and method, system includes including: fixed node;Motorized node;By fixed node and motorized node cooperate to cover blank area and observation detection blind area between node;Shipborne unmanned aerial vehicle, for carrying fixed node and motorized node and drop to preset position;Surface warship, surface warship is configured as when advancing by carrying sound source continuously emit underwater acoustic signal, and collect fixed node information and motorized node information, combine observation task to the navigation track of motorized node and the parameter of fixed node are regulated, to maintain the set coverage accuracy to target marine water body at any time.The present application constructs three-dimensional synchronous observation network by motorized node and fixed node cooperation, compared with traditional marine environment investigation method and marine observation network is more efficient, flexible, can obtain larger range, higher resolution and more complete space-time matching environmental data.

Description

Technical Field

[0001] This invention relates to the field of marine exploration technology, and in particular to a fixed and mobile coordinated three-dimensional synchronous marine observation system and method. Background Technology

[0002] Existing methods for collecting marine acoustic data or observation networks primarily rely on fixed buoy nodes, single-ship, single-buoy surveys, or underwater platforms to acquire information. In complex marine environments, fixed buoy node platforms suffer from blind spots in their three-dimensional observation networks, only acquiring underwater acoustic signals and environmental information around their deployment location. Their sensing range is limited, deployment and maintenance are difficult, and interconnection to form a network is challenging. Conventional single-platform observation methods or large-scale integrated platform observation or environmental survey methods suffer from limitations such as acquiring data only along fixed routes, insufficient sampling of acoustic field information, and significant transmission loss. Furthermore, a common problem with single-platform or statically deployed single-platform systems is the weak robustness of the entire sensor network, insufficient stability and accuracy of acoustic data, small observation coverage, and the existence of blind spots.

[0003] Meanwhile, due to the complexity of the marine environment and underwater acoustic propagation, environmental and underwater acoustic data in specific sea areas exhibit spatiotemporal variations. Conventional underway marine acoustic surveys, utilizing buoys, moorings, and vessels equipped with vertical scalar and vector hydrophone arrays, as well as hydrological sensors for temperature, salinity, and current velocity, and sound sources, can only acquire information from a limited number of measurement points within a specific time period. Furthermore, these surveys are time-consuming and costly. In terms of efficiently acquiring spatiotemporally matched data over a large area, they are insufficient to meet the needs of marine research and underwater target detection. Summary of the Invention

[0004] (a) Technical problems to be solved

[0005] In view of the above-mentioned shortcomings and deficiencies of the prior art, the present invention provides a fixed and mobile coordinated three-dimensional synchronous marine observation system and method, which solves the technical problems of existing single observation platforms having small observation coverage, easy blind spots, insufficient stability and accuracy of the detected observation data, and difficulty in matching the detected data with spatiotemporal information.

[0006] (II) Technical Solution

[0007] To achieve the above objectives, the main technical solutions adopted by the present invention include:

[0008] In a first aspect, embodiments of the present invention provide a fixed and mobile coordinated ocean three-dimensional synchronous observation system, comprising:

[0009] Fixed nodes are configured at fixed locations within the target ocean water body to form a rectangular network, used to collect fixed node information including underwater acoustic signals at fixed locations, environmental information, node positions, and timing information.

[0010] Mobile nodes are deployed between the surface and seabed of the target ocean body, constantly diving and surfacing and moving along a zigzag route. They are used to collect mobile node information along the route, including underwater acoustic signals, environmental information, current node position, and timing information. Fixed nodes and mobile nodes work together to cover the blank areas and blind spots between nodes.

[0011] Shipborne drones, using surface ships as bases, are used to carry fixed and mobile nodes and deploy them to preset locations;

[0012] Surface ships, carrying sound sources, are configured to continuously emit underwater acoustic signals as they advance along a zigzag route along a fixed node deployment path. They also collect information from fixed and mobile nodes via radio communication. Combined with observation missions, the navigation trajectories of mobile nodes and parameters of fixed nodes are adjusted to ensure that the ocean three-dimensional synchronous observation network formed by mobile and fixed nodes maintains the set coverage accuracy of the target ocean body at all times.

[0013] Optionally, the fixed nodes include acoustic buoys, and the mobile nodes include underwater gliders; wherein both the acoustic buoys and the underwater gliders are lightweight devices with cost and power consumption below a threshold.

[0014] Optionally, any fixed node can serve as a communication relay point between mobile nodes and other fixed nodes.

[0015] Optionally, cross-validation of the collected data can be performed between neighboring fixed nodes and mobile nodes, and only the data that passes the validation can be retained.

[0016] Optionally, it also includes: a shore station for issuing observation tasks to surface ships, including marine environmental noise, marine sound propagation, marine reverberation and synchronous marine environment, and for receiving and analyzing data transmitted back from surface ships via communication satellites and outputting corresponding control commands.

[0017] Secondly, embodiments of the present invention provide a method for coordinated fixed and mobile ocean three-dimensional synchronous observation, used to construct the system described above, including:

[0018] Based on the acquired observation mission, surface vessels, carrying drones, fixed nodes, and mobile nodes, arrive at the target ocean water body designated by the observation mission;

[0019] Surface ships control shipborne UAVs to deploy several fixed nodes to designated locations based on fixed node deployment instructions issued by shore stations, forming a rectangular network.

[0020] Based on the mobile node deployment instructions issued by the shore station, the surface ships control the shipborne UAVs to deploy several mobile nodes to the designated locations, so that the mobile nodes can move along a zigzag route according to the navigation trajectory contained in the mobile node deployment instructions, thereby enabling the fixed nodes and mobile nodes to cooperate to cover the blank areas and blind spots between the nodes.

[0021] When surface ships advance along a zigzag route along the deployment path of fixed nodes, they continuously emit underwater acoustic signals through preset sound sources, collect information on fixed nodes and mobile nodes through radio communication, and adjust the navigation trajectory of mobile nodes and the parameters of fixed nodes in conjunction with the observation mission, so that the ocean three-dimensional synchronous observation network formed by mobile nodes and fixed nodes can maintain the set coverage accuracy of the target ocean water body at all times.

[0022] Optionally, before the surface ship controls its shipborne UAV to deploy several fixed nodes to designated locations to form a rectangular network based on the fixed node deployment command issued by the shore station, the process also includes:

[0023] Based on the general environmental information of the target ocean water body, the starting point of the surface vessel, the basic parameters of the UAV, the basic parameters of the fixed node, and the basic parameters of the mobile node, the surface vessel receiving shore station uses an environment-aided planning algorithm with the objective function of covering the three-dimensional space of the surveyed sea area. The output includes the fixed node deployment command containing the deployment position of the fixed node and the mobile node deployment command containing the deployment position and navigation trajectory of the mobile node.

[0024] Optionally, based on the acquired general environmental information of the target ocean water body, the starting point of surface vessels, the basic parameters of the UAV, the basic parameters of fixed nodes, and the basic parameters of mobile nodes, the shore station uses an environment-aided planning algorithm, with the objective function of covering the three-dimensional space of the surveyed sea area, to output fixed node deployment instructions containing the deployment positions of fixed nodes and mobile node deployment instructions containing the deployment positions and navigation trajectories of mobile nodes, including:

[0025] A11. The general environmental information of the target ocean water body, the starting point of surface ships, the basic parameters of UAVs, the basic parameters of fixed nodes, and the basic parameters of maneuvering nodes are used as inputs to a pre-constructed multi-objective optimization model based on the genetic particle swarm optimization algorithm and the deep deterministic policy gradient algorithm.

[0026] A12. Initialize the population by randomly generating a set of particles, each representing a solution to a set of decision variables, and initialize the parameters of the genetic particle swarm optimization algorithm containing crossover probability, mutation probability, and inertia weight, as well as the parameters of the deep deterministic policy gradient containing the Actor and Critic networks, the learning rate and discount factor, and the memory cache used to store information containing state, action and reward.

[0027] A13. Based on the multiple objective functions corresponding to each particle, evaluate the fitness value of each particle on the multiple objective functions;

[0028] A14. Encode the parameters of the current genetic particle swarm optimization algorithm and the fitness values ​​of the particles on multiple objective functions into vectors as input to the deep deterministic policy gradient algorithm;

[0029] A15. Based on the fitness values ​​of the particles on multiple objective functions, perform non-dominated sorting of the particles and divide the particles into different frontiers;

[0030] A16. Calculate the crowding degree of particles in each front to determine the particle distribution density in the target space;

[0031] A17. The population is updated by genetic operations such as crossover and introducing new variations, based on the frontier obtained from non-dominated sorting and density calculated by crowding.

[0032] A18. To achieve local optimization and fast convergence, the population is updated by performing N particle swarm optimization iterations on the results of genetic operations.

[0033] A19. After each update iteration of the genetic particle swarm optimization algorithm, the action a predicted for the current iteration number is executed in the particle swarm. t And observe the reward r t and new state s t+1 , will the empirical tuple (s t ,a t ,r t ,s t+1 The parameters of the Actor and Critic networks are stored in the memory cache and then trained and updated using the deep deterministic policy gradient algorithm based on the actions, states, and rewards in the memory cache.

[0034] A20. Decode the parameter vector learned by the deep deterministic strategy gradient algorithm into the parameter settings of the genetic particle swarm algorithm, and combine the latest collected marine environmental information to optimize the navigation trajectory of the maneuvering node and the position of the fixed node in real time.

[0035] A21. Repeat steps A13 to A20 until the maximum number of iterations is reached or the convergence reaches a certain level. Output the optimal particle and its fitness value, and then obtain the optimization parameters output by the multi-objective optimization model, including the fixed node layout, UAV deployment path, optimal deployment position and optimal flight trajectory of the mobile node, and the real-time coverage of the synchronous observation network. Then generate the fixed node deployment command and the mobile node deployment command.

[0036] Optionally, performing N particle swarm update iterations to update the population depends on the following particle velocity update formula:

[0037] v ij (t+1)=ω t v ij (t)+c 1t r1(x pbestij -x ij (t))+c 2t r2(x gbest -x ij (t)) (1)

[0038] In equation (1), ω t c 1t c 2t Let r1 and r2 be the inertial weights after the t-th iteration and the two parameters controlling the acceleration, respectively, where r1 and r2 are random numbers in the interval [0,1]. pbestij x represents the historical best position of particle i. gbest The historical best position representing the population, v ij and x ij These represent the velocity and position of particle i in its j-th dimension, respectively.

[0039] Updating the parameters of the Actor network depends on the following implementation of the Actor network's loss function:

[0040]

[0041] In equation (2), Represents the parameters θ of the Actor network μ The gradient of the policy function J, N is the number of samples in the buffer, s is the current state, and θ Q For Critic network parameters, For the Critic network, for state s i and by Actor network in s i Deterministic action μ(s) output in the state i The gradient of the estimated Q value with respect to action a. For an Actor network with state s i The output deterministic action with respect to its parameter θ μ The gradient;

[0042] Updating the parameters of the Critic network depends on the following implementation of the Critic network's loss function:

[0043]

[0044] In equation (3), L is the loss function of the Critic network, and Q(s) t,a t / θ Q ) represents the sample in state s t The action taken at that time a t y t The target action value is calculated for the Critic network;

[0045] The parameters of the Critic network and Actor network are updated in the following ways:

[0046] θ Q′ ←τθ Q +(1-τ)θ Q′ ,θ μ′ ←τθ μ +(1-τ)θ μ′ (4)

[0047] In equation (4), θ Q For the Critic network parameters, θ Q′ For the updated Critic network parameters, θ μ Let θ be the Actor network parameter. μ′ Here are the updated Actor network parameters, and τ is the learning rate;

[0048] The actions predicted by the Actor network are implemented based on the following action prediction formula of the Actor network:

[0049] a=μ(s∣θμ) (5)

[0050] In equation (5), μ(s∣θμ) represents the Actor network's prediction of the action.

[0051] Optionally, the surface ship controls its shipborne UAV to deploy several mobile nodes to designated locations based on the mobile node deployment command issued by the shore station. This allows the mobile nodes to advance along a zigzag route following the trajectory included in the deployment command, enabling fixed and mobile nodes to coordinate and cover the gaps and blind spots between nodes. The system also includes:

[0052] A radio communication link is established between the surface ships and the fixed and mobile nodes after deployment for real-time communication.

[0053] Surface ships continuously acquire the location and timing information of each node through the interaction between mobile nodes, fixed nodes, and GPS. Ultimately, they obtain fixed node information and mobile node information that match the spatiotemporal information of the target ocean body by collecting data from both mobile and fixed nodes.

[0054] (III) Beneficial Effects

[0055] The beneficial effects of this invention are:

[0056] This invention enables rapid deployment via unmanned surface vessels or shipborne drones, resulting in low implementation costs, short construction cycles, multi-platform collaboration, and high spatiotemporal acquisition resolution and system robustness. Fixed nodes maintain their spatial location for long-term in-situ acquisition of underwater acoustic signals or environmental information at a fixed location; mobile nodes continuously submerge and surface between the sea surface and seabed, employing a zigzag path to acquire underwater acoustic signals or environmental information along their route. Distributed throughout the ocean observation space, the fixed and mobile nodes acquire unified spatiotemporal information via GPS positioning and timing, forming a distributed, three-dimensional underwater acoustic signal sensing network. This network continuously acquires marine acoustic environmental and underwater acoustic data, enabling large-scale, deep, and long-term synchronous and effective three-dimensional observation of the target area and ocean body.

[0057] Therefore, this invention addresses the shortcomings of existing technologies by coordinating fixed and mobile nodes to form a distributed three-dimensional underwater acoustic signal sensing network, creating a more comprehensive, stable, reliable, and high-performance ocean acoustic three-dimensional synchronous observation network. Compared to traditional ocean observation methods, this invention's synchronous observation network offers a wider observation range, higher observation accuracy, higher spatiotemporal resolution, and lower observation costs. Attached Figure Description

[0058] Figure 1 A schematic diagram illustrating the composition of a fixed and mobile coordinated ocean three-dimensional synchronous observation system provided by the present invention;

[0059] Figure 2 A schematic diagram illustrating the operation of a fixed and mobile coordinated ocean three-dimensional synchronous observation system provided by the present invention;

[0060] Figure 3 A schematic diagram illustrating the optimization process of a multi-objective optimization model based on the genetic particle swarm optimization algorithm and the depth deterministic strategy gradient algorithm for a fixed and maneuver-coordinated ocean three-dimensional synchronous observation method provided by the present invention.

[0061] Figure 4 This is a schematic diagram of the overall process of a fixed and mobile coordinated ocean three-dimensional synchronous observation method provided by the present invention.

[0062] [Explanation of Labels in the Attached Image]

[0063] 10: Fixed nodes;

[0064] 20: Maneuvering Node;

[0065] 30: Shipborne unmanned aerial vehicle; 31: Sound source:

[0066] 40: Surface ships;

[0067] 50: Communication satellite;

[0068] 60: Shore Station. Detailed Implementation

[0069] To better explain and facilitate understanding of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0070] like Figure 1 As shown in the embodiment of the present invention, a fixed and mobile coordinated ocean three-dimensional synchronous observation system includes: fixed nodes 10, configured at fixed positions within the target ocean water body to form a rectangular network, used to collect fixed node information including underwater acoustic signals, environmental information, node positions, and timing information; and mobile nodes 20, configured between the sea surface and seabed within the target ocean water body, continuously diving and surfacing and advancing along a zigzag route, used to collect mobile node information including underwater acoustic signals, environmental information, current node positions, and timing information along the route; the fixed nodes 10 and mobile nodes 20 cooperate to cover the blank areas between nodes and the observation area. Blind spots are detected; a shipborne UAV 30, based on a surface ship 40, is used to carry a fixed node 10 and a mobile node 20 and deploy them to a preset location; the surface ship 40 carries a sound source 31, which is configured to continuously emit underwater acoustic signals through the sound source 31 when moving along the deployment path of the fixed node 10 in a zigzag route, and to collect information on the fixed node and the mobile node through radio communication. Combined with the observation mission, the navigation trajectory of the mobile node 20 and the parameters of the fixed node 10 are adjusted so that the three-dimensional synchronous ocean observation network formed by the mobile node 20 and the fixed node 10 can maintain the set coverage accuracy of the target ocean water body at all times.

[0071] This invention utilizes unmanned surface vessels or shipborne drones (30) for rapid deployment, resulting in low implementation costs, short construction periods, multi-platform collaboration, and high spatiotemporal acquisition resolution and system robustness. Fixed nodes (10) remain spatially stationary, acquiring underwater acoustic signals or environmental information from a fixed location over extended periods. Mobile nodes (20) continuously submerge and surface between the sea surface and seabed, employing a zigzag path to acquire underwater acoustic signals or environmental information along their route. The fixed and mobile nodes (20) are distributed throughout the ocean observation space, acquiring unified spatiotemporal information via GPS positioning and timing. Working together, they form a distributed, three-dimensional underwater acoustic signal sensing network, continuously acquiring marine acoustic environmental and underwater acoustic data. This enables large-scale, deep, and long-term synchronous and effective three-dimensional observation of the target sea area and ocean body.

[0072] Therefore, addressing the shortcomings of existing technologies, this invention coordinates fixed node 10 and mobile node 20 to form a distributed three-dimensional underwater acoustic signal sensing network, creating a more comprehensive, stable, reliable, and high-performance ocean acoustic three-dimensional synchronous observation network. Compared to traditional ocean observation methods, the synchronous observation network of this invention offers a larger observation range, higher observation accuracy, higher spatiotemporal resolution, and lower observation costs.

[0073] To better understand the above technical solutions, exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that the present invention can be understood more clearly and thoroughly, and that the scope of the present invention can be fully conveyed to those skilled in the art.

[0074] The fixed node 10 includes an acoustic buoy, and the mobile node 20 includes an underwater glider carrying acoustic payloads and environmental observation components. The acoustic buoy observation network has a wide coverage area, enabling continuous monitoring of a large sea area and long-term acquisition of ocean data. The underwater glider is flexible and maneuverable, able to cover gaps between buoys and blind spots in observation, and can conduct high-precision or intensive observations of key areas as needed. Preferably, both the acoustic buoy and the underwater glider are lightweight devices with cost and power consumption below a threshold.

[0075] Next, cross-validation of the collected data is performed between the adjacent fixed node 10 and mobile node 20, retaining only the data that passes the validation. On the other hand, the fixed buoy and the underwater glider work together to cross-validate the observation data, improving data quality and accuracy. The integrated observation network fully leverages the advantages of both platforms; its modular structure allows for rapid reconfiguration and easier network expansion, providing superior responsiveness to different monitoring tasks.

[0076] Furthermore, any fixed node 10 can serve as a communication relay point between the mobile node 20 and other fixed nodes 10. Any fixed node 10 can provide communication guarantees for the mobile node 20 and other fixed nodes 10, enhancing the robustness of the network.

[0077] Furthermore, the ocean three-dimensional synchronous observation system also includes: a shore station 60, which is used to issue observation tasks including ocean environmental noise, ocean sound propagation, ocean reverberation and synchronous ocean environment to the surface ship 40, and to receive the data transmitted back from the surface ship 40 through the communication satellite 50, analyze and utilize it to output corresponding control commands.

[0078] In a specific embodiment, such as Figure 2As shown, the main body of the observation network consists of surface ships 40, several acoustic buoys, and several underwater gliders. The surface ships 40 and shipborne UAVs 30 enable rapid deployment, while GPS positioning and timing are used to acquire unified spatiotemporal information. During operation, the buoys are responsible for observation of the sea area near their fixed positions, while the underwater gliders operate along a zigzag trajectory, maneuvering to observe and fill in gaps in buoy coverage. They can also conduct intensive observations of key areas. The coordinated operation of these two systems allows for the acquisition of high spatiotemporal resolution, three-dimensional observation data.

[0079] The surface vessel 40 carries a sound source 31 and moves along a zigzag route along the deployment path of the acoustic buoy, continuously emitting acoustic signals. The acoustic buoy transmits time and location information and recorded acoustic signals to the surface vessel 40 via radio communication. When the underwater glider runs along the sawtooth trajectory (sea surface), it transmits real-time time and location information acquired by GPS, marine environmental information such as temperature, salinity, current, and density at different water depths, and recorded acoustic signals to the surface vessel 40 via radio communication. After receiving the data, the surface vessel 40 processes the data on board and transmits it back to the shore station 60 via a communication satellite 50.

[0080] Furthermore, embodiments of the present invention also provide a method for coordinated fixed and mobile ocean three-dimensional synchronous observation, used to construct the system described above, comprising:

[0081] S1. Based on the acquired observation mission, the surface vessel arrives at the target ocean water body designated by the observation mission, carrying UAVs, fixed node 10 and mobile node 20.

[0082] S2. The surface ship 40 controls the shipborne UAV 30 to deploy several fixed nodes 10 to designated positions according to the fixed node 10 deployment command issued by the shore station 60 to form a rectangular network.

[0083] S3. The surface ship 40 controls the shipborne UAV 30 to deploy several mobile nodes 20 to the designated location according to the mobile node deployment command issued by the shore station 60. This allows the mobile nodes 20 to move along a zigzag route according to the navigation trajectory included in the mobile node deployment command, thereby enabling the fixed node 10 and the mobile node 20 to cooperate to cover the blank areas and blind spots between nodes.

[0084] S4. When the surface ship 40 moves along the zigzag route of the fixed node 10, it continuously emits underwater acoustic signals through the preset sound source 31. It collects information on the fixed node and the mobile node through radio communication, and adjusts the navigation trajectory of the mobile node 20 and the parameters of the fixed node 10 in combination with the observation mission, so that the ocean three-dimensional synchronous observation network formed by the mobile node 20 and the fixed node 10 can maintain the set coverage accuracy of the target ocean water body at all times.

[0085] Furthermore, before step S2, the following steps are also included: A1, the surface vessel 40 receives the shore station 60 based on the general environmental information of the target ocean water body, the starting point of the surface vessel, the basic parameters of the UAV, the basic parameters of the fixed node, and the basic parameters of the mobile node, and outputs a fixed node deployment command containing the deployment position of the fixed node 10 and a mobile node deployment command containing the deployment position and navigation trajectory of the mobile node 20 through an environment-assisted planning algorithm with the objective function of covering the three-dimensional space of the surveyed sea area.

[0086] To address the optimization of the established marine three-dimensional synchronous observation network, this invention proposes a multi-objective optimization algorithm based on the genetic particle swarm optimization algorithm as an environment-aided planning algorithm. Furthermore, it introduces the depth deterministic policy gradient algorithm (DDPG) to achieve adaptive parameter adjustment of the genetic particle swarm optimization algorithm, thereby further enhancing the algorithm's global and local search capabilities.

[0087] The optimization criteria for the placement of fixed-node underwater acoustic buoys are as follows: 1. Dynamically plan the deployment positions of the fixed-node underwater acoustic buoys by combining the original underwater acoustic data of the sea area to be measured (such as seabed topography, seabed reverberation, sound velocity profile, etc.) and the parameters of the underwater acoustic buoys. 2. Further optimize the deployment by using real-time environmental information fed back by the buoys during the dynamic planning process. 3. Dynamically plan the optimal deployment positions of the fixed-node underwater acoustic buoys based on the principle of prioritizing collaborative detection in key areas and avoiding duplicate detection as much as possible. 4. Treat the deployment of underwater acoustic buoys by UAVs as a Traveler's Problem (TSP) for path planning.

[0088] The zigzag trajectory of the maneuvering nodes is optimized based on the monitoring blind spots, blank areas, and key areas of the fixed acoustic buoys. The main optimization parameters include glider spacing, turnaround timing, and turnaround depth to create an efficient three-dimensional observation network. The optimization principles are: 1. Allocate appropriate starting positions based on the area to be surveyed (length, width, depth), the number of unmanned gliders, and mission requirements to avoid redundant surveys and maximize efficiency. 2. To avoid efficiency degradation and data redundancy caused by redundant surveys, the unmanned gliders turn back immediately after entering the detection range of the acoustic buoys. For key areas, intensive observation is conducted to acquire high spatiotemporal resolution three-dimensional observation data. 3. Dynamically adjust the navigation parameters of the unmanned gliders, such as speed, depth, and turnaround angle, based on real-time environmental information and mission requirements.

[0089] Therefore, as Figure 3 As shown, the specific algorithm steps are as follows:

[0090] Algorithm input: general environmental information of the sea area to be detected, starting point of surface ships, basic parameters of acoustic buoys (including number and performance), basic parameters of UAVs (including number and performance), and basic parameters of gliders (including number and performance).

[0091] (1) Initialization: ① Initialize the population and genetic particle swarm optimization (GSO) parameters: Randomly generate a set of particles, each representing a solution to a set of decision variables. Provide the initial crossover probability, mutation probability, inertia weight, etc. ② Initialize the parameters of the DDPG algorithm: including Actor and Critic network parameters, learning rate, discount factor, etc. ③ Initialize the memory cache R to store information such as state, action, and reward.

[0092] (2) Fitness Evaluation: Fitness evaluation is performed for each particle. In the optimization of this problem, each particle has multiple objective functions (detection range of the acoustic buoy, path of the UAV to place the acoustic buoy, deployment position and flight path of the glider), and it is necessary to calculate the fitness of each particle on multiple objectives.

[0093] (3) Parameter encoding: Based on the parameter range and fitness of the genetic particle swarm algorithm, the parameters are encoded into vectors and used as inputs to DDPG.

[0094] (4) Non-dominated sorting: Based on the objective function value of the particles, the particles are sorted in a non-dominated manner to divide the particles into different Pareto Fronts.

[0095] (5) Calculate the crowding: Calculate the crowding of particles in each front to determine the distribution density of particles in the target space.

[0096] (6) Genetic operations: including selection, crossover, and mutation. These operations update the population by retaining the better solution, crossover, and introducing new mutations, based on non-dominated sorting and crowding calculation. The parameters such as selection probability, crossover probability, and mutation probability are optimized automatically with iteration.

[0097] (7) Update the particle swarm: Perform N updates to the particle swarm based on the results of the genetic operation to achieve local optimization and fast convergence. The improved particle velocity update formula is as follows. The parameters ω, c1, c2, r1, and r2 of the ordinary PSO are all given the subscript t, which is the parameter at the current iteration number.

[0098] v ij (t+1)=ω t v ij (t)+c 1t r1(x pbestij -x ij (t))+c 2t r2(x gbest -x ij (t)) (1)

[0099] In equation (1), ω t c 1t c2t Let r1 and r2 be the inertial weights after the t-th iteration and the two parameters controlling the acceleration, respectively, where r1 and r2 are random numbers in the interval [0,1]. pbestij x represents the historical best position of particle i. gbest The historical best position representing the population, v ij and x ij Let $j$ represent the velocity and position of particle $i$ in the $j$-th dimension, respectively.

[0100] (8) Execute the DDPG algorithm: After each update iteration of the genetic particle swarm optimization algorithm, execute action a in the particle population for the predicted current iteration number. t And observe the reward r t and new state s t+1 , will the empirical tuple (s t ,a t ,r t ,s t+1 The parameters of the Actor and Critic networks are stored in a memory cache and then trained and updated using a deep deterministic policy gradient algorithm based on the actions, states, and rewards in the memory cache.

[0101] Actor Network Update: The Actor network parameters are updated using the policy gradient algorithm. The loss function of the Actor network can be expressed as:

[0102]

[0103] In equation (2), Represents the parameters θ of the Actor network μ The gradient of the policy function J, N is the number of samples in the buffer, s is the current state, and θ Q For Critic network parameters, For the Critic network, for state s i and by Actor network in s i Deterministic action μ(s) output in the state i The gradient of the estimated Q value with respect to action a. For an Actor network with state s i The output deterministic action with respect to its parameter θ μ The gradient of the policy is obtained by multiplying and summing these two gradients. Approximate value.

[0104] Update target network parameters: θ Q′ ←τθ Q +(1-τ)θ Q′ ,θ μ′ ←τθ μ +(1-τ)θμ′ In equation (4), θ Q For the Critic network parameters, θ Q′ For the updated Critic network parameters, θ μ Let θ be the Actor network parameter. μ′ Let τ be the updated Actor network parameters, and τ be the learning rate. Using the "soft update" method shown in the above formula can reduce the update speed of the target value and improve the convergence stability of the algorithm.

[0105] Selecting and executing an action: The current action is obtained from the current state and the Actor network. The predicted action can be represented as: a = μ(s|θμ), where s represents the current state and μ(s|θμ) represents the Actor network's prediction of the action.

[0106] (9) Parameter decoding: Decode the parameter vector learned by DDPG into the parameter settings of the genetic particle swarm algorithm.

[0107] (10) Optimize the glider's flight path and the position of the acoustic buoy in real time by combining the latest marine environmental information.

[0108] (11) Repeated iteration: Repeatedly execute the above steps (2) to (10) until the maximum number of iterations M is reached or the convergence reaches a certain level.

[0109] (12) Output the optimal particle and its fitness value.

[0110] (13) Algorithm output: underwater acoustic buoy deployment, UAV deployment path, optimal deployment position and optimal flight trajectory of glider, and overall real-time coverage, thereby generating fixed node deployment instructions and mobile node deployment instructions.

[0111] Furthermore, after step S3, the process also includes: establishing a radio communication link between the surface vessel 40 and the fixed node 10 and the mobile node 20 after deployment for real-time communication; the surface vessel 40 continuously acquires the location and timing information of each node through the interaction between the mobile node 20, the fixed node 10 and the GPS, and finally obtains the fixed node information and the mobile node information of the target ocean water body matching spatiotemporal information through the mobile node 20 and the fixed node 10.

[0112] In the above steps, the rapid deployment of the observation network is as follows: First, a surface ship 40 carrying multiple fixed nodes 10 and mobile nodes 20 arrives at the vicinity of the mission area; second, given the general environmental information of the sea area, the shore station 60 uses an environment-aided planning algorithm to obtain the optimal deployment positions of the acoustic buoys and gliders, as well as the optimal flight path of the gliders; then, the UAV takes off sequentially from the surface ship 40 and deploys multiple sets of acoustic buoys to preset positions according to a straight flight path, forming a rectangular buoy network. The buoys obtain position and timing information in real time via GPS; furthermore, multiple sets of underwater gliders are deployed from the surface ship 40 or the UAV, maneuvering to preset positions on their own, or airdropped to preset positions by the UAV, and running along a zigzag trajectory according to preset flight paths and water depths.

[0113] To address the survey needs related to marine environmental noise, sound propagation, reverberation, and synchronous marine environment, a highly efficient three-dimensional synchronous marine observation network, combining fixed and mobile systems, can be deployed in various forms in local sea areas. This allows for the accumulation of a large amount of marine environmental data, gradually improving the spatiotemporal acquisition density and accuracy of marine environmental data. Specific embodiments are described below:

[0114] When conducting marine environmental noise surveys, the underwater acoustic buoys transmit time and location information and recorded acoustic signals to the surface ship 40 via radio communication. When the underwater glider runs along a sawtooth trajectory (on the sea surface), it transmits the real-time time and location information and recorded acoustic signals acquired by GPS to the surface ship 40 via radio communication. After receiving the data, the surface ship 40 processes the data on board and transmits it back to the shore station 60 via satellite communication.

[0115] When conducting ocean reverberation and underwater acoustic propagation surveys, surface vessel 40 carries sound source 31 and proceeds along a zigzag route following the deployment path of the acoustic buoys, continuously emitting underwater acoustic signals. Data transmitted from the acoustic buoys and underwater gliders is sent to surface vessel 40, processed on board, and then transmitted back to shore station 60 via satellite communication.

[0116] When performing synchronous marine environmental surveys, the underwater glider will transmit real-time GPS-acquired time and location information, as well as marine environmental information such as temperature, salinity, current, and density at different water depths, back to the shore station 60 via satellite communication when it travels along a sawtooth trajectory (sea surface).

[0117] In summary, this invention discloses a fixed and mobile coordinated three-dimensional synchronous ocean observation system and method, proposing a three-dimensional synchronous ocean observation network covering the sea surface and ocean water body, composed of lightweight underwater acoustic receiving nodes. The network uses two types of lightweight nodes: fixed nodes 10 and mobile nodes 20. Fixed nodes 10 remain spatially stationary, used for long-term in-situ acquisition of underwater acoustic signals or environmental information at a fixed location; mobile nodes 20 continuously dive and surface between the sea surface and seabed to achieve zigzag movement, used for long-term acquisition of underwater acoustic signals or environmental information along their travel route. Fixed nodes 10 and mobile nodes 20 are distributed throughout the entire ocean water body observation space, cooperating to form a distributed three-dimensional underwater acoustic signal sensing network. With known environmental information of the target sea area, the node deployment scheme is designed with the support of an environment-aided planning algorithm. According to the node deployment plan, the fixed node 10 and the mobile node 20 are handed over to the surface ship 40 and the shipborne UAV 30 for rapid deployment. By obtaining unified spatiotemporal information through GPS positioning and timing, it can collect marine acoustic environment data and underwater acoustic data with matching spatiotemporal information in a large area of ​​the ocean for a long time and efficiently.

[0118] like Figure 4 As shown, the detection service process using the above network is as follows:

[0119] (1) Given the survey sea area and the general environmental information within the area, the deployment positions of the acoustic buoys and underwater gliders are calculated using an environmental-assisted planning algorithm (preferably particle swarm optimization method) with the objective function of covering the three-dimensional space of the survey sea area, thus completing a customized sensing network scheme that takes into account environmental effects.

[0120] (2) Multiple shipborne UAVs 30 carrying multiple acoustic buoys take off from the surface ship 40, deploy the buoys to the designated positions, and underwater gliders are launched from the surface ship 40 or the shipborne UAVs 30 and move to the preset positions on their own.

[0121] (3) The underwater acoustic buoy serves as a fixed node 10, continuously collecting marine environmental noise, receiving underwater acoustic signals, and transmitting them to the surface ship 40 for signal processing via radio communication.

[0122] (4) The underwater glider acts as a maneuvering node 20, running along the planned path in a sawtooth trajectory, continuously collecting marine environmental information, environmental noise and receiving underwater acoustic signals, which are then transmitted to the surface ship 40 for signal processing via radio communication.

[0123] (5) After the fixed acoustic buoy and the mobile underwater glider are deployed, a radio communication link is established with the surface ship 40 for real-time communication. The mobile node 20 and the fixed node 10 cooperate to continuously acquire the location and timing information of each node through GPS, and continuously acquire environmental and underwater acoustic information at different locations and times in the local three-dimensional space. The fixed node 10 can continuously acquire target and environmental information at fixed locations, realizing long-term observation of fixed-point and fixed-depth location data. The mobile node 20 acquires information over a wide range of water depths, and can conduct targeted detection of special underwater areas, making up for the information blind spots detected by the fixed node 10.

[0124] (6) The surface ship 40 collects and processes the sensor data of each node through radio communication, and adjusts the underwater glider track and the parameters of the underwater acoustic buoy according to the observation needs to ensure that the ocean observation process covers the three-dimensional space of the target sea area.

[0125] (7) The information processed by the surface ship 40 is transmitted back to the shore station 60 for analysis and utilization via satellite communication.

[0126] Therefore, this invention organically combines fixed node monitoring platforms with mobile node monitoring platforms, realizing a marine three-dimensional observation network with superior performance in terms of observation efficiency, range, accuracy, timeliness, and cost.

[0127] Since the systems / devices described in the above embodiments of the present invention are systems / devices used to implement the methods of the above embodiments of the present invention, those skilled in the art can understand the specific structure and modifications of the systems / devices based on the methods described in the above embodiments of the present invention, and therefore will not be repeated here. All systems / devices used in the methods of the above embodiments of the present invention fall within the scope of protection of the present invention.

[0128] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0129] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions.

[0130] It should be noted that any reference numerals placed between parentheses in the claims should not be construed as limiting the claims. The word "comprising" does not exclude the presence of components or steps not listed in the claims. The word "a" or "an" preceding a component does not exclude the presence of a plurality of such components. The invention can be implemented by means of hardware comprising several different components and by means of a suitably programmed computer. In claims that enumerate several means, several of these means may be embodied by the same hardware. The use of the terms first, second, third, etc., is merely for convenience of expression and does not indicate any order. These terms can be understood as part of the component names.

[0131] Furthermore, it should be noted that in the description of this specification, the terms "one embodiment," "some embodiments," "embodiment," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Furthermore, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0132] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the claims should be interpreted to include both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0133] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, then this invention should also include these modifications and variations.

Claims

1. A three-dimensional synchronous ocean observation system that combines fixed and mobile surveillance, characterized in that, include: Fixed nodes are configured at fixed locations within the target ocean water body to form a rectangular network, used to collect fixed node information including underwater acoustic signals at fixed locations, environmental information, node positions, and timing information. Mobile nodes are deployed between the surface and seabed of the target ocean body, constantly diving and surfacing and moving along a zigzag route. They are used to collect mobile node information along the route, including underwater acoustic signals, environmental information, current node position, and timing information. Fixed nodes and mobile nodes work together to cover the blank areas and blind spots between nodes. Shipborne drones, using surface ships as bases, are used to carry fixed and mobile nodes and deploy them to preset locations; Surface ships carry sound sources and are configured to continuously emit underwater acoustic signals as they advance along a zigzag route along the deployment path of fixed nodes. They also collect information from fixed and mobile nodes via radio communication. Combined with the observation mission, the navigation trajectory of the mobile nodes and the parameters of the fixed nodes are adjusted so that the three-dimensional synchronous ocean observation network formed by the mobile and fixed nodes can maintain the set coverage accuracy of the target ocean water body at all times. The shore station is configured to, based on the acquired general environmental information of the target ocean body, the starting point of surface ships, the basic parameters of UAVs, the basic parameters of fixed nodes, and the basic parameters of mobile nodes, use an environment-aided planning algorithm with the objective function of covering the three-dimensional space of the surveyed sea area. It outputs fixed node deployment instructions containing the deployment locations of fixed nodes and mobile node deployment instructions containing the deployment locations and navigation trajectories of mobile nodes, specifically including: A11. The general environmental information of the target ocean water body, the starting point of the surface ships, the basic parameters of the UAV, the basic parameters of the fixed nodes, and the basic parameters of the maneuvering nodes are used as inputs to a pre-constructed multi-objective optimization model based on the genetic particle swarm optimization algorithm and the deep deterministic policy gradient algorithm. A12. Initialize the population by randomly generating a set of particles, each representing a solution to a set of decision variables, and initialize the parameters of the genetic particle swarm optimization algorithm containing crossover probability, mutation probability, and inertia weight, as well as the parameters of the deep deterministic policy gradient containing the Actor and Critic networks, the learning rate and discount factor, and the memory cache used to store information containing state, action and reward. A13. Based on the multiple objective functions corresponding to each particle, evaluate the fitness value of each particle on the multiple objective functions; A14. Encode the parameters of the current genetic particle swarm optimization algorithm and the fitness values ​​of the particles on multiple objective functions into vectors as input to the deep deterministic policy gradient algorithm; A15. Based on the fitness values ​​of the particles on multiple objective functions, perform non-dominated sorting of the particles and divide the particles into different frontiers; A16. Calculate the crowding degree of particles in each front to determine the particle distribution density in the target space; A17. The population is updated by genetic operations such as crossover and introducing new variations, based on the frontier obtained from non-dominated sorting and density calculated by crowding. A18. To achieve local optimization and fast convergence, the population is updated by performing N particle swarm optimization iterations on the results of genetic operations. A19. After each update iteration of the genetic particle swarm optimization algorithm, execute the action predicted for the current iteration number in the particle swarm. And observe the reward returns. and new status , experience tuple The parameters of the Actor and Critic networks are stored in the memory cache and then trained and updated using a deep deterministic policy gradient algorithm based on the actions, states, and rewards in the memory cache. A20. Decode the parameter vector learned by the deep deterministic strategy gradient algorithm into the parameter settings of the genetic particle swarm algorithm, and combine the latest collected marine environmental information to optimize the navigation trajectory of the maneuvering node and the position of the fixed node in real time. A21. Repeat steps A13 to A20 until the maximum number of iterations is reached or the convergence reaches a certain level. Output the optimal particle and its fitness value, and then obtain the optimization parameters output by the multi-objective optimization model, including the fixed node layout, UAV deployment path, optimal deployment position and optimal flight trajectory of the mobile node, and the real-time coverage of the synchronous observation network. Then generate the fixed node deployment command and the mobile node deployment command.

2. The fixed and mobile coordinated ocean three-dimensional synchronous observation system as described in claim 1, characterized in that, Fixed nodes include acoustic buoys, and mobile nodes include underwater gliders; both acoustic buoys and underwater gliders are lightweight devices with cost and power consumption below the threshold.

3. The fixed and mobile coordinated ocean three-dimensional synchronous observation system as described in claim 1, characterized in that, Any fixed node can serve as a communication relay point between mobile nodes and other fixed nodes.

4. The fixed and mobile coordinated ocean three-dimensional synchronous observation system as described in claim 1, characterized in that, Cross-validation of collected data is performed between nearby fixed nodes and mobile nodes, and only the data that passes the validation is retained.

5. The fixed and mobile coordinated ocean three-dimensional synchronous observation system as described in any one of claims 1-4, characterized in that, The shore station is also configured to issue observation tasks to surface ships, including marine environmental noise, ocean sound propagation, ocean reverberation, and synchronous marine environment, as well as to receive data transmitted back from surface ships via communication satellites, analyze and utilize the data, and output corresponding control commands.

6. A method for coordinated fixed and mobile three-dimensional synchronous ocean observation, used to construct a system as described in any one of claims 1-5, characterized in that, include: Based on the acquired observation mission, surface ships, carrying drones, fixed nodes, and mobile nodes, arrive at the target ocean waters designated by the observation mission. Surface ships control shipborne UAVs to deploy several fixed nodes to designated locations based on fixed node deployment instructions issued by shore stations, forming a rectangular network. Based on the mobile node deployment instructions issued by the shore station, the surface ships control the shipborne UAVs to deploy several mobile nodes to the designated locations, so that the mobile nodes can move along a zigzag route according to the navigation trajectory contained in the mobile node deployment instructions, thereby enabling the fixed nodes and mobile nodes to cooperate to cover the blank areas and blind spots between the nodes. When surface ships advance along a zigzag route along the deployment path of fixed nodes, they continuously emit underwater acoustic signals through preset sound sources, collect information on fixed nodes and mobile nodes through radio communication, and adjust the navigation trajectory of mobile nodes and the parameters of fixed nodes in conjunction with the observation mission, so that the ocean three-dimensional synchronous observation network formed by mobile nodes and fixed nodes can maintain the set coverage accuracy of the target ocean water body at all times.

7. The method for coordinated fixed and mobile oceanographic observation as described in claim 6, characterized in that, The population is updated by performing N iterations of particle swarm optimization, which relies on the following particle velocity update formula: （1） In equation (1), , , The first t The inertial weights and two parameters controlling acceleration after the next iteration. and A random number in the interval [0,1]. Representative particles i The best historical location x gbest Represents the best historical position of the population. and Representing particles i The j Dimensional velocity and position; Updating the parameters of the Actor network depends on the following implementation of the Actor network's loss function: （2） In equation (2), Indicates the parameters of the Actor network. strategy function gradient, N The number of samples in the buffer. s This is the current state. For Critic network parameters, For the Critic network, the state and by Actor network in Deterministic actions output in a given state The estimated Q Value related to action a gradient, For the Actor network, for the state The output deterministic action with respect to its parameters The gradient; Updating the parameters of the Critic network depends on the following implementation of the Critic network's loss function: （3） In equation (3), The loss function for the Critic network is... For the sample in state Actions taken at that time , The target action value is calculated for the Critic network; The parameters of the Critic network and Actor network are updated as follows: （4） In equation (4), For Critic network parameters, For the updated Critic network parameters, For Actor network parameters, For the updated Actor network parameters, The learning rate; The actions predicted by the Actor network are implemented based on the following action prediction formula of the Actor network: （5） In equation (5), This represents the Actor network's prediction of actions.

8. The method for coordinated fixed and mobile oceanographic observation as described in claim 6, characterized in that, Surface ships, based on mobile node deployment commands issued by shore stations, control shipborne UAVs to deploy several mobile nodes to designated locations. These mobile nodes then follow a zigzag path defined in the deployment commands, enabling fixed and mobile nodes to coordinate and cover gaps and blind spots in observation and detection. This process also includes: A radio communication link is established between the surface ships and the fixed and mobile nodes after deployment for real-time communication. Surface ships continuously acquire the location and timing information of each node through the interaction between mobile nodes, fixed nodes, and GPS. Ultimately, they obtain fixed node information and mobile node information that match the spatiotemporal information of the target ocean body by collecting data from both mobile and fixed nodes.

Images