Method, system, device and medium for underwater multi-modal hybrid cross-layer opportunistic transmission

Abstract

This invention provides a method, system, device, and medium for underwater multimodal hybrid cross-layer opportunistic transmission, enabling end-to-end differentiated transmission of multiple services under multiple heterogeneous transmission modes underwater. The method achieves accurate evaluation of the forwarding performance of heterogeneous transmission modes through an initial set of candidate relay nodes and a method for selecting the optimal candidate relay region. Addressing the need for differentiated service assurance across multiple network services, a hierarchical decision-making strategy is proposed. The upper-layer strategy selects the optimal transmission mode based on service state characteristics, while the lower-layer strategy completes fine-grained routing parameter configuration under the constraints of the selected mode, reducing the computational complexity of decision-making. To address the limitations of underwater resource-poor network environments, a reward update method with state-action space traversal is adopted, enabling nodes to simultaneously calculate simulated rewards corresponding to different service types and routing parameters based on feedback from a single service transmission, while simultaneously calculating rewards for actual actions.

Description

Technical Field

[0001] This invention belongs to the field of underwater communication network routing technology, specifically relating to a method, system, device, and medium for underwater multimodal hybrid cross-layer opportunistic transmission. Background Technology

[0002] Underwater wireless communication networks are a key technology for applications such as maritime security surveillance, marine resource exploration, and marine environmental monitoring. The reliability, real-time performance, and energy efficiency of end-to-end data transmission are of paramount importance. However, the complex and dynamic marine environment poses continuous challenges to robust and efficient routing design.

[0003] Currently, underwater routing solutions are mainly based on two physical layer technologies: underwater acoustics and underwater optical. Underwater acoustic links, with their excellent attenuation characteristics in seawater, have become the mainstream means of long-distance propagation. However, their low propagation speed and narrow bandwidth limit data rates, typically not exceeding a few kilobits per second. Furthermore, the high power consumption of underwater acoustic modems hinders their practical application due to high latency. Underwater optical links, on the other hand, offer high data transmission rates and low latency, but are limited by severe water absorption and scattering effects, resulting in a relatively short effective communication distance, generally tens to hundreds of meters, and requiring high alignment. To fully utilize the complementary advantages of heterogeneous underwater acoustic / underwater optical links, underwater acoustic-optical hybrid networks have been proposed, aiming to achieve high-bandwidth data transmission, low latency, and wide network coverage.

[0004] At the routing strategy level, the underwater environment poses challenges to the transmission reliability of both types of links. Underwater acoustic links, due to severe multipath propagation and Doppler shift, are prone to signal distortion and random link interruptions under time-varying / space-varying channel conditions. Underwater optical links are severely constrained by water turbidity; particulate matter and marine organisms in the water cause rapid attenuation of the optical signal, while dynamic ocean currents and node movement reduce the pointing accuracy of the underwater optical link. These environmental factors collectively reduce the reliability of underwater acoustic-optical hybrid networks, and this cannot be solved simply by mixing the two types of links. Opportunistic routing, through a large redundant set of relay candidate nodes, can effectively combat random link interruptions, improve data transmission reliability, and effectively cope with dynamic underwater environments. Therefore, combining underwater acoustic-optical hybrid networks with opportunistic routing has become a new direction in underwater routing research.

[0005] Chinese invention patent CN202311233679.4 proposes an underwater cross-layer transmission mechanism with a hybrid acoustic-optical design. This method leverages the high bandwidth and high-speed data transmission advantages of underwater optical links to forward data packets, and utilizes the omnidirectional and long-distance communication advantages of underwater acoustic links to send forwarding claims (ACKs), enabling relay candidate nodes to synchronously compete for data packet forwarding qualifications. This achieves fast, high-capacity, and reliable data transmission based on the combination of acoustic and optical technologies. While this protocol provides a new approach to underwater acoustic-optical fusion transmission, it still has some limitations: Firstly, the use of a fixed data packet transmission mode and static routing parameter configuration cannot fully adapt to dynamic changes in underwater network topology, making it difficult to dynamically and optimally utilize heterogeneous link resources in complex marine environments, thus limiting the overall network performance ceiling. Secondly, the use of a unified measurement standard for all data streams lacks the ability to perceive and guarantee the differentiated service quality requirements of multiple services, making it difficult to meet the complex needs of multi-scenario task collaboration underwater.

[0006] In summary, existing technologies lack an underwater opportunistic routing solution that can deeply integrate the characteristics of heterogeneous acoustic-optical links, intelligently adapt to dynamic environments, and accurately guarantee the needs of multiple services, which severely restricts the application potential of underwater acoustic-optical hybrid networks in complex and dynamic marine environments. Summary of the Invention

[0007] To address the technical problem of low efficiency in the collaborative utilization of heterogeneous links of underwater acoustic and underwater optical intensity in existing large-scale underwater monitoring / detection networks, which fails to guarantee the differentiated transmission needs of service-driven networks, this invention provides a method, system, device, and medium for underwater multimodal hybrid cross-layer opportunistic transmission. The aim is to dynamically combine end-to-end multimodal transmission paths according to differentiated service needs, thereby achieving highly reliable and low-latency data transmission.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] In a first aspect, embodiments of this disclosure provide a method for underwater multimodal hybrid interlayer opportunistic transport, comprising the following steps:

[0010] Step S1: Each network node periodically broadcasts beacon information carrying the node's unique ID and spatial coordinates in an underwater acoustic mode. Each network node parses and receives the beacon information to obtain the initial set of candidate relay nodes corresponding to each transmission mode.

[0011] Step S2: Based on the underwater mission scenario, the services to be forwarded are divided into three categories: latency-sensitive, bandwidth-sensitive, and periodic. Nodes in the network containing services to be forwarded are used as forwarding nodes. Based on the initial relay candidate node set and the mode transmission angle, the forwarding nodes divide the initial relay candidate regions for each transmission mode and select the best relay candidate region for each transmission mode from the initial relay candidate regions.

[0012] Step S3: The forwarding node constructs state features based on the target destination node and service classification of the service to be forwarded. Based on the return values ​​of each transmission mode action under the state features, it selects the actual transmission mode action using a "random-greedy" strategy as the upper-layer decision. Combining the state features and the upper-layer decision, and based on the return values ​​of each routing parameter action, it determines the actual routing parameters for service forwarding using a "random-greedy" strategy as the lower-layer decision. The forwarding node embeds the actual routing parameters corresponding to the lower-layer decision into the header of the data packet of the service to be forwarded, and completes the forwarding of the service to be forwarded according to the actual transmission mode action determined by the upper-layer decision.

[0013] Step S4: Nodes that successfully receive and forward services within the best relay candidate area are designated as relay candidate nodes. Each relay candidate node calculates its corresponding forwarding factor based on its best relay candidate area, remaining node energy, and queue length for each transmission mode, and determines the waiting interval based on the forwarding factor. After the waiting interval expires, the relay candidate node sends a forwarding declaration to participate in forwarding competition. The relay candidate node that is qualified to forward services in the forwarding competition is designated as the next-hop forwarding node.

[0014] Step S5: Based on the received forwarding declaration, the forwarding node calculates the actual reward for the actions actually performed by the upper-layer and lower-layer decisions, and calculates the simulated reward corresponding to different service types and routing parameter configuration combinations under the same transmission mode using the reward update method of state-action space traversal. Based on the actual reward and simulated reward, the reward Q value of each action of the lower-layer and upper-layer decisions is updated sequentially. If the service is successfully forwarded to the target destination node, the forwarding of the forwarding service is completed; otherwise, steps S2-S5 are executed repeatedly until the service is successfully forwarded to the target destination node.

[0015] Furthermore, in step S1, when each network node parses and receives beacon information, it records the information of adjacent nodes within the beacon communication radius. If adjacent nodes satisfy the transmission mode... lower condition set Then it is included in the transmission mode. The initial set of candidate relay nodes, the condition set include:

[0016] ;

[0017] In the formula, Indicates transmission mode Maximum service communication radius , Represents the set of transmission modes; Indicates from the forwarding node To adjacent nodes The Euclidean distance; , This is the set of all nodes in the underwater network.

[0018] Furthermore, when obtaining the optimal relay candidate region in step S2, the following steps are included:

[0019] Step A1: Based on the transmission mode launch angle The entire relay candidate area is divided into multiple equal parts to form a transmission mode. Initial Relay Candidate Region Set :

[0020] ;

[0021] Step A2: Calculate different transmission modes respectively Each initial relay candidate node in the initial relay candidate node set within the initial relay candidate region Distance factor :

[0022] ;

[0023] In the formula, Indicates from the forwarding node To the target destination node of the service to be forwarded The Euclidean distance; Indicates starting from the initial relay candidate node To the target destination node of the service to be forwarded The Euclidean distance;

[0024] Step A3: Aggregate Transmission Mode Distance factor of the initial relay candidate node set within the initial relay candidate region The transmission modes are obtained. The following are the initial relay candidate regions candidate factors :

[0025] ;

[0026] In the formula, Indicates a forwarding node In transmission mode Initial relay candidate region The initial set of candidate relay nodes within; , Indicates transmission mode The initial set of relay candidate regions is determined; candidate factor values ​​are used to describe the region's forwarding potential.

[0027] Step A4: Forwarding Node Select transmission mode The initial relay candidate region with the largest lower candidate factor is selected as the optimal relay candidate region for the corresponding transmission mode. ;

[0028] .

[0029] Furthermore, the state features in step S3 include:

[0030] ;

[0031] In the formula, Forwarding business The target node identifier ID; Forwarding business Business categories;

[0032] Based on state features The state characteristics of the responses to each transmission mode action are determined using a "random-greedy" strategy. Actual transmission mode action :

[0033] ;

[0034] In the formula, For the business to be forwarded Transmission modes; The optimal relay candidate region corresponding to the transmission mode;

[0035] The reward for each routing parameter action based on state characteristics and upper-level decision-making. The actual routing parameters are obtained using a "random-greedy" strategy. :

[0036] ;

[0037] In the formula, Indicates a forwarding node In transmission mode The corresponding candidate factor parameters in the next routing parameters, This represents the energy parameter under the same conditions. This represents the queue parameters under the same conditions. This indicates the corresponding waiting interval under the same conditions.

[0038] Furthermore, in step S4, the forwarding factor include:

[0039] ;

[0040] In the formula, Indicates a forwarding node The transmission mode used for forwarding; Indicates a forwarding node relay candidate nodes In transmission mode Distance factor below; It represents the shortest communication radius among the transmission modes of each node in the network, and is used to balance the differences in communication range between heterogeneous links; This indicates the number of transmission modes carried by each node in the network; This represents the initial energy of each node in the network; This indicates the maximum queue length for each node in the network. Indicates a relay candidate node The remaining energy, Indicates a relay candidate node The length of the queue; Candidate factors for the best relay candidate region;

[0041] The waiting interval include:

[0042] ;

[0043] In the formula, This indicates the fixed communication radius for sending and forwarding ACK statements. It has twice the maximum transmission mode communication radius and a higher queuing priority than service data packets. This indicates the propagation speed of the forwarding statement; For transmission mode The communication radius; For the forwarding node To relay candidate node The Euclidean distance; For transmission mode The speed of propagation;

[0044] The process of competing for forwarding during the waiting interval includes the following steps:

[0045] If the relay candidate node If the target node is a relay candidate node, then... ID and forwarding services The target node ID is consistent, and the relay candidate node is... In receiving and forwarding services After sending a forwarding declaration, the remaining relay candidate nodes, upon receiving the forwarding declaration, relinquish their forwarding competition and delete any forwarding services they have already received. ;

[0046] If relay candidate node If the node is not the target node, then it is a relay candidate node. The startup duration is a timer corresponding to the waiting interval. If a forwarding service is received before the timer expires... Forwarding statement, relay candidate node Abandon forwarding competition and remove forwarding services from the queue. If no forwarding statement is received by the timer expires, the relay candidate node... Send with its own forwarding factor The forwarded statement, and wait If no forwarding statement with a smaller forwarding factor value is received during the waiting period, the relay candidate node will be considered a relay node. Capable of forwarding services If a node is selected for forwarding, it will be used as the next-hop forwarding node to continuously execute steps S2-S5; otherwise, it will abandon the forwarding competition and remove the forwarding service from the queue. .

[0047] Furthermore, in step S5, when obtaining the actual reward and simulated reward from the upper-level decision, the actual reward is... :

[0048] ;

[0049] In the formula, This indicates that the upper-level decision-making uses the actual transmission mode action. The penalty for not receiving a forwarding statement is used to suppress the selection of unreliable transmission modes;

[0050] In the state-action space traversal reward update method, it is assumed that the forwarding declaration feedback result of the simulated action is consistent with the transmission modal action. To maintain consistency, calculate the simulated rewards for other service categories under the same target node and the same transmission mode action, based on the upper-level decision. State features In the formula, Classify as other business categories, and meet the following requirements: and , A set of service categories within the network;

[0051] When obtaining the actual and simulated rewards for lower-level decisions, the following steps are included:

[0052] The actual routing parameter actions of the lower-level decision-making The actual reward is :

[0053] ;

[0054] In the formula, This represents the relative cost per hop of receiving a forwarding declaration. This indicates the relative cost of a single hop when no forwarding statement is received; This indicates a penalty for not receiving a forwarding declaration after the lower layer uses the actual routing parameters; This is a reward for a forwarding node that does not receive a forwarding statement. The smaller the value, the higher the corresponding reward; The reward for a forwarding node when it receives a forwarding statement. The smaller the value, the higher the corresponding reward;

[0055] ;

[0056] ;

[0057] In the formula, This represents the actual single-hop delay. This represents the simulated single-hop delay. and During calculation, self-forwarding nodes are included. Forwarding service Transmission, propagation, and queuing delays; This represents the actual single-hop energy consumption. This represents the virtual single-hop energy consumption. and The calculation includes provisions for forwarding services. The energy consumption of sending and forwarding statements; Indicates the reference transmission mode The actual time delay per unit distance, Indicates the reference transmission mode Actual energy consumption per unit distance; Indicates a forwarding node With the best local node The Euclidean distance between the nodes, where the local best node is located. It is a forwarding node In transmission mode Next, regarding forwarding services The actual set of relay candidate nodes corresponding to the forwarding declaration The node with the smallest forwarding factor; if no forwarding declaration is received, the communication radius of the reference mode is used. replace; Indicates forwarding service The weighting of single-hop latency is used to reflect the differentiated requirements of different services for single-hop latency and energy consumption;

[0058] In the reward update method based on state-action space traversal, for the lower-level action space... In addition to actual routing parameter actions Other routing parameter actions ,satisfy ,and Based on the local best node State calculation simulation reward The calculation ideally assumes a forwarding declaration feedback number of 1. Under the same state characteristics and the same transmission mode action, the simulated reward corresponds to different routing parameter actions. include:

[0059] ;

[0060] Other service classifications under the same destination node and the same transmission mode operation ,satisfy and Other business categories Simulated rewards for all corresponding lower-level routing parameter actions ,include:

[0061] ;

[0062] In the formula, Indicates a forwarding node In transmission mode Below, corresponding relay candidate area The set of actual relay candidate nodes corresponding to the received forwarding declaration.

[0063] Furthermore, in step S5, the rewards for each action of the lower-level decision-making are updated. ,include:

[0064] ;

[0065] In the formula, This represents the learning rate of lower-level decisions. This represents the discount factor for lower-level decisions;

[0066] When the rewards for each action in the upper-level decision are updated, the maximum reward is extracted from the updated reward value of the lower-level decision and used as the reward for the current best action. And based on the best action reward Update the rewards for each action in the upper-level decision-making process. ,include:

[0067] ;

[0068] ;

[0069] In the formula, This represents the learning rate of upper-level decisions. This represents the discount factor for upper-level decisions. Indicates the local best node The value of.

[0070] A system for underwater multimodal hybrid interlayer opportunistic transport includes:

[0071] The initialization unit is configured so that each network node periodically broadcasts beacon information carrying the node's unique ID and spatial coordinates in an underwater acoustic mode. Each network node parses and receives the beacon information to obtain the initial relay candidate node set corresponding to each transmission mode.

[0072] The service classification and forwarding unit is configured to classify the services to be forwarded into three categories: latency-sensitive, bandwidth-sensitive, and periodic, based on the underwater mission scenario. Nodes in the network containing the services to be forwarded are used as forwarding nodes. The forwarding nodes divide the initial relay candidate regions for each transmission mode based on the initial relay candidate node set and the mode transmission angle, and select the best relay candidate region for each transmission mode from the initial relay candidate regions.

[0073] The action decision unit is configured such that the forwarding node constructs state features based on the target destination node and service classification of the service to be forwarded. Based on the reward value of each transmission mode action under the state features, it selects the actual transmission mode action using a "random-greedy" strategy as the upper-layer decision. Combining the state features and the upper-layer decision, it determines the actual routing parameters for service forwarding using a "random-greedy" strategy based on the reward value of each routing parameter action as the lower-layer decision. The forwarding node embeds the actual routing parameters corresponding to the lower-layer decision into the header of the data packet of the service to be forwarded and completes the forwarding of the service according to the actual transmission mode action determined by the upper-layer decision.

[0074] The forwarding contention unit is configured to use nodes that have successfully received forwarded services within the best relay candidate area as relay candidate nodes. Each relay candidate node calculates a corresponding forwarding factor based on its best relay candidate area under each transmission mode, its remaining energy, and queue length, and determines a waiting interval based on the forwarding factor. After the waiting interval expires, the relay candidate node sends a forwarding declaration to participate in the forwarding contention. The relay candidate node that is qualified to forward services in the forwarding contention becomes the next-hop forwarding node.

[0075] The update unit is configured so that the forwarding node calculates the actual reward for the actual actions of the upper-layer decision and the lower-layer decision based on the received forwarding declaration, and calculates the simulated reward corresponding to different service types and routing parameter configuration combinations under the same transmission mode through a reward update method based on state-action space traversal. Based on the actual reward and the simulated reward, the reward Q value of each action of the lower-layer decision and the upper-layer decision is updated sequentially. If the service is successfully forwarded to the target destination node, the forwarding of the service is completed; otherwise, the service classification and forwarding unit, the action decision unit, the forwarding contention unit and the update unit are executed in a loop until the service is successfully forwarded to the target destination node.

[0076] In a third aspect, embodiments of this disclosure provide an electronic device, characterized in that the electronic device comprises:

[0077] At least one processor; and,

[0078] A memory that is communicatively connected to at least one processor; wherein,

[0079] The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the underwater multimodal hybrid interlayer opportunistic transport method according to any one of claims 1 to 7.

[0080] In a fourth aspect, embodiments of this disclosure provide a non-transitory computer-readable storage medium, characterized in that the non-transitory computer-readable storage medium stores computer instructions for causing the computer to perform the method of underwater multimodal hybrid interlayer opportunistic transport.

[0081] Compared with the prior art, the present invention has the following beneficial technical effects:

[0082] This invention provides a method, system, device, and medium for underwater multimodal hybrid cross-layer opportunistic transmission. Addressing the heterogeneous mode hybrid transmission requirements of underwater acoustic-optical hybrid networks, this method discloses an initial relay candidate node set and optimal relay candidate region selection method adapted to the heterogeneous characteristics of multiple transmission modes, enabling accurate evaluation of the forwarding performance of heterogeneous transmission modes. To address the need for differentiated service assurance across multiple network services, a hierarchical decision-making strategy is proposed, combining the two-level decision-making characteristics between transmission mode selection and routing parameter configuration. The upper-layer strategy selects the optimal transmission mode based on service state characteristics, while the lower-layer strategy completes fine-grained routing parameter configuration under the constraints of the selected mode. This effectively overcomes the insufficient adaptability of traditional fixed transmission modes and static routing parameters in dynamic marine environments and reduces the computational complexity of decision-making. To address the limitations of underwater resource-poor network environments, a reward update method with low-overhead state-action space traversal is adopted. This allows nodes to calculate simulated rewards corresponding to different service types and different routing parameters simultaneously, based on feedback from a single service transmission, significantly reducing the service interaction cost required for the convergence of the hierarchical decision-making strategy. This invention achieves joint optimization of intelligent transmission mode selection and fine-grained routing parameter configuration through deep collaborative acousto-optic heterogeneous links, significantly improving the adaptability and differentiated transmission requirements guarantee capability of underwater acousto-optic hybrid networks in dynamic multi-service environments. It provides a highly reliable information transmission approach for meeting the transmission needs of multiple services in multi-scenario, large-scale dynamic marine environments. Attached Figure Description

[0083] Figure 1 A flowchart illustrating a method for underwater multimodal hybrid interlayer opportunistic transport according to an embodiment of this disclosure is shown.

[0084] Figure 2 A schematic diagram of the stages of an underwater multimodal hybrid interlayer opportunistic transport method according to an embodiment of the present disclosure is shown;

[0085] Figure 3 A schematic diagram of the underwater network node deployment according to an embodiment of this disclosure is shown;

[0086] Figure 4 A schematic diagram illustrating network service classification according to an embodiment of this disclosure is shown;

[0087] Figure 5 A convergence diagram of the method proposed in the embodiments of this disclosure is shown;

[0088] Figure 6 The overall performance results of the method proposed in the embodiments of this disclosure are shown in the figure;

[0089] Figure 7 The diagram shows the service network performance results of the method proposed in the embodiments of this disclosure;

[0090] Figure 8A diagram of an apparatus for an underwater multimodal hybrid interlayer opportunistic transport method according to an embodiment of the present disclosure is shown. Detailed Implementation

[0091] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0092] This embodiment includes four stages, such as Figure 2 The diagram illustrates the network initialization phase, the service transmission phase based on relay candidate areas, the relay candidate node contention phase, and the service transmission policy update phase. The main task of the network initialization phase is for nodes to broadcast beacon information via underwater acoustic links to assist each node in the network in obtaining initial neighbor information, thereby dividing the initial relay candidate node set under different transmission modes. The main task of the service transmission phase based on relay candidate areas is for forwarding nodes to intelligently select the transmission mode and routing parameters based on the state characteristics of the service to be forwarded, and to forward the service based on the best relay candidate area under the transmission mode. The main task of the relay candidate node contention phase is for relay candidate nodes within the best relay candidate area to successfully receive the forwarded service, calculate the waiting interval based on the forwarding factor, and send a forwarding declaration (ACK) after the waiting interval expires to participate in the inter-node forwarding competition and obtain the next-hop forwarding node. The main task of the service transmission policy update phase is for the previous-hop forwarding node to update the upper and lower layer action rewards layer by layer using a low-overhead reward update method based on state-action space traversal, based on the received ACK feedback, to achieve efficient online optimization of the routing policy.

[0093] Figure 1 A flowchart 100 of a method for underwater multimodal hybrid interlayer opportunistic transport according to an embodiment of this disclosure is shown, as follows: Figure 1 As shown, it includes the following steps:

[0094] In step S101, each network node periodically broadcasts beacon information carrying its unique node ID and spatial coordinates in an underwater acoustic mode. Each network node parses and receives the beacon information to obtain the initial relay candidate node set corresponding to each transmission mode.

[0095] Specifically, each network node uses the underwater acoustic mode with the longest communication distance to periodically broadcast beacon information containing the node's unique ID and coordinates in a time-division multiple access (TDMA) manner. Figure 3 This diagram illustrates the underwater network node deployment in an embodiment of the present disclosure. Each network node includes a water-light (...) ), medium frequency underwater acoustics ( ) and high-frequency underwater acoustics ( Three transmission modes are used, and the relevant parameters for each mode are shown in Table 1. The water-optical communication mode uses LED communication with a beamwidth of [missing information]. The angle is set to 30 degrees, and each node has three fan-shaped candidate regions for water-optical relay with the destination node as the center.

[0096] Table 1. Parameter configuration of transmission mode

[0097] Transmission mode Transmission rate Communication radius Transmit power Propagation speed m / s Water Light 20Mbps 100m 10w <![CDATA[2.25×10 8 ]]> Mid-frequency underwater acoustics 6kbps 500m 2.8w 1500 High-frequency underwater acoustics 15kbps 250m 8w 1500

[0098] Furthermore, in step S101, when each network node receives beacons from adjacent nodes within the beacon communication range, it parses the beacon information of the adjacent nodes and updates the adjacent node information table. If the adjacent nodes satisfy the transmission mode... lower condition set Then it is included in the transmission mode. The initial set of candidate relay nodes, the condition set :

[0099] ;

[0100] in, Indicates transmission mode Maximum service communication radius , Represents the set of transmission modes; Indicates from the forwarding node To adjacent nodes The Euclidean distance; , This is the set of all nodes in the underwater network.

[0101] In step S102, based on the mission scenario of the underwater monitoring or detection network, the services to be forwarded are divided into three categories: latency-sensitive, bandwidth-sensitive, and periodic. Nodes in the network containing services to be forwarded are designated as forwarding nodes. Based on the initial relay candidate node set and the mode transmission angle, the forwarding nodes divide the initial relay candidate regions for each transmission mode and select the best relay candidate region for each transmission mode from the initial relay candidate regions.

[0102] Specifically Figure 4 This illustration shows a network service classification diagram in an embodiment of the present disclosure. The node side identifies and classifies the converged service flows and assigns independent queues to three types of services. The service arrival process follows a Poisson distribution, while the generation probability of different service types follows a preset normal distribution to simulate the randomness and heterogeneity of service flows in a real network. The generation probability of periodic services is 55%, bandwidth-sensitive services is 30%, and latency-sensitive services is 15%. Nodes in the network with services to be forwarded are called forwarding nodes. These nodes, based on the initial relay candidate node set and the mode transmission angle, divide the initial relay candidate region for each transmission mode and select the best relay candidate region with the strongest forwarding potential for each transmission mode.

[0103] In this embodiment, for the service to be forwarded Forwarding node First, based on the condition set With modal emission angle Divide the transmission modes The initial relay candidate region is determined, and the optimal relay candidate region in this mode is calculated. .

[0104] Specifically, when obtaining the optimal relay candidate region in step S102, the following steps are included:

[0105] Step A1: Based on the transmission mode launch angle The entire relay candidate area is divided into multiple equal parts to form a transmission mode. Initial Relay Candidate Region Set :

[0106] ;

[0107] For water-optical modes, due to their narrow beam characteristics, the regions are divided into finer segments. The omnidirectional underwater acoustic mode can be considered as a single, unified region. .

[0108] Step A2: Calculate different transmission modes respectively Each initial relay candidate node in the initial relay candidate node set within the initial relay candidate region Distance factor :

[0109] ;

[0110] In the formula, Indicates from the forwarding node To the target destination node of the service to be forwarded The Euclidean distance; Indicates starting from the initial relay candidate node To the target destination node of the service to be forwarded The Euclidean distance; it should be noted that the distance factor Effectively quantified the nodes Location advantage.

[0111] Step A3: Aggregate Transmission Mode Distance factor of the initial relay candidate node set within the initial relay candidate region Calculate each transmission mode Candidate factors for the initial relay candidate region :

[0112] ;

[0113] In the formula, Indicates a forwarding node In transmission mode Below, initial relay candidate region The initial set of candidate relay nodes within; It should be noted that candidate factors The size reflects the corresponding initial relay candidate region Its comprehensive forwarding potential.

[0114] Step A4: Forwarding Node Select transmission mode The initial relay candidate region with the largest lower candidate factor is selected as the best relay candidate region. ;

[0115] .

[0116] Next, proceed to step S103.

[0117] In step S103, the forwarding node constructs state features based on the target destination node and service classification of the service to be forwarded. Based on the return values ​​of each transmission mode action under the state features, it selects the actual transmission mode action using a "random-greedy" strategy as the upper-layer decision. Combining the state features and the upper-layer decision, it determines the actual routing parameters for service forwarding using a "random-greedy" strategy based on the return values ​​of each routing parameter action as the lower-layer decision. The forwarding node embeds the actual routing parameters corresponding to the lower-layer decision into the header of the data packet of the service to be forwarded and completes the forwarding of the service according to the actual transmission mode action determined by the upper-layer decision.

[0118] Specifically, the state features in step S103 include:

[0119] ;

[0120] In the formula, Forwarding business Destination node identifier (ID); Forwarding business Business categories;

[0121] Based on state features The state characteristics of the responses to each transmission mode action are determined using a "random-greedy" strategy. Actual transmission mode action :

[0122] ;

[0123] In the formula, For the business to be forwarded Transmission modes; The optimal relay candidate region corresponding to the transmission mode;

[0124] The reward for each routing parameter action based on state characteristics and upper-level decision-making. A "random-greedy" strategy is used to select the lower-level routing parameters. :

[0125] ;

[0126] In the formula, Indicates a forwarding node In transmission mode The corresponding candidate factor parameters in the next routing parameters, This represents the energy parameter under the same conditions. This represents the queue parameters under the same conditions. This represents the waiting interval parameter under the same conditions.

[0127] Next, proceed to step S104.

[0128] In step S104, nodes that successfully receive and forward services within the best relay candidate area are designated as relay candidate nodes. Each relay candidate node calculates its corresponding forwarding factor based on its best relay candidate area under each transmission mode, its remaining energy, and its queue length. The waiting interval is determined based on the forwarding factor. After the waiting interval ends, the relay candidate node sends a forwarding declaration to participate in the forwarding competition. The relay candidate node that is qualified to forward services in the forwarding competition is designated as the next-hop forwarding node.

[0129] Specifically, the forwarding factor in step S104 is used to evaluate successor candidate nodes. Its own forwarding potential, forwarding factor include:

[0130] ;

[0131] In the formula, Indicates a forwarding node The transmission mode used for forwarding; Indicates a forwarding node relay candidate nodes In transmission mode Distance factor below; It represents the shortest communication radius among the transmission modes of each node in the network, and is used to balance the differences in communication range between heterogeneous links; This represents the initial energy of each node in the network. This indicates the maximum queue length for each node in the network. Indicates a relay candidate node The remaining energy, Indicates a relay candidate node The length of the queue; , and The routing parameters used in the calculation of the forwarding factor all take values ​​in the range [0,1], and The smaller the forwarding factor value, the stronger the relay candidate node. The stronger the overall forwarding potential.

[0132] The waiting interval include:

[0133] ;

[0134] In the formula, This indicates the fixed communication radius for sending forwarding claims (ACKs). Twice the radius of the maximum transmission mode communication, Furthermore, its queuing priority is higher than that of business data packets. This indicates the propagation speed of the forwarding statement. ; For transmission mode The communication radius; For from node To neighbor nodes The Euclidean distance; For transmission mode The speed of propagation; it should be noted that the stronger the relay candidate node's forwarding potential... waiting interval The smaller the value, the earlier the forwarding statement is sent.

[0135] The process of competing for forwarding during the waiting interval includes the following steps:

[0136] If the relay candidate node If the target node is a relay candidate node, then... ID and forwarding services The target node ID is consistent, and the relay candidate node is... In receiving and forwarding services Immediately afterwards, a forwarding declaration is sent. Upon receiving the forwarding declaration, the remaining relay candidate nodes relinquish their forwarding competition and delete any forwarding services they have already received. ;

[0137] If relay candidate node If the node is not the target node, then it is a relay candidate node. The startup duration is a timer corresponding to the waiting interval. If a forwarding service is received before the timer expires... Forwarding statement, relay candidate node Abandon forwarding competition and remove forwarding services from the queue. If no forwarding statement is received by the timer expires, the relay candidate node... Send with its own forwarding factor The forwarded statement, and wait If no forwarding declaration with a smaller forwarding factor value is received during the waiting period, the forwarding contention relay candidate node... Capable of forwarding services If a node is selected for forwarding, it will be used as the next-hop forwarding node to continuously execute steps S2-S5; otherwise, it will abandon the forwarding competition and remove the forwarding service from the queue. The core of the forwarding competition is that each candidate relay node that successfully receives the forwarding service sends a forwarding declaration based on its own calculated waiting interval, and confirms the final next-hop forwarding node by comparing the forwarding factors between nodes.

[0138] Next, proceed to step S105.

[0139] In step S105, the forwarding node calculates the actual reward for the actions actually performed by the upper-layer and lower-layer decisions based on the received forwarding declaration, and calculates the simulated reward corresponding to different service types and routing parameter configuration combinations under the same transmission mode using a reward update method based on state-action space traversal. Based on the actual reward and simulated reward, the reward Q value of each action of the lower-layer and upper-layer decisions is updated sequentially. If the service is successfully forwarded to the target destination node, the forwarding of the forwarding service is completed; otherwise, steps S102-S105 are executed repeatedly until the service is successfully forwarded to the target destination node.

[0140] Specifically, in step S105, when the actual execution reward and simulated reward of the upper-level decision are obtained, the actual reward is... :

[0141] ;

[0142] In the formula, This indicates that the upper layer uses transport mode actions. The penalty for not receiving a forwarding statement is used to suppress the selection of unreliable transmission modes;

[0143] In the low-overhead reward update method based on state-action space traversal, it is assumed that the forwarding declaration feedback result of the simulated action is consistent with the transmission modal action. To maintain consistency, calculate the simulated rewards for other service categories under the same target node and the same transmission mode action, based on the upper-level decision. State features In the formula, Classify as other business categories, and meet the following requirements: and , A set of service categories within the network;

[0144] When obtaining the actual and simulated rewards for lower-level decisions, the following steps are included:

[0145] The actual execution of routing parameters in the lower-level decision-making process The actual reward is :

[0146] ;

[0147] In the formula, This represents the relative cost per hop of receiving a forwarding declaration. This indicates the relative cost of a single hop when no forwarding statement is received; This indicates a penalty for not receiving a forwarding declaration after the lower layer uses the actual routing parameters; This is a reward for a forwarding node that does not receive a forwarding statement. The smaller the value, the higher the corresponding reward; The reward for a forwarding node when it receives a forwarding statement. The smaller the value, the higher the reward; the reward formula is designed to encourage forwarding nodes to choose actions with relatively lower single-hop costs in order to improve transmission efficiency.

[0148] ;

[0149] ;

[0150] In the formula, This represents the actual single-hop delay. This represents the simulated single-hop delay. and During calculation, self-forwarding nodes are included. Forwarding service Transmission, propagation, and queuing delays; ; In the formula, Indicates forwarding service The length of latency-sensitive services and periodic services is 128 bytes.

[0151] This represents the actual single-hop energy consumption. This represents the virtual single-hop energy consumption. and The calculation includes provisions for forwarding services. Sending and forwarding statements to report energy consumption; , In the formula, Indicates the energy consumption of sending and forwarding the declaration. Indicates the number of received forwarded statement feedbacks. The larger the value, the higher the relative cost per hop, and the lower the reward for the forwarding node to receive the forwarding declaration, in order to encourage the node to choose actions that are not redundant with the forwarding declaration.

[0152] Indicates the reference transmission mode The actual time delay per unit distance, Indicates the reference transmission mode Actual energy consumption per unit distance; Indicates a forwarding node With the following local best nodes The Euclidean distance between the nodes, where the local best node is located. It is a forwarding node In transmission mode Below, we have received requests for forwarding services. The actual set of relay candidate nodes corresponding to the forwarding statement In the context of the node with the smallest forwarding factor, it should be noted that, considering the unknown actual single-hop communication distance when no forwarding declaration is received, the communication radius of the baseline mode is used. replace; Indicates forwarding service The weighting of single-hop latency is used to reflect the differentiated requirements of different services for single-hop latency and energy consumption.

[0153] During the state-action space traversal, for the lower-level action space In addition to the actual actions performed, other routing parameter actions are also included. (satisfy and Based on the local best node State calculation simulation reward The calculation ideally assumes a forwarding declaration feedback number of 1. It calculates the simulated rewards for actions with different routing parameters under the same state characteristics and transmission mode actions. include:

[0154] ;

[0155] Other service classifications under the same destination node and the same transmission mode operation ,satisfy and Other business categories Simulated rewards for all corresponding lower-level routing parameter actions ,include:

[0156] ;

[0157] In the formula, In representing forwarding nodes Below, transmission mode Corresponding relay candidate area The system already contains the set of actual relay candidate nodes corresponding to the forwarding declarations received. It should be noted that due to the complex and dynamic marine environment and the actions taken by forwarding nodes, redundancy may exist in the actual forwarding declarations sent by each relay candidate node, resulting in multiple forwarding declarations being received by each node. To avoid this problem and improve transmission efficiency, forwarding nodes need to select the relay candidate node with the strongest forwarding potential from the actual set of relay candidate nodes based on the forwarding factor as the local best node. .

[0158] Furthermore, in step S105, the rewards for each action of the lower-level decision are updated as follows: ,include:

[0159] ;

[0160] In the formula, This represents the learning rate of lower-level decisions. ; This represents the discount factor for lower-level decisions. For other business categories The update logic is consistent with that described above for updating the rewards of lower-level decision actions;

[0161] When the rewards for each action in the upper-level decision are updated, the maximum reward is extracted from the updated reward value of the lower-level decision and used as the reward for the current best action. And based on the best action reward Update the rewards for each action in the upper-level decision-making process. ,include:

[0162] ;

[0163] ;

[0164] In the formula, This represents the learning rate of upper-level decisions. ; This represents the discount factor for upper-level decisions. ; Indicates the local best node The value of.

[0165] To verify the effectiveness of the method proposed in this embodiment, underwater sensor networks of different sizes were constructed using OPNET simulation software for testing. The simulation aimed to simulate the impact of different network sizes and dynamic network conditions on the performance of the proposed method; therefore, 500m networks were set for each network. 500m, 750m 750m, 1000m 1000m, 1250m In three underwater network spaces at a depth of 1250m, 90 underwater sensor nodes were randomly distributed in each space, including one source node and one destination node. The link packet loss rate in the network was used as a characterization parameter to reflect the degree of network state change. Figure 5 It shows 500m Convergence diagrams at different packet loss rates in a 500m network space. As can be seen from the diagrams, the method proposed in this embodiment can achieve effective convergence under varying degrees of network state changes, proving that the proposed method can stably learn and adapt to the network environment. Furthermore, the convergence speed and the average reward value after convergence differ depending on the severity of the network state changes, indicating that the method proposed in this embodiment is applicable to different dynamic environments.

[0166] Figure 6 The results show that in four different sizes of underwater network space, the packet loss rate is randomly distributed between 0 and 0.3, and the initial remaining energy is randomly distributed between 70% and 100% of the initial energy. Figure 6 (a) is a comparison chart of data packet delivery rates. Figure 6 (b) is a comparison chart of energy efficiency. Figure 6 (c) is a comparison chart of average end-to-end delay. Figure 6 (d) is a network throughput comparison chart. The network performance of the proposed method in this embodiment is compared with that of existing single-mode opportunistic routing. The comparison indicators include network energy efficiency, end-to-end delivery rate, end-to-end latency, and network throughput. Simulation results show that as the network scale increases, the network node density decreases, the relay candidate node set of each method decreases, leading to an increase in the number of end-to-end hops, a decrease in the end-to-end transmission success rate of services, and a downward trend in total energy consumption. Compared with pure underwater acoustic opportunistic routing, the method proposed in this embodiment, based on upper-layer and lower-layer decisions, can assist forwarding nodes in selecting the current optimal link mode and opportunistic routing parameters, thereby effectively achieving high-reliability, low-latency transmission of multi-service data. In large-scale network scenarios, the method proposed in this embodiment can dynamically utilize the wide range of underwater acoustic links and the low-latency advantages of underwater optical links, significantly improving the transmission reliability of acoustic-optical hybrid networks and expanding their application boundaries.

[0167] Figure 7 It shows a specific 750m Under the conditions of a 750m network scale, a packet loss rate randomly distributed between 0 and 0.3%, and initial remaining energy randomly distributed between 70% and 100% of the initial energy, this embodiment compares the network performance of the proposed method with that of existing single-mode opportunistic routing for transmitting different services. Figure 7 (a) is a comparison chart of average end-to-end latency for different services. Figure 7 (b) is a comparison chart of data packet delivery rates for different services. Figure 7 (c) is a comparison chart of network throughput for different services. The method proposed in this embodiment achieves the lowest end-to-end latency for latency-sensitive services. Although the utilization rate of the water-optical mode is higher for bandwidth-sensitive services, its latency is improved by 66.74% compared to latency-sensitive services. This is because the latency weight in the reward function of bandwidth-sensitive services is lower than that of latency-sensitive services, resulting in lower routing parameters (such as waiting interval parameters) even when using the same water-optical mode. The standard protocol, designed to balance energy consumption, exhibits a higher delivery rate when transmitting bandwidth-sensitive services due to its longer transmission time, which significantly increases the probability of transmission conflicts with other services or forwarding claims. The method proposed in this embodiment reduces such conflicts by equipping different services with differentiated transmission modes and routing parameter selections, thus maintaining a high delivery rate. Since the standard protocol does not employ differentiated transmission strategies for different services, the throughput difference between latency-sensitive and periodic services is not significant, and the high throughput of bandwidth-sensitive services mainly stems from the large data packet size of the service itself. The method proposed in this embodiment, with its superior latency and delivery rate, achieves throughput exceeding that of the standard protocol for all three types of services. Furthermore, the differences in throughput between services demonstrate the protocol's intelligent scheduling based on differentiated transmission requirements.

[0168] A second embodiment of the present invention also provides a system for underwater multimodal hybrid interlayer opportunistic transport, comprising:

[0169] The initialization unit is configured so that each network node periodically broadcasts beacon information carrying the node's unique ID and spatial coordinates in an underwater acoustic mode. Each network node parses and receives the beacon information to obtain the initial relay candidate node set corresponding to each transmission mode.

[0170] The service classification and forwarding unit is configured to classify the services to be forwarded into three categories—latency-sensitive, bandwidth-sensitive, and periodic—in combination with the underwater mission scenario. Nodes in the network containing the services to be forwarded are designated as forwarding nodes. Based on the initial relay candidate node set and the mode transmission angle, the forwarding nodes divide the initial relay candidate regions for each transmission mode and select the best relay candidate region with the strongest forwarding potential under each transmission mode from the initial relay candidate regions.

[0171] The action decision unit is configured such that the forwarding node constructs state features based on the target destination node and service classification of the service to be forwarded. Based on the reward value of each transmission mode action under the state features, it selects the actual transmission mode action using a "random-greedy" strategy as the upper-layer decision. Combining the state features and the upper-layer decision, it determines the actual routing parameters for service forwarding using a "random-greedy" strategy based on the reward value of each routing parameter action as the lower-layer decision. The forwarding node embeds the actual routing parameters corresponding to the lower-layer decision into the header of the data packet of the service to be forwarded and completes the forwarding of the service according to the actual transmission mode action determined by the upper-layer decision.

[0172] The forwarding contention unit is configured to use nodes that have successfully received forwarded services within the best relay candidate area as relay candidate nodes. Each relay candidate node calculates a corresponding forwarding factor based on its best relay candidate area under each transmission mode, its remaining energy, and queue length, and determines a waiting interval based on the forwarding factor. After the waiting interval expires, the relay candidate node sends a forwarding declaration to participate in the forwarding contention. The relay candidate node that is qualified to forward services in the forwarding contention becomes the next-hop forwarding node.

[0173] The update unit is configured so that the forwarding node calculates the actual reward for the actual actions of the upper-layer decision and the lower-layer decision based on the received forwarding declaration, and calculates the simulated reward corresponding to different service types and routing parameter configuration combinations under the same transmission mode through the reward update method of state-action space traversal. Based on the actual reward and simulated reward, the reward Q value of each action of the lower-layer decision and the upper-layer decision is updated in sequence. If the service is successfully forwarded to the target destination node, the forwarding of the service is completed. Otherwise, the service classification and forwarding unit, the action decision unit, the forwarding contention unit and the update unit are executed in a loop until the service is successfully forwarded to the target destination node.

[0174] The third embodiment of the present invention also provides an electronic device, the electronic device comprising:

[0175] At least one processor;

[0176] A memory that is communicatively connected to at least one processor; wherein,

[0177] The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the underwater multimodal hybrid interlayer opportunistic transport method according to any one of claims 1 to 7.

[0178] The fourth embodiment of the present invention also provides a non-transitory computer-readable storage medium storing computer instructions for causing the computer to perform the underwater multimodal hybrid interlayer opportunistic transport method described in any of the foregoing embodiments.

[0179] The fifth embodiment of the present invention also provides a computer program product, which includes a computing program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions that, when executed by a computer, cause the computer to perform the underwater multimodal hybrid interlayer opportunistic transport method of any of the foregoing embodiments.

[0180] Figure 8 The illustration shows a method or device 1000 implementing an embodiment of the present invention. In some embodiments, more or fewer devices may be included than illustrated. In some embodiments, it may be implemented using a single or multiple devices. In some embodiments, it may be implemented using cloud-based or distributed devices.

[0181] like Figure 8 As shown, device 1000 includes a processor 1001 for performing various appropriate operations and processes based on programs and / or data stored in read-only memory (ROM) 1002 or programs and / or data loaded from storage portion 1008 into random access memory (RAM) 1003. Processor 1001 may be a multi-core processor or may contain multiple processors. In some embodiments, processor 1001 may include a general-purpose main processor and one or more special coprocessors, such as a central processing unit (CPU), graphics processing unit (GPU), neural network processor (NPU), digital signal processor (DSP), etc. Various programs and data required for the operation of device 1000 are also stored in RAM 1003. Processor 1001, ROM 1002, and RAM 1003 are interconnected via bus 1004. Input / output (I / O) interface 1005 is also connected to bus 1004.

[0182] The processor and memory described above are used together to execute programs stored in the memory. When the program is executed by a computer, it can implement the methods, steps, or functions described in the above embodiments.

[0183] The following components are connected to I / O interface 1005: an input section 1006 including a keyboard, mouse, touchscreen, etc.; an output section 1007 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 1008 including a hard disk, etc.; and a communication section 1009 including a network interface card such as a LAN card, modem, etc. The communication section 1009 performs communication processing via a network such as the Internet. A drive 1010 is also connected to I / O interface 1005 as needed. A removable medium 1011, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 1010 as needed so that computer programs read from it can be installed into storage section 1008 as needed. Figure 8 The diagram only shows a portion of the components and does not imply that the device 1000 only includes... Figure 8 The components shown.

[0184] The systems, devices, modules, or units described in the above embodiments can be implemented by a computer or its associated components. The computer may be, for example, a mobile terminal, smartphone, personal computer, laptop computer, in-vehicle human-machine interface device, personal digital assistant, media player, navigation device, game console, tablet computer, wearable device, smart TV, Internet of Things system, smart home, industrial computer, server, or a combination thereof.

[0185] Although not shown, in this embodiment of the invention, a computer-readable storage medium is provided having a computer program / instruction stored thereon, which, when executed by a processor, implements the underwater multimodal hybrid interlayer opportunistic transport method described in the embodiment.

[0186] Storage media in embodiments of the present invention include articles that are permanent and non-permanent, removable and non-removable, capable of storing information by any method or technology. Examples of storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0187] Although not shown, embodiments of the present invention also provide a computer program product, including: a computer program / instructions that, when executed by a processor, implement the underwater multimodal hybrid interlayer opportunistic transport method described in the embodiments.

[0188] The methods, programs, systems, apparatuses, etc., in embodiments of the present invention can be executed or implemented in one or more networked computers, or practiced in a distributed computing environment. In the embodiments of this specification, in these distributed computing environments, tasks can be performed by remote processing devices connected via a communication network.

Claims

1. A method for underwater multimodal hybrid interlayer opportunistic transport, characterized in that, Includes the following steps: Step S1: Each network node periodically broadcasts beacon information carrying the node's unique ID and spatial coordinates in an underwater acoustic mode. Each network node parses and receives the beacon information to obtain the initial set of candidate relay nodes corresponding to each transmission mode. Step S2: Based on the underwater mission scenario, the services to be forwarded are divided into three categories: latency-sensitive, bandwidth-sensitive, and periodic. Nodes in the network containing services to be forwarded are used as forwarding nodes. Based on the initial relay candidate node set and the mode transmission angle, the forwarding nodes divide the initial relay candidate regions for each transmission mode and select the best relay candidate region for each transmission mode from the initial relay candidate regions. Step S3: The forwarding node constructs state features based on the target destination node and service classification of the service to be forwarded. Based on the return values ​​of each transmission mode action under the state features, it selects the actual transmission mode action using a "random-greedy" strategy as the upper-layer decision. Combining the state features and the upper-layer decision, and based on the return values ​​of each routing parameter action, it determines the actual routing parameters for service forwarding using a "random-greedy" strategy as the lower-layer decision. The forwarding node embeds the actual routing parameters corresponding to the lower-layer decision into the header of the data packet of the service to be forwarded, and completes the forwarding of the service to be forwarded according to the actual transmission mode action determined by the upper-layer decision. Step S4: Nodes that successfully receive and forward services within the best relay candidate area are designated as relay candidate nodes. Each relay candidate node calculates its corresponding forwarding factor based on its best relay candidate area, remaining node energy, and queue length for each transmission mode, and determines the waiting interval based on the forwarding factor. After the waiting interval expires, the relay candidate node sends a forwarding declaration to participate in forwarding competition. The relay candidate node that is qualified to forward services in the forwarding competition is designated as the next-hop forwarding node. Step S5: Based on the received forwarding declaration, the forwarding node calculates the actual reward for the actions actually performed by the upper-layer and lower-layer decisions, and calculates the simulated reward corresponding to different service types and routing parameter configuration combinations under the same transmission mode using the reward update method of state-action space traversal. Based on the actual reward and simulated reward, the reward Q value of each action of the lower-layer and upper-layer decisions is updated sequentially. If the service is successfully forwarded to the target destination node, the forwarding of the forwarding service is completed; otherwise, steps S2-S5 are executed repeatedly until the service is successfully forwarded to the target destination node.

2. The method for underwater multimodal hybrid interlayer opportunistic transport according to claim 1, characterized in that, In step S1, when each network node parses and receives beacon information, it records the information of adjacent nodes within the beacon communication radius. If the adjacent nodes satisfy the transmission mode... lower condition set Then it is included in the transmission mode. The initial set of candidate relay nodes, the condition set include: ； In the formula, Indicates transmission mode Maximum service communication radius , Represents the set of transmission modes; Indicates from the forwarding node To adjacent nodes The Euclidean distance; , This is the set of all nodes in the underwater network.

3. The method for underwater multimodal hybrid interlayer opportunistic transport according to claim 1, characterized in that, When obtaining the optimal relay candidate region in step S2, the following steps are included: Step A1: Based on the transmission mode launch angle The entire relay candidate area is divided into multiple equal parts to form a transmission mode. Initial Relay Candidate Region Set : ； Step A2: Calculate different transmission modes respectively Each initial relay candidate node in the initial relay candidate node set within the initial relay candidate region Distance factor : ； In the formula, Indicates from the forwarding node To the target destination node of the service to be forwarded The Euclidean distance; Indicates starting from the initial relay candidate node To the target destination node of the service to be forwarded The Euclidean distance; Step A3: Aggregate Transmission Mode Distance factor of the initial relay candidate node set within the initial relay candidate region The transmission modes are obtained. The following are the initial relay candidate regions candidate factors : ； In the formula, Indicates a forwarding node In transmission mode Initial relay candidate region The initial set of candidate relay nodes within; , Indicates transmission mode The initial set of relay candidate regions is determined; candidate factor values ​​are used to describe the region's forwarding potential. Step A4: Forwarding Node Select transmission mode The initial relay candidate region with the largest lower candidate factor is selected as the optimal relay candidate region for the corresponding transmission mode. ; 。 4. The underwater multimodal hybrid interlayer opportunistic transport method according to claim 1, characterized in that, The state features in step S3 include: ； In the formula, Forwarding business The target node identifier ID; Forwarding business Business categories; Based on state features The state characteristics of the responses to each transmission mode action are determined using a "random-greedy" strategy. Actual transmission mode action : ； In the formula, For the business to be forwarded Transmission modes; The optimal relay candidate region corresponding to the transmission mode; The reward for each routing parameter action based on state characteristics and upper-level decision-making. The actual routing parameters are obtained using a "random-greedy" strategy. : ； In the formula, Indicates a forwarding node In transmission mode The corresponding candidate factor parameters in the next routing parameters, This represents the energy parameter under the same conditions. This represents the queue parameters under the same conditions. This indicates the corresponding waiting interval under the same conditions.

5. The underwater multimodal hybrid interlayer opportunistic transport method according to claim 3, characterized in that, Forwarding factor in step S4 include: ； In the formula, Indicates a forwarding node The transmission mode used for forwarding; Indicates a forwarding node relay candidate nodes In transmission mode Distance factor below; It represents the shortest communication radius among the transmission modes of each node in the network, and is used to balance the differences in communication range between heterogeneous links; This indicates the number of transmission modes carried by each node in the network; This represents the initial energy of each node in the network; This indicates the maximum queue length for each node in the network. Indicates a relay candidate node The remaining energy, Indicates a relay candidate node The length of the queue; Candidate factors for the best relay candidate region; The waiting interval include: ； In the formula, This indicates the fixed communication radius for sending and forwarding ACK statements. It has twice the maximum transmission mode communication radius and a higher queuing priority than service data packets. This indicates the propagation speed of the forwarding statement; For transmission mode The communication radius; For the forwarding node To relay candidate node The Euclidean distance; For transmission mode The speed of propagation; The process of competing for forwarding during the waiting interval includes the following steps: If the relay candidate node If the target node is a relay candidate node, then... ID and forwarding services The target node ID is consistent, and the relay candidate node is... In receiving and forwarding services After sending a forwarding declaration, the remaining relay candidate nodes, upon receiving the forwarding declaration, relinquish their forwarding competition and delete any forwarding services they have already received. ; If relay candidate node If the node is not the target node, then it is a relay candidate node. The startup duration is a timer corresponding to the waiting interval. If a forwarding service is received before the timer expires... Forwarding statement, relay candidate node Abandon forwarding competition and remove forwarding services from the queue. If no forwarding statement is received by the timer expires, the relay candidate node... Send with its own forwarding factor The forwarded statement, and wait If no forwarding statement with a smaller forwarding factor value is received during the waiting period, the relay candidate node will be considered a relay node. Capable of forwarding services If a node is selected for forwarding, it will be used as the next-hop forwarding node to continuously execute steps S2-S5; otherwise, it will abandon the forwarding competition and remove the forwarding service from the queue. .

6. The method for underwater multimodal hybrid interlayer opportunistic transport according to claim 1, characterized in that, In step S5, when the actual reward and simulated reward of the upper-level decision are obtained, the actual reward is: : ； In the formula, This indicates that the upper-level decision-making uses the actual transmission mode action. The penalty for not receiving a forwarding statement is used to suppress the selection of unreliable transmission modes; In the state-action space traversal reward update method, it is assumed that the forwarding declaration feedback result of the simulated action is consistent with the transmission modal action. To maintain consistency, calculate the simulated rewards for other service categories under the same target node and the same transmission mode action, based on the upper-level decision. State features In the formula, Classify as other business categories, and meet the following requirements: and , A set of service categories within the network; When obtaining the actual and simulated rewards for lower-level decisions, the following steps are included: The actual routing parameter actions of the lower-level decision-making The actual reward is : ； In the formula, This represents the relative cost per hop of receiving a forwarding declaration. This indicates the relative cost of a single hop when no forwarding statement is received; This indicates a penalty for not receiving a forwarding declaration after the lower layer uses the actual routing parameters; This is a reward for a forwarding node that does not receive a forwarding statement. The smaller the value, the higher the corresponding reward; The reward for a forwarding node when it receives a forwarding statement. The smaller the value, the higher the corresponding reward; ； ； In the formula, This represents the actual single-hop delay. This represents the simulated single-hop delay. and During calculation, self-forwarding nodes are included. Forwarding service Transmission, propagation, and queuing delays; This represents the actual single-hop energy consumption. This represents the virtual single-hop energy consumption. and The calculation includes provisions for forwarding services. The energy consumption of sending and forwarding statements; Indicates the reference transmission mode The actual time delay per unit distance, Indicates the reference transmission mode Actual energy consumption per unit distance; Indicates a forwarding node With the best local node The Euclidean distance between the nodes, where the local best node is located. It is a forwarding node In transmission mode Next, regarding forwarding services The actual set of relay candidate nodes corresponding to the forwarding declaration The node with the smallest forwarding factor; if no forwarding declaration is received, the communication radius of the reference mode is used. replace; Indicates forwarding service The weighting of single-hop latency is used to reflect the differentiated requirements of different services for single-hop latency and energy consumption; In the state-action space traversal reward update method, for the lower-level action space... In addition to actual routing parameter actions Other routing parameter actions ,satisfy ,and Based on the local best node State calculation simulation reward The calculation ideally assumes a forwarding declaration feedback number of 1. Under the same state characteristics and the same transmission mode action, the simulated reward corresponds to different routing parameter actions. include: ； Other service classifications under the same destination node and the same transmission mode operation ,satisfy and Other business categories Simulated rewards for all corresponding lower-level routing parameter actions ,include: ； In the formula, Indicates a forwarding node In transmission mode Below, corresponding relay candidate area The set of actual relay candidate nodes corresponding to the received forwarding declaration.

7. The method for underwater multimodal hybrid interlayer opportunistic transport according to claim 6, characterized in that, In step S5, the rewards for each action of the lower-level decision-making are updated. ,include: ； In the formula, This represents the learning rate of lower-level decisions. This represents the discount factor for lower-level decisions; When the rewards for each action in the upper-level decision are updated, the maximum reward is extracted from the updated reward value of the lower-level decision and used as the reward for the current best action. And based on the best action reward Update the rewards for each action in the upper-level decision-making process. ,include: ； ； In the formula, This represents the learning rate of upper-level decision-making. This represents the discount factor for upper-level decisions. Indicates the local best node The value of.

8. A system for underwater multimodal hybrid interlayer opportunistic transport, characterized in that, The method for underwater multimodal hybrid interlayer opportunistic transport based on any one of claims 1-7 includes: The initialization unit is configured so that each network node periodically broadcasts beacon information carrying the node's unique ID and spatial coordinates in an underwater acoustic mode. Each network node parses and receives the beacon information to obtain the initial relay candidate node set corresponding to each transmission mode. The service classification and forwarding unit is configured to classify the services to be forwarded into three categories: latency-sensitive, bandwidth-sensitive, and periodic, based on the underwater mission scenario. Nodes in the network containing the services to be forwarded are used as forwarding nodes. The forwarding nodes divide the initial relay candidate regions for each transmission mode based on the initial relay candidate node set and the mode transmission angle, and select the best relay candidate region for each transmission mode from the initial relay candidate regions. The action decision unit is configured such that the forwarding node constructs state features based on the target destination node and service classification of the service to be forwarded. Based on the reward value of each transmission mode action under the state features, it selects the actual transmission mode action using a "random-greedy" strategy as the upper-layer decision. Combining the state features and the upper-layer decision, it determines the actual routing parameters for service forwarding using a "random-greedy" strategy based on the reward value of each routing parameter action as the lower-layer decision. The forwarding node embeds the actual routing parameters corresponding to the lower-layer decision into the header of the data packet of the service to be forwarded and completes the forwarding of the service according to the actual transmission mode action determined by the upper-layer decision. The forwarding contention unit is configured to use nodes that have successfully received forwarded services within the best relay candidate area as relay candidate nodes. Each relay candidate node calculates a corresponding forwarding factor based on its best relay candidate area under each transmission mode, its remaining energy, and queue length, and determines a waiting interval based on the forwarding factor. After the waiting interval expires, the relay candidate node sends a forwarding declaration to participate in the forwarding contention. The relay candidate node that is qualified to forward services in the forwarding contention becomes the next-hop forwarding node. The update unit is configured so that the forwarding node calculates the actual reward for the actual actions of the upper-layer decision and the lower-layer decision based on the received forwarding declaration, and calculates the simulated reward corresponding to different service types and routing parameter configuration combinations under the same transmission mode through the reward update method of state-action space traversal. Based on the actual reward and simulated reward, the reward Q value of each action of the lower-layer decision and the upper-layer decision is updated in sequence. If the service is successfully forwarded to the target destination node, the forwarding of the service is completed. Otherwise, the service classification and forwarding unit, the action decision unit, the forwarding contention unit and the update unit are executed in a loop until the service is successfully forwarded to the target destination node.

9. An electronic device, characterized in that, The electronic device includes: At least one processor; and, A memory that is communicatively connected to at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the underwater multimodal hybrid interlayer opportunistic transport method according to any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium, characterized in that, The non-transitory computer-readable storage medium stores computer instructions for causing the computer to perform the underwater multimodal hybrid interlayer opportunistic transport method as described in any one of claims 1 to 7.

Images