Service-driven underwater wireless sensor network composite routing agent

Abstract

The invention relates to a service-driven underwater wireless sensor network composite routing agent, which comprises a routing function module and a plugging platform, and is characterized in that the routing function module comprises a reinforcement learning routing component and a plurality of functional components; the routing component obtains an optimal transmission path according to an optimization target in a limited range during data transmission; the plurality of functional components are used for supporting data transmission requirements of different services; the plug-in platform generates a service QoS field according to the service demand, wherein the service QoS field has a mapping relationship with the plurality of functional components; and the routing function module selects and matches at least one function component from a plurality of function components according to the mapping relationship and the value of the service QoS field, and is used for correspondingly processing data and / or a next hop node set during data acquisition or data receiving or data sending. The function component is configured based on service requirements, and the forwarding capability can be dynamically output according to services in the underwater wireless sensor network.

Description

Technical Field

[0001] The invention relates to the technical field of network routing, and in particular to a service-driven underwater wireless sensor network composite routing intelligent body. Background Art

[0002] Underwater wireless sensor networks are widely used in many fields such as underwater disaster warning, pollutant monitoring, hydrological data monitoring, marine resource exploration, and auxiliary navigation. They are also important infrastructure for studying, building, and developing the ocean. Underwater wireless sensor networks are composed of sensor nodes, communication nodes, and aggregation nodes. Data is collected by underwater sensor nodes, forwarded hop by hop through communication nodes located at different depths, and transmitted to surface aggregation nodes, which are then forwarded to shore-based data centers by aggregation nodes. Underwater transmission channels have problems such as large transmission delay, limited transmission bandwidth, many interference factors, and serious multipath phenomena. Underwater communication nodes are affected by water flow, and their positions and node relationships change dynamically. The communication energy consumption is large and energy supply is difficult. The underwater transmission environment is open, and communication nodes are vulnerable to attacks by malicious nodes. However, there are many underwater monitoring applications, and the data transmission requirements vary greatly. This makes it extremely difficult for underwater wireless sensor networks to achieve efficient, reliable, and secure data communication.

[0003] Routing algorithms are the key to achieving efficient, reliable and secure communications for all types of services. Underwater wireless sensor network routing algorithms are used to construct transmission routes from underwater data collection nodes to surface data aggregation nodes. The core is to select "good" relay nodes. Routing algorithms should not only pay attention to whether the channel bandwidth between nodes meets business requirements, whether the transmission direction points to the target node, whether the relay node has sufficient energy, and whether there are more neighbor nodes to facilitate data forwarding, but also consider the topological changes caused by node movement, changes in the hydrological environment, and the impact of external noise on underwater wireless communications. At the same time, it is also necessary to meet the service quality requirements of different types of business data communications as much as possible, ensure the security, robustness, and reliability of routing, reduce data transmission delays, and improve data transmission success rates.

[0004] In order to meet the communication needs of different services, routing algorithms suitable for different application scenarios are proposed.

[0005] In order to meet the needs of high-throughput business data transmission, a variety of routing algorithms that can cope with congestion, adapt to different priorities, and have edge prediction and compression have been proposed. RCAR proposed by Jin et al. can achieve a method to avoid congestion of relay nodes when the business traffic is large. Through the design of reinforcement learning routing algorithm, the transmission delay and energy consumption distribution in underwater data communication are optimized. PB-ACR proposed by Chen et al. can distinguish the priorities of different applications. High-priority business applications choose the shortest route, and low-priority business applications choose routes with more remaining energy. ACOR proposed by Singhal et al. supports adaptive adjustment of data sampling frequency based on data relevance to reduce transmission energy consumption. TBDP proposed by Wang et al. is a two-level bidirectional prediction model suitable for 3D underwater wireless sensor networks with AUVs, which relies on AUVs to carry data for communication. For the needs of sequential data transmission, EP-ADTA proposed by Wang et al. adaptively adjusts the transmission accuracy of time-series monitoring data based on the transmission environment while selecting the transmission route. When the transmission environment is good and the transmission route can provide sufficient transmission bandwidth, high-precision monitoring data is transmitted. When the transmission environment is bad and the transmission route cannot provide sufficient bandwidth, the monitoring data transmission accuracy is appropriately reduced to prioritize the transmission of feature data to the surface aggregation node. In response to the needs of image data transmission, VRC-ADTA proposed by Wang et al. can decompose the transmitted image data into average value data and multi-level detail coefficient data, and realize the transmission of image data in a progressive transmission manner.

[0006] In order to meet the needs of highly reliable business data transmission, routing algorithms based on deterministic rules, reinforcement learning, multi-agent reinforcement learning and cooperative communication have been proposed. VBF proposed by Xie et al. establishes a virtual pipe on the vector between the source node and the sink node, limits the set of candidate forwarding nodes by controlling the radius of the virtual pipe, and selects the best relay node according to the distance between the node and the forwarding vector. QELAR proposed by Hu et al. is a routing algorithm based on Q-learning. When planning the route, it not only selects the shortest transmission path, but also comprehensively considers the remaining energy of the receiving node to avoid the problem of premature exhaustion of node energy on the optimal path. The DMARL algorithm proposed by Li et al. is a distributed multi-agent reinforcement learning routing algorithm that comprehensively considers factors such as remaining energy, link stability and data transmission quality to select routes for underwater wireless optical networks. The SA-FRL algorithm proposed by Zhang et al. realizes cooperative communication of underwater networks based on the Q-learning algorithm. SA-FRL defines the feedback channel transmission state and system mutual information as the state of the reinforcement learning algorithm, and defines the system mutual information and access delay as rewards. It selects cooperative nodes with good link quality and low access delay to improve the efficiency of cooperative communication. SQMCR proposed by Wang et al. selects next-hop nodes and cooperative communication nodes based on reinforcement learning and game theory methods, improving the reliability of data forwarding in multi-hop underwater wireless sensor networks.

[0007] In order to meet the needs of high-security business data transmission, routing algorithms based on encryption, authentication, beam control and trust evaluation have been proposed. SEEORVA proposed by Varun et al. uses lightweight encryption technology to encrypt the transmitted data. SecFUN proposed by Giuseppe et al. uses AES and short digital signature algorithms in Galois counter mode as encryption blocks to ensure the confidentiality, integrity, authentication and non-repudiation of forwarded data. SARP proposed by Manjula uses the direction of arrival of the signal to authenticate the security of neighboring nodes. ST-CJ proposed by Su et al. defines the security coverage of the network and uses the long propagation delay of underwater acoustic signals to make the interference signal based on cooperation conflict at the eavesdropper without affecting the reception of legitimate users. Yang et al. proposed a trusted routing based on blockchain and reinforcement learning. The blockchain manages the credibility of nodes and makes the routing information traceable. The intrusion detection model DOIDS for opportunistic routing proposed by Zhang et al. selects the energy consumption, forwarding and link quality information of candidate nodes as detection feature values, detects potential abnormal nodes through the DBSCAN clustering algorithm, and reduces the false detection rate. The GTR proposed by Wang et al. realizes trust evaluation of neighbor nodes based on GAN and performs routing selection based on reinforcement learning, which improves the security of data forwarding in multi-hop underwater wireless sensor networks.

[0008] However, underwater wireless sensor networks often need to support multiple services to achieve the expected construction goals. Routing algorithms that simply meet the data transmission requirements of specific services cannot meet the application requirements of underwater wireless sensor networks. Therefore, it is necessary to design a set of composite routing algorithms that can meet the data transmission requirements of different services.

[0009] In the existing underwater wireless sensor network routing algorithms, specific business data transmission requirements are generally met by building a variety of typical routing algorithm components, which lays the foundation for building a composite routing algorithm.

[0010] The EP-ADTA routing algorithm is based on edge prediction and reinforcement learning algorithms. It mainly provides transmission services for high-throughput sequence data in underwater wireless sensor networks. It can automatically adjust the data transmission accuracy according to the changes in the transmission channel conditions, support progressive transmission, and ensure that the monitoring sequence data that meets the business needs is efficiently transmitted to the destination. The EP-ADTA routing algorithm includes edge prediction function module, virtual pipeline function module and reinforcement learning routing function module. The edge prediction function module is used to realize the prediction and compression of sequence data. Based on the Arma algorithm, the prediction of new data is realized according to the previous data, prediction parameters and correction data. Under certain transmission accuracy requirements, the transmission of prediction parameters and correction data will be more efficient than the transmission of the original new data. The virtual pipeline function module is used to limit the selection of the next hop node to ensure that the shorter route is preferred and the transmission delay is reduced. The virtual pipeline function module can also dynamically adjust the pipeline radius according to the change of the remaining energy to ensure the balance of node energy consumption. The reinforcement learning routing function module realizes the selection of the next hop node and data accuracy based on the Q-learning algorithm. The optimization goals include reducing the number of hops, balancing energy consumption, reducing delay and improving transmission accuracy.

[0011] The VRC-ADTA routing algorithm is based on edge compression and reinforcement learning algorithms. It mainly provides transmission services for high-throughput image data in underwater wireless sensor networks. It can customize data transmission accuracy according to transmission channel conditions, support progressive transmission, and ensure that monitoring image data that meets business needs is efficiently transmitted to the destination. The VRC-ADTA routing algorithm includes an edge compression function module and a reinforcement learning routing function module. The edge compression function module is used to compress image data. Based on the Haar algorithm, the image data is decomposed into average value data and multi-level detail coefficient data. According to the transmission channel conditions and business requirements, only the average value data and certain levels of detail coefficient data can be transmitted, or progressive transmission can be used. The reinforcement learning routing function module selects the next hop node based on the Q-learning algorithm. The optimization goals include reducing the number of hops, balancing energy consumption, and reducing latency.

[0012] The SQMCR routing algorithm is based on cooperative communication and reinforcement learning algorithms. It mainly provides highly reliable data transmission services for underwater wireless sensor networks. It can select data transmission paths and cooperative communication nodes according to changes in transmission channels and the status of neighboring nodes, and can ensure reliable data transmission in multi-hop networks. The SQMCR algorithm includes cooperative communication function modules and reinforcement learning routing function modules. The reinforcement learning routing function module selects the next hop node based on the Q-learning algorithm. The optimization goals include reducing the number of hops, balancing energy consumption, reducing latency, and improving network topology robustness. The cooperative communication function module selects nodes and timings to participate in cooperative communication based on the selected transmission path based on the Stackelberg game theory and the Q-learning algorithm.

[0013] The GTR routing algorithm is implemented based on trust evaluation and reinforcement learning algorithms. It mainly provides data transmission services for high-security services in underwater wireless sensor networks. It can identify malicious nodes, avoid receiving data from malicious nodes or passing data to malicious nodes, and ensure that data is transmitted between trusted nodes. The GTR routing algorithm includes a trust evaluation function module and a reinforcement learning routing function module. The trust evaluation function module is used to identify malicious nodes among neighboring nodes. It identifies malicious nodes based on the GAN algorithm and the trust feature profile of trusted nodes, and can dynamically adjust the trust feature profile according to changes in the transmission channel to ensure the detection performance of malicious nodes. The reinforcement learning routing function module implements the selection of the next hop node based on the Q-learning algorithm. The optimization goals include reducing the number of hops, balancing energy consumption, and improving the robustness of the network topology.

[0014] The above routing algorithms can achieve good performance in meeting the data transmission requirements of specific services in underwater wireless sensor networks, but cannot meet the requirements of comprehensive service transmission. The routing algorithm encapsulates functional modules in a component manner, which can greatly improve its reusability. Based on business demand drive, the combined application of multiple components can meet the comprehensive service data transmission requirements of underwater wireless sensor networks.

[0015] The methods described in this section are not necessarily methods that have been previously conceived or employed. Unless otherwise indicated, it should not be assumed that any method described in this section is considered to be prior art simply because it is included in this section. Similarly, unless otherwise indicated, the issues mentioned in this section should not be considered to have been recognized in any prior art. Summary of the invention

[0016] The present invention aims at the technical problems existing in the prior art and provides a service-driven underwater wireless sensor network composite routing intelligent agent.

[0017] The technical solution of the present invention to solve the above technical problems is as follows:

[0018] A business-driven underwater wireless sensor network composite routing agent, comprising a routing function module and a plug-in platform;

[0019] The routing function module includes a routing component and multiple functional components; the routing component is used to obtain the best transmission path within a limited range according to the optimization target when sending data; the multiple functional components are used to support the data transmission requirements of different services;

[0020] The plug-in platform includes a database reading and writing module, a data packet processing module, an information transceiver module, a business demand library, an environment status library, a transmission strategy library, and a business database;

[0021] The database reading and writing module is used to read the business requirements from the business requirements library, and generate a business QoS field according to the business requirements, and the business QoS field has a mapping relationship with the multiple functional components; the routing function module selects at least one functional component from the multiple functional components according to the mapping relationship and the value of the business QoS field, and is used to perform corresponding processing on the data and / or the next hop node set when collecting data, receiving data, or sending data;

[0022] The data packet processing module is used to read and write data packet content during data acquisition, data reception or data transmission; the data packets include control data packets and business data packets;

[0023] The information transceiver module is used to receive or send data packets based on the hydroacoustic signal;

[0024] The business requirement library is used to store business types and business requirements related to the business types;

[0025] The environment status library is used to store status information of the current node, neighboring nodes and transmission channels;

[0026] The transmission strategy library is used to store the routing components and related parameters of multiple functional components in the routing function module;

[0027] The service database is used to store the service data queues that have been sent and the service data queues of different priorities to be sent.

[0028] Further, the functional components include: a virtual pipeline component, an edge prediction component, an edge compression component, a collaborative communication component, and a trust assessment component;

[0029] The virtual pipe component is used to limit the range of the next hop node set so that the transmission path is close to the best transmission path between the sending source node and the destination node;

[0030] The edge prediction component is used to compress the transmission amount required for sequence data;

[0031] The edge compression component is used to compress the transmission amount required for the image data;

[0032] The collaborative communication component is used to obtain nodes and timings participating in collaborative communication according to the transmission path and optimization target;

[0033] The trust evaluation component is used to evaluate whether a neighbor node is a malicious node.

[0034] Furthermore, the length of the service QoS field is 1 byte, including priority, delay, accuracy, throughput, reliability, security and service type fields; wherein,

[0035] The priority field occupies 2 bits and is used to represent 4 priorities, 00 is the lowest priority, 01 is low priority, 10 is high priority, and 11 is the highest priority;

[0036] The latency occupies 1 bit and is used to represent the service's demand for data forwarding latency. 0 represents normal latency and 1 represents low latency.

[0037] The precision occupies 1 bit and is used to represent the service's demand for data forwarding precision. 0 represents normal precision and 1 represents high precision.

[0038] Throughput occupies 1 bit and is used to represent the throughput requirement of business transmission data. 0 represents normal throughput and 1 represents high throughput.

[0039] Reliability occupies 1 bit and is used to characterize the service's demand for data forwarding reliability. 0 represents normal reliability and 1 represents high reliability.

[0040] Security occupies 1 bit and is used to represent the business's demand for data forwarding security. 0 represents normal security and 1 represents high security.

[0041] The service type occupies 1 bit and is used to characterize the classification of service data. 0 represents sequence data and 1 represents image data.

[0042] Further, the routing function module selects at least one functional component from a plurality of functional components according to the mapping relationship and the value of the service QoS field, including:

[0043] When the delay is low, the virtual pipeline component is configured; when the delay is normal, the composite routing is not configured with the virtual pipeline component;

[0044] When the service type is sequence data, when the throughput is high throughput and the accuracy is normal accuracy, the edge prediction component is configured to transmit the prediction parameters and the correction data that meets the accuracy requirements in direct transmission mode; when the throughput is high throughput and the accuracy is high accuracy, the edge prediction component is configured to transmit the prediction parameters and the correction data that meets the accuracy requirements in progressive transmission mode; when the throughput is normal throughput and the accuracy is normal accuracy, the edge prediction component is configured to transmit the prediction parameters and the correction data that meets the accuracy requirements in direct transmission mode; when the throughput is normal throughput and the accuracy is high accuracy, the edge prediction component is not configured;

[0045] When the service type is image data, when the throughput is high throughput and the precision is normal precision, the edge compression component is configured to transmit the average value data and some detail coefficient data that meet the precision requirements in progressive transmission mode; when the throughput is high throughput and the precision is high precision, the edge compression component is configured to transmit the average value data and all detail coefficient data in progressive transmission mode; when the throughput is normal throughput and the precision is normal precision, the edge compression component is configured to transmit the average value data and some detail coefficient data in progressive transmission mode; when the throughput is normal throughput and the precision is high precision, the edge compression component is not configured;

[0046] When the reliability is high reliability, the collaborative communication component is configured; when the reliability is normal reliability, the collaborative communication component is not configured;

[0047] When the security is normal, the trust evaluation component is configured according to the security of the transmission environment.

[0048] Furthermore, during data collection: the database read-write module reads the business requirements from the business requirements library, and generates a business QoS field based on the business requirements; the routing function module selects an edge prediction component or an edge compression component based on the values ​​of the throughput, accuracy and business type fields in the business QoS field; the database read-write module reads corresponding parameter data from the transmission policy library; the data packet processing module generates a business data packet using the data output by the edge prediction component or the edge compression component or the original collected data and the parameter data; the database read-write module writes the business data packet into the corresponding business data queue in the business database according to the priority in the business QoS field.

[0049] Furthermore, when data is sent: the data packet processing module reads the business data packet to be sent from the business database according to the priority order and the first-in-first-out principle; the routing function module selects the trust assessment component and the virtual pipe component according to the security and delay fields in the business QoS field in the business data packet to be sent; the plug-in platform selects the next hop node from the next hop node set through the routing component, the trust assessment component and the virtual pipe component; the data packet processing module modifies the receiving node in the business data packet to the next hop node selected by the routing function module; the information transceiver module sends the data packet to be sent modified by the data packet processing module according to the corresponding transmission mode.

[0050] Furthermore, when data is received: the information transceiver module receives the business data packet and sends it to the data packet processing module; the data packet processing module reads the business data packet header and obtains the business QoS field; the routing function module selects functional components according to the business QoS field; the plug-in platform uses the selected functional components to perform corresponding processing on the business data packet, including discarding the data packet, saving the data, or using the database read-write module to write the business data packet into the forwarding queue of the business database.

[0051] Furthermore, the plug-in platform utilizes the optional functional components to perform corresponding processing on the business data packet, and also includes: if a collaborative communication component is optional, the plug-in platform utilizes the collaborative communication component to determine whether the current composite routing intelligent entity participates in collaborative communication, and if so, writes the node ID of the current composite routing intelligent entity into the collaborative communication ID in the header of the business data packet, and the plug-in platform writes the business data packet into the highest priority queue of the business database through the database read-write module.

[0052] Furthermore, the control data packet is forwarded in a broadcast manner to synchronize the status information of the underwater communication node and the neighbor node information, including: node ID, node location, remaining energy, and the number of in-degree nodes, out-degree nodes and malicious nodes and their IDs. When a trust evaluation component is configured, the in-degree nodes and out-degree nodes become trusted in-degree nodes and trusted out-degree nodes.

[0053] Furthermore, the business data packet is forwarded in a point-to-point manner to forward business data, including: address information related to data sending and receiving, data sending information related to the routing component, and information related to the business data. When the edge prediction component is configured, the data content transmitted also includes prediction parameters, correction data sequence number and content, and correction data quantity.

[0054] The beneficial effects of the present invention are: 1. A composite routing agent is constructed based on a component approach, which can meet the needs of comprehensive business data transmission in underwater wireless sensor networks.

[0055] Underwater wireless sensor networks are widely used and usually have the need for comprehensive business data transmission. However, existing underwater wireless sensor network routing algorithms are often designed based on specific business needs, and it is difficult to meet the data transmission needs of comprehensive services. Based on the research on existing routing algorithms, the present invention decomposes the typical routing algorithm applied to specific services in a component manner, and then constructs a composite routing algorithm with components as units, designs the business processes between components, and forms a composite routing algorithm that can adapt to the data transmission of different services at the same time. The composite routing intelligent body built based on the composite routing algorithm is compatible with the data transmission of different services by selecting different functional components, so that the underwater wireless sensor network can carry comprehensive services.

[0056] 2. Based on the business configuration composite routing algorithm component, the composite routing agent can dynamically output forwarding capabilities according to the business in the underwater wireless sensor network.

[0057] The goal of routing algorithm design is to support services and provide the data forwarding capabilities they need. However, existing underwater wireless sensor network routing algorithms focus more on changes in transmission channels and lack adaptation to changes in business data transmission requirements. Based on the research on the main services of underwater wireless sensor networks, the present invention forms a classification of quality of service (QoS) required for typical business data transmission, designs a QoS field for characterizing business data transmission requirements, and then defines the configuration of composite routing algorithm components mapped by different QoS fields to drive the dynamic output of composite routing algorithm data forwarding capabilities. Based on the configuration of the composite routing algorithm components driven by the business QoS field, the underwater wireless sensor network can dynamically adjust the output of data forwarding capabilities according to different business requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 A structural diagram of an underwater wireless sensor network composite routing agent model provided by an embodiment of the present invention;

[0059] Figure 2 A schematic diagram of the service QoS field structure provided in an embodiment of the present invention;

[0060] Figure 3 A data packet structure diagram provided for an embodiment of the present invention;

[0061] Figure 4 A schematic diagram of a data collection process of a composite routing agent as an underwater sensor node provided by an embodiment of the present invention;

[0062] Figure 5A schematic diagram of a data transmission process of a composite routing agent as an underwater sensor node or a communication node provided in an embodiment of the present invention;

[0063] Figure 6 A schematic diagram of a data receiving process of a composite routing agent as an underwater communication node provided by an embodiment of the present invention;

[0064] Figure 7 A schematic diagram of the structure of an underwater wireless sensor network routing protocol simulation experiment system provided by an embodiment of the present invention;

[0065] Figure 8 is the packet delivery rate of sequence data in different test scenarios;

[0066] Fig. 9 It is the data packet delivery rate of image data in different scenarios;

[0067] Fig.10 is the average transmission delay of sequence data in different test scenarios;

[0068] Fig.11 is the average transmission delay of image data in different test scenarios. DETAILED DESCRIPTION

[0069] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0070] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, "plurality" means two or more, unless otherwise clearly and specifically defined.

[0071] In the description of the present application, the term "for example" is used to mean "used as an example, illustration or description". Any embodiment described as "for example" in the present application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid unnecessary details to obscure the description of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present application.

[0072] The embodiment of the present invention provides an underwater wireless sensor network composite routing agent, such as Figure 1 As shown, it includes a routing function module and a plug-in platform. The routing function module is used to realize routing selection and adapt to various business needs. The plug-in platform is used to access various routing function components and provide them with communication, data packet processing, database and other functions.

[0073] The routing function module is obtained by decomposing and encapsulating four underwater wireless sensor network routing algorithms, namely EP-ADTA, VRC-ADTA, SQMCR and GTR, and includes multiple functional components, such as routing components and virtual pipeline components, edge prediction components, edge compression components, collaborative communication components, and trust evaluation components. It supports data transmission requirements of sequence data monitoring services and image data monitoring services with different latency requirements, reliability requirements, and security requirements, and with different throughputs.

[0074] Among them, the routing component is a reinforcement learning routing component obtained by decomposing, integrating and encapsulating the reinforcement learning routing module functions of EP-ADTA, VRC-ADTA, SQMCR and GTR routing algorithms. The core algorithm of the reinforcement learning routing component is the Q-learning algorithm. The input is a set of candidate next-hop nodes, and the output is the selected next-hop node. The optimization goals include reducing the number of hops, balancing energy consumption, reducing latency and improving network topology robustness. The optimization strategy is obtained based on the iteration of the Bellman equation. The Q value initialization and dynamic exploration probability optimization method defined in the SQMCR algorithm are used to improve the convergence speed of the reinforcement learning routing component. The function of the reinforcement learning routing component is to obtain the best transmission path within a limited range according to the optimization goal. The reinforcement learning routing component is a required component of the composite routing algorithm.

[0075] It should be understood here that although the reinforcement learning routing component is used in this embodiment, those skilled in the art can also replace the core algorithm of the routing component with other routing algorithms of the same type according to actual needs.

[0076] The virtual pipeline component is obtained by decomposing and encapsulating the virtual pipeline functional module of the EP-ADTA routing algorithm. The input of the virtual pipeline component is the neighbor node set of the current node where the data packet is located, and the output is the candidate next-hop node set. The virtual pipeline component is established based on the vector between the sending source node and the destination node. The pipeline radius sets the initial value based on the underwater network and business requirements, and expands as the average residual energy of the neighboring nodes decreases. The function of the virtual pipeline component is to limit the range of the next-hop node set, so that the transmission path is closer to the optimal transmission path between the sending source node and the destination node, reducing the transmission delay. The virtual pipeline component is an optional component of the composite routing algorithm and is selected based on the delay requirements of business data transmission.

[0077] The edge prediction component is obtained by decomposing and encapsulating the edge prediction function model of the EP-ADTA routing algorithm. The input of the edge prediction component is sequence data and accuracy requirements, and the output is prediction parameters and correction data. The edge prediction component is implemented based on the Arma algorithm and can be configured in two modes: direct transmission and progressive transmission. Direct transmission means packaging the prediction parameters and correction data together and transmitting them simultaneously with the same priority. Progressive transmission means packaging the prediction parameters and correction data with different accuracy separately and transmitting them in sequence with different priorities. The function of the edge prediction component is to compress the transmission volume required for sequence data and improve the throughput of the underwater wireless sensor network carrying sequence data. The edge prediction component is an optional component of the composite routing algorithm and is selected based on the throughput requirements of sequence data transmission in the business.

[0078] The edge compression component is obtained by decomposing and encapsulating the edge compression function model of the VRC-ADTA routing algorithm. The input of the edge compression component is the image data and accuracy requirements, and the output is the average value data and multi-level detail coefficient data. The edge compression component is implemented based on the Haar algorithm and adopts a progressive transmission mode. During data transmission, the average value data is transmitted first, and then the multi-level detail coefficient data is transmitted with different priorities. The function of the edge compression component is to compress the transmission volume required for image data and improve the throughput of the underwater wireless sensor network carrying the atlas data. The edge compression component is an optional component of the composite routing algorithm and is selected according to the throughput requirements of image data transmission in the business.

[0079] The collaborative communication component is obtained by decomposing and encapsulating the collaborative communication function module of the SQMCR routing algorithm. The core algorithm of the collaborative communication component is the Stackelberg Q-learning algorithm. The input is the candidate collaborative communication node set, and the output is the collaborative communication node. The collaborative communication node is further determined based on the transmission path selected by the sending node. The optimization goal is to improve the overall data delivery rate and reduce the communication energy consumption. The optimization strategy is obtained based on the iteration of the Bellman equation. The function of the collaborative communication component is to obtain the nodes and timing participating in the collaborative communication based on the transmission path and optimization goal. The collaborative communication component is an optional component of the composite routing algorithm and is selected based on the reliability requirements of the business data.

[0080] The trust assessment component is obtained by decomposing and encapsulating the trust assessment function model of the GTR routing algorithm. The core algorithm of the trust assessment component is the GAN algorithm, with the input being the trust feature attributes of the neighboring node to be evaluated and the output being the credibility of the neighboring node. The trust assessment component generates a trust feature profile of a trusted node by regularly collecting the trust feature attributes of the transmission node, channel, and data, which is used to compare with the neighboring node to be evaluated to determine whether the neighboring node is a malicious node. The trust assessment component is an optional component of the composite routing algorithm and is selected based on the security requirements of the transmission environment and business data.

[0081] The plug-in platform includes a database reading and writing module, a data packet processing module, an information sending and receiving module, a business demand library, an environmental status library, a transmission strategy library, and a business database. It is responsible for communicating with other communication nodes, sending and receiving data packets, and storing the control information and business information of the data packets in the corresponding database, supporting various functional components to complete their functions.

[0082] The business requirements library stores business types and business requirements for data transmission priority, latency, throughput, security, and reliability.

[0083] The environment state library stores the state information of the current node, neighboring nodes and transmission channels, including the node's location information, remaining energy, out-degree nodes and in-degree nodes, packet error rate of data forwarding, channel transmission delay, packet loss rate, etc.

[0084] The transmission strategy library stores the Q table of the reinforcement learning routing component, the thresholds and hyperparameters of each functional component, including the initial radius of the virtual pipeline, the accuracy setting of the edge prediction and edge compression components, the benefit-cost adjustment parameters of the collaborative communication component, and the malicious node threshold and model degradation threshold of the trust evaluation component.

[0085] The business database stores the sent business data queues and the business data queues of different priorities to be sent.

[0086] The database read-write module is used to realize the communication and management functions between external components or modules and various databases.

[0087] The data packet processing module is used to process business data packets and control data packets, and read and write data packet contents.

[0088] The signal transceiver module receives and sends various data packets based on underwater acoustic signals to achieve communication with various communication nodes.

[0089] The database read-write module reads the business requirements from the business requirements library and generates a business QoS field according to the business requirements; the routing function module selects at least one functional component from multiple functional components according to the value of the business QoS field, and performs corresponding processing on the data and / or the next-hop node set during data collection, data reception or data sending.

[0090] The service QoS field is used to characterize the demand of service data for data transmission. Its length is 1 byte and includes priority, delay, accuracy, throughput, reliability, security and service type fields, such as Figure 2 As shown. Among them,

[0091] Priority (Pri 2bit): used to represent 4 priorities, where 00 is the lowest priority, 01 is low priority, 10 is high priority, and 11 is the highest priority. Data to be forwarded with different priorities are queued separately, and data in the high priority queue is sent first. Data in the same priority queue is sent according to the first-in, first-out principle.

[0092] Delay (D 1bit): used to characterize the service's demand for data forwarding delay, where 0 represents normal delay and 1 represents low delay. Delay requirements are only compared among services of the same priority.

[0093] Precision (P 1bit): used to characterize the service's demand for data forwarding precision, where 0 represents normal precision and 1 represents high precision. The precision requirements vary for different service types.

[0094] Throughput (T 1bit): used to characterize the throughput requirement of service transmission data, where 0 represents normal throughput and 1 represents high throughput.

[0095] Reliability (R 1 bit): used to characterize the service's demand for data forwarding reliability, where 0 represents normal reliability and 1 represents high reliability.

[0096] Security (S 1 bit): used to characterize the service's requirement for data forwarding security, where 0 represents normal security and 1 represents high security.

[0097] Service type (Type 1bit): used to characterize the classification of service data, where 0 represents sequence data and 1 represents image data.

[0098] Business-driven means selecting functional components in the routing function module based on the business QoS fields generated by business requirements to achieve multi-hop forwarding in underwater wireless sensor networks.

[0099] The intelligent agent as a communication node of the underwater wireless sensor network will select the functional components that should be configured for the composite routing algorithm based on the service QoS field in the forwarded data packet and the mapping relationship between the service QoS and the functional components. That is, at least one functional component is selected from multiple functional components according to the value of the service QoS field. The mapping relationship between the service QoS field and the functional component is shown in Table 1.

[0100] Table 1 Mapping relationship between service QoS and functional components

[0101]

[0102] Set up queues for data to be forwarded according to different priorities, and give priority to forwarding high-priority data. For service data of the same priority, configure functional components based on other fields of service QoS.

[0103] When the delay (D) is low delay (1), the virtual pipeline component is configured; when the delay (D) is normal delay (0), the virtual pipeline component is not configured.

[0104] When the service type (Type) is sequence data (0), when the throughput (T) is high throughput (1) and the precision (P) is normal precision (0), the edge prediction component is configured to transmit the prediction parameters and the correction data that meets the precision requirements in direct transmission mode. When the throughput (T) is high throughput (1) and the precision (P) is high precision (1), the edge prediction component is configured to transmit the prediction parameters and the correction data that meets the precision requirements in progressive transmission mode. When the throughput (T) is normal throughput (0) and the precision (P) is normal precision (0), the edge prediction component is configured to transmit the prediction parameters and the correction data that meets the precision requirements in direct transmission mode. When the throughput (T) is normal throughput (0) and the precision (P) is high precision (1), the edge prediction component is not configured.

[0105] When the service type (Type) is image data (1), when the throughput (T) is high throughput (1) and the precision (P) is normal precision (0), the edge compression component is configured to transmit the average value data and some detail coefficient data that meet the precision requirements in progressive transmission mode. When the throughput (T) is high throughput (1) and the precision (P) is high precision (1), the edge compression component is configured to transmit the average value data and all detail coefficient data in progressive transmission mode. When the throughput (T) is normal throughput (0) and the precision (P) is normal precision (0), the edge compression component is configured to transmit the average value data and some detail coefficient data in progressive transmission mode. When the throughput (T) is normal throughput (0) and the precision (P) is high precision (1), the edge compression component is not configured.

[0106] When the reliability (R) is high reliability (1), the cooperative communication component is configured; when the reliability (R) is normal reliability (0), the cooperative communication component is not configured.

[0107] When security (S) is high security (1), the transmission environment should also be high security, and the trust assessment component is configured; when security (S) is normal security (0), the trust assessment component is configured according to the security of the transmission environment.

[0108] The composite routing agent synchronizes the node status of the underwater wireless sensor network through data packets to realize business data forwarding. Its data packet structure is as follows: Figure 3 As shown. Among them, Figure 3 - (a) represents a service data packet structure used to forward service data, Figure 3 -(b) indicates the data content in the service data packet after the edge prediction component is configured, Figure 3 - (c) represents the structure of a control data packet used for node status synchronization. The control data packet and the service data packet are forwarded in a time-division manner on the control channel and the service channel respectively.

[0109] 1. Control data packet design

[0110] Control data packets are forwarded in a broadcast manner, mainly used to synchronize the status information of underwater communication nodes and neighbor node information, including: node ID, node location, remaining energy, as well as in-degree, out-degree and malicious node numbers and their IDs. When a trust evaluation component is configured, the in-degree and out-degree nodes become trusted in-degree and out-degree nodes.

[0111] 2. Business data package design

[0112] Business data packets are forwarded in a point-to-point manner, mainly used to forward business data, including: data sending and receiving related address information (sending node, receiving node, collaborative communication node and target node ID), data sending information related to the reinforcement learning routing component (V value, timestamp and number of transmissions), and information related to business data (data ID, business QoS and data content). When the edge prediction component is configured, the data content transmitted includes prediction parameters, correction data sequence number and content, and correction data quantity.

[0113] In the underwater wireless sensor network composed of the above-mentioned composite routing agent, the data forwarding process is implemented by each underwater communication node and underwater sensor node based on the composite routing agent Agent distributed in the network.

[0114] During data collection

[0115] In the initialization phase of the underwater wireless sensor network, each underwater communication node initializes the service requirements, environmental status, transmission strategy, service data database and component configuration of the agent-based underwater communication node and the composite routing algorithm according to the initial conditions of the transmission environment and the initial requirements of the transmission service.

[0116] The database read-write module reads the business requirements from the business requirements library and generates a business QoS field according to the business requirements; the routing function module selects the edge prediction component or the edge compression component according to the throughput, accuracy and business type field values ​​in the business QoS field. The database read-write module reads the corresponding parameter data from the transmission strategy library; the data packet processing module generates a business data packet using the data output by the edge prediction component or the edge compression component or the original collected data and the parameter data. If the business type is image data, the edge compression component and the corresponding transmission mode and the detail coefficient data that need to be transmitted are selected according to the throughput and accuracy. If the business type is sequence data, the edge prediction component and the corresponding transmission mode and the correction data that meets the accuracy requirements are selected according to the throughput and accuracy. The database read-write module writes the business data packet into the corresponding business data queue in the business database according to the priority in the business QoS field. The data acquisition process of underwater sensor nodes is as follows. Figure 4 shown.

[0117] When sending data

[0118] When the channel for sending the service data packet arrives, the underwater communication node in the data forwarding state periodically sends the service data packet according to the priority order and the first-in-first-out principle. Before sending the data packet, first check the security and delay of the service QoS field of the data packet, and confirm the configuration of the trust evaluation component and the virtual pipe component. The reinforcement learning routing component first filters the neighbor nodes located on the vector from the neighbor node to the target node as the candidate next hop node set. According to the security field, if the trust evaluation component is configured, the next hop node set becomes a trusted candidate next hop neighbor node set after being filtered by the trust evaluation component. According to the delay field, if the virtual pipe component is configured, the next hop node set is filtered by the virtual pipe component as a candidate next hop neighbor node set located in the virtual pipe. The reinforcement learning routing component calculates the Q value of each candidate next hop node based on the final candidate next hop node set and its state, based on the Q-learning algorithm and the historical Q table, and the node with the largest Q value is used as the next hop node, that is, the receiving node. Modify the ID of the receiving node in the data packet to the next hop node ID, and send the data packet. When the channel for sending control data packets arrives, the underwater communication node in the data forwarding state periodically broadcasts control data packets to the outside according to the status information of the database to achieve state synchronization between neighboring nodes. Figure 5 shown.

[0119] When receiving data

[0120] During the operation phase of the underwater wireless sensor network, the underwater communication node in the data forwarding state will constantly listen to whether there are new data packets on its transmission channel. If a new data packet arrives, it will first detect whether it is a business data packet or a control data packet. If it is a control data packet, the data packet content is read and the database content related to the neighbor sending node in the underwater communication node is updated. If the trust evaluation module is configured, the credibility of the node needs to be evaluated and the node status is marked in the database. If it is a business data packet, the data packet header is read. Check the business QoS field of the business data packet, check the type, priority, accuracy, throughput and security of the business data, and configure the required functional components for the composite routing algorithm according to the business QoS. If the trust evaluation module is configured, detect whether it comes from a malicious node based on the sending node ID in the data packet. If the data packet comes from a malicious node, it will be discarded. If the data packet comes from a trusted node, the database content related to the node is updated according to the data packet header information. Check whether the receiving node ID in the header is consistent with this node. If it is consistent, check whether the data is response data. If it is response data, identify the forwarded data according to the response result. If it is not response data, then check whether the target node ID is consistent with this node. If not, extract the content of the data packet, read the priority in the service data QoS, and store it in the forwarding queue of the corresponding priority according to the data priority; if consistent, extract the content in the data packet and save it according to the service type and service ID. After the complete data is collected, check the service data type. If it is image data, check whether the edge compression module is configured. If configured, restore the transmitted image data according to the Haar algorithm. If not configured, directly restore the image data; if it is sequence data, check whether the edge prediction module is configured. If configured, restore the transmitted sequence data according to the Arma algorithm. If not configured, directly restore the sequence data. Check whether the receiving node ID in the packet header is consistent with this node. If not, check the service QoS field of the service data packet, check the reliability field of the service data, and confirm whether the collaborative communication component is configured. If the collaborative communication component is configured, you need to choose whether to participate in the collaborative communication according to the collaborative communication component. If you participate in the collaborative communication, write the ID of this node into the collaborative communication ID and put the data packet into the forwarding data with the highest priority. If you do not participate in the collaborative communication, the data packet will be discarded; if the collaborative communication component is not configured, the data packet will be discarded. The data receiving process of the underwater communication node is as follows Figure 6 shown.

[0121] Figure 7 The invention provides an underwater wireless sensor network routing protocol simulation experiment system implemented by the underwater wireless sensor network composite routing agent based on the above-mentioned embodiment of the present invention.

[0122] The "Underwater Wireless Sensor Network Routing Protocol Simulation Experiment System" is a simulation experiment system for underwater wireless sensor networks based on Python. It can simulate various transmission environments and various business applications of underwater wireless sensor networks. It is suitable for different network scales and supports the collection, storage, analysis and management of information in the entire process of data packet forwarding. It can access various routing algorithms to realize the testing and evaluation of the data forwarding performance of routing algorithms.

[0123] The Python programming language is used to implement the service-driven underwater wireless sensor network composite routing algorithm proposed in the embodiment of the present invention, and the relevant functional components of the routing algorithms such as EP-ADTA, VRC-ADTA, SQMCR and GTR are deployed together in the "Underwater Wireless Sensor Network Routing Protocol Simulation Experiment System". The "Underwater Wireless Sensor Network Routing Protocol Simulation Experiment System" simulates and tests the performance of data forwarding of underwater wireless sensor network communication nodes under different transmission channels, different network scales and mixed transmission conditions of multiple business data. As a routing algorithm resident in underwater communication nodes, the service-driven composite routing algorithm can effectively schedule various functional components according to changes in business needs to ensure that the QoS requirements of business applications are met.

[0124] In this embodiment, the data packet delivery rate and average transmission delay of 150KBytes of data from the underwater sensor node to the surface aggregation node are tested under different network scales, business traffic, node position dynamic range, channel interruption probability and different malicious node proportions. The sequence data is an application based on underwater temperature monitoring, and the data comes from the KEO station in the NOAA database. One temperature data occupies 4 bytes. The image data is a single-channel picture of (300, 500) to reflect the situation of underwater fish. One pixel data occupies 1 byte. The size of the data content encapsulation of each data packet does not exceed 100 bytes. The three-dimensional underwater area is set to 500m×500m×500m. The speed of sound is 1500m / s and the bandwidth is 10kHz. The communication distance and perception distance of the communication node are set to 150m. The MAC layer protocol is implemented using S-FAMA. The parameters of each functional component are based on the parameters specified in the paper. The accuracy of the edge prediction of the sequence data is controlled at 0.1°C, and the image data is edge compressed using the "average value data + 1st and 2nd level detail coefficients" method. The data forwarding performance test scenarios include 9, which are set according to Table 2.

[0125] Table 2 Data performance test scenario configuration table

[0126]

[0127] The packet delivery rate refers to the ratio of the number of packets received by the surface node to the number of packets sent by the underwater sensor node. A higher packet delivery rate reflects that the routing algorithm can overcome the impact of dynamic transmission channels and changing business traffic, as well as the interference of malicious nodes, and use fewer packets to complete the transmission of business data. Figure 8 and Fig. 9 It reflects the packet delivery rate of sequence data and image data in the underwater wireless sensor network configured with the composite routing algorithm under different test scenarios. Figure 8 and Fig. 9 It is not difficult to find that the routing algorithm proposed by this technology can adapt to different business requirements and transmission channels in different test scenarios, and realize reliable, efficient and secure data packet transmission.

[0128] The average transmission delay refers to the ratio of the time required to transmit all data packets to the number of data packets. Shorter transmission delay reflects that the routing algorithm can overcome the impact of dynamic transmission channels and changing business traffic, as well as interference from malicious nodes, and take up less transmission time to complete the transmission of business data. Fig.10 and Fig.11 It reflects the average transmission delay of sequence data and image data in the underwater wireless sensor network configured with the composite routing algorithm under different test scenarios. Fig.10 and Fig.11 It is not difficult to find that the routing algorithm proposed by this technology can adapt to different business requirements and transmission channels in different test scenarios, and realize efficient, reliable and secure data packet transmission.

[0129] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be performed in parallel, sequentially or in a different order, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document is not limited here.

[0130] Although the embodiments or examples of the present disclosure have been described with reference to the accompanying drawings, it should be understood that the above-mentioned methods, systems and devices are merely exemplary embodiments or examples, and the scope of the present invention is not limited by these embodiments or examples, but only by the claims after authorization and their equivalent scope. Various elements in the embodiments or examples can be omitted or replaced by their equivalent elements. In addition, each step can be performed in an order different from that described in the present disclosure. Further, the various elements in the embodiments or examples can be combined in various ways. It is important that with the evolution of technology, many elements described herein can be replaced by equivalent elements that appear after the present disclosure.

Claims

1. A business-driven underwater wireless sensor network composite routing agent, characterized in that: Including routing function modules and plug-in platforms, The routing function module includes a routing component and multiple functional components; the routing component is used to obtain the best transmission path within a limited range according to the optimization target when sending data; the multiple functional components are used to support the data transmission requirements of different services; The plug-in platform includes a database reading and writing module, a data packet processing module, an information transceiver module, a business demand library, an environment status library, a transmission strategy library, and a business database; The database reading and writing module is used to read the business requirements from the business requirements library, and generate a business QoS field according to the business requirements, and the business QoS field has a mapping relationship with the multiple functional components; the routing function module selects at least one functional component from the multiple functional components according to the mapping relationship and the value of the business QoS field, and is used to perform corresponding processing on the data and / or the next hop node set when collecting data, receiving data, or sending data; The data packet processing module is used to read and write data packet content during data acquisition, data reception or data transmission; the data packets include control data packets and business data packets; The information transceiver module is used to receive or send data packets based on the hydroacoustic signal; The business requirement library is used to store business types and business requirements related to the business types; The environment status library is used to store status information of the current node, neighboring nodes and transmission channels; The transmission strategy library is used to store the routing components and related parameters of multiple functional components in the routing function module; The service database is used to store the service data queues that have been sent and the service data queues of different priorities to be sent.

2. The composite routing agent according to claim 1, characterized in that: The functional components include: a virtual pipeline component, an edge prediction component, an edge compression component, a collaborative communication component, and a trust assessment component; The virtual pipe component is used to limit the range of the next hop node set so that the transmission path is close to the best transmission path between the sending source node and the destination node; The edge prediction component is used to compress the transmission amount required for sequence data; The edge compression component is used to compress the transmission amount required for the image data; The collaborative communication component is used to obtain nodes and timings participating in collaborative communication according to the transmission path and optimization target; The trust evaluation component is used to evaluate whether a neighbor node is a malicious node.

3. The composite routing agent according to claim 2, characterized in that: The length of the service QoS field is 1 byte, including priority, delay, accuracy, throughput, reliability, security and service type fields; The priority field occupies 2 bits and is used to represent 4 priorities, 00 is the lowest priority, 01 is low priority, 10 is high priority, and 11 is the highest priority; The latency occupies 1 bit and is used to represent the service's demand for data forwarding latency. 0 represents normal latency and 1 represents low latency. The precision occupies 1 bit and is used to represent the service's demand for data forwarding precision. 0 represents normal precision and 1 represents high precision. Throughput occupies 1 bit and is used to represent the throughput requirement of business transmission data. 0 represents normal throughput and 1 represents high throughput. Reliability occupies 1 bit and is used to characterize the service's demand for data forwarding reliability. 0 represents normal reliability and 1 represents high reliability. Security occupies 1 bit and is used to represent the business's demand for data forwarding security. 0 represents normal security and 1 represents high security. The service type occupies 1 bit and is used to characterize the classification of service data. 0 represents sequence data and 1 represents image data.

4. The composite routing agent according to claim 3, characterized in that: The routing function module selects at least one functional component from a plurality of functional components according to the mapping relationship and the value of the service QoS field, including: When the delay is low, the virtual pipeline component is configured; when the delay is normal, the composite routing is not configured with the virtual pipeline component; When the service type is sequence data, when the throughput is high throughput and the accuracy is normal accuracy, the edge prediction component is configured to transmit the prediction parameters and the correction data that meets the accuracy requirements in direct transmission mode; when the throughput is high throughput and the accuracy is high accuracy, the edge prediction component is configured to transmit the prediction parameters and the correction data that meets the accuracy requirements in progressive transmission mode; when the throughput is normal throughput and the accuracy is normal accuracy, the edge prediction component is configured to transmit the prediction parameters and the correction data that meets the accuracy requirements in direct transmission mode; when the throughput is normal throughput and the accuracy is high accuracy, the edge prediction component is not configured; When the service type is image data, when the throughput is high throughput and the precision is normal precision, the edge compression component is configured to transmit the average value data and some detail coefficient data that meet the precision requirements in progressive transmission mode; when the throughput is high throughput and the precision is high precision, the edge compression component is configured to transmit the average value data and all detail coefficient data in progressive transmission mode; when the throughput is normal throughput and the precision is normal precision, the edge compression component is configured to transmit the average value data and some detail coefficient data in progressive transmission mode; when the throughput is normal throughput and the precision is high precision, the edge compression component is not configured; When the reliability is high reliability, the collaborative communication component is configured; when the reliability is normal reliability, the collaborative communication component is not configured; When the security is normal, the trust evaluation component is configured according to the security of the transmission environment.

5. The composite routing agent according to claim 3, characterized in that: During data collection: the database read-write module reads the business requirements from the business requirements library and generates a business QoS field based on the business requirements; the routing function module selects the edge prediction component or the edge compression component according to the throughput, accuracy and business type field values ​​in the business QoS field; the database read-write module reads the corresponding parameter data from the transmission strategy library; the data packet processing module generates a business data packet using the data output by the edge prediction component or the edge compression component or the original collected data and the parameter data; the database read-write module writes the business data packet into the corresponding business data queue in the business database according to the priority in the business QoS field.

6. The composite routing agent according to claim 3, characterized in that: When data is sent: the data packet processing module reads the service data packet to be sent from the service database according to the priority order and the first-in-first-out principle; the routing function module selects the trust evaluation component and the virtual pipe component according to the security and delay fields in the service QoS field in the service data packet to be sent; The patching platform selects the next hop node from the next hop node set through the routing component, the trust evaluation component, and the virtual pipeline component; the data packet processing module modifies the receiving node in the service data packet to the next hop node selected by the routing function module; The information transceiver module sends out the data packet to be sent modified by the data packet processing module according to the corresponding transmission mode.

7. The composite routing agent according to claim 3, characterized in that: When receiving data: the information transceiver module receives the business data packet and sends it to the data packet processing module; the data packet processing module reads the business data packet header and obtains the business QoS field; the routing function module selects functional components according to the business QoS field; the plug-in platform uses the selected functional components to perform corresponding processing on the business data packet, including discarding the data packet, saving the data, or using the database read-write module to write the business data packet into the forwarding queue of the business database.

8. The composite routing agent according to claim 7, characterized in that: The plug-in platform utilizes the optional functional components to perform corresponding processing on the business data packet, and also includes: if a collaborative communication component is selected, the plug-in platform utilizes the collaborative communication component to determine whether the current composite routing intelligent entity participates in collaborative communication, and if so, writes the node ID of the current composite routing intelligent entity into the collaborative communication ID in the header of the business data packet, and the plug-in platform writes the business data packet into the highest priority queue of the business database through the database read and write module.

9. The composite routing agent according to claim 2, characterized in that: The control data packet is forwarded in a broadcast manner to synchronize the status information of the underwater communication node and the neighbor node information, including: node ID, node location, remaining energy, and the number of in-degree nodes, out-degree nodes and malicious nodes and their IDs. When a trust evaluation component is configured, the in-degree nodes and out-degree nodes become trusted in-degree nodes and trusted out-degree nodes.

10. The composite routing agent according to claim 2, characterized in that: The business data packet is forwarded in a point-to-point manner and is used to forward business data, including: address information related to data sending and receiving, data sending information related to the routing component, and information related to the business data. When the edge prediction component is configured, the data content transmitted also includes prediction parameters, correction data sequence number and content, and correction data quantity.

Images