Intelligent ship network optimization method, system and device based on multi-link convergence and medium

Abstract

The invention relates to an intelligent ship network optimization method, system and device based on multi-link convergence and a medium. The method comprises the following steps: firstly, acquiring multi-network link connection parameters of the intelligent ship, constructing an adaptive environment, modifying a protocol stack, and integrating a physical link into a virtual link to realize resource convergence; and based on the real-time parameters of the virtual link and the physical link, dynamically distributing traffic by adopting an improved aggregation protocol, and generating a load balancing scheme. And then, regularly sending a detection packet quality index, and predicting a link quality trend in combination with historical data and navigation longitude and latitude. And finally, dynamically adjusting flow distribution according to a prediction result, an application real-time demand, a bandwidth demand and a current position tariff characteristic, and generating a ship-shore communication scheme considering both efficiency and cost. By adopting the method, the stability, reliability and economy of ship-shore communication of the intelligent ship can be effectively enhanced, and the method is suitable for multi-scene network monitoring optimization.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of intelligent ship communication, and particularly relates to an intelligent ship network optimization method, system, device and medium based on multi-link aggregation. BACKGROUND

[0002] With the rapid development of intelligent ship technology, ship-shore communication puts forward higher requirements on the reliability, real-time performance and bandwidth resources of the network. The traditional single-link communication mode has problems such as limited coverage, insufficient bandwidth and weak anti-interference ability, and is difficult to meet the communication needs of intelligent ships in remote control, video monitoring, navigation data transmission and other business scenarios. Although the existing technology realizes a certain degree of link redundancy by introducing 4G / 5G, satellite, microwave and other multi-type network links, there are still defects in multi-link parameter adaptation, dynamic flow distribution and link quality prediction, such as the inability of each link resource to be flexibly distributed according to the real-time state, the difficulty in balancing communication real-time performance and cost, and the lack of forward-looking monitoring of network environment changes in the ship navigation area, resulting in low efficiency and even interruption of ship-shore communication. SUMMARY

[0003] Therefore, it is necessary to provide an intelligent ship network optimization method, system, device and medium based on multi-link aggregation, which can realize logical aggregation of multi-type network resources, improve link load balancing capability and resource utilization, and effectively improve the stability, reliability and economy of intelligent ship ship-shore communication.

[0004] In a first aspect, the application provides an intelligent ship network optimization method based on multi-link aggregation, comprising:

[0005] Obtaining connection parameters of multiple network links required by the intelligent ship, modifying the network protocol stack according to the parameters, and integrating multiple physical links to obtain a virtual link.

[0006] Based on the real-time parameters of the virtual link and each physical link, an improved link aggregation control protocol is used to dynamically distribute data flow, and a load-balanced flow distribution scheme is obtained.

[0007] Timely sending probe data packets to each link and calculating real-time link quality indicators, combining historical link quality data and the current latitude and longitude position of the ship to predict the link quality trend, and obtaining a link quality prediction result.

[0008] According to the link quality prediction result and the flow distribution scheme, combining the real-time performance and bandwidth demand of different applications and the link cost characteristics corresponding to the current position of the ship, dynamically adjusting the flow distribution to obtain a ship-shore communication scheme.

[0009] In one of the embodiments, the connection parameters of multiple network links required by the intelligent ship are acquired, the network protocol stack is modified according to the parameters to build an adaptive interactive environment, and the multiple physical links are integrated to obtain a virtual link, including:

[0010] The interface parameter information and communication protocol information of multiple network links required by the intelligent ship are acquired, the multi-link access device is adapted to the open source operating system, and an interactive environment capable of identifying multiple types of links is obtained.

[0011] Based on the adapted interactive environment, the network protocol stack of the open source operating system kernel is customized and modified, and a multi-link cooperative scheduling logic is added to obtain an optimized network protocol stack supporting unified access, management and data flow of multiple physical links; the cooperative scheduling logic includes protocol conversion rules, data fragmentation and recombination algorithms of different links.

[0012] According to the optimized network protocol stack and the bandwidth characteristics and delay threshold of each link in the intelligent ship navigation scene, the multiple physical links accessed are logically grouped and resource-pooled integrated to build a virtual link framework with dynamic bandwidth superposition function.

[0013] The virtual link framework matches the availability of each physical link according to the real-time navigation position of the intelligent ship, and the bandwidth resources of each physical link are elastically allocated to the virtual link according to the demand, to obtain a virtual link with bandwidth convergence capability and scenario-based link scheduling capability.

[0014] In one of the embodiments, based on the real-time parameters of the virtual link and each physical link, an improved link aggregation control protocol is used to dynamically distribute data traffic to obtain a load-balanced traffic distribution scheme, including:

[0015] The running parameters of the virtual link and each physical link are acquired in real time to obtain real-time quantitative data of link bandwidth, delay, packet loss rate and jitter.

[0016] Based on the real-time quantitative data, an improved link aggregation control protocol is used to determine the traffic bearing weight coefficient of each physical link through a parameter weighting algorithm.

[0017] According to the traffic bearing weight coefficient and the total bandwidth demand of the virtual link, the traffic distribution proportion of each physical link is calculated to obtain an initial traffic distribution scheme.

[0018] The initial traffic distribution scheme is subjected to load verification, and if the load of a certain physical link exceeds a preset threshold, the distribution proportion of the corresponding link is adjusted downward by the weight coefficient proportion, and the downward adjusted part of the traffic distribution is supplemented to the low-load link to obtain the adjusted traffic distribution proportion of each link.

[0019] A traffic allocation scheme adapted to the multi-link characteristics of intelligent ships is generated based on the traffic allocation ratio of each link. The traffic allocation scheme includes the specific quotas of high-priority traffic and non-high-priority traffic carried by physical links, the traffic switching trigger threshold, and the execution strategy to maximize the bandwidth utilization of low-cost links.

[0020] In one embodiment, the traffic allocation percentage for each physical link is calculated using the following formula:

[0021]

[0022] in, Indicates the first Traffic allocation percentage for each physical link Indicates the first The weighting factor of the link bandwidth. Indicates the first The weighting factor for link delay. Indicates the first The weighting coefficients for packet loss rate across links. Indicates the first The weighting coefficient of link jitter, Indicates the first Real-time bandwidth of each link Indicates the first Real-time latency of the link Indicates the first Real-time packet loss rate of the link, Indicates the first Real-time jitter of the link, This indicates the total number of physical links. This indicates the total bandwidth requirement of the virtual link.

[0023] In one embodiment, probe data packets are periodically sent to each link and real-time link quality indicators are calculated. The link quality change trend is predicted by combining historical link quality data with the ship's current navigation latitude and longitude position, resulting in a link quality prediction result, including:

[0024] A probe data packet transmission plan is generated based on the characteristics of the intelligent ship's navigation area and the link type; the data packet transmission frequency in the transmission plan is set differently according to the link's real-time requirements.

[0025] Based on the transmission plan, a probe data packet containing a precise timestamp and link identification information is created; the probe data packet carries the basic verification fields required for link quality detection.

[0026] The probe data packets returned by each link are received, the link delay is calculated by comparing the sending or receiving time stamps, the packet loss rate is calculated by counting the number of packet losses in a unit of time, the jitter is calculated by analyzing the fluctuation amplitude of the packet transmission time, and the link quality indicators including delay, packet loss rate and jitter are output.

[0027] The link quality indicators and the real-time latitude and longitude data of the ship are read, the key environmental variables affecting the link quality are screened out by combining the historical link quality characteristics of the same period, and the structured link quality trend prediction data is generated.

[0028] The link quality trend prediction data is input into the trained link quality prediction model, the correlation between the historical link quality change rule and the current environmental variables is learned, and the change threshold of the link quality indicators of each link in the future preset time period is output, and the link quality prediction result is obtained.

[0029] In one embodiment, according to the link quality prediction result and the traffic distribution scheme, combined with the real-time and bandwidth requirements of different applications and the link cost characteristics corresponding to the current position of the ship, the ship-shore communication scheme is dynamically adjusted, including:

[0030] Combined with the real-time requirement threshold of the high-priority application of the intelligent ship, according to the link quality prediction result and the link quality indicator of each link, the weight assignment algorithm is used to generate the traffic distribution weight of the high-priority application.

[0031] Based on the traffic distribution weight and the real-time available bandwidth data of the low-cost link, combined with the utilization rate of the low-cost link in the traffic distribution scheme, the deep reinforcement learning algorithm is input to generate a dynamic adjustment strategy.

[0032] According to the dynamic adjustment strategy, the optimized traffic distribution scheme is obtained by calculating the model combined with the historical optimal traffic distribution parameters; the traffic distribution scheme includes the carrying capacity of high-priority and non-high-priority traffic of each link.

[0033] According to the traffic distribution scheme, the delay characteristic data and the quality prediction result of each link are extracted, the links whose delay meets the high-priority requirement and whose predicted quality does not deteriorate are screened and sorted as candidate paths, and the low-delay transmission path is determined combined with the bandwidth availability of the candidate path.

[0034] According to the link characteristics of the low-delay transmission path and the carrying capacity of the path in the traffic distribution scheme, the ship communication data sending frequency and the data packet size are adjusted, and the ship-shore communication scheme considering high-priority real-time transmission and resource efficient utilization is generated.

[0035] In one embodiment, the calculation model is represented by the following formula:

[0036]

[0037] wherein, denotes the link allocated to the class of traffic, denotes the traffic allocation weight of the denotes the real-time available bandwidth of the denotes the total number of physical links, denotes the historical optimal parameter correction coefficient, , denotes the optimal proportion of the class of traffic carried by the denotes the total demand quota of the class of traffic, , denotes the link quality prediction correction factor.

[0038] In a second aspect, the present application also provides an intelligent ship network monitoring system based on multi-link aggregation, which comprises:

[0039] A link parameter adaptation module is configured to obtain connection parameters of multiple network links required by the intelligent ship, modify the network protocol stack according to the parameters to build an adaptive interactive environment, and integrate the multiple physical links to obtain a virtual link.

[0040] A traffic dynamic allocation module is configured to dynamically allocate data traffic based on the virtual link and real-time parameters of each physical link by using an improved link aggregation control protocol to obtain a load-balanced traffic allocation scheme.

[0041] A link quality prediction module is configured to send probe data packets to each link at regular intervals and calculate real-time link quality indicators, combine historical link quality data and the current latitude and longitude position of the ship to predict the link quality trend, and obtain a link quality prediction result.

[0042] A communication scheme generation module is configured to dynamically adjust the traffic allocation based on the link quality prediction result and the traffic allocation scheme, combine the real-time and bandwidth requirements of different applications and the link cost characteristics corresponding to the current position of the ship, and obtain a ship-shore communication scheme.

[0043] In a third aspect, the present application also provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the foregoing method when executing the computer program.

[0044] In a fourth aspect, the present application also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to implement the foregoing method.

[0045] The above-mentioned intelligent ship network optimization method, system, computer device and storage medium based on multi-link aggregation first acquire connection parameters of multi-network links (such as 4G / 5G, satellite, microwave, etc.) of an intelligent ship, construct an adaptive interaction environment and modify a network protocol stack based on the parameters, integrate multiple physical links into a virtual link, and realize logical abstraction and resource aggregation of the underlying link. Based on the real-time parameters (bandwidth, delay, packet loss rate, etc.) of the virtual link and each physical link, an improved link aggregation control protocol is used to dynamically allocate data traffic, generate a load-balanced traffic distribution scheme, and ensure that the resources of each link are elastically scheduled as needed. The real-time link quality indicators (delay, packet loss rate, jitter) are calculated by sending probe data packets to each link at regular intervals, and the link quality prediction results are formed by combining the historical link quality data and the current latitude and longitude position of the ship to predict the trend of the link quality, thereby realizing forward-looking analysis of the network state. According to the link quality prediction results and the traffic distribution scheme, combined with the real-time requirements (such as remote control instruction delay threshold) of different applications, bandwidth requirements, and the link cost characteristics corresponding to the current position of the ship, the traffic distribution strategy is dynamically adjusted, and finally a ship-shore communication scheme is generated to realize the collaborative optimization of communication efficiency and cost. This method realizes the logical aggregation of multiple types of network resources through multi-link parameter integration and protocol stack adaptation, improves the link load balancing capability and resource utilization rate by using an improved link aggregation control protocol to dynamically allocate traffic, and can provide data support for traffic scheduling by using a link quality prediction mechanism based on probe data packets and sailing positions to perceive network state changes in advance. The dynamic adjustment strategy takes into account the real-time performance and cost control of communication by combining application requirements and cost characteristics. Overall, this method effectively improves the stability, reliability and economy of ship-shore communication of an intelligent ship, and is suitable for efficient network monitoring and optimization in multiple scenarios such as open sea and near shore. BRIEF DESCRIPTION OF DRAWINGS

[0046] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related art, the drawings needed to be used in the embodiments or related art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0047] Figure 1 A flowchart of an intelligent ship network optimization method based on multi-link aggregation provided by an embodiment of the present application;

[0048] Figure 2 A structural block diagram of an intelligent ship network monitoring system based on multi-link aggregation provided by an embodiment of the present application. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application is further described in detail below in combination with the drawings and examples. It should be understood that the specific examples described herein are only used to explain the present application and do not limit the present application.

[0050] In one embodiment, as shown in Figure 1 The present application provides a multi-link aggregation-based intelligent ship network optimization method, which can include the following steps:

[0051] Step S101, obtain the connection parameters of multiple network links required by the intelligent ship, modify the network protocol stack according to the parameters, and integrate the multiple physical links to obtain a virtual link.

[0052] Specifically, the connection parameters (including IP address, port protocol, transmission standard, etc.) of the heterogeneous network links (such as 4G / 5G, satellite, microwave, etc.) accessed by the intelligent ship are collected through the interface, the adaptive interaction environment is constructed based on the parameter characteristics, the link layer and network layer of the network protocol stack are modified (such as the multi-link aggregation logic of the TCP / IP protocol stack), the bandwidth, address, etc. of the physical link are logically integrated through the link bundling technology, and the virtual link with a unified data transmission interface is formed, providing a standardized network communication channel for the upper layer application.

[0053] Step S102, based on the real-time parameters of the virtual link and each physical link, a modified link aggregation control protocol is used to dynamically allocate data traffic to obtain a load-balanced traffic distribution scheme.

[0054] Using the real-time collected operating parameters (including bandwidth utilization, delay, packet loss rate, etc.) of the virtual link and each physical link, the dynamic weight algorithm of the improved link aggregation control protocol (LACP) is used to calculate the traffic bearing weight coefficient of each physical link according to the current load state of the link and the quality of service (QoS) demand; through the proportional operation of the weight coefficient and the total bandwidth demand of the virtual link, a load-balanced traffic distribution scheme is generated, realizing the collaborative transmission and resource scheduling of different service data on multiple links.

[0055] Step S103, periodically send probe data packets to each link and calculate real-time link quality indicators, combine historical link quality data and the current latitude and longitude position of the ship to predict the link quality trend, and obtain the link quality prediction result.

[0056] According to a preset period (such as 50ms-100ms), the time-stamped probe data packets are sent to each physical link, and the real-time delay, packet loss rate, jitter and other quality indicators are calculated by analyzing the return packets; the real-time indicators are associated with the historically stored link quality data (classified according to time, latitude and longitude, etc.), and the geographical environment characteristics (such as the signal coverage law of near shore / sea) corresponding to the current sailing latitude and longitude position of the ship are combined, and a machine learning model (such as an LSTM time series model) is used to predict the link quality change trend in the future period. During model training and prediction, multi-dimensional data will be fused first, including link historical quality data classified and stored according to time period, sailing latitude and longitude, and current delay, packet loss rate, jitter and other indicators obtained by sending real-time probe data packets, and the geographical characteristics (such as the signal coverage intensity of 4G / 5G in near shore area and the law of atmospheric attenuation of satellite signal in sea area) of the current position of the ship. Through model learning, the internal correlation between link quality and time and position changes (for example, when sailing near shore, the fluctuation trend of link delay with the increase of distance from the base station; when sailing in the sea, the rising mode of satellite packet loss rate in severe weather) is learned. Finally, the quality change trend of each physical link in the future set period (such as 1-5 minutes in the future) is accurately predicted, and the link quality prediction result containing the specific delay fluctuation range (such as 60-80ms), the packet loss rate threshold (such as ≤2%) and the prediction result confidence interval is output.

[0057] Step S104, according to the link quality prediction result and the traffic distribution scheme, combined with the real-time and bandwidth requirements of different applications and the link cost characteristics corresponding to the current position of the ship, dynamically adjust the traffic distribution to obtain the ship-shore communication scheme.

[0058] Specifically, based on the link quality prediction result (such as the delay of a certain link will exceed the threshold in the next 5 minutes) and the traffic distribution scheme, the real-time requirements (such as remote control instructions require delay ≤100ms) and bandwidth requirements (such as video backhaul requires 10Mbps) of different applications, and the link cost standards (such as satellite link charges by traffic, near shore 4G / 5G charges by package) corresponding to the current position of the ship, the cost-performance balancing algorithm is used to dynamically adjust the traffic distribution proportion of each link, and finally the ship-shore communication scheme considering communication quality and cost control is generated, realizing the optimal path transmission of business data.

[0059] The above-mentioned intelligent ship network optimization method based on multi-link aggregation first acquires connection parameters of multiple network links (such as 4G / 5G, satellite, microwave, etc.) of the intelligent ship, constructs an adaptive interaction environment based on the parameters, modifies the network protocol stack, integrates the multiple physical links into a virtual link, realizes logical abstraction and resource aggregation of the underlying link, dynamically allocates data flow based on the real-time parameters (bandwidth, delay, packet loss rate, etc.) of the virtual link and each physical link by using an improved link aggregation control protocol, generates a load-balanced traffic distribution scheme, ensures that the resources of each link are elastically scheduled as needed, sends probe data packets to each link at regular intervals, calculates real-time link quality indicators (delay, packet loss rate, jitter), combines historical link quality data and the current latitude and longitude position of the ship, predicts the link quality trend, forms a link quality prediction result, realizes forward-looking analysis of the network state, dynamically adjusts the traffic distribution strategy according to the link quality prediction result and the traffic distribution scheme, combines the real-time requirements of different applications (such as remote control instruction delay threshold), bandwidth requirements, and the link cost characteristics corresponding to the current position of the ship, and finally generates a ship-shore communication scheme to realize the coordinated optimization of communication efficiency and cost. This method realizes logical aggregation of multiple types of network resources through multi-link parameter integration and protocol stack adaptation, improves link load balancing capability and resource utilization rate by dynamically allocating traffic using an improved link aggregation control protocol, and can provide data support for traffic scheduling by using a link quality prediction mechanism based on probe data packets and sailing position to perceive network state changes in advance. The dynamic adjustment strategy takes into account the real-time performance and cost control of communication. Overall, this method effectively improves the stability, reliability and economy of ship-shore communication of intelligent ships, and is suitable for efficient network monitoring and optimization in multiple scenarios such as open sea and near shore.

[0060] In one embodiment, acquiring connection parameters of multiple network links required by an intelligent ship, modifying the network protocol stack according to the parameters, and integrating multiple physical links to obtain a virtual link can include the following steps:

[0061] Step S201: Acquire interface parameter information and communication protocol information of multiple network links required by an intelligent ship, adapt a multi-link access device to an open source operating system to obtain an interactive environment that can identify multiple types of links.

[0062] Step S202: Based on the adapted interactive environment, customize and modify the network protocol stack of the kernel of the open source operating system, add multi-link cooperative scheduling logic, and obtain an optimized network protocol stack that supports unified access, management and data flow of multiple physical links; the cooperative scheduling logic includes protocol conversion rules, data fragmentation and recombination algorithms of different links.

[0063] Step S203, according to the optimization of the network protocol stack, the bandwidth characteristics and delay threshold of each link in the intelligent ship navigation scene are combined, the accessed multiple physical links are logically grouped and resource pooled and integrated, and a virtual link framework with dynamic bandwidth superposition function is constructed.

[0064] Step S204, the availability of each physical link is matched according to the real-time navigation position of the intelligent ship in the virtual link framework, and the bandwidth resources of each physical link are flexibly allocated to the virtual link according to the demand, so that a virtual link with bandwidth convergence capability and scenario-based link scheduling capability is obtained.

[0065] Firstly, the interface parameter information and communication protocol information of the required multiple network links (covering 4G / 5G, satellite, microwave, etc.) of the intelligent ship are systematically collected: the interface parameter information includes interface type (such as RJ45), maximum transmission rate (such as 4G / 5G link 100Mbps, satellite link 2Mbps), signal modulation mode, etc.; the communication protocol information includes link layer protocol (such as IEEE 802.3 (Ethernet) for Ethernet link, RLC / MAC layer protocol for mobile cellular (4G / 5G) adaptation), network layer protocol (such as IPv4 / IPv6) and application layer adaptation protocol (such as SCPS protocol for satellite link). Based on the above parameters and information, through the development of an adaptive driver, the multi-link access device (4G / 5G CPE, satellite communication terminal, microwave transceiver, etc.) is compatible with the open source operating system (such as Linux), the device node mapping and data interaction logic configuration are completed, and finally a standardized interaction environment that can automatically identify and access multiple types of links is built, realizing the collaborative work of the underlying hardware and the system.

[0066] Secondly, based on the adapted interaction environment, the network protocol stack of the open source operating system kernel is customized and modified: for the link layer, a heterogeneous protocol conversion module is added to realize the bidirectional conversion of satellite SCPS protocol and TCP protocol, microwave private protocol and IP protocol; for the network layer, a dynamic routing logic based on link characteristics is developed to support data transmission path selection according to link quality priority; for the transport layer, the TCP congestion control algorithm is optimized to adapt to the retransmission mechanism of high delay links (such as satellite). At the same time, multi-link collaborative scheduling logic is implanted, including dynamic data fragmentation algorithm based on link bandwidth (splitting large data packets according to the real-time bandwidth proportion of each link) and reassembly algorithm based on timestamp (ensuring that the fragmented data is reassembled in order at the receiving end), finally forming an optimized network protocol stack that supports unified access, centralized management and efficient data flow of multiple physical links.

[0067] Then, based on the optimized network protocol stack and combined with the inherent characteristics of each link in typical navigation scenarios of intelligent ships (nearshore, offshore, bridge areas, etc.): in nearshore scenarios, 4G / 5G links have high bandwidth (20-100Mbps) and low latency (50-100ms), while in offshore scenarios, satellite links have wide coverage but limited bandwidth (1-5Mbps) and high latency (500-1000ms), and microwave links are suitable for short-distance (5-50km) and high bandwidth (50-100Mbps) transmission. The connected physical links are logically grouped according to function and scenario (e.g., the "nearshore communication group" includes 4G / 5G and microwave, and the "offshore communication group" includes satellite and self-organizing network). Through resource pooling technology, the bandwidth, address and other resources of the links in each group are abstracted into a unified resource pool to build a virtual link framework with dynamic bandwidth superposition function. When multiple links are available at the same time, the framework can automatically aggregate the bandwidth of each link (e.g., 100Mbps of 4G / 5G and 50Mbps of microwave are superimposed to 150Mbps) to meet the high bandwidth service requirements.

[0068] Finally, the virtual link framework obtains real-time navigation latitude and longitude through ship positioning modules (such as GPS / BeiDou), and combines it with pre-stored link coverage maps (such as 4G / 5G coverage within 30km of the nearshore area and satellite coverage in the open sea area) to dynamically match the availability status of each physical link (such as automatically excluding microwave links that are unavailable due to obstruction when entering the bridge area); based on service requirements (such as low latency for remote control and high bandwidth for video backhaul), it elastically allocates corresponding bandwidth resources from available physical links to virtual links (such as prioritizing the allocation of 4G / 5G links with latency ≤100ms for control commands, and allocating superimposed bandwidth of 4G / 5G and microwave for video backhaul), ultimately forming a virtual link that combines bandwidth aggregation capability and scenario-based link scheduling capability, realizing on-demand scheduling and efficient utilization of multi-link resources.

[0069] This embodiment addresses the compatibility issues between various heterogeneous links and operating systems by systematically collecting interface parameter information and communication protocols, and adapting devices, ensuring that underlying devices can be uniformly identified and accessed. Through customized modifications to the network protocol stack and the implantation of collaborative scheduling logic, it overcomes protocol barriers between different links, enabling efficient cross-link data flow. Based on logical grouping and resource pooling integration according to navigation scenarios and link characteristics, it constructs a virtual link framework with dynamic bandwidth overlay, improving the overall utilization of bandwidth resources. Combined with link availability matching based on the ship's real-time location and elastic resource allocation, the virtual link can adapt to the communication needs of different navigation scenarios, effectively enhancing the reliability, flexibility, and bandwidth supply capacity of the intelligent ship network, providing stable and efficient underlying support for ship-to-shore multi-service communication.

[0070] In one embodiment, based on the real-time parameters of virtual links and each physical link, an improved link aggregation control protocol is used to dynamically allocate data traffic to obtain a load-balanced traffic allocation scheme, which may include the following steps:

[0071] Step S301: Obtain the operating parameters of the virtual link and each physical link in real time to obtain real-time quantitative data of link bandwidth, latency, packet loss rate and jitter.

[0072] Step S302: Based on real-time quantitative data, an improved link aggregation control protocol is adopted, and the traffic carrying weight coefficient of each physical link is determined through a parameter weighting algorithm.

[0073] Step S303: Calculate the traffic allocation ratio of each physical link based on the traffic carrying weight coefficient and the total bandwidth requirement of the virtual link to obtain the initial traffic allocation scheme.

[0074] Step S304: Perform load verification on the initial traffic allocation scheme. If the load of a physical link exceeds the preset threshold, reduce the corresponding link allocation ratio according to the weight coefficient ratio and supplement the reduced traffic allocation to the low-load link to obtain the adjusted traffic allocation ratio of each link.

[0075] Step S305: Generate a traffic allocation scheme adapted to the multi-link characteristics of smart ships based on the traffic allocation ratio of each link; the traffic allocation scheme includes the specific quotas of high-priority traffic and non-high-priority traffic carried by physical links respectively, the traffic switching trigger threshold, and the execution strategy to maximize the bandwidth utilization of low-cost links.

[0076] Specifically, the system first acquires real-time operational parameters of the virtual links and each physical link (4G / 5G, satellite, microwave, etc.) through a monitoring module, obtaining real-time quantitative data on link bandwidth, latency, packet loss rate, and jitter. Based on this data, an improved link aggregation control protocol is used to calculate the traffic carrying weight coefficient of each physical link through a parameter weighting algorithm (weighting factors include latency as a percentage of 30%, bandwidth availability as 40%, and packet loss rate as 30%). Then, according to the weight coefficients and the total bandwidth requirement of the virtual links, the traffic share of each physical link is allocated proportionally to form an initial traffic allocation scheme.

[0077] Subsequently, load verification is performed on the initial scheme: if the load on a certain link exceeds a preset threshold (e.g., 80% bandwidth utilization), the allocation ratio of that link is reduced according to its weight, and the reduced portion is supplemented by allocating it according to the weight ratio of low-load links, resulting in an adjusted traffic allocation ratio. Finally, a traffic allocation scheme is generated based on the allocation ratio of each link, specifying the specific quotas for high-priority traffic (e.g., remote control commands) and non-high-priority traffic (e.g., video backhaul) carried by each physical link, setting a traffic switching trigger threshold (e.g., initiating switching when a link's latency exceeds 100ms), and formulating an execution strategy to maximize the bandwidth utilization of low-cost links (e.g., near-shore 4G / 5G).

[0078] This embodiment achieves dynamic weight allocation of multi-link resources through real-time parameter acquisition and weighting algorithms, ensuring precise matching between traffic distribution and real-time link status. A load verification mechanism prevents single-link overload, improving overall network stability. The combination of priority traffic and pricing strategies guarantees real-time transmission of critical services while reducing communication costs. Specifically, the latency compliance rate of high-priority traffic is improved, the bandwidth utilization rate of low-cost links is increased on average, and the multi-link load balancing is controlled within ±10%. This effectively solves the problems of resource waste, uneven load, and insufficient support for critical services in multi-link communication on intelligent ships, providing an efficient and economical traffic management solution for ship-to-shore communication.

[0079] In one embodiment, the traffic allocation percentage for each physical link can be calculated using the following formula:

[0080]

[0081] in, Indicates the first Traffic allocation percentage for each physical link Indicates the first The weighting factor of the link bandwidth. Indicates the first The weighting factor for link delay. Indicates the first The weighting coefficients for packet loss rate across links. Indicates the first The weighting coefficient of link jitter, Indicates the first Real-time bandwidth of each link Indicates the first Real-time latency of the link Indicates the first Real-time packet loss rate of the link, Indicates the first Real-time jitter of the link, This indicates the total number of physical links. This indicates the total bandwidth requirement of the virtual link.

[0082] Preferably, for typical business data of intelligent ships (such as high-definition surveillance video and navigation log file transmission), a sample set with data requirement feature annotations is constructed—video data is labeled "sensitive to latency ≤100ms and jitter ≤20ms", and file transmission is labeled "sensitive to bandwidth ≥20Mbps". Historical bandwidth (B), latency (D), packet loss rate (L), and jitter (J) data of 4G / 5G and satellite links under different navigation scenarios such as nearshore and offshore are incorporated. The model is trained using a random forest or neural network model, allowing the model to deeply learn the differences in preference for various link parameters among different data types (e.g., video has a much lower tolerance for latency fluctuations than file transmission). After the model training is completed, dynamic parameters of each physical link are collected in real time during actual navigation (e.g., nearshore 4G...). Suddenly, due to factors such as bandwidth dropping from 100Mbps to 30Mbps due to bridge obstruction and latency of distant satellites increasing from 600ms to 900ms due to cloud cover, along with changes in the network environment (such as signal blockage and weather interference), real-time data is input into the model. The model automatically adjusts the weights of each parameter based on the correlation patterns learned during training. For example, when transmitting video, the latency weight is increased from the basic 30% to 50%, and the jitter weight from 10% to 25%, while the bandwidth weight is reduced. When transmitting files, the bandwidth weight is increased from 40% to 60%, and the latency weight is reduced to 15%. The model can also accurately capture non-linear relationships between parameters (e.g., when satellite latency exceeds 800ms, the impact of packet loss rate on link capacity increases non-linearly, and the model can dynamically amplify the adjustment range of the packet loss rate weight). Finally, the optimized link parameter weights and the real-time collected link indicators are substituted into the traffic allocation ratio formula. The calculation yields an allocation result that adapts to the current data requirements and network status. This avoids the problem of fixed weights failing to match different data preferences, and can quickly respond to dynamic changes in the link. It also significantly reduces allocation deviations caused by rigid parameters or failure to consider nonlinear relationships, ensuring that traffic allocation always aligns with business needs and the actual network status.

[0083] This embodiment calculates the traffic allocation ratio for each physical link using a formula. It incorporates four core operating parameters—real-time bandwidth, latency, packet loss rate, and jitter—into the allocation logic. Furthermore, it flexibly adjusts the impact of different parameters on the allocation results through independent weighting coefficients (e.g., increasing latency weight to prioritize high-priority services with low latency requirements). It also achieves a quantitative balance in the allocation ratio by combining the total number of physical links with the total bandwidth requirements of virtual links. This calculation method avoids allocation deviations caused by single-parameter decisions and ensures that traffic allocation accurately matches the real-time performance of each link with the overall bandwidth requirements. Ultimately, it effectively improves the load balancing of multiple links in intelligent ships, the stability of critical service transmission, and the utilization rate of link resources. It solves the problems of incomplete parameter consideration and insufficient adaptability in traditional allocation methods, providing quantitative calculation support for the efficient operation of ship-shore communication.

[0084] In one embodiment, probe data packets are periodically sent to each link and real-time link quality indicators are calculated. By combining historical link quality data with the ship's current navigation latitude and longitude position, the link quality change trend is predicted to obtain the link quality prediction result. This may include the following steps:

[0085] Step S401: Generate a probe data packet transmission plan based on the characteristics of the intelligent ship's navigation area and the link type; the data packet transmission frequency in the transmission plan is set differently according to the link's real-time requirements.

[0086] Step S402: Create a probe data packet containing a precise timestamp and link identification information according to the sending plan; the probe data packet carries the basic verification fields required for link quality detection.

[0087] Step S403: Receive probe data packets returned by each link, calculate link latency by comparing sending or receiving timestamps, calculate packet loss rate by counting the number of data packets lost per unit time, calculate jitter by analyzing the fluctuation range of data packet transmission time, and output link quality indicators including latency, packet loss rate, and jitter.

[0088] Step S404: Read the link quality indicators and real-time latitude and longitude data of the ship, combine them with the historical link quality characteristics of the same period of latitude and longitude, screen out the key environmental variables that affect the link quality, and generate structured link quality trend prediction data.

[0089] Step S405: Input the link quality trend prediction data into the trained link quality prediction model. By learning the correlation between historical link quality change patterns and current environmental variables, output the change threshold of link quality indicators for each link in the future preset time period to obtain the link quality prediction result.

[0090] First, a probe data packet transmission plan is generated based on the characteristics of the intelligent ship's navigation area (nearshore, offshore, bridge area, etc.) and link type (4G / 5G, satellite, microwave): a higher transmission frequency (100ms / time) is set for links with high real-time requirements (such as satellite links carrying remote control commands), and a lower frequency (500ms / time) is set for non-high-priority links (such as 4G links for video backhaul). Second, probe data packets are created according to the plan, embedding nanosecond-level precise timestamps, unique link identification information (such as link ID, transmission standard), and carrying basic detection fields such as CRC check fields to ensure that the data packets can be used for link quality quantitative analysis.

[0091] Then, the system receives data packets from each link, calculates real-time latency using the difference between send and receive timestamps, calculates the packet loss rate by counting the number of sent and lost data packets per unit time (e.g., 10 seconds), analyzes the fluctuation range of data packet transmission time to calculate jitter, and outputs link quality data containing three indicators. Next, the system reads the link quality indicators and the ship's real-time latitude and longitude, matches them with historical link quality data from the same period and region (e.g., packet loss rate characteristics of 4G links in nearshore areas during summer), filters out key environmental variables, and generates structured trend prediction data. Finally, the prediction data is input into a trained LSTM neural network model, which learns the correlation between historical link quality changes and environmental variables, outputs the threshold values ​​for changes in latency and packet loss rate of each link in the next 10 minutes, forming the link quality prediction results.

[0092] This embodiment achieves precise matching of detection frequency and real-time requirements through differentiated transmission plans based on navigation area and link type, avoiding bandwidth waste caused by invalid detection. Nanosecond-level timestamps and CRC check fields ensure the calculation accuracy of indicators such as latency and packet loss rate, providing a reliable data foundation for link quality analysis. A key environmental variable screening mechanism combining latitude, longitude, and historical data makes link quality trend prediction more closely aligned with actual navigation scenarios (such as the impact of solar storms on offshore satellite links). The LSTM model's ability to learn historical patterns improves the accuracy of predicting future quality indicator change thresholds (error rate ≤ 8%). Overall, this method provides forward-looking quality prediction support for dynamic multi-link traffic allocation on intelligent ships, proactively avoiding high-latency and high-packet-loss links, improving the stability of critical service transmissions, reducing the risk of communication interruptions due to sudden changes in link quality, and optimizing resource utilization and cost control in ship-to-shore communication.

[0093] In one embodiment, based on the link quality prediction results and the traffic allocation scheme, and taking into account the real-time and bandwidth requirements of different applications and the link tariff characteristics corresponding to the ship's current location, the traffic allocation is dynamically adjusted to obtain the ship-to-shore communication scheme, which may include the following steps:

[0094] Step S501: Combining the real-time requirement threshold of high-priority applications of intelligent ships, and based on the link quality prediction results and link quality indicators of each link, a weight assignment algorithm is used to generate traffic allocation weights specific to high-priority applications.

[0095] Step S502: Based on the traffic allocation weight and the real-time available bandwidth data of low-cost links, and combined with the utilization rate of low-cost links in the traffic allocation scheme, a deep reinforcement learning algorithm is used to generate a dynamic adjustment strategy.

[0096] Step S503: Based on the dynamic adjustment strategy and the historical best traffic allocation parameters, the optimized traffic allocation scheme is obtained through the calculation model; the traffic allocation scheme includes the carrying capacity of high and non-high priority traffic for each link.

[0097] Step S504: Extract the latency characteristic data and quality prediction results of each link according to the traffic allocation scheme, filter the links whose latency meets the high priority requirements and whose predicted quality has not deteriorated, and sort them as candidate paths. Combine the bandwidth availability of the candidate paths to determine the low latency transmission path.

[0098] Step S505: Based on the link characteristics of the low-latency transmission path and the path capacity in the traffic allocation scheme, adjust the ship communication data transmission frequency and data packet size to generate a ship-shore communication scheme that balances high-priority real-time transmission and efficient resource utilization.

[0099] Specifically, firstly, based on the real-time requirement threshold (e.g., latency ≤ 100ms) of high-priority applications of intelligent ships (e.g., remote control commands), and according to the link quality prediction results of each link (e.g., latency fluctuation range in the next 5 minutes) and real-time quality indicators (current latency, packet loss rate), a dedicated traffic allocation weight for high-priority applications is generated through a weighted assignment algorithm (where latency indicators account for 60% and packet loss rate accounts for 40%).

[0100] Secondly, based on the traffic allocation weight, the real-time available bandwidth data of low-cost links (such as nearshore 4G / 5G), and the utilization rate of low-cost links in the traffic allocation scheme (such as the current utilization rate of 70%), the above parameters are input into a deep reinforcement learning algorithm (such as the DQN model), and a dynamic adjustment strategy (such as the traffic migration rule when the utilization rate of low-cost links exceeds 85%) is generated through iterative training.

[0101] Next, based on the dynamic adjustment strategy and combined with the historical best traffic allocation parameters (such as link load balancing parameters in similar navigation scenarios), the optimized traffic allocation scheme is obtained through a calculation model (such as the traffic quota allocation formula), which clarifies the specific quotas for high-priority traffic (such as satellite links carrying 50% of control commands) and non-high-priority traffic (such as 4G links carrying 80% of video backhaul) carried by each link.

[0102] Then, the latency characteristic data and quality prediction results of each link are extracted from the optimization scheme. Links whose latency meets the high priority requirement (≤100ms) and whose predicted quality does not deteriorate (e.g., future packet loss rate ≤1%) are selected and sorted into candidate paths from low to high latency. The final low-latency transmission path is determined by combining the real-time bandwidth availability of the candidate paths (e.g., remaining bandwidth ≥ 30% of the requirement).

[0103] Finally, based on the link characteristics of the low-latency transmission path (such as the signal attenuation coefficient of the satellite link) and the carrying capacity of the path in the scheme (such as the need to carry 1Mbps control commands), the data transmission frequency and data packet size of the ship's communication equipment are adjusted to generate a ship-to-shore communication scheme that takes into account both real-time transmission of high-priority services and efficient utilization of communication resources.

[0104] This embodiment ensures precise alignment between traffic allocation and the real-time requirements of core services by generating dedicated weights for high-priority applications; dynamic adaptation of low-cost link utilization using deep reinforcement learning algorithms reduces costs while maintaining communication quality; optimization of the calculation model based on historical optimal parameters makes the traffic allocation scheme both scientific and based on practical experience; screening of low-latency paths and verification of bandwidth availability further improve the transmission stability of high-priority services; and targeted adjustments to device transmission frequency and power achieve refined utilization of communication resources. Overall, the real-time performance of high-priority applications is improved, the utilization rate of low-cost links is increased, and communication resource waste is reduced. This effectively addresses the problems of insufficient core service support, low resource utilization, and difficulty in cost control in multi-link communication on intelligent ships, providing a comprehensive optimization solution for efficient ship-shore communication.

[0105] In one embodiment, the computational model can be represented by the following formula:

[0106]

[0107] in, Indicates the first Link assigned to the first Traffic quota, Indicates the first Traffic allocation weights for each link Indicates the first The real-time available bandwidth of each link. This indicates the total number of physical links. This represents the historical optimal parameter correction coefficient. , Indicates the first The link carried the first in the historical scheme The optimal proportion of traffic of this type Indicates the first Total demand for traffic types , This represents the link quality prediction correction factor.

[0108] This embodiment's calculation model integrates link traffic allocation weights, real-time available bandwidth, historical optimal parameters, and link quality prediction correction factors to construct a traffic allocation quantification mechanism that combines real-time adaptability with the inheritance of historical experience. The basic allocation quota is determined by the traffic allocation weights and real-time bandwidth ratio. Historical optimal parameter correction coefficients balance the current link status with historical optimization experience, avoiding allocation deviations caused by real-time data fluctuations. Historical optimal ratios are used to introduce optimal resource allocation experience in similar scenarios, improving the practical reliability of the allocation scheme. The allocation results are dynamically adjusted using link quality prediction correction factors, ensuring that traffic quotas match future link quality trends in advance. This model effectively solves the problems of delayed real-time status response, insufficient utilization of historical experience, and lack of forward-looking prediction in traditional allocation methods. It achieves dynamic optimization allocation of multi-link traffic on intelligent ships, increasing the real-time compliance rate of high-priority services by more than 20% and improving the average link resource utilization rate by 15%, providing quantitative calculation support for ship-to-shore communication that balances real-time performance, stability, and historical optimization experience.

[0109] In one embodiment, such as Figure 2 As shown, this application also provides an intelligent ship network monitoring system based on multi-link convergence, the system may include:

[0110] The link parameter adaptation module 601 is used to obtain the connection parameters of various network links required by the intelligent ship, construct an adaptation interaction environment based on the parameters to modify the network protocol stack, and integrate multiple physical links to obtain a virtual link.

[0111] The traffic dynamic allocation module 602 is used to dynamically allocate data traffic based on the real-time parameters of virtual links and each physical link, using an improved link aggregation control protocol, to obtain a load-balanced traffic allocation scheme.

[0112] The link quality prediction module 603 is used to periodically send probe data packets to each link and calculate real-time link quality indicators. It combines historical link quality data with the ship's current navigation latitude and longitude position to predict the link quality change trend and obtain the link quality prediction result.

[0113] The communication scheme generation module 604 is used to dynamically adjust the traffic allocation to obtain the ship-shore communication scheme based on the link quality prediction results and traffic allocation scheme, combined with the real-time and bandwidth requirements of different applications and the link tariff characteristics corresponding to the ship's current position.

[0114] The aforementioned intelligent ship network monitoring system based on multi-link aggregation firstly involves a link parameter adaptation module acquiring connection parameters of multiple network links on the intelligent ship, constructing an adaptation interaction environment, and modifying the network protocol stack to integrate multiple physical links into virtual links, thereby achieving logical abstraction of underlying resources. Secondly, a dynamic traffic allocation module dynamically allocates traffic based on real-time parameters (bandwidth, latency, etc.) of the virtual and physical links, using an improved link aggregation control protocol to generate a load balancing scheme and ensure on-demand resource scheduling. Next, a link quality prediction module periodically sends probe data packets to calculate real-time quality indicators (latency, packet loss rate, etc.) and predicts link quality trends by combining historical data and the ship's latitude and longitude. Finally, a communication scheme generation module dynamically adjusts traffic allocation based on the prediction results and traffic allocation scheme, combined with application requirements (real-time performance, bandwidth) and the link cost characteristics of the current location, generating a ship-to-shore communication scheme.

[0115] This embodiment of the system solves the compatibility problem of heterogeneous links through a link parameter adaptation module, forming a unified communication channel; a dynamic traffic allocation module uses real-time parameters to achieve load balancing and improve resource utilization; a link quality prediction module combines navigation position and historical data to provide forward-looking support for traffic scheduling; and a communication scheme generation module comprehensively considers application requirements and tariff characteristics, balancing communication quality and cost control. Overall, the system achieves full-process optimization of multi-link communication on intelligent ships, improving the real-time compliance rate of high-priority services, reducing link resource waste, and effectively enhancing the stability, economy, and scenario adaptability of ship-to-shore communication.

[0116] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0117] In one embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the intelligent ship network optimization method, system, device, and medium based on multi-link convergence as described above.

[0118] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0119] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The components described as separate parts may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this disclosure according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0120] The above-described embodiments are merely illustrative of several implementation methods of the embodiments of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the embodiments of this application, and these modifications and improvements all fall within the protection scope of the embodiments of this application.

Claims

1. A method for optimizing intelligent ship networks based on multi-link convergence, characterized in that, The method includes: The system obtains the connection parameters of various network links required by intelligent ships, constructs an adaptive interactive environment based on the parameters, modifies the network protocol stack, and integrates multiple physical links to obtain virtual links. Based on the real-time parameters of the virtual links and each physical link, an improved link aggregation control protocol is used to dynamically allocate data traffic, resulting in a load-balanced traffic allocation scheme. The system periodically sends probe data packets to each link and calculates real-time link quality indicators. It then combines historical link quality data with the ship's current latitude and longitude position to predict the link quality change trend and obtain the link quality prediction results. Based on the link quality prediction results and traffic allocation scheme, and taking into account the real-time and bandwidth requirements of different applications and the link tariff characteristics corresponding to the ship's current location, the traffic allocation is dynamically adjusted to obtain the ship-to-shore communication scheme.

2. The method according to claim 1, characterized in that, The process of obtaining connection parameters for various network links required by the intelligent ship, constructing an adaptive interactive environment based on the parameters to modify the network protocol stack, and integrating multiple physical links to obtain a virtual link includes: The interface parameter information and communication protocol information of various network links required by intelligent ships are obtained, and the multi-link access devices are adapted to the open source operating system to obtain an interactive environment that can identify multiple types of links. Based on the adapted interactive environment, the network protocol stack of the open-source operating system kernel is customized and modified, and multi-link collaborative scheduling logic is added to obtain an optimized network protocol stack that supports unified access, management and data flow of multiple physical links; the collaborative scheduling logic includes protocol conversion rules for different links, data fragmentation and reassembly algorithms; Based on the optimized network protocol stack and the bandwidth characteristics and latency thresholds of each link in the intelligent ship navigation scenario, the multiple physical links that have been connected are logically grouped and resource pooled to build a virtual link framework with dynamic bandwidth superposition function. The virtual link framework matches the availability of each physical link based on the real-time navigation position of the intelligent ship, and flexibly allocates the bandwidth resources of each physical link to the virtual link according to demand, thereby obtaining a virtual link that has both bandwidth aggregation capability and scenario-based link scheduling capability.

3. The method according to claim 1, characterized in that, The method of dynamically allocating data traffic based on the real-time parameters of the virtual links and each physical link, using an improved link aggregation control protocol, to obtain a load-balanced traffic allocation scheme includes: The operating parameters of the virtual link and each physical link are acquired in real time to obtain real-time quantitative data of link bandwidth, latency, packet loss rate and jitter. Based on the real-time quantized data, an improved link aggregation control protocol is adopted, and the traffic carrying weight coefficient of each physical link is determined through a parameter weighting algorithm. Based on the traffic carrying weight coefficient and the total bandwidth requirement of the virtual link, the traffic allocation ratio of each physical link is calculated to obtain the initial traffic allocation scheme; The initial traffic allocation scheme is load-checked. If the load of a physical link exceeds a preset threshold, the corresponding link allocation ratio is reduced according to the weight coefficient ratio, and some traffic is allocated to the low-load link to obtain the adjusted traffic allocation ratio of each link. A traffic allocation scheme adapted to the multi-link characteristics of intelligent ships is generated based on the traffic allocation ratio of each link. The traffic allocation scheme includes the specific quotas of high-priority traffic and non-high-priority traffic carried by physical links, the traffic switching trigger threshold, and the execution strategy to maximize the bandwidth utilization of low-cost links.

4. The method according to claim 3, characterized in that, The traffic allocation ratio for each physical link is calculated using the following formula: in, Indicates the first Traffic allocation percentage for each physical link Indicates the first The weighting factor of the link bandwidth. Indicates the first The weighting factor for link delay. Indicates the first The weighting coefficients for packet loss rate across links. Indicates the first The weighting coefficient of link jitter, Indicates the first Real-time bandwidth of each link Indicates the first Real-time latency of the link Indicates the first Real-time packet loss rate of the link, Indicates the first Real-time jitter of the link, This indicates the total number of physical links. This indicates the total bandwidth requirement of the virtual link.

5. The method according to claim 1, characterized in that, The process involves periodically sending probe data packets to each link and calculating real-time link quality indicators. By combining historical link quality data with the ship's current latitude and longitude position, the link quality change trend is predicted, resulting in a link quality prediction outcome, including: A probe data packet transmission plan is generated based on the characteristics of the intelligent ship's navigation area and the link type; the data packet transmission frequency in the transmission plan is set differently according to the link's real-time requirements. A probe data packet containing a precise timestamp and link identification information is created according to the sending plan; the probe data packet carries the basic verification fields required for link quality detection. The system receives the probe data packets returned by each link, calculates the link delay by comparing the sending or receiving timestamps, calculates the packet loss rate by counting the number of data packets lost per unit time, calculates the jitter by analyzing the fluctuation range of data packet transmission time, and outputs link quality indicators including delay, packet loss rate, and jitter. Read the link quality indicators and real-time latitude and longitude data of the ship, combine them with the historical link quality characteristics of the same period of latitude and longitude, screen out the key environmental variables that affect the link quality, and generate structured link quality trend prediction data. The link quality trend prediction data is input into the trained link quality prediction model. By learning the correlation between historical link quality change patterns and current environmental variables, the model outputs the change threshold of the link quality indicators for each link in the future preset time period, thus obtaining the link quality prediction result.

6. The method according to claim 1, characterized in that, The process of dynamically adjusting traffic allocation based on the link quality prediction results and traffic allocation scheme, combined with the real-time and bandwidth requirements of different applications and the link tariff characteristics corresponding to the ship's current position, to obtain a ship-to-shore communication scheme includes: Combining the real-time requirement threshold of high-priority applications of intelligent ships, and based on the link quality prediction results and link quality indicators of each link, a weighting algorithm is used to generate traffic allocation weights specific to the high-priority applications. Based on the traffic allocation weights and the real-time available bandwidth data of low-cost links, and combined with the utilization rate of low-cost links in the traffic allocation scheme, a deep reinforcement learning algorithm is used to generate a dynamic adjustment strategy. Based on the dynamic adjustment strategy and historical optimal traffic allocation parameters, an optimized traffic allocation scheme is obtained through a calculation model; the traffic allocation scheme includes the carrying capacity of high and non-high priority traffic on each link; Based on the traffic allocation scheme, extract the latency characteristic data and quality prediction results of each link, filter the links whose latency meets the high priority requirements and whose predicted quality has not deteriorated, and sort them as candidate paths. Combine the bandwidth availability of the candidate paths to determine the low-latency transmission path. Based on the link characteristics of the low-latency transmission path and the carrying capacity of the path in the traffic allocation scheme, the ship communication data transmission frequency and data packet size are adjusted to generate a ship-shore communication scheme that balances high-priority real-time transmission and efficient resource utilization.

7. The method according to claim 6, characterized in that, The computational model is expressed by the following formula: in, Indicates the first Link assigned to the first Traffic quota, Indicates the first Traffic allocation weights for each link Indicates the first The real-time available bandwidth of each link. This indicates the total number of physical links. This represents the historical optimal parameter correction coefficient. , Indicates the first The link carried the first in the historical scheme The optimal proportion of traffic of this type Indicates the first Total demand for traffic types , This represents the link quality prediction correction factor.

8. An intelligent ship network monitoring system based on multi-link convergence, characterized in that, The system includes: The link parameter adaptation module is used to obtain the connection parameters of various network links required by the intelligent ship, build an adaptation interaction environment based on the parameters to modify the network protocol stack, and integrate multiple physical links to obtain a virtual link. The traffic dynamic allocation module is used to dynamically allocate data traffic based on the real-time parameters of the virtual link and each physical link, using an improved link aggregation control protocol, to obtain a load-balanced traffic allocation scheme. The link quality prediction module is used to periodically send probe data packets to each link and calculate real-time link quality indicators. It combines historical link quality data with the ship's current navigation latitude and longitude position to predict the link quality change trend and obtain the link quality prediction result. The communication scheme generation module is used to dynamically adjust the traffic allocation to obtain a ship-to-shore communication scheme based on the link quality prediction results and traffic allocation scheme, combined with the real-time and bandwidth requirements of different applications and the link tariff characteristics corresponding to the ship's current position.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images