SLA level monitoring system of ocean vessel navigation system

Abstract

The invention relates to an SLA level monitoring system of an ocean vessel navigation system, belongs to the technical field of remote monitoring, and solves the problem that the SLA level of a navigation system service provider cannot be accurately and reliably monitored in the prior art. The system comprises a ship end acquisition module used for acquiring navigation state data of an ocean vessel navigation system to form SLA service data; the communication module is used for transmitting the SLA service data to a shore-end monitoring center by adopting an optimal link based on the ocean vessel position and meteorological information; and the shore end monitoring center is used for receiving the SLA service data sent by the communication module and calculating the SLA service level of the ocean vessel navigation system according to the SLA service data. And reliable and stable SLA service monitoring and evaluation are realized.

Description

Technical Field

[0001] The present invention relates to the field of remote monitoring technology, and in particular to a SLA level monitoring system for an ocean-going ship navigation system. Background Art

[0002] With the development trend of informatization and intelligence in the ocean shipping industry, ships use sensors, communications, the Internet of Things, the Internet and other technical means to automatically sense and obtain information and data on the ship itself, the marine environment, logistics, ports, etc., and use computer technology, automatic control technology and big data processing and analysis technology to achieve intelligent operations in ship navigation, management, maintenance, cargo transportation, etc. At present, ship-to-shore communication is mainly carried out through ship satellites, which is unstable and not timely, and there are certain difficulties in ship-to-shore communication.

[0003] SLA, or Service Level Agreement, refers to the service commitment provided by the system service provider to the customer. The agreement focuses on the timeliness and stability of the service. Therefore, when there are difficulties in ship-to-shore communication, how the service operator can provide an SLA recognized by the customer is an inevitable problem and challenge that needs to be solved. At present, ship-to-shore communication is mainly carried out through ship satellites, which has the phenomenon of unstable communication and low timeliness, and ship-to-shore communication has certain difficulties;

[0004] Currently, ocean-going ship system service providers can only determine SLA by analyzing and calculating the data transmitted by ships to shore. However, the amount of data transmitted from ocean-going ships to shore is large, and due to the instability and untimeliness of satellite communications, data loss or abnormalities may occur. Therefore, it is impossible to accurately and reliably monitor the SLA level of navigation system service providers. Summary of the invention

[0005] In view of the above analysis, an embodiment of the present invention aims to provide a SLA level monitoring system for an ocean-going ship navigation system, so as to solve the problem that the SLA level of the navigation system service provider cannot be accurately and reliably monitored in the prior art.

[0006] On the one hand, an embodiment of the present invention provides an SLA level monitoring system for an ocean-going ship navigation system, comprising:

[0007] The ship-side acquisition module is used to collect the navigation status data of the ocean-going ship's navigation system to form SLA service data;

[0008] A communication module, used to transmit the SLA service data to a shore monitoring center using an optimal link based on the ocean-going vessel's position and weather information;

[0009] The shore monitoring center is used to receive the SLA service data sent by the communication module and calculate the SLA service level of the ocean-going ship's navigation system based on the SLA service data.

[0010] Based on the further improvement of the above system, the optimal link is determined in the following way:

[0011] Calculate the position of the ocean-going ship at each moment in the next unit time according to the current position and speed of the ocean-going ship;

[0012] According to the position of the ocean-going ship at each moment in the next unit time and the meteorological information of the position of the ocean-going ship at each moment in the next unit time, the signal strength and communication quality information of each satellite link at each moment in the next unit time are predicted based on the trained neural network model;

[0013] The optimal link for the next unit time is determined based on the signal strength, communication quality information, link energy consumption and transmission cost of each satellite link at each moment in the next unit time.

[0014] Based on the further improvement of the above system, the optimal link for the next unit time is determined based on the signal strength, communication quality information, link energy consumption and transmission cost of each satellite link at each moment in the next unit time, including:

[0015] Calculate the availability of each satellite link at each moment in the next unit time based on the signal strength, communication quality information, link energy consumption and transmission cost of each satellite link at each moment in the next unit time;

[0016] The availability of each satellite link in the next unit time is calculated based on the availability of each satellite link at each moment in the next unit time, and the communication link with the highest availability is selected as the optimal link.

[0017] Based on the further improvement of the above system, the communication quality information includes: bit error rate, throughput, packet loss rate and delay;

[0018] The following formula is used to calculate the availability of each satellite link at each moment in the next unit time:

[0019] A i =α1×SS+α2(ω1×TP-ω2×BER-ω3×LT-ω4×PLR)-α3

[0020] ×EC-α4×TC

[0021] Where SS represents the signal strength of the satellite link, TP represents the throughput of the satellite link, BER represents the bit error rate of the satellite link, LT represents the delay of the satellite link, PLR represents the packet loss rate of the satellite link, EC represents the link energy consumption of the satellite link, TC represents the transmission cost, α1, α2, α3 and α4 represent the weights of signal strength, communication quality, link energy consumption and transmission cost respectively, ω1, ω2, ω3 and ω4 represent the weights of throughput, bit error rate, delay and packet loss rate respectively, A i represents the availability of the satellite link at the i-th moment.

[0022] Based on further improvement of the above system, the average value of the availability of each satellite link at each moment in the next unit time is calculated to obtain the availability of each satellite link in the next unit time.

[0023] Based on the further improvement of the above system, the neural network model for predicting signal strength and communication quality is trained in the following way:

[0024] For each satellite link, the location information, meteorological information, signal strength and communication quality information of the ocean-going ship at the historical sailing time are obtained, and a training sample set corresponding to the satellite link is constructed; a multi-task regression neural network model corresponding to the satellite link is constructed, and the constructed multi-task regression neural network model is trained based on the training sample set corresponding to the satellite link to obtain a trained neural network model for predicting the signal strength and communication quality of the satellite link.

[0025] Based on the further improvement of the above system, the following formula is used to calculate the training loss of the multi-task regression neural network model:

[0026]

[0027]

[0028] Among them, L ij represents the loss of the jth regression data of the i-th sample in the current training batch, β j represents the loss weight of the jth regression data, y ij represents the true value of the jth regression data of the i-th sample in the current training batch, represents the predicted value of the jth regression data of the i-th sample in the current training batch, δ j represents the threshold of the jth regression data, γ represents the penalty coefficient, λ represents the decay rate, m represents the number of predicted data, and n represents the number of samples in the current training batch.

[0029] Based on the further improvement of the above system, the following formula is used to calculate the threshold of the jth regression data:

[0030] Calculate the median regression error ME of the j-th regression data of the training sample of the previous training batch j ;

[0031] Calculate the absolute deviation of the median regression error of the jth regression data of the training sample of the previous training batch

[0032] Calculate the dynamic threshold δ based on the absolute deviation from the median j =c×MAD j ; where c is the magnification factor, median(·) indicates the median calculation, represents the true value of the jth regression data of the i-th sample in the previous training batch, Represents the predicted value of the jth regression data of the ith sample in the previous training batch.

[0033] Based on the further improvement of the above system, the multi-task regression neural network model includes an input layer, a shared feature extraction layer, a task feature extraction layer and an output layer;

[0034] The input layer is used to receive input sample data and output initial feature data; the shared feature extraction layer is used to extract multi-task shared features based on the initial feature data; the task feature extraction layer includes multiple groups of fully connected layers, each group of fully connected layers corresponds to a regression task; and is used to extract features corresponding to each regression task;

[0035] The output layer includes a plurality of output heads, each of which corresponds to a group of fully connected layers of the task feature extraction layer, and is used to output the regression task result according to the features corresponding to each regression task;

[0036] The regression tasks include a signal strength regression task, a bit error rate regression task, a throughput regression task, a packet loss rate regression task and a delay regression task.

[0037] Based on the further improvement of the above system, the SLA service data includes navigation data;

[0038] The position of the ocean-going ship at each moment in the next unit time is calculated in the following way:

[0039] The estimated arrival time of each track point after the current ocean-going vessel's position in the navigation data is calculated based on the navigation data in the SLA service data, the current position and speed of the ocean-going vessel; for each moment in the next unit time, if the estimated arrival time is the same as the estimated arrival time of the track point, the position of the ocean-going vessel at that moment is the position of the track point with the same estimated arrival time; otherwise, the position of the ocean-going vessel at that moment is calculated based on the positions of the two track points with the estimated arrival times closest to that moment.

[0040] Compared with the prior art, the SLA level monitoring system of the ocean-going vessel's navigation system provided in this embodiment collects the navigation status data of the navigation system to form SLA data through the ship-side acquisition module, and then determines the optimal link based on the position of the ocean-going vessel and meteorological information through the communication module, and sends the SLA service data to the shore-side monitoring center through the optimal link, thereby stably sending the SLA service data to the shore-side monitoring center, which is convenient for the shore-side monitoring center to accurately and stably monitor and reliably evaluate the SLA level of the navigation system.

[0041] In the present invention, the above-mentioned technical solutions can also be combined with each other to achieve more preferred combination solutions. Other features and advantages of the present invention will be described in the subsequent description, and some advantages can become obvious from the description, or can be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained through the contents particularly pointed out in the description and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The accompanying drawings are only used for the purpose of illustrating specific embodiments and are not to be considered as limiting the present invention. In the entire drawings, the same reference symbols represent the same components;

[0043] Figure 1 The block diagram of the SLA level monitoring system of the ocean-going ship navigation system according to the embodiment of the present invention. DETAILED DESCRIPTION

[0044] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.

[0045] A specific embodiment of the present invention discloses a SLA level monitoring system for an ocean-going ship navigation system, such as Figure 1 As shown, including:

[0046] The ship-side acquisition module is used to collect the navigation status data of the ocean-going ship's navigation system to form SLA service data;

[0047] A communication module, used to transmit the SLA service data to a shore monitoring center using an optimal link based on the ocean-going vessel's position and weather information;

[0048] The shore monitoring center is used to receive the SLA service data sent by the communication module and calculate the SLA service level of the ocean-going ship's navigation system based on the SLA service data.

[0049] During implementation, a ship-side acquisition module is installed on the ocean-going ship with a navigation system installed. The navigation system is used to provide meteorological, hydrological, positioning and navigation services for the ship. The ship-side acquisition module collects the navigation status data of the ocean-going ship's navigation system as the navigation system's meteorological service, hydrological service, positioning and navigation service data, etc., which are used for subsequent service level evaluation and monitoring of the services provided.

[0050] During implementation, multiple antennas are usually installed on the ocean-going vessel to connect to different satellite links, such as the Asia-Pacific network VSAT, the international network VAST, etc. The communication stability and efficiency of satellite links are different in different locations and weather conditions, and the traffic charges of different links are different.

[0051] Compared with the prior art, the SLA level monitoring system of the ocean-going vessel's navigation system provided in this embodiment collects the navigation status data of the navigation system to form SLA data through the ship-side acquisition module, and then determines the optimal link based on the position of the ocean-going vessel and meteorological information through the communication module, and sends the SLA service data to the shore-side monitoring center through the optimal link, thereby stably sending the SLA service data to the shore-side monitoring center, which is convenient for the shore-side monitoring center to accurately and stably monitor and reliably evaluate the SLA level of the navigation system.

[0052] During implementation, the ship-side acquisition module collects the navigation status data provided by the navigation system in real time. The navigation status data at each moment is encapsulated as SLA service data, including the SLA service data identifier UUID, the ocean-going ship ID, timestamp and the navigation status data provided by the navigation system.

[0053] The ship-side acquisition module sends the collected SLA service data to the communication module. The communication module encrypts and compresses the received SLA service data, stores it and puts it into the transmission queue, and sends it to the shore monitoring center through the optimal link in turn.

[0054] The communication stability and efficiency of satellite links vary under different locations and weather conditions. In order to ensure stable transmission of SLA service data, the communication module uses the following method to determine the optimal link:

[0055] S11, calculating the position of the ocean-going ship at each moment in the next unit time according to the current position and speed of the ocean-going ship;

[0056] S12, according to the position of the ocean-going ship at each moment in the next unit time and the meteorological information of the position of the ocean-going ship at each moment in the next unit time, the signal strength and communication quality information of each satellite link at each moment in the next unit time are predicted based on the trained neural network model;

[0057] S13. Determine the optimal link for the next unit time based on the signal strength, communication quality information, link energy consumption and transmission cost of each satellite link at each moment in the next unit time.

[0058] During implementation, the unit time can be set according to the navigation speed and navigation route, for example, it can be 12 hours. Each moment in the unit time can be set according to the SLA service data transmission frequency requirement, for example, if the SLA service data is sent every 10 minutes, then every 10 minutes can be set as a moment.

[0059] During implementation, within each unit time, the position of the ocean-going ship at each moment in the next unit time is predicted based on the current position, speed and heading of the ocean-going ship, so as to determine the reliable satellite link for the content of the next unit time, and select the optimal link for data transmission, thereby improving the stability and efficiency of transmission.

[0060] During implementation, since the SLA service data contains navigation information, that is, the sailing route planned by the navigation system, the position of the ocean-going ship at each moment in the next unit time can be calculated based on its planned trajectory points and the current position and speed of the ocean-going ship.

[0061] During implementation, the position of the ocean-going vessel at each moment in the next unit time is calculated in the following way:

[0062] The estimated arrival time of each track point after the current ocean-going vessel's position in the navigation data is calculated based on the navigation data in the SLA service data, the current position and speed of the ocean-going vessel; for each moment in the next unit time, if the estimated arrival time is the same as the estimated arrival time of the track point, the position of the ocean-going vessel at that moment is the position of the track point with the same estimated arrival time; otherwise, the position of the ocean-going vessel at that moment is calculated based on the positions of the two track points with the estimated arrival times closest to that moment.

[0063] During implementation, since the latest SLA service data is closest to the current navigation status of the ocean-going vessel, the estimated arrival time of each track point after the current position of the ocean-going vessel is first calculated based on the navigation data in the latest SLA service data, the current position and speed of the ocean-going vessel.

[0064] During implementation, the first track point p1(x1,y1) after the current position in the navigation data can be found based on the current ship's positioning data. Assuming that the current position of the ship is p0(x0,y0) and the speed is v, the estimated arrival time of the first track point can be calculated. t0 is the current time. The estimated arrival time of the second trajectory point p2(x2,y2) can be calculated based on the estimated arrival time t1 of the first trajectory point. By analogy, the estimated arrival time t of each trajectory point can be calculated i .

[0065] For each moment in the next unit time, if it is the same as the estimated arrival time of a certain track point, the position of the ship is directly obtained as the position of the track point. If it is not the same as the estimated arrival time of any track point, find the two nearest estimated arrival times t i and t j , the corresponding trajectory point is p i (x i ,y i ) and p j (x j ,y j ), then according to the formula The position of the ocean-going vessel at that moment is calculated.

[0066] After calculating the position of the ocean-going ship at each moment in the next unit time, the signal strength and communication quality corresponding to each moment are predicted based on the trained neural network model according to the satellite orbit parameter information of different satellite links, the meteorological information at each moment in the next unit time, and the position of the ocean-going ship at each moment in the next unit time, thereby predicting the signal strength and communication quality information of each satellite link at each moment in the next unit time of the ocean-going ship, and determining the optimal link.

[0067] In order to accurately predict the signal lightness and communication quality information of the satellite link, the neural network model is trained in the following way:

[0068] For each satellite link, the location information, meteorological information, signal strength and communication quality information of the ocean-going ship at the historical sailing time are obtained, and a training sample set corresponding to the satellite link is constructed; a multi-task regression neural network model corresponding to the satellite link is constructed, and the constructed multi-task regression neural network model is trained based on the training sample set corresponding to the satellite link to obtain a trained neural network model for predicting the signal strength and communication quality of the satellite link.

[0069] During implementation, a corresponding neural network model is trained for each satellite link. Satellite orbit parameters include semi-major axis, eccentricity, orbit inclination, right ascension of ascending node, perigee angular distance, perigee angular distance, true anomaly, period, rate, and ascending node time. These orbit parameters are usually provided by satellite manufacturers, launch service providers, and space agencies, and can be obtained through space orbit analysis software and database queries.

[0070] The location information of the ocean-going vessel at the historical sailing time can be obtained through the positioning module during the historical sailing, such as GPS positioning, AIS positioning, satellite positioning, etc.

[0071] Since signals will attenuate in the atmosphere and rainfall and clouds will also affect signal attenuation, meteorological information includes temperature, air pressure, rainfall, humidity, cloud thickness and cloud height.

[0072] During implementation, the signal strength may be measured using an existing signal strength test tool.

[0073] Communication quality is a concept that needs to be measured through multi-dimensional information. The communication quality information of the present invention includes bit error rate, throughput, packet loss rate and delay.

[0074] The bit error rate is the ratio of the number of received erroneous bits to the total number of transmitted bits. Throughput refers to the amount of data successfully transmitted per unit time, usually expressed in bits per second (bps). The packet loss rate refers to the ratio of the number of packets lost during data transmission to the total number of packets sent. Latency refers to the time required for data to travel from the sender to the receiver, including propagation delay, processing delay, and queuing delay. The bit error rate, throughput, packet loss rate, and latency can be used to comprehensively measure the quality of communication and provide a basis for determining the optimal link.

[0075] During implementation, the bit error rate, throughput, packet loss rate and latency can be tested using existing testing tools.

[0076] During implementation, the positioning, location information, and meteorological information of the ocean-going ships' historical voyages are used as sample input data, and the corresponding signal strength, bit error rate, throughput, packet loss rate, and delay of a certain satellite link are used as labels to construct training samples of the satellite link.

[0077] During implementation, a multi-task regression neural network model is constructed for each satellite link, wherein the multi-task regression neural network model includes an input layer, a shared feature extraction layer, a task feature extraction layer and an output layer;

[0078] The input layer is used to receive input sample data and output initial feature data;

[0079] The shared feature extraction layer is used to extract multi-task shared features based on the initial feature data;

[0080] The task feature extraction layer includes multiple groups of fully connected layers, each group of fully connected layers corresponds to a regression task, and is used to extract features corresponding to each regression task;

[0081] The output layer includes a plurality of output heads, each of which corresponds to a group of fully connected layers of the task feature extraction layer, and is used to output the regression task result according to the features corresponding to each regression task;

[0082] The regression tasks include a signal strength regression task, a bit error rate regression task, a throughput regression task, a packet loss rate regression task and a delay regression task.

[0083] Specifically, the following formula is used to calculate the training loss of the multi-task regression neural network model:

[0084]

[0085]

[0086] Among them, L ij represents the loss of the jth regression data of the i-th sample in the current training batch, β j represents the loss weight of the jth regression data, y ij represents the true value of the jth regression data of the i-th sample in the current training batch, represents the predicted value of the jth regression data of the i-th sample in the current training batch, δ j represents the threshold of the jth regression data, γ represents the penalty coefficient, λ represents the decay rate, m represents the number of predicted data, and n represents the number of samples in the current training batch.

[0087] The regression data are signal strength, bit error rate, throughput, packet loss rate and latency.

[0088] When implemented, when the sample error is small, that is, The squared error is used to achieve sensitivity to small errors. When the sample error is large, the linear error is used to reduce the impact of outliers on the loss. In order to further increase the multiplication of outliers, when the error is large, a penalty term is added to the loss. This imposes a greater penalty on large errors, thereby improving the convergence efficiency of training.

[0089] In order to improve the adaptability of the loss function, the threshold of the j-th regression data adopts a dynamic threshold. Specifically, the dynamic threshold of the j-th regression data is calculated using the following formula:

[0090] Calculate the median regression error ME of the j-th regression data of the training sample of the previous training batch j ;

[0091] Calculate the absolute deviation of the median regression error of the jth regression data of the training sample of the previous training batch

[0092] Calculate the dynamic threshold δ based on the absolute deviation from the median j =c×MAD j ; where c is the magnification factor, median(·) indicates the median calculation, represents the true value of the jth regression data of the i-th sample in the previous training batch, Represents the predicted value of the jth regression data of the ith sample in the previous training batch.

[0093] During implementation, the threshold of the current training batch can be dynamically determined based on the error of the previous batch. In order to improve the robustness of the threshold, the median error of the training samples of the previous batch is first calculated. Then according to the formula The absolute deviation of the median regression error is calculated and then amplified by a magnification factor to obtain an adaptive threshold to improve the efficiency of model training. In implementation, the magnification factor can be 1.5.

[0094] When the model reaches the training times or the predetermined training accuracy, the training is stopped to obtain a trained neural network model for predicting signal strength and communication quality.

[0095] During implementation, the position information and weather information of the ocean-going ship at each moment in the next unit time are input into the trained neural network model corresponding to each satellite link, so as to predict the signal strength and communication quality information of each satellite link at each moment.

[0096] Then, the optimal link for the next unit time is determined based on the signal strength, communication quality information, link energy consumption and transmission cost of each satellite link at each moment in the next unit time. Specifically, it includes:

[0097] S131, calculating the availability of each satellite link at each moment in the next unit time based on the signal strength, communication quality information, link energy consumption and transmission cost of each satellite link at each moment in the next unit time;

[0098] S132. Calculate the availability of each satellite link in the next unit time based on the availability of each satellite link at each moment in the next unit time, and select the communication link with the highest availability as the optimal link.

[0099] In order to comprehensively consider and select the optimal satellite link, the availability of each satellite link at each moment in the next unit time is calculated based on the signal strength, communication quality information, link energy consumption and transmission cost of each satellite link at each moment in the next unit time. Specifically, the following formula is used to calculate the availability of each satellite link at each moment in the next unit time:

[0100] A i =α1×SS+α2(ω1×TP-ω2×BER-ω3×LT-ω4×PLR)-α3

[0101] ×EC-α4×TC

[0102] Where SS represents the signal strength of the satellite link, TP represents the throughput of the satellite link, BER represents the bit error rate of the satellite link, LT represents the delay of the satellite link, PLR represents the packet loss rate of the satellite link, EC represents the link energy consumption of the satellite link, TC represents the transmission cost, α1, α2, α3 and α4 represent the weights of signal strength, communication quality, link energy consumption and transmission cost respectively, ω1, ω2, ω3 and ω4 represent the weights of throughput, bit error rate, delay and packet loss rate respectively, A i represents the availability of the satellite link at the i-th moment.

[0103] When implemented, the link energy consumption of the satellite link may be the energy consumption of the antenna.

[0104] During implementation, the sum of α1, α2, α3 and α4 should be 1, and the sum of ω1, ω2, ω3 and ω4 should be 1. The values ​​of each weight parameter can be set according to actual conditions. By comprehensively considering signal strength, communication quality, link energy consumption and transmission cost, the optimal satellite link can be determined through comprehensive evaluation.

[0105] After obtaining the availability of each satellite link at each moment in the next unit time, the average value of the availability of each satellite link at each moment in the next unit time is calculated to obtain the availability of each satellite link in the next unit time.

[0106] By evaluating the availability of each communication link in the next unit time, the communication link with the highest availability is selected as the optimal link as the link for transmitting SLA service data in the next unit time, thereby improving the transmission stability of SLA service data.

[0107] The shore monitoring center receives the SLA and decrypts and decompresses it, and calculates the SLA service level of the ocean-going ship's navigation system based on the SLA service data. During implementation, for example, the accuracy of the positioning service can be calculated based on the positioning data, and the accuracy and timeliness of the meteorological service can be calculated based on the meteorological data, thereby monitoring and evaluating the service level of the ocean-going ship's navigation system.

[0108] During implementation, in order to further improve transmission reliability, the shore monitoring center returns the UUID information of the received SLA service to the communication module after receiving the SLA service data. The communication module considers the transmission successful when receiving the confirmation information; if the reception confirmation of the sent SLA service data is not received within the first threshold time, the SLA service data is resent. Thus, the transmission reliability is improved.

[0109] Those skilled in the art will appreciate that all or part of the processes of the above-mentioned embodiments can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, wherein the computer-readable storage medium is a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0110] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A SLA level monitoring system for an ocean-going ship navigation system, characterized in that: include: The ship-side acquisition module is used to collect the navigation status data of the ocean-going ship's navigation system to form SLA service data; A communication module, used to transmit the SLA service data to a shore monitoring center using an optimal link based on the ocean-going vessel's position and weather information; The shore monitoring center is used to receive the SLA service data sent by the communication module and calculate the SLA service level of the ocean-going ship's navigation system based on the SLA service data.

2. The SLA level monitoring system for an ocean-going ship navigation system according to claim 1, characterized in that: The optimal link is determined in the following way: Calculate the position of the ocean-going ship at each moment in the next unit time according to the current position and speed of the ocean-going ship; According to the position of the ocean-going ship at each moment in the next unit time and the meteorological information of the position of the ocean-going ship at each moment in the next unit time, the signal strength and communication quality information of each satellite link at each moment in the next unit time are predicted based on the trained neural network model; The optimal link for the next unit time is determined based on the signal strength, communication quality information, link energy consumption and transmission cost of each satellite link at each moment in the next unit time.

3. The SLA level monitoring system of the ocean-going ship navigation system according to claim 2, characterized in that: Determine the optimal link for the next unit time based on the signal strength, communication quality information, link energy consumption and transmission cost of each satellite link at each moment in the next unit time, including: Calculate the availability of each satellite link at each moment in the next unit time based on the signal strength, communication quality information, link energy consumption and transmission cost of each satellite link at each moment in the next unit time; The availability of each satellite link in the next unit time is calculated based on the availability of each satellite link at each moment in the next unit time, and the communication link with the highest availability is selected as the optimal link.

4. The SLA level monitoring system for an ocean-going ship navigation system according to claim 3, characterized in that: The communication quality information includes: bit error rate, throughput, packet loss rate and delay; The following formula is used to calculate the availability of each satellite link at each moment in the next unit time: A i =α1×SS+α2(ω1×TP-ω2×BER-ω3×LT-ω4×PLR)-α3 ×EC-α4×TC Where SS represents the signal strength of the satellite link, TP represents the throughput of the satellite link, BER represents the bit error rate of the satellite link, LT represents the delay of the satellite link, PLR represents the packet loss rate of the satellite link, EC represents the link energy consumption of the satellite link, TC represents the transmission cost, α1, α2, α3 and α4 represent the weights of signal strength, communication quality, link energy consumption and transmission cost respectively, ω1, ω2, ω3 and ω4 represent the weights of throughput, bit error rate, delay and packet loss rate respectively, A i represents the availability of the satellite link at the i-th moment.

5. The SLA level monitoring system for an ocean-going vessel navigation system according to claim 3, characterized in that: The availability of each satellite link in the next unit time is obtained by calculating the average value of the availability of each satellite link at each moment in the next unit time.

6. The SLA level monitoring system for the ocean-going ship navigation system according to claim 5, characterized in that: The neural network model for predicting signal strength and communication quality is trained in the following way: For each satellite link, the location information, weather information, signal strength and communication quality information of the ocean-going ship at the historical sailing time are obtained to construct a training sample set corresponding to the satellite link; A multi-task regression neural network model corresponding to the satellite link is constructed, and the constructed multi-task regression neural network model is trained based on the training sample set corresponding to the satellite link to obtain a trained neural network model for predicting the signal strength and communication quality of the satellite link.

7. The SLA level monitoring system for an ocean-going vessel navigation system according to claim 6, characterized in that: The training loss of the multi-task regression neural network model is calculated using the following formula: Among them, L ij represents the loss of the jth regression data of the i-th sample in the current training batch, β j represents the loss weight of the jth regression data, y ij represents the true value of the jth regression data of the i-th sample in the current training batch, represents the predicted value of the jth regression data of the i-th sample in the current training batch, δ j represents the threshold of the jth regression data, γ represents the penalty coefficient, λ represents the decay rate, m represents the number of predicted data, and n represents the number of samples in the current training batch.

8. The SLA level monitoring system for the ocean-going ship navigation system according to claim 7, characterized in that: The threshold of the jth regression data is calculated using the following formula: Calculate the median regression error ME of the j-th regression data of the training sample of the previous training batch j ; Calculate the absolute deviation of the median regression error of the jth regression data of the training sample of the previous training batch Calculate the dynamic threshold δ based on the absolute deviation from the median j =c×MAD j ; where c is the magnification factor, median(·) indicates the median calculation, represents the true value of the jth regression data of the i-th sample in the previous training batch, Represents the predicted value of the jth regression data of the ith sample in the previous training batch.

9. The SLA level monitoring system for an ocean-going vessel navigation system according to claim 6, characterized in that: The multi-task regression neural network model includes an input layer, a shared feature extraction layer, a task feature extraction layer and an output layer; The input layer is used to receive input sample data and output initial feature data; The shared feature extraction layer is used to extract multi-task shared features based on the initial feature data; The task feature extraction layer includes multiple groups of fully connected layers, each group of fully connected layers corresponds to a regression task, and is used to extract features corresponding to each regression task; The output layer includes a plurality of output heads, each of which corresponds to a group of fully connected layers of the task feature extraction layer, and is used to output the regression task result according to the features corresponding to each regression task; The regression tasks include a signal strength regression task, a bit error rate regression task, a throughput regression task, a packet loss rate regression task and a delay regression task.

10. The SLA level monitoring system of the ocean-going ship navigation system according to claim 2, characterized in that: The SLA service data includes navigation data; The position of the ocean-going ship at each moment in the next unit time is calculated in the following way: Calculate the estimated arrival time of each track point after the current position of the ocean-going vessel in the navigation data based on the navigation data in the SLA service data, the current position and speed of the ocean-going vessel; For each moment in the next unit time, if it is the same as the estimated arrival time of the trajectory point, the position of the ocean-going ship at that moment is the position of the trajectory point with the same estimated arrival time; otherwise, the position of the ocean-going ship at that moment is calculated based on the positions of the two trajectory points with the estimated arrival times closest to that moment.

Images