Shore-end multimode link decision-making method based on communication service quality

Abstract

The invention discloses a shore-end multimode link decision-making method based on communication service quality, and relates to the technical field of shore ship communication. The method comprises the following steps: acquiring a first service data packet, a second service data packet, a third service data packet and data packet characteristics from a shore ship communication service; inputting the data packet features into a first agent, wherein the first agent is used for maximizing the transmission gain of the scheduled data packet and minimizing the queue length; attribute parameters of different networks are obtained, the attribute parameters of the different networks and the communication service data packet type output by the first intelligent agent are input into a second intelligent agent, and the second intelligent agent is used for obtaining the action output according to the second intelligent agent to make a link decision according to the minimum network switching frequency. According to the method, the channel state, the network load and the service QoS demand are sensed in real time, the intelligent decision algorithm is combined, service data packet scheduling is carried out, the optimal communication mode is dynamically selected, and therefore efficient adaptation of communication resources and service demands is achieved.

Description

Technical Field

[0001] This invention relates to the field of ship-to-shore communication technology, and more specifically, to a shore-based multi-mode link decision-making method based on communication service quality. Background Technology

[0002] The concept of autonomous surface vessels (ASDs), proposed by the International Maritime Organization (IMO) in 2018, has accelerated the automation and intelligentization of ships and promoted the innovative development of integrated ship-shore communication equipment. Against the backdrop of rapid development in intelligent shipping, ship-shore communication systems face more stringent requirements for stability and high throughput. Faced with increasingly complex business scenarios, traditional single communication modes are no longer sufficient to meet the needs of modern shipping. Specifically: traditional VDES (Vehicle Dedicated Surface Equipment) possesses dedicated maritime frequency bands and good maritime coverage, but is susceptible to bandwidth limitations in high-density communication scenarios; while Mesh self-organizing networks offer dynamic topology adaptation capabilities and high throughput advantages, they are more sensitive to node distribution and mobility; 5G networks can provide ultra-low latency and high bandwidth support, but their maritime coverage is limited.

[0003] To adapt to diverse ship-shore communication needs, heterogeneous converged communication network architectures are gradually becoming a development trend. However, ship-shore communication services not only have differentiated priorities, but their quality of service requirements also vary significantly. This diversity of service characteristics makes achieving dynamic resource allocation and optimal transmission path selection in a multi-network environment a key challenge. Especially when network status fluctuates due to complex sea conditions, ensuring the communication reliability of high-priority services while also considering the transmission efficiency of ordinary services directly affects the overall effectiveness of ship-shore collaboration. Therefore, achieving intelligent network selection and resource scheduling based on service characteristics has become a core issue in improving the service capabilities of integrated ship-shore communication systems.

[0004] Existing technologies have proposed various scheduling mechanisms and heterogeneous network selection algorithms. Resource scheduling typically employs deadline-based or priority-based scheduling mechanisms; for heterogeneous network selection, traditional algorithms such as the analytic hierarchy process (AHP) or intelligent algorithms based on deep learning are commonly used. Resource scheduling and heterogeneous network selection are closely related. Resource scheduling focuses on micro-level resource allocation to ensure efficient data packet transmission, while heterogeneous network selection addresses macro-level access decisions. Together, they can improve end-to-end performance. However, current research on resource scheduling and network selection suffers from significant single-dimensional limitations. This single-dimensional research mechanism is insufficient to effectively address the complex scenarios of concurrent scheduling of multiple services and dynamic switching between multiple networks on ships. Summary of the Invention

[0005] In view of the shortcomings of existing technologies, this invention proposes a shore-based multi-mode link decision-making method based on communication service quality, applied to ship-shore integrated communication equipment. This method achieves efficient matching of communication resources and service requirements by real-time sensing of channel status, network load, and service QoS requirements, combined with intelligent decision-making methods, to schedule service data packets and dynamically select the optimal communication mode.

[0006] The technical means employed in this invention are as follows: A shore-side multimode link decision-making method based on communication service quality includes the following steps: From shore-to-ship communication services, a first service data packet representing life safety services, a second service data packet representing ship operation services, and a third service data packet representing commercial communication services are obtained, with the priority of the first service data packet, the second service data packet, and the third service data packet decreasing in that order. Obtain the data packet characteristics of the first service data packet, the second service data packet, and the third service data packet, wherein the data packet characteristics include queue length, priority, and waiting time; The data packet characteristics are input into a first agent, which is used to optimize data packet scheduling by maximizing data packet transmission gain and minimizing queue length. The state space of the first agent includes data packet characteristics, the action space includes different communication service data packet types, and the reward function is the difference between data packet transmission gain and queue length. The attribute parameters of different networks are obtained, including latency and signal-to-noise ratio. The attribute parameters of different networks and the communication service data packet type output by the first agent are input into the second agent. The second agent is used to optimize network selection by minimizing the number of network switching. The state space of the second agent includes the communication service data packet type and the attribute parameters of the network, the action space includes candidate networks and whether to switch, and the reward function is the difference between the network evaluation score and the switching penalty. Make link decisions based on the actions output by the second agent.

[0007] Furthermore, the formula for calculating the data packet transmission gain is as follows:

[0008] in, b represents the data packet transmission gain, and b represents ship-to-shore communication services. This refers to businesses related to human safety. This refers to ship operation services. Indicates commercial communication services. For positive gain weights, Prioritizing different service data packets This refers to the relaxation of the waiting time.

[0009] Furthermore, the first The formula for calculating the queue length of different service data packets in each time slot is as follows:

[0010] in, For the first The length of the data packet queue for each time slot. and For different service data packets The number of data packets transmitted and received in each time slot.

[0011] Furthermore, the formula for calculating network latency is as follows:

[0012] in, Indicates communication method Total latency, , Communication methods Transmission delay and propagation delay Furthermore, the formula for calculating the network signal-to-noise ratio is as follows:

[0013] in, Indicates communication method The signal-to-noise ratio, Indicates communication method The transmission power, and Indicates communication method The transmit and receive antenna gain, Indicates communication method Propagation loss, This represents noise power.

[0014] Furthermore, the formula for calculating the network evaluation score is as follows:

[0015] in, This indicates the online evaluation score. and Preset weights for different industries and different network attributes.

[0016] Compared with the prior art, the present invention has the following advantages.

[0017] 1. The shore-based multi-mode link decision-making method proposed in this invention employs an IPPO architecture to jointly optimize packet scheduling and network selection, and achieves efficient decision-making through asynchronously and independently trained agents. The packet scheduling agent dynamically schedules packets based on packet priority, while the network selection agent selects networks for the scheduled packets based on network status and QoS requirements. This scheme can improve system throughput and packet delivery rate, providing a more reliable guarantee for ship-shore communication.

[0018] 2. This invention introduces an invalid action mask into the packet scheduling agent, which improves the scheduling efficiency of packets and effectively shortens the packet waiting time.

[0019] 3. In the network selection agent, the present invention adopts a joint action space to select the optimal network for data packets and decide whether to switch; it introduces an attention mechanism to dynamically adjust feature weights based on current business needs, and automatically assigns higher weights to key features, thereby reducing unnecessary network switching. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a schematic diagram illustrating an application scenario of the shore-side multimode link decision-making method based on communication service quality according to the present invention.

[0022] Figure 2 This is a model architecture diagram of a shore-side multi-mode link decision-making method based on communication service quality according to the present invention. Detailed Implementation

[0023] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0024] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0025] This invention provides a shore-side multimode link decision-making method based on communication service quality, which mainly includes the following steps.

[0026] S1. From the shore-to-ship communication service, obtain the first service data packet representing life safety services, the second service data packet representing ship operation services, and the third service data packet representing commercial communication services, with the priority of the first service data packet, the second service data packet, and the third service data packet decreasing in that order.

[0027] This application, based on relevant standards of the International Association of Lighthouse Authorities (IALA), constructs a ship-to-shore service model, classifies services, and determines the QoS requirements for each service. Ship-to-shore communication service b is broadly categorized into life-saving safety services. Ship operation business Commercial communication services Human safety-related services, including distress alerts and navigation communications, are extremely sensitive to latency; excessive latency can lead to serious safety incidents. Ship operation services, primarily involving meteorological information and hydrological data, have a certain tolerance for latency, but signal interruptions can affect navigation safety, thus requiring high communication quality. Commercial communication services, including cargo information, commercial port services, and entertainment data for passengers or crew, only require a basic, stable connection.

[0028] Different communication services have different QoS requirements. To better measure QoS requirements, this invention introduces a set of indicators. ,in Represents delay. This represents the signal-to-noise ratio (SNR). Different services have different sensitivities to different metrics; life-saving services, for example, are more sensitive to certain indicators. Most sensitive are ship operation businesses. The most sensitive indicator is the balance between the two, while commercial communications businesses pay more attention to the balance between the two indicators.

[0029] As a preferred embodiment of the present invention, this application also constructs a packet scheduling model that includes a scheduling gain model and a service packet queue model. The scheduling gain model constructs a scheduling gain function by quantifying service priority and latency sensitivity. The service packet queue model is used to construct the packet queue length.

[0030] Specifically, the scheduling gain model sets different priorities and time sensitivities for data packets of different service types. Human safety data packets have the highest priority and must be sent with the least possible waiting time; ship operation data packets have medium priority and can tolerate some delay; commercial communication data packets have low priority and are not time-sensitive. When data packets of different service types arrive, they are allocated to the corresponding queues according to their type. Following the "high priority first" principle, data packets in the high-priority queue are processed first. When the high-priority data packet queue is empty, low-priority data packets are scheduled. The arrival time slot of a data packet is recorded, and a waiting time slack is defined for different service types of data packets, as shown in the following formula:

[0031] in, Measuring different types of data packets The remaining tolerable waiting time, Maximum allowed waiting time for different types of data packets and , The maximum waiting time for services related to human safety. The maximum waiting time for services related to human safety. The maximum waiting time for services related to human safety. This represents the time that different types of data packets have been waiting in the queue.

[0032] To meet the time-sensitive requirements of each service type's data packets, we select data packets with the minimum latency slack. Since only one data packet is sent per time slot, different output gains can be obtained after successfully transmitting data packets of different service types:

[0033] when hour, This represents positive gain; the higher the priority, the greater the gain. Among these, For positive gain weights, Prioritize different service data packets; the higher the priority, the better. The larger, that is , Prioritize services related to human safety. Prioritizing ship operation-related businesses Prioritization is applied to commercial communication services. To prevent low-priority data packets from being "starved," a penalty gain is set, i.e., when... hour, This is the penalty gain. The higher the priority, the greater the penalty gain. , Penalties and benefits for businesses related to human safety. The penalty gain for ship operation-related businesses, Penalty gains for commercial communication services.

[0034] Furthermore, the following settings are made for the business data packet queue model.

[0035] Assuming that the amount of incoming service type data packets within one frame (unit of time) is The amount of data packets transmitted is The arrival of service data packets follows the parameter: The Poisson distribution appears A data packet arrived. The probability of a data packet being transmitted is:

[0036] in, For time intervals, for A data packet arrived. One data packet was transmitted.

[0037] Dividing each frame into time slots, then in the first... The queue length for different service data packets in each time slot is:

[0038] in, For the first The length of the data packet queue for each time slot. and For different service data packets The number of data packets transmitted and received in each time slot.

[0039] Furthermore, the following constraints were set for the packet scheduling model.

[0040] At the packet scheduling level, it is necessary to balance the priority of service packets with waiting time and minimize queue length, thus it is a multi-objective optimization problem. The mathematical expression of the optimization problem is shown below.

[0041]

[0042]

[0043]

[0044] .

[0045] As a preferred embodiment of the present invention, this application also constructs a shore-based communication network model, which is mainly used to simulate ship motion and calculate the time delay and signal-to-noise ratio of candidate networks based on the ship-shore distance.

[0046] Specifically, during ship-to-shore collaborative communication, the shore-based communication system is always situated within a heterogeneous wireless network coverage environment. Due to environmental factors, candidate shore-based networks... The attributes of the network are constantly fluctuating. Based on the QoS requirements of ship-to-shore communication services, this patent considers the network attributes as follows: .

[0047] The end-to-end latency of this patented network is the total data packet transmission time, as shown below.

[0048]

[0049] in, , , , These are processing delay, queuing delay, sending delay, and propagation delay, respectively.

[0050] In the marine environment, the propagation of radio waves is significantly influenced by multiple factors, including the unique refraction effect caused by evaporating waveguides. Traditional logarithmic path loss models used in wireless communication are insufficient to accurately characterize the unique characteristics of maritime wireless channels. To accurately calculate the SNR of each network, this invention employs a dual diffraction model to simulate the propagation characteristics of marine radio waves. The propagation loss formula is shown below.

[0051]

[0052]

[0053] in, and These represent the heights of the transmitting and receiving antennas for different communication methods. The distance between the ship and the shore. and These refer to the radio wavelengths and frequencies used in different communication methods. It is the speed of light.

[0054] The SNR calculation formulas for different networks are shown below.

[0055]

[0056] in, For the transmission power of different communication methods, and For the transmit and receive antenna gain of different communication methods, For the propagation loss of different communication methods, This represents noise power.

[0057] Since different ship-to-shore communication services have different QoS requirements, a network evaluation scoring function is established to measure the quality of the current network, as shown below.

[0058]

[0059] in, This indicates the online evaluation score. and Preset weights for different industries and different network attributes.

[0060] In the process of selecting communication methods, multiple communication method attributes should be comprehensively evaluated based on service QoS requirements to achieve joint optimization of communication quality and throughput. Optimal network access selection can be achieved by establishing a network comprehensive score function and solving for its maximum value. Therefore, the mathematical expression of the optimization problem is:

[0061]

[0062] in, and These represent the weights of signal-to-noise ratio and latency for networks related to human safety, respectively. and These represent the weights of the signal-to-noise ratio and latency for the ship operation network, respectively.

[0063] S2. Obtain the data packet characteristics of the first service data packet, the second service data packet, and the third service data packet, wherein the data packet characteristics include queue length, priority, and waiting time.

[0064] S3. Input the data packet characteristics into the first intelligent agent. The first intelligent agent is used to optimize data packet scheduling by maximizing data packet transmission gain and minimizing queue length. The state space of the first intelligent agent includes data packet characteristics, the action space includes different communication service data packet types, and the reward function is the difference between data packet transmission gain and queue length.

[0065] Specifically, the first intelligent agent is a packet scheduling intelligent agent, defined as follows: .

[0066] intelligent agent The state space at time slot t includes queue length, packet priority, and packet waiting time, as shown below:

[0067] Therefore, the state space is .

[0068] intelligent agent The action space in time slot t contains different communication service data packet types, as shown below:

[0069] in, Data packets representing human life safety This represents data packets related to ship operations. This represents data packets used in commercial communications.

[0070] Therefore, the action space is .

[0071] For intelligent agents The rewards for maximizing the gain of sending data packets and minimizing the queue length are set, and the rewards for time slot t are shown below.

[0072]

[0073] S4. Obtain the attribute parameters of different networks, including latency and signal-to-noise ratio. Input the attribute parameters of different networks and the communication service data packet type output by the first agent into the second agent. The second agent is used to optimize network selection by minimizing the number of network switching. The state space of the second agent includes the communication service data packet type and the attribute parameters of the network. The action space includes candidate networks and whether to switch. The reward function is the difference between the network evaluation score and the switching penalty.

[0074] Specifically, the second agent is the agent selected by the network, defined as follows: .

[0075] intelligent agent The state space in time slot t includes the type of data packet selected by the agent and the network attributes, as shown below:

[0076] Therefore, the state space is , and These represent the time delay and signal-to-noise ratio of different networks at time slot t, respectively.

[0077] intelligent agent The action space is a joint action space. In time slot t, the action space includes the candidate network and whether to switch, as shown below:

[0078] in, Representing candidate networks, Indicates whether to switch.

[0079] Therefore, the action space is

[0080] For intelligent agents To maximize the network evaluation score function, a reward is set. To reduce the number of network switches and lower the switching cost, a network switch penalty is introduced, as shown below:

[0081] in, To switch penalties, The value is a random number between (0, 1). The switching penalty is 0 if the network selected in the previous time step is the same as the currently selected network; otherwise, the switching penalty is... .

[0082] When using, the intelligent agent The initial state is defined as The state at time slot t is defined as The state at time slot t+1 is defined as intelligent agent The initial state is defined as The state at time slot t is defined as The state at time slot t+1 is defined as .

[0083] intelligent agent The state space at time slot t includes the queue length. Data packet priority Data packet waiting time , denoted as: Intelligent agents The action space in time slot t contains different types of communication service data packets, denoted as: in, Data packets representing human life safety This represents data packets related to ship operations. This represents data packets used in commercial communications. For intelligent agents... A reward is set for maximizing the gain of data packet transmission and minimizing the queue length. The reward in time slot t is denoted as... .

[0084] intelligent agent The state space in time slot t includes the type of data packet selected by the agent. Network latency and signal-to-noise ratio , recorded as Intelligent agents The action space is a joint action space. In time slot t, the action space includes the candidate network and whether a switch occurs, denoted as: ,in, Representing candidate networks, Indicates whether to switch. For intelligent agents... To maximize the network evaluation score function, we set a reward. In order to reduce the number of network switches and reduce the switching cost, we introduce a network switching penalty, as shown below.

[0085]

[0086] in, To switch penalties, The value is a random number between (0, 1). The switching penalty is 0 if the network selected in the previous time step is the same as the currently selected network; otherwise, the switching penalty is... .

[0087] The training steps for the two agents are described below. First, obtain the experience pools for both agents. (Agents) The experience pool is recorded as intelligent agent Experience pool is recorded as

[0088] intelligent agent The formula for calculating the objective function of the actor network is shown below.

[0089]

[0090] in, Represents the dominant function; The output probability of the policy after masking. This represents the probability output by the old strategy after masking. It is a constant. Representing intelligent agents Actions during time slot t Representing intelligent agents The state in time slot t.

[0091] intelligent agent The objective function of the critic network is calculated as follows.

[0092]

[0093] in, For the state value function, This is a discount reward.

[0094] intelligent agent The formula for calculating the objective function of the actor network is shown below.

[0095]

[0096] in, For intelligent agents The state after the attention mechanism in time slot t Output the probability for the new strategy. The probability of the output being the old strategy; It is a constant.

[0097] intelligent agent The objective function of the critic network is calculated as follows.

[0098]

[0099] Specifically, regarding intelligent agents The state after the attention mechanism in time slot t The method for obtaining is shown below.

[0100] Suppose that at time t, the agent... The input sequence is ,in For intelligent agents In the nth state of time slot t, the weight matrix is: , , For each element The query vector is The key vector is The value vector is .

[0101] The formula for calculating attention weights is shown below.

[0102]

[0103] in, The attention score represents the input. right Attention level This is the scaling factor.

[0104] Then, the value vector of each element is multiplied by its corresponding attention weight, and the final output is obtained by summing them, as shown in the following formula.

[0105]

[0106] Furthermore, as a preferred embodiment of this application, training the agent mainly includes the following steps.

[0107] Step 1: Initialize the policy network and value network.

[0108] Step 2: Initialize the experience pool.

[0109] Step 3: Set the total number of training iterations.

[0110] Step 4: Set the initial state of the agent: , .

[0111] Step 5: For the intelligent agent Input status ,if If the value is 0, then the action mask is used, and then the action is executed according to the strategy. Receive rewards And the state in the next moment. and put Store in experience pool middle.

[0112] Step 6: For the intelligent agent Input status as well as Receive rewards And the state in the next moment. , and then Store in experience pool middle.

[0113] Step 7: From the experience pool Select group B data for the agent Conduct training.

[0114] Step 8: Intelligent Agent The objective function of the critic network is The critic network is updated using gradient descent.

[0115] Step 9: Intelligent Agent The objective function of the actor network is The actor network is updated using gradient ascent.

[0116] Step 10: From the experience pool Select group B data for the agent Conduct training.

[0117] Step 11: Intelligent Agent The objective function of the critic network is The critic network is updated using gradient descent.

[0118] Step 12: Intelligent Agent The objective function of the actor network is The actor network is updated using gradient ascent.

[0119] Step 13: Empty the experience pool and .

[0120] S5. Obtain the link decision based on the action output by the second agent.

[0121] To achieve joint optimization of resource scheduling and network selection in ship-shore integrated communication equipment, this application constructs a ship service model, a packet scheduling model, and a shore-side communication network model, and proposes a shore-side multi-mode link decision-making method based on communication service quality. A dual-agent collaborative mechanism is adopted, with agents responsible for packet scheduling and network selection decisions respectively. At the packet scheduling level, the scheduling agent... Data packets are scheduled based on their state and priority. Simultaneously, by introducing an invalid action masking mechanism, the selection of empty queues is effectively shielded, ensuring the effectiveness of scheduling decisions. At the network selection level, the agent... Employing a joint action space, the agent is defined based on the network state. The scheduled data packets select the optimal network and decide whether to perform a network switch. To reduce the number of switchovers, an attention mechanism is incorporated, which dynamically assigns higher weights to key features based on current business needs and uses a weighted averaging strategy to suppress short-term fluctuations, thereby significantly reducing the frequency of network switchovers.

[0122] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A shore-side multi-mode link decision-making method based on communication service quality, characterized in that, Includes the following steps: From shore-to-ship communication services, a first service data packet representing life safety services, a second service data packet representing ship operation services, and a third service data packet representing commercial communication services are obtained, with the priority of the first service data packet, the second service data packet, and the third service data packet decreasing in that order. Obtain the data packet characteristics of the first service data packet, the second service data packet, and the third service data packet, wherein the data packet characteristics include queue length, priority, and waiting time; The data packet characteristics are input into a first agent, which is used to optimize data packet scheduling by maximizing data packet transmission gain and minimizing queue length. The state space of the first agent includes data packet characteristics, the action space includes different communication service data packet types, and the reward function is the difference between data packet transmission gain and queue length. The attribute parameters of different networks are obtained, including latency and signal-to-noise ratio. The attribute parameters of different networks and the communication service data packet type output by the first agent are input into the second agent. The second agent is used to optimize network selection by minimizing the number of network switching. The state space of the second agent includes the communication service data packet type and the attribute parameters of the network, the action space includes candidate networks and whether to switch, and the reward function is the difference between the network evaluation score and the switching penalty. Make link decisions based on the actions output by the second agent.

2. The shore-side multi-mode link decision-making method based on communication service quality according to claim 1, characterized in that, The formula for calculating the data packet transmission gain is as follows: in, b represents the data packet transmission gain, and b represents ship-to-shore communication services. This refers to businesses related to human safety. This refers to ship operation services. Indicates commercial communication services. For positive gain weights, Prioritizing different service data packets This refers to the relaxation of the waiting time.

3. The shore-side multi-mode link decision-making method based on communication service quality according to claim 1, characterized in that, No. The formula for calculating the queue length of different service data packets in each time slot is as follows: in, For the first The length of the data packet queue for each time slot. and For different service data packets The number of data packets transmitted and received in each time slot.

4. The shore-side multi-mode link decision-making method based on communication service quality according to claim 1, characterized in that, The formula for calculating network latency is as follows: in, Indicates communication method Total latency, , Communication methods The transmission delay and propagation delay.

5. The shore-side multi-mode link decision-making method based on communication service quality according to claim 1, characterized in that, The formula for calculating the network signal-to-noise ratio is as follows: in, Indicates communication method signal-to-noise ratio, Indicates communication method The transmission power, and Indicates communication method The transmit and receive antenna gain, Indicates communication method Propagation loss, This represents noise power.

6. The shore-side multi-mode link decision-making method based on communication service quality according to claim 1, characterized in that, The formula for calculating the network evaluation score is as follows: in, This indicates the online evaluation score. and Preset weights for different industries and different network attributes.

Images