Port intelligent control system based on network communication

Abstract

The present application discloses a port intelligent control system based on network communication, which utilizes network communication technology and artificial intelligence technology to perform natural language text semantic analysis and understanding on ship berthing requests sent by users, capture the semantic meaning of the ship's berthing requirements, and match them with the real-time status semantic information of the currently available berthing points to find the berthing point that best suits the ship.

Description

Technical Field

[0001] The present application relates to the field of intelligent control, and more specifically, to a port intelligent control system based on network communication. Background Art

[0002] With the continuous development of global trade, ports, as important hubs for cargo circulation, are crucial to the smooth operation of the entire supply chain in terms of operational efficiency and management level. Among them, berth allocation plays a vital role in the control and management of ports.

[0003] Specifically, berths are limited resources in ports, and each ship needs a suitable berth to load and unload cargo. Therefore, optimizing the berth allocation process can maximize the use of these limited resources and ensure that ships can dock smoothly. In addition, reasonable berth allocation can optimize the terminal operation process. If berth allocation is improper, it may cause ships to wait in line, delay loading and unloading time, and increase operation time and cost.

[0004] Traditional port management methods mainly rely on manual berth allocation, which cannot effectively cope with the complex and changing berthing needs of ships, resulting in low resource utilization and low operational efficiency. Therefore, an optimized port intelligent control system is expected. Summary of the invention

[0005] In order to solve the above technical problems, this application is proposed. The embodiment of this application provides a port intelligent control system based on network communication, which uses network communication technology and artificial intelligence technology to perform natural language text semantic analysis and understanding on the ship berthing request sent by the user, captures the semantic meaning of the ship's berthing demand, and matches it with the real-time status semantic information of the current available berthing points to find the berthing point that best suits the ship.

[0006] According to one aspect of the present application, a port intelligent control system based on network communication is provided, which includes:

[0007] The ship berthing request transmission module is used for the user to send the ship berthing request to the port intelligent control server through the network communication transmission module;

[0008] A ship berthing request semantic analysis module, used for performing global semantic feature analysis on the ship berthing request to obtain a ship berthing request semantic feature vector;

[0009] An available berthing point status extraction module is used to extract real-time status descriptions of multiple available berthing points from the database of the port intelligent control server;

[0010] An available anchorage point state understanding module, used for semantically encoding the real-time state descriptions of the plurality of available anchorage points to obtain a sequence of available anchorage point state semantic feature vectors;

[0011] The berthing point allocation module is used to obtain a berthing point allocation result based on the semantic matching degree between the sequence of the ship berthing request semantic feature vector and the available berthing point status semantic feature vector.

[0012] Compared with the prior art, the present application provides a port intelligent control system based on network communication, which utilizes network communication technology and artificial intelligence technology to perform natural language text semantic analysis and understanding of ship berthing requests sent by users, capture the semantic meaning of the ship's berthing requirements, and match it with the real-time status semantic information of the current available berthing points to find the berthing point that best suits the ship. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0014] Figure 1 is a block diagram of a port intelligent control system based on network communication according to an embodiment of the present application;

[0015] Figure 2 is a system architecture diagram of a port intelligent control system based on network communication according to an embodiment of the present application;

[0016] Figure 3 A block diagram of the training phase of the port intelligent control system based on network communication according to an embodiment of the present application;

[0017] Figure 4 A block diagram of a semantic analysis module for ship berthing request in a port intelligent control system based on network communication according to an embodiment of the present application;

[0018] Figure 5 It is a block diagram of a berthing point allocation module in a port intelligent control system based on network communication according to an embodiment of the present application. DETAILED DESCRIPTION

[0019] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0020] As shown in this application and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0021] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0022] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. Meanwhile, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.

[0023] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0024] Traditional port management methods mainly rely on manual berth allocation, which cannot effectively cope with the complex and changing berthing needs of ships, resulting in low resource utilization and low operational efficiency. Therefore, an optimized port intelligent control system is expected.

[0025] Network communication refers to the process of transmitting and exchanging information, data, sound, images and other data through computer networks or other communication devices. Considering that in the field of port control, network communication plays a key role. For example, ports need to monitor the status of ships, containers, quay cranes and other equipment in real time. Network communication can quickly transmit this data to help port managers make timely decisions. Network communication supports the loading and real-time updating of high-definition video, allowing ports to have more accurate positioning and navigation services. In addition, network communication can also be used to monitor port security equipment, such as smart unlocking stations. The development and application of network communication in the field of port control provides a new solution for optimizing the berth allocation process of ports.

[0026] In the technical solution of the present application, a port intelligent control system based on network communication is proposed. Figure 1 It is a block diagram of a port intelligent control system based on network communication according to an embodiment of the present application. Figure 2 FIG. 1 is a system architecture diagram of a port intelligent control system based on network communication according to an embodiment of the present application. Figure 1 and Figure 2 As shown, according to an embodiment of the present application, a port intelligent control system 300 based on network communication includes: a ship berthing request transmission module 310, which is used by a user to send a ship berthing request to a port intelligent control server through a network communication transmission module; a ship berthing request semantic analysis module 320, which is used to perform a global semantic feature analysis on the ship berthing request to obtain a ship berthing request semantic feature vector; an available berthing point status extraction module 330, which is used to extract a real-time status description of multiple available berthing points from the database of the port intelligent control server; an available berthing point status understanding module 340, which is used to perform semantic encoding on the real-time status description of the multiple available berthing points to obtain a sequence of available berthing point status semantic feature vectors; a berthing point allocation module 350, which is used to obtain a berthing point allocation result based on the degree of semantic matching between the ship berthing request semantic feature vector and the sequence of the available berthing point status semantic feature vectors.

[0027] In particular, the ship berthing request transmission module 310 is used for the user to send a ship berthing request to the port intelligent control server through the network communication transmission module. It is worth mentioning that here, the network communication transmission module refers to a device or software module used to transmit data and information in the network. Its function is to connect the user end and the server end through the network, and is responsible for transmitting the data sent by the user (such as the ship berthing request) to the target server through the network. This module can be a software program, a hardware device or an implementation of a network protocol, which is used to ensure the reliable transmission of data in the network. The port intelligent control server is a server system with intelligent processing capabilities, which is used to receive, process and respond to ship berthing requests sent by users, and perform intelligent control operations such as berth allocation. In this way, the user can send a ship berthing request to the port intelligent control server through the network communication transmission module to achieve fast, safe and accurate transmission of information.

[0028] In particular, the ship berthing request semantic analysis module 320 is used to perform global semantic feature analysis on the ship berthing request to obtain a ship berthing request semantic feature vector, wherein the ship berthing request includes the ship size, the ship draft, the type of cargo carried by the ship, the processing time of the cargo carried by the ship, the estimated arrival time of the ship and the berthing restriction conditions. Specifically, the size of the ship directly affects the space required for the ship to berth; the draft of the ship refers to the vertical distance from the bottom of the ship to the water surface, which affects the choice of berthing position of the ship under specific water depth conditions. According to the draft of the ship, the berth water depth suitable for the ship to berth can be determined to avoid the problem of the ship running aground or insufficient water depth; different types of cargo may require different loading and unloading equipment and operation processes. The type and processing time of the cargo carried by the ship can help determine the appropriate berth so as to provide the ship with appropriate loading and unloading equipment and operating conditions to improve the operation efficiency; the estimated arrival time of the ship refers to the time when the ship is scheduled to arrive at the berth. This information is crucial to the scheduling and operation plan of the berth. According to the estimated arrival time of the ship, the allocation and operation plan of the berth can be reasonably arranged to avoid the situation of berth being idle or crowded; and the berth restriction conditions can include the special needs of the ship and the specific restriction information on the berth, such as the loading and unloading equipment of the berth, etc. Based on the above information, the various information contained in the ship berthing request can help the port intelligent control system better understand the berthing needs of the ship, so as to achieve the best match between the ship and the berth and optimize the berth allocation process. In particular, in a specific example of the present application, Figure 4 As shown, the ship berthing request semantic analysis module 320 includes: a request item semantic encoding unit 321, used for semantically encoding the ship berthing request to obtain a sequence of ship berthing request item semantic feature vectors; a global semantic interaction unit 322, used for passing the sequence of ship berthing request item semantic feature vectors through a deep semantic feature capture module based on a global dependent attention interaction mechanism to obtain the ship berthing request semantic feature vector.

[0029] Specifically, the request item semantic encoding unit 321 is used to semantically encode the ship berthing request to obtain a sequence of semantic feature vectors of the ship berthing request item. That is, by means of semantic encoding, various information in the ship berthing request can be standardized into a computer-understandable form, the data format and structure can be unified, the consistency and accuracy of the data can be ensured, and the semantic information and meaning in the ship berthing request can be extracted and understood. More specifically, the ship berthing request is semantically encoded to obtain a sequence of semantic feature vectors of the ship berthing request item, including: performing word segmentation processing on each text item in the ship berthing request to obtain a sequence of multiple berthing request words; using a word embedding layer to process the sequence of multiple berthing request words to obtain a sequence of multiple berthing request word embedding vectors; passing the sequence of the multiple berthing request word embedding vectors through a contextual text semantic encoder to obtain a sequence of semantic feature vectors of the ship berthing request item. It is worth mentioning that the word embedding layer is a commonly used layer in deep learning models, which is used to convert discrete word representations into dense continuous vector representations. This vector representation can capture the semantic and grammatical relationships between words, thus providing a better representation of input features. A text semantic encoder is a neural network model that converts text data into a low-dimensional vector representation with semantic information. This encoder is often used in natural language processing tasks such as text classification, semantic matching, sentiment analysis, etc.

[0030] Specifically, the global semantic interaction unit 322 is used to pass the sequence of semantic feature vectors of the ship berthing request items through a deep semantic feature capture module based on a global dependency attention interaction mechanism to obtain the semantic feature vector of the ship berthing request. That is, in the technical solution of the present application, the sequence of semantic feature vectors of the ship berthing request items is passed through a deep semantic feature capture module based on a global dependency attention interaction mechanism to capture the semantic associations and dependencies between different request items within the sequence range of the entire ship berthing request item semantic feature vector, and fully understand the importance and relevance between various information in the ship berthing request, thereby obtaining the semantic feature vector of the ship berthing request. More specifically, the sequence of semantic feature vectors of the ship berthing request items is processed with the following global dependency attention interaction formula to obtain the semantic feature vector of the ship berthing request; wherein the global dependency attention interaction formula is:

[0031]

[0032]

[0033] Among them, v i is the i-th semantic feature vector of the ship berthing request item in the sequence of semantic feature vectors of the ship berthing request item, vj is the jth semantic feature vector of the ship berthing request item in the sequence of semantic feature vectors of the ship berthing request item, ||·||1 represents the 1-norm of the feature vector, M is the length of the sequence of semantic feature vectors of the ship berthing request item -1, V k is the representation of the sequence of semantic feature vectors of the ship berthing request item, D i represents the characteristic difference coefficient, exp(·) represents the natural exponential function operation, N s represents the total number of characteristic difference coefficients, P k A semantic feature vector for the ship's berthing request.

[0034] It is worth mentioning that in other specific examples of the present application, the global semantic feature analysis of the ship berthing request can also be performed in other ways to obtain a semantic feature vector of the ship berthing request, for example: input the ship berthing request; perform text cleaning on the berthing request text, such as removing special characters, punctuation marks and stop words; use a pre-trained word embedding model (such as Word2Vec, GloVe, FastText) to convert the text into a dense vector representation; perform word embedding mapping on each word in the berthing request to obtain a word vector sequence; fill or truncate the word vector sequence so that the word vector sequence length of each berthing request is the same ; Use models such as recurrent neural network (RNN), long short-term memory network (LSTM), or Transformer to capture the semantic information in the sequence; input the padded word vector sequence into the neural network model; add a global pooling layer (such as global average pooling) or an attention mechanism before the output layer of the model to capture the semantic information of the entire sequence; obtain the global semantic feature representation of each parking request through the output of the model; convert the global semantic feature representation of each parking request into a semantic feature vector of fixed length; use a fully connected layer or other techniques to further compress the obtained global semantic features into a feature vector of fixed dimension.

[0035] In particular, the available berthing point status extraction module 330 is used to extract the real-time status description of multiple available berthing points from the database of the port intelligent control server. The real-time extraction of the berthing point status description can help monitor the situation of each berthing point, including whether it is idle, whether there are ships docked, the type and status of docked ships, etc. By understanding the real-time status of each berthing point, effective resource scheduling and management can be carried out.

[0036] In particular, the available berthing point status understanding module 340 is used to semantically encode the real-time status descriptions of the multiple available berthing points to obtain a sequence of available berthing point status semantic feature vectors. Here, semantic understanding of the real-time status descriptions of each available berthing point by semantic encoding can understand the status semantic information of each available berthing point, so as to measure and match it with the semantic meaning of the berthing demand of the ship, thereby making a reasonable berthing allocation decision. Specifically, in a specific example of the present application, semantic encoding of the real-time status descriptions of the multiple available berthing points to obtain a sequence of available berthing point status semantic feature vectors includes: performing word segmentation processing on the real-time status descriptions of the multiple available berthing points to obtain a sequence of real-time status description words of the multiple available berthing points; using a word embedding layer to process the sequence of real-time status description words of the multiple available berthing points to obtain a sequence of real-time status description word embedding vectors of the multiple available berthing points; passing the sequence of real-time status description word embedding vectors of the multiple available berthing points through a context text semantic encoder to obtain a sequence of available berthing point status semantic feature vectors.

[0037] In particular, the berthing point allocation module 350 is used to obtain a berthing point allocation result based on the semantic matching degree between the sequence of the ship berthing request semantic feature vector and the available berthing point status semantic feature vector. In particular, in a specific example of the present application, Figure 5 As shown, the berthing point allocation module 350 includes: a cross-modal matching unit 351, which is used to pass the sequence of the ship berthing request semantic feature vector and the available berthing point status semantic feature vector through a cross-modal metric matching module to obtain multiple metric matching coefficients; an allocation result generation and transmission unit 352, which is used to use the available berthing point corresponding to the maximum value of the multiple metric matching coefficients as the berthing point allocation result, and transmit the berthing point allocation result to the user through the network communication transmission module.

[0038] Specifically, the cross-modal matching unit 351 is used to pass the sequence of the ship's berthing request semantic feature vector and the available berthing point status semantic feature vector through a cross-modal metric matching module to obtain a plurality of metric matching coefficients. That is, in the technical solution of the present application, the sequence of the ship's berthing request semantic feature vector and the available berthing point status semantic feature vector are passed through a cross-modal metric matching module to obtain a plurality of metric matching coefficients. That is, the cross-modal metric matching module is used to quantify and measure the degree of match between the user's berthing demand and the status information of each available berthing point. More specifically, the sequence of the ship's berthing request semantic feature vector and the available berthing point status semantic feature vector are processed with the following cross-modal metric matching formula to obtain the plurality of metric matching coefficients; wherein, the cross-modal metric matching formula is:

[0039]

[0040] Where h is the semantic feature vector of the ship state, v i is the i-th anchorage state semantic feature vector in the sequence of anchorage state semantic feature vectors, e i is the metric matching coefficient of the i-th, W i is the i-th weight coefficient matrix, B i is the i-th bias vector, tanh(·) represents the hyperbolic tangent function processing, (·) T represents the transpose of the eigenvector.

[0041] Specifically, the allocation result generation and transmission unit 352 is used to use the available berthing point corresponding to the maximum value of the multiple metric matching coefficients as the berthing point allocation result, and transmit the berthing point allocation result to the user through the network communication transmission module. In this way, by selecting the most matching berthing point, unnecessary search and calculation time can be reduced, thereby improving the efficiency of berthing point allocation. This helps to speed up the berthing process of the ship and reduce waiting time.

[0042] It is worth mentioning that in other specific examples of the present application, the berthing point allocation result can also be obtained by other methods based on the degree of semantic matching between the sequence of the ship berthing request semantic feature vector and the available berthing point status semantic feature vector, for example: for each ship berthing request, the semantic feature is extracted from the request and represented as a semantic feature vector; for the real-time status description of each available berthing point, the semantic feature is also extracted and represented as a semantic feature vector; for each ship berthing request semantic feature vector and each available berthing point status semantic feature vector, the semantic matching degree between them is calculated. Commonly used calculation methods include cosine similarity, Euclidean distance, etc.; use a neural network model to learn semantic matching, such as using a Siamese network or an attention mechanism to capture the association between semantics; for each ship berthing request, match its semantic feature vector with the semantic feature vectors of all available berthing point states to obtain a matching degree sequence; normalize this matching degree sequence to ensure that the matching degree is within an appropriate range; based on the matching degree sequence, select the available berthing point with the highest matching degree as the allocated berthing point for the ship; set a threshold to determine whether to accept the matching result to ensure that the allocation result meets the expected requirements; once the berthing point allocation is completed, update the corresponding status information, including the ship berthing status and the berthing point status; transmit the berthing point allocation result to the user through the network communication transmission module.

[0043] It should be understood that before using the above neural network model for inference, the deep semantic feature capture module based on the global dependent attention interaction mechanism and the cross-modal attitude measurement matching module need to be trained. That is, according to the port intelligent control system 300 based on network communication of the present application, it also includes a training stage 400 for training the deep semantic feature capture module based on the global dependent attention interaction mechanism and the cross-modal attitude measurement matching module.

[0044] Figure 3 FIG. 1 is a block diagram of the training phase of the port intelligent control system based on network communication according to an embodiment of the present application. Figure 3 As shown, according to the embodiment of the present application, the port intelligent control system 300 based on network communication includes: a training stage 400, including: a training data acquisition unit 410, used to acquire training data, wherein the training data includes a training ship berthing request sent by a user through a network communication transmission module, and the true values ​​of multiple metric matching coefficients; a training request item semantic encoding unit 420, used to semantically encode the training ship berthing request to obtain a sequence of training ship berthing request item semantic feature vectors; a training global semantic interaction unit 430, used to pass the sequence of training ship berthing request item semantic feature vectors through the deep semantic feature capture module based on the global dependent attention interaction mechanism to obtain a training ship berthing request semantic feature vector; a training available berthing point state extraction unit 440, used to extract a training real-time state description of multiple available berthing points from the database of the port intelligent control server; a training available berthing point state understanding unit 450, used to semantically encode the multiple The invention relates to a method for semantically encoding the training real-time status descriptions of the available berthing points to obtain a sequence of training available berthing point status semantic feature vectors; a training cross-modal matching unit 460, which is used to pass the training ship berthing request semantic feature vector and the sequence of the training available berthing point status semantic feature vectors through the cross-modal attitude metric matching module to obtain a plurality of training metric matching coefficients; a cross-entropy calculation unit 470, which is used to calculate the cross-entropy function value between the training metric matching vector composed of the plurality of training metric matching coefficients and the true metric matching vector composed of the true values ​​of the plurality of metric matching coefficients; a training unit 480, which is used to train the deep semantic feature capture module based on the global dependent attention interaction mechanism and the cross-modal attitude metric matching module using the cross-entropy function value as the loss function value, wherein in each round of iteration of the training, each training available berthing point status semantic feature vector in the sequence of the training available berthing point status semantic feature vectors is optimized.

[0045] In particular, in the technical solution of the present application, the training ship berthing request semantic feature vector expresses the global semantic feature of the ship berthing request based on the global dependent attention interaction of the local encoded semantics, and each training available berthing point state semantic feature vector in the sequence of the training available berthing point state semantic feature vectors expresses the encoded semantic feature of the real-time state description of the corresponding available berthing point. Therefore, considering the source semantic differences of the real-time state descriptions of the available berthing points and the corresponding encoded semantic feature expression differences, the sequence of the training available berthing point state semantic feature vectors will also have inconsistent overall feature distribution of feature vectors between the training available berthing point state semantic feature vectors, thereby affecting the correlation between the multiple metric matching coefficients obtained by the cross-modal metric matching module of the training ship berthing request semantic feature vector and the sequence of the training available berthing point state semantic feature vectors, thereby affecting the calculation accuracy of the multiple metric matching coefficients. Therefore, the applicant of the present application optimizes each training available berthing point state semantic feature vector in the sequence of the training available berthing point state semantic feature vectors at each iteration during the training process of the model, which is expressed as:

[0046]

[0047] where v i，j ∈V i is the i-th training available anchorage state semantic feature vector V in the sequence of training available anchorage state semantic feature vectors i The eigenvalue of the jth position, φ(v i，j ) represents the probabilistic function of the eigenvalue, that is, the eigenvalue v i，j The probabilistic function mapped to the interval [0, 1], L is the length of the semantic feature vector of the training available anchor point state, p i is the semantic feature vector V of the available anchor point state for training i The probability value obtained by dividing the corresponding metric matching coefficient by the sum of the metric matching coefficients corresponding to all training available anchorage state semantic feature vectors, and α is a weight hyperparameter. That is, for each training available anchorage state semantic feature vector V in the sequence of training available anchorage state semantic feature vectors i The corresponding feature scene, through the probability distribution prospect constraint and the relative probability mapping response hypothesis to undertake the metric matching probability reasoning logic association of scene saturation, so as to provide each training available anchorage state semantic feature vector V in the sequence of the training available anchorage state semantic feature vectors iThe feature set of gives the scene concept ontology cognition, that is, the overall distribution is internally aligned with the scene-based metric matching probability logic reasoning under the metric matching process to improve each training available anchorage state semantic feature vector V in the sequence of the training available anchorage state semantic feature vectors i The feature vector scene distribution of the semantic metric cognitive understanding ability. In this way, the coefficient ω i For each training available anchorage state semantic feature vector V in the sequence of training available anchorage state semantic feature vectors i By performing weighted optimization, the overall feature distribution inconsistency between the sequences of the training available anchorage state semantic feature vectors based on the metric matching scenario can be improved, thereby improving the calculation accuracy of the multiple metric matching coefficients.

[0048] As described above, the port intelligent control system 300 based on network communication according to the embodiment of the present application can be implemented in various wireless terminals, such as a server with a port intelligent control algorithm based on network communication. In a possible implementation, the port intelligent control system 300 based on network communication according to the embodiment of the present application can be integrated into the wireless terminal as a software module and / or a hardware module. For example, the port intelligent control system 300 based on network communication can be a software module in the operating system of the wireless terminal, or can be an application developed for the wireless terminal; of course, the port intelligent control system 300 based on network communication can also be one of the many hardware modules of the wireless terminal.

[0049] Alternatively, in another example, the port intelligent control system 300 based on network communication and the wireless terminal may also be separate devices, and the port intelligent control system 300 based on network communication may be connected to the wireless terminal via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0050] The embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A port intelligent control system based on network communication, characterized in that: include: The ship berthing request transmission module is used for the user to send the ship berthing request to the port intelligent control server through the network communication transmission module; A ship berthing request semantic analysis module, used for performing global semantic feature analysis on the ship berthing request to obtain a ship berthing request semantic feature vector; An available berthing point status extraction module is used to extract real-time status descriptions of multiple available berthing points from the database of the port intelligent control server; An available anchorage point state understanding module, used for semantically encoding the real-time state descriptions of the plurality of available anchorage points to obtain a sequence of available anchorage point state semantic feature vectors; A berthing point allocation module, used to obtain a berthing point allocation result based on the semantic matching degree between the sequence of the ship berthing request semantic feature vector and the available berthing point status semantic feature vector; Wherein, the ship berthing request semantic analysis module includes: A request item semantic encoding unit, used for semantically encoding the ship berthing request to obtain a sequence of semantic feature vectors of the ship berthing request item; A global semantic interaction unit, used for passing the sequence of semantic feature vectors of the ship berthing request item through a deep semantic feature capture module based on a global dependent attention interaction mechanism to obtain the semantic feature vector of the ship berthing request; Wherein, the global semantic interaction unit is used to: The sequence of the semantic feature vectors of the ship berthing request items is processed by the following global dependency attention interaction formula to obtain the semantic feature vector of the ship berthing request; wherein the global dependency attention interaction formula is: in, is the first in the sequence of semantic feature vectors of the ship berthing request item The semantic feature vector of the ship berthing request, is the first in the sequence of semantic feature vectors of the ship berthing request item The semantic feature vector of the ship berthing request, represents the 1-norm of the eigenvector, is the length of the sequence of semantic feature vectors of the ship berthing request item - 1, is the representation of the sequence of semantic feature vectors of the ship berthing request item, represents the characteristic difference coefficient, represents the natural exponential function operation, represents the total number of feature difference coefficients, A semantic feature vector for the ship's berthing request.

2. The port intelligent control system based on network communication according to claim 1 is characterized in that: The ship berthing request includes the ship size, the ship draft, the type of cargo carried by the ship, the processing time of the cargo carried by the ship, the estimated arrival time of the ship and berthing restrictions.

3. The port intelligent control system based on network communication according to claim 2 is characterized in that: The request item semantic encoding unit includes: a word segmentation subunit, configured to perform word segmentation processing on each text item in the ship berthing request to obtain a sequence of multiple berthing request words; a word embedding subunit, configured to process the sequence of the plurality of parking request words using a word embedding layer to obtain a sequence of a plurality of parking request word embedding vectors; The text semantic encoding subunit is used to pass the sequence of the plurality of berthing request word embedding vectors through a contextual text semantic encoder to obtain a sequence of semantic feature vectors of the ship berthing request items.

4. The port intelligent control system based on network communication according to claim 3 is characterized in that: The berthing point allocation module comprises: A cross-modal matching unit, used for passing the sequence of the ship berthing request semantic feature vector and the available berthing point status semantic feature vector through a cross-modal metric matching module to obtain a plurality of metric matching coefficients; The allocation result generating and transmitting unit is used to use the available parking point corresponding to the maximum value of the multiple metric matching coefficients as the parking point allocation result, and transmit the parking point allocation result to the user through the network communication transmission module.

5. The port intelligent control system based on network communication according to claim 4 is characterized in that: The cross-modal matching unit is used to: The sequence of the ship berthing request semantic feature vector and the available berthing point status semantic feature vector is processed by the following cross-modal attitude metric matching formula to obtain the multiple metric matching coefficients; wherein the cross-modal attitude metric matching formula is: in, is the semantic feature vector of the ship berthing request, is the first in the sequence of available parking point state semantic feature vectors Available parking point state semantic feature vectors, For the The metric matching coefficient, For the A weight coefficient matrix, For the A bias vector, represents the hyperbolic tangent function processing, represents the transpose of the eigenvector.

6. The port intelligent control system based on network communication according to claim 5 is characterized in that: It also includes a training module for training the deep semantic feature capture module based on the global dependent attention interaction mechanism and the cross-modal attitude measurement matching module.

7. The port intelligent control system based on network communication according to claim 6 is characterized in that: The training module comprises: A training data acquisition unit, used to acquire training data, wherein the training data includes a training ship berthing request sent by a user through a network communication transmission module, and true values ​​of a plurality of metric matching coefficients; A training request item semantic encoding unit, used for semantically encoding the training ship berthing request to obtain a sequence of semantic feature vectors of the training ship berthing request items; A training global semantic interaction unit is used to pass the sequence of semantic feature vectors of the training ship berthing request items through the deep semantic feature capture module based on the global dependent attention interaction mechanism to obtain the training ship berthing request semantic feature vector; A training available berthing point status extraction unit is used to extract the training real-time status description of multiple available berthing points from the database of the port intelligent control server; A training available anchoring point state understanding unit is used to semantically encode the training real-time state descriptions of the plurality of available anchoring points to obtain a sequence of training available anchoring point state semantic feature vectors; A training cross-modal matching unit, used for passing the sequence of the training ship berthing request semantic feature vector and the training available berthing point status semantic feature vector through the cross-modal attitude metric matching module to obtain a plurality of training metric matching coefficients; A cross entropy calculation unit, used to calculate a cross entropy function value between a training metric matching vector composed of the multiple training metric matching coefficients and a true metric matching vector composed of true values ​​of the multiple metric matching coefficients; A training unit is used to train the deep semantic feature capture module based on the global dependent attention interaction mechanism and the cross-modal attitude measurement matching module using the cross entropy function value as the loss function value, wherein in each round of iteration of the training, each training available anchoring point state semantic feature vector in the sequence of training available anchoring point state semantic feature vectors is optimized.

Images