Intelligent berthing method for inland river anchorage ground

Abstract

The invention relates to an intelligent berthing method for an inland river anchorage ground, and the method comprises the following steps: S1, a ship sends a berthing application, and obtains the current information of the ship; s2, predicting the time when the ship arrives at each mooring anchor ground according to the ship information; s3, carrying out statistics on information of all current berthing ships on the anchor ground, and carrying out berth distribution on the berthing ships on the basis of the current anchoring ground berthing condition; s4, after berthing of the ship is completed, whether the ship is berthed at the designated position according to requirements or not is checked; s5, after the ship receives the off-berthing information, counting the off-berthing time of the ship for berth distribution of the later berthing ship; and S6, after the preset off-berthing time is reached, the ship is subjected to off-berthing operation. The method can predict the arrival time of the ship and carry out intelligent berth distribution on the berthing ship, thereby reducing the management cost of the anchor ground, improving the operation efficiency of the anchor ground, and improving the intelligent construction of the anchor ground.

Description

Technical Field

[0001] The present invention relates to the technical field of ship dispatching, and more specifically to an intelligent berthing method for an inland river anchorage. Background Art

[0002] As an important public facility for water transport production and shipping economic development, anchorages play a vital role in the safe operation of ships and directly affect the smooth passage of waterways and navigation safety. With the rapid development of waterway transportation, the inland river ship traffic volume has become denser and the ship types have become larger, and the demand for anchorages is in urgent need of an increase. Restricted by shoreline resources, the number and scale of inland waterway anchorages has tended to be saturated. How to utilize the current limited anchorage resources to maximize the efficiency of anchorage utilization has become an urgent problem to be solved. In addition, inland waterway anchorages are geographically dispersed and relatively remote, with low coverage of information, digitalization and networking, and low degree of centralized management of anchorages. The current anchorage management model still relies on on-site guards, which greatly increases the difficulty and cost of anchorage management. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide an intelligent mooring method for an inland anchorage, which can realize remote centralized control and management of the anchorage, significantly reduce the anchorage management cost, improve the anchorage operation efficiency, and enhance the intelligent construction of the anchorage.

[0004] The technical solution adopted by the present invention to solve the technical problem is: constructing an intelligent mooring method for an inland anchorage, comprising the following steps: S1. The ship sends a berth application and obtains the current information of the ship; S2. Predicting the time for the ship to arrive at each available anchorage based on the ship information; S3, collecting information of all ships currently moored at the anchorage, and allocating berths to incoming ships based on the current mooring situation at the anchorage; S4. After the ship has moored, check whether the ship is moored at the designated location as required; S5. When the ship receives the departure information, the departure time of the ship is counted for the berth allocation of the next ship to berth; S6. After reaching the scheduled unberthing time, the vessel performs unberthing operations; In step S3, by constructing a mathematical model, based on the information of all the moored ships at the anchorage and the current occupancy of the anchorage, a genetic algorithm is used to solve the result of the ship berth allocation. The model objective function F minimizes the average berthing and unberthing effects of the ships as follows: (1) The constraints of this model include the number of anchorage points and berths, ship berthing conditions, and ship arrival and departure times, which are expressed as follows: (2) (3) (4) (5) N: represents the anchor point set, , n is the total number of anchor points, : indicates the total number of available berths at the nth anchorage point, , , m represents the mth berth at the nth anchor point, I: indicates the number of ships expected to arrive, i indicates the i-th ship, V: indicates the number of ships assigned berths, : represents the length of ship i, , : represents the length of the inner ship adjacent to the berth i, , : represents the length of the outer ship adjacent to the berth i of the ship, , : indicates the estimated arrival time of ship i, , : represents the departure time of ship i, , : indicates the number of berthing times required for ship i during berthing maneuvers, , : Indicates the number of berthing times required for the ship inside the berth to leave after the ship i berths. , : indicates the number of times the ships outside the berth need to move when the ship i leaves the berth. , : represents the total number of berthing times caused by ship i's berthing and leaving. , The impact of ship i's berthing and leaving berth includes three parts: 1) When ship i is berthing, is it necessary for the ships outside the berth to move? Indicates the number of times berthing is required when berthing; 2) After ship i berths, when the ship on the inner side of the berth leaves, is it necessary for ship i to move berth? Indicates the number of times a berth needs to be moved after berthing; 3) When ship i leaves the berth, whether the ship outside the berth needs to move to the berth for ship i, Indicates the number of times the berth needs to be moved when leaving the berth; The total number of berthing times caused by ship i berthing and leaving berth is express.

[0005] According to the above scheme, in step S1, the ship automatic identification system data is used to obtain the basic information of the ship through decoding, and the current information of the ship is obtained in combination with the ship declaration information. The ship automatic identification system is Automatic Identification System, referred to as AIS.

[0006] According to the above scheme, the ship declaration information includes: ship name, ship type, cargo type, ship length, ship width, ship height, draft ship size, actual cargo / passenger capacity, actual displacement, declaration time, and navigation scheduling information.

[0007] According to the above scheme, in step S2, a long short-term memory network deep learning model is used to predict the time when the ship arrives at each available anchorage. The long short-term memory network is Long Short-Term Memory, referred to as LSTM. Building the LSTM model includes the following steps: S201. Define model architecture: Use the deep learning framework TensorFlow to define the LSTM model architecture; S202. Select the appropriate number of layers and neurons according to the input data: use multiple LSTM layers, each layer contains a number of neurons; S203, add a Dense layer: add multiple Dense layers after the LSTM layer, and map the output of the LSTM layer to the target variable; S204. Compile the model: use the mean square error to formulate the loss function and use the Adam optimizer to optimize the model.

[0008] According to the above scheme, before the construction and training phase of the LSTM deep learning model, the historical ship navigation data, AIS data, weather data, and anchorage data are collected, and the ship position data, timestamp data, environmental data, ship information, and ship arrival time are obtained from the collected data; After obtaining the collected data, preprocess the data for model training; In the training phase of the model, the training input is time series data with multi-dimensional features, and the training output is the estimated arrival time of the ship; When using the LSTM deep learning model to predict the time for a ship to arrive at each available anchorage, the current ship type, cargo type, ship length, ship width, ship height, draft, actual cargo / passenger capacity, actual displacement, current weather data, anchorage data, and AIS data at the current time and several time steps before are preprocessed and used as input.

[0009] According to the above scheme, the data preprocessing includes time series processing, missing value processing and normalization processing; the time series processing organizes the data into a time series format, the missing value processing processes or fills in the missing data, and the normalization processing normalizes or standardizes the feature data.

[0010] According to the above scheme, in step S4, the specific steps of checking whether the ship is moored at the designated location as required are as follows: S401, obtaining AIS data: the data includes static data and dynamic data of the ship. The data used when checking the berthing position of the ship includes the position, heading, speed, ship name, and MMSI number of the ship; S402, identifying the target ship: identifying the target ship from the AIS data, and ensuring that the selected ship is consistent with the inspected ship through the MMSI number corresponding to each ship; S403, verifying the ship's position information: checking the position information in the AIS data, including longitude and latitude, to verify whether the information is consistent with the requirements of the designated berthing area. If the ship's position information is within the allocated berthing area, the ship's berthing position is correct; S404, AIS data timestamp check: check the timestamp of AIS data to ensure that the data is current and real-time, so as to confirm the exact position of the ship in real time; S405, ship communication status check: To ensure that the ship's AIS equipment is in normal working condition, check the communication status in the AIS data.

[0011] The intelligent mooring method for inland river anchorages of the present invention has the following beneficial effects: 1. The present invention comprehensively utilizes modern information technology to carry out research on key technologies for intelligent operation of unmanned anchorages, realize remote centralized control and management of anchorages, and provide effective and reliable solutions for unmanned and intelligent operation of anchorages; 2. The present invention can streamline the number of anchorage personnel, reduce anchorage management costs, improve anchorage management efficiency, enhance anchorage business management level, tap into inland anchorage resources, innovate anchorage management models, promote intelligent digital construction of anchorages, and improve the intelligence level of shipping supporting facilities and equipment; 3. The present invention can realize remote centralized control and management of anchorages, which can help anchorage management departments streamline anchorage duty personnel, significantly reduce anchorage management costs, improve anchorage operation efficiency, and enhance the intelligent construction of anchorages; 4. The present invention uses the long short-term memory network deep learning model to predict the arrival time of ships, which can estimate the arrival of multiple ships in advance, realize the comprehensive management of multiple ships, improve the practicality and rationality of ship management, and improve the accuracy of ship berthing; 5. The present invention establishes a mathematical model based on the requirements for ships to berth, which can increase the safety of ships during berthing, and uses a genetic algorithm to seek the optimal solution of the model, which can improve the comprehensive utilization rate of the anchorage and reduce the impact of ships on other ships during berthing. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The present invention will be further described below with reference to the accompanying drawings and embodiments, in which: Figure 1 It is a flow chart of the intelligent mooring method for inland river anchorage of the present invention; Figure 2 It is a flow chart of checking the berthing position of a ship according to the present invention. DETAILED DESCRIPTION

[0013] In order to have a clearer understanding of the technical features, purposes and effects of the present invention, specific embodiments of the present invention are now described in detail with reference to the accompanying drawings.

[0014] like Figure 1 As shown, the inland river anchorage intelligent mooring method of the present invention comprises the following steps: S1. Predict the time when the ship will arrive at each available anchorage through ship information. Use the ship's Automatic Identification System (AIS) data to obtain the basic information of the ship through decoding, and combine it with the ship's declared information to obtain the current information of the ship. The ship's declared information at least includes: ship name, ship type, cargo type, length, ship width, ship height, draft and other ship dimensions, actual displacement, actual cargo / passenger capacity, declaration time, and other information required for navigation scheduling.

[0015] S2: After entering the lock waters, the ship sends the berth application and ship information to the cloud server. The Long Short-Term Memory (LSTM) deep learning model is used to predict the time for the ship to arrive at each available anchorage.

[0016] Before the model is built and trained, historical ship navigation data, AIS data, weather data, and anchorage data are collected. The ship position data, timestamp data, environmental data, ship information, and ship arrival time are obtained from the collected data. The ship position data includes longitude, latitude, speed, heading, etc. The timestamp data is used to record the time of the ship position data. Environmental data includes weather conditions, anchorage conditions, etc. The ship information includes ship type, cargo type, length, width, height, draft, actual displacement, actual cargo / passenger capacity, etc. The arrival time refers to the actual arrival time of the ship.

[0017] After obtaining the above data, the data is preprocessed for model training. Data preprocessing includes time series processing, that is, organizing the data into a time series format; missing value processing, that is, processing or filling missing data; normalization processing, that is, normalizing or standardizing feature data to improve the model training effect. And convert the time series data into a format acceptable to the model. For LSTM, the data needs to be converted into a 3D array format: [number of samples, time steps, number of features].

[0018] The method of building an LSTM deep learning model includes the following steps: 1) Define the model architecture: Use the deep learning framework TensorFlow to define the LSTM model architecture.

[0019] 2) Choose the appropriate number of layers and neurons based on the input data: Use multiple LSTM layers, each containing a number of neurons.

[0020] 3) Add Dense layers: Add multiple Dense layers after the LSTM layer to map the output of the LSTM layer to the target variable.

[0021] 4) Compile the model: use the mean square error to formulate the loss function and use the Adam optimizer to optimize the model.

[0022] In the training phase of the model, the training input is time series data with multi-dimensional features, i.e., ship position data, timestamp data, environmental data, ship information, etc. at several past time steps. The training output is the estimated arrival time of the ship.

[0023] When using the LSTM model to predict the time of a ship's arrival at each available anchorage, the current ship type, cargo type, ship length, ship width, ship height, draft, actual cargo / passenger capacity, actual displacement, current weather data, anchorage data, and AIS data at the current time and several time steps before are preprocessed as input, that is, time series data in the same format as during training, and the output is the predicted value, that is, the time of the ship's arrival at each available anchorage.

[0024] S3. Count all the information of the ships currently moored at the anchorage, and allocate berths to the ships based on the current mooring situation at the anchorage.

[0025] According to the ship type and the type of cargo loaded, ships are divided into five categories: Class I flammable and explosive dangerous goods ships, Class II flammable and explosive dangerous goods ships, non-flammable and explosive dangerous goods ships, bulk carriers, and container ships. When allocating berths, for safety reasons, different types of ships are allocated to corresponding anchorages. Class I flammable and explosive dangerous goods ships and Class II flammable and explosive dangerous goods ships are allocated to different dangerous goods anchorages, non-flammable and explosive dangerous goods ships and bulk carriers are allocated to bulk carrier anchorages, and container ships are allocated to container ship anchorages. There are usually multiple anchorages of the same type, and ships generally have multiple anchorages where they can berth.

[0026] When ships are berthing together, in order to provide more contact area between the ships, the contact area between the inner ship and the anchor point is increased, thereby improving the stability of the ship when berthing. When berthing together, the length of the inner berth is required to be greater than the length of the outer berth, which helps to reduce the shaking and swinging of the ship during the berthing process. Therefore, the requirements for berthing together are: (1) the outer berth can only be docked after the inner berth is docked; (2) the length of the inner berth must be greater than the length of the outer berth.

[0027] Since other ships at the same anchorage need to perform berthing operations when a ship berths or leaves, in order to reduce the impact of the ship on other ships when the ship berths or leaves, a mathematical model is constructed, and the genetic algorithm is used to solve the result of ship berth allocation based on the information of all anchored ships and the current occupancy of the anchorage.

[0028] The model objective function F minimizes the average berthing and departure effects of ships and is expressed as follows: (1) The constraints of this model include the number of anchorage points and berths, ship berthing conditions, and ship arrival and departure times, which are expressed as follows: (2) (3) (4) (5) N: represents the anchor point set, , n is the total number of anchor points.

[0029] : indicates the total number of available berths at the nth anchorage point, , , m represents the mth berth at the nth anchor point.

[0030] I: represents the number of ships expected to arrive, i represents the i-th ship.

[0031] V: represents the number of ships assigned berths.

[0032] : represents the length of ship i, .

[0033] : represents the length of the inner ship adjacent to the berth i, .

[0034] : represents the length of the outer ship adjacent to the berth i of the ship, .

[0035] : indicates the estimated arrival time of ship i, .

[0036] : represents the departure time of ship i, .

[0037] : indicates the number of berthing times required for ship i during berthing maneuvers, .

[0038] : Indicates the number of berthing times required for the ship inside the berth to leave after the ship i berths. .

[0039] : indicates the number of times the ships outside the berth need to move when the ship i leaves the berth. .

[0040] : represents the total number of berthing times caused by ship i's berthing and leaving. .

[0041] The impact of ship i on berthing and leaving berth mainly includes three parts: 1) When ship i is berthing, is it necessary for the ships outside the berth to move? Indicates the number of times berthing is required when berthing; 2) After ship i berths, when the ship on the inner side of the berth leaves, is it necessary for ship i to move berth? Indicates the number of times a berth needs to be moved after berthing; 3) When ship i leaves the berth, whether the ship outside the berth needs to move to the berth for ship i, Indicates the number of times the vessel needs to move berth when leaving a berth.

[0042] The total number of berthing times caused by ship i berthing and leaving berth is express.

[0043] The model is solved using a genetic algorithm. Depending on the specific problem, the genetic algorithm first encodes each feasible solution into an individual through the genetic algorithm encoding method, and a population is composed of multiple individuals. Next, by simulating the evolutionary laws of "natural selection" and "survival of the fittest" in nature, the fitness function of each individual is calculated to determine whether to retain it; if the individual passes the crossover probability judgment, the selected individual will cross over to produce a new individual and join the population; if the individual passes the mutation probability judgment, the selected individual will mutate to produce a new individual to improve the diversity of the population and join the population; this continues generation by generation until the algorithm termination condition is met. Finally, the individual with the largest fitness value in the current population is selected as the approximate optimal solution to the problem.

[0044] S4. Check whether the ship is moored at the designated location as required, such as Figure 2 As shown, the following steps are included: S401. Obtaining ship AIS data: These data include ship static data and ship dynamic data. The data mainly used when checking the berthing position of a ship are the ship's position, heading, speed, ship name, MMSI number and other data.

[0045] S402, identifying the target ship: identifying the target ship from the AIS data, and ensuring that the selected ship is consistent with the ship to be inspected through the MMSI number corresponding to each ship.

[0046] S403, verifying the ship's position information: checking the position information in the AIS data, including longitude and latitude, and verifying whether the longitude and latitude are consistent with the requirements of the designated berthing area. If the ship's position information is within the allocated berthing area, the ship's berthing position is correct.

[0047] S404, AIS data timestamp check: Check the timestamp of AIS data to ensure that the data is current and real-time, so as to confirm the exact position of the ship in real time.

[0048] S405, ship communication status check: Check the communication status in the AIS data to ensure that the ship's AIS equipment is in normal working condition, which helps to maintain real-time position information and smooth communication in abnormal situations.

[0049] S5. In the constructed mathematical model of ship berth allocation, information such as the ship's departure time is required when solving the ship berth. When the ship receives the departure information, the ship's departure time is counted for the berth allocation of subsequent ships.

[0050] S6. After arriving at the scheduled unberthing time, the vessel performs unberthing operations.

[0051] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, all of which are within the protection of the present invention.

Claims

1. An intelligent mooring method for inland anchorages, characterized in that: The following steps are involved: S1. The ship sends a berth application and obtains the current information of the ship; S2. Predicting the time for the ship to arrive at each available anchorage based on the ship information; S3, collecting information of all ships currently moored at the anchorage, and allocating berths to incoming ships based on the current mooring situation at the anchorage; S4. After the ship has moored, check whether the ship is moored at the designated location as required; S5. When the ship receives the departure information, the departure time of the ship is counted for the berth allocation of the next ship to berth; S6. After reaching the scheduled departure time, the vessel performs departure operations; In step S3, by constructing a mathematical model, based on the information of all the moored ships at the anchorage and the current occupancy of the anchorage, a genetic algorithm is used to solve the result of the ship berth allocation. The model objective function F minimizes the average berthing and unberthing effects of the ships as follows: （1） The constraints of this model include the number of anchorage points and berths, ship berthing conditions, and ship arrival and departure times, which are expressed as follows: （2） （3） （4） （5） N: represents the anchor point set, , n is the total number of anchor points, : indicates the total number of available berths at the nth anchorage point, , , m represents the mth berth at the nth anchor point, I: indicates the number of ships expected to arrive, i indicates the i-th ship, V: indicates the number of ships assigned berths, : represents the length of ship i, , : represents the length of the inner ship adjacent to the berth i, , : represents the length of the outer ship adjacent to the berth i of the ship, , : indicates the estimated arrival time of ship i, , : represents the departure time of ship i, , : indicates the number of berthing times required for ship i during berthing maneuvers, , : Indicates the number of berthing times required for the ship inside the berth to leave after the ship i berths. , : indicates the number of times the ships outside the berth need to move when the ship i leaves the berth. , : represents the total number of berthing times caused by ship i's berthing and leaving. , The impact of ship i's berthing and leaving berth includes three parts: 1) When ship i is berthing, is it necessary for the ships outside the berth to move? Indicates the number of times berthing is required when berthing; 2) After ship i berths, when the ship on the inner side of the berth leaves, is it necessary for ship i to move berth? Indicates the number of times a berth needs to be moved after berthing; 3) When ship i leaves the berth, whether the ship outside the berth needs to move to the berth for ship i, Indicates the number of times the berth needs to be moved when leaving the berth; The total number of berthing times caused by ship i berthing and leaving berth is express.

2. The intelligent mooring method for inland river anchorage according to claim 1 is characterized in that: In the step S1, the ship's automatic identification system (AIS) data is used to obtain basic information of the ship through decoding, and the current information of the ship is obtained in combination with the ship's declared information.

3. The intelligent mooring method for inland river anchorage according to claim 2 is characterized in that: The ship declaration information includes: ship name, ship type, cargo type, ship length, ship width, ship height, draft and ship size, actual cargo / passenger capacity, actual displacement, declaration time, and navigation scheduling information.

4. The intelligent mooring method for inland river anchorage according to claim 1 is characterized in that: In step S2, a long short-term memory network (LSTM) deep learning model is used to predict the time when the ship arrives at each available anchorage. The method for constructing the LSTM deep learning model includes the following steps: S201. Define model architecture: Use the deep learning framework TensorFlow to define the LSTM model architecture; S202. Select the appropriate number of layers and neurons according to the input data: use multiple LSTM layers, each layer contains a number of neurons; S203, add a Dense layer: add multiple Dense layers after the LSTM layer, and map the output of the LSTM layer to the target variable; S204. Compile the model: use the mean square error to formulate the loss function and use the Adam optimizer to optimize the model.

5. The intelligent mooring method for inland river anchorage according to claim 4 is characterized in that: Before the construction and training phase of the LSTM deep learning model, historical ship navigation data, AIS data, weather data, and anchorage data are collected, and ship position data, timestamp data, environmental data, ship information, and ship arrival time are obtained from the collected data; After obtaining the collected data, preprocess the data for model training; In the training phase of the model, the training input is time series data with multi-dimensional features, and the training output is the estimated arrival time of the ship; When using the LSTM deep learning model to predict the time for a ship to arrive at each available anchorage, the current ship type, cargo type, ship length, ship width, ship height, draft, actual cargo / passenger capacity, actual displacement, current weather data, anchorage data, and AIS data at the current time and several time steps before are preprocessed and used as input.

6. The intelligent mooring method for inland river anchorage according to claim 5 is characterized in that: The data preprocessing includes time series processing, missing value processing and normalization processing; the time series processing organizes the data into a time series format, the missing value processing processes or fills in the missing data, and the normalization processing normalizes or standardizes the feature data.

7. The intelligent mooring method for inland river anchorage according to claim 1, characterized in that: In step S4, the specific steps of checking whether the ship is moored at the designated location as required are as follows: S401, obtaining AIS data: the data includes static data and dynamic data of the ship. The data used when checking the berthing position of the ship includes the position, heading, speed, ship name, and MMSI number of the ship; S402, identifying the target ship: identifying the target ship from the AIS data, and ensuring that the selected ship is consistent with the inspected ship through the MMSI number corresponding to each ship; S403, verifying the ship's position information: checking the position information in the AIS data, including longitude and latitude, to verify whether the information is consistent with the requirements of the designated berthing area. If the ship's position information is within the allocated berthing area, the ship's berthing position is correct; S404, AIS data timestamp check: check the timestamp of AIS data to ensure that the data is current and real-time, so as to confirm the exact position of the ship in real time; S405, ship communication status check: To ensure that the ship's AIS equipment is in normal working condition, check the communication status in the AIS data.

Images