A buoy collision early warning method based on an artificial intelligence time series prediction algorithm

Abstract

The application provides a navigation mark collision early warning method based on an artificial intelligence time sequence prediction algorithm, and comprises the following steps: determining a navigation mark rotation area according to navigation mark historical data, wherein the navigation mark historical data comprises navigation mark accident historical data and navigation mark historical latitude and longitude information; obtaining ship navigation track data to be measured; inputting the ship navigation track data to be measured into a constructed track prediction model to obtain track prediction data; judging whether the ship to be measured collides with a target navigation mark based on the navigation mark rotation area and the track prediction data; if yes, sending a navigation mark collision warning signal to the ship to be measured through an AIS base station and leaving the navigation mark rotation area. The application solves the problem of low prediction accuracy of ships and navigation marks in a target sea area in the prior art, which causes a large safety hazard.

Description

Technical Field

[0001] This invention relates to the field of maritime traffic safety technology, and in particular to a method for early warning of navigational aid collisions based on an artificial intelligence time series prediction algorithm. Background Technology

[0002] As an important type of navigational aid, navigation marks indicate the direction and extent of a navigable area, while also marking various operational areas or dangerous waters, creating fundamental conditions for the safe and stable navigation of ships. In my country, the basic functions of navigation marks include ship positioning, navigation, and other specialized purposes. With increasing demands for shipping management and technological advancements, the functions of navigation marks have become increasingly sophisticated, evolving from basic to multifunctional. In addition to basic functions, they now include features such as three-dimensional ship monitoring, water pollution monitoring, early warning and decision support, historical information retrieval, and trend analysis and prediction. However, while providing navigational assistance, navigation marks are frequently subjected to collisions with passing vessels. On average, dozens of ship-navigation mark collisions occur annually in various regions, resulting in emergency repair costs reaching millions of yuan each year. Especially in busy ports, ship-navigation mark collisions occur almost monthly, negatively impacting maritime traffic safety.

[0003] Navigational aid collisions occur because navigational aids serve as guides for ships entering coastal areas. Due to factors such as sea conditions and limited visibility, collisions are common when ships approach navigational aids, resulting in damage to the main structure of the aids. This damage primarily includes broken paint on navigational aids, damaged or malfunctioning solar panels, and damaged or malfunctioning navigational lights. These collisions severely impact navigational safety in nearby waters, requiring relevant departments to dispatch vessels for repairs promptly, thus consuming valuable maritime resources.

[0004] The turning radius of a navigation buoy is determined by the fact that the main structure of a navigation buoy is anchored to a concrete block on the seabed via anchor chains. Since the length of the anchor chains is generally greater than the seabed depth and the effects of waves, the main structure of the buoy will reciprocate within a circular area around the concrete block. The radius of this circular area is called the turning radius. Current technologies have relatively low accuracy in predicting the movement of ships and navigation buoys within target sea areas, posing significant safety hazards. Summary of the Invention

[0005] To overcome the shortcomings of existing technologies, the purpose of this invention is to provide a navigation mark collision warning method based on an artificial intelligence time series prediction algorithm. This invention solves the problem that the prediction accuracy of ships and navigation marks in the target sea area is relatively low in existing technologies, resulting in significant safety hazards.

[0006] To achieve the above objectives, the present invention provides the following solution:

[0007] A navigational aid collision warning method based on an artificial intelligence time series prediction algorithm includes:

[0008] The navigation mark turning area is determined based on historical navigation mark data, which includes historical navigation mark accident data and historical latitude and longitude information of the navigation mark.

[0009] Acquire the navigation trajectory data of the vessel under test;

[0010] The navigation trajectory data of the ship to be measured is input into the constructed trajectory prediction model to obtain trajectory prediction data;

[0011] Based on the buoy turning area, and according to the trajectory prediction data, it is determined whether the vessel under test has collided with the target buoy. If so, a buoy collision alarm signal is sent to the vessel under test via the AIS base station, and the vessel leaves the buoy turning area.

[0012] Preferably, determining the turning area of ​​the navigation mark based on historical navigation mark data includes:

[0013] Identify target navigation marks based on historical data of navigation mark accidents;

[0014] The historical trajectory of the target navigation beacon is determined based on its historical latitude and longitude information.

[0015] The turning radius and center of the target beacon are determined based on the historical trajectory.

[0016] The turning area of ​​the target beacon is determined based on its turning radius and center.

[0017] Preferably, the method for constructing the trajectory prediction model is as follows:

[0018] Obtain a sample dataset, which includes a training set and a test set;

[0019] An initial model is constructed based on the row LSTM and GRU algorithms and the training set.

[0020] The initial model is validated based on the test set to obtain the trajectory prediction model.

[0021] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0022] This invention provides a navigation mark collision warning method based on an artificial intelligence time series prediction algorithm, comprising: determining the navigation mark turning area based on historical navigation mark data, wherein the historical navigation mark data includes historical navigation mark accident data and historical latitude and longitude information of the navigation mark; acquiring the navigation trajectory data of the vessel to be tested; inputting the navigation trajectory data of the vessel to be tested into a pre-constructed trajectory prediction model to obtain trajectory prediction data; and, based on the navigation mark turning area and the trajectory prediction data, determining whether the vessel to be tested has collided with the target navigation mark. If so, a navigation mark collision warning signal is sent to the vessel to be tested via an AIS base station, and the vessel is ordered to leave the navigation mark turning area. This invention constructs navigation mark area and vessel prediction data through historical data, thereby improving the accuracy of risk prediction based on the prediction data and navigation mark area. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 A flowchart of a navigation mark collision warning method based on an artificial intelligence time series prediction algorithm is provided in an embodiment of the present invention;

[0025] Figure 2 This is a schematic diagram of the turning radius of a navigation mark provided in an embodiment of the present invention;

[0026] Figure 3 This is a schematic diagram of a vessel in the waters near a navigation mark, provided in an embodiment of the present invention. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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 are within the scope of protection of the present invention.

[0028] The purpose of this invention is to provide a navigation mark collision warning method based on an artificial intelligence time series prediction algorithm. This invention solves the problem that the prediction accuracy of ships and navigation marks in the target sea area is relatively low in the existing technology, which causes a large safety hazard.

[0029] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0030] like Figure 1 As shown, this invention provides a navigational aid collision warning method based on an artificial intelligence time series prediction algorithm, comprising:

[0031] Step 100: Determine the turning area of ​​the navigation mark based on the historical data of the navigation mark, wherein the historical data of the navigation mark includes historical data of navigation mark accidents and historical latitude and longitude information of the navigation mark;

[0032] Step 200: Obtain the navigation trajectory data of the vessel under test;

[0033] Step 300: Input the navigation trajectory data of the ship to be tested into the constructed trajectory prediction model to obtain trajectory prediction data;

[0034] Step 400: Based on the buoy turning area, determine whether the vessel under test has collided with the target buoy according to the trajectory prediction data. If so, send a buoy collision alarm signal to the vessel under test through the AIS base station and sail away from the buoy turning area.

[0035] Furthermore, such as Figure 2-3 As shown, determining the turning area of ​​a navigation mark based on historical navigation mark data includes:

[0036] Identify target navigation marks based on historical data of navigation mark accidents;

[0037] The historical trajectory of the target navigation beacon is determined based on its historical latitude and longitude information.

[0038] The turning radius and center of the target beacon are determined based on the historical trajectory.

[0039] The turning area of ​​the target beacon is determined based on its turning radius and center.

[0040] Specifically, Figure 2 In the diagram, the center is the center of the circumference radius determined by the buoy's historical trajectory in the telemetry and remote control system. The circumference radius is the longest straight-line distance from the center of the buoy's historical trajectory. The cement block is a device used to fix the buoy and prevent it from drifting.

[0041] Figure 3 In the diagram, the smaller circle represents the area indicated by the circumference radius of the navigation mark, while the larger circle represents the area where ships navigate near the navigation mark based on the center of the smaller circle.

[0042] Specifically, based on historical information of navigation marks in the navigation mark telemetry and control system of the South China Sea Navigation Safety Center of the China Maritime Safety Administration, bidding points prone to navigation mark collision accidents are searched, and target navigation marks are selected based on historical navigation mark collision data of the past three years. Based on the historical latitude and longitude information of navigation marks in the navigation mark telemetry and control system of the South China Sea Navigation Safety Center of the China Maritime Safety Administration, the movement trajectory of the navigation mark within the turning radius with ocean tides and sea winds is plotted, and the turning radius and center of the navigation mark are determined according to the historical trajectory. Based on the center of the turning radius of the navigation mark, a length much larger than the turning radius is selected as the navigation area for ships near the navigation mark.

[0043] Furthermore, the method for constructing the trajectory prediction model is as follows:

[0044] Obtain a sample dataset, which includes a training set and a test set;

[0045] An initial model is constructed based on the row LSTM and GRU algorithms and the training set.

[0046] The initial model is validated based on the test set to obtain the trajectory prediction model.

[0047] Specifically, based on the AIS system, historical vessel trajectories are acquired in the waters near selected navigation marks. For each vessel, preprocessing is performed first. After processing the raw data, the vessel navigation data is evenly distributed across the time series. For the same vessel's navigation data, it is divided into training and testing data. The navigation data from the earlier period is used as the training data, and the data from the later period is used as the testing data. After the dataset is divided, the mean and variance are calculated on the training data to standardize it. Following standardization, LSTM and GRU algorithms are applied. After multiple iterations, the calculation results converge. By comparing the predicted vessel trajectories from the training set with those from the test set, the algorithm is considered successful when the prediction accuracy reaches 99% or higher.

[0048] Specifically, the telemetry and remote control system of the South China Sea Navigation Support Center of the China Maritime Safety Administration obtains the current location information (latitude and longitude) of the navigation mark through dual-mode positioning using the US GPS and the Chinese Beidou BD, transmits the information to the system through the SIM card on the navigation mark, and updates and sends the current location information at regular intervals.

[0049] The AIS system mainly consists of an AIS shipboard station and AIS base stations. These base stations can also be networked to form an AIS shore-based network system to expand the service range of AIS. AIS operates primarily on two dedicated VHF radio frequencies, and according to the performance standards of the International Maritime Organization, the system's reporting capacity is at least 2000 reports per minute.

[0050] The artificial intelligence time series prediction algorithms mentioned mainly refer to LSTM and GRU algorithms. LSTM (Long Short-Term Memory) is a special RNN architecture designed to address the limitations of traditional RNNs in handling long-term dependencies. The core of LSTM lies in the introduction of a gate mechanism to control the entry, retention, and exit of information, thereby better capturing long-term temporal dependencies. GRU (Gated Recurrent Unit) is a simplified variant of LSTM, which simplifies the three gates (input gate, forget gate, and output gate) of LSTM to two gates (update gate and output gate).

[0051] The core of LSTM lies in the introduction of a gate mechanism, including an input gate, a forget gate, and an output gate. These gates are responsible for controlling the input, retention, and output of information, respectively. The mathematical model of LSTM is as follows:

[0052] i t =σ(W xi *x t +W hi *h t-1 +b i )

[0053] f t =σ(W xf *x t +W hf *h t-1 +b f )

[0054] o t =σ(W xo *x t +W ho *h t-1 +b o )

[0055] g t =tanh(W xg *x t +W hg *h t-1 +b g )

[0056] c t =f t *c t-1 +i t *g t

[0057] h t =o t *tanh(c t )

[0058] Among them, i t f t o t and g t These represent the input gate, forget gate, output gate, and gate state, respectively. t h represents the hidden state at the current time step. t This indicates the output status at the current time step.

[0059] The core of GRU lies in the introduction of update gates and output gates, which are responsible for updating and outputting control information, respectively.

[0060] The mathematical model for GRU is as follows:

[0061] z t =σ(W xz *x t +W hz *h t-1 +b z )

[0062] r t =σ(W xr *x t +W hr *h t-1 +b r )

[0063]

[0064]

[0065] Among them, z t Indicates the update gate, r t This indicates that the door is being reset. h represents the candidate hidden state. t This indicates the hidden state at the current time step.

[0066] Furthermore, an algorithm with a trained accuracy of 99% is used to predict the navigation trajectory of ships in the waters near the buoy. When it is predicted that a ship is about to enter the buoy's turning radius, a buoy collision warning signal is sent to the ship via the station-based AIS, allowing the ship to leave the buoy's turning radius area.

[0067] The prediction of the vessel's trajectory in the waters near the selected navigation mark in the future refers to determining the trajectory of the selected navigation mark and...

[0068] Given its turning radius, a certain length is selected as the radius of the navigation beacon, centered on the point of its turning radius.

[0069] In near-water areas, based on the form data of ships in the AIS system within the designated waters, artificial intelligence time series algorithms are used to predict the future navigation trajectory of ships. If it is predicted that a ship is about to enter the turning radius area of ​​the navigation mark, the navigation mark is likely to be hit by the ship. Warning information is sent to the ship through the shore-based AIS to make it leave the turning radius area of ​​the navigation mark.

[0070] The beneficial effects of this invention are as follows:

[0071] Based on existing navigation aid telemetry and control systems and AIS systems, this system uses historical navigation aid and ship data to predict the likelihood of navigation aid collisions, enabling high-precision early warning of potential navigation aid collisions and preventing abnormal failures and damage to navigation aids. This is of great significance for maintaining maritime traffic safety and ensuring the high-quality development of maritime navigation support.

[0072] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0073] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A buoy collision early warning method based on an artificial intelligence time series prediction algorithm, characterized in that, The method comprises the following steps: According to the navigation mark historical data, the navigation mark rotation area is determined, and the navigation mark historical data includes: navigation mark accident historical data and navigation mark historical latitude and longitude information; Obtain the ship navigation track data to be tested; The ship navigation track data to be tested is input into the constructed trajectory prediction model to obtain trajectory prediction data; Based on the navigation mark rotation area, it is judged whether the ship to be tested collides with the target navigation mark according to the trajectory prediction data, specifically: it is judged whether the ship is about to travel to the navigation mark rotation area, if yes, the navigation mark collision warning signal is sent to the ship to be tested through the AIS base station, and the ship leaves the navigation mark rotation area; According to the navigation mark historical data, the navigation mark rotation area is determined, which comprises: According to the navigation mark accident historical data, the target navigation mark is determined; According to the navigation mark historical latitude and longitude information, the historical track of the target navigation mark is determined; According to the historical track, the rotation radius and the center of the target navigation mark are determined; Based on the rotation radius and the center of the target navigation mark, the navigation mark rotation area is determined.

2. The buoy collision warning method based on the artificial intelligence time series prediction algorithm according to claim 1, characterized in that, The construction method of the trajectory prediction model is: Obtain the sample data set, which includes the training set and the test set; Based on the algorithm of row LSTM and GRU, the initial model is constructed according to the training set; According to the test set, the initial model is verified to obtain the trajectory prediction model.

Images