CN113408711A - Ship motion extremely-short-term forecasting method and system based on LSTM neural network - Google Patents
Ship motion extremely-short-term forecasting method and system based on LSTM neural network Download PDFInfo
- Publication number
- CN113408711A CN113408711A CN202110803068.3A CN202110803068A CN113408711A CN 113408711 A CN113408711 A CN 113408711A CN 202110803068 A CN202110803068 A CN 202110803068A CN 113408711 A CN113408711 A CN 113408711A
- Authority
- CN
- China
- Prior art keywords
- data
- neural network
- ship motion
- lstm neural
- initial data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a method and a system for forecasting ship motion in an extremely short period based on an LSTM neural network, and relates to the forecasting of the ship motion in the extremely short period. The method for forecasting the ship motion in the extremely short term based on the LSTM neural network comprises the steps of obtaining initial data, wherein the initial data comprises ship motion data, other ship motion attitude data and wave height data, preprocessing the initial data to generate effective initial data, forecasting the ship motion by adopting various preset LSTM neural network models according to the effective initial data to generate various forecasting results, comparing the various forecasting results to generate a comparison result, and obtaining final forecasting result information according to the comparison result, so that the forecasting results are forecasted in multiple aspects, the forecasting results are obtained by comparing the forecasting results, the smaller the difference among the forecasting results is, the more reliable the forecasting results are indicated, and the accuracy of the extremely short term forecasting of the ship motion is improved.
Description
Technical Field
The invention relates to the field of ship motion extremely-short-term forecasting, in particular to a ship motion extremely-short-term forecasting method and system based on an LSTM neural network.
Background
The extremely short-term forecasting of the ship motion refers to forecasting of the motion attitude of the ship from several seconds to ten and several seconds in the future, and the effective realization of the method is favorable for guaranteeing the safety of the ship in sailing and operation on the sea. In recent years, with the rapid development of artificial neural network algorithms, more and more learners use artificial neural networks to forecast the ship motion in a very short period.
At present, the problem of inaccurate prediction exists in the extremely short-term prediction of ship motion by adopting an artificial neural network.
Disclosure of Invention
The invention aims to provide a method and a system for forecasting ship motion in an extremely short period based on an LSTM neural network, which are used for solving the problem that forecasting of the ship motion in the extremely short period by adopting an artificial neural network in the prior art is not accurate enough.
In a first aspect, an embodiment of the present application provides a ship motion very short-term forecasting method based on an LSTM neural network, which includes the following steps:
acquiring initial data, wherein the initial data comprises ship motion data, other ship motion attitude data and wave height data;
preprocessing the initial data to generate effective initial data;
respectively adopting a plurality of preset LSTM neural network models to predict the ship motion according to the effective initial data so as to generate a plurality of prediction results;
and comparing the various prediction results to generate a comparison result, and obtaining final prediction result information according to the comparison result.
In the implementation process, initial data is obtained, the initial data comprises ship motion data, other ship motion attitude data and wave height data, the initial data is preprocessed to generate effective initial data, noise data can be removed through preprocessing the initial data, the influence of data magnitude difference can be reduced, meanwhile, the training convergence speed is accelerated, the calculation efficiency is improved, the ship motion is predicted by adopting a plurality of preset LSTM neural network models respectively according to the effective initial data to generate a plurality of prediction results, the prediction is carried out by adopting different input data from several aspects, the prediction results are compared to generate a comparison result, and the final prediction result information is obtained according to the comparison result, so that the prediction result is predicted according to one input data, but the forecasting is carried out from multiple aspects, and the forecasting results are compared, so that the smaller the difference between the forecasting results is, the more ready the forecasting results are, and the accuracy of the extremely short-term forecasting of the ship motion is improved.
Based on the first aspect, in some embodiments of the present invention, the step of predicting the ship motion by using a plurality of preset LSTM neural network models according to the valid initial data to generate a plurality of prediction results includes the following steps:
extracting and inputting the ship motion data in the effective initial data into a preset first LSTM neural network model to generate a first prediction result;
extracting and inputting the ship motion data and other ship motion attitude data in the effective initial data into a preset second LSTM neural network model to generate a second prediction result;
and extracting and inputting the ship motion data and the wave height data in the effective initial data into a preset third LSTM neural network model to generate a third prediction result.
Based on the first aspect, in some embodiments of the present invention, the step of preprocessing the initial data to generate valid initial data includes the following steps:
inputting the initial data into a preset smoothing filter to generate preprocessed data;
and substituting the preprocessed data into a preset normalized mathematical expression for calculation to generate effective initial data.
Based on the first aspect, in some embodiments of the present invention, the normalized mathematical expression is:
Based on the first aspect, in some embodiments of the present invention, the method further comprises the following steps:
acquiring sample initial data; the initial data comprises ship motion attitude data and wave height data;
obtaining model parameters;
and constructing an LSTM neural network framework according to the model parameters and the sample initial data and carrying out iterative training to obtain an LSTM neural network model.
Based on the first aspect, in some embodiments of the present invention, the step of obtaining model parameters includes the following steps:
setting initial model parameters according to the initial data of the sample;
and performing optimization calculation on the initial model parameters by adopting a particle swarm optimization algorithm to generate optimized model parameters, and taking the optimized model parameters as the model parameters.
Based on the first aspect, in some embodiments of the invention, the following steps are further included;
extracting and carrying out multilayer LSTM neural network calculation on a data vector group in the sample initial data to generate target data;
substituting the target data into a preset loss function expression for calculation to obtain a loss value;
and optimizing the LSTM neural network model by adopting a preset optimizer according to the loss value so as to generate the optimized LSTM neural network model.
Based on the first aspect, in some embodiments of the present invention, the step of comparing the plurality of prediction results to generate a comparison result, and obtaining final prediction result information according to the comparison result includes the following steps:
comparing any two prediction results in the multiple prediction results to obtain a plurality of comparison results;
comparing the comparison results with a preset threshold value respectively, and if the comparison results are smaller than the preset threshold value, taking any one of the prediction results as final prediction result information; and if at least one comparison result is not smaller than a preset threshold value, acquiring initial data.
In a second aspect, an embodiment of the present application provides a ship motion very short term forecasting system based on an LSTM neural network, including:
the data acquisition module is used for acquiring initial data; the initial data comprises ship motion data, other ship motion attitude data and wave height data;
the data preprocessing module is used for preprocessing the initial data to generate effective initial data;
the LSTM neural network model prediction module is used for predicting ship motion by adopting various preset LSTM neural network models according to effective initial data so as to generate various prediction results;
and the prediction result comparison module is used for comparing various prediction results to generate and obtain final prediction result information according to the comparison result.
In the implementation process, initial data is obtained through a data obtaining module, the initial data comprises ship motion data, other ship motion attitude data and wave height data, then the data preprocessing module preprocesses the initial data to generate effective initial data, noise data can be removed through preprocessing the initial data, the influence of data magnitude difference can be reduced, meanwhile, the training convergence speed is accelerated, and the calculation efficiency is improved, then an LSTM neural network model prediction module predicts the ship motion by respectively adopting a plurality of preset LSTM neural network models according to the effective initial data to generate a plurality of prediction results, the prediction results are predicted by adopting different input data from several aspects, then a prediction result comparison module compares the plurality of prediction results to generate a comparison result, and final prediction result information is obtained according to the comparison result, therefore, the prediction results are not only predicted according to input data, but also predicted from multiple aspects, and the prediction results are compared, so that the smaller the difference between the prediction results is, the more ready the prediction results are, and the accuracy of the extremely short-term prediction of the ship motion is improved.
In a third aspect, an embodiment of the present application provides an electronic device, which includes a memory for storing one or more programs; a processor. The one or more programs, when executed by the processor, implement the method as described in any of the first aspects above.
The embodiment of the invention at least has the following advantages or beneficial effects:
the embodiment of the invention provides a ship motion extremely short-term forecasting method and a system based on an LSTM neural network, which can eliminate noise data by acquiring initial data comprising ship motion data, other ship motion attitude data and wave height data, then preprocess the initial data to generate effective initial data, remove the influence of data magnitude difference by preprocessing the initial data, accelerate the training convergence speed and improve the calculation efficiency, then respectively predict the ship motion by adopting a plurality of preset LSTM neural network models according to the effective initial data to generate a plurality of forecasting results, predict by adopting different input data from several aspects, compare the plurality of forecasting results to generate a comparison result, and obtain final forecasting result information according to the comparison result, therefore, the prediction results are not only predicted according to input data, but also predicted from multiple aspects, and the prediction results are compared, so that the smaller the difference between the prediction results is, the more ready the prediction results are, and the accuracy of the extremely short-term prediction of the ship motion is improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present invention and therefore should not be considered as limiting the scope, and for those skilled in the art, other related drawings can be obtained according to the drawings without inventive efforts.
Fig. 1 is a flowchart of a very short-term forecasting method for ship motion based on an LSTM neural network according to an embodiment of the present invention;
fig. 2 is a block diagram of a very short-term forecasting system for ship motion based on an LSTM neural network according to an embodiment of the present invention;
fig. 3 is a block diagram of an electronic device according to an embodiment of the present invention.
Icon: 110-a data acquisition module; 120-a data pre-processing module; 130-LSTM neural network model prediction module; 140-prediction result comparison module; 101-a memory; 102-a processor; 103-communication interface.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments. The components of the embodiments of the present application, generally described and illustrated in the figures herein, can be arranged and designed in a wide variety of different configurations.
Thus, the following detailed description of the embodiments of the present application, presented in the accompanying drawings, is not intended to limit the scope of the claimed application, but is merely representative of selected embodiments of the application. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
Examples
Some embodiments of the present application will be described in detail below with reference to the accompanying drawings. The embodiments described below and the individual features of the embodiments can be combined with one another without conflict.
Referring to fig. 1, fig. 1 is a flowchart of a very short-term ship motion forecasting method based on an LSTM neural network according to an embodiment of the present invention. The ship motion extremely short-term forecasting method based on the LSTM neural network comprises the following steps:
step S110: acquiring initial data; the initial data comprises ship motion data, other ship motion attitude data and wave height data; the ship motion data comprises ship rolling motion duration data, and the other ship motion attitude data comprises ship six-degree-of-freedom motion data; the wave height data includes wave height data at different locations. The initial data can be obtained by setting wave height monitoring points and various sensors on the ship.
Step S120: preprocessing the initial data to generate effective initial data; the data preprocessing is to process each group of data distribution, and the preprocessing comprises the following steps:
firstly, inputting initial data into a preset smoothing filter to generate preprocessed data; the smoothing filter is mainly used for smoothing and filtering the initial data to remove noise in the initial data, so that useful initial data is obtained. The smoothing filter can be obtained by adopting a moving average smoothing filtering algorithm and a wavelet decomposition filtering algorithm, which belong to the prior art and are not described herein again.
The preprocessed data is then substituted into a preset normalized mathematical expression for computation to generate valid initial data. The preprocessed data is nonlinear data, and a smooth preprocessed data curve is obtained through normalization processing, wherein the normalization mathematical expression is as follows:
Step S130: respectively adopting a plurality of preset LSTM neural network models to predict the ship motion according to the effective initial data so as to generate a plurality of prediction results; the LSTM neural network is a special circulating neural network, can effectively solve the problem of gradient explosion of the traditional circulating neural network, can process long-time sequence data, and has long-time memory capacity. The prediction of the ship motion by utilizing the multiple preset LSTM neural network models mainly comprises the following steps:
firstly, extracting and inputting ship motion data in effective initial data into a preset first LSTM neural network model to generate a first prediction result; the first LSTM neural network model is an LSTM neural network model established by using ship motion data, and data prediction is carried out by inputting the ship motion data. For example, a set of ship rolling motion duration data acquired from the effective initial data is input into the first LSTM neural network model to predict rolling, and rolling data of ship motion is obtained through prediction. The forecasting method carries out ship motion forecasting by mining the internal rule of the data of the forecasting method.
Secondly, extracting and inputting the ship motion data and other ship motion attitude data in the effective initial data into a preset second LSTM neural network model to generate a second prediction result; the second LSTM neural network model is an LSTM neural network model established by using ship motion attitude data. Considering that six free motions of the ship are mutually coupled, the motion postures of a plurality of ships are used as input by utilizing the characteristic of LSTM multi-feature input to carry out extremely short-term prediction on certain motion of the ship, so that the prediction precision can be improved.
And thirdly, extracting and inputting the ship motion data and the wave height data in the effective initial data into a preset third LSTM neural network model to generate a third prediction result. The third LSTM neural network model is mainly an LSTM neural network model established by using a mapping relationship between a wave height and a ship. In the operation process of the ship in the waves, because the waves are the main excitation factors of the swing motion generated by the ship, the mapping relation between the wave height and the ship motion can be established through the LSTM neural network, and because the waves need a certain time for being propagated in the space and have the memory effect, the optimal response relation between the waves and the ship motion has a certain time delay, the ship motion is forecasted in an extremely short period through the waves, and the forecasting duration can be effectively prolonged. Considering that the wave height may be lost in a short time in the acquisition process, the LSTM neural network can be used for predicting the wave height to fill up the missing segments, so that more accurate predicted data can be obtained.
The three LSTM neural network model prediction modes can be carried out independently or simultaneously.
Step S140: and comparing the various prediction results to generate a comparison result, and obtaining final prediction result information according to the comparison result. The above comparison process includes the following processes:
firstly, comparing any two prediction results in a plurality of prediction results to obtain a plurality of comparison results; the comparison means that the difference value of any two prediction results in a plurality of prediction results is compared, for example, the prediction result A obtained by predicting by using the first LSTM neural network model is 0.028, the prediction result B obtained by predicting by using the second LSTM neural network model is 0.025, and the prediction result C obtained by predicting by using the third LSTM neural network model is 0.027; and comparing the prediction result A with the prediction result B to obtain a difference value between the prediction result A and the prediction result B of 0.002, comparing the prediction result A with the prediction result C to obtain a difference value between the prediction result A and the prediction result C of 0.001, and comparing the prediction result B with the prediction result C to obtain a difference value between the prediction result B and the prediction result C of 0.002.
Then, the comparison results are respectively compared with the preset threshold value, due to the fact that ship motion is subjected to extremely short-term prediction, the data have strong continuity, the predicted value and the actual value of the previous moment cannot generate jumping change, an effective judgment threshold value can be set by using a slope correlation theory between the data, and the effective judgment threshold value can be used as the preset threshold value. If the comparison results are all smaller than a preset threshold value, taking any one of the prediction results as final prediction result information; and if at least one comparison result is not smaller than a preset threshold value, acquiring initial data. For example: and obtaining a prediction result A, a prediction result B and a prediction result C, wherein the difference value between the prediction result A and the prediction result B is 0.001, the difference value between the prediction result A and the prediction result C is 0.002, the difference value between the prediction result B and the prediction result C is 0.002, and the preset threshold value is 0.005. For example: and obtaining a prediction result D, a prediction result E and a prediction result F, wherein the difference value between the prediction result D and the prediction result E is 0.003, the difference value between the prediction result D and the prediction result F is 0.007, the difference value between the prediction result E and the prediction result F is 0.009, and the preset threshold value is 0.004.
In the implementation process, initial data is obtained, the initial data comprises ship motion data, other ship motion attitude data and wave height data, the initial data is preprocessed to generate effective initial data, noise data can be removed by preprocessing the initial data, the influence of data magnitude difference can be reduced, the training convergence speed is accelerated, the calculation efficiency is improved, the ship motion is predicted by adopting a plurality of preset LSTM neural network models respectively according to the effective initial data to generate a plurality of prediction results, the prediction is carried out by adopting different input data from several aspects, then the prediction results are compared to generate a comparison result, and the final prediction result information is obtained according to the comparison result, so that the prediction result is predicted according to one input data, but the forecasting is carried out from multiple aspects, and the forecasting results are compared, so that the smaller the difference between the forecasting results is, the more ready the forecasting results are, and the accuracy of the extremely short-term forecasting of the ship motion is improved.
The first LSTM neural network model, the second LSTM neural network model and the third LSTM neural network model all belong to LSTM neural network models, and the construction of the LSTM neural network model mainly comprises the following processes:
firstly, obtaining sample initial data; the number of the sample initial data may be set according to actual conditions, for example, the number of the sample initial data may be 4000 or 2000.
Then, obtaining model parameters; the model parameters mainly refer to the main parameters of the LSTM model, and include input step number, neuron node number, hidden layer number, optimizer type and learning rate, training iteration number, training batch size, and the like. Some of the parameter settings not only affect the model training efficiency, but also affect the prediction accuracy, so that the training efficiency is effectively improved while the prediction accuracy is ensured by selecting appropriate model parameters. For example: the number of neuron nodes is 16, the number of hidden layer layers is 2, the optimizer selects adam (adaptive motion estimation), the learning rate is set to be 0.0006, the number of training iteration rounds is 30000, and the size of a training batch is 32. The number of training samples is 4000 and the duration is 1600 seconds. Considering the difference between the sea state periods of the third, fourth and fifth levels, the number of the test samples is 100, 150 and 200 respectively, and the test duration is 40, 60 and 80 seconds respectively.
The LSTM model parameter setting is generally judged by experience to find the optimal parameter, and the model parameter can be obtained by performing optimization calculation by adopting a particle swarm optimization algorithm to improve the accuracy of data. The method comprises the following steps:
firstly, setting initial model parameters according to sample initial data; the initial model parameters can be set according to experience and mainly comprise input step number, neuron node number, hidden layer number, optimizer type and learning rate, training iteration round number, training batch size and the like.
And secondly, performing optimization calculation on the initial model parameters by adopting a particle swarm optimization algorithm to generate optimized model parameters, and taking the optimized model parameters as the model parameters. The optimization calculation mainly includes optimizing initial model parameters by a particle swarm optimization algorithm to obtain optimized model parameters, and firstly, allocating initial random positions and initial random speeds to all particles in a space. The position of each particle is then advanced in turn based on its velocity, the known optimal global position in the problem space, and the known optimal position of the particle. As the computation progresses, the particles are clustered or aggregated around one or more optimal points by exploring and utilizing known vantage points in the search space. The particle swarm optimization algorithm belongs to the prior art, and is not described herein again.
And finally, constructing an LSTM neural network framework according to the model parameters and the sample initial data and carrying out iterative training to obtain an LSTM neural network model. And constructing an LSTM neural network framework through the obtained model parameters, and then bringing the initial sample data into the LSTM neural network framework for iterative training, thereby obtaining the LSTM neural network model.
The establishment of the LSTM neural network model further comprises the following steps;
firstly, extracting and carrying out multilayer LSTM neural network calculation on a data vector group in sample initial data to generate target data; the multilayer LSTM neural network means that an LSTM unit structure consists of a forgetting gate, an input gate, an output gate and unit states, and at the current moment, three LSTM network input parameters are as follows: input value (such as wave height) at the current moment, output value (such as ship motion) at the last moment and unit state at the last moment. The output parameters are two: the current time output value and the current time unit state. The LSTM realizes the control of three gates through activating functions, thereby realizing the retention and forgetting of history information. For example: the wave height time series data are (h1, h2, h 3.),ht) and the ship rolling angle data is (theta)1，θ2，θ3......，θt) And inputting the data into a multilayer LSTM neural network for calculation to obtain ship rolling angle data corresponding to the nth data in the future.
Then, bringing the target data into a preset loss function expression for calculation to obtain a loss value; the above-mentioned expression of the loss function may adopt a root-mean-square loss function, and the expression of the root-mean-square loss function is:
wherein: prediction represents the target data, label represents the theoretical data, and N represents the total number of samples. The root mean square loss function represents the overall prediction error case.
And finally, optimizing the LSTM neural network model by adopting a preset optimizer according to the loss value so as to generate the optimized LSTM neural network model. The loss function, i.e. the function of the difference between the value of the objective function and the true value, is actually a function of the parameter to be optimized. The task of the optimizer is to calculate the gradient of the loss function in each time segment and then update the model parameters. The optimizer comprises an Adam optimizer, the model parameters are optimized according to the loss values, so that new model parameters are obtained, and then iterative training is carried out according to the new model parameters, so that an optimized LSTM neural network model is obtained. The Adam optimizer is prior art and will not be described herein.
Based on the same inventive concept, the invention further provides a ship motion very-short term forecasting system based on the LSTM neural network, please refer to fig. 2, and fig. 2 is a structural block diagram of a ship motion very-short term forecasting system based on the LSTM neural network provided by the embodiment of the invention. The ship motion extremely short-term forecasting system based on the LSTM neural network comprises:
a data obtaining module 110, configured to obtain initial data; the initial data comprises ship motion data, other ship motion attitude data and wave height data;
a data preprocessing module 120, configured to preprocess the initial data to generate valid initial data;
the LSTM neural network model prediction module 130 is configured to predict ship motions by using a plurality of preset LSTM neural network models according to the valid initial data, so as to generate a plurality of prediction results;
and the prediction result comparison module 140 is configured to compare the multiple prediction results to generate and obtain final prediction result information according to the comparison result.
In the implementation process, the initial data is obtained by the data obtaining module 110, the initial data includes the ship motion data, other ship motion attitude data and wave height data, then the data preprocessing module 120 preprocesses the initial data to generate effective initial data, noise data can be removed by preprocessing the initial data, the influence of data magnitude difference can be reduced, meanwhile, the training convergence speed is accelerated, and the calculation efficiency is improved, then the LSTM neural network model predicting module 130 predicts the ship motion by adopting a plurality of preset LSTM neural network models according to the effective initial data to generate a plurality of prediction results, the prediction results are predicted by adopting different input data from several aspects, then the prediction result comparing module 140 compares the plurality of prediction results to generate a comparison result, and the final prediction result information is obtained according to the comparison result, therefore, the prediction results are not only predicted according to input data, but also predicted from multiple aspects, and the prediction results are compared, so that the smaller the difference between the prediction results is, the more ready the prediction results are, and the accuracy of the extremely short-term prediction of the ship motion is improved.
Referring to fig. 3, fig. 3 is a schematic structural block diagram of an electronic device according to an embodiment of the present disclosure. The electronic device comprises a memory 101, a processor 102 and a communication interface 103, wherein the memory 101, the processor 102 and the communication interface 103 are electrically connected with each other directly or indirectly to realize the transmission or interaction of data. For example, the components may be electrically connected to each other via one or more communication buses or signal lines. The memory 101 may be used to store software programs and modules, such as program instructions/modules corresponding to the LSTM neural network based ship motion very short-term forecasting system provided in the embodiments of the present application, and the processor 102 executes the software programs and modules stored in the memory 101, so as to execute various functional applications and data processing. The communication interface 103 may be used for communicating signaling or data with other node devices.
The Memory 101 may be, but is not limited to, a Random Access Memory (RAM), a Read Only Memory (ROM), a Programmable Read-Only Memory (PROM), an Erasable Read-Only Memory (EPROM), an electrically Erasable Read-Only Memory (EEPROM), and the like.
The processor 102 may be an integrated circuit chip having signal processing capabilities. The Processor 102 may be a general-purpose Processor, including a Central Processing Unit (CPU), a Network Processor (NP), and the like; but also Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field Programmable Gate Arrays (FPGAs) or other Programmable logic devices, discrete Gate or transistor logic devices, discrete hardware components.
It will be appreciated that the configuration shown in fig. 3 is merely illustrative and that the electronic device may include more or fewer components than shown in fig. 3 or have a different configuration than shown in fig. 3. The components shown in fig. 3 may be implemented in hardware, software, or a combination thereof.
In the embodiments provided in the present application, it should be understood that the disclosed system and method may be implemented in other ways. The above-described system embodiments are merely illustrative, and for example, the flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
In addition, functional modules in the embodiments of the present application may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.
The above-described functions, if implemented in the form of software functional modules and sold or used as a separate product, may be stored in a computer-readable storage medium. Based on such understanding, the technical solution of the present application or portions thereof that substantially contribute to the prior art may be embodied in the form of a software product stored in a storage medium and including instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the above-described method according to the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
To sum up, the LSTM neural network-based ship motion very short term forecasting method and system provided by the embodiments of the present application obtains initial data including ship motion data, other ship motion attitude data, and wave height data, pre-processes the initial data to generate effective initial data, removes noise data by pre-processing the initial data, reduces the influence of data magnitude difference, increases the training convergence speed, improves the calculation efficiency, predicts ship motion by using a plurality of preset LSTM neural network models according to the effective initial data to generate a plurality of prediction results, predicts by using different input data from several aspects, compares the plurality of prediction results to generate a comparison result, and the final forecast result information is obtained according to the comparison result, so that the forecast result is not only forecasted according to input data, but also forecasted from multiple aspects, and the forecast results are compared to obtain, the smaller the difference among the forecast results is, the more ready the forecast result is, and the accuracy of the extremely short-term forecast of the ship motion is improved.
The above description is only a preferred embodiment of the present application and is not intended to limit the present application, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present application shall be included in the protection scope of the present application.
It will be evident to those skilled in the art that the present application is not limited to the details of the foregoing illustrative embodiments, and that the present application may be embodied in other specific forms without departing from the spirit or essential attributes thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the application being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference sign in a claim should not be construed as limiting the claim concerned.
Claims (10)
1. A ship motion extremely short-term forecasting method based on an LSTM neural network is characterized by comprising the following steps:
acquiring initial data; the initial data comprises ship motion data, other ship motion attitude data and wave height data;
preprocessing the initial data to generate effective initial data;
respectively adopting a plurality of preset LSTM neural network models to predict the ship motion according to the effective initial data so as to generate a plurality of prediction results;
and comparing the various prediction results to generate a comparison result, and obtaining final prediction result information according to the comparison result.
2. The LSTM neural network based ship motion ultra-short term forecasting method as claimed in claim 1, wherein the step of predicting ship motion by using a plurality of preset LSTM neural network models respectively according to valid initial data to generate a plurality of prediction results comprises the following steps:
extracting and inputting the ship motion data in the effective initial data into a preset first LSTM neural network model to generate a first prediction result;
extracting and inputting the ship motion data and other ship motion attitude data in the effective initial data into a preset second LSTM neural network model to generate a second prediction result;
and extracting and inputting the ship motion data and the wave height data in the effective initial data into a preset third LSTM neural network model to generate a third prediction result.
3. The LSTM neural network-based ship motion very short term prediction method of claim 1, wherein the step of preprocessing the initial data to generate valid initial data comprises the steps of:
inputting the initial data into a preset smoothing filter to generate preprocessed data;
and substituting the preprocessed data into a preset normalized mathematical expression for calculation to generate effective initial data.
4. The LSTM neural network-based ship motion very-short-term forecasting method according to claim 3, wherein the normalized mathematical expression is as follows:
5. The LSTM neural network-based ship motion very-short-term forecasting method according to claim 1, comprising the following steps:
acquiring sample initial data;
obtaining model parameters;
and constructing an LSTM neural network framework according to the model parameters and the sample initial data and carrying out iterative training to obtain an LSTM neural network model.
6. The LSTM neural network based ship motion very short term forecasting method of claim 5, wherein the step of obtaining model parameters comprises the steps of:
setting initial model parameters according to the initial data of the sample;
and performing optimization calculation on the initial model parameters by adopting a particle swarm optimization algorithm to generate optimized model parameters, and taking the optimized model parameters as the model parameters.
7. The LSTM neural network-based ship motion extreme short-term forecasting method according to claim 6, further comprising the steps of;
extracting and carrying out multilayer LSTM neural network calculation on a data vector group in the sample initial data to generate target data;
substituting the target data into a preset loss function expression for calculation to obtain a loss value;
and optimizing the LSTM neural network model by adopting a preset optimizer according to the loss value so as to generate the optimized LSTM neural network model.
8. The LSTM neural network-based ship motion very short-term forecasting method of claim 1, wherein the step of comparing the plurality of prediction results to generate a comparison result and obtaining final forecasting result information according to the comparison result comprises the steps of:
comparing any two prediction results in the multiple prediction results to obtain a plurality of comparison results;
comparing the comparison results with a preset threshold value respectively, and if the comparison results are smaller than the preset threshold value, taking any one of the prediction results as final prediction result information; and if at least one comparison result is not smaller than a preset threshold value, acquiring initial data.
9. A ship motion extremely short-term forecasting system based on an LSTM neural network is characterized by comprising the following components:
the data acquisition module is used for acquiring initial data; the initial data comprises ship motion data, other ship motion attitude data and wave height data;
the data preprocessing module is used for preprocessing the initial data to generate effective initial data;
the LSTM neural network model prediction module is used for predicting ship motion by adopting various preset LSTM neural network models according to effective initial data so as to generate various prediction results;
and the prediction result comparison module is used for comparing various prediction results to generate and obtain final prediction result information according to the comparison result.
10. An electronic device, comprising:
a memory for storing one or more programs;
a processor;
the one or more programs, when executed by the processor, implement the method of any of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110803068.3A CN113408711A (en) | 2021-07-15 | 2021-07-15 | Ship motion extremely-short-term forecasting method and system based on LSTM neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110803068.3A CN113408711A (en) | 2021-07-15 | 2021-07-15 | Ship motion extremely-short-term forecasting method and system based on LSTM neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113408711A true CN113408711A (en) | 2021-09-17 |
Family
ID=77686647
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110803068.3A Pending CN113408711A (en) | 2021-07-15 | 2021-07-15 | Ship motion extremely-short-term forecasting method and system based on LSTM neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113408711A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116842474A (en) * | 2023-06-14 | 2023-10-03 | 青岛哈尔滨工程大学创新发展中心 | Ship motion extremely short-term forecasting method and system based on TFT model |
| CN116842381A (en) * | 2023-06-13 | 2023-10-03 | 青岛哈尔滨工程大学创新发展中心 | Ship motion extremely-short-term prediction model generalization optimization method based on data fusion |
| CN117892886A (en) * | 2024-03-18 | 2024-04-16 | 青岛哈尔滨工程大学创新发展中心 | Ship motion extremely short-term probability forecasting method and system based on confidence interval |
| CN117909665A (en) * | 2024-03-18 | 2024-04-19 | 青岛哈尔滨工程大学创新发展中心 | Ship motion envelope forecast data processing method and system based on Fourier filtering |
-
2021
- 2021-07-15 CN CN202110803068.3A patent/CN113408711A/en active Pending
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116842381A (en) * | 2023-06-13 | 2023-10-03 | 青岛哈尔滨工程大学创新发展中心 | Ship motion extremely-short-term prediction model generalization optimization method based on data fusion |
| CN116842474A (en) * | 2023-06-14 | 2023-10-03 | 青岛哈尔滨工程大学创新发展中心 | Ship motion extremely short-term forecasting method and system based on TFT model |
| CN117892886A (en) * | 2024-03-18 | 2024-04-16 | 青岛哈尔滨工程大学创新发展中心 | Ship motion extremely short-term probability forecasting method and system based on confidence interval |
| CN117909665A (en) * | 2024-03-18 | 2024-04-19 | 青岛哈尔滨工程大学创新发展中心 | Ship motion envelope forecast data processing method and system based on Fourier filtering |
| CN117892886B (en) * | 2024-03-18 | 2024-05-28 | 青岛哈尔滨工程大学创新发展中心 | Ship motion extremely short-term probability forecasting method and system based on confidence interval |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113408711A (en) | Ship motion extremely-short-term forecasting method and system based on LSTM neural network | |
| CN109978228B (en) | PM2.5 concentration prediction method, device and medium | |
| CN109120462B (en) | Method and device for predicting opportunistic network link and readable storage medium | |
| CN109544598B (en) | Target tracking method and device and readable storage medium | |
| CN111989696A (en) | Neural network for scalable continuous learning in domains with sequential learning tasks | |
| CN110132282B (en) | Unmanned aerial vehicle path planning method and device | |
| US20220147877A1 (en) | System and method for automatic building of learning machines using learning machines | |
| US11375176B2 (en) | Few-shot viewpoint estimation | |
| CN109462564B (en) | Underwater communication modulation mode identification method and system based on deep neural network | |
| CN110826692B (en) | Automatic model compression method, device, equipment and storage medium | |
| CN114091594A (en) | Model training method and device, equipment and storage medium | |
| Alam | Recurrent neural networks in electricity load forecasting | |
| US20220383073A1 (en) | Domain adaptation using domain-adversarial learning in synthetic data systems and applications | |
| Abbas et al. | Geometric Convolutional Neural Networks–A Journey to Surrogate Modelling of Maritime CFD | |
| CN109800873B (en) | Image processing method and device | |
| WO2021262139A1 (en) | Distributed machine learning models | |
| CN114826951B (en) | Service automatic degradation method, device, computer equipment and storage medium | |
| Santosa et al. | Safe distance reminder system on ship against port for the standing process using image processing | |
| US20240127075A1 (en) | Synthetic dataset generator | |
| US20230385687A1 (en) | Estimating optimal training data set size for machine learning model systems and applications | |
| CN115859508B (en) | Flow field analysis method, element model generation method, training method and device | |
| CN116933896B (en) | Super-parameter determination and semantic conversion method, device, equipment and medium | |
| CN116384442A (en) | Ship motion extremely short-term forecasting method based on long-short-term neural network | |
| CN113221448A (en) | Method and system for generating pseudo online sequence of galloping data of overhead transmission line | |
| KR102250354B1 (en) | Method and apparatus for determining delay possibility of shipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |