CN117048802A - Ship future motion attitude prediction method and system based on real sea state strong adaptation - Google Patents
Ship future motion attitude prediction method and system based on real sea state strong adaptation Download PDFInfo
- Publication number
- CN117048802A CN117048802A CN202310461492.3A CN202310461492A CN117048802A CN 117048802 A CN117048802 A CN 117048802A CN 202310461492 A CN202310461492 A CN 202310461492A CN 117048802 A CN117048802 A CN 117048802A
- Authority
- CN
- China
- Prior art keywords
- real
- ship
- model
- prediction model
- prediction
- 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
- 230000033001 locomotion Effects 0.000 title claims abstract description 61
- 238000000034 method Methods 0.000 title claims abstract description 30
- 230000006978 adaptation Effects 0.000 title claims abstract description 19
- 238000012549 training Methods 0.000 claims abstract description 43
- 239000010410 layer Substances 0.000 claims description 32
- 238000003062 neural network model Methods 0.000 claims description 29
- 230000006870 function Effects 0.000 claims description 15
- 210000002569 neuron Anatomy 0.000 claims description 6
- 238000005457 optimization Methods 0.000 claims description 5
- 230000004913 activation Effects 0.000 claims description 4
- 239000002356 single layer Substances 0.000 claims description 3
- 238000013528 artificial neural network Methods 0.000 description 18
- 230000004044 response Effects 0.000 description 7
- 210000004556 brain Anatomy 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000000306 recurrent effect Effects 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 238000004880 explosion Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 230000003446 memory effect Effects 0.000 description 2
- 230000001052 transient effect Effects 0.000 description 2
- 206010063385 Intellectualisation Diseases 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008033 biological extinction Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 210000004027 cell Anatomy 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 210000002364 input neuron Anatomy 0.000 description 1
- 238000002955 isolation Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000015654 memory Effects 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000036632 reaction speed Effects 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Classifications
-
- G—PHYSICS
- G07—CHECKING-DEVICES
- G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
- G07C5/00—Registering or indicating the working of vehicles
- G07C5/08—Registering or indicating performance data other than driving, working, idle, or waiting time, with or without registering driving, working, idle or waiting time
- G07C5/0841—Registering performance data
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B63—SHIPS OR OTHER WATERBORNE VESSELS; RELATED EQUIPMENT
- B63B—SHIPS OR OTHER WATERBORNE VESSELS; EQUIPMENT FOR SHIPPING
- B63B79/00—Monitoring properties or operating parameters of vessels in operation
- B63B79/30—Monitoring properties or operating parameters of vessels in operation for diagnosing, testing or predicting the integrity or performance of vessels
-
- 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
- G06N3/044—Recurrent networks, e.g. Hopfield networks
- G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
-
- 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)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Combustion & Propulsion (AREA)
- Mechanical Engineering (AREA)
- Ocean & Marine Engineering (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a ship future motion attitude prediction method and a system based on real sea state strong adaptation, wherein the method comprises the following steps: performing iterative training on the initial prediction model by using real ship real-time data acquired at different moments to obtain a real-time prediction model; the initial prediction model is obtained by training historical motion data of the ship; real-time data of the real ship obtained at the current moment are input into the real-time prediction model to be predicted, and real-time data of the real ship, which are obtained by continuously updating, are used as training data to continuously train and optimize the model, so that the model can realize self-updating, the strong adaptability of the model to real sea conditions is realized, and the problem that prediction errors caused by incapability of adapting to instant sea wave environments of prediction model parameters in the prior art is solved.
Description
Technical Field
The invention relates to the technical field of ship motion gestures, in particular to a ship future motion gesture prediction method based on real sea condition strong adaptation.
Background
With the continuous development of economy and society, the world-wide countries are more closely connected, and the rapid development of shipping industry is driven. In order to protect the offshore trade and energy channel safety of the country and maintain the overseas interests of the country, a large number of newly-emerging combat ships are in service in the next water in recent years in China, and further research on ship motion modeling and ship hydrodynamic performance is necessary. The carrier-based helicopter has the characteristics of high reaction speed, small required field and strong maneuverability, does not need a special runway for lifting, and can be an aircraft carrier or a plurality of common ships. The carrier-based helicopter can follow the ship to reach any sea area to execute tasks, is more suitable for carrying out functions of offshore rescue, offshore transportation, offshore law enforcement and the like, and can also be used for carrying out various missions such as reconnaissance rescue, anti-submarine, amphibious assault, air early warning, electronic warfare and the like on the sea. Because the marine environment and the gas image condition of the ship and the ship-based helicopter are changed and tested when the ship and the ship-based helicopter execute tasks, the ship-based helicopter frequently rises and falls on the deck of the ship which continuously swings.
Because the ship always moves irregularly during running, the landing point of the helicopter is a three-dimensional moving point in space, and the phenomenon of single-point landing or slipping of the helicopter during landing affects the safety of the ship and the helicopter. Accordingly, the helicopter must find a "rest period" in which the movement of the ship is relatively gentle during the final phase of landing, during which safe landing is completed. The rest period is 3-5 seconds short and 10 seconds long. Therefore, a ship estimation algorithm needs to be researched, a set of ship estimation models are established, and ship motion gestures or motion trends are estimated. Fig. 3 is a view of a carrier-based helicopter landing on a deck of a ship.
The rapid increase in computer capabilities has led to the rapid rise of intelligent technology since the 80 s of the 20 th century, and has made a significant contribution in various fields. In order to meet the requirements of automation and intellectualization of modern large ships in offshore navigation and the requirement of specific ships for executing specific tasks, many students at home and abroad begin to apply machine learning to the field of marine hydrodynamic force from the sixty of the twentieth century, but the effect is not ideal.
With further development of computer technology, students imitate a neural network system of the human brain, and hope to make the computer learn the thinking mode of the human brain to solve the practical problem. Neural Networks (NNs) are simplifications and abstractions of the human brain nervous system, and have strong automatic learning and nonlinear mapping capabilities. From the eighth nineties of the twentieth century, research into recurrent neural networks (Recurrent Neural Network, RNN) began. The RNN is a recurrent neural network, and takes sequence data as input, and each RNN unit is connected in a chained manner to form a closed loop which continuously recurs in the sequence evolution direction. However, as the time interval increases, the RNN cannot establish connections for the neural network elements for longer time intervals. The relevant scholars have pointed out that when RNN models are used to predict long time series problems such as vessel motions, the training efficiency of the model may be reduced and even gradient explosions and gradient vanishes may result.
The scholars then develop Long Short Term Memory (LSTM) neural networks based on this. In order to store longer-time data information, the LSTM unit improves the neural node of the RNN, and three structures of a forgetting gate, an input gate and an output gate are added in a neural network unit, wherein the three structures control the operation, refreshing and storage of data at the historical moment. This improvement makes LSTM neural network units also known as Memory cells. Thus, LSTM neural networks can memorize longer time history data than conventional RNNs. Gers et al point out that when solving the weight matrix gradient, the forgetting gate of the LSTM neural network model effectively avoids gradient extinction and gradient explosion, and has better treatment effect on long sequence problems.
It is well known that the time series of movements of a marine structure has a significant memory effect, i.e. the waves generated by the current movements will significantly affect the subsequent movements of the vessel. Therefore, the LSTM neural network algorithm based on deep learning has been increasingly applied to the problem of time series modeling of marine structure motion. Qiao et al predict dynamic response of mooring lines based on vessel motion. The result shows that the mooring rope response predicted based on the LSTM model has higher precision. Sun et al propose a new ship motion attitude hybrid prediction model, which adopts LSTM neural network and Gaussian process regression technology to predict and experiment ship rolling and pitching motions, and verifies the effectiveness and advancement of the hybrid model. Guo et al do deterministic predictions of heave motions of the floating semi-submersible based on LSTM neural networks, and then quantify the uncertainty of the predicted time series. However, it is easy to see that the accuracy of the prediction result obtained based on deep learning at present depends on the volume of the earlier training set, but no matter how large the earlier training data volume is, the fixed prediction model parameters obtained through training are difficult to adapt to the instantaneous and changeable sea wave environment, which also brings difficulty for how to apply the theoretical research result to the actual engineering.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention aims to provide a ship future motion attitude prediction method and system based on real sea condition strong adaptation, and solves the problem that prediction errors caused by incapability of adapting to transient sea wave environments in the prior art.
In order to achieve the above object, the present invention provides the following solutions:
a ship future motion attitude prediction method based on real sea state strong adaptation comprises the following steps:
performing iterative training on the initial prediction model by using real ship real-time data acquired at different moments to obtain a real-time prediction model; the initial prediction model is obtained by training historical motion data of the ship;
and inputting real-time data of the real ship obtained at the current moment into the real-time prediction model for prediction to obtain real-time ship motion prediction data.
Preferably, the network structure of the initial prediction model is an LSTM neural network model; the LSTM neural network model comprises an input layer, a hidden layer and an output layer which are sequentially connected.
Preferably, the method for constructing the initial prediction model includes:
training each parameter of the LSTM neural network model based on the ship historical motion data to obtain a trained LSTM neural network model;
and adjusting weight function coefficients among the input layer, the hidden layer and the output layer of the trained LSTM neural network model to obtain an initial prediction model.
Preferably, the parameters of the LSTM neural network model include:
the number of single-layer neurons, the number of hidden neuron layers, the dimension of an input vector, the dimension of an output vector, the learning rate, the loss function, the activation function and the number of learning rounds.
A real sea condition strong adaptation based future motion attitude prediction system for a ship, comprising:
the training module is used for performing iterative training on the initial prediction model by using real ship real-time data acquired at different moments to obtain a real-time prediction model
And the prediction module is used for carrying out iterative training on the initial prediction model by using real ship real-time data acquired at different moments to obtain a real-time prediction model.
Preferably, the training module further comprises:
the initial prediction model building unit is used for training each parameter of the LSTM neural network model based on the ship historical motion data to obtain a trained LSTM neural network model.
And the adjusting unit is used for adjusting weight function coefficients among the input layer, the hidden layer and the output layer of the trained LSTM neural network model to obtain an initial prediction model.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
the invention provides a ship future motion attitude prediction method and system based on real sea condition strong adaptation.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions of the prior art, the drawings that are needed in the embodiments will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a schematic diagram of a ship future motion attitude prediction method based on real sea state strong adaptation provided by an embodiment of the invention;
FIG. 2 is a frame diagram of a real-time prediction program based on an LSTM model according to an embodiment of the present invention;
fig. 3 is a structural diagram of an LSTM unit provided in an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The invention aims to provide a ship future motion attitude prediction method and system based on real sea condition strong adaptation, and solves the problem that prediction errors caused by incapability of adapting to transient sea wave environments of prediction model parameters in the prior art.
In order that the above-recited objects, features and advantages of the present invention will become more readily apparent, a more particular description of the invention will be rendered by reference to the appended drawings and appended detailed description.
As shown in fig. 1, the invention provides a ship future motion attitude prediction method based on real sea state strong adaptation, which comprises the following steps:
step 100: performing iterative training on the initial prediction model by using real ship real-time data acquired at different moments to obtain a real-time prediction model; the initial prediction model is obtained by training historical motion data of the ship;
step 200: and inputting real-time data of the real ship obtained at the current moment into the real-time prediction model for prediction to obtain real-time ship motion prediction data.
Further, the network structure of the initial prediction model is an LSTM neural network model; the LSTM neural network model comprises an input layer, a hidden layer and an output layer which are sequentially connected.
The ship motion prediction model constructed based on the LSTM neural network algorithm is a numerical model for predicting a motion response curve in a future period by fitting a time history curve based on the past history motion data of the ship. The method takes historical motion data as model input data and motion data of a period of time in the future as model output data. In order to be able to memorize the data information for a longer time, the LSTM unit is added with a forgetting gate f t Input gate i t And an output gate o t Three configurations are shown in fig. 2.
LSTM forgets the long-lived and small-lived parts of the history data through a forget gate. Input data is selectively input through an input gate. The output data is selectively output through the output gate, so that the memory effect of new data is enhanced, and the specific expression is as follows:
wherein: w (W) i 、W f 、W o The weights of the input gate, the forget gate and the output gate are respectively connected;
b i 、b f 、b o the bias conditions of the input door, the forget door and the output door are respectively;
h t-1 and h t The LSTM unit states of the last moment and the moment are respectively;
x t is the information of the input neuron at the current moment;
sigma (x) and tanh (x) are respectively activation functions,
the invention constructs a relation model of ship motion response and load response based on the LSTM neural network algorithm, and realizes codes by means of Pytorch. Pytorch is an open-source Python machine learning library. As a deep learning framework with Python priority, pythorch can not only achieve powerful GPU acceleration, but also support dynamic neural networks.
Specifically, the method for constructing the initial prediction model comprises the following steps:
training each parameter of the LSTM neural network model based on the ship historical motion data to obtain a trained LSTM neural network model;
and adjusting weight function coefficients among the input layer, the hidden layer and the output layer of the trained LSTM neural network model to obtain an initial prediction model.
Specifically, the parameters of the LSTM neural network model include:
the number of single-layer neurons, the number of hidden neuron layers, the dimension of an input vector, the dimension of an output vector, the learning rate, the loss function, the activation function and the number of learning rounds. The parameter adjustment of the model is the adjustment of the super parameter of the model, the super parameter is the parameter manually adjusted before or during training, and the super parameter adjustment is the performance of fitting independence problem for optimizing the algorithm, and meanwhile, the model is prevented from being too complex and falling into over fitting.
The data set needs to be preprocessed before training the LSTM prediction model. The invention sets 75% of the data set as the training set; 25% of the dataset was set as the test set. Meanwhile, mutual isolation of the training set and the testing set is ensured, so that the data of the training set is prevented from polluting the testing set.
The invention aims to adjust the weight of each hidden layer during transmission in a training link, so that the rule of the prediction model is closed to the rule of the current wave environment of the real ship, and finally, the real-time prediction of the future motion response of the ship body adapting to the real sea condition is realized. The method comprises the following specific steps:
(a) Constructing an LSTM neural network model, collecting motion calendar data of a target ship type under various sea conditions, and training various parameters of the neural network;
(b) After a certain iterative training, a relation model of historical motion and future motion data of the ship body is subjected to preliminary adaptation by adjusting weight function coefficients among layers, and the model is stored;
(c) During actual application, the stored model is loaded, and the model is retrained by means of real ship motion data monitored in current sea conditions, so that the model parameters are ensured to be adjusted for the latest motion data all the time;
(d) At intervals (set manually), the model parameters retrained based on the actually measured motion data are synchronized to a prediction module, so that the prediction model is continuously updated and optimized;
(e) At this time, the prediction module can rapidly predict the motion response at the future moment according to the motion data of the ship at the latest time in the current sea state. The method is repeated in a circulating way, the constructed prediction model is continuously modified to ensure that the parameters of the prediction call are the latest model which is most suitable for the current sea state, and the program framework is shown in figure 3.
The embodiment also provides a ship future motion attitude prediction system based on real sea state strong adaptation, which comprises the following steps:
the training module is used for performing iterative training on the initial prediction model by using real ship real-time data acquired at different moments to realize automatic optimization and obtain a real-time prediction model with stronger adaptability to real sea conditions
And the prediction module is used for carrying out iterative training on the initial prediction model by using real ship real-time data acquired at different moments to obtain a real-time prediction model.
Specifically, the training module further includes:
the initial prediction model building unit is used for training each parameter of the LSTM neural network model based on the ship historical motion data to obtain a trained LSTM neural network model.
And the adjusting unit is used for adjusting weight function coefficients among the input layer, the hidden layer and the output layer of the trained LSTM neural network model to obtain an initial prediction model.
The beneficial effects of the invention are as follows:
the key core is that the model can be updated by continuously training real-time data of a real ship under real sea conditions, so that the adaptability of a prediction model to the real sea conditions is enhanced, the prediction precision is greatly improved compared with a fixed parameter model, and the thought has great advancement and innovation; the motion prediction process is operated in parallel with the model training process, namely only the model training optimization part is carried out when the prediction is not carried out; if the prediction is performed, the prediction module is called, and the training optimization is performed simultaneously, so that excessive consumption of computing resources is avoided. In addition, the invention plans to use a direct output multi-step prediction method aiming at the limit of the current single-step prediction method in practical engineering application. The data input by the method is X= { X 1 ,x 2 ,…,x M-1 ,x M Output data is y= { Y } 1 ,y 2 ,…,y N-1 ,y N I.e. given historical data with a step size of M, future data with a step size of N is predicted. This eliminates the need to limit the prediction value at the previous time to predict the next step, and avoids the problem of error transfer.
In the present specification, each embodiment is described in a progressive manner, and each embodiment is mainly described in a different point from other embodiments, and identical and similar parts between the embodiments are all enough to refer to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant points refer to the description of the method section.
The principles and embodiments of the present invention have been described herein with reference to specific examples, the description of which is intended only to assist in understanding the methods of the present invention and the core ideas thereof; also, it is within the scope of the present invention to be modified by those of ordinary skill in the art in light of the present teachings. In view of the foregoing, this description should not be construed as limiting the invention.
Claims (6)
1. The ship future motion attitude prediction method based on real sea condition strong adaptation is characterized by comprising the following steps of:
performing iterative training on the initial prediction model by using real ship real-time data acquired at different moments to obtain a real-time prediction model; the initial prediction model is obtained by training historical motion data of the ship;
and inputting real-time data of the real ship obtained at the current moment into the real-time prediction model for prediction to obtain real-time ship motion prediction data.
2. The method for predicting the future motion attitude of the ship based on the strong adaptation of real sea conditions according to claim 1, wherein the network structure of the initial prediction model is an LSTM neural network model; the LSTM neural network model comprises an input layer, a hidden layer and an output layer which are sequentially connected.
3. The method for predicting the future motion attitude of the ship based on the strong adaptation of the real sea condition according to claim 2, wherein the method for constructing the initial prediction model comprises the following steps:
training each parameter of the LSTM neural network model based on the ship historical motion data to obtain a trained LSTM neural network model;
and adjusting weight function coefficients among the input layer, the hidden layer and the output layer of the trained LSTM neural network model to obtain an initial prediction model.
4. A method for predicting future motion attitude of a vessel based on true sea state adaptation as claimed in claim 3, wherein the parameters of the LSTM neural network model include:
the number of single-layer neurons, the number of hidden neuron layers, the dimension of an input vector, the dimension of an output vector, the learning rate, the loss function, the activation function and the number of learning rounds.
5. A ship future motion attitude prediction system based on real sea state strong adaptation, characterized by comprising:
the training module is used for performing iterative training on the initial prediction model by using real ship real-time data acquired at different moments to realize automatic optimization and obtain a real-time prediction model with stronger adaptability to real sea conditions
And the prediction module is used for updating the real-time prediction model after iterative training optimization into the real-time prediction model of the prediction module.
6. The real sea state based strong adaptation of marine vessel future motion gesture prediction system of claim 5, wherein the training module further comprises:
the initial prediction model building unit is used for training various parameters of the LSTM neural network model based on the ship historical motion data to obtain a trained LSTM neural network model;
and the adjusting unit is used for adjusting weight function coefficients among the input layer, the hidden layer and the output layer of the trained LSTM neural network model to obtain an initial prediction model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310461492.3A CN117048802A (en) | 2023-04-26 | 2023-04-26 | Ship future motion attitude prediction method and system based on real sea state strong adaptation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310461492.3A CN117048802A (en) | 2023-04-26 | 2023-04-26 | Ship future motion attitude prediction method and system based on real sea state strong adaptation |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117048802A true CN117048802A (en) | 2023-11-14 |
Family
ID=88656043
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310461492.3A Pending CN117048802A (en) | 2023-04-26 | 2023-04-26 | Ship future motion attitude prediction method and system based on real sea state strong adaptation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117048802A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112800075A (en) * | 2021-02-01 | 2021-05-14 | 上海海事大学 | Ship operation forecast database updating method based on real ship six-degree-of-freedom attitude data |
| CN114021441A (en) * | 2021-10-28 | 2022-02-08 | 江苏科技大学 | Ship motion attitude prediction method based on CNN-BilSTM |
| CN114330828A (en) * | 2021-11-29 | 2022-04-12 | 北京遥测技术研究所 | Method for forecasting movement rest period of ship |
| CN115578546A (en) * | 2022-09-23 | 2023-01-06 | 浙江大华技术股份有限公司 | Ship attitude prediction method, equipment, device and system |
| CN115906619A (en) * | 2022-11-04 | 2023-04-04 | 武汉理工大学深圳研究院 | PSO-BP optimization algorithm-based energy management method for hybrid double electric ships |
-
2023
- 2023-04-26 CN CN202310461492.3A patent/CN117048802A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112800075A (en) * | 2021-02-01 | 2021-05-14 | 上海海事大学 | Ship operation forecast database updating method based on real ship six-degree-of-freedom attitude data |
| CN114021441A (en) * | 2021-10-28 | 2022-02-08 | 江苏科技大学 | Ship motion attitude prediction method based on CNN-BilSTM |
| CN114330828A (en) * | 2021-11-29 | 2022-04-12 | 北京遥测技术研究所 | Method for forecasting movement rest period of ship |
| CN115578546A (en) * | 2022-09-23 | 2023-01-06 | 浙江大华技术股份有限公司 | Ship attitude prediction method, equipment, device and system |
| CN115906619A (en) * | 2022-11-04 | 2023-04-04 | 武汉理工大学深圳研究院 | PSO-BP optimization algorithm-based energy management method for hybrid double electric ships |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111351488B (en) | Intelligent trajectory reconstruction reentry guidance method for aircraft | |
| Zhang et al. | Ship motion attitude prediction based on an adaptive dynamic particle swarm optimization algorithm and bidirectional LSTM neural network | |
| CN108762281A (en) | It is a kind of that intelligent robot decision-making technique under the embedded Real-time Water of intensified learning is associated with based on memory | |
| Wang et al. | Vessel trajectory prediction based on AIS data and bidirectional GRU | |
| Cheng et al. | Research status of artificial neural network and its application assumption in aviation | |
| CN112034711A (en) | Unmanned ship sea wave interference resistance control method based on deep reinforcement learning | |
| CN107121926A (en) | A kind of industrial robot Reliability Modeling based on deep learning | |
| CN114330644B (en) | Neural network model compression method based on structure search and channel pruning | |
| CN113033118B (en) | Autonomous floating control method of underwater vehicle based on demonstration data reinforcement learning technology | |
| CN114818477A (en) | Aircraft aerodynamic characteristic prediction method, device and medium | |
| Piao et al. | Adaptive backstepping sliding mode dynamic positioning system for pod driven unmanned surface vessel based on cerebellar model articulation controller | |
| CN117634661A (en) | Ship maneuvering motion forecasting method based on self-attention two-way long-short-term memory network | |
| CN117048802A (en) | Ship future motion attitude prediction method and system based on real sea state strong adaptation | |
| Wang et al. | MUTS-based cooperative target stalking for a multi-USV system | |
| CN111797474A (en) | Method, device and medium for establishing and predicting ship dynamic stability performance prediction model | |
| Xu et al. | Real-time planning and collision avoidance control method based on deep reinforcement learning | |
| CN116341612A (en) | AUV drift track prediction method based on ABiLSTM-QSOA network | |
| CN114715331B (en) | Floating ocean platform power positioning control method and system | |
| CN115718497A (en) | Multi-unmanned-boat collision avoidance decision method | |
| CN115933712A (en) | Bionic fish leader-follower formation control method based on deep reinforcement learning | |
| Xu et al. | Surf-riding and broaching prediction of ship sailing in regular waves by LSTM based on the data of ship motion and encounter wave | |
| Sun et al. | Deep Learning‐Based Trajectory Tracking Control forUnmanned Surface Vehicle | |
| Li et al. | Prioritized experience replay based reinforcement learning for adaptive tracking control of autonomous underwater vehicle | |
| Wang et al. | Networked and Deep Reinforcement Learning-Based Control for Autonomous Marine Vehicles: A Survey | |
| CN117111594B (en) | Self-adaptive track control method for unmanned surface vessel |
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 |