CN103177290B - Identification method for model of ship domain based on online self-organization neural network - Google Patents
Identification method for model of ship domain based on online self-organization neural network Download PDFInfo
- Publication number
- CN103177290B CN103177290B CN201310116653.1A CN201310116653A CN103177290B CN 103177290 B CN103177290 B CN 103177290B CN 201310116653 A CN201310116653 A CN 201310116653A CN 103177290 B CN103177290 B CN 103177290B
- Authority
- CN
- China
- Prior art keywords
- ship
- model
- neural network
- input
- layer
- 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.)
- Expired - Fee Related
Links
Abstract
The invention provides an identification method for a model of ship domain based on an online self-organization fuzzy neural network. The method comprises the following steps of: selecting a ship safety zone model and determining a function, an input variable and a desired output value of the model of the ship safety zone model; establishing a dynamic fuzzy neural network comprising an input layer, a membership function layer, a T-norm layer and an output layer; training the dynamic fuzzy neural network by a training dataset comprising an input variable and an output value of the model till achieving accuracy requirements; and taking sailing parameters of two corresponding ships as input variables and inputting the input variables into the trained ship safety zone model so as to obtain ship safety areas of the two ships. By adopting the technical scheme, the safety model modified by using the method has better accuracy and higher safety compared with the conventional model of the ship domain.
Description
Technical field
The present invention relates to a kind of modification method of ship domain model, more particularly to it is a kind of based on online self-organizing feature map
The discrimination method of the ship domain model of network.
Background technology
Marine intelligent transportation traffic has been increasingly becoming vessel traffic as the important component part of China's science and technology development strategy
With the emerging crossing research focus of effective and science fusion.And the research of the individual ship behavior for maritime traffic system,
Then it is particularly important.20th century six the seventies, Japan plus rattan [1] propose the concept in navigation safety field so far,
Understand in document [2] [3] [4] [5], researcher proposes the navigation safety domain model of various different shapes and sizes.
Modern ships have a wide range of applications in field.But, a unified model, the master for causing the problems referred to above cannot be formed all the time
Reason is wanted to have：（1）The factor of different navigation environments causes to produce the model of different shapes and sizes；（2）Most model is all
Formed according to statistical data or the method for simulation experiment；（3）Existing model is all it can be readily appreciated that but be difficult to be applied to reality
In.Document [3] [4] proposes a kind of hexagon ship domain model of complexity, each with ship's speed and ship convolution parameter determination
Side size, the model causes the ship in the case of collision prevention to be easy to be optimized its flight path using evolution algorithm, but its complicated journey
Degree is higher, and the physics meaning is more ambiguous, is not easy to understand and practical application.[5] provide several with reference to factors such as ship turning performances
In the case of ship domain border quantization method, the maneuvering performance of ship is embodied in the method, but moulded dimension and shadow
Functional relationship between the factor of sound is artificial given a kind of rough estimation equation.It should be noted that [2] proposition is " transversal
Area " model is put together by former and later two semiellipses, is determined by Ship Controling parameter and headway etc., is classical model
One of.
On the other hand, fuzzy system, neutral net and fuzzy neural network are fast-developing, and because it has extraordinary forcing
Closely, generalization ability, is applied to rapidly in industrial circle.When design fuzzy system, neutral net and fuzzy neural network, it is necessary to
Regular number work Hidden nodes are first determined, while the mode learning algorithm of application error back propagation is trained.It is well known that should
Method pace of learning is slow, is easily trapped into local minimum point.Therefore, in the urgent need to finding a Fast Learning for real-time application
Method.To solve the above problems, researcher proposes dynamic neural network.But D-FNN is suffered from the drawback that：
Dynamic fuzzy neural network（DFNN）It is to use standard gaussian function to divide in the input space, and input in its rule becomes
The width of all Gaussian functions of amount is all identical, and this point is generally inconsistent with real, particularly when input variable has
Very different operation interval when.
Dynamic fuzzy neural network（DFNN）No matter how its membership function to be distributed, its membership function and fuzzy rule
Quantity it is all identical.This causes some membership function overlappings, the fuzzy rule indigestion for extracting.
Dynamic fuzzy neural network（DFNN）In the Gaussian function width of first fuzzy rule randomly select.
Dynamic fuzzy neural network（DFNN）It is middle to there is too many parameter set in advance, and these parameters all lack physics meaning
Justice, so as to relatively difficult when these special parameters are selected.For dynamic fuzzy neural network（DFNN）Input quantity must be carried out
Normalization and the renormalization of output, will so add intensive, and the time for reaching preferable Approximation effect is long.
Therefore, it is of the invention then a novel intelligent ship is proposed based on " cross-sectional area " model and Self-organized Fuzzy Neural Network
The model in oceangoing ship field.
The content of the invention
Proposition for problem above of the invention, and a kind of ship domain based on online self organizing neural network developed
The discrimination method of model, with following steps：
- safety of ship regional model is selected, determine function, input variable and the desired output of the model；
- set up the dynamic fuzzy neural network comprising input layer, membership function layer, T- norms layer and output layer；
- using the input variable comprising the model and the training dataset of output valve, to the dynamic fuzzy nerve net
Network is trained until reaching required precision；
- by the navigational parameter of two correspondence ships, as the safety of ship region mould that input variable is input into after training is finished
Type, obtains the safety of ship region of two ships.
The safety of ship regional model is cross-sectional area model：The model is approximately by former and later two semiellipse splits
Into the function of the model is shown below：
Wherein, Rbf、RbaAnd SbThe forward and backward oval radius and transversal radius in region, T are represented respectively90For ship turning 90
Time, D needed for degreeTIt is ambient parameter for tactical diameter, s and t；
Input variable is：Pk=[Lk, Bk, U1 k, U2 k, αk], desired output is：K=1,2......n,
U2 kRepresent object ship ship's speed；αkRepresent two ship angles.
Described dynamic fuzzy neural network is specifically included：
Input layer：With multiple nodes, the linguistic variable of one input of each node on behalf；
Membership function layer：With multiple nodes, one membership function of each node on behalf, the membership function uses Gauss
Function is expressed as follows：
Wherein, i=1,2......r, j=1,2......u, wherein μijFor xiJ-th membership function, cijIt is xiJth
The center of individual Gaussian function, σj L,σj RX is represented respectivelyiJ-th membership function left and right width, r is input variable number, and u is to be
The total regular number of system, x represents the linguistic variable of an input, x=[x1,x2,......,xi], xiRepresent numbers of the x in i-th dimension
Value；
T- norm layers：With multiple nodes, the IF- parts of one possible fuzzy rule of each node on behalf, j-th rule
Then it is output as：
Wherein：
Output layer：At least there is a node, each node in this layer represents respectively an output variable, and the output is
The superposition of all input signals：
Y is output variable, ωjIt is THEN- parts, for TSK models：ωj=α0j+α1jx1+α2jx2,......,+αrjxr, j=1,2 ..., u.
Three independent broad sense self organizing neural networks are built for three variables of cross-sectional area model：
The dynamic fuzzy neural network training is comprised the steps of：
- calculate degree of membership of the data to each center, expectation quality k with asymmetric Gaussian functioneWith can accommodate effectively half
Footpath kd, ke=max { emaxβk-1,emin, kd=max { dmaxγk-1,dmin}；
- computing system error：||ek| |=| | tk-yk||；Calculate mahalanobis distance：
WhereinAnd find most pony
Family name is apart from dmin, find out from the nearest node of the sample；
When（1）||ei| | ＞ ke,dmin＞ kdWhen, need to increase a fuzzy rule；
（2）||ei||≤ke,dmin≤kdWhen, broad sense self organizing neural network can completely accommodate the data；
（3）||ei||≤ke,dmin＞ kdWhen, there is broad sense self organizing neural network preferable generalization ability need to only adjust knot
Fruit parameter；
（4）||ei| | ＞ ke,dmin≤kdWhen, adjustment covers the RBF nodes of the data and updates result parameter；
- define error rate of descent：
Err is converted to：ERR=[ρ1,ρ2,…ρu],ERR∈R(r +1)u；
Define and calculate the importance of every rule：
Judge：If sigj＞ ks, then j-th strip rule, wherein k are deletedsIt is the parameter for judging sensitivity to predefine parameter；
- repetition training, until network output error reaches requirement, terminates training, enters when having barrier or other ships
When entering the safety of ship field to obtained by the present invention, the data that deck officer can obtain according to the present invention are judging this ship
New ship trajectory, it is to avoid collide.
Training is obtained after dynamic fuzzy neural network, and using inspection data collection the performance of dynamic neural network is checked：Respectively
To Rbf、RbaAnd SbMake root-mean-square error.
As a result of above-mentioned technical proposal, the present invention is proposing a kind of ship based on online self organizing neural network
The discrimination method of domain model, it is proposed that OSFNN, is functionally equivalent to the online Self-Fuzzy nerve of TSK fuzzy systems
Network.It is, in principle, that the OSFNN compared with the D-FNN based on symmetrical Gaussian function, based on asymmetric Gaussian function
Receive domain to provide more flexible, wider nonlinear transformation to approach in any one nonlinear system, therefore algorithm more
Complexity, with more generality；From using for upper, the fuzzy rule that the algorithm is extracted has good intelligibility.It is worth carrying
Go out, although OSFNN is more complicated than D-FNN, parameter set in advance required for it is but few than D-FNN, therefore it
Implement than D-FNN and be more prone to.Relatively with traditional ship domain model, through the security model of present invention amendment, tool
There is more preferable precision, safety is also higher.
Description of the drawings
For clearer explanation embodiments of the invention or the technical scheme of prior art, below will be to embodiment or existing
The accompanying drawing to be used needed for having technology description does one and simply introduces, it should be apparent that, drawings in the following description are only
Some embodiments of the present invention, for those of ordinary skill in the art, on the premise of not paying creative work, may be used also
To obtain other accompanying drawings according to these accompanying drawings.
Fig. 1 is the topology diagram of broad sense self organizing neural network of the present invention
Fig. 2 is the algorithm flow chart of broad sense self organizing neural network of the present invention
Fig. 3 is the cross-sectional area vessel area model that the present invention is adopted
Fig. 4 is R of the present inventionbfNeuronal quantity and the graph of a relation of training set sample size
Fig. 5 is R of the present inventionbfRoot-mean-square error and training set sample size graph of a relation
Fig. 6 is R of the present inventionbaNeuronal quantity and the graph of a relation of training set sample size
Fig. 7 is R of the present inventionbaRoot-mean-square error and training set sample size graph of a relation
Fig. 8 is S of the present inventionbNeuronal quantity and the graph of a relation of training set sample size
Fig. 9 is S of the present inventionbRoot-mean-square error and training set sample size graph of a relation
Specific embodiment
To make purpose, technical scheme and the advantage of embodiments of the invention clearer, with reference to the embodiment of the present invention
In accompanying drawing, clearly complete description is carried out to the technical scheme in the embodiment of the present invention：
As shown in Fig. 1-Fig. 9：A kind of discrimination method of the ship domain model based on online self organizing neural network, mainly
Comprise the steps：
First, safety of ship regional model is selected, determines function, input variable and the desired output of the model：
The vessel area model for adopting is " cross-sectional area " model, wherein Rbf、RbaAnd SbRepresent respectively region before and after half
Footpath and transversal radius, model is determined by equation below：
Wherein, L, B, U represent respectively the length and width and speed T of ship90Time, D for needed for 90 degree of ship turningTTo return
Turn diameter, s and t and represent ambient parameter, environment has：Two ships are relative to travel（Stem to stem）, two ships travel in opposite directions（Stern is to ship
Head）, two ships intersect traveling（Ship side is to fore）.（The former represents this ship, and the latter represents object ship）：
When two fores are correct,
During two ship decussations,
When two ships are end to end,
Then, the dynamic fuzzy neural network comprising input layer, membership function layer, T- norms layer and output layer is set up：
If OSFNN dynamic fuzzy neural networks are divided into four layers（Such as Fig. 1）, respectively：Input layer, membership function layer, T- models
Several layers, output layer,
Input layer：Each node represents respectively the linguistic variable of an input.
Membership function layer：Each node DGF represents respectively a membership function, and the membership function is with following Gaussian function
Represent：
Wherein, i=1,2......r, j=1,2......u, wherein μijFor xiJ-th membership function, cijIt is xiJth
The center of individual Gaussian function, σj L,σj RX is represented respectivelyiJ-th membership function left and right width, r is input variable number, and u is
The total regular number of system.
T- norm layers：Each node in this layer represents respectively the IF- parts in a possible fuzzy rule.
Therefore j-th rule is output as：
Wherein
Output layer：Each node in this layer represents respectively an output variable, if only one of which output node, for list
Output, if there is multiple output nodes, for the output of many results, the output is the superposition of all input signals：
Wherein, y is output variable, ωjIt is THEN- parts, for TSK models：ωj=α0j+α1jx1+α2jx2,......,
+αrjxr, j=1,2 ..., u.
The present invention need to build three independent OSFNN, respectively：
Using the input variable comprising the model and the training dataset of output valve, to the dynamic fuzzy neural network
It is trained until reaching required precision：
（1）Calculate degree of membership of the data to each Gauss center with asymmetric Gaussian function, initialization system it is predefined
Parameter.Expectation quality ke with can accommodate effective radius kd.
Wherein ke=max { emaxβk-1,emin, kd=max { dmaxγk-1,dmin}。
（2）Computing system error：||ek| |=| | tk-yk||；Calculate mahalanobis distance：
Wherein
And minimum mahalanobis distance is found, find out from the nearest node of the sample.
（3）Judged according to mahalanobis distance：
1. | | ei| | ＞ ke,dmin＞ kdWhen, need to increase a fuzzy rule.
2. | | ei||≤ke,dmin≤kdWhen, OSFNN can completely accommodate the data.
3. | | ei||≤ke,dmin＞ kdWhen, there is OSFNN preferable generalization ability need to only adjust result parameter.
4. | | ei| | ＞ ke,dmin≤kdWhen, adjustment covers the RBF nodes of the data and updates result parameter.
（4）Decompose Ψ=PQ, wherein T with equation of linear regression T=Ψ A+E and QR：Desired output vector；A：Quan Xiang
Amount；Ψ：Regression matrix；E error vectors；P：Orthogonal matrix, P ∈ Rn*v；Q：Upper triangular matrix, Q ∈ Rv*v。
Define error rate of descent：
Err is converted to：ERR=[ρ1,ρ2,…ρu],ERR∈R(r+1)u。
Define and calculate the importance of every rule：
Judge：If sigj＞ ks, then j-th strip rule is deleted.Repetition training, until network output error reaches requirement, terminates
Training.
Inspection data collection is retrieved by model by above-mentioned, for checking the performance of OSFNN dynamic fuzzy neural networks, and
Respectively to Rbf、RbaAnd SbMake root-mean-square error, respectively such as Fig. 4, shown in 5,6.
During ship's navigation, to object ship（Possibly into this ship security fields）The R for obtainingbf、RbaAnd SbNumerical value, build new
Ship domain.
The above, the only present invention preferably specific embodiment, but protection scope of the present invention is not limited thereto,
Any those familiar with the art the invention discloses technical scope in, technology according to the present invention scheme and its
Inventive concept equivalent or change in addition, all should be included within the scope of the present invention.
Claims (3)
1. a kind of discrimination method of the ship domain model based on online Self-organized Fuzzy Neural Network, with following steps：
- safety of ship regional model is selected, determine function, input variable and the desired output of the model；
- set up the dynamic fuzzy neural network comprising input layer, membership function layer, T- norms layer and output layer；
- using the input variable comprising the model and the training dataset of output valve, the dynamic fuzzy neural network is entered
Row training is until reach required precision；
- safety of ship the regional model for the navigational parameter of this ship being input into after training is finished as input variable, obtains this ship
Navigation safety region；
The safety of ship regional model is cross-sectional area model：The model is approximately put together by former and later two semiellipses, should
The function of model is shown below：
Wherein, Rbf、RbaAnd SbThe forward and backward oval radius and transversal radius in region, T are represented respectively90For 90 degree of institutes of ship turning
Time, the D for needingTIt is ambient parameter for tactical diameter, s and t；
Input variable is：Pk=[Lk, Bk, U1 k, U2 k, αk], desired output is：K=1,2......n, U2 k
Represent object ship ship's speed；αkRepresent two ship angles；For this ship's speed；
Three independent online self organizing neural networks are built for three variables of cross-sectional area model：
Wherein, L, B, U represent respectively the length and width and speed of ship, and a is ship angle.
2. the identification side of a kind of ship domain model based on online Self-organized Fuzzy Neural Network according to claim 1
Method, is further characterized in that：Described dynamic fuzzy neural network is specifically included：
Input layer：With multiple nodes, the linguistic variable of one input of each node on behalf；
Membership function layer：With multiple nodes, one membership function of each node on behalf, the membership function uses Gaussian function
It is expressed as follows：
Wherein, i=1,2......r, j=1,2......u, wherein μijFor xiJ-th membership function, cijIt is xiJ-th
The center of Gaussian function, σij L,σij RX is represented respectivelyiJ-th membership function left and right width, r is input variable number, and u is to be
The total regular number of system, x represents the linguistic variable of an input, x=[x1,x2,......,xi], xiRepresent numbers of the x in i-th dimension
Value；
T- norm layers：With multiple nodes, the IF- parts of one possible fuzzy rule of each node on behalf, j-th rule
It is output as：
Wherein
Output layer：At least there is a node, each node in this layer represents respectively an output variable, and the output is all
The superposition of input signal：
Y is output variable, ωjIt is THEN- parts, for TSK models：ωj=α0j+α1jx1+α2jx2,......,+αijxr, j=
1,2 ..., u, αijFor the matrix element in TSK models；
The dynamic fuzzy neural network training is comprised the steps of：
- calculate degree of membership of the data to each Gauss center, expectation quality k with asymmetric Gaussian functioneWith can accommodate effectively half
Footpath kd, ke=max { emaxβk-1,emin, kd=max { dmaxγk-1,dmin}；
- computing system error：||ek| |=| | tk-yk||；Calculate mahalanobis distance：
WhereinAnd find minimum mahalanobis distance
dmin, find out from the nearest GEBF nodes of sample；
When (1) | | ei||>ke,dmin>kdWhen, need to increase a fuzzy rule；
(2)||ei||≤ke,dmin≤kdWhen, broad sense self organizing neural network can completely accommodate the data；
(3)||ei||≤ke,dmin>kdWhen, there is broad sense self organizing neural network preferable generalization ability need to only adjust result ginseng
Number；
(4)||ei||>ke,dmin≤kdWhen, adjustment covers the RBF nodes of the data and updates result parameter；
- define error rate of descent：
Decompose Ψ=PQ, wherein T with equation of linear regression T=Ψ A+E and QR：Desired output vector；A：Weight vector；Ψ：Return
Return matrix；E error vectors；P：Orthogonal matrix, P ∈ Rn*v；Q：Upper triangular matrix, Q ∈ Rv*v；
err∈R1*(r+1)v；Err is converted to：ERR=[ρ1,ρ2,…ρu],ERR∈R(r+1)*u；
Define and calculate the susceptiveness of every rule：
Judge：If sigj>ks, then j-th strip rule, wherein k are deletedsIt is the parameter for judging sensitivity to predefine parameter；
- repetition training, until network output error reaches requirement, terminates training, enters into when having barrier or other ships
During resulting safety of ship field, deck officer can judge the new ship trajectory of this ship according to the data for obtaining, and keep away
Exempt to collide.
3. the identification side of a kind of ship domain model based on online Self-organized Fuzzy Neural Network according to claim 2
Method, is further characterized in that：Training is obtained after dynamic fuzzy neural network, and using inspection data collection the property of dynamic neural network is checked
Energy：Respectively to Rbf、RbaAnd SbMake root-mean-square error, represent the performance of this dynamic fuzzy neural network.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310116653.1A CN103177290B (en) | 2013-04-03 | 2013-04-03 | Identification method for model of ship domain based on online self-organization neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310116653.1A CN103177290B (en) | 2013-04-03 | 2013-04-03 | Identification method for model of ship domain based on online self-organization neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103177290A CN103177290A (en) | 2013-06-26 |
| CN103177290B true CN103177290B (en) | 2017-05-03 |
Family
ID=48637130
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310116653.1A Expired - Fee Related CN103177290B (en) | 2013-04-03 | 2013-04-03 | Identification method for model of ship domain based on online self-organization neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103177290B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104050329B (en) * | 2014-06-25 | 2017-07-11 | 哈尔滨工程大学 | A kind of method for detecting Collision Risk Index |
| CN106527568A (en) * | 2016-12-15 | 2017-03-22 | 中国人民解放军后勤工程学院 | Maximum power tracker control method based on generalized dynamic fuzzy neural network |
| CN107545785A (en) * | 2017-07-21 | 2018-01-05 | 华南理工大学 | A kind of river channel running method based on big data |
| CN109189071B (en) * | 2018-09-25 | 2021-03-26 | 大连海事大学 | Robust self-adaptive unmanned ship path tracking control method based on fuzzy observer |
| CN109213174B (en) * | 2018-10-24 | 2021-10-01 | 北京工业大学 | Sewage treatment plant intelligent patrol obstacle avoidance method based on fuzzy neural network |
| CN110796899B (en) * | 2019-10-30 | 2020-11-10 | 青岛科技大学 | Ship-shore relative field acquisition method based on ship cluster situation in limited water area |
| CN112149237A (en) * | 2020-10-15 | 2020-12-29 | 北京海兰信数据科技股份有限公司 | Real-time ship collision avoidance method and system |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6301572B1 (en) * | 1998-12-02 | 2001-10-09 | Lockheed Martin Corporation | Neural network based analysis system for vibration analysis and condition monitoring |
| CN101825871A (en) * | 2010-04-09 | 2010-09-08 | 哈尔滨工程大学 | Intelligent adaptive control method for heave and pitch device for oblique rudder ship |
| CN102331717A (en) * | 2011-10-10 | 2012-01-25 | 哈尔滨工程大学 | Intelligent control method of navigational speed of ship |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6493689B2 (en) * | 2000-12-29 | 2002-12-10 | General Dynamics Advanced Technology Systems, Inc. | Neural net controller for noise and vibration reduction |
-
2013
- 2013-04-03 CN CN201310116653.1A patent/CN103177290B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6301572B1 (en) * | 1998-12-02 | 2001-10-09 | Lockheed Martin Corporation | Neural network based analysis system for vibration analysis and condition monitoring |
| CN101825871A (en) * | 2010-04-09 | 2010-09-08 | 哈尔滨工程大学 | Intelligent adaptive control method for heave and pitch device for oblique rudder ship |
| CN102331717A (en) * | 2011-10-10 | 2012-01-25 | 哈尔滨工程大学 | Intelligent control method of navigational speed of ship |
Non-Patent Citations (1)
| Title |
|---|
| 基于模糊神经网络的一种船舶碰撞危险度计算方法;陈建华 等;《船舶科学技术》;20080430;第30卷(第2期);第136-138页，附图1-2 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103177290A (en) | 2013-06-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103177290B (en) | Identification method for model of ship domain based on online self-organization neural network | |
| CN103198720B (en) | A kind of modification method of the ship domain model based on broad sense self organizing neural network | |
| Liu et al. | Fuzzy comprehensive evaluation for the motion performance of autonomous underwater vehicles | |
| Zhang et al. | Data mining approach for automatic ship-route design for coastal seas using AIS trajectory clustering analysis | |
| CN103207937B (en) | A kind of quaternary safety of ship domain model system and collision avoidance method | |
| CN111063218A (en) | Ship collision avoidance decision method | |
| CN113156947B (en) | Method for planning path of ship in dynamic environment | |
| Wang et al. | Local path optimization method for unmanned ship based on particle swarm acceleration calculation and dynamic optimal control | |
| Wang et al. | Transferable and adaptable driving behavior prediction | |
| Hu et al. | Intelligent anomaly detection of trajectories for IoT empowered maritime transportation systems | |
| Alarcin et al. | Rudder roll stabilization for fishing vessel using neural network approach | |
| Yuan et al. | Prioritized experience replay-based deep q learning: Multiple-reward architecture for highway driving decision making | |
| CN103186815B (en) | Method for identifying ship domain model based on on-line quick self-organization fuzzy neural network | |
| Ren et al. | A novel high-dimensional trajectories construction network based on multi-clustering algorithm | |
| Liu et al. | Black-box modeling of ship maneuvering motion based on Gaussian process regression with wavelet threshold denoising | |
| CN106846916A (en) | A kind of navigation method of control | |
| Nguyen et al. | Automatic collision avoiding support system for ships in congested waters and at open sea | |
| Zhuo et al. | A ship based intelligent anti-collision decision-making support system utilizing trial manoeuvres | |
| Higaki et al. | Human-like route planning for automatic collision avoidance using generative adversarial imitation learning | |
| Hiraoka et al. | Two-stage prediction method of typhoon position by fuzzy modeling. Fusion of outline prediction and detailed prediction | |
| Zhang et al. | Time-delay wavelet network predictor based on sensitivity analysis with application to predictive ship course control | |
| Jiang et al. | A joint multi-model machine learning prediction approach based on confidence for ship stability | |
| Zhang et al. | Path planning of ship collision avoidance based on stochastic schemata exploiter | |
| Zhang et al. | Using the CGAN Model Extend Encounter Targets Image Training Set. | |
| Sarabi-Jamab et al. | PiLiMoT: A modified combination of LoLiMoT and PLN learning algorithms for local linear neurofuzzy modeling |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170503 Termination date: 20210403 |