CN107918393B - Marine Autopilot based on depth confidence network - Google Patents
Marine Autopilot based on depth confidence network Download PDFInfo
- Publication number
- CN107918393B CN107918393B CN201711222632.2A CN201711222632A CN107918393B CN 107918393 B CN107918393 B CN 107918393B CN 201711222632 A CN201711222632 A CN 201711222632A CN 107918393 B CN107918393 B CN 107918393B
- Authority
- CN
- China
- Prior art keywords
- module
- ship
- depth confidence
- network model
- confidence network
- 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.)
- Active
Links
- 238000003062 neural network model Methods 0.000 claims abstract description 26
- 238000004891 communication Methods 0.000 claims abstract description 25
- 230000007613 environmental effect Effects 0.000 claims abstract description 16
- 238000012549 training Methods 0.000 claims description 12
- 238000013528 artificial neural network Methods 0.000 claims description 5
- 238000000034 method Methods 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 2
- 238000011161 development Methods 0.000 abstract description 6
- 239000010410 layer Substances 0.000 description 52
- 230000000694 effects Effects 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 230000004308 accommodation Effects 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 2
- 210000004556 brain Anatomy 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 230000003466 anti-cipated effect Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 239000003344 environmental pollutant Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 239000011229 interlayer Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 231100000719 pollutant Toxicity 0.000 description 1
- 238000004064 recycling Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/0206—Control of position or course in two dimensions specially adapted to water vehicles
Landscapes
- Engineering & Computer Science (AREA)
- Aviation & Aerospace Engineering (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to the Marine Autopilots for being based on depth confidence network (DBN), comprising: microprocessor module, memory module, communication module and human-machine interface module；The input/output terminal of microprocessor module is connected with the output input of memory module；Microprocessor module is connected with the output input of communication module；Microprocessor module is connected with the output end of human-machine interface module, the Examination for the Crew data information of microprocessor module reception human-machine interface module, and depth confidence network model is established according to Examination for the Crew data information, depth confidence network model is converted into general neural network model again, communication module receives ship navigation environment information, and microprocessor module is sent by ship's navigation environmental information, microprocessor module calculate to ship's navigation environmental information by general neural network model obtains rudder angle data.The Marine Autopilot can carry out automatic Pilot to ship under various circumstances, mitigate crewman's workload, promote shipping business development.
Description
Technical field
The present invention relates to ship automatic control technical fields, in particular to the Marine Autopilot based on depth confidence network.
Background technique
Ship changes and keeps the speed of a ship or plane or course by the thrust of propeller and the rudder power of rudder, realizes from the port of departure to mesh
Port sail plan.With the development of science and technology, the following ship must be changed towards structure is complicated, the direction hair of operation automation
Exhibition.And Ship Steering System is very important control system, is the equipment for controlling ship course, it, which can overcome, makes ship
The various influences for deviateing prearranged heading keep ship automatically stable and run up in scheduled boat, therefore the automation of rudder is real
The key of existing shipboard automation, performance directly affect the maneuverability, economy and safety of ship's navigation, therefore autopilot
Technology is always to be taken as the science and technology with higher economic value and social benefit, since nineteen twenty-two autopilot comes out,
Engineers and technicians from one generation to the next constantly explore and study to the performance for how improving the system.
Practice have shown that excellent control algolithm is applied in Marine Autopilot, the maneuverability of ship can be greatly improved
The respond of energy and ship, while the energy consumed by ship's navigation can be effectively saved, the discharge of pollutant is reduced, is reduced
Crewman's workload promotes the development of shipping business.
Autopilot precision is low in the prior art, adaptability is weaker.
Summary of the invention
The present invention provides a kind of Marine Autopilot based on depth confidence network, solves or part solves existing skill
The technical problem that autopilot precision is low in art, adaptability is weaker.
Marine Autopilot provided by the invention based on depth confidence network, comprising: microprocessor module, memory mould
Block, communication module and human-machine interface module；
The input/output terminal of the microprocessor module is connected with the output input of the memory module；Micro- place
Reason device module is connected with the output input of the communication module；The microprocessor module is defeated with the human-machine interface module
Outlet is connected；
The microprocessor module receives the Examination for the Crew data information of the human-machine interface module, and according to the crewman
Examination data information establishes depth confidence network model, then the depth confidence network model is converted into general neural network mould
Type；
The memory module stores the general neural network model；
The communication module receives ship navigation environment information, and sends the ship's navigation environmental information to described micro-
Processor module, the microprocessor module carry out the ship's navigation environmental information by the general neural network model
It calculates and obtains rudder angle data.
Preferably, the depth confidence network model includes: one layer of input layer, four layers of hidden layer and one layer of output
Layer；
The number of nodes of input layer is 70, and the quantity of the first hiding node layer is input layer number 1/3~2/3, the
The quantity of two hiding node layers is the 1/3~2/3 of the first hiding layer number, the quantity of third and fourth layer of hiding node layer is equal to the
Two, the quantity of a hiding node layer, output node quantity are 7.
Preferably, the Examination for the Crew data information is officer/pilot's test and training system information；
The Examination for the Crew data information includes: external environment, ship's speed and the rudder angle of ship's navigation, the ship's navigation
External environment includes at least: wave, stream and weather.
Preferably, establishing depth confidence network model according to the Examination for the Crew data information, then the depth is set
Communication network model conversion is at general neural network model, comprising:
The limited Boltzmann machine for forming the depth confidence network model is successively instructed using contrast divergence algorithm
Practice, obtains the primary data of the depth confidence network model；
The depth confidence network model is converted into the general neural network model；
Using back-propagation algorithm, and in conjunction in the Examination for the Crew data information grasp the preferable historical data of ship achievement,
The general neural network model is finely adjusted.
Preferably, the ship's navigation environmental information includes at least: stream, wave and weather.
Preferably, the rudder angle data are sent to ship after the microprocessor module obtains the rudder angle data
Main controller, the ship main controller operating ship steering engine work.
Preferably, the microprocessor module is embedded microprocessor or mobile desktop microprocessor.
Preferably, the communication module includes but is not limited to: serial RS232 interface, 422/485 interface, USB interface with
And network interface.
One or more technical solutions provided herein, have at least the following technical effects or advantages:
Due to using the ship being made of microprocessor module, memory module, communication module and human-machine interface module
Autopilot, microprocessor module receive the Examination for the Crew data information of human-machine interface module, and according to Examination for the Crew data information
Depth confidence network model is established, then depth confidence network model is converted into general neural network model；Memory module is deposited
Store up general neural network model；Communication module receives ship navigation environment information, and sends ship's navigation environmental information to micro-
Processor module, microprocessor module calculate to ship's navigation environmental information by general neural network model obtains rudder angle
Data；By depth confidence network application in Marine Autopilot, the working principle of human brain, realization pair can be more really simulated
Accommodation more accurately controls, and can be improved the learning ability of autopilot and the adaptive ability for environment.In this way,
The technical problem that autopilot precision is low, adaptability is weaker in the prior art is efficiently solved, it is right under various circumstances to realize
Ship carries out automatic Pilot, mitigates crewman's workload to a certain extent, promotes the technical effect of shipping business development.
Detailed description of the invention
Fig. 1 is the information communication schematic diagram of the Marine Autopilot provided by the invention based on depth confidence network.
Specific embodiment
Marine Autopilot provided by the embodiments of the present application based on depth confidence network is solved or is partially solved existing
The technical problem that autopilot precision is low in technology, adaptability is weaker, by setting by microprocessor module, memory module,
The Marine Autopilot of communication module and human-machine interface module composition, microprocessor module receive the Examination for the Crew of human-machine interface module
Data information, and depth confidence network model is established according to Examination for the Crew data information, then depth confidence network model is converted
At general neural network model, microprocessor module calculates ship's navigation environmental information by general neural network model
Obtain rudder angle data；It realizes and automatic Pilot is carried out to ship under various circumstances, mitigate crewman's workload to a certain extent,
Promote the technical effect of shipping business development.
Marine Autopilot provided by the invention based on depth confidence network, comprising: microprocessor module, memory mould
Block, communication module and human-machine interface module；The output input phase of the input/output terminal of microprocessor module and memory module
Even；Microprocessor module is connected with the output input of communication module；The output end of microprocessor module and human-machine interface module
It is connected.
Microprocessor module receives the Examination for the Crew data information of human-machine interface module, and according to Examination for the Crew data information
Depth confidence network model is established, then depth confidence network model is converted into general neural network model.
Memory module stores general neural network model.
Communication module receives ship navigation environment information, and sends microprocessor module for ship's navigation environmental information,
Microprocessor module calculate to ship's navigation environmental information by general neural network model obtains rudder angle data.Wherein,
Ship's navigation environmental information includes at least: stream, wave and weather.
Further, depth confidence network model includes: one layer of input layer, four layers of hidden layer and one layer of output layer；It is defeated
The number of nodes for entering layer is 70, and the quantity of the first hiding node layer is input layer number 1/3~2/3, the second hidden layer section
The quantity of point is the 1/3~2/3 of the first hiding layer number, and the quantity of third and fourth layer of hiding node layer is equal to second, one hidden layer
The quantity of node, output node quantity are 7.
Further, Examination for the Crew data information is officer/pilot's test and training system information；Examination for the Crew
Data information includes: external environment, ship's speed and the rudder angle of ship's navigation, the external environment of ship's navigation include at least: wave, stream and
Weather.
Further, depth confidence network model is established according to Examination for the Crew data information, then by depth confidence network mould
Type is converted into general neural network model, comprising:
It is successively trained, is obtained using limited Boltzmann machine of the contrast divergence algorithm to composition depth confidence network model
To the primary data of depth confidence network model；
Depth confidence network model is converted into general neural network model；
It using back-propagation algorithm, and combines and grasps the preferable historical data of ship achievement in Examination for the Crew data information, to general
Logical neural network model is finely adjusted.
Further, after microprocessor module obtains rudder angle data, rudder angle data are sent to ship main controller, ship master
The work of control machine operating ship steering engine.Microprocessor module is embedded microprocessor or mobile desktop microprocessor.Communicate mould
Block includes but is not limited to: serial RS232 interface, 422/485 interface, USB interface and network interface.
Below by specific embodiment to the working principle of the Marine Autopilot based on depth confidence network of the application into
Row is discussed in detail:
Referring to attached drawing 1, whole system is by microprocessor module, memory module, communication module and human-machine interface module
Composition.
The input/output terminal of its microprocessor module is connected by the output input of bus and memory module, and communicates mould
The output input of block is connected, and is connected with the output end of human-machine interface module；The input/output terminal of memory module by bus with
The output input of microprocessor module is connected；The input/output module of communication module passes through the output of bus and microprocessor module
Input is connected；The output end of human-machine interface module is connected by bus with the input terminal of microprocessor module.
This system is realized the training to DBN model and is based on when it is implemented, using microprocessor as main control chip
The calculating of DBN model.In the training stage, the microprocessor module receives the information from man-machine interface, and in memory mould
6 layers of DBN model are established in block.6 layers of DBN model include one layer of input layer, four layers of hidden layer and one layer of output layer.Each layer of mind
Quantity through network is determined by operator by human-computer interaction module.The number of nodes of each hidden layer has following constraint: defeated
The number of nodes for entering layer is 70, and the quantity of the first hiding node layer is input layer number 1/3~2/3, the second hidden layer section
The quantity of point is the 1/3~2/3 of the first hiding layer number, and the quantity of third and fourth layer of hiding node layer is equal to second, one hidden layer
The quantity of node, output node quantity are 7；Examination for the Crew data information, including ship fortune are stored in the memory module
Row external environment (distinguished and admirable wave etc.) and ship's speed and rudder angle.There are also the DBN models of no initializtion in simultaneous memory module, once
Information from the user is received, microprocessor is just trained it, Examination for the Crew number of the training sample in memory
According to.The specific training algorithm of DBN model are as follows: first using contrast divergence algorithm (CD-K) to the limited bohr of composition DBN model
Hereby graceful machine (RBM) is successively trained, and obtains the primary data of DBN model；Then DBN model is converted into general neural network
(NN) model is allowed it to by normal use；Back-propagation algorithm (BP) finally is utilized, and behaviour's ship achievement is combined preferably to go through
History data, are finely adjusted DBN model, and specific algorithm is as follows:
1, J cycle of training, parameter K (K=1) and sample size n in learning rate η, contrast divergence algorithm are givens。
2, weight vector W and amount of bias a, b are initialized, Δ W=Δ a=Δ b=0 is initialized.
3, Δ W, Δ a, Δ b are updated by recycling to sample using contrast divergence algorithm algorithm.
A, the conditional probability that k-th of neuron of hidden layer is 1, i.e. P (h are calculated using following formula (1)k=1 | v)
Calculated result in (1) formula and the random number r between [0,1] are compared, if r < P (hk=1 | v), then hk=
1,0 is otherwise taken, updated hidden layer h is finally obtained0。
B, the conditional probability that k-th of neuron of visual layers is 1, i.e. P (v are calculated using following formula (2)k=1 | h)
Calculated result in (2) formula and the random number r between [0,1] are compared, if
R < P (vk=1 | h), then vk=1, otherwise take 0.Finally obtain updated visual layers v1。
C, Δ W, Δ a, Δ b are updated using formula (3)~(5).
Δbj=Δ bj+[p(hj=1 | v0)-p(hj|v1)] (5)
D, using algorithm a, tri- step of b, c completes the layer-by-layer training to RBM model.
4, network parameter W, a, b are updated using formula (6)~(8).
5, into algorithm steps 2, requirement until meeting cycle of training, then algorithm stops.
The parameter trained by the above process is only local optimum, needs to carry out reversely to finely tune that can be only achieved DBN whole
Body optimal effectiveness.When reversed fine tuning, DBN trains the BP network of top using BP network with having supervision, arrives to RBM e-learning
Feature classify, by the layer-by-layer back transfer of error between the reality output and anticipated output of DBN to all RBM networks,
The parameter for finely tuning RBM network interlayer, obtains optimal DBN, so far, completes model training.
The specific implementation of communication module determines by running environment, including but not limited to RS232,422/485, USB with
And network interface.The work of communication module are as follows: be connected with ship main controller, receive and come from ship's navigation environment, comprising: stream, wave
And weather etc., and as input data, it is input in the NN model after the completion of training, and rudder angle number is therefrom calculated
According to, and ship main controller is sent it to, to manipulate ship steering engine work.
One or more technical solutions provided herein, have at least the following technical effects or advantages:
Due to using the ship being made of microprocessor module, memory module, communication module and human-machine interface module
Autopilot, microprocessor module receive the Examination for the Crew data information of human-machine interface module, and according to Examination for the Crew data information
Depth confidence network model is established, then depth confidence network model is converted into general neural network model；Memory module is deposited
Store up general neural network model；Communication module receives ship navigation environment information, and sends ship's navigation environmental information to micro-
Processor module, microprocessor module calculate to ship's navigation environmental information by general neural network model obtains rudder angle
Data；By depth confidence network application in Marine Autopilot, the working principle of human brain, realization pair can be more really simulated
Accommodation more accurately controls, and can be improved the learning ability of autopilot and the adaptive ability for environment.In this way,
The technical problem that autopilot precision is low, adaptability is weaker in the prior art is efficiently solved, it is right under various circumstances to realize
Ship carries out automatic Pilot, mitigates crewman's workload to a certain extent, promotes the technical effect of shipping business development.
Above-described specific embodiment has carried out further the purpose of the present invention, technical scheme and beneficial effects
It is described in detail, it should be understood that being not limited to this hair the foregoing is merely a specific embodiment of the invention
Bright, all within the spirits and principles of the present invention, any modification, equivalent substitution, improvement and etc. done should be included in the present invention
Protection scope within.
Claims (8)
1. a kind of Marine Autopilot based on depth confidence network characterized by comprising microprocessor module, memory mould
Block, communication module and human-machine interface module；
The input/output terminal of the microprocessor module is connected with the output input of the memory module；The microprocessor
Module is connected with the output input of the communication module；The output end of the microprocessor module and the human-machine interface module
It is connected；
The microprocessor module receives the Examination for the Crew data information of the human-machine interface module, and according to the Examination for the Crew
Data information establishes depth confidence network model, then the depth confidence network model is converted into general neural network model；
The memory module stores the general neural network model；
The communication module receives ship navigation environment information, and sends the micro process for the ship's navigation environmental information
Device module, the microprocessor module calculate the ship's navigation environmental information by the general neural network model
Obtain rudder angle data.
2. the Marine Autopilot as described in claim 1 based on depth confidence network, which is characterized in that
The depth confidence network model includes: one layer of input layer, four layers of hidden layer and one layer of output layer；
The number of nodes of input layer is 70, and the quantity of the first hiding node layer is input layer number 1/3~2/3, and second is hidden
The quantity for hiding node layer is the 1/3~2/3 of the first hiding layer number, and the quantity of third and fourth layer of hiding node layer is equal to second, one
The quantity of node layer is hidden, output node quantity is 7.
3. the Marine Autopilot as described in claim 1 based on depth confidence network, which is characterized in that
The Examination for the Crew data information is officer/pilot's test and training system information；
The Examination for the Crew data information includes: external environment, ship's speed and the rudder angle of ship's navigation, the outside of the ship's navigation
Environment includes at least: wave, stream and weather.
4. the Marine Autopilot as described in claim 1 based on depth confidence network, which is characterized in that examined according to the crewman
Examination data information establishes depth confidence network model, then the depth confidence network model is converted into general neural network mould
Type, comprising:
The limited Boltzmann machine for forming the depth confidence network model is successively trained using contrast divergence algorithm, is obtained
To the primary data of the depth confidence network model；
The depth confidence network model is converted into the general neural network model；
Using back-propagation algorithm, and in conjunction with the preferable historical data of ship achievement is grasped in the Examination for the Crew data information, to institute
General neural network model is stated to be finely adjusted.
5. the Marine Autopilot as described in claim 1 based on depth confidence network, which is characterized in that
The ship's navigation environmental information includes at least: stream, wave and weather.
6. the Marine Autopilot as described in claim 1 based on depth confidence network, which is characterized in that
After the microprocessor module obtains the rudder angle data, the rudder angle data are sent to ship main controller, the ship
The work of oceangoing ship main controller operating ship steering engine.
7. the Marine Autopilot as described in claim 1 based on depth confidence network, which is characterized in that
The microprocessor module is embedded microprocessor or mobile desktop microprocessor.
8. the Marine Autopilot as described in claim 1 based on depth confidence network, which is characterized in that
The communication module includes but is not limited to: serial RS232 interface, 422/485 interface, USB interface and network interface.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711222632.2A CN107918393B (en) | 2017-11-29 | 2017-11-29 | Marine Autopilot based on depth confidence network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711222632.2A CN107918393B (en) | 2017-11-29 | 2017-11-29 | Marine Autopilot based on depth confidence network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107918393A CN107918393A (en) | 2018-04-17 |
| CN107918393B true CN107918393B (en) | 2019-10-18 |
Family
ID=61897132
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711222632.2A Active CN107918393B (en) | 2017-11-29 | 2017-11-29 | Marine Autopilot based on depth confidence network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107918393B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109557913A (en) * | 2018-10-19 | 2019-04-02 | 江汉大学 | Marine Autopilot based on interacting depth neural network |
| CN110687906A (en) * | 2019-09-20 | 2020-01-14 | 江汉大学 | Intelligent automatic rudder of ship based on pruning technology |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104730920A (en) * | 2015-04-16 | 2015-06-24 | 大连海事大学 | Adaptive dynamic surface controller structure of neural network and method for designing adaptive dynamic surface controller structure |
| CN105511464A (en) * | 2015-12-01 | 2016-04-20 | 南京长峰航天电子科技有限公司 | Ship autonomous navigation control system and control method |
| CN106355151A (en) * | 2016-08-30 | 2017-01-25 | 电子科技大学 | Recognition method, based on deep belief network, of three-dimensional SAR images |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101615000A (en) * | 2009-06-19 | 2009-12-30 | 大连海事大学 | Robust adaptive autopilot control algorithm for navigation simulator |
| CN107145072A (en) * | 2017-05-31 | 2017-09-08 | 威海海洋职业学院 | Ship rudder roll damping adaptive control system |
-
2017
- 2017-11-29 CN CN201711222632.2A patent/CN107918393B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104730920A (en) * | 2015-04-16 | 2015-06-24 | 大连海事大学 | Adaptive dynamic surface controller structure of neural network and method for designing adaptive dynamic surface controller structure |
| CN105511464A (en) * | 2015-12-01 | 2016-04-20 | 南京长峰航天电子科技有限公司 | Ship autonomous navigation control system and control method |
| CN106355151A (en) * | 2016-08-30 | 2017-01-25 | 电子科技大学 | Recognition method, based on deep belief network, of three-dimensional SAR images |
Non-Patent Citations (3)
| Title |
|---|
| 基于受限玻尔兹曼机的深度学习模型及应用;张艳霞;《中国优秀硕士学位论文全文数据库》;20170215;I140-127 * |
| 基于神经网络的船舶自动舵设计;刘百顺等;《海洋工程》;20020228;36-40 * |
| 神经网络与自动舵设计研究;杨鸿;《中国优秀硕士学位论文全文数据库》;20020115;I140-117 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107918393A (en) | 2018-04-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108445892A (en) | A kind of drive lacking unmanned boat formation control device structure and design method | |
| Zhang et al. | Ship motion attitude prediction based on an adaptive dynamic particle swarm optimization algorithm and bidirectional LSTM neural network | |
| CN109214107A (en) | A kind of ship's navigation behavior on-line prediction method | |
| CN110083983A (en) | A kind of boat segmental speed of a ship or plane optimization method and intelligent management system | |
| Xue et al. | Proximal policy optimization with reciprocal velocity obstacle based collision avoidance path planning for multi-unmanned surface vehicles | |
| CN107918393B (en) | Marine Autopilot based on depth confidence network | |
| CN111400864B (en) | Ship collision avoidance decision optimization method based on prospect theory | |
| CN108710373A (en) | A kind of networking drive lacking unmanned boat distribution formation control device and its design method | |
| Lan et al. | Path planning for underwater gliders in time-varying ocean current using deep reinforcement learning | |
| CN110095120A (en) | Biology of the Autonomous Underwater aircraft under ocean circulation inspires Self-organizing Maps paths planning method | |
| CN106444359A (en) | Human-simulated intelligent control method for autonomous region keeping of water-jet propulsion unmanned ship | |
| CN109799818A (en) | The unmanned boat collaboration handling controller and its design method of multi-parameter route guidance | |
| CN108459602A (en) | The autonomous method of docking of drive lacking unmanned boat under more obstacle complex environments | |
| CN110778398B (en) | Marine diesel engine fuel management control system | |
| CN109144099A (en) | Unmanned aerial vehicle group action scheme fast evaluation method based on convolutional neural networks | |
| Yang et al. | Learning-aided intelligent cooperative collision avoidance mechanism in dynamic vessel networks | |
| CN102289675A (en) | Method for intelligently predicting ship course | |
| Amendola et al. | Navigation in restricted channels under environmental conditions: Fast-time simulation by asynchronous deep reinforcement learning | |
| Zhang et al. | Hybrid threshold event-triggered control for sail-assisted USV via the nonlinear modified LVS guidance | |
| Wang et al. | MUTS-based cooperative target stalking for a multi-USV system | |
| Zhou et al. | A system for the validation of collision avoidance algorithm performance of autonomous ships | |
| CN109557913A (en) | Marine Autopilot based on interacting depth neural network | |
| Amendola et al. | Batch reinforcement learning of feasible trajectories in a ship maneuvering simulator | |
| CN115903820A (en) | Multi-unmanned-boat pursuit and escape game control method | |
| Lin et al. | Path planning of unmanned surface vehicle based on improved q-learning algorithm |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |