CN111738500B - Navigation time prediction method and device based on deep learning - Google Patents
Navigation time prediction method and device based on deep learning Download PDFInfo
- Publication number
- CN111738500B CN111738500B CN202010530842.3A CN202010530842A CN111738500B CN 111738500 B CN111738500 B CN 111738500B CN 202010530842 A CN202010530842 A CN 202010530842A CN 111738500 B CN111738500 B CN 111738500B
- Authority
- CN
- China
- Prior art keywords
- navigation
- time
- ship
- voyage
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 34
- 238000013135 deep learning Methods 0.000 title claims abstract description 21
- 238000012549 training Methods 0.000 claims abstract description 16
- 238000012545 processing Methods 0.000 claims abstract description 11
- 238000005516 engineering process Methods 0.000 claims abstract description 10
- 238000013528 artificial neural network Methods 0.000 claims description 18
- 238000013527 convolutional neural network Methods 0.000 claims description 17
- 125000004122 cyclic group Chemical group 0.000 claims description 17
- 230000003068 static effect Effects 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 6
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 5
- 230000006399 behavior Effects 0.000 claims description 4
- 230000006870 function Effects 0.000 claims description 4
- 230000004913 activation Effects 0.000 claims description 3
- 230000008520 organization Effects 0.000 claims description 3
- 238000011144 upstream manufacturing Methods 0.000 claims description 3
- 230000002123 temporal effect Effects 0.000 claims description 2
- 230000008901 benefit Effects 0.000 abstract description 5
- 238000010586 diagram Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 230000008569 process Effects 0.000 description 3
- 230000008878 coupling Effects 0.000 description 2
- 238000010168 coupling process Methods 0.000 description 2
- 238000005859 coupling reaction Methods 0.000 description 2
- 238000007726 management method Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000010219 correlation analysis Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
- G06Q10/083—Shipping
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Tourism & Hospitality (AREA)
- Life Sciences & Earth Sciences (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- General Engineering & Computer Science (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Business, Economics & Management (AREA)
- Development Economics (AREA)
- Biomedical Technology (AREA)
- Software Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Health & Medical Sciences (AREA)
- Game Theory and Decision Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention provides a navigation time prediction method based on deep learning, which comprises the following steps: acquiring AIS data; processing the acquired AIS data to obtain navigation time data of different navigation sections and different time sections; constructing a navigation time prediction model, and inputting the navigation time data into the constructed navigation time prediction model for training; combining the trained navigation time prediction model with a route planning technology to obtain an accurate navigation time prediction value. According to the technical scheme, the navigation time can be accurately predicted, so that the ship knows the estimated time of passing through a certain navigation section in advance, and the optimal path is selected in advance, so that the navigation efficiency and economic benefit are improved, and a foundation is laid for intelligent navigation of the ship.
Description
Technical Field
The invention relates to the technical field of navigation time prediction, in particular to a navigation time prediction method and device based on deep learning.
Background
The inland navigation is taken as an important component of a modern comprehensive transportation system in China, has the outstanding advantages of large transportation capacity, light pollution, low cost, low energy consumption and the like, plays a great role in promoting the rapid coordinated development of regional economy, and is counted: five thousands of rivers exist in China, the total length is more than 42 ten thousand kilometers, the total length of a inland waterway is about 13.51 ten thousand kilometers, wherein the high-grade waterway accounts for 46.56%, the total length is about 6.29 ten thousand kilometers, the number of inland ports is up to 1300, the productive berths are about 2.6 ten thousand, and the number of transport ships is up to 20 ten thousand.
With the continuous increase of inland canal traffic, the congestion situation is increasingly severe. Traffic management departments are also exploring traffic management methods all the time to optimize travel experience. The navigation time is taken as one of the inland traffic information, the effect is quite important, and the accurate navigation time prediction can help the ship to know the estimated time of passing through a certain navigation section in advance, and the optimal route is selected in advance, so that the navigation efficiency and economic benefit are improved.
However, the current navigation time estimation method for inland river includes:
1) Estimating according to experience;
2) And dividing the ship route distance and the initial speed to obtain the estimated time of ship navigation.
The method for estimating the navigation time of the inland river does not fully consider the periodicity and regularity of the time sequence of the navigation time and the spatial correlation between channels, and can not accurately estimate the navigation time of the ship.
Disclosure of Invention
According to the technical problem, a navigation time prediction method based on deep learning is provided. The invention mainly utilizes a navigation time prediction model and fuses channel static information (channel length, water depth, historical average navigation time and the like) to improve the prediction precision of the model.
The invention adopts the following technical means:
a method of voyage time prediction based on deep learning, comprising:
acquiring AIS data;
processing the acquired AIS data to obtain navigation time data of different navigation sections and different time sections;
constructing a navigation time prediction model, and inputting the navigation time data into the constructed navigation time prediction model for training;
combining the trained navigation time prediction model with a route planning technology to obtain an accurate navigation time prediction value.
Further, the processing the received AIS data to obtain voyage time data of different voyages and different time periods includes:
cutting and segmenting the channels according to channel characteristics, wherein ships with similar types in each segment of channel have similar navigation behaviors and navigation time;
calculating the average sailing speed of the ship in each sailing section by using the ship sailing speed information in the AIS data, and further calculating the average sailing time required by the ship to pass through the whole sailing section according to the average sailing speed;
according to traffic flow characteristics, setting the time interval as n hours, and carrying out time-division navigation time statistics on each navigation segment.
Further, the AIS data comprise ship static data, ship dynamic data, ship voyage data and voyage safety information;
the ship static data comprise ship names, call signs, marine mobile service identification codes (MMSI), international Maritime Organization (IMO) numbers, ship lengths, ship widths and ship types;
the ship dynamic data comprise ship position data, ground speed/course and ship head direction information;
the ship voyage data comprise ship state, draft, destination and ETA information;
the voyage safety information comprises voyage warning and weather report information.
Further, the average sailing time required by the ship to pass through the whole voyage section is calculated according to the average speed, and the calculation formula is as follows:
wherein T is average sailing time, V i The navigation speed of the ith ship is n, the number of ships in the navigation section is n, and l is the mileage of the navigation section.
Further, the constructed voyage time prediction model comprises a convolutional neural network for capturing spatial correlation between adjacent voyages and a cyclic neural network for capturing time correlation.
Further, the inputting the navigation time data into the constructed navigation time prediction model for training comprises:
setting L1, L2 and L3 to respectively represent the navigation time sequence of an upstream navigation section, the navigation time sequence of a target navigation section and the navigation time sequence of a downstream navigation section;
adopting Concat operation to the L1, L2 and L3, splicing the navigation time sequences of the three navigation segments into a characteristic, and marking the characteristic as L; L=Concat [ L1, L2, L3]
Extracting the spliced characteristic L by adopting a one-dimensional convolutional neural network, marking the extracted characteristic as F, and carrying out one-dimensional convolutional operation according to the following steps: f=f (Σ) i∈M H i *W i +b), wherein H is the voyage time sequence, W is the weight of the convolution shift operator, b is the paranoid, f (·) is the activation function;
inputting the extracted feature F into the cyclic neural network to obtain a feature extracted by the navigation time prediction model; the cyclic neural network is provided with 20 inputs, which are respectively predicted navigation time characteristics of the first 20 time periods, and each navigation time characteristic is obtained by comprehensively extracting three navigation segments according to space through the one-dimensional convolutional neural network;
fusing the space characteristics of the channel with the characteristics extracted by the navigation time prediction model, wherein the space characteristics of the channel participating in prediction comprise three elements of a navigation section length, a navigation width and a water depth;
and then fusing the 1 channel history similarity feature with the spatial features of the three channels and the features extracted by the navigation time prediction model, inputting the fused features into a full-connection layer, and finally outputting the navigation time of the designated navigation section in the next time section.
Further, the convolutional neural network adopts a one-dimensional convolutional neural network; the cyclic neural network adopts a GRU structure cyclic neural network.
Further, each of the navigation time sequences is composed of the navigation time of the navigation section in the current time section and the first 20 time sections.
The invention also provides a navigation time prediction device based on deep learning, which comprises the following steps:
the acquisition unit is used for acquiring AIS data;
the computing unit is used for processing the AIS data acquired by the acquisition unit to obtain navigation time data of different navigation sections and different time sections;
the training unit is used for constructing a navigation time prediction model, and inputting the navigation time data into the constructed navigation time prediction model for training;
and the generating unit is used for combining the trained navigation time prediction model with the route planning technology to obtain an accurate navigation time prediction value.
A computer-readable storage medium having a set of computer instructions stored therein; the set of computer instructions, when executed by the processor, implement the deep learning based voyage time prediction method described above.
Compared with the prior art, the invention has the following advantages:
1. according to the invention, a navigation time prediction model integrating various characteristic information is constructed by using a more advanced deep learning technology, and compared with other ship navigation time prediction methods, the prediction method can obtain better prediction precision.
2. The navigation time prediction method based on deep learning provided by the invention can accurately predict the navigation time, so that the ship knows the estimated time passing through a certain navigation section in advance, and the optimal path is selected in advance, thereby improving the navigation efficiency and economic benefit of navigation and laying a foundation for intelligent navigation of the ship.
3. The navigation time prediction model can be deployed as a public service interface and is combined with a route planning technology, so that navigation time prediction service is provided for the public and crews through a mobile phone APP, and the channel information service level is improved.
Based on the reasons, the method can be widely popularized in the fields of navigation time prediction and the like.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings may be obtained according to the drawings without inventive effort to a person skilled in the art.
FIG. 1 is a flow chart of the method of the present invention.
Fig. 2 is a schematic diagram of an AIS data structure according to an embodiment of the present invention.
Fig. 3 is a schematic diagram of a leg cut according to an embodiment of the present invention.
Fig. 4 is a schematic diagram of a navigation time prediction model according to an embodiment of the present invention.
Fig. 5 is a schematic structural diagram of a recurrent neural network according to an embodiment of the present invention.
FIG. 6 is a graph showing the relationship between voyage time provided by an embodiment of the present invention.
Detailed Description
In order that those skilled in the art will better understand the present invention, a technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in which it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, shall fall within the scope of the present invention.
It should be noted that the terms "first," "second," and the like in the description and the claims of the present invention and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments of the invention described herein may be implemented in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
At present, the method for estimating the navigation time of the inland river does not fully consider the periodicity and regularity of the time sequence of the navigation time and the spatial correlation between channels, and can not accurately estimate the navigation time of the ship.
The invention provides a navigation time prediction method based on deep learning, which specifically comprises the following steps: acquiring AIS data; processing the acquired AIS data to obtain navigation time data of different navigation sections and different time sections; constructing a navigation time prediction model, and inputting the navigation time data into the constructed navigation time prediction model for training; combining the trained navigation time prediction model with a route planning technology to obtain an accurate navigation time prediction value. The method constructs a navigation time prediction model, fully considers the spatial correlation and the time periodicity of navigation time in the navigation channel, and fuses the navigation channel characteristics and the navigation channel historical average time into the model to obtain better prediction precision.
Embodiments of the present invention are described in further detail below with reference to the accompanying drawings.
A navigation time prediction method based on deep learning, as shown in figure 1, comprises the following steps:
step 1: acquiring AIS data; as shown in fig. 2, the AIS data includes ship static data, ship dynamic data, ship voyage data and voyage safety information;
the ship static data comprise ship names, call signs, marine mobile service identification codes (MMSI), international Maritime Organization (IMO) numbers, ship lengths, ship widths and ship types;
the ship dynamic data comprise ship position data, ground speed/course and ship head direction information;
the ship voyage data comprise ship state, draft, destination and ETA information;
the voyage safety information comprises voyage warning and weather report information.
It can be seen that the AIS data does not have direct information about the time of flight, and therefore, when analyzing the time of flight of a ship using the AIS, it is necessary to perform processing and calculation by means of the marine mobile service identification code (MMSI) in the AIS data, the longitude and latitude of the ship, the speed of the ship, the time stamp of the AIS data, and the like.
The voyage time is aimed at the interval of the ship voyage from the starting point to the ending point of the route, however, because the planned route is too long, the time for directly counting and estimating the whole route according to the historical data generates larger errors. In order to improve the precision of the navigation time prediction, the scheme adopts a method of segment prediction to carry out statistical analysis and segmentation on the navigation time. Specific:
step 2: processing AIS data to obtain navigation time data of different navigation sections and different time sections, including:
cutting and segmenting the channel according to channel characteristics (channel trend, channel water depth, channel width and the like), wherein ships with similar types in each segment of channel have similar navigation behaviors and navigation time; as shown in fig. 3, a schematic representation of leg cuts is shown, each leg having its own number.
Calculating the average sailing speed of the ship in each sailing section by using the ship sailing speed information in the AIS data, and further calculating the average sailing time required by the ship to pass through the whole sailing section according to the average sailing speed; the average voyage time required by the ship to pass through the whole voyage section is calculated according to the average speed, and the calculation formula is as follows:
wherein T is average sailing time, V i The navigation speed of the ith ship is n, the number of ships in the navigation section is n, and l is the mileage of the navigation section.
According to traffic flow characteristics, average sailing speeds of ships have large differences in different time periods in each sailing section. Therefore, in order to improve the statistical accuracy of the voyage time, further time-division voyage time statistics is performed for each voyage segment, in this embodiment, 2 hours are used as time intervals, and the following table shows the voyage time-division voyage time statistical result obtained based on the AIS data statistics.
Step 3: constructing a navigation time prediction model, and inputting the navigation time data into the constructed navigation time prediction model for training; as shown in fig. 4, the constructed voyage time prediction model includes a convolutional neural network for capturing spatial correlation between adjacent voyages and a cyclic neural network for capturing temporal correlation. The convolutional neural network adopts a one-dimensional convolutional neural network; the cyclic neural network adopts a GRU structure cyclic neural network.
Inputting the navigation time data into the constructed navigation time prediction model for training, wherein the method comprises the following steps of:
l1, L2, L3 in fig. 4 represent the voyage time series of the upstream leg, the voyage time series of the target leg, and the voyage time series of the downstream leg, respectively; each voyage time sequence consists of voyage time of the voyage segment in the current time segment and the first 20 time segments.
Adopting Concat operation on L1, L2 and L3, splicing navigation time sequences of three navigation segments into a characteristic, and marking the characteristic as L; l=concat [ L1, L2, L3];
extracting the spliced characteristic L by adopting a one-dimensional convolutional neural network, marking the extracted characteristic as F, and carrying out one-dimensional convolutional operation according to the following steps: f=f (Σ) i∈M H i *W i +b), wherein H is the voyage time sequence, W is the weight of the convolution shift operator, b is the paranoid, f (·) is the activation function; in the one-dimensional convolutional neural network, each convolutional shift operator represents a system for extracting navigation time sequence features, the weight parameters of the convolutional shift operators are continuously adjusted in the training process through an error counter-propagation mode, and finally the best spatial correlation features among three navigation segments are learned.
Inputting the extracted feature F into the cyclic neural network to obtain a feature extracted by the navigation time prediction model; the cyclic neural network is provided with 20 inputs, which are respectively predicted navigation time characteristics of the first 20 time periods, and each navigation time characteristic is obtained by comprehensively extracting three navigation segments according to space through the one-dimensional convolutional neural network; when the invention predicts the navigation time, the cyclic neural network with the GRU structure is adopted, and the structure is shown in figure 5. It comprises 1 GRU layer, input is 20, output 1, i.e. the voyage time of the voyage section for one period of time in the future.
Because the space characteristics of the channel comprise the length, the width, the depth and the like of the navigation section and the traffic capacity of the channel are related, the traffic speed of the ship is greatly influenced, and the navigation time of the ship passing through the navigation section is influenced. Therefore, in order to increase the comprehensiveness of the consideration factors of the prediction model and improve the prediction effect of the model, the spatial characteristics of the channel are further fused with the characteristics extracted by the navigation time prediction model, and the spatial characteristics of the channel participating in prediction comprise three elements of the length of a navigation section, the width of the navigation and the depth of water; in this embodiment, their values are shown in the following table:
in addition, for a certain period of time in a day, the traffic behaviors of a voyage have certain similarity, and the similarity is helpful for improving the prediction effect, so that 1 historical similarity feature of the voyage is fused with the spatial features of three voyages and the features extracted by the voyage time prediction model, and then the three historical similarity features are input into a fully-connected layer, and finally the voyage time of the specified voyage in the next period of time is output.
To evaluate how long the historical data was taken to calculate the historical correlation characteristics, the voyage time for 12 time periods per day (2 hours per period) was subjected to correlation analysis with the average voyage time for the corresponding time period for the previous n days. Taking the 6 th navigation segment as an example, the navigation time of each time segment in 2018/5/7 days is compared with the navigation time corresponding to each time segment in the previous n days, and the result shows that the average value in the previous 7 days has better similarity and the calculation consumption brought by the average value is more reasonable. The following table shows specific corresponding values, while figure 6 reflects their similarity. In fig. 6, the abscissa indicates each time period of the day, the ordinate indicates the voyage time, the blue curve indicates the voyage time of 2018/5/7 day in each time period, and the yellow curve indicates the average voyage time of the first seven days in each time period. As can be seen from the graph, the two curves have similar trend and higher relevance.
In summary, the features extracted by the voyage time prediction model (convolutional neural network+GRU cyclic neural network) are fused with 3 channel space features and 1 channel history similarity feature, and are input into a full connection layer together, and finally the voyage time of the designated voyage section in the next time period is output.
Corresponding to the navigation time prediction method based on the deep learning in the application, the application also provides a navigation time prediction device based on the deep learning, which comprises the following steps: the device comprises an acquisition unit, a calculation unit, a training unit and a generation unit; wherein:
the acquisition unit is used for acquiring AIS data;
the computing unit is used for processing the AIS data acquired by the acquisition unit to obtain navigation time data of different navigation sections and different time sections;
the training unit is used for constructing a navigation time prediction model, and inputting the navigation time data into the constructed navigation time prediction model for training;
and the generating unit is used for combining the trained navigation time prediction model with the route planning technology to obtain an accurate navigation time prediction value.
For the embodiments of the present invention, since they correspond to those in the above embodiments, the description is relatively simple, and the relevant similarities will be found in the description of the above embodiments, and will not be described in detail herein.
In order to verify the effect of the prediction model, besides the prediction model fusing various characteristics, a prediction model based on LSTM alone, a prediction model based on GRU alone and a prediction model only containing the dynamic characteristics of ship traffic are realized for comparison analysis.
And training the AIS data and the leg data of the No. 2-8 legs by using 11 months in 2018-2019 respectively by using the four models, performing a test experiment by using 1 month data as a verification set, and adopting a relative error as an evaluation standard, wherein the final experimental result is shown in the following table. It can be seen from the following table that the multi-feature fused prediction model has the smallest relative error, i.e. the highest prediction accuracy.
The embodiments also disclose a computer readable storage medium having stored therein a set of computer instructions which when executed by a processor implement the deep learning based voyage time prediction method as provided in any of the embodiments above.
In the several embodiments provided in the present application, it should be understood that the disclosed technology content may be implemented in other manners. The above-described embodiments of the apparatus are merely exemplary, and the division of the units, for example, may be a logic function division, and may be implemented in another manner, for example, a plurality of units or components may be combined or may be integrated into another system, or some features may be omitted, or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be through some interfaces, units or modules, or may be in electrical or other forms.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional unit in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.
The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied essentially or in part or all of the technical solution or in part in the form of a software product stored in a storage medium, including instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a removable hard disk, a magnetic disk, or an optical disk, or other various media capable of storing program codes.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some or all of the technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the invention.
Claims (6)
1. A method for predicting voyage time based on deep learning, comprising:
acquiring AIS data;
processing the acquired AIS data to obtain navigation time data of different navigation sections and different time sections; comprising the following steps:
cutting and segmenting the channels according to channel characteristics, wherein ships with similar types in each segment of channel have similar navigation behaviors and navigation time;
calculating the average sailing speed of the ship in each sailing section by using the ship sailing speed information in the AIS data, and further calculating the average sailing time required by the ship to pass through the whole sailing section according to the average sailing speed; the calculation formula is as follows:
wherein T is average sailing time, V i The navigation speed of the ith ship is n, the number of the ships in the navigation section is n, and l is the mileage of the navigation section;
according to traffic flow characteristics, setting the time interval as n hours, and carrying out time-division navigation time statistics on each navigation segment;
constructing a navigation time prediction model, wherein the constructed navigation time prediction model comprises a convolutional neural network for capturing the spatial correlation between adjacent voyages and a cyclic neural network for capturing the temporal correlation; inputting the navigation time data into the constructed navigation time prediction model for training; comprising the following steps:
setting L1, L2 and L3 to respectively represent the navigation time sequence of an upstream navigation section, the navigation time sequence of a target navigation section and the navigation time sequence of a downstream navigation section;
adopting Concat operation to the L1, L2 and L3, splicing the navigation time sequences of the three navigation segments into a characteristic, and marking the characteristic as L; L=Concat [ L1, L2, L3]
Extracting the spliced characteristic L by adopting a one-dimensional convolutional neural network, marking the extracted characteristic as F, and carrying out one-dimensional convolutional operation according to the following steps: f=f (Σ) i∈M H i *W i +b), wherein H is the voyage time sequence, W is the weight of the convolution shift operator, b is the paranoid, and f (g) is the activation function;
inputting the extracted feature F into a cyclic neural network to obtain a feature extracted by the navigation time prediction model; the cyclic neural network is provided with 20 inputs, which are respectively predicted navigation time characteristics of the first 20 time periods, and each navigation time characteristic is obtained by comprehensively extracting three navigation segments according to space through the one-dimensional convolutional neural network;
fusing the space characteristics of the channel with the characteristics extracted by the navigation time prediction model, wherein the space characteristics of the channel participating in prediction comprise three elements of a navigation section length, a navigation width and a water depth;
fusing the history similarity features of the 1 channel with the spatial features of the three channels and the features extracted by the navigation time prediction model, inputting the fused features into a full-connection layer, and finally outputting the navigation time of the appointed navigation section in the next time section;
combining the trained navigation time prediction model with a route planning technology to obtain an accurate navigation time prediction value.
2. The deep learning based voyage time prediction method of claim 1, wherein the AIS data comprises ship static data, ship dynamic data, ship voyage data and voyage safety information;
the ship static data comprise ship names, call signs, marine mobile service identification codes (MMSI), international Maritime Organization (IMO) numbers, ship lengths, ship widths and ship types;
the ship dynamic data comprise ship position data, ground speed/course and ship head direction information;
the ship voyage data comprise ship state, draft, destination and ETA information;
the voyage safety information comprises voyage warning and weather report information.
3. The deep learning-based voyage time prediction method as claimed in claim 1, wherein the convolutional neural network adopts a one-dimensional convolutional neural network; the cyclic neural network adopts a GRU structure cyclic neural network.
4. The deep learning based voyage time prediction method of claim 1, wherein each voyage time sequence consists of voyage time of the voyage in the current time period and the previous 20 time periods.
5. A voyage time prediction device based on the voyage time prediction method based on deep learning as claimed in any one of claims 1 to 4, comprising:
the acquisition unit is used for acquiring AIS data;
the computing unit is used for processing the AIS data acquired by the acquisition unit to obtain navigation time data of different navigation sections and different time sections;
the training unit is used for constructing a navigation time prediction model, and inputting the navigation time data into the constructed navigation time prediction model for training;
and the generating unit is used for combining the trained navigation time prediction model with the route planning technology to obtain an accurate navigation time prediction value.
6. A computer-readable storage medium having a set of computer instructions stored therein; the set of computer instructions, when executed by a processor, implements a deep learning based voyage time prediction method as claimed in any of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010530842.3A CN111738500B (en) | 2020-06-11 | 2020-06-11 | Navigation time prediction method and device based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010530842.3A CN111738500B (en) | 2020-06-11 | 2020-06-11 | Navigation time prediction method and device based on deep learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111738500A CN111738500A (en) | 2020-10-02 |
| CN111738500B true CN111738500B (en) | 2024-01-12 |
Family
ID=72648851
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010530842.3A Active CN111738500B (en) | 2020-06-11 | 2020-06-11 | Navigation time prediction method and device based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111738500B (en) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112330297A (en) * | 2020-11-18 | 2021-02-05 | 深圳壹账通智能科技有限公司 | Block chain-based barge berthing method, block chain-based barge berthing device and storage medium |
| CN112118143B (en) * | 2020-11-18 | 2021-02-19 | 迈普通信技术股份有限公司 | Traffic prediction model training method, traffic prediction method, device, equipment and medium |
| CN112561452B (en) * | 2021-02-19 | 2021-05-14 | 深圳普罗空运有限公司 | Goods tracking method based on international air transportation logistics APP |
| CN113221450B (en) * | 2021-04-27 | 2024-03-12 | 中国科学院国家空间科学中心 | Space-time prediction method and system for sparse non-uniform time sequence data |
| CN113096481B (en) * | 2021-05-19 | 2022-09-02 | 江苏科技大学 | Ship path planning training system for laboratory |
| CN113393048B (en) * | 2021-06-24 | 2023-09-15 | 武汉长江船舶设计院有限公司 | Electric cruise ship navigation energy consumption prediction and control method |
| CN113610347B (en) * | 2021-07-05 | 2024-04-05 | 上海海事大学 | Ship shipping market capacity assessment method |
| CN114118604B (en) * | 2021-12-02 | 2024-05-28 | 武汉理工大学 | Method and system for predicting ferry crossing channel behavior based on machine learning |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103531043A (en) * | 2013-09-27 | 2014-01-22 | 重庆大学 | Point-to-point ship sailing time estimation method based on ship route matching |
| CN107316501A (en) * | 2017-06-28 | 2017-11-03 | 北京航空航天大学 | A kind of SVMs Travel Time Estimation Method based on grid search |
| CN109035761A (en) * | 2018-06-25 | 2018-12-18 | 复旦大学 | Travel time estimation method based on back-up surveillance study |
| CN109214107A (en) * | 2018-09-26 | 2019-01-15 | 大连海事大学 | A kind of ship's navigation behavior on-line prediction method |
| CN110110943A (en) * | 2019-05-21 | 2019-08-09 | 大连海事大学 | A kind of fleet's efficiency comprehensive intelligent Optimal Management System and optimization method based on big data |
| CN110321785A (en) * | 2019-05-09 | 2019-10-11 | 北京尚文金泰教育科技有限公司 | A method of introducing ResNet deep learning network struction dermatoglyph classification prediction model |
| CN110555568A (en) * | 2019-09-12 | 2019-12-10 | 重庆交通大学 | Road traffic running state real-time perception method based on social network information |
| GB201917510D0 (en) * | 2019-11-29 | 2020-01-15 | Percy Iain Bryden | Navigation guidance method and system |
| CN110827543A (en) * | 2019-11-11 | 2020-02-21 | 重庆邮电大学 | Short-term traffic flow control method based on deep learning and spatio-temporal data fusion |
| CN111091708A (en) * | 2019-12-13 | 2020-05-01 | 中国科学院深圳先进技术研究院 | Vehicle track prediction method and device |
| CN111242344A (en) * | 2019-12-11 | 2020-06-05 | 大连海事大学 | Intelligent water level prediction method based on cyclic neural network and convolutional neural network |
-
2020
- 2020-06-11 CN CN202010530842.3A patent/CN111738500B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103531043A (en) * | 2013-09-27 | 2014-01-22 | 重庆大学 | Point-to-point ship sailing time estimation method based on ship route matching |
| CN107316501A (en) * | 2017-06-28 | 2017-11-03 | 北京航空航天大学 | A kind of SVMs Travel Time Estimation Method based on grid search |
| CN109035761A (en) * | 2018-06-25 | 2018-12-18 | 复旦大学 | Travel time estimation method based on back-up surveillance study |
| CN109214107A (en) * | 2018-09-26 | 2019-01-15 | 大连海事大学 | A kind of ship's navigation behavior on-line prediction method |
| CN110321785A (en) * | 2019-05-09 | 2019-10-11 | 北京尚文金泰教育科技有限公司 | A method of introducing ResNet deep learning network struction dermatoglyph classification prediction model |
| CN110110943A (en) * | 2019-05-21 | 2019-08-09 | 大连海事大学 | A kind of fleet's efficiency comprehensive intelligent Optimal Management System and optimization method based on big data |
| CN110555568A (en) * | 2019-09-12 | 2019-12-10 | 重庆交通大学 | Road traffic running state real-time perception method based on social network information |
| CN110827543A (en) * | 2019-11-11 | 2020-02-21 | 重庆邮电大学 | Short-term traffic flow control method based on deep learning and spatio-temporal data fusion |
| GB201917510D0 (en) * | 2019-11-29 | 2020-01-15 | Percy Iain Bryden | Navigation guidance method and system |
| CN111242344A (en) * | 2019-12-11 | 2020-06-05 | 大连海事大学 | Intelligent water level prediction method based on cyclic neural network and convolutional neural network |
| CN111091708A (en) * | 2019-12-13 | 2020-05-01 | 中国科学院深圳先进技术研究院 | Vehicle track prediction method and device |
Non-Patent Citations (4)
| Title |
|---|
| Roll Motion Prediction of Unmanned Surface Vehicle Based on Coupled CNN and LSTM;Wenjie Zhang等;《Future internet》;第1-15页 * |
| Water Level Prediction Model Based on GRU and CNN;Mingyang pan等;《ACCESS》;第60090-60100页 * |
| 信息化时代的内河船员服务;钱正锋;;中国海事(第07期);第13-14页 * |
| 基于改进A*算法的内河水网航线规划及应用;潘明阳等;《上海海事大学学报》;第40-45页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111738500A (en) | 2020-10-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111738500B (en) | Navigation time prediction method and device based on deep learning | |
| Zhang et al. | A predictive analytics method for maritime traffic flow complexity estimation in inland waterways | |
| Xiao et al. | Maritime traffic probabilistic forecasting based on vessels’ waterway patterns and motion behaviors | |
| CN111160753B (en) | Road network node importance assessment method based on knowledge graph | |
| Wang et al. | A novel approach for generating routable road maps from vehicle GPS traces | |
| Xin et al. | A simulation model for ship navigation in the “Xiazhimen” waterway based on statistical analysis of AIS data | |
| Miller et al. | Mining traffic incidents to forecast impact | |
| CN114936332A (en) | Method and device for analyzing ship traffic flow in water area of wind power plant and storage medium | |
| Liu et al. | A data mining method to extract traffic network for maritime transport management | |
| CN114139642A (en) | Ship route extraction method combining ship journey semantic object and graph theory | |
| Eljabu et al. | Anomaly detection in maritime domain based on spatio-temporal analysis of ais data using graph neural networks | |
| He et al. | Mining channel water depth information from IoT-based big automated identification system data for safe waterway navigation | |
| Xu et al. | Trajectory clustering for SVR-based Time of Arrival estimation | |
| CN102376025A (en) | Method for simulating mobile phone data and evaluating urban road network traffic condition | |
| CN114490913A (en) | Method and device for determining state of ship entering port and electronic equipment | |
| CN116823572B (en) | Population flow data acquisition method and device and computer readable storage medium | |
| CN114565176B (en) | Long-term ship track prediction method | |
| Molavipour et al. | Causality graph of vehicular traffic flow | |
| Zohoori et al. | A VECTORIZED ALGORITHM FOR WATERWAY TRAFFIC ANALYSIS USING AIS DATA. | |
| Fuentes et al. | A spatial framework for extracting Suez Canal transit information from AIS | |
| CN112800038B (en) | Water environment data analysis method | |
| Mannarini et al. | Preliminary inter-comparison of AIS data and optimal ship tracks | |
| CN114118604A (en) | Ferry channel crossing behavior prediction method and system based on machine learning | |
| Lawlor et al. | Time-varying mixtures of Markov chains: an application to road traffic modeling | |
| Sun et al. | Alleviating data sparsity problems in estimated time of arrival via auxiliary metric learning |
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 |