CN113393027B - Navigation mark drift intelligent prediction method based on deep learning - Google Patents
Navigation mark drift intelligent prediction method based on deep learning Download PDFInfo
- Publication number
- CN113393027B CN113393027B CN202110648282.6A CN202110648282A CN113393027B CN 113393027 B CN113393027 B CN 113393027B CN 202110648282 A CN202110648282 A CN 202110648282A CN 113393027 B CN113393027 B CN 113393027B
- Authority
- CN
- China
- Prior art keywords
- navigation mark
- data
- navigation
- drift
- time
- 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 41
- 238000013135 deep learning Methods 0.000 title claims abstract description 20
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims abstract description 37
- 230000007246 mechanism Effects 0.000 claims abstract description 17
- 238000007781 pre-processing Methods 0.000 claims abstract description 12
- 230000033001 locomotion Effects 0.000 claims abstract description 9
- 238000012549 training Methods 0.000 claims abstract description 6
- 230000006870 function Effects 0.000 claims description 16
- 238000004364 calculation method Methods 0.000 claims description 12
- 230000004913 activation Effects 0.000 claims description 5
- 238000012217 deletion Methods 0.000 claims description 3
- 230000037430 deletion Effects 0.000 claims description 3
- 210000002569 neuron Anatomy 0.000 claims description 3
- 238000003672 processing method Methods 0.000 claims description 3
- 238000013528 artificial neural network Methods 0.000 description 5
- 238000001914 filtration Methods 0.000 description 4
- 230000002265 prevention Effects 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 125000004122 cyclic group Chemical group 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000007726 management method Methods 0.000 description 3
- 238000010606 normalization Methods 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 210000004556 brain Anatomy 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 238000010219 correlation analysis Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000000306 recurrent effect Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 239000000725 suspension Substances 0.000 description 2
- 230000005856 abnormality Effects 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 210000004027 cell Anatomy 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000010921 in-depth analysis Methods 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 238000004321 preservation Methods 0.000 description 1
- 238000000611 regression analysis Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012360 testing method 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
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/29—Geographical information databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- 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
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/067—Enterprise or organisation modelling
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- General Physics & Mathematics (AREA)
- Economics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Marketing (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biophysics (AREA)
- Databases & Information Systems (AREA)
- Artificial Intelligence (AREA)
- Game Theory and Decision Science (AREA)
- Molecular Biology (AREA)
- General Business, Economics & Management (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Educational Administration (AREA)
- Remote Sensing (AREA)
Abstract
The invention provides a navigation mark drift intelligent prediction method based on deep learning, which relates to the technical field of navigation mark drift prediction, and comprises the following steps: acquiring sample data and preprocessing the sample data; the sample data comprises navigation mark historical position data and water level historical data; constructing a time sequence based on the preprocessed sample data; establishing and training a navigation mark drift prediction model by utilizing a navigation mark motion rule reflected by a time sequence; the navigation mark drift prediction model comprises a gating circulation unit for extracting time sequence characteristics in the time sequence and a attention mechanism for giving different weights to each time step of the time sequence; inputting preset time data of a navigation mark, and predicting by using a trained navigation mark drift prediction model; and outputting the navigation mark predicted position. The method based on deep learning learns time sequence characteristics from a large number of navigation marks and water level historical data, and further improves model prediction accuracy by considering water level factor influence and an attention adding mechanism.
Description
Technical Field
The invention relates to the technical field of navigation mark drift prediction, in particular to an intelligent navigation mark drift prediction method based on deep learning.
Background
A navigation mark is a facility or system that helps guide a ship to navigate, locate and mark obstacles and represent a warning, providing safety information for various water activities. Are often placed in or near navigable areas to indicate the location of tunnels, anchors, beach hazards, and other obstacles. The buoy is a buoy floating on the water surface, and has the greatest quantity in the buoy and wide application. Buoys are usually arranged along the boundaries of the channel, are tethered to the water surface by tie-downs and anchor chains, are affected by natural and human factors such as tide fluctuation, past ship collision, self-failure of the tethered device and the like, can generate offset phenomena, and directly affect the channel dimension. If the position of the buoy is too far from being timely perceived, wrong navigation aiding information is sent to the past ship, and navigation safety is threatened.
In recent years, deep learning techniques have rapidly developed, and are widely used in the fields of numerical prediction, image recognition, and the like. In the field of numerical prediction, theories and methods based on cyclic neural networks are excellent in performance, have strong data adaptation capability, can capture the connection between a plurality of characteristics, are good at short-term prediction, and are common methods in the problem of time sequence prediction. The digital channel telemetry and remote control system is also a typical time series data for collecting accumulated navigation mark position GPS observation data and water level data for a long time. A large amount of historical observation data provides data support for in-depth analysis of factors affecting the navigation mark drift and mining of the law of navigation mark drift movement.
If the latest deep learning method can be utilized, a high-applicability navigation mark drift prediction method is constructed by combining the movement rule of navigation mark drift, the accurate and rapid prediction of the navigation mark drift position and drift distance is carried out, and early warning is carried out in advance, so that the method has important significance in navigation ship route planning, collision prevention and other aspects.
At present, the traditional navigation mark drift prediction method comprises the following steps:
journal paper: inland navigation mark drift characteristic research based on Kalman filtering and K-means++ algorithm (Zhou Yumeng, junior citizens, jiang Zhonglian, zhong Cheng. Inland navigation mark drift characteristic research based on Kalman filtering and K-means++ algorithm [ J ]. University of martial arts university (traffic science and engineering edition), 2019,43 (01): 81-85.).
The paper proposes that the Kalman filtering is used for preprocessing the navigation mark data, the K-means++ algorithm is combined to obtain a navigation mark clustering center, a correlation analysis method is adopted to establish a inland navigation mark drift-water level quadratic fit model, the navigation mark drift distance can be estimated according to the water level value, and the model can better reflect the influence of the water level of the river reach in the Yangtze river on the navigation mark drift.
This method has the following disadvantages: the Kalman filtering algorithm and the K-means++ algorithm (the flow chart of which is shown in figure 1) can only process the processing history data, cannot perform real-time calculation, and the rationality of taking the clustering center as a reference position is yet to be verified; the constructed linear regression model has a simple structure and large prediction errors.
Journal paper: the calculation method of the navigation mark drift under the action of the tidal current field is researched (Zhou Chunhui, zhao Junnan, gan Langxiong, xu Yanmin, xu Caiyun) the calculation method of the navigation mark drift under the action of the tidal current field is researched [ J ]. Safety and environmental school report, 2021,21 (01): 217-223 ].
The paper proposes preprocessing navigation mark telemetry data through Laida mathematical criteria, clustering the preprocessed data by combining a K-means++ algorithm and an ISODATA algorithm, selecting a clustering center with higher accuracy as a reference point for calculating navigation mark drift amount, and calculating navigation mark drift distance. The method for analyzing the navigation mark drift by adopting the Person correlation analysis method and the regression analysis method can be used for constructing a navigation mark drift model under the action of a tide field, reducing the false alarm behavior caused by navigation mark drift in the tide field, and providing a stronger practical reference for maintenance and management of a remote control and telemetry system of the marine navigation mark.
According to the method, the precision of the navigation mark drift prediction result is improved by analyzing the relation between the flow direction and flow velocity of the tide and the navigation mark drift amount, but the model structure is simpler, the time sequence characteristics of navigation mark data are ignored, and the prediction precision needs to be further improved.
Disclosure of Invention
In view of the above, the invention provides an intelligent navigation mark drift prediction method based on deep learning, and provides a navigation mark drift prediction model based on the combination of an attention mechanism and a cyclic neural network, which can predict the navigation mark drift position and drift distance in a short period through navigation mark GPS observation data, thereby improving the prediction precision, and the prediction result can be used for navigation departments and navigation ships to refer to, thereby being beneficial to improving the navigation safety level of the navigation channel.
For this purpose, the invention provides the following technical scheme:
the invention provides a navigation mark drift intelligent prediction method based on deep learning, which comprises the following steps:
s1, acquiring sample data, and preprocessing the sample data; wherein the sample data includes navigation mark historical position data and water level historical data;
s2, constructing a time sequence based on the preprocessed sample data;
s3, establishing and training a navigation mark drift prediction model by utilizing a navigation mark motion rule reflected by a time sequence; the navigation mark drift prediction model comprises a gating circulation unit for extracting time sequence characteristics in the time sequence and a attention mechanism for giving different weights to each time step of the time sequence;
s4, inputting preset time data of the navigation mark, and predicting by using a trained navigation mark drift prediction model;
s5, outputting the navigation mark predicted position.
Further, the network structure of the navigation mark drift prediction model includes: the device comprises an input layer, 2 GRU layers, an Attention layer and an output layer;
the network parameters of the navigation mark drift prediction model comprise:
the input vector shape of the input layer is [64,24,3],64 is the data size of the primary input network, namely, batch_size,24 is a time step, and 3 is a characteristic quantity, namely, a longitude value, a latitude value and a water level value;
the number of neurons of the two GRU layers is 128 and 64 respectively, and the activation functions are tanh functions.
Further, the Attention mechanism calculation formula is as follows:
h k′ ＝tanh(W c [C；h k ])；
wherein x is 1 ,x 2 ,...,x k Represents the input sequence, h 1 ,h 2 ,...,h k Representing state values of hidden layers corresponding to an input sequence, scoring function s kj By dot product, alpha ki Attention weight of current input for hidden layer of history input, C is intermediate vector; h is a k′ And the final output current time hidden layer state value is represented.
Further, preprocessing the sample data includes:
for data repetition, finding a repeated record item and deleting the repeated record item;
for data deletion, linear interpolation is adopted to fill data;
for data anomalies, detecting outliers by using a box graph, and filling by using a data missing processing method after eliminating the outliers.
Further, the linear interpolation formula includes:
wherein t is i And x i Respectively filling time point and filling value, x i+1 、x i-1 Respectively adjacent time point observations.
Further, the data preprocessing further comprises:
extracting navigation marks and water level data as a data set according to a preset time interval;
normalization is performed for each column in the dataset.
Further, the loss function of the navigation mark drift prediction model adopts an average absolute error.
The technical scheme has the following beneficial effects: the invention creatively provides a navigation mark drift prediction model, a time sequence characteristic is learned from a large amount of navigation marks and water level historical data based on a deep learning method, and the model prediction precision is further improved by considering the influence of water level factors and an attention adding mechanism, so that the prediction precision meets the actual application requirements.
The navigation method and the navigation system can be applied to guiding the navigation of the ship, the ship can refer to the navigation mark drifting predicted position to conduct route planning, and navigation efficiency of the navigation channel is improved. The navigation mark drift pre-warning method can be further applied to the aspect of navigation mark management, and the supervision department can better master the navigation mark drift dynamics according to the navigation mark drift prediction position, so that drift pre-warning is carried out, and the navigation safety level of the navigation mark is improved. The predicted navigation mark drifting position has important significance in the aspect of collision prevention, and the ship can take collision prevention measures in advance according to the predicted position, so that navigation safety is ensured, and property loss is reduced.
In conclusion, the technical scheme of the invention is used for realizing the navigation mark drift prediction, has important application in various aspects such as navigation planning, navigation mark management, collision prevention and the like, and has obvious economic and social benefits.
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 flowchart of a K-means++ algorithm of the prior art;
FIG. 2 is a flow chart of a method for intelligent prediction of navigation mark drift based on deep learning in an embodiment of the invention;
FIG. 3 is a schematic diagram of a navigation mark drift prediction model based on the Attention-GRU in the embodiment of the invention;
FIG. 4 is a schematic diagram of a GRU unit structure according to an embodiment of the invention;
FIG. 5 is a schematic diagram of the structure of an Attention unit according to 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.
The invention aims to predict the drift position and the drift distance of a navigation mark in a short period by constructing time sequence characteristics among high-precision model learning data by using a deep learning method based on a large amount of navigation mark and water level historical data.
There are generally two factors for buoy displacement: firstly, the navigation mark is shifted under the influence of water flow, tide and wind; secondly, the navigation mark shift is caused by human factors, namely collision or dragging of the past ship. Under the influence of tide, river water makes reciprocating motion to cause periodic change of water level. The navigation mark is used as a suspension body floating on the water surface and can do reciprocating motion along with the suspension body. Thus, water level variation is an important factor affecting the drift of the navigation mark.
The basic idea of the invention is that a time sequence is constructed based on navigation mark historical GPS position data, namely longitude value and latitude value, and water level characteristics are considered, then a time sequence characteristic of the sequence is extracted by utilizing a deep cyclic neural network, different weights are further given to each time step of an input sequence by using an attention mechanism, and the influence of more key moments is highlighted to help a model to make more accurate judgment. And finally, predicting the longitude value and the latitude value of the navigation mark position in a short period by using the trained deep learning model.
Referring to fig. 2, a flowchart of a navigation mark drift intelligent prediction method based on deep learning in an embodiment of the invention is shown, and the method comprises the following steps:
s1, acquiring sample data, and preprocessing the sample data;
wherein the sample data includes navigation mark historical position data and water level historical data;
due to the reasons of measurement accuracy, data transmission faults and the like, a small amount of poor quality exists in the navigation mark historical position data and the water level historical data. For time series prediction, the quality of data is directly related to the availability and robustness of a model, and the prediction accuracy of the model is greatly affected. Therefore, the preprocessing work on the raw data is important.
The water level and navigation mark observation data mainly have the problems of data repetition, data deletion, data abnormality and the like.
The data repetition refers to a plurality of records with the same navigation mark or water level station appearance time stamp, and the recorded information is the same. For data repetition, only duplicate entries need to be found and deleted.
Data missing refers to the lack of data records at several points in time. The observation navigation mark and the original observation data of the water level find that the data is less missing. Therefore, data filling is performed by linear interpolation, as shown in formula (1), t i And x i Respectively filling time point and filling value, x i+1 、x i-1 Respectively adjacent time point observations.
Data anomalies (error values) refer to extremely small or large values in a time series, where the data values are in error or deviate significantly from normal levels. The embodiment of the invention uses a box graph to detect abnormal values, and uses a data missing processing method to fill after eliminating the abnormal values.
After data preprocessing, the navigation mark and water level data are extracted according to a preset time interval to be used as experimental data sets, and the final data sets are formed as shown in table 1, wherein the preset time interval in table 1 is 2 hours.
TABLE 1
| Sequence number | Time | Water level | Longitude and latitude | Latitude of latitude |
| 0 | 2020/2/20 22：00 | 1.78 | 120.559559 | 32.009901 |
| 1 | 2020/2/21 0:00 | 2.83 | 120.559503 | 32.009912 |
| 2 | 2020/2/21 2:00 | 2.98 | 120.559422 | 32.009961 |
| 3 | 2020/2/21 4:00 | 2.32 | 120.559445 | 32.009959 |
| 4 | 2020/2/21 6:00 | 1.65 | 120.559524 | 32.009917 |
| 5 | 2020/2/21 8:00 | 1.24 | 120.559565 | 32.009914 |
| 6 | 2020/2/21 10:00 | 1.98 | 120.559559 | 32.009908 |
Because the data values of the water level, the longitude and the latitude have larger difference, normalization processing is needed for each column in the data set, and the normalization calculation formula is shown as formula (2):
where x is a value to be normalized, x min 、x max When the columns are the minimum and maximum, respectively, x' is the normalized value.
S2, constructing a time sequence based on the preprocessed sample data;
s3, establishing and training a navigation mark drift prediction model by utilizing a navigation mark motion rule reflected by a time sequence; the navigation mark drift prediction model comprises a gating circulation unit for extracting time sequence characteristics in the time sequence and a attention mechanism for giving different weights to each time step of the time sequence;
and constructing a navigation mark drift prediction model by utilizing a navigation mark motion rule, namely constructing the navigation mark drift prediction model by considering the water level characteristics, wherein the prediction model is integrated with an attention mechanism. The navigation mark drift prediction model is based on an Attention mechanism (Attention) and a gating circulating unit (GRU), and the construction of the prediction model comprises an overall network structure, super parameters, parameters of an input layer, 2 GRU layers, an Attention layer, an output layer and the like, and a loss function and an optimization method of the network. The GRU (Gate Recurrent Unit, gated loop unit) is one type of loop neural network (Recurrent Neural Network, RNN) that has been proposed to address long-term memory and gradients in back propagation.
S4, inputting preset time data of the navigation mark, and predicting by using a trained navigation mark drift prediction model;
in the embodiment of the invention, the preset time data are 24 time data.
S5, outputting the navigation mark predicted position.
In the embodiment of the invention, a plurality of navigation marks of navigation sections at the lower reaches of the Yangtze river are selected for comparison experiments, the RMSE value of the drift distance predicted value is used as an evaluation index, and the calculation result is shown in table 2.
TABLE 2
The embodiment of the invention provides a navigation mark drift prediction model based on Attention-GRU, a deep learning-based method learns time sequence characteristics from a large number of navigation marks and water level historical data, and the model prediction precision is further improved by considering the influence of water level factors and an Attention adding mechanism, and meets the actual application requirements.
For easy understanding, the following describes in detail the establishment of the model for the navigation mark drift prediction in step S2 in the embodiment of the present invention.
(1) Construction of a model
The navigation mark drift prediction model structure based on the Attention and the GRU is shown in figure 3.
In this model, the core parameters are as follows:
1) An input layer. The input vector shape is [64,24,3],64 is the data size of one input network, namely, batch_size,24 is a time step, and 3 is the characteristic quantity, namely, a longitude value, a latitude value and a water level value, wherein the batch_size and the time step can be adjusted by self during subsequent testing.
2) And a GRU layer. The number of neurons of the two GRU layers is 128 and 64 respectively, and the activation functions are tanh functions. The GRU introduces a gating mechanism, makes full use of history information, and can solve the long-term dependence problem existing in long distance. The GRU unit structure is shown in FIG. 4.
The GRU contains only an update gate and a reset gate, the update gate controlling the extent to which state information from a previous time is brought into the current state, a larger value indicating more information was brought from the previous time. The reset gate controls the extent to which the current state is combined with the previous information, the larger the value the more information is retained. The specific calculation formula is as follows:
in the formula (3), r t 、z t Representing the values of reset gate, update gate, x, respectively t For the input at the time t,representing candidate hidden state values at time t, h t Conceal the layer state value for time t, W r 、W z 、W h For training the parameter matrix, σ and tanh represent the sigmoid activation function and hyperbolic tangent activation function, respectively, and, as indicated by the radix, the corresponding elements of the matrix are multiplied. Reset gate r t For controlling hiding state h from previous time step t-1 If r t Approaching 0 means resetting the corresponding hidden state h t-1 0, i.e. discard the hidden state of the last time step, if r t Approaching 1 indicates a greater preservation of the hidden state of the previous time step. Updating door z t For controlling the hidden state to be includedThe update degree of the candidate hidden state of the previous time step information, on the one hand, z t Determine the last time step hidden state h t-1 Inflow into the current hidden state h t On the other hand, 1-z t Deciding candidate hidden state->To the current hidden state h t Is updated by the update degree of the above-mentioned program. Assume that the update gate is in time steps t 'to t (t'<t) is always approximately 1, i.e. the input information between time steps t' to t hardly flows into the hidden state h of time step t t Can be regarded as a hidden state h at an earlier time t‘-1 Always save and pass to the current time step t.
3) An attach layer. The Attention mechanism is a model for simulating the Attention of the human brain, and the Attention of the human brain to things at a specific moment is considered to be concentrated to a specific place, and the Attention of other parts is reduced or even ignored. In practical training, the GRU needs to capture sequence information step by step, and the performance of the GRU on a long sequence gradually decays along with the increase of the number of steps, so that the GRU has difficulty in retaining all useful information. The Attention mechanism gives different weights to each time step of the input sequence, so that the influence of more key time is highlighted, the model is helped to make more accurate judgment, and calculation and storage cost of the model cannot be increased. The Attention cell structure is shown in fig. 5.
In FIG. 5, x 1 ,x 2 ,...,x k Represents the input sequence, h 1 ,h 2 ,...,h k Representing state values of hidden layers corresponding to an input sequence, scoring function s kj By dot product, alpha ki Attention weight for the current input for hidden layer of historical input, C is the intermediate vector. h is a k′ And the final output current time hidden layer state value is represented. The Attention mechanism calculation formula is as follows:
h k′ ＝tanh(W c [C；h k ]) (7)
(2) Loss function
Loss calculation is required according to the predicted result and the actual result of forward propagation, so that the weight parameters are updated in the process of backward propagation. Here, the loss function uses MAE (Mean Absolute Error, average absolute error), that is, the average value of the absolute error value of the observed value and the true value, and its calculation formula is shown in equation (8).
In the formula (8), m represents the total number of samples, y i Andthe true position (longitude and latitude coordinates) and the predicted position (longitude and latitude coordinates) of the navigation mark are respectively.
The embodiment of the invention also discloses a computer readable storage medium, wherein a computer instruction set is stored in the computer readable storage medium, and when the computer instruction set is executed by a processor, the intelligent navigation mark drift prediction method based on the deep learning provided by any embodiment is realized.
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 (5)
1. The intelligent navigation mark drift prediction method based on deep learning is characterized by comprising the following steps of:
s1, acquiring sample data, and preprocessing the sample data; wherein the sample data includes navigation mark historical position data and water level historical data;
s2, constructing a time sequence based on the preprocessed sample data;
s3, establishing and training a navigation mark drift prediction model by utilizing a navigation mark motion rule reflected by a time sequence; the navigation mark drift prediction model comprises a gating circulation unit for extracting time sequence characteristics in the time sequence and a attention mechanism for giving different weights to each time step of the time sequence;
s4, inputting preset time data of the navigation mark, and predicting by using a trained navigation mark drift prediction model;
s5, outputting a navigation mark predicted position;
the network structure of the navigation mark drift prediction model comprises: an input layer, 2 GRU layers, an Attention layer and an output layer;
the network parameters of the navigation mark drift prediction model comprise:
the input vector shape of the input layer is [64,24,3],64 is the data size of the primary input network, namely, batch_size,24 is a time step, and 3 is a characteristic quantity, namely, a longitude value, a latitude value and a water level value;
the number of neurons of the two GRU layers is 128 and 64 respectively, and the activation functions are tanh functions; the attention mechanism calculation formula is as follows:
h k′ ＝tanh(W c [C；h k ])；
wherein h is 1 ,h 2 ,…,h k Representing state values of hidden layers corresponding to an input sequence, scoring function s kj By dot product, alpha ki Attention weight of current input for hidden layer of history input, C is intermediate vector; h is a k′ And the final output current time hidden layer state value is represented.
2. The intelligent prediction method for navigation mark drift based on deep learning as claimed in claim 1, wherein preprocessing the sample data comprises:
for data repetition, finding a repeated record item and deleting the repeated record item;
for data deletion, linear interpolation is adopted to fill data;
for data anomalies, detecting outliers by using a box graph, and filling by using a data missing processing method after eliminating the outliers.
3. The intelligent prediction method for navigation mark drift based on deep learning according to claim 2, wherein the linear interpolation formula comprises:
wherein t is i And x i Respectively filling time point and filling value, x i+1 、x i-1 Respectively adjacent time point observations.
4. The intelligent prediction method for navigation mark drift based on deep learning as set forth in claim 1, wherein the data preprocessing further includes:
extracting navigation marks and water level data as a data set according to a preset time interval;
normalizing each column in the data set;
accordingly, constructing a time series based on the preprocessed sample data, comprising:
a time series is constructed based on the normalized dataset.
5. The intelligent prediction method of navigation mark drift based on deep learning as set forth in claim 1, wherein the loss function of the navigation mark drift prediction model uses average absolute error.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110648282.6A CN113393027B (en) | 2021-06-10 | 2021-06-10 | Navigation mark drift intelligent prediction method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110648282.6A CN113393027B (en) | 2021-06-10 | 2021-06-10 | Navigation mark drift intelligent prediction method based on deep learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113393027A CN113393027A (en) | 2021-09-14 |
| CN113393027B true CN113393027B (en) | 2023-11-07 |
Family
ID=77620220
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110648282.6A Active CN113393027B (en) | 2021-06-10 | 2021-06-10 | Navigation mark drift intelligent prediction method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113393027B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113932817B (en) * | 2021-12-15 | 2022-03-08 | 思创数码科技股份有限公司 | Navigation mark offset evaluation method and system |
| CN114580309B (en) * | 2022-05-09 | 2022-08-05 | 深圳市朗诚科技股份有限公司 | Drift speed prediction method, drift speed prediction device, calculation equipment and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104913892A (en) * | 2015-05-08 | 2015-09-16 | 京杭运河江苏省交通运输厅苏北航务管理处 | Beacon collision detection system based on navigation positioning system and method |
| KR101778724B1 (en) * | 2016-12-05 | 2017-09-14 | 동국대학교 산학협력단 | Device and method for reconition of road markings |
| CN111414442A (en) * | 2020-01-08 | 2020-07-14 | 重庆大学 | Navigation mark position checking method based on navigation channel topographic map and water level data |
| CN111488940A (en) * | 2020-04-15 | 2020-08-04 | 大连海事大学 | Intelligent navigation mark image classification method based on deep convolutional neural network |
| CN112288193A (en) * | 2020-11-23 | 2021-01-29 | 国家海洋信息中心 | Ocean station surface salinity prediction method based on GRU deep learning of attention mechanism |
| CN112364119A (en) * | 2020-12-01 | 2021-02-12 | 国家海洋信息中心 | Ocean buoy track prediction method based on LSTM coding and decoding model |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11256983B2 (en) * | 2017-07-27 | 2022-02-22 | Waymo Llc | Neural networks for vehicle trajectory planning |
| CN108416695B (en) * | 2018-02-24 | 2020-07-07 | 合肥工业大学 | Power load probability density prediction method, system and medium based on deep learning |
| CN112684701A (en) * | 2020-12-01 | 2021-04-20 | 哈尔滨工程大学 | Ship motion forecasting method based on long-time and short-time memory network and Gaussian process regression |
-
2021
- 2021-06-10 CN CN202110648282.6A patent/CN113393027B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104913892A (en) * | 2015-05-08 | 2015-09-16 | 京杭运河江苏省交通运输厅苏北航务管理处 | Beacon collision detection system based on navigation positioning system and method |
| KR101778724B1 (en) * | 2016-12-05 | 2017-09-14 | 동국대학교 산학협력단 | Device and method for reconition of road markings |
| CN111414442A (en) * | 2020-01-08 | 2020-07-14 | 重庆大学 | Navigation mark position checking method based on navigation channel topographic map and water level data |
| CN111488940A (en) * | 2020-04-15 | 2020-08-04 | 大连海事大学 | Intelligent navigation mark image classification method based on deep convolutional neural network |
| CN112288193A (en) * | 2020-11-23 | 2021-01-29 | 国家海洋信息中心 | Ocean station surface salinity prediction method based on GRU deep learning of attention mechanism |
| CN112364119A (en) * | 2020-12-01 | 2021-02-12 | 国家海洋信息中心 | Ocean buoy track prediction method based on LSTM coding and decoding model |
Non-Patent Citations (3)
| Title |
|---|
| Traffic Sign Recognition on Video Sequence Using Deep Neural Networks and Matching Algorithm;Ilya Belkin 等;2019 International Conference on Artificial Intelligence: Applications and Innovations;35-39 * |
| 基于LSTM的船舶航迹预测模型;权波 等;计算机科学(第S2期);136-141 * |
| 智能水位预测服务系统研究;潘明阳 等;大连海事大学学报;第46卷(第3期);31-37 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113393027A (en) | 2021-09-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Tang et al. | A model for vessel trajectory prediction based on long short-term memory neural network | |
| Feng et al. | Dynamic land use change simulation using cellular automata with spatially nonstationary transition rules | |
| CN113393027B (en) | Navigation mark drift intelligent prediction method based on deep learning | |
| CN109359557B (en) | SAR remote sensing image ship detection method based on transfer learning | |
| CN109214107A (en) | A kind of ship's navigation behavior on-line prediction method | |
| Üneş et al. | Prediction of millers ferry dam reservoir level in USA using artificial neural network | |
| CN112418521B (en) | Short-term marine fish school and fish quantity prediction method | |
| CN111695299B (en) | Mesoscale vortex trajectory prediction method | |
| CN110070144A (en) | A kind of lake water quality prediction technique and system | |
| CN110082717A (en) | A kind of underwater wireless sensor node positioning method | |
| CN115618251B (en) | Ship track prediction method and device, electronic equipment and storage medium | |
| CN108008099A (en) | A kind of pollution sources localization method | |
| CN103020485B (en) | Based on the short-term wind speed forecasting method of beta noise core ridge regression technology | |
| CN110866477A (en) | Track motion mode identification method and device | |
| CN114399073A (en) | Ocean surface temperature field prediction method based on deep learning | |
| Chen et al. | A cellular automata approach of urban sprawl simulation with Bayesian spatially-varying transformation rules | |
| Yang et al. | An approach to ship behavior prediction based on AIS and RNN optimization model | |
| CN117076893A (en) | Sound velocity distribution forecasting method based on long-term and short-term memory neural network | |
| Abozari et al. | Comparison performance of artificial neural network based method in estimation of electric conductivity in wet and dry periods: Case study of Gamasiab river, Iran | |
| Rochac et al. | Streamflow prediction using a hybrid methodology based on convolutional neural network and long short-term memory | |
| Xue et al. | Improved BSO for ship path planning with grounding alert using interpolated Delaunay triangulation | |
| Sharma et al. | Hydrologic simulation approach for El Niño Southern Oscillation (ENSO)-affected watershed with limited raingauge stations | |
| Li et al. | Prediction and Analysis of the Temperature of British Seas in the Next 50 Years Based on the Grey Model | |
| Li et al. | An approach for traffic pattern recognition integration of ship AIS data and port geospatial features | |
| Huang et al. | Ship trajectory anomaly detection based on multi-feature fusion |
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 |