CN111882925A - Shipping traffic flow prediction system based on information propagation diagram and recurrent neural network - Google Patents

Shipping traffic flow prediction system based on information propagation diagram and recurrent neural network Download PDF

Info

Publication number
CN111882925A
CN111882925A CN202010731978.0A CN202010731978A CN111882925A CN 111882925 A CN111882925 A CN 111882925A CN 202010731978 A CN202010731978 A CN 202010731978A CN 111882925 A CN111882925 A CN 111882925A
Authority
CN
China
Prior art keywords
channel
prediction
time
prediction unit
shipping
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202010731978.0A
Other languages
Chinese (zh)
Inventor
耿雄飞
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
China Waterborne Transport Research Institute
Original Assignee
China Waterborne Transport Research Institute
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by China Waterborne Transport Research Institute filed Critical China Waterborne Transport Research Institute
Priority to CN202010731978.0A priority Critical patent/CN111882925A/en
Publication of CN111882925A publication Critical patent/CN111882925A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G3/00—Traffic control systems for marine craft
    • 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
    • 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
    • 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
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/40—Business processes related to the transportation industry

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Strategic Management (AREA)
  • Human Resources & Organizations (AREA)
  • Economics (AREA)
  • Software Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Mathematical Physics (AREA)
  • Data Mining & Analysis (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • General Business, Economics & Management (AREA)
  • Biomedical Technology (AREA)
  • Tourism & Hospitality (AREA)
  • Artificial Intelligence (AREA)
  • Molecular Biology (AREA)
  • Marketing (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Game Theory and Decision Science (AREA)
  • Development Economics (AREA)
  • Ocean & Marine Engineering (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Primary Health Care (AREA)
  • Traffic Control Systems (AREA)

Abstract

The invention relates to the technical field of shipping traffic, in particular to a shipping traffic flow prediction system based on an information propagation diagram and a recurrent neural network, which comprises a controller, an AIS data module and a prediction module; the AIS data module is used for acquiring current real-time channel flow data and historical channel flow data and feeding back the current real-time channel flow data and the historical channel flow data to the controller; the prediction module comprises a channel network graph and a prediction unit, the channel network graph is a directed graph formed by detection points with correlation in the shipping network so as to determine the driving direction and the driving distance between the detection points, the prediction unit comprises a spatial correlation prediction unit and a time correlation prediction unit, and the spatial correlation prediction unit and the time correlation prediction unit are both connected with the controller so as to comprehensively predict future channel flow information, provide basis for the control of the waterway shipping and improve the waterway shipping efficiency.

Description

Shipping traffic flow prediction system based on information propagation diagram and recurrent neural network
Technical Field
The invention relates to the technical field of shipping traffic, in particular to a shipping traffic flow prediction system based on an information propagation diagram and a recurrent neural network.
Background
With the rapid development of economy, the traffic volume of waterway shipping is also rapidly increased, and the congestion phenomenon of the waterway is increasingly serious. Because the waterway channel mainly depends on the system of the river, cannot be randomly expanded and is a limited resource, the reasonable control of the waterway channel is particularly important for improving the shipping efficiency. Meanwhile, the traffic flow is a function changing along with time and space, and the short-time traffic flow shows strong dynamic property, nonlinearity, uncertainty, periodicity, non-stationarity and spatial correlation.
The purpose of shipping management and control and shipping optimal scheduling is to relieve or avoid traffic flow congestion in a future period of time and improve traffic efficiency. Therefore, the prediction of traffic flow for a period of time in the future is an important basis for guiding shipping management and control and optimizing scheduling.
The short-term shipping traffic flow prediction problem is mainly a problem of predicting the change of the network traffic flow of the channel in a short term by researching the historical data record of the network traffic condition of the channel and the current real-time data of the channel condition, and is a problem of predicting the change of the short-term dynamic traffic flow based on the analysis and mining of big data.
Since the AIS of the ship, the GPS on the ship, the weather and the abundant information of the hydrological data can be fully recorded and kept at the present stage, the method for predicting the shipping traffic flow based on the historical and real-time data becomes possible.
Disclosure of Invention
Technical problem to be solved
In order to solve the problems, the invention provides a shipping traffic flow prediction system based on an information propagation diagram and a recurrent neural network, so that the shipping traffic flow of a waterway is predicted, a basis is provided for the control of the waterway shipping, and the shipping efficiency of the waterway is improved.
(II) technical scheme
In order to achieve the purpose, the invention provides the following technical scheme: the shipping traffic flow prediction system based on the information propagation diagram and the recurrent neural network comprises a controller, an AIS data module and a prediction module; the AIS data module is used for acquiring current real-time channel flow data and historical channel flow data and feeding back the current real-time channel flow data and the historical channel flow data to the controller; the prediction module comprises a channel network graph and a prediction unit, the channel network graph is a directed graph formed by detection points with correlation in the shipping network so as to determine the driving direction and the driving distance between the detection points, the prediction unit comprises a spatial correlation prediction unit and a time correlation prediction unit, and the spatial correlation prediction unit and the time correlation prediction unit are both connected with the controller so as to comprehensively predict future channel flow information.
Preferably, the spatial correlation prediction unit predicts the steady-state probability of a transition from a single detection point to other detection points on the channel network graph, i.e. the probability of a transition from a single detection point to another detection point
Figure BDA0002603648260000021
Thus, the propagation convolution of the channel traffic is obtained:
Figure BDA0002603648260000022
and further obtaining a propagation convolution layer of the channel flow:
Figure BDA0002603648260000023
so as to predict the spatial correlation of the channel traffic.
Preferably, the temporal correlation prediction unit includes an LSTM prediction model and a GRU prediction model to make a state prediction of future channel traffic information using current and historical channel traffic information.
Preferably, the LSTM prediction model is trained by using a historical time sequence as an input and using a current time as an output, training a parameter matrix and a bias vector, and establishing a model of a recurrent neural network to perform time correlation prediction on channel traffic, and the calculation process is as follows:
i(t)＝(Wi[x(t)，h(t-1)]+bi)
Figure BDA0002603648260000031
f(t)＝(Wf[x(t)，h(t-1)]+bf)
Figure BDA0002603648260000032
o(t)＝(Wo[x(t)，h(t-1)]+bo)
Figure BDA0002603648260000033
where x (t) is the input time signal,
Figure BDA0002603648260000036
is multiplication of corresponding elements of the matrix, Wi，Wf，Wc，WoIs a parameter matrix of the model, bi，bf，bc，boIs a bias vector.
Preferably, the GRU prediction model is trained by using a historical time sequence as an input and using a current time as an output, so as to train a model parameter matrix and a bias vector, and establish a model of a recurrent neural network of a time sequence signal, so as to perform time correlation prediction on the channel traffic, and the calculation process is as follows:
r(t)＝(Wr[x(t)，h(t-1)]+br)
z(t)＝(Wz[x(t)，h(t-1)]+bz)
Figure BDA0002603648260000034
Figure BDA0002603648260000035
where x (t) is the input time signal,
Figure BDA0002603648260000041
is multiplication of corresponding elements of the matrix, Wr，Wz，WcIs a parameter matrix of the model, bi，bz，bcIs a bias vector.
Preferably, the comprehensive prediction execution flow of the spatial correlation prediction unit and the GRU prediction model is as follows:
establishing a channel network graph G ═ V, E, W, and obtaining a W matrix
Figure BDA0002603648260000042
Wherein DOIs the sum of out-of-degree weights for each vertex;
in each GRU unit, the input is XtAnd H(t-1)The propagation convolution is used to replace the parameter matrix and [ X ]t，H(t-1)]The product of (a) and (b), namely:
Figure BDA0002603648260000043
wherein theta is a parameter matrix and thetakIs the kth parameter of Θ;
further, using the GRU prediction model, one can obtain:
r(t)＝(Θr*[X(t)，H(t-1)]+br)
z(t)＝(Θz*[X(t)，H(t-1)]+bz)
Figure BDA0002603648260000044
Figure BDA0002603648260000045
(III) advantageous effects
The invention provides a shipping traffic flow prediction system based on an information propagation diagram and a recurrent neural network, which has the following beneficial effects: the AIS data module records current real-time channel flow data and historical channel flow data and feeds the current real-time channel flow data and the historical channel flow data back to the controller or further feeds the current real-time channel flow data and the historical channel flow data back to the space correlation prediction unit and the time correlation prediction unit, and the space correlation prediction unit and the time correlation prediction unit feed predicted information of the waterway shipping traffic flow back to the controller after calculating the feedback data so as to be consulted by related workers, provide basis for the control of waterway shipping and improve the waterway shipping efficiency.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention without limiting the invention in which:
FIG. 1 shows a block diagram of an embodiment of the present invention;
FIG. 2 illustrates a flow diagram of an LSTM predictive model of an embodiment of the invention;
FIG. 3 shows a flow diagram of a GRU prediction model of an embodiment of the present invention;
fig. 4 shows a flow diagram of a GRU recurrent neural network of an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1-4, an embodiment of the present invention provides a system for predicting shipping traffic flow based on an information propagation graph and a recurrent neural network, including a controller, an AIS data module and a prediction module; the AIS data module is used for acquiring current real-time channel flow data and historical channel flow data and feeding back the current real-time channel flow data and the historical channel flow data to the controller; the prediction module comprises a channel network graph and a prediction unit, the channel network graph is a directed graph formed by detection points with correlation in the shipping network so as to determine the driving direction and the driving distance between the detection points, the prediction unit comprises a spatial correlation prediction unit and a time correlation prediction unit, and the spatial correlation prediction unit and the time correlation prediction unit are both connected with the controller so as to comprehensively predict future channel flow information.
According to the scheme, the AIS data module obtains current real-time channel flow data and historical channel flow data and feeds the current real-time channel flow data and the historical channel flow data back to the controller, or further feeds the current real-time channel flow data and the historical channel flow data back to the space correlation prediction unit and the time correlation prediction unit, and after the space correlation prediction unit and the time correlation prediction unit calculate the feedback data, prediction information of the waterway shipping traffic flow is fed back to the controller to be consulted by related workers, a basis is provided for management and control of waterway shipping, and waterway shipping efficiency is improved.
Further, the spatial correlation prediction unit predicts the steady-state probability of a transition from a single detection point to other detection points on the channel network graph, i.e. the probability of the steady-state transition
Figure BDA0002603648260000061
Thus, the propagation convolution of the channel traffic is obtained:
Figure BDA0002603648260000062
and further obtaining a propagation convolution layer of the channel flow:
Figure BDA0002603648260000063
so as to predict the spatial correlation of the channel traffic.
Specifically, the channel network graph is G ═ V, E, W, where V represents traffic flow detection points of the channel, | V | ═ N, which indicates N traffic flow detection points in the transit network; e represents the direct edge-connecting relation among the traffic flow detection points, namely whether each detection point has a directly connected channel; (i, j) belongs to E, if a direct channel exists between the vertexes i and j, namely a closed channel section exists between the vertexes i and j; the navigation channel network graph is a directed graph, namely (i, j) represents the edges from i to j; w is formed as RN×NRepresenting the distance relationship among all the flow detection points in the channel network; w (i, j) represents the distance to be traveled from i to j along the route, and if there is no straight edge between i and j, W (i, j) is 0; given the flow at each vertex at time t as X(t)Then get
DO＝diag(W·1)
The sum of the degrees of each vertex is used as a diagonal matrix of the diagonal. Then due to
Figure BDA0002603648260000071
Therefore, it is not only easy to use
Figure BDA0002603648260000074
Is the state transition matrix of the steady state markov process.
The propagation probability of the information at the point v propagating to other points in the navigation network graph G can be characterized by random walk with the re-departure probability of alpha E [0,1 ]. The random walk with a re-starting probability α is as follows: assuming that the random walk is initiated by point v, is currently in u, and the position of the previous step is s: at this step, the probability of alpha is used to return to s from the point u; one-hop reachable neighbor that randomly walks from point u to point u with a probability of 1- α: and t ∈ N [ u ], and the probability of transition is determined by the transition probability between u and t.
Due to the fact that
Figure BDA0002603648260000075
Described is a state transition matrix for a steady state markov process. Therefore, after several steps of random walk, the steady-state probability of the information propagation process is:
Figure BDA0002603648260000072
based on the steady-state probability, a propagation convolution method of the traffic flow of the shipping network can be designed. Setting current flow information in network as X(t)∈RN×PLet the convolution filter function be fθThen the information propagation formula for traffic vectors propagating on the network can be described as:
Figure BDA0002603648260000073
where θ ∈ RKAre the parameters of the convolution filter and are,
Figure BDA0002603648260000085
is the state transition matrix of the information transfer process.
We define a convolution function, X f, of the flow propagation over the transit network GθThe result of (2) is still a traffic flow state vector of NxP dimension, and the P dimension convolution result is input to the activation function and mapped into Q dimension output according to the concept of the propagation convolution layer. The design method of the propagation convolution layer is as follows:
Figure BDA0002603648260000081
the spatial correlation prediction of the channel flow can be completed.
Further, the time correlation prediction unit comprises an LSTM prediction model and a GRU prediction model so as to use current and historical channel traffic information to perform state prediction of future channel traffic information.
Further, the LSTM prediction model takes a historical time sequence as an input, takes the current time as an output training, trains a parameter matrix and a bias vector, establishes a model of a recurrent neural network, and performs time correlation prediction on channel traffic, and the calculation process is as follows:
i(t)＝(Wi[x(t)，h(t-1)]+bi)
Figure BDA0002603648260000082
f(t)＝(Wf[x(t)，h(t-1)]+bf)
Figure BDA0002603648260000083
o(t)＝(Wo[x(t)，h(t-1)]+bo)
Figure BDA0002603648260000084
where x (t) is the input time signal,
Figure BDA0002603648260000091
is multiplication of corresponding elements of the matrix, Wi，Wf，Wc，WoIs a parameter matrix of the model, bi，bf，bc，boIs a bias vector.
Furthermore, the GRU prediction model takes the historical time sequence as input, takes the current time as output training, trains a model parameter matrix and a bias vector, establishes a model of a recurrent neural network of a time sequence signal, and performs time correlation prediction on the channel flow, and the calculation process is as follows:
r(t)＝(Wr[x(t)，h(t-1)]+br)
z(t)＝(Wz[x(t)，h(t-1)]+bz)
Figure BDA0002603648260000092
Figure BDA0002603648260000093
where x (t) is the input time signal,
Figure BDA0002603648260000094
is multiplication of corresponding elements of the matrix, Wr，Wz，WcIs a parameter matrix of the model, bi，bz，bcIs a bias vector.
Further, the comprehensive prediction execution flow of the spatial correlation prediction unit and the GRU prediction model is as follows:
establishing a channel network graph G ═ V, E, W, and obtaining a W matrix
Figure BDA0002603648260000095
Wherein DOIs the sum of out-of-degree weights for each vertex;
in each GRU unit, the input is XtAnd H(t-1)Again using propagation convolution instead of parametersMatrix and [ X ]t，H(t-1)]The product of (a) and (b), namely:
Figure BDA0002603648260000101
wherein theta is a parameter matrix and thetakIs the kth parameter of Θ;
further, using the GRU prediction model, one can obtain:
r(t)＝(Θr*[x(t)，H(t-1)]+br)
z(t)＝(Θz*[x(t)，H(t-1)]+bz)
Figure BDA0002603648260000102
Figure BDA0002603648260000103
specifically, the propagation of traffic flow has both spatial correlation and temporal correlation. The information propagation convolution network model on the water route shipping network diagram is mainly used for capturing the spatial correlation relationship. And the recurrent neural network can better capture the correlation in time.
By combining the respective characteristics of the two, a traffic flow prediction model combining propagation convolution and a recurrent neural network can be designed. In the recurrent neural network, the process of calculating the current input and output at each step can be understood as a process of carrying out information propagation convolution on the basis of the edge weight relation of the graph in the propagation network, so that the propagation convolution calculation can be used for replacing the calculation process of multiplying the weight matrix by [ x (t), h (t-1) ].
Furthermore, because the propagation convolution process extracts the spatial correlation of the traffic flow well, and the recurrent neural network can extract the time correlation of the traffic flow well, the propagation convolution is introduced into each GRU unit to obtain the GRU recurrent neural network (DCRNN), so that a prediction result can be obtained well. In the prediction process, the traffic flow information and the current traffic flow information are used as input to predict the traffic flow information at the next moment, the predicted result is used as input to predict the traffic flow at the later moment again, and the model parameters of each GRU unit combined with propagation convolution are trained by using back propagation, so that the number of steps required to be predicted can be controlled by a user.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments according to the present application. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.
The relative arrangement of the components and steps, the numerical expressions, and numerical values set forth in these embodiments do not limit the scope of the present application unless specifically stated otherwise. Meanwhile, it should be understood that the sizes of the respective portions shown in the drawings are not drawn in an actual proportional relationship for the convenience of description. Techniques, methods, and apparatus known to those of ordinary skill in the relevant art may not be discussed in detail but are intended to be part of the specification where appropriate. In all examples shown and discussed herein, any particular value should be construed as merely illustrative, and not limiting. Thus, other examples of the exemplary embodiments may have different values. It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus, once an item is defined in one figure, further discussion thereof is not required in subsequent figures.
In the description of the present application, it is to be understood that the orientation or positional relationship indicated by the directional terms such as "front, rear, upper, lower, left, right", "lateral, vertical, horizontal" and "top, bottom", etc., are generally based on the orientation or positional relationship shown in the drawings, and are used for convenience of description and simplicity of description only, and in the case of not making a reverse description, these directional terms do not indicate and imply that the device or element being referred to must have a particular orientation or be constructed and operated in a particular orientation, and therefore, should not be considered as limiting the scope of the present application; the terms "inner and outer" refer to the inner and outer relative to the profile of the respective component itself.
Spatially relative terms, such as "above … …," "above … …," "above … …," "above," and the like, may be used herein for ease of description to describe one device or feature's spatial relationship to another device or feature as illustrated in the figures. It will be understood that the spatially relative terms are intended to encompass different orientations of the device in use or operation in addition to the orientation depicted in the figures. For example, if a device in the figures is turned over, devices described as "above" or "on" other devices or configurations would then be oriented "below" or "under" the other devices or configurations. Thus, the exemplary term "above … …" can include both an orientation of "above … …" and "below … …". The device may be otherwise variously oriented (rotated 90 degrees or at other orientations) and the spatially relative descriptors used herein interpreted accordingly.
It should be noted that the terms "first", "second", and the like are used to define the components, and are only used for convenience of distinguishing the corresponding components, and the terms have no special meanings unless otherwise stated, and therefore, the scope of protection of the present application is not to be construed as being limited.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.

Claims (6)

1. The shipping traffic flow prediction system based on the information propagation diagram and the recurrent neural network comprises a controller, and is characterized by further comprising an AIS data module and a prediction module;
the AIS data module is used for acquiring current real-time channel flow data and historical channel flow data and feeding back the current real-time channel flow data and the historical channel flow data to the controller;
the prediction module comprises a channel network graph and a prediction unit, wherein the channel network graph is a directed graph formed by detection points with correlation in a shipping network so as to determine the driving direction and the driving distance between the detection points, the prediction unit comprises a spatial correlation prediction unit and a time correlation prediction unit, and the spatial correlation prediction unit and the time correlation prediction unit are both connected with the controller so as to comprehensively predict future channel flow information.
2. The system for predicting shipping traffic flow based on information propagation graph and recurrent neural network as claimed in claim 1, wherein said spatial correlation prediction unit predicts the steady-state probability of transition from a single detection point to other detection points on the navigation channel network graph by using the steady-state probability
Figure FDA0002603648250000011
Thus, the propagation convolution of the channel traffic is obtained:
Figure FDA0002603648250000012
and further obtaining a propagation convolution layer of the channel flow:
Figure FDA0002603648250000013
so as to predict the spatial correlation of the channel traffic.
3. The system of claim 2, wherein the temporal correlation prediction unit comprises an LSTM prediction model and a GRU prediction model to make a state prediction of future channel traffic information using current and historical channel traffic information.
4. The system of claim 3, wherein the LSTM prediction model is trained by using a historical time sequence as an input and using a current time as an output to train a parameter matrix and a bias vector, and a model of a recurrent neural network is established to perform time-dependent prediction on the channel traffic, and the calculation process is as follows:
i(t)＝(Wi[x(t)，h(t-1)]+bi)
Figure FDA0002603648250000021
f(t)＝(Wf[x(t)，h(t-1)]+bf)
Figure FDA0002603648250000022
o(t)＝(Wo[x(T)，h(t-1)]+bo)
Figure FDA0002603648250000023
where x (t) is the input time signal,
Figure FDA0002603648250000024
is multiplication of corresponding elements of the matrix, Wi,Wf,Wc,WoIs a parameter matrix of the model, bi,bf,bc,boIs a bias vector.
5. The system of claim 3, wherein the GRU prediction model is trained with historical time series as input and current time as output to train model parameter matrix and offset vector, and establishes model of recurrent neural network of time series signal to perform time correlation prediction of channel traffic, and the calculation procedure is as follows:
r(t)＝(Wr[x(t)，h(t-1)]+br)
z(t)＝(Wz[x(t)，h(t-1)]+bz)
Figure FDA0002603648250000031
Figure FDA0002603648250000032
where x (t) is the input time signal,
Figure FDA0002603648250000033
is multiplication of corresponding elements of the matrix, Wr,Wz,WcIs a parameter matrix of the model, bi,bz,bcIs a bias vector.
6. The system for predicting shipping traffic flow based on information propagation diagram and recurrent neural network as claimed in claim 5, wherein said integrated prediction execution flow of spatial correlation prediction unit and GRU prediction model is:
establishing a channel network graph G ═ V, E, W, and obtaining a W matrix
Figure FDA0002603648250000034
Wherein DoIs the sum of out-of-degree weights for each vertex;
in each GRU unit, the input is XtAnd H(t-1)The propagation convolution is used to replace the parameter matrix and [ X ]t，H(t-1)]The product of (a) and (b), namely:
Figure FDA0002603648250000035
wherein theta is a parameter matrix and thetakIs the kth parameter of Θ;
further, using the GRU prediction model, one can obtain:
r(t)＝(Θr*[X(t)，H(t-1)]+br)
z(t)＝(Θz*[X(t)，H(t-1)]+bz)
Figure FDA0002603648250000036
Figure FDA0002603648250000041
。
CN202010731978.0A 2020-07-27 2020-07-27 Shipping traffic flow prediction system based on information propagation diagram and recurrent neural network Pending CN111882925A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010731978.0A CN111882925A (en) 2020-07-27 2020-07-27 Shipping traffic flow prediction system based on information propagation diagram and recurrent neural network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010731978.0A CN111882925A (en) 2020-07-27 2020-07-27 Shipping traffic flow prediction system based on information propagation diagram and recurrent neural network

Publications (1)

Publication Number Publication Date
CN111882925A true CN111882925A (en) 2020-11-03

Family

ID=73200575

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010731978.0A Pending CN111882925A (en) 2020-07-27 2020-07-27 Shipping traffic flow prediction system based on information propagation diagram and recurrent neural network

Country Status (1)

Country Link
CN (1) CN111882925A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113658452A (en) * 2021-07-31 2021-11-16 海南超船电子商务有限公司 Course distance measuring and calculating method and system

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104091470A (en) * 2014-07-15 2014-10-08 南京大学 Channel traffic information prediction method and application based on multidata fusion
KR20180077571A (en) * 2016-12-29 2018-07-09 대우조선해양 주식회사 Apparatus and method for analyzing operation status of ship
CN110009937A (en) * 2019-03-21 2019-07-12 武汉理工大学 A kind of ship collision risk analysis method based on AIS data
CN110491129A (en) * 2019-09-24 2019-11-22 重庆城市管理职业学院 The traffic flow forecasting method of divergent convolution Recognition with Recurrent Neural Network based on space-time diagram
CN110991713A (en) * 2019-11-21 2020-04-10 杭州电子科技大学 Irregular area flow prediction method based on multi-graph convolution sum GRU
CN111241466A (en) * 2020-01-15 2020-06-05 上海海事大学 Ship flow prediction method based on deep learning
CN111325976A (en) * 2020-02-24 2020-06-23 交通运输部水运科学研究所 Short-term traffic flow prediction method and system
CN111339229A (en) * 2020-02-24 2020-06-26 交通运输部水运科学研究所 Ship autonomous navigation aid decision-making system

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104091470A (en) * 2014-07-15 2014-10-08 南京大学 Channel traffic information prediction method and application based on multidata fusion
KR20180077571A (en) * 2016-12-29 2018-07-09 대우조선해양 주식회사 Apparatus and method for analyzing operation status of ship
CN110009937A (en) * 2019-03-21 2019-07-12 武汉理工大学 A kind of ship collision risk analysis method based on AIS data
CN110491129A (en) * 2019-09-24 2019-11-22 重庆城市管理职业学院 The traffic flow forecasting method of divergent convolution Recognition with Recurrent Neural Network based on space-time diagram
CN110991713A (en) * 2019-11-21 2020-04-10 杭州电子科技大学 Irregular area flow prediction method based on multi-graph convolution sum GRU
CN111241466A (en) * 2020-01-15 2020-06-05 上海海事大学 Ship flow prediction method based on deep learning
CN111325976A (en) * 2020-02-24 2020-06-23 交通运输部水运科学研究所 Short-term traffic flow prediction method and system
CN111339229A (en) * 2020-02-24 2020-06-26 交通运输部水运科学研究所 Ship autonomous navigation aid decision-making system

Non-Patent Citations (6)

* Cited by examiner, † Cited by third party
Title
DAVIDE ANDREOLETTI 等: "Network Traffic Prediction based on Diffusion Convolutional Recurrent Neural Networks", 《2019 IEEE INFOCOM WKSHPS:NI 2019:NETWORK INTELLIGENCE:MACHINE LEARNING FOR NETWORKING》 *
方炯焜 等: "结合GloVe和GRU的文本分类模型", 《计算机工程与应用》 *
牟军敏 等: "《内河水网船舶通过能力研究》", 30 April 2019, 武汉理工大学出版社 *
王根生 等: "多元特征融合的GRU神经网络文本情感分类模型", 《小型微型计算机系统》 *
蔡学龙: "多航道交汇水域船舶交通流的建模与仿真", 《中国优秀博硕士学位论文全文数据库(硕士)·工程科技Ⅱ辑》 *
谢培元 等: "一种基于堆叠门控循环单元的暂态稳定智能评估系统", 《电气自动化》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113658452A (en) * 2021-07-31 2021-11-16 海南超船电子商务有限公司 Course distance measuring and calculating method and system
CN113658452B (en) * 2021-07-31 2022-11-22 海南超船电子商务有限公司 Course distance measuring and calculating method and system

Similar Documents

Publication Publication Date Title
CN107230351B (en) A kind of Short-time Traffic Flow Forecasting Methods based on deep learning
CN104134351B (en) A kind of Short-time Traffic Flow Forecasting Methods
CN107180530B (en) A kind of road network trend prediction method based on depth space-time convolution loop network
CN111860951B (en) Rail transit passenger flow prediction method based on dynamic hypergraph convolutional network
CN102110365B (en) Road condition prediction method and road condition prediction system based on space-time relationship
CN112489426B (en) Urban traffic flow space-time prediction scheme based on graph convolution neural network
CN109636049A (en) A kind of congestion index prediction technique of combination road network topology structure and semantic association
Osorio et al. A surrogate model for traffic optimization of congested networks: an analytic queueing network approach
CN103314543A (en) Forecasting of dynamic environmental parameters to optimize operation of a wireless communication system
CN106017496A (en) Real-time navigation method based on road condition
CN106879032A (en) A kind of outdoor seamless and system based on pattern classification
CN113673749B (en) Vehicle-pedestrian map track prediction method and device
Wu Spatiotemporal dynamic forecasting and analysis of regional traffic flow in urban road networks using deep learning convolutional neural network
Liu et al. A method for short-term traffic flow forecasting based on GCN-LSTM
Moghimi et al. Short-term prediction of signal cycle on an arterial with actuated-uncoordinated control using sparse time series models
CN111882925A (en) Shipping traffic flow prediction system based on information propagation diagram and recurrent neural network
CN113344470B (en) Intelligent power supply system management method based on block chain
CN115762147B (en) Traffic flow prediction method based on self-adaptive graph meaning neural network
Li et al. A sequence and network embedding method for bus arrival time prediction using GPS trajectory data only
Panovski et al. Real-time public transportation prediction with machine learning algorithms
CN111524358A (en) Regional radiation sexual communication flow prediction method
CN113903171B (en) Vehicle crowd sensing node optimization method based on spatial-temporal characteristics of highway network
Bouyahia et al. Real-time traffic data smoothing from GPS sparse measures using fuzzy switching linear models
CN112101677B (en) Public transport travel path planning method, device, equipment and storage medium
Li et al. MGC-GAN: Multi-graph convolutional generative adversarial networks for accurate citywide traffic flow prediction

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
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20201103