CN114582131B - Monitoring method and system based on ramp intelligent flow control algorithm - Google Patents
Monitoring method and system based on ramp intelligent flow control algorithm Download PDFInfo
- Publication number
- CN114582131B CN114582131B CN202210263636.XA CN202210263636A CN114582131B CN 114582131 B CN114582131 B CN 114582131B CN 202210263636 A CN202210263636 A CN 202210263636A CN 114582131 B CN114582131 B CN 114582131B
- Authority
- CN
- China
- Prior art keywords
- ramp
- flow
- flow control
- data
- expressway
- 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
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 60
- 238000000034 method Methods 0.000 title claims abstract description 35
- 238000012544 monitoring process Methods 0.000 title claims abstract description 27
- 239000002245 particle Substances 0.000 claims abstract description 21
- 238000012549 training Methods 0.000 claims abstract description 8
- 238000010586 diagram Methods 0.000 claims description 33
- 230000007246 mechanism Effects 0.000 claims description 24
- 238000007781 pre-processing Methods 0.000 claims description 24
- 230000015654 memory Effects 0.000 claims description 23
- 238000004364 calculation method Methods 0.000 claims description 16
- 230000002159 abnormal effect Effects 0.000 claims description 6
- 230000005540 biological transmission Effects 0.000 claims description 4
- 238000012546 transfer Methods 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 2
- 230000006870 function Effects 0.000 description 7
- 239000011159 matrix material Substances 0.000 description 7
- 230000008569 process Effects 0.000 description 7
- 238000013527 convolutional neural network Methods 0.000 description 5
- 238000013528 artificial neural network Methods 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 230000001105 regulatory effect Effects 0.000 description 4
- 230000004913 activation Effects 0.000 description 3
- 230000006399 behavior Effects 0.000 description 3
- 230000001276 controlling effect Effects 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 230000000306 recurrent effect Effects 0.000 description 3
- 238000005096 rolling process Methods 0.000 description 3
- 230000002123 temporal effect Effects 0.000 description 3
- 238000000586 desensitisation Methods 0.000 description 2
- 238000004880 explosion Methods 0.000 description 2
- 230000007774 longterm Effects 0.000 description 2
- 230000006386 memory function Effects 0.000 description 2
- 230000006403 short-term memory Effects 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008033 biological extinction Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000008034 disappearance Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000001939 inductive effect Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000000547 structure data Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000011144 upstream manufacturing Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- 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
- G06N3/044—Recurrent networks, e.g. Hopfield 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
- G06N3/04—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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/048—Activation functions
-
- 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
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0125—Traffic data processing
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/065—Traffic control systems for road vehicles by counting the vehicles in a section of the road or in a parking area, i.e. comparing incoming count with outgoing count
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention provides a monitoring method and a monitoring system based on an intelligent ramp flow control algorithm, which are characterized in that traffic data are firstly obtained and preprocessed, then an ST-GCN model is built according to the preprocessed data by adopting a flow prediction algorithm and iterative training is carried out to obtain an optimal ST-GCN model so as to predict the main road section flow of an expressway, the influence degree of the inlet flow of each ramp on the main road section flow of the expressway is calculated, then the overall expected flow control is calculated according to the predicted main road section flow of the expressway by combining with an expressway ramp toll gate channel number regulation algorithm, and then the optimal flow control scheme of each ramp in a road network is calculated by adopting a particle swarm algorithm according to the calculated overall expected flow control and the influence degree of the inlet flow of each ramp on the main road section flow, and finally the optimal flow control scheme is displayed, so that the traffic efficiency of the expressway is effectively improved. The invention improves the prediction precision, is convenient to use, only needs to be connected with the existing information board equipment on the expressway, does not need new equipment, and has high accuracy.
Description
Technical Field
The invention relates to the technical field of intelligent highways, in particular to a monitoring method and a system based on a ramp intelligent flow control algorithm.
Background
The intelligent highway cloud platform mainly provides an informatization and intelligent management means for the safe and smooth of the expressway in the provincial area and provides an informatization support for the emergency treatment of the expressway; the system can comprehensively control the road network state, improve the road network management capability, improve the public information service level and finally realize the 'knowing, measurable, controllable and serviceable' of the expressway management, so that the road network operation is safer and more efficient, the public travel is more convenient and comfortable, the traffic management is more scientific and intelligent, and the intelligent road is more green and economical.
The intelligent ramp flow control is an important application in the technical field of intelligent roads, and the current flow control process of the ramp of the highway is still low in traffic and long in congestion time.
Disclosure of Invention
The invention provides a monitoring method based on a ramp intelligent flow control algorithm, which aims to solve the problems of lower traffic volume, long congestion time and the like in the existing ramp intelligent flow control, dynamically predicts the traffic flow in a road network and calculates the influence degree of the inlet flow of each ramp on the flow of a main road section by establishing a space-time diagram convolution network model, namely an ST-GCN model, and calculates by adopting a highway ramp toll gate channel number regulation algorithm and a particle swarm algorithm to obtain a multi-toll gate ramp coordinated optimal flow control scheme, thereby inducing the behavior of a driver and a passenger and improving the traffic efficiency of a highway. The invention also relates to a monitoring system based on the ramp intelligent flow control algorithm.
The technical scheme of the invention is as follows:
the monitoring method based on the ramp intelligent flow control algorithm is characterized by comprising the following steps of:
and (3) data acquisition and preprocessing: acquiring traffic data and preprocessing the acquired traffic data;
model establishment and calculation steps: according to the preprocessed data, a space-time diagram convolution network and a long and short time memory network are utilized, a space-time diagram convolution network model is established by combining an attention mechanism, iterative training is carried out on the established space-time diagram convolution network model, an optimal space-time diagram convolution network model is obtained, expressway main road section flow is predicted according to the optimal space-time diagram convolution network model, and the influence degree of each ramp inlet flow on the main road section flow is calculated;
and a flow calculating step: calculating overall expected control flow by adopting a highway ramp toll station channel number regulation algorithm according to the predicted main road section flow of the highway;
scheme calculation step: according to the calculated overall expected flow control and the influence degree of the inlet flow of each ramp on the flow of the main road section, calculating to obtain the actual flow control of each ramp by adopting a particle swarm algorithm, and according to the actual flow control of each ramp, calculating to obtain the optimal flow control scheme of each ramp in the road network by combining the manual toll lane traffic capacity and the electronic toll lane traffic capacity of each toll station;
information display step: and displaying the optimal flow control scheme through an information board arranged on the expressway.
Preferably, in the model building and calculating step, whether the real-time ramp flow control is needed or the ramp flow control is needed in advance is judged according to the real-time expressway main road section flow in the traffic data and the predicted expressway main road section flow and the congestion judging condition, if the congestion condition occurs, the real-time ramp flow control is carried out, and the flow calculating step is entered.
Preferably, in the model building and calculating step, the space-time diagram convolution network model consists of a diagram convolution network, a long-term and short-term memory network and an attention mechanism;
firstly, using preprocessed historical time series data as input data, capturing a topological structure of a highway network by using a graph convolution network to obtain space correlation characteristics, inputting the obtained time series with the space correlation characteristics into a long-period memory network, obtaining dynamic changes through information transfer among units to obtain the time characteristics, and calculating the influence of the input data on current prediction through a full connection layer and an attention mechanism to obtain the main road section flow of the highway.
Preferably, in the data acquisition and preprocessing step, the preprocessing includes deleting abnormal data and repeated data, filling the missing data, and removing noise and normalizing the data.
Preferably, in the data acquisition and preprocessing step, the traffic data includes real-time traffic data and historical traffic data.
A monitoring system based on ramp intelligent flow control algorithm is characterized by comprising a data acquisition and preprocessing module, a model building and calculating module, a flow calculating module, a scheme calculating module and an information display module,
and the data acquisition and preprocessing module is used for: acquiring traffic data and preprocessing the acquired traffic data;
and a model building and calculating module: according to the preprocessed data, a space-time diagram convolution network and a long and short time memory network are utilized, a space-time diagram convolution network model is established by combining an attention mechanism, iterative training is carried out on the established space-time diagram convolution network model, an optimal space-time diagram convolution network model is obtained, expressway main road section flow is predicted according to the optimal space-time diagram convolution network model, and the influence degree of each ramp inlet flow on the main road section flow is calculated;
the flow calculation module is used for: calculating overall expected control flow by adopting a highway ramp toll station channel number regulation algorithm according to the predicted main road section flow of the highway;
the scheme calculation module: according to the calculated overall expected flow control and the influence degree of the inlet flow of each ramp on the flow of the main road section, calculating to obtain the actual flow control of each ramp by adopting a particle swarm algorithm, and according to the actual flow control of each ramp, calculating to obtain the optimal flow control scheme of each ramp in the road network by combining the manual toll lane traffic capacity and the electronic toll lane traffic capacity of each toll station;
information display module: and displaying the optimal flow control scheme through an information board arranged on the expressway.
Preferably, the model building and calculating module further judges whether to perform real-time ramp flow control or to perform ramp flow control in advance according to the traffic data real-time expressway main road section flow and the predicted expressway main road section flow and the congestion judging conditions, if the congestion condition occurs, the real-time ramp flow control is performed, and if the congestion condition does not occur, the current opening and closing state of the ramp is maintained.
Preferably, the space-time diagram convolution network model consists of a diagram convolution network, a long-period memory network and an attention mechanism;
the model building and calculating module firstly uses the preprocessed historical time series data as input data, captures the topological structure of the expressway network by using a graph convolution network to obtain space correlation characteristics, then inputs the obtained time series with the space correlation characteristics into a long-period memory network, obtains dynamic changes through information transmission among units to obtain time characteristics, and then calculates the influence of the input data on current prediction through a full connection layer and an attention mechanism to obtain the main road section flow of the expressway.
Preferably, the preprocessing includes deleting abnormal data and repeated data, filling the missing data, and removing noise and normalizing the data.
Preferably, the traffic data includes real-time traffic data and historical traffic data.
The beneficial effects of the invention are as follows:
the invention provides a monitoring method based on an intelligent ramp flow control algorithm, which is characterized in that a space-time graph convolutional neural network model, namely an ST-GCN model, is built to predict future traffic flow, and is combined with two parts of data to comprehensively judge whether ramp flow control is needed or not. The particle swarm algorithm with self-organization, evolution and memory functions is also adopted to calculate and obtain the multi-toll station ramp coordination optimal flow control scheme, and the algorithm has fewer parameters to be regulated, simple operation and high convergence speed. When the invention is realized, only the existing information board equipment on the expressway is needed to be docked, no new equipment is needed, and the invention is convenient to use and high in accuracy. After the method is started, the traffic of the main line of the expressway can be improved by 30%, and the congestion time of the main line can be reduced by 30%.
The invention also relates to a monitoring system based on the ramp intelligent flow control algorithm, which corresponds to the monitoring method based on the ramp intelligent flow control algorithm, and can be understood as a system for realizing the monitoring method based on the ramp intelligent flow control algorithm.
Drawings
FIG. 1 is a flow chart of a monitoring method based on a ramp intelligent flow control algorithm.
FIG. 2 is a schematic diagram of the overall framework of the ST-GCN model of the present invention.
Fig. 3 is a flow chart of the particle swarm algorithm of the present invention.
Detailed Description
The present invention will be described below with reference to the accompanying drawings.
The invention relates to a monitoring method based on a ramp intelligent flow control algorithm, which is shown in a flow chart of fig. 1 and sequentially comprises the following steps:
and (3) data acquisition and preprocessing: firstly, acquiring real-time and historical traffic data, preprocessing the acquired traffic data, removing abnormal and repeated data in the traffic data, filling the missing data, removing noise, normalizing the data and the like, and then performing desensitization processing on the preprocessed traffic data.
Model establishment and calculation steps: according to the traffic data after pretreatment and desensitization, a graph convolution network and a long-short-time memory network are utilized, a flow prediction algorithm is adopted to carry out space correlation and time correlation modeling by combining an attention mechanism, a space-time graph convolution network model-an ST-GCN model is established, the established ST-GCN model is trained to obtain an optimal ST-GCN model, the main road section flow of the expressway is predicted according to the optimal ST-GCN model, and the influence degree of the inlet flow of each ramp on the main road section flow is calculated; the traffic data is time-series data, and has time correlation, namely, the traffic data of a certain time period is related to the traffic data of a plurality of previous time periods, and the closer to the current time period, the higher the data correlation. In order to analyze the time correlation of traffic data, a deep learning correlation algorithm is adopted to explore the correlation between the current moment and the lag moment. Considering that a single road section can not well show the trend of the change of the running condition of the vehicle and the relevance among road sections in the whole expressway network, the traffic condition of a certain road section is easy to be influenced by surrounding road sections in terms of traffic flow, for example, the speed of a downstream road section is likely to be slowed down when the upstream road section is blocked, and the spatial relevance degree, namely the spatial relevance, of traffic data among road sections can be found through analysis of the traffic spatial dependence in the road network.
Specifically, a flow prediction algorithm is adopted to establish an ST-GCN model for capturing the spatial correlation and the time correlation of traffic data and realizing traffic speed prediction based on expressways, and the ST-GCN model consists of a graph rolling network (Graph Convolutional Network, GCN), a long short-Term Memory (LSTM) network and an attention mechanism. As shown in fig. 2, an adjacency matrix constructed according to a network, a functional similarity matrix constructed according to POIs and real-time and historical traffic data related to an expressway network are used as input data, a GCN model is used for capturing the topological structure of the expressway network to obtain the spatial characteristics of the traffic data, the obtained traffic data with the spatial characteristics is input into an LSTM network model, dynamic changes are obtained through information transfer among units to obtain the temporal characteristics of the traffic data, and finally the influence of the input data on current prediction is calculated through a full connection layer and an attention mechanism to obtain a prediction result, namely the main road section flow of the expressway.
Acquiring complex spatial correlations is a key issue for traffic prediction. A conventional Convolutional Neural Network (CNN) can obtain local spatial features, but it can only be used in euclidean space, such as images, regular grids, etc. However, the highway network is in the form of a graph, not a two-dimensional grid, which means that the CNN model cannot reflect the complex topology of the highway network, and thus cannot accurately capture spatial correlation. In recent years, attention has been paid to the popularization of CNNs to graph rolling networks (GCNs) capable of processing binary graph structure data. The GCN model has been successfully applied in many applications including document classification, unsupervised learning, and image classification. The GCN model constructs a filter in the Fourier domain, the filter acts on nodes of the graph and first-order neighbors thereof, the spatial features between the nodes are captured, and then the GCN model is built by superposing a plurality of evolution layers.
The GCN model can obtain the topological relation between the central road and its peripheral roads, the topological structure of the encoded road network and the attributes on the roads, and then obtain the spatial correlation, such as the spatial feature information of intersections, adjacent road segments, distant road segments, and the like. The spatial features are thus learned from traffic data using the GCN model. A two-layer GCN model is represented by the following formula:
in the above-mentioned method, the step of,representing an adjacency matrix, X representing a feature matrix, W 0 And W is 1 Representing the weight matrix in the first and second layers, σ, relu representing the activation function.
Wherein,,representing a preprocessing step->Is a matrix with self-connecting structure, +.>Is a degree matrix->
Acquiring temporal correlation is another key issue for traffic prediction. Currently, the most widely used neural network model for processing sequence data is the current Recurrent Neural Network (RNN). However, conventional recurrent neural networks have limitations for long-term prediction due to defects such as gradient extinction and gradient explosion. LSTM and GRU models are variants of recurrent neural networks that have proven to solve the above problems. The basic principles of LSTM and GRU are about the same, they use gating mechanisms to memorize as much long-term information as possible and are equally effective for various tasks, while LSTM adds a gating mechanism to control the transfer of information and the updating of states and memory cells compared to the GRU model. Thus, the LSTM model is selected to obtain a time dependence, i.e., a temporal correlation, from the traffic data. The LSTM neural network has a memory cell and a state cell, which are constantly updated by a gating mechanism. The gating mechanisms are respectively an input gate, a forget gate and an output gate, and the three gates are mutually independent and respectively process the input, forget and output processes of the time sequence characteristic information. In addition, through the combined action of the gating mechanism and the memory unit, the information transmission can be selectively controlled, so that the problems of gradient disappearance and gradient explosion can be effectively solved, and the method has better processing and predicting capabilities for longer sequence data. The training process of the LSTM model is as follows:
selectively forgetting the information c stored in the memory unit i-1 Selecting a sigma (sigmoid) function as an activation function of a forgetting gate by storing information c in a memory cell i-1 Forgetting door f i Multiplication will leave a part of the information forgotten, thus requiring input gate i according to the present i Updating information in the memory cell. Similar to a forgetting gate, an input gate i i Also, important characteristic information is selected to update the memory cell c i . The output gate also uses sigma function as the activation function, responsible for memorizing the cell information c i And finally, connecting the output result of the LSTM to the full connection layer to obtain a prediction result.
It should be noted that the importance of each short subsequence feature of the long sequence is different, and the ability to give LSTM more important features of interest may better enable prediction of short traffic speed. Therefore, the LSTM model extracts the salient features of the short sequence by using an attention mechanism, takes the importance of traffic speeds at different times to the final output result into consideration, and inputs the result into the full-connection layer after calculating the attention coefficient and multiplying the result by the matrix to obtain the prediction result.
In order to capture the spatial and time dependence of traffic data at the same time, a space-time diagram convolution network model (namely ST-GCN model) based on a Graph Convolution Network (GCN) and a long-short-term memory (LSTM) network is provided, and the specific calculation process is as follows:
u t ＝σ(W u [f(A，X t )，h t-1 ]+b u ) (2)
r t ＝σ(W r [f(A，X t )，h t-1 ]+b r ) (3)
c t ＝tanh(W c [f(A，X t )，(r t *h t-1 )]+b c ) (4)
h t ＝u t *h t-1 +(1-u t )*c t (5)
in the above, u t 、r t Is an update gate and a reset gate at time t, h t Represents the vehicle speed output at time t, h t-1 Represents the vehicle speed output at time t-1, f (A, X t ) The graph convolution process is represented, defined as inequalities 1, w and b represent weights and deviations during training, and tanh is a hyperbolic tangent curve function.
In the intelligent ramp flow control, on one hand, an ST-GCN model is used for predicting the flow of a main road section of an expressway; on the other hand, the influence degree (proportion) of the inlet flow of each ramp on the flow of the main road section when the main line congestion is relieved is calculated by using the model, and the key ramp is identified.
And a flow calculating step: calculating overall expected control flow by adopting a highway ramp toll station channel number regulation algorithm according to the predicted main road section flow of the highway; preferably, the intelligent control stream data update period is less than 2 minutes.
The method is characterized in that the total expected control flow is calculated by combining a feedback control-based expressway ramp toll gate channel number adjustment algorithm (the feedback control-based toll gate channel number adjustment is used as a closed-loop control method, the main line traffic state of the expressway is required to be periodically detected, the difference value between the main line traffic state and the expected traffic state is used as a control input, constraints such as queuing capacity, passing rate boundary and the like of the toll gate channel are comprehensively considered, the main line traffic state is maintained near the expected value by adjusting the toll gate open channel, and then the road network operation efficiency is improved), and the channel number adjustment model is as follows:
M＝max(0，l i，t -l i，max ) (7)
in the above, N i，t For regulating traffic flow in next period of toll station i, Q i，t-Δt For the actual incoming traffic flow, k, in a period over the toll station i i，t For the actual road section density, k of the main road in the regulating and controlling direction of the toll station i des，i Desired road section density for main road in the control direction of toll station i, v des For the expected road section density of the main road in the regulating and controlling direction of the toll station i, deltat is the flow control period, M is the queuing overrun penalty term, cap i For the single passage capacity of toll stations, l i，t Queuing length for toll station, l i，max The queuing length is limited for toll stations.
Because the traditional ramp flow control model only controls a single ramp, however, the main road congestion is not necessarily caused by overlarge inlet flow of the adjacent ramp, and the effect of controlling the single ramp is often not achieved, the total expected flow Q can be calculated by adopting the channel number adjustment model desire The method comprises the following steps:
Q desire ＝(k i，t -k des，i )*v des *Δt+M (8)
scheme calculation step: according to the calculated overall expected flow control and the influence degree of the calculated inlet flow of each ramp on the flow of the main road section, calculating to obtain the actual flow control of each ramp by adopting a particle swarm algorithm, and according to the actual flow control of each ramp, calculating to obtain the optimal flow control scheme of each ramp in the road network by combining the manual toll lane traffic capacity and the electronic toll lane traffic capacity of each toll station;
after the degree (proportion) of influence of the inlet flow of each ramp on the flow of the main road section is determined when the overall expected flow control and the main line congestion relief are carried out, an optimization algorithm can be selected to calculate and formulate a multi-charging station ramp coordination optimal flow control scheme in the road network. The particle swarm algorithm is a method for searching the global optimal value by searching the current searched optimal value, and solves the optimal through information interaction in the population. Compared with other intelligent optimization algorithms, the particle swarm optimization method has the advantages of self-organization, evolutionary and memory functions, fewer parameters to be adjusted, simplicity in operation, high convergence speed and the like. The particles solved by the particle swarm algorithm are the actual traffic flow which needs to be controlled by each ramp.
Specifically, as shown in fig. 3, the method comprises initializing, calculating the particle fitness, searching for the individual extremum and the population extremum continuously, updating the speed and the position, calculating the particle fitness again after updating, and updating the individual extremum and the population extremum until the termination condition is met. Definition of particle swarmX＝(X 1 ，X 2 ，...，X n ) Wherein the ith particle represents a vector x= (X) i1 ，X i2 ，...，X iD ) D corresponds to the number of current-controlling ramps and X i1 Representing the traffic flow that the first ramp needs to control, the velocity V (V i1 ，V i2 ，...，V iD ) The individual extremum is P i (P i1 ，P i2 ，...，P iD ) Population extremum is P g (P g1 ，P g2 ，...，P gD )。
The particle update rate is according toAnd position->
In the above formula, ω is an inertial weight, C 1 ，C 2 Is an acceleration factor, k represents the number of iterations, R 1 ，R 2 Is a random number greater than 0 and less than 1.
The actual control flow of each ramp is multiplied by the influence degree (proportion) of each ramp given by an ST-GCN algorithm, the sum is the total actual control flow, the objective function is the difference between the total actual control flow and the total expected control flow, and the smaller the difference is, the better is, namely:
F(X)＝αX i1 +βX i2 +γx i3 +...κX iD -Q desire (11)
in the above formula, alpha, beta and gamma are weights of each ramp, and represent the influence degree (proportion) of each ramp, X i1 、X i2 、X i3 ...X iD Representing each ofThe actual flow control of the ramp, D, represents the total number of ramps needing to be subjected to flow control. The calculation is terminated when the difference between the total actual control flow and the total desired control flow is less than 2%.
After the particle swarm algorithm is solved, solving the opening number of the specific lanes of each ramp according to the manual toll lane traffic capacity and the electronic toll lane traffic capacity of each toll station, and obtaining a coordinated optimal flow control scheme of a plurality of toll stations in the road network.
Preferably, in the model building and calculating step, whether the real-time ramp flow control is needed or the ramp flow control is needed in advance is judged according to the real-time expressway main road section flow in the traffic data and the predicted expressway main road section flow and the congestion judging condition, if the congestion condition occurs, the real-time ramp flow control is carried out, and the flow calculating step is carried out; if no congestion condition exists, the current opening and closing state of the ramp is maintained, and the following relevant steps such as flow calculation, scheme calculation and the like are not carried out.
In the same test road network, whether ramp flow control is implemented or not is performed in VISSIM software for vehicle flow simulation. Specifically, two scenes are set: firstly, no control measures are taken; secondly, adopting an intelligent ramp flow control algorithm to control the flow of the ramp; and comparing the traffic volume of the main line of the corresponding road network with the congestion time of the main line by arranging a detector on the road section.
Information display step: the intelligent ramp flow control algorithm can output the recommended switching condition of each ramp of each toll station in detail, and the optimal flow control scheme is displayed through the information board arranged on the expressway, so that the behavior of a driver and a passenger is induced, and the expressway passing efficiency is improved.
The invention also relates to a monitoring system based on the ramp intelligent flow control algorithm, which corresponds to the monitoring method based on the ramp intelligent flow control algorithm, and can be understood as a system for realizing the method, and the system comprises a data acquisition and preprocessing module, a model building and calculating module, a flow calculating module, a scheme calculating module and an information display module, in particular,
the data acquisition and preprocessing module acquires traffic data and preprocesses the acquired traffic data;
the model building and calculating module is used for building a space-time diagram convolution network model, namely an ST-GCN model, by utilizing a diagram convolution network and a long-short-time memory network according to the preprocessed data and adopting a flow prediction algorithm in combination with an attention mechanism, performing iterative training on the built ST-GCN model to obtain an optimal ST-GCN model, predicting the main road section flow of the expressway according to the optimal ST-GCN model, and calculating the influence degree of the inlet flow of each ramp on the main road section flow;
the flow calculation module is used for: calculating overall expected control flow by adopting a highway ramp toll station channel number regulation algorithm according to the predicted main road section flow of the highway;
the scheme calculation module calculates actual control flow of each ramp by adopting a particle swarm algorithm according to the calculated overall expected control flow and the influence degree of the inlet flow of each ramp on the flow of the main road section, and calculates the optimal flow control scheme of each ramp in the road network according to the actual control flow of each ramp and the combination of the manual toll lane traffic capacity and the electronic toll lane traffic capacity of each toll station;
and the information display module displays the optimal flow control scheme through an information board arranged on the expressway.
Preferably, the model building and calculating module further judges whether to perform real-time ramp flow control or ramp flow control in advance according to the traffic data real-time expressway main road section flow and the predicted expressway main road section flow and the congestion judging condition, if the congestion condition occurs, the real-time ramp flow control is performed, and if the congestion condition does not occur, the current opening and closing state of the ramp is maintained.
Preferably, the ST-GCN model consists of a graph rolling network, a long-short-term memory network and an attention mechanism;
the model building and calculating module firstly uses the preprocessed historical time series data as input data, captures the topological structure of the expressway network by using a graph convolution network to obtain space correlation characteristics, then inputs the obtained time series with the space correlation characteristics into a long-short-period memory network, obtains dynamic changes through information transmission among units to obtain time characteristics, and then calculates the influence of the input data on the current prediction through a full connection layer and an attention mechanism to obtain a prediction result, namely the main road section flow of the expressway.
Preferably, the preprocessing includes deleting abnormal data and repeated data, filling up the missing data, and removing noise and data normalization.
The invention provides an objective and scientific monitoring method and system based on an intelligent ramp flow control algorithm, which are characterized in that the spatial correlation and the time correlation of traffic data are analyzed by establishing an ST-GCN model, the traffic road condition of a certain period of time in the future is predicted, the optimal flow control scheme of each ramp in a road network is obtained by adopting a particle swarm algorithm, manual intervention is not needed in the whole process, the working intensity of personnel is greatly reduced, the optimal flow control scheme is pushed to an information board arranged on a highway, and the behavior of a driver and a passenger is induced by the information of the information board, so that the traffic efficiency of the highway is effectively improved.
It should be noted that the above-described embodiments will enable those skilled in the art to more fully understand the invention, but do not limit it in any way. Therefore, although the present invention has been described in detail with reference to the drawings and examples, it will be understood by those skilled in the art that the present invention may be modified or equivalent, and in all cases, all technical solutions and modifications which do not depart from the spirit and scope of the present invention are intended to be included in the scope of the present invention.
Claims (10)
1. The monitoring method based on the ramp intelligent flow control algorithm is characterized by comprising the following steps of:
and (3) data acquisition and preprocessing: acquiring traffic data and preprocessing the acquired traffic data;
model establishment and calculation steps: according to the preprocessed data, a space-time diagram convolution network and a long and short time memory network are utilized, a space-time diagram convolution network model is established by combining an attention mechanism, iterative training is carried out on the established space-time diagram convolution network model, an optimal space-time diagram convolution network model is obtained, expressway main road section flow is predicted according to the optimal space-time diagram convolution network model, and the influence degree of each ramp inlet flow on the main road section flow is calculated;
and a flow calculating step: according to the predicted flow of the main road section of the expressway, a closed loop control channel number regulation model is constructed by adopting a feedback control-based expressway ramp toll gate channel number regulation algorithm to calculate the total expected control flow so as to maintain the main traffic state near an expected value;
scheme calculation step: according to the calculated overall expected flow control and the influence degree of the inlet flow of each ramp on the flow of the main road section, calculating to obtain the actual flow control of each ramp by adopting a particle swarm algorithm, and according to the actual flow control of each ramp, calculating to obtain the optimal flow control scheme of each ramp in the road network by combining the manual toll lane traffic capacity and the electronic toll lane traffic capacity of each toll station;
information display step: and displaying the optimal flow control scheme through an information board arranged on the expressway.
2. The method for monitoring the intelligent ramp flow control algorithm according to claim 1, wherein in the model building and calculating steps, whether the real-time ramp flow control is needed or performed in advance is judged according to the congestion judging conditions and the traffic data, if the congestion condition occurs, the real-time ramp flow control is performed, and the flow calculating step is performed.
3. The monitoring method based on ramp intelligent flow control algorithm according to claim 1, wherein in the model building and calculating steps, the space-time diagram convolution network model consists of a diagram convolution network, a long-period memory network and an attention mechanism;
firstly, using preprocessed historical time series data as input data, capturing a topological structure of a highway network by using a graph convolution network to obtain space correlation characteristics, inputting the obtained time series with the space correlation characteristics into a long-period memory network, obtaining dynamic changes through information transfer among units to obtain the time characteristics, and calculating the influence of the input data on current prediction through a full connection layer and an attention mechanism to obtain the main road section flow of the highway.
4. The method for monitoring the intelligent ramp-based flow control algorithm according to claim 1, wherein in the data acquisition and preprocessing step, the preprocessing includes deleting abnormal data and repeated data, filling the missing data, and removing noise and normalizing the data.
5. The method for monitoring a ramp-based intelligent flow control algorithm according to claim 1, wherein in the data acquisition and preprocessing step, the traffic data includes real-time traffic data and historical traffic data.
6. A monitoring system based on ramp intelligent flow control algorithm is characterized by comprising a data acquisition and preprocessing module, a model building and calculating module, a flow calculating module, a scheme calculating module and an information display module,
and the data acquisition and preprocessing module is used for: acquiring traffic data and preprocessing the acquired traffic data;
and a model building and calculating module: according to the preprocessed data, a space-time diagram convolution network and a long and short time memory network are utilized, a space-time diagram convolution network model is established by combining an attention mechanism, iterative training is carried out on the established space-time diagram convolution network model, an optimal space-time diagram convolution network model is obtained, expressway main road section flow is predicted according to the optimal space-time diagram convolution network model, and the influence degree of each ramp inlet flow on the main road section flow is calculated;
the flow calculation module is used for: according to the predicted flow of the main road section of the expressway, a closed loop control channel number regulation model is constructed by adopting a feedback control-based expressway ramp toll gate channel number regulation algorithm to calculate the total expected control flow so as to maintain the main traffic state near an expected value;
the scheme calculation module: according to the calculated overall expected flow control and the influence degree of the inlet flow of each ramp on the flow of the main road section, calculating to obtain the actual flow control of each ramp by adopting a particle swarm algorithm, and according to the actual flow control of each ramp, calculating to obtain the optimal flow control scheme of each ramp in the road network by combining the manual toll lane traffic capacity and the electronic toll lane traffic capacity of each toll station;
information display module: and displaying the optimal flow control scheme through an information board arranged on the expressway.
7. The system of claim 6, wherein the model building and calculating module further determines whether real-time ramp control is needed or performed in advance according to congestion determination conditions according to real-time highway main road section flow and predicted highway main road section flow in traffic data, and if congestion occurs, real-time ramp control is performed, and if congestion does not occur, current opening and closing states of the ramp are maintained.
8. The monitoring system based on the ramp intelligent flow control algorithm according to claim 6, wherein the space-time diagram convolution network model consists of a diagram convolution network, a long-period memory network and an attention mechanism;
the model building and calculating module firstly uses the preprocessed historical time series data as input data, captures the topological structure of the expressway network by using a graph convolution network to obtain space correlation characteristics, then inputs the obtained time series with the space correlation characteristics into a long-period memory network, obtains dynamic changes through information transmission among units to obtain time characteristics, and then calculates the influence of the input data on current prediction through a full connection layer and an attention mechanism to obtain the main road section flow of the expressway.
9. The intelligent ramp-control-algorithm-based monitoring system of claim 6, wherein the preprocessing includes deleting abnormal data and repeated data, filling in missing data, and removing noise and data normalization.
10. The ramp intelligent flow control algorithm based monitoring system of claim 6, wherein the traffic data includes real-time traffic data and historical traffic data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210263636.XA CN114582131B (en) | 2022-03-17 | 2022-03-17 | Monitoring method and system based on ramp intelligent flow control algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210263636.XA CN114582131B (en) | 2022-03-17 | 2022-03-17 | Monitoring method and system based on ramp intelligent flow control algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114582131A CN114582131A (en) | 2022-06-03 |
| CN114582131B true CN114582131B (en) | 2023-08-29 |
Family
ID=81780710
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210263636.XA Active CN114582131B (en) | 2022-03-17 | 2022-03-17 | Monitoring method and system based on ramp intelligent flow control algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114582131B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115116222B (en) * | 2022-06-21 | 2023-05-23 | 湖南省醴潭高速公路建设开发有限公司 | Multi-lane remote control method and system |
| CN115691144B (en) * | 2023-01-03 | 2023-03-31 | 西南交通大学 | Abnormal traffic state monitoring method, device and equipment and readable storage medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017140175A1 (en) * | 2016-02-18 | 2017-08-24 | 华南理工大学 | Toll road network traffic information collection and guidance system based on route identification system |
| CN111091712A (en) * | 2019-12-25 | 2020-05-01 | 浙江大学 | Traffic flow prediction method based on cyclic attention dual graph convolution network |
| CN111540199A (en) * | 2020-04-21 | 2020-08-14 | 浙江省交通规划设计研究院有限公司 | High-speed traffic flow prediction method based on multi-mode fusion and graph attention machine mechanism |
| CN111563611A (en) * | 2020-04-13 | 2020-08-21 | 北京工业大学 | Cloud data center renewable energy space-time prediction method for orientation graph convolutional network |
| CN112819213A (en) * | 2021-01-22 | 2021-05-18 | 华南理工大学 | Expressway freight volume prediction method and system based on deep learning network |
| CN113505536A (en) * | 2021-07-09 | 2021-10-15 | 兰州理工大学 | Optimized traffic flow prediction model based on space-time diagram convolution network |
| CN113744526A (en) * | 2021-08-25 | 2021-12-03 | 江苏大学 | Expressway risk prediction method based on LSTM and BF |
-
2022
- 2022-03-17 CN CN202210263636.XA patent/CN114582131B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017140175A1 (en) * | 2016-02-18 | 2017-08-24 | 华南理工大学 | Toll road network traffic information collection and guidance system based on route identification system |
| CN111091712A (en) * | 2019-12-25 | 2020-05-01 | 浙江大学 | Traffic flow prediction method based on cyclic attention dual graph convolution network |
| CN111563611A (en) * | 2020-04-13 | 2020-08-21 | 北京工业大学 | Cloud data center renewable energy space-time prediction method for orientation graph convolutional network |
| CN111540199A (en) * | 2020-04-21 | 2020-08-14 | 浙江省交通规划设计研究院有限公司 | High-speed traffic flow prediction method based on multi-mode fusion and graph attention machine mechanism |
| CN112819213A (en) * | 2021-01-22 | 2021-05-18 | 华南理工大学 | Expressway freight volume prediction method and system based on deep learning network |
| CN113505536A (en) * | 2021-07-09 | 2021-10-15 | 兰州理工大学 | Optimized traffic flow prediction model based on space-time diagram convolution network |
| CN113744526A (en) * | 2021-08-25 | 2021-12-03 | 江苏大学 | Expressway risk prediction method based on LSTM and BF |
Non-Patent Citations (1)
| Title |
|---|
| 基于图卷积网络的路网短时交通流预测研究;陈喜群;周凌霄;曹震;;交通运输系统工程与信息(04);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114582131A (en) | 2022-06-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112241814B (en) | Traffic prediction method based on reinforced space-time diagram neural network | |
| CN112216108B (en) | Traffic prediction method based on attribute-enhanced space-time graph convolution model | |
| CN110164128B (en) | City-level intelligent traffic simulation system | |
| CN111612243B (en) | Traffic speed prediction method, system and storage medium | |
| CN113313947B (en) | Road condition evaluation method of short-term traffic prediction graph convolution network | |
| CN114582131B (en) | Monitoring method and system based on ramp intelligent flow control algorithm | |
| Huang et al. | Diffusion convolutional recurrent neural network with rank influence learning for traffic forecasting | |
| CN113053115B (en) | Traffic prediction method based on multi-scale graph convolution network model | |
| CN111899510A (en) | Intelligent traffic system flow short-term prediction method and system based on divergent convolution and GAT | |
| CN115273466B (en) | Monitoring method and system based on flexible lane management and control algorithm | |
| CN114664091A (en) | Early warning method and system based on holiday traffic prediction algorithm | |
| CN111553535B (en) | Navigation reference method based on AE-LSTM-BO traffic flow prediction | |
| Coşkun et al. | Deep reinforcement learning for traffic light optimization | |
| CN115148019A (en) | Early warning method and system based on holiday congestion prediction algorithm | |
| Dahal et al. | GA-based learning for rule identification in fuzzy neural networks | |
| CN112766603A (en) | Traffic flow prediction method, system, computer device and storage medium | |
| CN115376317B (en) | Traffic flow prediction method based on dynamic graph convolution and time sequence convolution network | |
| CN110675632A (en) | Vehicle short-time trajectory prediction control method aiming at multi-feature space and data sparseness | |
| CN113487856A (en) | Traffic flow combination prediction model based on graph convolution network and attention mechanism | |
| CN111242395A (en) | Method and device for constructing prediction model for OD (origin-destination) data | |
| Rahman | SHORT TERM TRAFFIC FLOW PREDICTION USING MACHINE LEARNING-KNN, SVM AND ANN WITH WEATHER INFORMATION. | |
| CN114565187A (en) | Traffic network data prediction method based on graph space-time self-coding network | |
| CN115206092A (en) | Traffic prediction method of BiLSTM and LightGBM model based on attention mechanism | |
| CN113362637A (en) | Regional multi-field-point vacant parking space prediction method and system | |
| CN115907122A (en) | Regional electric vehicle charging load prediction method |
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 |