CN116699730A - Road group fog prediction method based on edge calculation - Google Patents
Road group fog prediction method based on edge calculation Download PDFInfo
- Publication number
- CN116699730A CN116699730A CN202310973181.5A CN202310973181A CN116699730A CN 116699730 A CN116699730 A CN 116699730A CN 202310973181 A CN202310973181 A CN 202310973181A CN 116699730 A CN116699730 A CN 116699730A
- Authority
- CN
- China
- Prior art keywords
- fog
- cluster
- grid
- vehicle
- mist
- 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.)
- Granted
Links
- 238000004364 calculation method Methods 0.000 title claims abstract description 49
- 238000000034 method Methods 0.000 title claims abstract description 36
- 239000003595 mist Substances 0.000 claims abstract description 58
- 238000007781 pre-processing Methods 0.000 claims abstract description 7
- 230000008859 change Effects 0.000 claims description 15
- 230000006870 function Effects 0.000 claims description 9
- 238000000605 extraction Methods 0.000 claims description 6
- 230000007246 mechanism Effects 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 6
- 230000003014 reinforcing effect Effects 0.000 claims description 6
- 230000002159 abnormal effect Effects 0.000 claims description 3
- 238000013459 approach Methods 0.000 claims description 3
- 238000013528 artificial neural network Methods 0.000 claims description 3
- 125000004122 cyclic group Chemical group 0.000 claims description 3
- 239000000284 extract Substances 0.000 claims description 3
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 239000013598 vector Substances 0.000 claims description 3
- 238000012544 monitoring process Methods 0.000 description 5
- 230000009286 beneficial effect Effects 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 206010039203 Road traffic accident Diseases 0.000 description 3
- 230000009471 action Effects 0.000 description 3
- 238000001514 detection method Methods 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01W—METEOROLOGY
- G01W1/00—Meteorology
- G01W1/10—Devices for predicting weather conditions
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/09—Arrangements for giving variable traffic instructions
- G08G1/0962—Arrangements for giving variable traffic instructions having an indicator mounted inside the vehicle, e.g. giving voice messages
- G08G1/0967—Systems involving transmission of highway information, e.g. weather, speed limits
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Atmospheric Sciences (AREA)
- Environmental & Geological Engineering (AREA)
- Theoretical Computer Science (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Probability & Statistics with Applications (AREA)
- Biodiversity & Conservation Biology (AREA)
- Ecology (AREA)
- Environmental Sciences (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention discloses a road group fog prediction method based on edge calculation, which comprises the following steps: a vehicle-mounted device on the vehicle collects vehicle passing data every t seconds and performs data preprocessing to obtain mist index data; calculating a global optimal value of the edge server, and uploading the group fog index data to the edge server with the maximum global optimal value; calculating the density of the fog clusters and marking according to the longitude and latitude of the vehicle to obtain a density map of the fog clusters; constructing a KP clustering model, and clustering the mist density map to obtain a mist region cluster and a cluster center; updating the cluster center; constructing a space-time position prediction model to predict the position of the center of the cluster to obtain a cluster fog movement trend; the edge server sends the mass fog movement trend to the center server, and the center server gives an early warning to vehicles around the mass fog area; according to the method, the problem of real-time prediction and safety early warning of the road mist is effectively solved by predicting the mist movement trend.
Description
Technical Field
The invention relates to the field of computer edge calculation, in particular to a road group fog prediction method based on edge calculation.
Background
The mist is also mist in nature, is the mist which is influenced by the microclimate environment of local areas, and has stronger and lower visibility in the local range of tens to hundreds of meters in the large mist; the cluster fog has the characteristics of burst property, locality, small scale and large concentration, is difficult to forecast, has good external vision, and is hazy in the cluster fog, so that serious traffic accidents are easy to cause; the prediction of the movement trend of the mass fog is beneficial to reducing or avoiding the loss of highway traffic weather disasters, particularly for the turning extreme weather of the mass fog, the early warning information of the weather bureau is aimed at the prediction in a large range, the prediction of specific places at specific time is not accurate, and the guidance reliability of traffic roads is not high; the traditional prediction method still has a plurality of problems in the actual process, such as low efficiency, high cost, large risk, poor stability, lack of an effective prediction method, difficulty in timely and accurately predicting the group fog, and great trouble and potential safety hazard for road traffic; in view of these drawbacks, it is necessary to design a road fog prediction method based on edge calculation.
The publication number is CN107240281A, the patent literature of a group fog warning prediction system is named, the system is applied to the domestic intelligent traffic field through a road traffic safety information acquisition unit, a road traffic safety information analysis unit, an information release system and a traffic command center warning system, meteorological data and traffic safety information are combined and converted into road traffic safety information, warning and prediction are carried out aiming at group fog formation, group fog grade, group fog range and group fog moving area, the road safety force is increased, but the transmission of monitoring data is slower, and a warning information release channel is narrower, the change trend of the group fog cannot be monitored and judged in real time, and the timeliness is poor.
Patent literature with publication number of CN111443019A and name of a group fog detection device and a group fog moving direction judging method sends laser through a laser, the laser is received by an illuminance sensor, the illuminance sensor outputs a voltage value to a controller, the group fog concentration is judged according to the output voltage value, the Beidou positioning module obtains latitude and longitude information of a collector, the GSM communication module sends the latitude and longitude information to a host computer management center, and the transmitted laser intensity is utilized to detect the group fog concentration.
Disclosure of Invention
In order to solve the technical problems, the invention adopts a technical scheme that: the method for predicting the road fog based on edge calculation comprises the following steps:
s1, collecting vehicle passing data at intervals of t seconds by a vehicle-mounted device on a vehicle, and carrying out data preprocessing on the vehicle passing data to obtain mist index data;
s2, calculating a global optimal value of an edge server with the vehicle as a central radius within the range R, and uploading the group fog index data to the edge server with the maximum global optimal value;
s3, calculating the mass fog density according to the mass fog index data, and marking the mass fog density according to the longitude and latitude of the vehicle to obtain a mass fog density map;
s4, constructing a KP clustering model, and clustering the group fog density map by using the KP clustering model according to the group fog density to obtain group fog area clusters and cluster centers;
s5, updating the cluster-like center;
s6, constructing a space-time position prediction model, and predicting the position of the cluster center through the space-time position prediction model to obtain a cluster fog movement trend;
s7, the edge server sends the mass fog movement trend to a central server, and the central server gives early warning to traffic management departments and vehicles around the mass fog area;
the in-vehicle apparatus includes: the system comprises a meteorological sensor, a position module, a calculation module and a storage module;
the vehicle approach data includes: atmospheric data and location information;
the atmospheric data includes: atmospheric humidity, temperature, PM2.5 concentration, atmospheric pressure, wind speed, wind direction and air pressure reverse temperature;
the location information includes: longitude, latitude, time;
the data preprocessing comprises the following steps: removing the repeated value, the missing value and the abnormal value, and performing normalization operation;
and R is 5km.
Further, the step S2 includes:
s21, calculating a global optimal value X of an edge server in a range with a radius R by taking a vehicle as a center;
the calculation formula of the global optimal value X is as follows:
；
wherein ,is weight value->Representing the distance between the vehicle and the edge server, < >>For the load case of edge server i +.>Indicating the variation of the mist index data, +.>Weather factors are divided into four weather, namely sunny days, cloudy days, rainy days and snowy days, and the weather factor values from sunny days to snowy days are gradually increased;
the calculation formula of the distance between the vehicle and the edge server is as follows:
；
wherein ,representing longitude,>representing latitude,>representing the longitude of the edge server,representing the latitude of the edge server;
the calculation formula of the variation of the mist index data is as follows:
；
wherein ,indicating that the vehicle is +.>Data quantity of time of day->Indicating that the vehicle is +.>The amount of data at the moment;
s22, creating an edge server queue, and sequentially storing the edge servers into the edge server queue after the edge servers are ordered according to a global optimal value descending order;
s23, the vehicle selects a first edge server in the edge server queue to upload the mist index data.
Further, the step S3 includes:
s31, calculating the mass fog density according to the mass fog index data;
the calculation formula of the mass fog density S is as follows:
；
wherein ,is the temperature and humidity weight>Is the comprehensive weight, w 1 +w 2 =1，/>For the current atmospheric humidity, T is the current temperature, < >>For the concentration of PM2.5 after normalization, +.>For the concentration of PM2.5 at the current moment, wp is wind speed, f is a fuzzy factor, and f takes a value of 3;
s32, judging whether the mass fog density is larger than or equal to a threshold value Q, if so, marking the mass fog density on the current longitude and the current latitude of the vehicle according to the vehicle position information; if not, the mass fog density of the position point of the vehicle is set as None;
s33, marking the mass fog density of all vehicles according to the longitude and latitude of the vehicles to obtain a mass fog density map;
s34, dividing the mass fog density map into W multiplied by W grid units;
s35, if the grid unit only contains a mark with the mist density of None, setting the mist grid density to 0, otherwise, setting the maximum value of the mist density in the grid unit to be the mist grid density;
s36, creating a grid set, sorting the grid units in descending order according to the density of the mist grid, and sequentially storing the grid units into the grid set;
the magnitude of the threshold Q is；
And W is a positive integer greater than or equal to 1.
Further, the step S4 includes:
s41, selecting a grid cell with the maximum mass fog grid density in the mass fog density map as an initial grid center;
s42, merging the target grid unit with the initial grid center to obtain a new grid center;
s43, deleting the combined grid cells from the grid set, and stopping grid combining operation until the target grid cells do not exist to obtain a first cluster fog region cluster;
s44, selecting a grid cell with the maximum current group fog grid density in the grid set as a second initial grid center, and carrying out grid merging operation to obtain a second group fog region cluster;
s45, repeating the steps S41-S44 until the number of the grid centers reaches k, and ending clustering to obtain k cluster fog area clusters and k cluster center;
s46, calculating an optimal discrimination function value, and obtaining an optimal clustering result when the optimal discrimination function value is minimum;
the target grid cell is a grid cell adjacent to the initial grid center and having a cluster fog grid density greater than 0;
the calculation formula of the optimal discrimination function value is as follows:
；
wherein ki represents the cluster center, k represents the number of clusters,representing the difference between the density of the clusters of clusters in each cluster, +.>Representing the differences between each cluster of classes.
Further, the step S5 includes:
s51, calculating to obtain the priority SV of the center point according to the mist index data uploaded every t seconds;
the calculation formula of the center point priority SV is as follows:
；
wherein ,vehicle with number i>Group fog density for vehicle number i, < >>For the difference degree between the group fog index data of each cluster center at different moments, ki is the cluster center and is added with->Weight of haze density +.>Weight of degree of difference +.>The greater the priority of the center point is, the higher the priority of the center point is for the total number of vehicles;
s52, creating a preselected class cluster center point queue, and sequentially storing k class cluster centers into the preselected class cluster center point queue after descending and sequencing according to the priority of the center point;
s53, selecting a cluster center with the highest priority of the center point from a preselected cluster center point queue, and updating the cluster center closest to the vehicle marking position to obtain a new cluster center;
s54, continuously carrying out position update on the cluster center point of each cluster of the cluster fog area based on the cluster fog index data at different moments, creating a position update sequence, and storing the position information of the cluster center updated each time into the position update sequence;
the location update sequence is expressed as tnew= { L1, L2, …, ln }, where Ln is the location information of the cluster center of the cluster fog area after the nth update.
Further, the space-time position prediction model comprises a feature extraction module, a feature splicing module and a position prediction module, wherein the space-time position prediction model is based on a position prediction algorithm of a space-time cyclic neural network, the space features and the time features are directly fed into the network, and the network is responsible for learning the internal representation of the network; the feature extraction module converts original feature sequence data into a group of vectors or matrixes representing data features, and extracts valuable feature information in the sequence; the feature splicing module utilizes the relation among different features to fuse a plurality of input features and output a more comprehensive feature representation; the position prediction module comprises a long-period memory network and a multi-head self-attention mechanism, wherein the long-period memory network is used for capturing long-period time dependence, and the multi-head self-attention mechanism is used for sensing key features in each sequence and performing self-adaptive adjustment, and position prediction is generated after capturing long-period dependence.
Further, the step S6 includes:
s61, extracting relation features among the positions of the centers of k class clusters from the position updating sequence, creating a position feature sequence, and adding the relation features into the position feature sequence;
the relation featuresRepresenting the relation between the positions, said relation being characterized by +.>The calculation formula of (2) is shown as follows：
；
；
wherein ,is the position information of the cluster center of the m-th cluster fog area cluster, and is +.>，/>The SV is the center point priority, which is the position information of the cluster center of the jth cluster fog area cluster>TNew is a position updating sequence for the similarity of positions among the centers of various clusters;
s62, counting time intervals of position change when the cluster center is updated each time to form a time interval sequence G= {}；
S63, counting the occurrence frequency of each time interval in the time interval sequence, calculating the frequency of each time interval, and adding the occurrence frequency and the frequency of the time interval into the time interval sequence to obtain a time feature sequence;
；
；
；
；
wherein ,for every time interval +.>Frequency of->For the frequency of occurrence of time intervals, N is the frequency of all time intervals, < >>Is a time feature sequence;
s64, creating an influence characteristic sequence, and adding the current wind speed, time, position, wind direction, atmospheric humidity and air pressure inverse temperature into the influence characteristic sequence;
the influencing characteristic sequence is, wherein />For wind speed>For time (I)>For the position of->Wind direction, the->Atmospheric humidity, < >>The current wind speed, time, position, wind direction, atmospheric humidity and air pressure inverse temperature are determined by the centerObtaining by a server;
s65, extracting features of the position feature sequence, the time feature sequence and the influence feature sequence, splicing and fusing various features to serve as reinforcing features, and inputting the reinforcing features into a space-time position prediction model to accurately predict the position of a central point at the next moment;
s66, creating a position prediction sequence, and adding central point position information obtained by each prediction into the position prediction sequence;
s67, marking the position information of the central point in the position prediction sequence on a mass fog density chart to obtain a prediction track, wherein the prediction track is the movement track of the mass fog;
s68, obtaining the change trend of the mist through the mist movement track.
Further, the step S7 includes:
s71, the edge server sends the cluster fog change trend to a central server;
s72, marking an area to be covered by the mist as a mist area by the central server, and recording the position information of the mist area and the arrival time of the mist;
s73, the central server sends the position information of the group fog area and the arrival time of the group fog to a traffic management department, and synchronously gives vehicles in the area around the group fog to carry out group fog early warning.
In summary, due to the adoption of the technical scheme, the beneficial effects of the invention are as follows:
1. the road group fog prediction method based on the edge calculation provided by the invention avoids traffic accidents caused by the influence of the group fog on the sight line, is beneficial to a vehicle driver to know the change trend of the group fog, and is low in cost, large in detection range, convenient for a user to operate and capable of improving the reliability of group fog monitoring;
2. according to the road group fog prediction method based on edge calculation, the change trend of group fog is predicted by using a space-time position prediction model based on deep learning, prediction is accurate, a pre-plan can be set in advance, and when the numerical value is monitored to reach a certain threshold value, early warning information is automatically sent, so that the purposes of early warning and safe trip are achieved;
3. according to the road group fog prediction method based on edge calculation, the group fog prediction information is updated in real time through the data collected by the vehicle, and meanwhile, the early warning information is updated in real time, so that the instantaneity and timeliness of the early warning information can be effectively improved, and the problems of real-time monitoring and safety early warning of the road group fog can be effectively solved.
Drawings
Fig. 1 is a flowchart of a road mass fog prediction method based on edge calculation.
Fig. 2 is a flowchart of creating a cluster fog density map of the road cluster fog prediction method based on edge calculation.
Fig. 3 is a block fog clustering flow chart of the road block fog prediction method based on edge calculation.
Fig. 4 is a flowchart of an updating clustering center of the road mass fog prediction method based on edge calculation.
Fig. 5 is a flowchart of a predicted cluster fog change trend of the road cluster fog prediction method based on edge calculation.
Detailed Description
The preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings so that the advantages and features of the present invention can be more easily understood by those skilled in the art, thereby making clear and defining the scope of the present invention.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention, but the present invention may be practiced otherwise than as described herein; it will be apparent that the embodiments in the specification are only some, but not all, embodiments of the invention.
Fig. 1 is a flowchart of a road mass fog prediction method based on edge calculation, which includes:
s1, collecting vehicle passing data at intervals of t seconds by a vehicle-mounted device on a vehicle, and carrying out data preprocessing on the vehicle passing data to obtain mist index data;
the in-vehicle apparatus includes: the system comprises a meteorological sensor, a position module, a calculation module and a storage module;
the vehicle approach data includes: atmospheric data and location information;
the atmospheric data includes: atmospheric humidity, temperature, PM2.5 concentration, atmospheric pressure, wind speed, wind direction and air pressure reverse temperature;
the location information includes: longitude, latitude, time;
the data preprocessing comprises the following steps: and removing the repeated value, the missing value and the abnormal value, and performing normalization operation.
S2, calculating a global optimal value of an edge server with the vehicle as a central radius within the range R, and uploading the group fog index data to the edge server with the maximum global optimal value;
and R is 5km.
Further, the step S2 includes:
s21, calculating a global optimal value X of an edge server in a range with a radius R by taking a vehicle as a center;
the calculation formula of the global optimal value X is as follows:
；
wherein ,is weight value->Representing the distance between the vehicle and the edge server, < >>For the load case of edge server i +.>Indicating the variation of the mist index data, +.>For weather reasonThe method comprises the steps of dividing the weather into four days of sunny days, cloudy days, rainy days and snowy days, wherein the weather factor value from sunny days to snowy days is gradually increased;
the calculation formula of the distance between the vehicle and the edge server is as follows:
；
wherein ,representing longitude,>representing latitude,>representing the longitude of the edge server,representing the latitude of the edge server;
the calculation formula of the variation of the mist index data is as follows:
；
wherein ,indicating that the vehicle is +.>Data quantity of time of day->Indicating that the vehicle is +.>The amount of data at the moment;
s22, creating an edge server queue, and sequentially storing the edge servers into the edge server queue after the edge servers are ordered according to a global optimal value descending order;
s23, the vehicle selects a first edge server in the edge server queue to upload the mist index data.
S3, calculating the mass fog density according to the mass fog index data, and marking the mass fog density according to the longitude and latitude of the vehicle to obtain a mass fog density map.
Further, referring to fig. 2, the step S3 includes:
s31, calculating the mass fog density according to the mass fog index data;
the calculation formula of the mass fog density S is as follows:
；
wherein ,is the temperature and humidity weight>Is the comprehensive weight, w 1 +w 2 =1，/>For the current atmospheric humidity, T is the current temperature, < >>For the concentration of PM2.5 after normalization, +.>For the concentration of PM2.5 at the current moment, wp is wind speed, f is a fuzzy factor, and f takes a value of 3;
s32, judging whether the mass fog density is larger than or equal to a threshold value Q, if so, marking the mass fog density on the current longitude and the current latitude of the vehicle according to the vehicle position information; if not, the mass fog density of the position point of the vehicle is set as None;
s33, marking the mass fog density of all vehicles according to the longitude and latitude of the vehicles to obtain a mass fog density map;
s34, dividing the mass fog density map into W multiplied by W grid units;
s35, if the grid unit only contains a mark with the mist density of None, setting the mist grid density to 0, otherwise, setting the maximum value of the mist density in the grid unit to be the mist grid density;
s36, creating a grid set, sorting the grid units in descending order according to the density of the mist grid, and sequentially storing the grid units into the grid set;
the magnitude of the threshold Q is；
And W is a positive integer greater than or equal to 1.
S4, constructing a KP clustering model, and clustering the group fog density map by using the KP clustering model according to the group fog density to obtain group fog area clusters and cluster centers.
Further, referring to fig. 3, the S4 includes:
s41, selecting a grid cell with the maximum mass fog grid density in the mass fog density map as an initial grid center;
s42, merging the target grid unit with the initial grid center to obtain a new grid center;
s43, deleting the combined grid cells from the grid set, and stopping grid combining operation until the target grid cells do not exist to obtain a first cluster fog region cluster;
s44, selecting a grid cell with the maximum current group fog grid density in the grid set as a second initial grid center, and carrying out grid merging operation to obtain a second group fog region cluster;
s45, repeating the steps S41-S44 until the number of the grid centers reaches k, and ending clustering to obtain k cluster fog area clusters and k cluster center;
s46, calculating an optimal discrimination function value, and obtaining an optimal clustering result when the optimal discrimination function value is minimum;
the target grid cell is a grid cell adjacent to the initial grid center and having a cluster fog grid density greater than 0;
the calculation formula of the optimal discrimination function value is as follows:
；
wherein ki represents the cluster center, k represents the number of clusters,representing the difference between the density of the clusters of clusters in each cluster, +.>Representing the differences between each cluster of classes.
S5, updating the cluster-like center.
Further, referring to fig. 4, the step S5 includes:
s51, calculating to obtain the priority SV of the center point according to the mist index data uploaded every t seconds;
the calculation formula of the center point priority SV is as follows:
；
wherein ,vehicle with number i>Group fog density for vehicle number i, < >>For the difference degree between the group fog index data of each cluster center at different moments, ki is the cluster center and is added with->Weight of haze density +.>Weight of degree of difference +.>The greater the priority of the center point is, the higher the priority of the center point is for the total number of vehicles;
s52, creating a preselected class cluster center point queue, and sequentially storing k class cluster centers into the preselected class cluster center point queue after descending and sequencing according to the priority of the center point;
s53, selecting a cluster center with the highest priority of the center point from a preselected cluster center point queue, and updating the cluster center closest to the vehicle marking position to obtain a new cluster center;
s54, continuously carrying out position update on the cluster center point of each cluster of the cluster fog area based on the cluster fog index data at different moments, creating a position update sequence, and storing the position information of the cluster center updated each time into the position update sequence;
the location update sequence is expressed as tnew= { L1, L2, …, ln }, where Ln is the location information of the cluster center of the cluster fog area after the nth update.
S6, constructing a space-time position prediction model, and predicting the position of the cluster center through the space-time position prediction model to obtain a cluster fog movement trend.
Further, the space-time position prediction model comprises a feature extraction module, a feature splicing module and a position prediction module, wherein the space-time position prediction model is based on a position prediction algorithm of a space-time cyclic neural network, the space features and the time features are directly fed into the network, and the network is responsible for learning the internal representation of the network; the feature extraction module converts original feature sequence data into a group of vectors or matrixes representing data features, and extracts valuable feature information in the sequence; the feature splicing module utilizes the relation among different features to fuse a plurality of input features and output a more comprehensive feature representation; the position prediction module comprises a long-period memory network and a multi-head self-attention mechanism, wherein the long-period memory network is used for capturing long-period time dependence, and the multi-head self-attention mechanism is used for sensing key features in each sequence and performing self-adaptive adjustment, and position prediction is generated after capturing long-period dependence.
Further, referring to fig. 5, the step S6 includes:
s61, extracting relation features among the positions of the centers of k class clusters from the position updating sequence, creating a position feature sequence, and adding the relation features into the position feature sequence;
the relation featuresRepresenting the relation between the positions, said relation being characterized by +.>The calculation formula of (2) is as follows:
；
；
wherein ,is the position information of the cluster center of the m-th cluster fog area cluster, and is +.>，/>The SV is the center point priority, which is the position information of the cluster center of the jth cluster fog area cluster>TNew is a position updating sequence for the similarity of positions among the centers of various clusters;
s62, counting time intervals of position change when the cluster center is updated each time to form a time interval sequence G= {}；
S63, counting the occurrence frequency of each time interval in the time interval sequence, calculating the frequency of each time interval, and adding the occurrence frequency and the frequency of the time interval into the time interval sequence to obtain a time feature sequence;
；
；
；
；
wherein ,for every time interval +.>Frequency of->For the frequency of occurrence of time intervals, N is the frequency of all time intervals, < >>Is a time feature sequence;
s64, creating an influence characteristic sequence, and adding the current wind speed, time, position, wind direction, atmospheric humidity and air pressure inverse temperature into the influence characteristic sequence;
the influencing characteristic sequence is, wherein />For wind speed>For time (I)>For the position of->Wind direction, the->Atmospheric humidity, < >>The current wind speed, time, position, wind direction, atmospheric humidity and air pressure reverse temperature are obtained by a central server;
s65, extracting features of the position feature sequence, the time feature sequence and the influence feature sequence, splicing and fusing various features to serve as reinforcing features, and inputting the reinforcing features into a space-time position prediction model to accurately predict the position of a central point at the next moment;
s66, creating a position prediction sequence, and adding central point position information obtained by each prediction into the position prediction sequence;
s67, marking the position information of the central point in the position prediction sequence on a mass fog density chart to obtain a prediction track, wherein the prediction track is the movement track of the mass fog;
s68, obtaining the change trend of the mist through the mist movement track.
S7, the edge server sends the mass fog movement trend to a central server, and the central server gives an early warning to traffic management departments and vehicles around the mass fog area.
Further, the step S7 includes:
s71, the edge server sends the cluster fog change trend to a central server;
s72, marking an area to be covered by the mist as a mist area by the central server, and recording the position information of the mist area and the arrival time of the mist;
s73, the central server sends the position information of the group fog area and the arrival time of the group fog to a traffic management department, and synchronously gives vehicles in the area around the group fog to carry out group fog early warning.
The road group fog prediction method based on the edge calculation provided by the invention avoids traffic accidents caused by the influence of the group fog on the sight line, is beneficial to a vehicle driver to know the change trend of the group fog, and is low in cost, large in detection range, convenient for a user to operate and capable of improving the reliability of group fog monitoring; the change trend of the mist is predicted by using a space-time position prediction model based on deep learning, the prediction is accurate, a plan can be set in advance, and when the numerical value is monitored to reach a certain threshold value, early warning information is automatically sent, so that the purposes of early warning and safe trip are achieved; the cluster fog prediction information is updated in real time through the data collected by the vehicle, and meanwhile, the early warning information is updated in real time, so that the real-time performance and timeliness of the early warning information can be effectively improved, and the problems of real-time monitoring and safety early warning of the road cluster fog can be effectively solved.
In the present specification, each embodiment is described in a progressive manner, and each embodiment is mainly described in a different point from other embodiments, and identical and similar parts between the embodiments are all enough to refer to each other.
It should be noted that in this document, relational terms such as "first" and "second" and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises an element.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (8)
1. The road fog prediction method based on edge calculation is characterized by comprising the following steps of:
s1, collecting vehicle passing data at intervals of t seconds by a vehicle-mounted device on a vehicle, and carrying out data preprocessing on the vehicle passing data to obtain mist index data;
s2, calculating a global optimal value of an edge server with the vehicle as a central radius within the range R, and uploading the group fog index data to the edge server with the maximum global optimal value;
s3, calculating the mass fog density according to the mass fog index data, and marking the mass fog density according to the longitude and latitude of the vehicle to obtain a mass fog density map;
s4, constructing a KP clustering model, and clustering the group fog density map by using the KP clustering model according to the group fog density to obtain group fog area clusters and cluster centers;
s5, updating the cluster-like center;
s6, constructing a space-time position prediction model, and predicting the position of the cluster center through the space-time position prediction model to obtain a cluster fog movement trend;
s7, the edge server sends the mass fog movement trend to a central server, and the central server gives early warning to traffic management departments and vehicles around the mass fog area;
the in-vehicle apparatus includes: the system comprises a meteorological sensor, a position module, a calculation module and a storage module;
the vehicle approach data includes: atmospheric data and location information;
the atmospheric data includes: atmospheric humidity, temperature, PM2.5 concentration, atmospheric pressure, wind speed, wind direction and air pressure reverse temperature;
the location information includes: longitude, latitude, time;
the data preprocessing comprises the following steps: removing the repeated value, the missing value and the abnormal value, and performing normalization operation;
and R is 5km.
2. The road bolus prediction method based on edge calculation according to claim 1, wherein S2 includes:
s21, calculating a global optimal value X of an edge server in a range with a radius R by taking a vehicle as a center;
the calculation formula of the global optimal value X is as follows:
；
wherein ,is weight value->Representing the distance between the vehicle and the edge server, < >>For the load case of edge server i +.>Indicating the variation of the mist index data, +.>Weather factors are divided into four weather, namely sunny days, cloudy days, rainy days and snowy days, and the weather factor values from sunny days to snowy days are gradually increased;
the calculation formula of the distance between the vehicle and the edge server is as follows:
；
wherein ,representing longitude,>representing latitude,>representing longitude of edge server, ++>Representing the latitude of the edge server;
the calculation formula of the variation of the mist index data is as follows:
；
wherein ,indicating that the vehicle is +.>Data quantity of time of day->Indicating that the vehicle is +.>The amount of data at the moment;
s22, creating an edge server queue, and sequentially storing the edge servers into the edge server queue after the edge servers are ordered according to a global optimal value descending order;
s23, the vehicle selects a first edge server in the edge server queue to upload the mist index data.
3. The road bolus prediction method based on edge calculation according to claim 1, wherein S3 includes:
s31, calculating the mass fog density according to the mass fog index data;
the calculation formula of the mass fog density S is as follows:
；
wherein ,is the temperature and humidity weight>Is the comprehensive weight, w 1 +w 2 =1，/>For the current atmospheric humidity, T is the current temperature,for the concentration of PM2.5 after normalization, +.>For the concentration of PM2.5 at the current moment, wp is wind speed, f is a fuzzy factor, and f takes a value of 3;
s32, judging whether the mass fog density is larger than or equal to a threshold value Q, if so, marking the mass fog density on the current longitude and the current latitude of the vehicle according to the vehicle position information; if not, the mass fog density of the position point of the vehicle is set as None;
s33, marking the mass fog density of all vehicles according to the longitude and latitude of the vehicles to obtain a mass fog density map;
s34, dividing the mass fog density map into W multiplied by W grid units;
s35, if the grid unit only contains a mark with the mist density of None, setting the mist grid density to 0, otherwise, setting the maximum value of the mist density in the grid unit to be the mist grid density;
s36, creating a grid set, sorting the grid units in descending order according to the density of the mist grid, and sequentially storing the grid units into the grid set;
the magnitude of the threshold Q is；
And W is a positive integer greater than or equal to 1.
4. The road bolus prediction method based on edge calculation according to claim 1, wherein S4 includes:
s41, selecting a grid cell with the maximum mass fog grid density in the mass fog density map as an initial grid center;
s42, merging the target grid unit with the initial grid center to obtain a new grid center;
s43, deleting the combined grid cells from the grid set, and stopping grid combining operation until the target grid cells do not exist to obtain a first cluster fog region cluster;
s44, selecting a grid cell with the maximum current group fog grid density in the grid set as a second initial grid center, and carrying out grid merging operation to obtain a second group fog region cluster;
s45, repeating the steps S41-S44 until the number of the grid centers reaches k, and ending clustering to obtain k cluster fog area clusters and k cluster center;
s46, calculating an optimal discrimination function value, and obtaining an optimal clustering result when the optimal discrimination function value is minimum;
the target grid cell is a grid cell adjacent to the initial grid center and having a cluster fog grid density greater than 0;
the calculation formula of the optimal discrimination function value is as follows:
；
wherein ki represents the cluster center, k represents the number of clusters,representing the difference between the density of the clusters of clusters in each cluster, +.>Representing the differences between each cluster of classes.
5. The road bolus prediction method based on edge calculation according to claim 1, wherein S5 includes:
s51, calculating to obtain the priority SV of the center point according to the mist index data uploaded every t seconds;
the calculation formula of the center point priority SV is as follows:
；
wherein ,vehicle with number i>Group fog density for vehicle number i, < >>For the difference degree between the group fog index data of each cluster center at different moments, ki is the cluster center and is added with->Weight of haze density +.>Weight of degree of difference +.>The greater the priority of the center point is, the higher the priority of the center point is for the total number of vehicles;
s52, creating a preselected class cluster center point queue, and sequentially storing k class cluster centers into the preselected class cluster center point queue after descending and sequencing according to the priority of the center point;
s53, selecting a cluster center with the highest priority of the center point from a preselected cluster center point queue, and updating the cluster center closest to the vehicle marking position to obtain a new cluster center;
s54, continuously carrying out position update on the cluster center point of each cluster of the cluster fog area based on the cluster fog index data at different moments, creating a position update sequence, and storing the position information of the cluster center updated each time into the position update sequence;
the location update sequence is expressed as tnew= { L1, L2, …, ln }, where Ln is the location information of the cluster center of the cluster fog area after the nth update.
6. The road fog prediction method based on edge calculation as claimed in claim 1, wherein the space-time position prediction model comprises a feature extraction module, a feature splicing module and a position prediction module, the space-time position prediction model is based on a position prediction algorithm of a space-time cyclic neural network, the space features and the time features are directly fed into the network, and the network is responsible for learning the internal representation of the network; the feature extraction module converts original feature sequence data into a group of vectors or matrixes representing data features, and extracts valuable feature information in the sequence; the feature splicing module utilizes the relation among different features to fuse a plurality of input features and output a more comprehensive feature representation; the position prediction module comprises a long-period memory network and a multi-head self-attention mechanism, wherein the long-period memory network is used for capturing long-period time dependence, and the multi-head self-attention mechanism is used for sensing key features in each sequence and performing self-adaptive adjustment, and position prediction is generated after capturing long-period dependence.
7. The method for predicting road fog based on edge calculation of claim 5, wherein S6 comprises:
s61, extracting relation features among the positions of the centers of k class clusters from the position updating sequence, creating a position feature sequence, and adding the relation features into the position feature sequence;
the relation featuresRepresenting the relation between the positions, said relation being characterized by +.>The calculation formula of (2) is as follows:
；
；
wherein ,is the position information of the cluster center of the m-th cluster fog area cluster, and is +.>，/>The SV is the center point priority, which is the position information of the cluster center of the jth cluster fog area cluster>TNew is the position more for the similarity of the positions between the centers of various clustersA new sequence;
s62, counting time intervals of position change when the cluster center is updated each time to form a time interval sequence G= {}；
S63, counting the occurrence frequency of each time interval in the time interval sequence, calculating the frequency of each time interval, and adding the occurrence frequency and the frequency of the time interval into the time interval sequence to obtain a time feature sequence;
；
；
；
；
wherein ,for every time interval +.>Frequency of->For the frequency of occurrence of time intervals, N is the frequency of all time intervals, < >>Is a time feature sequence;
s64, creating an influence characteristic sequence, and adding the current wind speed, time, position, wind direction, atmospheric humidity and air pressure inverse temperature into the influence characteristic sequence;
the influencing characteristic sequence is, wherein />For the wind speed of the wind,for time (I)>For the position of->Wind direction, the->Atmospheric humidity, < >>The current wind speed, time, position, wind direction, atmospheric humidity and air pressure reverse temperature are obtained by a central server;
s65, extracting features of the position feature sequence, the time feature sequence and the influence feature sequence, splicing and fusing various features to serve as reinforcing features, and inputting the reinforcing features into a space-time position prediction model to accurately predict the position of a central point at the next moment;
s66, creating a position prediction sequence, and adding central point position information obtained by each prediction into the position prediction sequence;
s67, marking the position information of the central point in the position prediction sequence on a mass fog density chart to obtain a prediction track, wherein the prediction track is the movement track of the mass fog;
s68, obtaining the change trend of the mist through the mist movement track.
8. The road bolus prediction method based on edge calculation according to claim 1, wherein S7 includes:
s71, the edge server sends the cluster fog change trend to a central server;
s72, marking an area to be covered by the mist as a mist area by the central server, and recording the position information of the mist area and the arrival time of the mist;
s73, the central server sends the position information of the group fog area and the arrival time of the group fog to a traffic management department, and synchronously gives vehicles in the area around the group fog to carry out group fog early warning.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310973181.5A CN116699730B (en) | 2023-08-04 | 2023-08-04 | Road group fog prediction method based on edge calculation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310973181.5A CN116699730B (en) | 2023-08-04 | 2023-08-04 | Road group fog prediction method based on edge calculation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116699730A true CN116699730A (en) | 2023-09-05 |
| CN116699730B CN116699730B (en) | 2023-10-27 |
Family
ID=87837818
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310973181.5A Active CN116699730B (en) | 2023-08-04 | 2023-08-04 | Road group fog prediction method based on edge calculation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116699730B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2891904A1 (en) * | 2014-01-07 | 2015-07-08 | ABB Technology AB | Solar irradiance forecasting |
| US20170109634A1 (en) * | 2014-12-26 | 2017-04-20 | Matthew Kuhns | System and method for predicting sunset vibrancy |
| KR101880616B1 (en) * | 2017-08-03 | 2018-07-23 | 한국해양과학기술원 | Method of sea fog prediction based on Sea surface winds and sea fog information from satellite |
| CN111553405A (en) * | 2020-04-24 | 2020-08-18 | 青岛杰瑞工控技术有限公司 | Clustering fog recognition algorithm based on pixel density K-means |
| CN213780406U (en) * | 2021-01-15 | 2021-07-23 | 招商局重庆交通科研设计院有限公司 | Group fog early warning elimination system |
| CN116153095A (en) * | 2023-04-20 | 2023-05-23 | 江西师范大学 | Expressway mass fog early warning method based on edge calculation |
-
2023
- 2023-08-04 CN CN202310973181.5A patent/CN116699730B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2891904A1 (en) * | 2014-01-07 | 2015-07-08 | ABB Technology AB | Solar irradiance forecasting |
| US20170109634A1 (en) * | 2014-12-26 | 2017-04-20 | Matthew Kuhns | System and method for predicting sunset vibrancy |
| KR101880616B1 (en) * | 2017-08-03 | 2018-07-23 | 한국해양과학기술원 | Method of sea fog prediction based on Sea surface winds and sea fog information from satellite |
| CN111553405A (en) * | 2020-04-24 | 2020-08-18 | 青岛杰瑞工控技术有限公司 | Clustering fog recognition algorithm based on pixel density K-means |
| CN213780406U (en) * | 2021-01-15 | 2021-07-23 | 招商局重庆交通科研设计院有限公司 | Group fog early warning elimination system |
| CN116153095A (en) * | 2023-04-20 | 2023-05-23 | 江西师范大学 | Expressway mass fog early warning method based on edge calculation |
Non-Patent Citations (2)
| Title |
|---|
| 刘志;: "车路协同技术在智慧高速建设中应用展望", 汽车工业研究, no. 01 * |
| 赵北辰;杨卓敏;张亚洲;: "基于视频图像的团雾检测技术浅析", 中国交通信息化, no. 05 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116699730B (en) | 2023-10-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10850731B2 (en) | Apparatus and method for automated traffic and driving pattern recognition and location-dependent measurement of absolute and/or relative risk probabilities for car accidents | |
| Yang et al. | Understanding and predicting travel time with spatio-temporal features of network traffic flow, weather and incidents | |
| CN110268454B (en) | Determining a customized safe speed for a vehicle | |
| Robinson et al. | Modeling urban link travel time with inductive loop detector data by using the k-NN method | |
| Li et al. | Identifying important variables for predicting travel time of freeway with non-recurrent congestion with neural networks | |
| CN114662797A (en) | Intelligent city flood early warning method and system based on Internet of things | |
| CN113436433B (en) | Efficient urban traffic outlier detection method | |
| CN108648445A (en) | Dynamic traffic Tendency Prediction method based on traffic big data | |
| CN106780266B (en) | Principal component contribution degree parameter-based accident hotspot internal characteristic analysis and driving guidance method | |
| Liu et al. | A UAV allocation method for traffic surveillance in sparse road network | |
| Jonnalagadda et al. | Forecasting atmospheric visibility using auto regressive recurrent neural network | |
| CN112884014A (en) | Traffic speed short-time prediction method based on road section topological structure classification | |
| CN106384507A (en) | Travel time real-time estimation method based on sparse detector | |
| CN116434523A (en) | Vehicle active safety control method and device based on constraint degree in information perception scene | |
| CN116597642A (en) | Traffic jam condition prediction method and system | |
| CN202373020U (en) | Intelligent coach operation system realizing device based on cloud computing platform | |
| CN114067566A (en) | Road accident black point discrimination and accident influence characteristic analysis method and system | |
| CN117238126A (en) | Traffic accident risk assessment method under continuous flow road scene | |
| CN116699730B (en) | Road group fog prediction method based on edge calculation | |
| CN116597649A (en) | Road network traffic flow deduction method based on expressway charging system | |
| Fahs et al. | Traffic congestion prediction based on multivariate modelling and neural networks regressions | |
| Samal et al. | Speedpro: A predictive multi-model approach for urban traffic speed estimation | |
| CN114139984A (en) | Urban traffic accident risk prediction method based on flow and accident collaborative perception | |
| Šemanjski | Analysed potential of big data and supervised machine learning techniques in effectively forecasting travel times from fused data | |
| Rouphail et al. | Guidance for Field and Sensor-Based Measurement of HCM and Simulation Performance Measures |
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 |