CN112797997B - Emergency path planning architecture and method based on grid road network - Google Patents
Emergency path planning architecture and method based on grid road network Download PDFInfo
- Publication number
- CN112797997B CN112797997B CN202011509992.2A CN202011509992A CN112797997B CN 112797997 B CN112797997 B CN 112797997B CN 202011509992 A CN202011509992 A CN 202011509992A CN 112797997 B CN112797997 B CN 112797997B
- Authority
- CN
- China
- Prior art keywords
- grid
- road network
- road
- data
- vehicle
- 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
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/26—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
- G01C21/34—Route searching; Route guidance
- G01C21/3407—Route searching; Route guidance specially adapted for specific applications
- G01C21/343—Calculating itineraries, i.e. routes leading from a starting point to a series of categorical destinations using a global route restraint, round trips, touristic trips
-
- 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
- Y02A30/00—Adapting or protecting infrastructure or their operation
- Y02A30/60—Planning or developing urban green infrastructure
Abstract
The invention discloses an emergency path planning architecture and a method based on a grid road network, wherein the architecture is divided into three parts, namely: the system comprises a data modeling layer, a road network preprocessing layer and a path planning layer. The method comprises the steps of modeling the existing related road network data and the trajectory data of the vehicles, rasterizing the original road network, assigning values according to the traffic capacity of the roads in the grids of the original road network, constructing a road network structure with space-time characteristics, sequencing each grid through links among the roads, and finally effectively evacuating the vehicles in an emergency area through the roads with the maximum traffic capacity. The vehicle can run out of the emergency area in the shortest time, so that the vehicle in the emergency area can be quickly evacuated to a safe area, and the rescue vehicle can quickly run into the emergency area to be effectively rescued.
Description
Technical Field
The invention belongs to the field of emergency disaster management, and relates to an integrated historical vehicle data monitoring result architecture and an emergency path planning method of a grid road network based on the integrated historical vehicle data monitoring result architecture.
Background
In the face of earthquake disasters, the traditional emergency evacuation technology usually depends on past data and experience support, lacks flexibility and relevance, and has difficulty in coping with increasingly variable emergencies. In the context of information data, emergency path planning is transitioning from the traditional empirical mode to a multivariate analysis mode driven by data. Urban major disasters often contain large amounts of heterogeneous data. Such as traffic data, meteorological data, geographic data, etc. These massive data growth, frequent and fast information interaction modes provide opportunities and challenges for traditional emergency management. On one hand, the emergency path planning strategy is required to make macro decisions such as evacuation and evasion area selection, evacuation route selection and the like on the basis of mass data analysis, and plan road condition characteristics to be planned to reasonably arrange micro decisions for rapid and accurate directional planning, but the existing data analysis methods, such as machine learning and the like, and emergency evacuation strategies are difficult to meet the requirements. And when a disaster happens, the vehicle is selected as an evacuation tool, so that the convenience is realized, and the holding amount of the vehicle is larger for people in urban life. Some current research focuses on planning evacuation paths by investigating various factors, including driver characteristics, evacuation path characteristics, etc. There have been some studies on path planning by means of advance prediction of road traffic flow, but none of these methods take into account the emergency of an emergency event, i.e. a road congestion condition that may accompany a period of time after an event occurs, and the emergency event is not regularly followed. Therefore, a real-time path planning technology facing emergency situations is provided.
The invention provides an urban road path planning technology after an earthquake disaster. According to the method, the initial urban road traffic flow condition can be obtained by backtracking the historical track data, and the original road network is subjected to grid partitioning, so that the vehicles can be managed and organized conveniently to be effectively evacuated when an emergency occurs, the route search area is reduced, and the route planning time is shortened. According to the method, based on data and analysis technology and combined with characteristics of multiple targets, multiple factors and the like in the emergency evacuation process, researches are carried out on modeling of heterogeneous characteristic data of the emergency disaster, vector grid division of a space-time characteristic road network, vehicle evacuation and evasion routes and the like according to actual characteristic conditions of emergency path planning after the earthquake disaster. The method comprises the steps of firstly extracting vehicle historical tracks, matching the vehicle historical tracks to an actual road network, extracting and analyzing the vehicle tracks of different roads at different time, and restoring the road network condition of the roads by adopting a hierarchical backtracking method and a machine learning method. And establishing a space-time road network association model according to the road network condition and the road network characteristic data, carrying out reasonable situation road network path evacuation navigation and dynamic evasion route planning according to the divided space-time vector rasterization map, and further realizing accurate and rapid evacuation of vehicles according to the traffic capacity of roads. The method provides powerful support and help for emergency evacuation navigation and avoidance route planning of urban vehicles after the earthquake.
In conclusion, the invention carries out backtracking modeling on the existing vehicle historical track data to form a road network space-time dynamic graph with space-time characteristics. The road network map is divided into areas with equal size by performing rasterization division at equal intervals based on the map, and the types, the number, the grades and the like of intersections contained in each area are different. The divided grids are sorted for commuting capacity through a sorting algorithm, importance sorting and assignment are carried out on each grid, and finally vehicles in an emergency area are effectively evacuated through a path planning algorithm based on a grid network.
Disclosure of Invention
The invention provides an emergency path planning architecture and method based on a grid road network. Fig. 1 is a general architecture of vehicle evacuation technology for emergency events. The framework is divided into three parts, from bottom to top: the system comprises a data modeling layer, a road network preprocessing layer and a path planning layer. Firstly, the lowest layer obtains various data of a road network, for example; trajectory data, microwave data (traffic flow conditions obtained by a road network monitoring camera), road network data, weather data, and the like. After processing and integrating the data, mapping the data to a two-dimensional plane map. The middle layer is a road network pretreatment layer, the road network is divided into a plurality of grids with equal intervals after being subjected to rasterization, a plurality of unequal intersections and roads are respectively stored in each grid, the whole road network is subjected to grid sequencing through a sequencing algorithm, and a characteristic value is given to the whole road network, and the characteristic value reflects the current road network traffic capacity of the grids. The uppermost layer is a path planning layer which is processed by the first two parts.
The invention designs an emergency path planning method based on a grid road network, which comprises two path planning methods, namely an evacuation path planning method based on a grid map, and realizes reasonable evacuation of vehicles in an emergency area.
(1) Data modeling layer
The layer can complement the data of the missing GPS points by using a hierarchical backtracking method. The hierarchical backtracking method is to perform superposition processing on the traffic flow of the same geographical position at different times. For example, traffic on a road section in the same day changes with the passage of time, two peaks appear in the morning and evening, and traffic flow is in a stable state in the rest of the day. Through the conclusion of the regularity, the traffic flow of the same road at different time is superposed to form a more perfect overall traffic flow condition.
The urban road network can be effectively modeled by utilizing the historical track data of vehicles to carry out road flow reduction, the traditional road network modeling only simply considers the distance G (V, E) between intersections, wherein w: = distance (V) i ,v j ). Other factors such as road network speed, road grade, number of lanes, and weather are not taken into account. These factors are collectively referred to as situational information, which affects the travel time of the vehicle. For example, if an area is exposed to heavy rain, the speed of the vehicle passing through the area will be much lower than the speed of other vehicles that are not affected by the heavy rain. Therefore, the driving flow of the road network vehicles is restored according to the situation information, a situation vector space-time road network graph is established, and the structural change of the road network after disasters is also considered.
Firstly, processing the taxi track data of six months in different time periods, and separating the data through a given time interval, wherein the traffic flow operation conditions of the whole road network in different time periods are obtained. Next, by performing road network matching processing on the existing GPS data, since the track points may be deviated to both sides of the road, nearest neighbor matching work needs to be performed on the existing data. And finally, integrating the longitude and latitude points of each road in the city, and restoring each GPS data to each corresponding road. And backtracking the traffic flow condition of each road in the city through the six-month track data.
(2) Road network pretreatment layer
By backtracking the whole traffic flow of the urban road, the change of the traffic flow of each road of the city can be grasped, and the change is changed along with the time. Therefore, each region in the city has a semantic tag to define the current condition of the region according to different time. For example, during the peak hours, the main traffic trunk in the city may be congested to different degrees, and this phenomenon does not only occur at one intersection, but usually occurs at a plurality of intersections in succession. And these intersections are necessarily geographically connected intersections of several consecutive adjacent roads. The area made up of these intersections can be defined as a congested area.
The equal interval division grid enables the road network data to be divided into a plurality of region sub-blocks through reasonable equal interval size, and the condition of the whole road network intersection node can be effectively judged through management of the region sub-blocks. The number of intersections and the access degree contained in each area are represented as the communication capacity of the area, and the areas with large number of intersections and access degrees obtain higher ranking values, which means that the probability of guiding to other grids when the vehicle drives into the areas is higher. PageRank will be improved and used as a ranking algorithm, called Grid PageRank.
The entire road network is mapped into a graph structure.
Where GPR is the ordered value of the corresponding grid, grid2 will be recommended to the user if GPR (grid 1) < GPR (grid 2) because its value is higher. k is the number of meshes directly connecting vertices, meaning that vertices in a mesh are not necessarily connected to adjacent meshes. E is a damping coefficient, and N is the number of grids with intersections in the whole road network.
Wherein R is i Is the GPR value of the ith gridI represents the ID of the mesh, and com is defined as the set of different vertices in a mesh pointing to the same mesh. L is the number of lanes of the road.
Wherein grid i Represents the ith mesh and V is the total number of vertices in the entire mesh. V gridi Is the total number of vertices in the ith mesh. Fig. 2 shows the effect of rasterization on the entire road network. Fig. 3 shows the connectivity division of the road network.
(3) Path planning layer
The path planning layer is divided into two parts, wherein one part is an emergency evacuation method of the vehicle, and the other part is a detour path planning method of the vehicle. The selection of evacuation routes is a key problem in vehicle evacuation, and the planned result directly influences the evacuation effect. Vehicle evacuation is a multipoint-to-multipoint route planning problem, and on one hand, a smooth road section needs to be selected for evacuation, and on the other hand, the vehicle evacuation needs to be uniformly distributed to the road section with strong traffic capacity. The route condition can be backtracked and modeled by historical road network data, namely the state of a road section is obtained by vector extraction of monitored vehicle data. The magnitude of the density state of the vehicles in the road segment reflects the road segment congestion condition to some extent. The higher the vehicle density of the road section is, the more the traffic is congested; conversely, the smaller the vehicle density of the road section is, the better the traffic condition of the road section is.
And carrying out evacuation according to the importance sequence of the divided grid regions by the route evacuation, and calculating a reasonable evacuation route by the vehicle according to the comprehensive weighting of the multi-state potential information. When calculating the route, due to the characteristic that situation information changes constantly, a planning algorithm is required to plan the route and reach a safety area in a manner that the time factor is considered, namely the time of the vehicle reaching a certain point is as accurate as possible. Fig. 4 depicts a grid network based evacuation path indication map.
The path planning algorithm utilizes a Grid Bidirectional Dijkstra algorithm (GBD), which is obtained by improving according to the Bidirectional Dijkstra algorithm.
FIG. 5 depicts a flowchart of the overall framework operation. The emergency path planning framework steps based on the rasterized road network are described as follows:
s1, reading track data and road network data; analyzing each piece of track data, extracting longitude and latitude coordinates of each piece of information, matching the data to a corresponding road network according to the road network coordinates, and acquiring the speed of the road network from the track data. The road network speed is obtained by a historical backtracking method, historical track data of six months are expanded according to days, backtracking is carried out on each road according to time, and superposition calculation is carried out according to the traffic flow of each day so as to obtain the traffic flow of each road according to different times.
S2, acquiring a grid ID (current space-time scene) where the emergency event is located; the longitude and latitude coordinates of the intersection are mapped into a two-dimensional rectangular coordinate system, so that the whole road network data is converted into a two-dimensional coordinate format, the converted intersection becomes a point in a plane, and then the intersection point which is closest to the origin of the coordinates in the whole road network is found through the Euclidean distance to replace the origin, so that the whole road network coordinates are translated. And carrying out equidistant division on the grids according to the translated node coordinates.
S3, sorting and assigning values to each Grid after the grids are divided (by using Grid PageRank); firstly, calculating the number of nodes of each Grid (as the grids are divided at equal intervals, the number of intersections contained in the grids is different, but the calculation of the Grid traffic capacity is not influenced), then calculating the number of road lanes connected to other grids in each Grid, and calculating the traffic capacity of each Grid according to the number of intersections and the number of lanes in the grids, wherein each Grid is ranked by using a Grid PageRank algorithm, the algorithm is obtained by improving the PageRank. This process is a sort. And finally, after multiple iterations are carried out, the ranking value of each Grid is compared with the value of the Grid in the last time, and the value is smaller than a threshold value, the Grid PageRank algorithm is ended, and the final ranking value of each Grid is obtained, wherein the value represents the traffic capacity value of the current Grid.
And S4, planning the path of the vehicle in the emergency area through GBD, and selecting the road with larger current traffic capacity to provide the vehicle. And the grids can obtain corresponding sequencing values after sequencing, and emergency path planning is carried out on the vehicle through a GBD algorithm. The GBD algorithm is divided into two parts: firstly, a path planning is carried out between each grid, and secondly, the path planning is carried out in each grid. And recommending the grids which are communicated with each other around the grids and have strong trafficability by using the grid sequencing values as judgment bases among the grids, wherein the recommended grids are provided with a label which indicates whether the grids are recommended or not, and the setting is to ensure that the vehicles do not walk on the repeated grids. A bidirectional Dijkstra algorithm is adopted in the grids, the starting point is the node which leads the previous grid to be connected with the grid, and the terminal point is the last node which leads the next grid, so that the path planning time is accelerated. The vehicle is eventually provided with a set of grid sets plus all road sets within each grid.
The invention creatively provides that the existing related road network data and the trajectory data of the vehicles are modeled, the original road network is rasterized, the traffic capacity of the roads in the grids is assigned, a road network structure with space-time characteristics is constructed, each grid is sequenced through the links among the roads, and finally the vehicles in the emergency area can be effectively evacuated through the road with the maximum traffic capacity. The vehicle can run out of the emergency area in the shortest time, so that the vehicle in the emergency area can be quickly evacuated to a safe area, and the rescue vehicle can quickly run into the emergency area to be effectively rescued.
Drawings
Fig. 1 is an overall architecture of emergency evacuation based on a rasterized road network;
FIG. 2 is a rasterized road network effect graph;
FIG. 3 is a schematic diagram of a grid road network traffic capacity;
FIG. 4 is an emergency routing graph based on a rasterized road network;
fig. 5 is a flow chart of emergency evacuation based on a rasterized road network;
FIG. 6 urban road network grid results;
FIG. 7 rasterization effects of the sorting algorithm.
Detailed Description
The invention develops a simulation program by using Java language, collects a large amount of track data and simulates the traffic flow condition of an emergency area, and verifies the effectiveness of the evacuation path planning architecture and the method. And the equal-distance division effect of the road network is simulated through experiments, historical track data is compared, and vehicle evacuation time of an emergency area is simulated, so that the effectiveness of the emergency evacuation path method based on the rasterized road network is verified. The disaster scene is the vehicle evacuation process after the urban earthquake. The destruction degree and magnitude of earthquake are related to the released energy, and the earthquake is divided into five types by the geophysical institute of Chinese academy of sciences: micro-earthquakes (generally without induction), light earthquakes (generally with induction), strong earthquakes (also classified into harmless strong earthquakes, harmful strong earthquakes and destructive strong earthquakes), strong earthquakes (major destructive earthquakes) and major disastrous earthquakes (destructive earthquakes). The condition of a strong earthquake is concerned, and the corresponding magnitude is 4.5.
The invention firstly divides the whole road network into six interval grids, namely grids of 18 × 18, 36 × 36, 72 × 72, 144 × 144, 288 × 288 and 576 × 576. The efficiency of the sequencing algorithm under the six divisions and the efficiency of the emergency path planning algorithm are compared, and then the results are provided for vehicles (since the vehicles in the road network follow the first-in first-out principle, namely, after the vehicles A and the vehicles B respectively drive into a uniform road from the same intersection at intervals of 5 minutes and drive for a period of time, when the same intersection is turned out, the default vehicle A can drive away from the road first, namely, the driver is considered to drive according to the traffic rules). Fig. 6 shows the rasterization effect of the urban road network. It can be seen from fig. 6 that different grid sizes will have different numbers of intersection nodes (black dots in the figure) contained in each grid.
The experiment is carried out three times, the test is respectively carried out under the lengths of 6.5km, 12.6km and 21.3km, and the path efficiency of the bidirectional Dijkstra algorithm of the grids under different grid conditions is compared during the test. The path planning results of the GBD algorithm for a path length of 6.5km are shown in table 1. It can be seen from the table that as the number of grids increases, the number of searched grids also increases. However, the number of the search nodes is continuously reduced, because the number of the nodes in each grid is continuously reduced, when the algorithm searches for the nodes, the algorithm stops searching for the nodes connected to the external grid when the boundary nodes of the grid are searched, so that the smaller the grid division is, the higher the search efficiency in each grid is. The same results were obtained in tables 2 and 3.
TABLE 1 GBD algorithm operating results at a path length of 6.5km
| Grid map | Number of search grids | Number of search nodes | Run time (ms) |
| 18×18 | 8 | 2358 | 42.6 |
| 36×36 | 15 | 1904 | 40.3 |
| 72×72 | 22 | 1128 | 32.5 |
| 144×144 | 43 | 951 | 21.8 |
| 288×288 | 76 | 783 | 18.2 |
| 576×576 | 131 | 374 | 26.4 |
TABLE 2 GBD algorithm operating results at a path length of 12.6km
| Grid map | Number of search grids | Number of search nodes | Run time (ms) |
| 18×18 | 8 | 4581 | 73.8 |
| 36×36 | 15 | 3358 | 63.6 |
| 72×72 | 29 | 2874 | 48.2 |
| 144×144 | 50 | 1927 | 36.4 |
| 288×288 | 120 | 1029 | 26.8 |
| 576×576 | 245 | 712 | 38.5 |
TABLE 3 GBD Algorithm operating results at Path Length 21.3km
| Grid map | Number of search grids | Number of search nodes | Run time (ms) |
| 18×18 | 8 | 11572 | 179.5 |
| 36×36 | 22 | 7823 | 125.7 |
| 72×72 | 50 | 5629 | 93.8 |
| 144×144 | 99 | 3041 | 72.1 |
| 288×288 | 213 | 2508 | 60.9 |
| 576×576 | 857 | 2196 | 82.3 |
And the Grid PageRank sorting algorithm is improved based on PageRank. Since each grid can be seen as a whole, the roads between nodes of the grid inside and outside grids can be seen as a link, the weight of which is calculated according to the peer ability of the roads, and table 4 shows the sorting results based on 144 × 144 and 288 × 288 grids, which will change with time. The 5 min, 15 min, 30 min and 45 min will be selected as time windows, and the Grid PageRank algorithm will rank all the grids present one by one when the ranking interval reaches the time window. FIG. 7 shows the sorting efficiency of Grid PageRank under different Grid conditions, and it can be seen that the number of iterations becomes gradually gentle after reaching 25 iterations.
Table 4 partial calculation results of GridPageRank under different rasterization results
| 144×144 | |
288×288 | Rank value |
| 14-46 | 0.02403406605964532 | 28-91 | 0.00063961350785751 |
| 20-48 | 0.01689090111204018 | 57-103 | 0.00035618516562940 |
| 19-47 | 0.01265170851251942 | 138-61 | 0.00031077176150319 |
| 39-91 | 0.00950338786211769 | 160-35 | 0.00030765424485485 |
| 59-86 | 0.00879868442408170 | 161-109 | 0.00030728136329326 |
| 15-46 | 0.00874443282679572 | 136-126 | 0.00030685079492260 |
| 29-52 | 0.00846635288234325 | 171-106 | 0.00028400636699139 |
| 21-49 | 0.00846595871111033 | 103-193 | 0.00028075261107912 |
The results in tables 5, 6, and 7 show the evacuation effect of the vehicle at different time windows. Evacuation is simulated by simulating 100 vehicles, 200 vehicles and 300 vehicles (evacuation process is to evacuate from the grid to the safety zone designated by the simulation, the safety zone is selected to be a large parking lot, and all the safety zones set in the simulation experiment are the same). The evacuation time is in minutes.
TABLE 5.100 evacuation time in different time windows
| Grid map | Window for 5 minutes | Window for 15 | Window | 30 minutes | Window time 45 |
| 18×18 | 72.6 | 68.2 | 69.5 | 71.8 | |
| 36×36 | 55.8 | 53.7 | 59.4 | 62.2 | |
| 72×72 | 46.1 | 42.5 | 51.8 | 51.4 | |
| 144×144 | 32.6 | 29.1 | 36.7 | 38.4 | |
| 288×288 | 25.5 | 19.2 | 23.7 | 30.6 | |
| 576×576 | 38.2 | 36.6 | 39.4 | 42.3 |
TABLE 6.200 evacuation time in different time windows
TABLE 7.300 evacuation time in different time windows
| Grid map | Window for 5 minutes | Window for 15 minutes | Window for 30 minutes | Window for 45 |
| 18×18 | 83.2 | 85.6 | 86.7 | 89.2 |
| 36×36 | 80.1 | 78.1 | 76.4 | 82.5 |
| 72×72 | 76.1 | 72.5 | 73.1 | 71.4 |
| 144×144 | 67.6 | 63.5 | 65.2 | 68.7 |
| 288×288 | 58.2 | 53.3 | 60.8 | 62.2 |
| 576×576 | 68.8 | 67.1 | 69.2 | 73.9 |
As shown in tables 5, 6, and 7, the evacuation time becomes shorter as the grid pitch of the grid becomes smaller. But when the evacuation time of the vehicle increases in the case of the grid 576 x 576 compared to 288 x 288, that is because the increase in the number of grids results in the algorithm needing to explore more grids and the value of the grid changes when the time window arrives.
Through three times of vehicle evacuation test experiments, the evacuation time is slightly increased along with the increase of the vehicles, because the load of the lane is constant, and the vehicles sequentially run through the road according to the principle of first-in first-out of the road. And according to the different number of the grids, the smaller the number of the grids, the larger the space in the grids, the more intersection nodes are included, so that the number of intersections in the 18 × 18 grids is too large, the time consumption of the algorithm in searching the path is too long, and the traffic flow cannot be effectively judged, so that the vehicles can travel to the congested area. The same reasoning is in the 36 × 36, 72 × 72 and 144 × 144 rasterized road networks. The 288 × 288 grid network has a good evacuation effect because the number of intersections included in the interior of the grid facilitates the area search by the route planning algorithm. However, the reason why the evacuation time of 576 × 576 increases is that the number of grids is too large, the number of intersections included in each grid is small, and the search for each grid by the algorithm is equivalent to the search for a plurality of intersections, so that the significance of the grid is lost.
The algorithm provided by the invention can be used for emergency evacuation management of urban road vehicles, the emergency path planning algorithm based on the rasterized road network has ideal evacuation effect on the vehicles under 288 × 288 grids, the higher the value is, the stronger the traffic capacity of the grids is, and the vehicles are recommended according to the given value, so that the evacuation efficiency of the vehicles is improved.
Claims (2)
1. An emergency path planning architecture based on a grid network is characterized in that: the framework is divided into three parts, from bottom to top: the system comprises a data modeling layer, a road network preprocessing layer and a path planning layer; firstly, acquiring various data of a road network at the lowest layer, processing and integrating the various data, and mapping the various data to a two-dimensional plane map; the middle layer is a road network pretreatment layer, the road network is divided into a plurality of grids with equal intervals after being subjected to rasterization, a plurality of unequal intersections and roads are respectively stored in each grid, the whole road network is subjected to grid sequencing through a sequencing algorithm, and a characteristic value is given to the whole road network, wherein the characteristic value reflects the current road network traffic capacity of the grids; the uppermost layer is a path planning layer;
in the data modeling layer, the data of the missing GPS points can be completed by utilizing a hierarchical backtracking method; the hierarchical backtracking method is to perform superposition processing on the traffic flow of the same geographical position at different times; overlapping the traffic flow of the same road at different time to form the whole traffic flow condition;
the method comprises the steps of utilizing historical track data of vehicles to carry out road flow reduction to effectively model an urban road network, firstly carrying out time-interval processing on six-month taxi track data, separating the data through a given time interval, and obtaining the traffic flow operation conditions of the whole road network at different time intervals; next, by performing road network matching processing on the existing GPS data, since the track points may deviate to both sides of the road, nearest neighbor matching work needs to be performed on the existing data; finally, integrating longitude and latitude points of each road in the city, and restoring each GPS data to each corresponding road; backtracking the traffic flow condition of each road in the city through six months of track data;
in the road network preprocessing layer, by backtracking the whole traffic flow of urban roads, each region in the city has a semantic tag to define the current condition of the region according to different time; the equal interval division grid enables road network data to be divided into a plurality of region sub-blocks through reasonable equal interval size, and the condition of the whole road network intersection node can be judged through management of the region sub-blocks; the number and the access degree of intersections in each area are represented as the communication capacity of the area, and areas with large number of intersections and access degrees obtain higher ranking values, which means that the probability of guiding to other grids is higher when vehicles drive into the areas; the PageRank is improved and used as a sorting algorithm, and is called Grid PageRank; mapping the whole road network into a graph structure;
in the path planning layer, the emergency evacuation method of the vehicle is divided into two parts, one part is the detour path planning method of the vehicle, and the other part is the detour path planning method of the vehicle; the selection of evacuation routes is a key problem in vehicle evacuation, and the planned result directly influences the evacuation effect; vehicle evacuation is a multipoint-to-multipoint path planning problem, on one hand, a smooth road section is selected for evacuation, and on the other hand, the vehicle evacuation is uniformly distributed to the road section with stronger traffic capacity; the route condition is backtracked and modeled by historical road network data, namely the state of a road section is obtained by vector extraction of monitored vehicle data; the density state of the vehicles in the road section reflects the road section congestion condition to a certain extent; the higher the vehicle density of the road section is, the more the traffic is congested; on the contrary, the smaller the vehicle density of the road section is, the better the traffic condition of the road section is; the route evacuation is carried out according to the importance ranking of the divided grid regions, and the evacuation route is calculated by the vehicles according to the comprehensive weighting of the multi-state potential information; when the path is calculated, due to the characteristic that situation information is changed continuously, a planning algorithm is required to consider time factors, namely, a route is planned in a mode that the time of the vehicle reaching a certain point is accurate and the vehicle reaches a safety area; the path planning algorithm utilizes a grid bidirectional Dijkstra algorithm, and the algorithm is obtained by improving the bidirectional Dijkstra algorithm.
2. A method for emergency path planning based on a raster network using the architecture of claim 1, wherein: comprises the following steps of (a) carrying out,
s1, reading track data and road network data; analyzing each piece of track data, extracting longitude and latitude coordinates of each piece of information, matching the data to a corresponding road network according to the road network coordinates, and acquiring the speed of the road network from the track data; the road network speed is obtained by a historical backtracking method, historical track data of six months are developed according to the days, backtracking is carried out on each road according to the time, and the traffic flow of each road according to different times is obtained by carrying out superposition calculation according to the traffic flow of each day;
s2, acquiring a grid ID where the emergency event is located; mapping the longitude and latitude coordinates of the intersection into a two-dimensional rectangular coordinate system to convert the whole road network data into a two-dimensional coordinate format, changing the converted intersection into a point in a plane, and then finding an intersection point which is closest to the origin of the coordinates in the whole road network through Euclidean distance to replace the origin, so that the whole road network coordinates are translated; dividing the grids at equal intervals according to the translated node coordinates;
s3, sorting and assigning values to each grid after the grids are divided; firstly, calculating the number of nodes of each Grid, then calculating the number of road lanes connected to other grids in each Grid, calculating the traffic capacity of each Grid according to the number of intersections and the number of lanes in each Grid, and sequencing each Grid by using a Grid PageRank algorithm, wherein the algorithm is obtained by improving the PageRank, and the idea is that each Grid is initially pre-assigned, and is sequentially assigned according to the weight linked to the roads of other grids in the Grid; the process is a sequencing; finally, after multiple iterations are carried out, the ranking value of each Grid is compared with the value of the Grid in the last time, and the value is smaller than a threshold value, the Grid PageRank algorithm is ended, and the final ranking value of each Grid is obtained, wherein the value represents the traffic capacity value of the current Grid;
s4, planning a path of the vehicle in the emergency area through GBD, and selecting a road with larger current traffic capacity to provide for the vehicle; the grids can obtain corresponding sequencing values after sequencing, and emergency path planning is carried out on the vehicles through a GBD algorithm; the GBD algorithm is divided into two parts: firstly, performing path planning among each grid, and secondly, performing path planning in each grid; recommending grids which are communicated with the periphery of the grid and have strong trafficability by using the grid sorting value as a judgment basis among the grids, wherein the recommended grid is provided with a label which indicates whether the recommended grid is recommended or not, and the setting is to ensure that the vehicle does not walk on the repeated grid; a bidirectional Dijkstra algorithm is adopted in the grids, the starting point is the node leading the previous grid to the grid and connected with the grid, and the end point is the last node leading the next grid, so that the path planning time is shortened; the vehicle is eventually provided with a set of grids set plus all the roads set within each grid.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011509992.2A CN112797997B (en) | 2020-12-19 | 2020-12-19 | Emergency path planning architecture and method based on grid road network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011509992.2A CN112797997B (en) | 2020-12-19 | 2020-12-19 | Emergency path planning architecture and method based on grid road network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112797997A CN112797997A (en) | 2021-05-14 |
| CN112797997B true CN112797997B (en) | 2022-12-16 |
Family
ID=75807044
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011509992.2A Active CN112797997B (en) | 2020-12-19 | 2020-12-19 | Emergency path planning architecture and method based on grid road network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112797997B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116451507B (en) * | 2023-06-13 | 2023-11-10 | 浙江大学 | Typhoon track path simulation method based on non-uniform grid division |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1948913A (en) * | 2006-08-25 | 2007-04-18 | 北京航空航天大学 | Heuristic path culculating method for treating large scale floating vehicle data |
| CN107144286A (en) * | 2016-03-01 | 2017-09-08 | 阿里巴巴集团控股有限公司 | Air navigation aid and device |
| CN107195177A (en) * | 2016-03-09 | 2017-09-22 | 中国科学院深圳先进技术研究院 | Based on Forecasting Methodology of the distributed memory Computational frame to city traffic road condition |
| CN111710162A (en) * | 2020-07-07 | 2020-09-25 | 深圳市数字城市工程研究中心 | Urban road network traffic operation condition monitoring method and system |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10394234B2 (en) * | 2017-12-18 | 2019-08-27 | The Boeing Company | Multi-sensor safe path system for autonomous vehicles |
-
2020
- 2020-12-19 CN CN202011509992.2A patent/CN112797997B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1948913A (en) * | 2006-08-25 | 2007-04-18 | 北京航空航天大学 | Heuristic path culculating method for treating large scale floating vehicle data |
| CN107144286A (en) * | 2016-03-01 | 2017-09-08 | 阿里巴巴集团控股有限公司 | Air navigation aid and device |
| CN107195177A (en) * | 2016-03-09 | 2017-09-22 | 中国科学院深圳先进技术研究院 | Based on Forecasting Methodology of the distributed memory Computational frame to city traffic road condition |
| CN111710162A (en) * | 2020-07-07 | 2020-09-25 | 深圳市数字城市工程研究中心 | Urban road network traffic operation condition monitoring method and system |
Non-Patent Citations (1)
| Title |
|---|
| A Novel Urban Emergency Path Planning Method Based on Vector Grid Map;BOWEN YANG等;《IEEE ACCESS》;20200902;第154338-154352页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112797997A (en) | 2021-05-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Cova et al. | Modelling community evacuation vulnerability using GIS | |
| CN105070042A (en) | Modeling method of traffic prediction | |
| CN113065074B (en) | Track destination prediction method based on knowledge graph and self-attention mechanism | |
| CN109959388A (en) | A kind of intelligent transportation fining paths planning method based on grid extended model | |
| Shen et al. | Optimized public parking location modelling for green intelligent transportation system using genetic algorithms | |
| CN112797995B (en) | Vehicle emergency navigation method with space-time characteristic situation information | |
| Şen et al. | A GIS approach to fire station location selection | |
| CN112784000B (en) | Passenger searching method based on taxi track data | |
| CN112801399B (en) | Path generation method and device, terminal equipment and storage medium | |
| CN112629533A (en) | Refined path planning method based on road network rasterization road traffic flow prediction | |
| CN113804209B (en) | High-precision long-distance off-road path planning method for quadrangle grid | |
| CN106780739A (en) | A kind of intelligent substation patrol three-dimension GIS system method for building up | |
| Yang et al. | A novel urban emergency path planning method based on vector grid map | |
| CN112797997B (en) | Emergency path planning architecture and method based on grid road network | |
| CN112991743A (en) | Real-time traffic risk AI prediction method based on driving path and system thereof | |
| Olayode et al. | Traffic flow prediction at signalized road intersections: a case of Markov chain and artificial neural network model | |
| Halim et al. | On finding optimum commuting path in a road network: a computational approach for smart city traveling | |
| Cosido et al. | Automatic calculation of bicycle routes by combining meta-heuristics and GIS techniques within the framework of smart cities | |
| Otsuka et al. | Bayesian optimization for crowd traffic control using multi-agent simulation | |
| CN115355922A (en) | Travel path planning method and system based on improved ant colony algorithm | |
| US20220180745A1 (en) | Systems and methods for system optimal traffic routing | |
| CN114048913A (en) | Rescue planning method based on particle swarm algorithm and genetic algorithm mixing | |
| Luo et al. | A new memetic algorithm to solve the stochastic tsp | |
| Cai et al. | Multi-vehicles dynamic navigating method for large-scale event crowd evacuations | |
| Khodadadi et al. | Improve traffic management in the vehicular ad hoc networks by combining ant colony algorithm and fuzzy system |
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 |