Unmanned vehicle rescue point deployment method, device, equipment and medium

Abstract

The invention relates to the technical field of intelligent traffic, and discloses an unmanned online car-hailing rescue point deployment method, and the method comprises the steps: S10, obtaining a data set for constructing a cost flow network model; s20, splitting road nodes; s30, obtaining coordinates of the road nodes according to the road nodes, and carrying out clustering analysis on the road nodes, the operation data and the road test data to obtain the number of accidents around the road nodes and the average passing time between the adjacent road nodes; s40, a cost flow network model is constructed according to the data set, the number of accidents around the road nodes and the average passing time between the adjacent road nodes, and the position of each rescue point is determined; s50, based on the cost flow network model, solving the minimum cost and maximum flow meeting rescue point flow limiting conditions; and S60, outputting the transport capacity of each rescue point according to the minimum cost and the maximum flow.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent transportation, and particularly to a method, device, electronic device and medium for deploying rescue points of driverless vehicles. Background Art

[0002] With the rapid development of driverless technology and the growing scale of the online car-hailing market, driverless online car-hailing has become an important trend for future travel. However, due to the immaturity of technology and the complexity of road conditions, driverless online car-hailing is prone to accidents or failures during operation. To ensure the safety of passengers and improve service quality, timely and effective rescue response is particularly important.

[0003] Currently, the existing deployment scheme for rescue points of driverless online car-hailing usually conducts grid management, divides the urban area into multiple grids, and evenly sets a certain number of rescue points and allocates transportation capacity in each grid. Such a management method cannot effectively and reasonably allocate transportation capacity because the accident locations cannot be evenly distributed. In high-incidence areas of accidents, the existing rescue points and transportation capacity may not meet the demand, resulting in too long response time; while in low-incidence areas of accidents, there may be a situation of resource waste. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to solve the problem of unreasonable deployment of rescue points for existing driverless vehicles.

[0005] To solve the above technical problem, the present invention provides a method for deploying rescue points of driverless vehicles, and the method includes:

[0006] S10. Obtain a data set for constructing a cost flow network model, where the data set includes map data, road test data, operation data, alternative rescue point information, and total transportation capacity, and the map data includes lane information;

[0007] S20. Split road nodes according to the lane information and a preset distance threshold;

[0008] S30. Obtain road node coordinates according to the road nodes, and perform clustering analysis on the road nodes, operation data, and road test data to obtain the number of accidents around the road nodes and the average passing time between adjacent road nodes;

[0009] S40. Construct a cost flow network model according to the data set, the number of accidents around the road nodes, and the average passing time between adjacent road nodes, and determine the location of each rescue point;

[0010] S50. Solve the minimum cost maximum flow that meets the flow limit condition of the rescue points based on the cost flow network model;

[0011] S60. Output the transport capacity of each rescue point according to the minimum cost maximum flow.

[0012] Furthermore, the map information includes lane structure, lane longitude and latitude coordinates, and lane connection relationship. The data set includes an accident coordinate set Acci, and the accident longitude and latitude coordinate set Acci is obtained through the road test data and the operation data.

[0013] Among them, the steps of obtaining the accident longitude and latitude coordinate set Acci specifically include:

[0014] S11. Set a time window as the previous D days.

[0015] S12. Obtain the vehicle trajectory information of the previous D days according to the data set, denoted as set Traj.

[0016] S13. Obtain the accident coordinate set within D days according to the data set. The accidents include: manual takeover, remote takeover, sudden braking and sharp steering wheel turning, and actual collision accidents, denoted as the accident longitude and latitude coordinate set Acci.

[0017] Furthermore, the steps of constructing the cost flow network model and determining the location of each rescue point in S40 include:

[0018] S41. Set a distance threshold Len and divide the road according to the distance threshold and the road forward direction to obtain road nodes v1, v2,..., v n ;

[0019] S42. According to the longitude and latitude coordinate distance formula

[0020]

[0021] Calculate the longitude and latitude coordinates of the road nodes is the latitude difference between two longitude and latitude coordinates, and Δλ = λ1 - λ2 is the longitude difference between two longitude and latitude coordinates;

[0022] S43. Obtain the topological relationship of the road nodes, denoted as successor(v) represents the set of all successor nodes of v;

[0023] S44. For the road node v ∈ {v1, v2,..., v n}, denote where dis(a, v) is the distance between the accident point a and the road node v, and the collision(v) is the radius around the road node v The number of accidents within the area, for adjacent road nodes v i , v j , the average travel time from v i to v j is statistically counted through the set Traj, denoted as time(v i , v j );

[0024] S45. According to the average travel time time(v i , v j ) and the number of accidents construct a cost flow network G = (V, A), where V is the set of nodes of network G and A is the set of edges of network G.

[0025] Furthermore, in S50, the solution of the minimum cost maximum flow that satisfies the rescue point flow limit condition based on the cost flow network model includes the following steps:

[0026] S51. Build a source point s and a sink point t and connect the source point s to each rescue point. For the edge (s, u i ), configure its capacity as the transport capacity l i of the rescue point u i , that is, cap(s, u i ) = l i , and configure its cost as the cost c i per vehicle at the rescue point, that is, cost(s, u i ) = c i ;

[0027] S52. Connect each rescue point u i to the nearest adjacent road node v j , that is, the vehicle at the rescue point u i will first depart to the road node v j . Set the capacity as cap(u i , v j ) = ∞, and the cost as the travel time time(u i to v j ) multiplied by the weight w i , that is, cost(u j , v time ) = w i ·time(u j . v time ) i . v j ;

[0028] S53. Connect each road node v i to all its successor nodes successor(v i), for edge (v i , v j ), set its capacity to cap(v i , v j )=∞, the cost is point v i To point v j The travel time time(v i , v j ) multiplied by the weight w time , that is, cost(v i , v j )=w time ·time(v i , v j );

[0029] S54, connect all road nodes to the sink t, for the edge (v i , t), set the capacity to point v i rescue needs i ), where need(v i )=collision(v i ) / T, T is the average response time for handling a single incident, that is, cap(v i , t) = need(v i ), set the cost cost(v i , t) = 0;

[0030] S55, using the cost flow algorithm to solve the minimum cost flow from s to t with a maximum flow not exceeding L: A→R;

[0031] The restrictions include:

[0032] The flow rate from the source does not exceed L, that is, ∑ 1≤i≤m flow(s,u i )≤L;

[0033] The flow on any edge does not exceed the capacity, that is, This ensures that:

[0034] Arrive at each rescue point i The flow rate will not exceed the capacity of the rescue point l i ;

[0035] Each road node v j The flow to the sink will not exceed the rescue demand of the road node need(v j );

[0036] For any node except the source and sink, the inflow and outflow are the same, that is,

[0037] S56, Configure the transport capacity of the rescue point based on the flow. For the rescue point u i , configure the transport capacity flow(s, u i ).

[0038] Furthermore, when obtaining the accident longitude and latitude coordinate set, it also includes classifying the accident types. The accident types include vehicle collisions and vehicle failures, and different weights are assigned to each accident type.

[0039] Furthermore, when solving the minimum cost maximum flow that meets the rescue point flow limit condition, the cooperation ability between rescue points is considered to maximize the overall rescue efficiency.

[0040] According to another aspect of the present invention, there is provided a deployment device for a rescue point of an autonomous vehicle. The device includes:

[0041] A data acquisition module for obtaining a data set for constructing a cost flow network model. The data set includes map data, road test data, operation data, alternative rescue point information, and total transport capacity. The map data includes lane information;

[0042] A road node splitting module for splitting road nodes according to the lane information and a preset distance threshold;

[0043] A data analysis module for obtaining road node coordinates according to the road nodes and performing cluster analysis on the road nodes, operation data, and road test data to obtain the number of accidents around the road nodes and the average passing time between adjacent road nodes;

[0044] A cost flow model construction module for constructing a cost flow network model according to the data set, the number of accidents around the road nodes, and the average passing time between adjacent road nodes and determining the location of each rescue point;

[0045] A calculation module for solving the minimum cost maximum flow that meets the rescue point flow limit condition based on the cost flow network model;

[0046] An output module for outputting the transport capacity of each rescue point according to the minimum cost maximum flow.

[0047] According to another aspect of the present invention, there is provided an electronic device, including: at least one processor, and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute any one of the deployment methods of rescue points for driverless online car-hailing in the embodiments of the present invention.

[0048] According to another aspect of the present invention, there is provided a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to execute any one of the deployment methods of rescue points for driverless online car-hailing in the embodiments of the present invention.

[0049] Compared with the prior art, the beneficial effect of the deployment method of rescue points for driverless vehicles in the embodiments of the present invention lies in:

[0050] In the embodiments of the present invention, by analyzing real-time road test and accident data, rescue points can be configured near high-accident areas based on the minimum cost flow model, making the distribution of rescue resources more reasonable. This data-driven method can ensure that rescue vehicles are concentrated in the areas where they are most needed, avoiding waste of resources.

[0051] In the embodiments of the present invention, the minimum-cost resource allocation method can be found through the cost flow model, including minimizing the number of rescue points, vehicle deployment costs, and maintenance costs, so that cost data such as the operation and maintenance costs of rescue points and the daily vehicle costs can minimize the overall operation costs while meeting service requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 is a flowchart of the deployment method of rescue points provided by the embodiments of the present invention;

[0053] Figure 2 is a flowchart of the deployment method of rescue points provided by another embodiment of the present invention;

[0054] Figure 3 is a schematic diagram of the distribution of road nodes provided by another embodiment of the present invention;

[0055] Figure 4 is a schematic diagram of the number of accidents around road nodes provided by another embodiment of the present invention;

[0056] Figure 5 is a cost flow network diagram of the cost flow situation in road nodes provided by another embodiment of the present invention;

[0057] Figure 6 is a schematic diagram of the deployment device of rescue points provided by the embodiments of the present invention;

[0058] Figure 7It is a block diagram of the electronic device for implementing the embodiments of the present invention.

[0059] In the figure, 10 is the data acquisition module; 20 is the road node splitting module; 30 is the data analysis module; 40 is the cost flow model construction module; 50 is the calculation module; 60 is the output module; 600 is the electronic device; 601 is the calculation unit; 602 is the ROM; 603 is the RAM; 604 is the bus; 605 is the I / O interface; 606 is the input unit; 607 is the output unit; 608 is the storage unit; 609 is the communication unit. Detailed implementation manners

[0060] The following describes exemplary embodiments of the present invention with reference to the accompanying drawings, including various details of the embodiments of the present invention to facilitate understanding, which should be considered merely exemplary. Therefore, those of ordinary skill in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope of the present invention. Similarly, for clarity and conciseness, the description below omits the description of well-known functions and structures.

[0061] As Figures 1-5 shown, in an alternative embodiment of the present invention, the method for deploying the rescue points of driverless vehicles includes:

[0062] S10. Obtain a data set for constructing a cost flow network model, where the data set includes map data, road test data, operation data, alternative rescue point information, and total transport capacity, and the map data includes lane information;

[0063] S20. Split road nodes according to the lane information and a preset distance threshold;

[0064] S30. Obtain the coordinates of road nodes according to the road nodes and perform clustering analysis on the road nodes, operation data, and road test data to obtain the number of accidents around the road nodes and the average passing time between adjacent road nodes;

[0065] S40. Construct a cost flow network model according to the data set, the number of accidents around the road nodes, and the average passing time between adjacent road nodes, and determine the location of each rescue point;

[0066] S50. Solve the minimum cost maximum flow that satisfies the flow limit condition of the rescue points based on the cost flow network model;

[0067] S60. Output the transport capacity of each rescue point according to the minimum cost maximum flow.

[0068] Among them, the cost flow network model is a network flow model that not only considers the flow of traffic but also the costs required for the flow (such as time, resources, etc.). Lane information is the specific information about road lanes in map data, such as the number of lanes, direction, location, etc. The distance threshold is a preset distance value used to split road nodes, and road sections exceeding this distance will be regarded as different nodes. Cluster analysis is a data analysis method used to divide a data set into multiple groups or clusters, such that data points within the same group are similar to each other, while data points in different groups are quite different. The minimum-cost maximum-flow is to find a flow allocation scheme in the network flow problem that minimizes the total cost while meeting all traffic demands.

[0069] Specifically, the role of S10 is to collect relevant data required for constructing the cost flow network model to ensure the integrity and accuracy of the basic data for model construction. The role of S20 is to split road nodes. According to lane information and the distance threshold, the road is divided into multiple nodes for more refined analysis. In S30, through cluster analysis, the number of accidents around road nodes and the average travel time between adjacent nodes are understood, providing key parameters for constructing the cost flow network model and reflecting the actual traffic conditions and accident risks of road nodes. In S40, using the data set and the results of cluster analysis, a cost flow network model is constructed, and the location of rescue points is determined, providing a basic model for solving the minimum-cost maximum-flow problem and optimizing the allocation of rescue resources. In S50, the minimum-cost maximum-flow that meets the flow limit conditions of rescue points is found through the cost flow network model to ensure that the rescue cost (such as time, resources, etc.) is minimized while meeting rescue requirements. In S60, according to the solution results of the minimum-cost maximum-flow, the transport capacity of each rescue point is determined, providing specific guidance for the actual deployment of rescue resources to ensure rescue efficiency.

[0070] Specifically, it is necessary to combine Figures 2-5 To further illustrate with a complete implementation method:

[0071] This embodiment is mainly divided into three levels: the input layer, the model layer, and the output layer.

[0072] Input layer

[0073] Data is collected through five dimensions, namely:

[0074] Map data, including information such as several lane information, lane structure, lane longitude and latitude coordinates, and lane connection relationships.

[0075] Road test data and operation data. Set a time window as the previous D days; the vehicle trajectory information of the previous D days is denoted as the set Traj; the accident coordinate set within D days, including accidents such as manual takeover, remote takeover, sudden braking and sharp steering, and actual collision accidents, is denoted as the longitude and latitude coordinate set Acci.

[0076] Alternative rescue point information, with rescue points marked as u1, u2, …, u m ; The corresponding longitude and latitude coordinates are Transport capacity limits are l1, l2, …, l m ; The cost per vehicle for each order is c1, c2, …, c m .

[0077] The total transport capacity is L.

[0078] Model layer

[0079] Construct the model layer, which is divided into 5 steps. Split road nodes:

[0080] Set the distance threshold Len;

[0081] Divide the road according to the distance threshold and the road forward direction to obtain road nodes v1, v2, …, v n , where the longitude and latitude distance calculation formula is required:

[0082]

[0083] where R is the radius of the earth, taken as 6378.137 km, Δφ = φ1 - φ2 is the latitude difference between two longitude and latitude coordinates, and Δλ = λ1 - λ2 is the longitude difference between two longitude and latitude coordinates;

[0084] The longitude and latitude coordinates of the road nodes are:

[0085] Node topological relationship, denoted as successor(v) represents the set of all successor nodes of v.

[0086] The road nodes are as Figure 3 shown by the blue dots. Perform clustering analysis. For road node v ∈ {v1, v2,..., v n}, denoted as:

[0087]

[0088] where dis(a, v) is the distance between accident point a and road node v, which needs to be calculated using the longitude and latitude distance calculation formula described above. collision(v) refers to the number of accidents within the radius area around road node v. As Figure 4 shown, for adjacent road nodes v i , v j , the average travel time from v i to v j is statistically calculated through the trajectory set Traj. Denoted as time(v i vj ) Construct a cost flow network \(G=(V, A)\), where \(V\) is the set of nodes in network \(G\) and \(A\) is the set of edges in network \(G\). Construct a source node \(s\) and a sink node \(t\); Connect the source node \(s\) to each rescue point. For the edge \((s, u\) i ), configure its capacity as the transport capacity \(l\) of rescue point \(u\) i , that is, \(cap(s, u\) i ) = \(l\) i ; Configure its cost as the cost \(c\) of a single vehicle at the rescue point, that is, \(cost(s, u\) i ) = \(c\) i . Connect each rescue point \(u\) i to the nearest road node \(v\) nearby, that is, the vehicle at rescue point \(u\) i will first depart to road node \(v\) i . Set the capacity to \(cap(u\) j \(v\) i ) = \(\infty\), and the cost is the travel time \(time(u\) j , \(v\) i ) from rescue point \(u\) j to \(v\) i multiplied by the weight \(w\) j , that is, \(cost(u\) i \(v\) j ) = \(w\) time ·\(time(u\) i . \(v\) j ). Connect each road node \(v\) time to all its successor nodes \(successor(v\) i ). For the edge \((v\) j \(v\) i ), set its capacity to \(cap(v\) i , \(v\) i \(v\) j ) = \(\infty\), and the cost is the travel time \(time(v\) i , \(v\) j ) from point \(v\) i to point \(v\) j multiplied by the weight \(w\) i , that is, \(cost(v\) j \(v\) time ) = \(w\) i ·\(time(v\) j , \(v\) time ). Connect all road nodes to the sink node \(t\). For the edge \((v\) i , \(t)\), set the capacity to the rescue demand \(need(v\) j ) of point \(v\) i , where \(need(v\) i ) = \(collision(v\) i ) i ), where \(need(v\) i) / T, where T is the average response time for handling a single accident, i.e., cap(v i , t) = need(v i ); Set the cost cost(v i , t) = 0.

[0089] The cost flow network is as Figure 5 shown. Using the cost flow algorithm, solve the minimum cost flow from s to t with the maximum flow not exceeding L which satisfies:

[0090] The flow departing from the source point does not exceed L, i.e., ∑ 1≤i≤m flow(s, u i ) ≤ L;

[0091] The flow on any edge does not exceed the capacity, i.e., Thus, it can be ensured that:

[0092] The flow reaching each rescue point u i does not exceed the transport capacity l i

[0093] of the rescue point. j For each road node v j to the sink point, the flow does not exceed the rescue demand need(v

[0094] For any node other than the source point and the sink point, the inflow and outflow flows are the same, i.e.,

[0095] Configure the transport capacity of the rescue point based on flow;

[0096] For the rescue point u i , configure the transport capacity flow(s, u i ).

[0097] Output layer

[0098] The output layer contains the transport capacity required for each rescue point output by the model layer.

[0099] In the embodiment of the present invention, by analyzing real-time road test and accident data, based on the minimum cost flow model, rescue points can be configured near high accident-prone areas, making the distribution of rescue resources more reasonable. This data-driven method can ensure that rescue vehicles are concentrated in the areas where they are most needed, avoiding waste of resources. Through the cost flow model, the embodiment of the present invention can find the resource allocation method with the lowest total cost, including minimizing the number of rescue points, vehicle deployment costs, and maintenance costs, so that cost data such as the operation and maintenance costs of rescue points and the daily vehicle costs can minimize the overall operation cost while meeting service requirements.

[0100] In an alternative embodiment of the present invention, the map information includes lane structure, lane longitude and latitude coordinates, and lane connection relationships. The data set includes an accident coordinate set Acci, and the accident longitude and latitude coordinate set Acci is obtained through the roadside data and the operation data.

[0101] Among them, the steps of obtaining the accident longitude and latitude coordinate set Acci specifically include:

[0102] S11, set a time window as the previous D days;

[0103] S12, obtain the vehicle trajectory information of the previous D days according to the data set, denoted as set Traj;

[0104] S13, obtain the accident coordinate set within D days according to the data set. The accidents include: manual takeover, remote takeover, emergency braking and sharp steering wheel and actual collision accidents, denoted as the accident longitude and latitude coordinate set Acci.

[0105] In an alternative embodiment of the present invention, in step S40, the steps of constructing the cost flow network model and determining the location of each rescue point include:

[0106] S41, set a distance threshold Len and divide the road according to the distance threshold and the road forward direction to obtain road nodes v1, v2,..., v n ;

[0107] S42, according to the longitude and latitude coordinate distance formula:

[0108]

[0109] Calculate the longitude and latitude coordinates of the road nodes Δφ = φ1 - φ2 is the latitude difference between two longitude and latitude coordinates, and Δλ = λ1 - λ2 is the longitude difference between two longitude and latitude coordinates;

[0110] S43, obtain the topological relationship of the road nodes, denoted as successor(v) represents the set of all successor nodes of v;

[0111] S44, for the road node v ∈ {v1, v2,..., v n}, denote where dis(av) is the distance between the accident point a and the road node v, and the collision(v) is the number of accidents within the radius area around the road node v. For adjacent road nodes v i , v j, by collecting Traj statistics v i to v j The average travel time is denoted as time(v i , v j );

[0112] S45, based on the average travel time time(v i v j ) and number of accidents Construct a cost flow network G = (VA), where V is the point set of network G and A is the edge set of network G.

[0113] In an optional embodiment of the present invention, in S50, solving the minimum cost maximum flow that meets the flow restriction condition of the rescue point based on the cost flow network model includes the following steps:

[0114] S51, build source point s and sink point t and connect source point s to each rescue point, where for the edge (s, u i ), configure its capacity as rescue point u i Capacity i , that is, cap(s,u i )=l i , and configure its cost as the cost of a single vehicle at the rescue point c i , that is, cost(s,u i )=c i ;

[0115] S52, connect each rescue point u i To the nearest road node v j , i.e. rescue point u i The vehicle will first depart to the road node v j , set the capacity to cap(u i , v j )=∞, the cost is the rescue point u i to v j The travel time time(u i , v j ) multiplied by the weight w time , that is, cost(u i v j )=w time ·time(u i .v j );

[0116] S53, connect each road node v i To all its successor nodes successor(v i ), for edge (v i , v j) Set its capacity to cap(v i , v j ) = ∞, and the cost is the travel time time(v i to point v j ) multiplied by the weight w i , v j ), that is, cost(v time , v i ) = w j · time(v time , v i ); j );

[0117] S54. Connect all road nodes to the sink point t. For the edge (v i , t), set the capacity to the rescue demand need(v i ) of point v i ), where need(v i ) = collision(v i ) / T, and T is the average response time for handling a single accident, that is, cap(v i , t) = need(v i ), and set the cost cost(v i , t) = 0;

[0118] S55. Use the minimum - cost flow algorithm to solve the minimum - cost flow flow: A → R from s to t with the maximum flow not exceeding L;

[0119] Among them, the constraints include:

[0120] The flow starting from the source point does not exceed L, that is, ∑ 1≤i≤m flow(s, u i ) ≤ L;

[0121] The flow on any edge does not exceed the capacity, that is Thus, it can be ensured that:

[0122] The flow reaching each rescue point u i does not exceed the transport capacity l i of the rescue point;

[0123] The flow from each road node v j to the sink point does not exceed the rescue demand need(v j );

[0124] For any node other than the source point and the sink point, the inflow and outflow flows are the same, that is

[0125] S56. Configure the rescue point capacity based on the flow. For the rescue point u i , configure the capacity flow(s, u i ).

[0126] In an alternative embodiment of the present invention, when obtaining the accident longitude and latitude coordinate set, it further includes classifying the accident types, where the accident types include vehicle collisions and vehicle failures, and different weights are assigned to each accident type.

[0127] In an alternative embodiment of the present invention, when solving the minimum cost maximum flow that satisfies the rescue point flow limit condition, the cooperation ability between rescue points is considered to maximize the overall rescue efficiency.

[0128] As Figure 6 shown, according to another aspect of the present invention, there is provided a deployment device for a rescue point of an autonomous vehicle, and the device includes:

[0129] A data acquisition module 10, which is used to obtain a data set for constructing a cost flow network model. The data set includes map data, road test data, operation data, alternative rescue point information, and total capacity. The map data includes lane information;

[0130] A road node splitting module 20, which is used to split road nodes according to the lane information and a preset distance threshold;

[0131] A data analysis module 30, which is used to obtain the road node coordinates according to the road nodes and perform cluster analysis on the road nodes, operation data, and road test data to obtain the number of accidents around the road nodes and the average passing time between adjacent road nodes;

[0132] A cost flow model construction module 40, which is used to construct a cost flow network model and determine the location of each rescue point according to the data set, the number of accidents around the road nodes, and the average passing time between adjacent road nodes;

[0133] A calculation module 50, which is used to solve the minimum cost maximum flow that satisfies the rescue point flow limit condition based on the cost flow network model;

[0134] An output module 60, which is used to output the capacity size of each rescue point according to the minimum cost maximum flow.

[0135] The embodiments of the present invention have the following seven beneficial effects

[0136] 1. Precise positioning and resource optimization

[0137] By analyzing real-time road test and accident data, the minimum cost flow model can allocate rescue points near high-accident areas, making the distribution of rescue resources more reasonable. This data-driven approach can ensure that rescue vehicles are concentrated in the areas where they are most needed, avoiding resource waste.

[0138] The average travel time and accident probability between road nodes provide real-time data support for the model, making the allocation of rescue points more targeted and responsive in real time.

[0139] II. Cost Minimization

[0140] A core objective of the minimum cost flow model is to control the costs of rescue point configuration and capacity allocation. Through network cost flow calculation, the model can find the resource allocation method with the lowest total cost, including minimizing the number of rescue points, vehicle deployment costs, and maintenance costs.

[0141] Using cost data such as the operation and maintenance costs of rescue points and the daily vehicle costs, the model can minimize the overall operation cost while meeting service requirements.

[0142] III. Response Time Optimization

[0143] In case of an accident, rescue vehicles can quickly depart from the nearest rescue point through optimal configuration, greatly shortening the response time. The minimum cost flow model configures rescue resources by minimizing paths and distances, improving the timeliness of emergency response.

[0144] The average travel time, as the "edge cost" in the network, can effectively reflect road congestion, dynamically adjust subsequent rescue configurations, and better conform to the real traffic environment.

[0145] IV. Strong Dynamic Adjustment and Scalability

[0146] With the development of autonomous driving technology and the changes in the road network, the model can be adjusted according to new data (such as new accident data and road condition data) to maintain its dynamic response ability. Adding new rescue points or adjusting the configuration of existing rescue points can be quickly achieved in the model.

[0147] If new areas are opened or road conditions improve in the future, the minimum cost flow model can be flexibly expanded without the need to reconstruct the entire network, enhancing the operability and scalability of the model.

[0148] V. Data-Driven Scientific Decision Support

[0149] Through clustering analysis and traffic statistics, the model can intuitively identify accident-prone sections and areas with high vehicle demand, providing a scientific basis for decision-making and making the layout of rescue points more reasonable and forward-looking.

[0150] This configuration method based on big data analysis can provide detailed accident statistics and rescue demand prediction for managers, helping to improve the safety and service quality of autonomous driving.

[0151] VI. Improving the Social Recognition and Safety of Autonomous Driving Systems

[0152] Ensuring the operational safety of autonomous vehicles through a fast and efficient rescue system can significantly enhance the public's trust and acceptance of autonomous driving technology.

[0153] The reasonable configuration and resource allocation of rescue points can effectively reduce the potential impact of accidents, improve the overall safety factor of autonomous driving, and contribute to the industry's promotion.

[0154] VII. Supporting Multi-Objective Optimization

[0155] The minimum cost flow model can handle multiple optimization objectives simultaneously, such as the shortest rescue response time, the minimum cost, and the most reasonable resource allocation, making the model more in line with the actual operational requirements.

[0156] When a specific objective needs to be prioritized (for example, shortening the response time during peak hours), the model can make trade-offs between different objectives to achieve the optimal balance state.

[0157] Figure 7 FIG. shows a schematic block diagram of an exemplary electronic device 600 that can be used to implement an embodiment of the present invention. The electronic device 600 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely exemplary and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0158] As Figure 7 shown, the electronic device 600 includes a computing unit 601, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 602 or a computer program loaded from a storage unit 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the device 600 can also be stored. The computing unit 601, the ROM 602, and the RAM 603 are connected to each other through a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0159] The components in the electronic device 600 are connected to the I / O interface 605, including: an input unit 606, such as a keyboard, a mouse, etc.; an output unit 607, such as various types of displays, speakers, etc.; a storage unit 608, such as a magnetic disk, an optical disc, etc.; and a communication unit 609, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 609 allows the electronic device 600 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0160] The computing unit 601 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 601 include but are not limited to a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 601 executes the various methods and processes described above, such as a method for deploying rescue points of a driverless online car-hailing service. For example, in some embodiments, a method for deploying rescue points of a driverless online car-hailing service can be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as the storage unit 608. In some embodiments, part or all of the computer program can be loaded and / or installed onto the device 600 via the ROM 602 and / or the communication unit 609. When the computer program is loaded into the RAM 603 and executed by the computing unit 601, one or more steps of the method for deploying rescue points of a driverless online car-hailing service described above can be executed. Alternatively, in other embodiments, the computing unit 601 can be configured to execute a method for deploying rescue points of a driverless online car-hailing service in any other suitable manner (e.g., by means of firmware).

[0161] The various embodiments of the systems and technologies described above in this article can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-chip (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special-purpose or general-purpose programmable processor, and can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit the data and instructions to the storage system, the at least one input device, and the at least one output device.

[0162] The program code for implementing the method of the present invention can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing devices, such that when the program codes are executed by the processor or controller, the functions / operations specified in the flowchart and / or block diagram are implemented. The program codes can be executed entirely on the machine, partially on the machine, executed partially on the machine and partially on a remote machine as an independent software package, or executed entirely on a remote machine or server.

[0163] In the context of the present invention, a machine-readable medium can be a tangible medium that can contain or store a program for use by or in connection with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0164] In order to provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or a trackball) through which the user can provide input to the computer. Other kinds of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0165] The systems and techniques described herein can be implemented in a computing system including backend components (e.g., as a data server), or a computing system including middleware components (e.g., an application server), or a computing system including frontend components (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described herein), or a computing system including any combination of such backend components, middleware components, or frontend components. The components of the system can be interconnected to each other by digital data communication in any form or medium (e.g., a communication network). Examples of communication networks include: local area network (LAN), wide area network (WAN), and the Internet.

[0166] A computer system can include a client and a server. The client and the server are generally far from each other and typically interact through a communication network. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, can also be a server of a distributed system, or a server incorporating a blockchain.

[0167] It should be understood that various forms of the processes shown above can be used, steps can be reordered, added, or deleted. For example, the steps recited in the present invention can be executed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions disclosed in the present invention can be achieved, and no limitation is imposed herein.

[0168] The above specific embodiments do not constitute a limitation on the protection scope of the present invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for deploying rescue points for unmanned online car-hailing vehicles, characterized in that: The method comprises: S10, obtaining a data set for constructing a cost flow network model, the data set including map data, road test data, operation data, alternative rescue point information, and total transport capacity, the map data including lane information; S20, splitting the road node according to the lane information and a preset distance threshold; S30, acquiring the coordinates of the road nodes according to the road nodes and performing cluster analysis on the road nodes, operation data, and road test data to obtain the number of accidents around the road nodes and the average travel time between adjacent road nodes; S40, constructing a cost flow network model and determining the location of each rescue point based on the data set, the number of accidents around the road node, and the average travel time between adjacent road nodes; S50, solving the minimum cost maximum flow that meets the flow restriction condition of the rescue point based on the cost flow network model; S60, outputting the transport capacity of each rescue point according to the minimum cost maximum flow.

2. The method for deploying rescue points according to claim 1, characterized in that: The map information includes lane structure, lane longitude and latitude coordinates, and lane connection relationship, and the data set includes an accident coordinate set Acci, and the accident longitude and latitude coordinate set Acci is obtained through the road test data and the operation data; The step of obtaining the accident latitude and longitude coordinate set Acci specifically includes: S11, set a time window as the previous D days; S12, obtaining vehicle trajectory information for the previous D days according to the data set, recorded as a set Traj; S13, obtaining a set of accident coordinates within D days according to the data set, where the accidents include: manual takeover, remote takeover, sudden braking and steering, and actual collision accidents, recorded as the accident latitude and longitude coordinate set Acci.

3. The method for deploying rescue points according to claim 2, characterized in that: The steps of constructing the cost flow network model and determining the location of each rescue point in S40 include: S41, set the distance threshold Len and divide the road according to the distance threshold and the road direction to obtain road nodes v1, v2, ..., v n ; S42, according to the longitude and latitude coordinate distance formula Calculate the longitude and latitude coordinates of road nodes is the latitude difference between two longitude and latitude coordinates, Δλ=λ1-λ2 is the longitude difference between two longitude and latitude coordinates; S43, obtain the road node topology relationship, denoted as successor: {v1, v2, ..., v n }→2{v1, v2, …v n }, successor(v) represents the set of all successor nodes of v; S44, for a road node v∈{v1,v2,…,v n },remember Where dis(a, v) is the distance between the accident point a and the road node v, and collision(v) is the radius around the road node v. The number of accidents in the area, for adjacent road nodes v i ,v j , by collecting Traj statistics v i to v j The average travel time is denoted as time(v i , v j ); S45, based on the average travel time time(v i , v j ) and number of accidents Construct a cost flow network G = (V, A), where V is the point set of network G and A is the edge set of network G.

4. The method for deploying rescue points according to claim 3, characterized in that: In S50, solving the minimum cost maximum flow that meets the flow restriction condition of the rescue point based on the cost flow network model includes the following steps: S51, build source point s and sink point t and connect source point s to each rescue point, where for the edge (s, u i ), configure its capacity as rescue point u i Capacity i , that is, cap(s,u i )=l i , and configure its cost as the cost of a single vehicle at the rescue point c i , that is, cost(s,u i )=c i ; S52, connect each rescue point u i To the nearest road node v j , i.e. rescue point u i The vehicle will first depart to the road node v j , set the capacity to cap(u i , v j )=∞, the cost is the rescue point u i to v j The travel time time(u i , v j ) multiplied by the weight w time , that is, cost(u i , v j )=w time ·time(u i .v j ); S53, connect each road node v i To all its successor nodes successor(v i ), for edge (v i , v j ), set its capacity to cap(v i , v j )=∞, the cost is point v i To point v j The travel time time(v i , v j ) multiplied by the weight w time , that is, cost(v i , v j )=w time ·time(v i , v j ); S54, connect all road nodes to the sink t, for the edge (v i , t), set the capacity to point v i rescue needs i ), where need(v i )=collision(v i ) / T, T is the average response time for handling a single incident, that is, cap(v i , t) = need(v i ), set the cost cost(v i , t) = 0; S55, using the cost flow algorithm to solve the minimum cost flow from s to t with a maximum flow not exceeding L: A→R; The restrictions include: The flow from the source does not exceed L, that is, ∑ 1≤i≤m flow(s,u i )≤L; The flow on any edge does not exceed the capacity, that is, This ensures that: Arrive at each rescue point i The flow rate will not exceed the capacity of the rescue point l i ; Each road node v j The flow to the sink will not exceed the rescue demand of the road node need(v j ); For any node other than the source and sink, the inflow and outflow are the same, that is, S56, configure the rescue point capacity based on flow, for rescue point u i , configure the capacity flow (s, u i ).

5. The method for deploying rescue points according to claim 1, characterized in that: When obtaining the latitude and longitude coordinate set of the accident, the accident type is also classified, and the accident type includes vehicle collision and vehicle failure, and different weights are assigned to each accident type.

6. The method for deploying rescue points according to claim 4, characterized in that: When solving the minimum cost maximum flow that meets the flow restriction conditions of the rescue points, the cooperation ability between the rescue points is taken into consideration to maximize the overall rescue efficiency.

7. A deployment device for a rescue point of an unmanned online car-hailing vehicle, characterized in that: The device comprises: A data acquisition module, the data acquisition module is used to obtain a data set for building a cost flow network model, the data set includes map data, road test data, operation data, alternative rescue point information, total transport capacity, and the map data includes lane information; A road node splitting module, the road node splitting module is used to split the road nodes according to the lane information and a preset distance threshold; A data analysis module, the data analysis module is used to obtain the coordinates of the road nodes according to the road nodes and perform cluster analysis on the road nodes, operation data, and road test data to obtain the number of accidents around the road nodes and the average travel time between adjacent road nodes; A cost flow model building module, the cost flow model building module is used to build a cost flow network model and determine the location of each rescue point based on the data set, the number of accidents around the road node, and the average travel time between adjacent road nodes; A calculation module, the calculation module is used to solve the minimum cost maximum flow that meets the flow restriction condition of the rescue point based on the cost flow network model; An output module is used to output the transport capacity of each rescue point according to the minimum cost maximum flow.

8. An electronic device, characterized in that: include: at least one processor, and a memory communicatively coupled to the at least one processor; The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method described in any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-6.

Images