Vehicle rescue management method and system based on cloud platform and edge computing

Abstract

The application provides a vehicle rescue management method and system based on a cloud platform and edge computing, and relates to the field of data processing.The method comprises the following steps: determining a vehicle rescue service station setting scheme and an edge computing configuration scheme of a target area according to historical vehicle rescue data of the target area, setting multiple vehicle rescue service stations in the target area, and setting multiple edge computing stations in the target area; determining a vehicle rescue resource dynamic scheduling scheme and an edge computing dynamic scheduling scheme according to real-time road information of the target area through a cloud platform; scheduling the multiple edge computing stations to determine vehicle rescue demand of the target area according to the real-time road information of the target area based on the edge computing dynamic scheduling scheme; receiving vehicle rescue requests through the multiple edge computing stations; and performing vehicle rescue through the multiple edge computing stations according to the vehicle rescue demand and the vehicle rescue requests of the target area, so that the vehicle rescue efficiency and quality are improved.

Description

Technical Field

[0001] This invention relates to the field of vehicle rescue management, and in particular to a vehicle rescue management method and system based on cloud platform and edge computing. Background Technology

[0002] Roadside assistance refers to emergency roadside assistance for vehicles, providing on-site minor repairs and other services to owners of disabled vehicles. It also includes roadside assistance following traffic accidents, such as medical treatment for the injured and traffic control. With the rapid development of the automotive industry, the demand for roadside assistance services is also growing rapidly, highlighting the importance of roadside assistance management.

[0003] The traditional process of providing vehicle rescue services in existing technologies typically involves the vehicle owner initiating a rescue request, and the rescue provider offering corresponding assistance based on the request. After receiving the rescue information, the rescue provider then assigns the service to the nearest rescue vehicle station, which in turn dispatches the information to a rescue vehicle. This process involves lengthy message transmission, making the dispatch of rescue information cumbersome and slow, and failing to meet the need for immediate assistance to vehicle owners. Furthermore, the allocation of rescue resources is often inefficient, hindering the improvement of rescue efficiency.

[0004] Therefore, there is a need to provide vehicle rescue management methods and systems based on cloud platforms and edge computing to improve the efficiency and quality of vehicle rescue. Summary of the Invention

[0005] This invention provides a vehicle rescue management method based on a cloud platform and edge computing, comprising: acquiring historical vehicle rescue data for a target area; determining a vehicle rescue service station setup scheme and an edge computing configuration scheme for the target area based on the historical vehicle rescue data; setting up multiple vehicle rescue service stations in the target area according to the vehicle rescue service station setup scheme; setting up multiple edge computing stations in the target area according to the edge computing configuration scheme; acquiring real-time road information for the target area; determining a dynamic scheduling scheme for vehicle rescue resources and a dynamic scheduling scheme for edge computing based on the real-time road information of the target area through a cloud platform; scheduling vehicle rescue resources to multiple vehicle rescue service stations according to the dynamic scheduling scheme for vehicle rescue resources; scheduling multiple edge computing stations based on the dynamic scheduling scheme for edge computing to determine the vehicle rescue needs of the target area according to the real-time road information of the target area; receiving vehicle rescue requests through multiple edge computing stations; and performing vehicle rescue through multiple edge computing stations based on the vehicle rescue needs and vehicle rescue requests of the target area.

[0006] Furthermore, based on historical vehicle rescue data of the target area, a vehicle rescue service station setup plan for the target area is determined, including: determining multiple risk locations and the risk coefficient of each risk location in the target area based on historical vehicle rescue data; determining the driving distance between any two risk locations based on an electronic map of the target area; determining the accident correlation coefficient between any two risk locations based on historical vehicle rescue data of the target area; grouping the multiple risk locations in the target area into multiple first risk location groups based on the driving distance and accident correlation coefficient of any two risk locations; determining the first central location and rescue resource requirements corresponding to each first risk location group based on the risk coefficient of each risk location included in the first risk location group and the accident correlation coefficient between any two risk locations included in the first risk location group; and determining the vehicle rescue service station setup plan for the target area based on the first central location and rescue resource requirements corresponding to each first risk location group.

[0007] Furthermore, based on the driving distance and accident correlation coefficient between any two risk locations, multiple risk locations in the target area are grouped to determine multiple first risk location groups. This includes: using the K-means algorithm to cluster multiple risk locations in the target area based on the driving distance between any two risk locations and the maximum driving distance threshold to determine multiple first risk location clusters; for each first risk location cluster, using the K-means algorithm to cluster multiple risk locations included in the first risk location cluster based on the accident correlation coefficient between any two risk locations and the maximum accident correlation coefficient threshold to determine the first risk location group included in the first risk location cluster.

[0008] Furthermore, based on historical vehicle rescue data of the target area, an edge computing configuration scheme for the target area is determined, including: for any two risk locations, determining the straight-line distance between the two risk locations according to the electronic map of the target area; clustering multiple risk locations in the target area based on the straight-line distance and the maximum threshold of the straight-line distance using the K-means algorithm to determine multiple second risk location clusters; for each second risk location cluster, clustering multiple risk locations included in the second risk location cluster based on the accident correlation coefficient and the maximum threshold of the accident correlation coefficient using the K-means algorithm to determine the second risk location group included in the second risk location cluster; for each second risk location group, determining the second center location and edge computing requirements corresponding to the second risk location group based on the risk coefficient of each risk location included in the second risk location group and the accident correlation coefficient between any two risk locations included in the second risk location group; and determining the edge computing configuration scheme for the target area based on the second center location and edge computing requirements corresponding to each second risk location group.

[0009] Further, acquiring real-time road information for the target area includes: acquiring meteorological data corresponding to historical vehicle rescue data in the target area; determining relevant meteorological elements based on the historical vehicle rescue data and the meteorological data corresponding to the historical vehicle rescue data in the target area; acquiring traffic flow data corresponding to the historical vehicle rescue data in the target area; for each risk location, determining the relevant roads corresponding to the risk location based on the historical vehicle rescue data and the traffic flow data corresponding to the historical vehicle rescue data in the target area; for each risk location, determining the relevant risk location corresponding to the risk location based on the historical vehicle rescue data in the target area; for each risk location, acquiring single-point real-time road information corresponding to the risk location based on relevant meteorological elements, the relevant roads corresponding to the risk location, and the relevant risk locations corresponding to the risk location, wherein the real-time road information for the target area includes single-point real-time road information corresponding to each risk location, and the single-point real-time road information corresponding to the risk location includes real-time meteorological information acquired based on relevant meteorological elements, real-time traffic flow information of the relevant roads corresponding to the risk location, real-time vehicle rescue information of the relevant risk locations corresponding to the risk location, and real-time status information of the risk location.

[0010] Furthermore, based on real-time road information in the target area, a dynamic scheduling plan for vehicle rescue resources is determined through the cloud platform. This includes: for each risk location, determining the single-point vehicle rescue demand parameters corresponding to the risk location based on real-time meteorological information obtained from relevant meteorological elements, real-time traffic flow information of the relevant roads corresponding to the risk location, and real-time vehicle rescue information of the relevant risk locations corresponding to the risk location; for each first risk location group, determining the group vehicle rescue demand parameters corresponding to the first risk location group based on the single-point vehicle rescue parameters corresponding to each risk location included in the first risk location group; and determining a dynamic scheduling plan for vehicle rescue resources based on the group vehicle rescue demand parameters corresponding to each first risk location group.

[0011] Furthermore, based on real-time road information in the target area, the cloud platform determines a dynamic edge computing scheduling scheme, including: for each risk location, determining the single-point computing power requirement parameters corresponding to the risk location based on real-time meteorological information obtained from relevant meteorological elements, real-time traffic flow information of the relevant roads corresponding to the risk location, and real-time vehicle rescue information of the relevant risk locations corresponding to the risk location; for each second risk location group, determining the group computing power requirement parameters corresponding to the second risk location group based on the single-point computing power requirement parameters corresponding to each risk location included in the first risk location group; and determining the dynamic edge computing scheduling scheme based on the group computing power requirement parameters corresponding to each second risk location group, wherein the dynamic edge computing scheduling scheme includes the risk location corresponding to each edge computing station.

[0012] Furthermore, the real-time status information of the risk location includes at least the audio and image information of the risk location; based on the edge computing dynamic scheduling scheme, multiple edge computing stations are scheduled to determine the vehicle rescue needs of the target area according to the real-time road information of the target area, including: for each edge computing station, the single-point vehicle rescue needs of the risk location are determined according to the audio and image information of the corresponding risk location, wherein the single-point vehicle rescue needs of the risk location include the single-point vehicle rescue needs of each risk location.

[0013] Furthermore, vehicle rescue is carried out through multiple edge computing stations based on the vehicle rescue needs and requests in the target area. This includes: for each edge computing station, determining the optimal vehicle rescue service station and target vehicle rescue resources from multiple vehicle rescue service stations based on the vehicle rescue needs and requests in the corresponding risk location; and the optimal vehicle rescue service station carrying out vehicle rescue based on the target vehicle rescue resources.

[0014] This invention provides a vehicle rescue management system based on a cloud platform and edge computing. Applying the aforementioned vehicle rescue management method based on a cloud platform and edge computing, the system includes: a resource configuration module for acquiring historical vehicle rescue data for a target area; determining a vehicle rescue service station setup scheme and an edge computing configuration scheme for the target area based on the historical vehicle rescue data; setting up multiple vehicle rescue service stations in the target area based on the vehicle rescue service station setup scheme; and setting up multiple edge computing stations in the target area based on the edge computing configuration scheme; an information acquisition module for acquiring real-time road information for the target area; a dynamic scheduling module for determining a dynamic scheduling scheme for vehicle rescue resources and a dynamic scheduling scheme for edge computing based on the real-time road information of the target area via a cloud platform; and scheduling vehicle rescue resources to multiple vehicle rescue service stations based on the dynamic scheduling scheme; and a rescue management module for scheduling multiple edge computing stations based on the dynamic scheduling scheme; determining vehicle rescue needs in the target area based on the real-time road information; receiving vehicle rescue requests through multiple edge computing stations; and performing vehicle rescue based on the vehicle rescue needs and requests in the target area through multiple edge computing stations.

[0015] Compared with existing technologies, the vehicle rescue management method and system based on cloud platform and edge computing provided by this invention has at least the following beneficial effects:

[0016] Edge computing stations can acquire and process road information and vehicle rescue requests in the target area in real time, thereby quickly determining rescue needs. This approach reduces the time delay of data transmission to the cloud platform for processing, resulting in a faster rescue response.

[0017] The cloud platform's dynamic vehicle rescue resource dispatch plan, based on real-time road information, can quickly dispatch the nearest rescue resources to the accident scene.

[0018] Edge computing stations further optimize resource allocation based on real-time road information and rescue needs, ensuring the efficient use of rescue resources.

[0019] By analyzing historical vehicle rescue data, the optimal setup and edge computing configuration for vehicle rescue service stations in a target area can be determined. This allows for more precise allocation of rescue resources, meeting the diverse rescue needs of different regions.

[0020] Based on changes in real-time road information and rescue requests, the cloud platform can dynamically adjust the dispatching plan for vehicle rescue resources. Edge computing stations can then fine-tune rescue resources according to real-time conditions to ensure effective resource utilization.

[0021] Rapid rescue response and precise resource matching can significantly reduce the waiting time for stranded vehicles. This helps improve satisfaction with rescue services and public trust in the rescue system. Optimized resource allocation and dynamic dispatching schemes can ensure that rescue resources arrive at the accident scene at critical moments. This helps improve the success rate of rescues and reduce losses caused by untimely rescue. Attached Figure Description

[0022] This specification will be further described by way of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. These embodiments are not limiting; in these embodiments, the same reference numerals denote the same structures, wherein:

[0023] Figure 1 This is a flowchart illustrating a vehicle rescue management method based on cloud platform and edge computing, according to some embodiments of this specification.

[0024] Figure 2 This is a flowchart illustrating a vehicle rescue service station setup scheme for determining a target area, as shown in some embodiments of this specification.

[0025] Figure 3 This is a flowchart illustrating the process of determining multiple first risk location clusters according to some embodiments of this specification;

[0026] Figure 4 This is a schematic diagram of a vehicle rescue management system based on cloud platform and edge computing, as shown in some embodiments of this specification. Detailed Implementation

[0027] To more clearly illustrate the technical solutions of the embodiments in this specification, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are merely some examples or embodiments of this specification. For those skilled in the art, these drawings can be applied to other similar scenarios without creative effort. Unless obvious from the context or otherwise specified, the same reference numerals in the drawings represent the same structures or operations.

[0028] Figure 1 This is a flowchart illustrating a vehicle rescue management method based on cloud platform and edge computing, as shown in some embodiments of this specification. Figure 1 As shown, the vehicle rescue management method based on cloud platform and edge computing may include the following steps.

[0029] Step 111: Obtain historical vehicle rescue data for the target area.

[0030] Specifically, the target area can be any region requiring vehicle rescue management. Historical vehicle rescue data for the target area includes information on vehicle rescues that occurred at specific historical points in time. This includes, for example: date and time: recording the exact date and time of each rescue, facilitating understanding the frequency and temporal distribution of rescue activities; rescue location: indicating the specific location of each rescue, aiding in analyzing the geographical distribution of rescue needs; vehicle information: including the type of vehicle rescued (e.g., sedan, truck, motorcycle), brand, model, and vehicle condition (e.g., whether the engine is off, whether a tire has blown out); and rescue type: categorized according to the nature of the rescue activity, such as accident rescue, breakdown towing, battery drain rescue, tire replacement, etc.

[0031] Step 112: Based on the historical vehicle rescue data of the target area, determine the vehicle rescue service station setup plan and edge computing configuration plan for the target area.

[0032] Figure 2 This is a flowchart illustrating a vehicle rescue service station setup scheme for determining a target area, as shown in some embodiments of this specification. Figure 2 As shown, in some embodiments, a vehicle rescue service station setup plan for the target area is determined based on historical vehicle rescue data of the target area, including:

[0033] Based on historical vehicle rescue data for the target area, multiple risk locations within the target area and the risk coefficient for each risk location are determined.

[0034] For any two risk locations, determine the driving distance between the two risk locations based on the electronic map of the target area;

[0035] For any two risk locations, determine the accident correlation coefficient between the two risk locations based on historical vehicle rescue data of the target area;

[0036] Based on the driving distance and accident correlation coefficient between any two risk locations, multiple risk locations in the target area are grouped to determine multiple first risk location groups;

[0037] For each first risk location group, the first center location and rescue resource requirements corresponding to the first risk location group are determined based on the risk coefficient of each risk location included in the first risk location group and the accident correlation coefficient between any two risk locations included in the first risk location group.

[0038] Based on the primary center location and rescue resource needs corresponding to each primary risk location group, determine the vehicle rescue service station setup plan for the target area.

[0039] Specifically, the following process can be used to determine multiple risk locations within the target area and the risk coefficient for each location:

[0040] S11. Based on the historical vehicle rescue data of the target area, determine the location where vehicle rescue occurred in the target area. The location where vehicle rescue occurred in the target area can be a location where the number of times vehicle rescue occurred in the target area is greater than a threshold.

[0041] S12. For each location in the target area where vehicle rescue occurs, calculate the location weight based on the historical vehicle rescue data of the target area;

[0042] S13. Select locations in the target area where vehicle rescue has occurred with a location weight greater than the location weight threshold as candidate locations;

[0043] S14. Based on the electronic map of the target area, determine the driving distance between two candidate locations, wherein the driving distance between two candidate locations can be the length of the road from one candidate location to another.

[0044] S15. Based on the K-means algorithm, cluster multiple candidate locations in the target area according to the driving distance between any two candidate locations and the second maximum driving distance threshold to determine multiple candidate location clusters. Among them, the driving distance between any candidate location included in the candidate location cluster and the candidate location corresponding to the cluster center of the candidate location cluster is less than the second maximum driving distance threshold.

[0045] S16. For each candidate location cluster, calculate the central coefficient corresponding to each candidate location included in the candidate location cluster, and take the candidate location with the largest central coefficient as the risk location.

[0046] S17. For each risk location, determine the risk coefficient based on historical vehicle rescue data of the target area.

[0047] For example, the location weight of a vehicle rescue location in the target area can be calculated using the following formula:

[0048]

[0049] Among them, w (i,1) N represents the location weight of the i-th location in the target area where vehicle rescue occurs. i Let N be the total number of vehicle rescues occurring at the i-th location within the target area. j Let J be the total number of times vehicle rescue occurs at the j-th location in the target area, where J is the total number of locations in the target area where vehicle rescue occurs.

[0050] The centrality coefficient corresponding to each candidate location within the candidate location cluster can be calculated using the following formula:

[0051]

[0052] Where, γ i Let a1 and a2 be the center coefficients corresponding to the i-th candidate position included in the candidate position cluster, and let a1 and a2 be the weights, where a1 + a2 = 1 and a1 and a2 are greater than 0. (m,i) Let M be the driving distance between the m-th candidate position and the ith candidate position included in the candidate position cluster, and M be the total number of candidate positions included in the candidate position cluster.

[0053] The risk coefficient for a risky location can be calculated using the following formula:

[0054]

[0055] Among them, R i Let N be the risk coefficient of the i-th risk location in the target area. i N represents the sum of the total number of vehicle rescues occurring at the candidate location cluster corresponding to the i-th risk location in the target area. e Let E be the sum of the total number of vehicle rescues occurring at the candidate locations included in the candidate location cluster corresponding to the e-th risk location in the target area, where E is the total number of risk locations in the target area.

[0056] The accident correlation coefficient between two risk locations can be calculated using the following formula:

[0057]

[0058] Among them, C (i,j) Let N be the accident correlation coefficient between the i-th risk location and the j-th risk location. (i,t) N represents the sum of the total number of vehicle rescues that occurred at the candidate location cluster corresponding to the i-th risk location in the target area during the t-th historical time period. (j,t) Let T be the sum of the total number of vehicle rescues that occurred at the candidate location cluster corresponding to the j-th risk location in the target area during the t-th historical time period, where T is the total number of historical time periods sampled.

[0059] In some embodiments, multiple risk locations in a target area are grouped according to the driving distance and accident correlation coefficient between any two risk locations to determine multiple first risk location groups, including:

[0060] Based on the K-means algorithm, multiple risk locations in the target area are clustered according to the driving distance between any two risk locations and the maximum driving distance threshold, thus determining multiple first risk location clusters;

[0061] For each first risk location cluster, the K-means algorithm is used to cluster multiple risk locations included in the first risk location cluster based on the accident correlation coefficient between any two risk locations and the maximum threshold of the accident correlation coefficient, thereby determining the first risk location group included in the first risk location cluster.

[0062] Figure 3 This is a flowchart illustrating the process of determining multiple first risk location clusters according to some embodiments of this specification, such as... Figure 3 As shown, specifically, based on the following process, multiple risk locations in the target area can be clustered according to the driving distance between any two risk locations and the maximum driving distance threshold to determine multiple first risk location clusters:

[0063] S21. Initialize the total number of cluster centers K1, and randomly select K1 risk locations from multiple risk locations as the initial cluster centers;

[0064] S22. For each risk location that is not a cluster center, the cluster centers whose driving distance from the risk location is less than the maximum driving distance threshold are taken as candidate cluster centers. If the number of candidate cluster centers for the risk location is greater than 0, the risk location is clustered into the cluster to which the cluster center with the smallest driving distance belongs. If the number of candidate cluster centers for the risk location is 0, the risk location is taken as a risk location to be clustered.

[0065] S23. Determine whether the number of risk locations to be clustered is greater than the number threshold. If yes, proceed to S24; otherwise, proceed to S25.

[0066] S24. Let K1 = K1 + 1, calculate the average driving distance corresponding to each risk location to be clustered, and take the risk location with the smallest average driving distance as the new cluster center, and execute S22.

[0067] S25. For each cluster, calculate the variance of the driving distance corresponding to the cluster based on the driving distance between each risk location included in the cluster and the cluster center of the cluster;

[0068] S26. Determine whether the driving distance variance corresponding to each cluster is less than the driving distance variance threshold. If yes, complete the grouping and treat each cluster as a first risk location cluster. If not, proceed to S27.

[0069] S27. Let K1 = K1 + 1. Select the cluster with the largest variance of driving distance that is greater than or equal to the driving distance variance threshold as the cluster to be optimized. Calculate the mean driving distance for each risk location included in the cluster to be optimized. Select the risk location with the largest mean driving distance as the new cluster center and execute S22.

[0070] The mean driving distance corresponding to the risk location to be clustered can be calculated using the following formula:

[0071]

[0072] Among them, D (i,1) Let D be the mean driving distance corresponding to the i-th risk location to be clustered. (h,i) Let H be the driving distance between the h-th risk location to be clustered and the i-th risk location to be clustered, where H is the total number of risk locations to be clustered.

[0073] The mean travel distance for each risk location within the cluster to be optimized can be calculated using the following formula:

[0074]

[0075] Among them, D (i,2) Let D be the mean driving distance corresponding to the i-th risk location in the cluster to be optimized. (g,i) Let G be the driving distance between the g-th risk location and the i-th risk location included in the cluster to be optimized, and let G be the total number of risk locations included in the cluster to be optimized.

[0076] Understandably, by initializing cluster centers and iteratively optimizing, this method can efficiently cluster multiple risk locations within a target area. This approach avoids the tediousness and subjectivity of manual classification, improving both efficiency and accuracy. Steps S23 and S27 allow for dynamic adjustment of the number of cluster centers based on the number of risk locations to be clustered and the variance of the cluster's travel distance. This adaptive mechanism ensures that the clustering results can adapt well to risk location sets of different sizes and distributions, improving the flexibility and applicability of the clustering.

[0077] The first risk location group, which includes the first risk location cluster, can be determined according to the following procedure:

[0078] S31. Initialize the total number of cluster centers K2, and randomly select K2 risk locations from the multiple risk locations included in the first risk location cluster as the initial cluster centers;

[0079] S32. For each risk location that is not a cluster center included in the first risk location cluster, the cluster center with an accident correlation coefficient less than the maximum threshold of the accident correlation coefficient with the risk location is taken as a candidate cluster center. If the number of candidate cluster centers of the risk location is greater than 0, the risk location is clustered into the group to which the cluster center with the smallest accident correlation coefficient belongs. If the number of candidate cluster centers of the risk location is 0, the risk location is taken as a risk location to be clustered.

[0080] S33. Determine whether the number of risk locations to be clustered is greater than the number threshold. If yes, proceed to S34; otherwise, proceed to S35.

[0081] S34. Let K2 = K2 + 1, calculate the mean accident correlation coefficient corresponding to each risk location to be clustered, and take the risk location with the smallest mean accident correlation coefficient as the new cluster center, and execute S32.

[0082] S35. For each group, calculate the variance of the accident correlation coefficients for the group based on the accident correlation coefficients between each risk location included in the group and the cluster center of the group.

[0083] S36. Determine whether the variance of the accident correlation coefficient corresponding to each group is less than the threshold of the accident correlation coefficient variance. If yes, complete the grouping and treat each group as a first risk position group. If not, proceed to S37.

[0084] S37. Let K2 = K2 + 1. The group with the largest variance of accident correlation coefficient that is greater than or equal to the threshold of accident correlation coefficient variance is taken as the group to be optimized. Calculate the mean of accident correlation coefficients corresponding to each risk location in the group to be optimized. Take the risk location with the largest mean of accident correlation coefficients as the new cluster center and execute S32.

[0085] Understandably, by considering accident correlation coefficients, this process allows for more precise segmentation of risk locations. Different accident correlation coefficients reflect the likelihood of simultaneous vehicle rescue needs at two risk locations. Through cluster analysis, risk locations with lower accident correlation coefficients within the first risk location cluster can be grouped together, effectively preventing congestion of vehicle rescue needs at individual vehicle rescue service stations. By automating and iteratively optimizing the cluster center process, this process can efficiently handle large amounts of risk location data. This reduces the time and effort required for manual classification, improving the overall efficiency of risk management.

[0086] The position weight of each risk location in the first risk location group can be determined based on the risk coefficient of each risk location in the first risk location group and the accident correlation coefficient between any two risk locations in the first risk location group. The risk location with the largest position weight is then taken as the first center location corresponding to the first risk location group.

[0087] The location weight of each risk location included in the first risk location group can be calculated using the following formula:

[0088]

[0089] Among them, w (i,2) Let b1, b2, and b3 be the position weights of the i-th risk position included in the first risk position group, where b1 + b2 + b3 = 1, b1, b2, and b3 are greater than 0, and C is the position weight of the i-th risk position. (i,l) Let L be the accident correlation coefficient between the l-th and i-th risk locations included in the first risk location group, and D be the total number of risk locations included in the first risk location group.(i,l) The distance between the l-th and i-th risk locations included in the first risk location group is denoted as .

[0090] The rescue resource demand corresponding to the first risk location group can be determined by the first demand prediction model based on the risk coefficient of each risk location included in the first risk location group, the accident correlation coefficient between any two risk locations included in the first risk location group, and the historical vehicle rescue data of each risk location included in the first risk location group. The first demand prediction model can be a convolutional neural network model.

[0091] In some embodiments, the edge computing configuration scheme for the target area is determined based on historical vehicle rescue data of the target area, including:

[0092] For any two risk locations, determine the straight-line distance between the two risk locations based on the electronic map of the target area;

[0093] Based on the K-means algorithm, multiple risk locations in the target area are clustered according to the straight-line distance between any two risk locations and the maximum threshold of the straight-line distance, and multiple second risk location clusters are determined.

[0094] For each second risk location cluster, the K-means algorithm is used to cluster multiple risk locations included in the second risk location cluster based on the accident correlation coefficient between any two risk locations and the maximum threshold of the accident correlation coefficient, to determine the second risk location group included in the second risk location cluster.

[0095] For each second risk location group, the second center location and edge computing requirements corresponding to the second risk location group are determined based on the risk coefficient of each risk location included in the second risk location group and the accident correlation coefficient between any two risk locations included in the second risk location group.

[0096] Based on the second center location and edge computing requirements corresponding to each second risk location group, determine the edge computing configuration scheme for the target area.

[0097] Specifically, the method for determining multiple second risk location clusters is similar to that for determining multiple first risk location clusters, and the method for determining the second risk location group included in the second risk location cluster is similar to that for determining the first risk location group included in the first risk location cluster, which will not be elaborated here.

[0098] The method for determining the second center location corresponding to the second risk location group is similar to the method for determining the first center location corresponding to the first risk location group, and will not be repeated here.

[0099] The edge computing demand corresponding to the second risk location group can be determined by the second demand prediction model based on the risk coefficient of each risk location included in the first risk location group and the accident correlation coefficient between any two risk locations included in the first risk location group. The second demand prediction model can be a convolutional neural network model.

[0100] Step 113: Set up multiple vehicle rescue service stations in the target area according to the vehicle rescue service station setup plan for the target area.

[0101] Step 114: Based on the edge computing configuration scheme of the target area, set up multiple edge computing stations in the target area.

[0102] Step 115: Obtain real-time road information for the target area.

[0103] In some embodiments, step 115 specifically includes:

[0104] Obtain meteorological data corresponding to historical vehicle rescue data in the target area;

[0105] Based on the historical vehicle rescue data of the target area and the corresponding meteorological data, determine the relevant meteorological elements (e.g., temperature, humidity, air pressure, wind speed, wind direction, precipitation, etc.).

[0106] Obtain traffic flow data corresponding to historical vehicle rescue data in the target area;

[0107] For each risk location, the relevant roads corresponding to the risk location are determined based on the historical vehicle rescue data of the target area and the traffic flow data corresponding to the historical vehicle rescue data of the target area.

[0108] For each risk location, the relevant risk locations are determined based on the historical vehicle rescue data of the target area;

[0109] For each risk location, real-time road information for a single point is obtained based on relevant meteorological elements, relevant roads corresponding to the risk location, and relevant risk locations corresponding to the risk location. The real-time road information for the target area includes real-time road information for a single point corresponding to each risk location. The real-time road information for a single point corresponding to a risk location includes real-time meteorological information obtained based on relevant meteorological elements, real-time traffic flow information of relevant roads corresponding to the risk location, real-time vehicle rescue information of relevant risk locations corresponding to the risk location, and real-time status information of the risk location.

[0110] Specifically, for each meteorological element, the correlation coefficient between the meteorological element and vehicle rescue can be calculated based on the meteorological element values ​​at multiple historical time points and the total number of vehicle rescues in the target area at those multiple historical time points. Meteorological elements with a correlation coefficient greater than the correlation coefficient threshold are considered as relevant meteorological elements.

[0111] For each road in the target area, the correlation coefficient between the road and the risk location can be calculated based on the traffic flow of the road at multiple historical time points and the total number of vehicle rescues at the risk location at those multiple historical time points. Roads with a correlation coefficient greater than the correlation coefficient threshold are identified as the relevant roads corresponding to the risk location.

[0112] For any two risk locations, the correlation coefficient between the two risk locations can be calculated based on the total number of vehicle rescues at multiple historical time points. Two risk locations with a correlation coefficient greater than the correlation coefficient threshold are considered to be related risk locations.

[0113] In some embodiments, the real-time status information of a risk location includes at least audio and image information of the risk location.

[0114] Step 116: Based on the real-time road information of the target area, determine the dynamic scheduling scheme for vehicle rescue resources and the dynamic scheduling scheme for edge computing through the cloud platform.

[0115] In some embodiments, step 116 specifically includes:

[0116] For each risk location, a single-point vehicle rescue demand parameter is determined based on real-time meteorological information obtained from relevant meteorological elements, real-time traffic flow information of relevant roads corresponding to the risk location, and real-time vehicle rescue information of relevant risk locations corresponding to the risk location. The single-point vehicle rescue demand parameter is used to measure the rescue demand intensity of the risk location at the current moment. For example, the single-point vehicle rescue demand parameter can be determined by a first parameter determination model based on real-time meteorological information obtained from relevant meteorological elements, real-time traffic flow information of relevant roads corresponding to the risk location, and real-time vehicle rescue information of relevant risk locations corresponding to the risk location. The first parameter determination model can be a convolutional neural network model.

[0117] For each first risk location group, the group vehicle rescue demand parameter corresponding to the first risk location group is determined based on the single-point vehicle rescue parameters corresponding to each risk location included in the first risk location group. The average value of the single-point vehicle rescue parameters corresponding to each risk location included in the first risk location group can be used as the group vehicle rescue demand parameter corresponding to the first risk location group.

[0118] Based on the vehicle rescue demand parameters corresponding to each first-risk location group, a dynamic scheduling scheme for vehicle rescue resources is determined.

[0119] Specifically, a dynamic vehicle rescue resource scheduling scheme can be determined based on the group vehicle rescue demand parameters corresponding to each first-risk location group using a dynamic vehicle rescue resource scheduling model. This model can be a reinforcement learning model. Resources such as rescue vehicles, personnel, and equipment are rationally allocated according to the rescue demand intensity of each risk location group. For risk location groups with higher rescue demands, more rescue resources should be allocated preferentially. By comprehensively considering factors such as real-time weather information, real-time traffic information, and real-time vehicle rescue information, the single-point vehicle rescue demand parameters for each risk location and the group vehicle rescue demand parameters for each first-risk location group can be determined, thereby formulating a dynamic scheduling scheme for vehicle rescue resources. This scheme can ensure the rational allocation and efficient utilization of rescue resources, improving the efficiency and quality of rescue operations.

[0120] In some embodiments, a dynamic edge computing scheduling scheme is determined by a cloud platform based on real-time road information of the target area, including:

[0121] For each risk location, the single-point computing power requirement parameter is determined based on real-time meteorological information obtained from relevant meteorological elements, real-time traffic flow information of relevant roads corresponding to the risk location, and real-time vehicle rescue information of relevant risk locations. The single-point computing power requirement parameter is used to measure the computing power demand intensity of the risk location at the current moment. It can be determined by the second parameter determination model based on real-time meteorological information obtained from relevant meteorological elements, real-time traffic flow information of relevant roads corresponding to the risk location, and real-time vehicle rescue information of relevant risk locations. The second parameter determination model can be a convolutional neural network model.

[0122] For each second risk location group, the group computing power requirement parameter corresponding to the second risk location group is determined based on the single-point computing power requirement parameter corresponding to each risk location included in the first risk location group. For example, the average value of the single-point computing power requirement parameter corresponding to each risk location included in the first risk location group can be used as the group computing power requirement parameter corresponding to the second risk location group.

[0123] Based on the group computing power demand parameters corresponding to each second risk location group, a dynamic scheduling scheme for edge computing is determined, wherein the dynamic scheduling scheme for edge computing includes the risk location corresponding to each edge computing station.

[0124] Specifically, an edge computing scheduling scheme can be determined based on the group computing power demand parameters corresponding to each second-risk location group using an edge computing scheduling model, where the edge computing scheduling model can be a reinforcement learning model. By comprehensively considering factors such as real-time weather information, real-time traffic information, and real-time vehicle rescue information, the single-point computing power demand parameters for each risk location and the group computing power demand parameters for each second-risk location group can be determined. Furthermore, a dynamic scheduling scheme for edge computing resources can be formulated based on these parameters to ensure rapid response and sufficient computing support during peak computing power demand periods.

[0125] Step 117: According to the dynamic dispatching plan for vehicle rescue resources, dispatch vehicle rescue resources to multiple vehicle rescue service stations.

[0126] Step 118: Based on the edge computing dynamic scheduling scheme, multiple edge computing stations are scheduled to determine the vehicle rescue needs of the target area according to the real-time road information of the target area.

[0127] In some embodiments, step 118 specifically includes:

[0128] For each edge computing station, the single-point vehicle rescue needs at the corresponding risk location are determined based on the sound and image information of the risk location. The single-point vehicle rescue needs at the risk location include the single-point vehicle rescue needs of each risk location.

[0129] Specifically, the demand judgment model can determine the single-point vehicle rescue demand at the risk location based on the sound and image information of the risk location at multiple consecutive time points. The demand judgment model can be a long short-term memory artificial neural network model.

[0130] Step 119: Receive vehicle rescue requests through multiple edge computing stations.

[0131] Specifically, vehicle rescue requests can be initiated from the user's client.

[0132] Step 120: Vehicle rescue is carried out through multiple edge computing stations based on the vehicle rescue needs and requests in the target area.

[0133] In some embodiments, step 120 specifically includes:

[0134] For each edge computing station, based on the vehicle rescue needs and vehicle rescue requests at the corresponding risk location, the optimal vehicle rescue service station and target vehicle rescue resources are determined from multiple vehicle rescue service stations.

[0135] The optimal vehicle rescue service station will provide vehicle rescue based on the available rescue resources for the target vehicle.

[0136] Specifically, the edge computing station first collects real-time vehicle rescue needs and requests for the corresponding risk location. This information may include accident type, vehicle type, fault severity, and geographical location. Next, the edge computing station uses built-in data processing and analysis algorithms to quickly process this information to assess the urgency of the rescue need and resource requirements. Based on the collected information, the edge computing station evaluates the capabilities, locations, and available resources of multiple vehicle rescue service stations. These service stations may have different rescue specialties, vehicle types, and numbers of rescue personnel, thus requiring matching based on actual needs. Based on the above assessment, the edge computing station uses optimization algorithms (such as shortest path algorithms and resource matching algorithms) to determine the optimal vehicle rescue service station. The optimal service station should be able to respond to rescue requests in the shortest possible time, provide necessary rescue resources, and meet specific rescue needs. The edge computing station analyzes the required types of vehicle rescue resources (such as tow trucks, ambulances, and technicians) based on the urgency of the rescue need and resource requirements. From the optimal vehicle rescue service stations, the edge computing station selects and allocates target vehicle rescue resources. These resources should be able to meet the rescue needs while ensuring the efficiency and safety of the rescue operation. Once the optimal vehicle rescue service station and target vehicle rescue resources are identified, the edge computing station immediately notifies the service station to prepare for rescue. Simultaneously, the edge computing station may also need to maintain communication with the stranded vehicle or relevant personnel, providing necessary rescue guidance and information. Following the instructions from the edge computing station, the optimal vehicle rescue service station dispatches the target vehicle rescue resources to the accident scene. Upon arrival, rescue personnel execute appropriate rescue operations based on the accident situation and rescue needs.

[0137] Figure 4 These are schematic diagrams of modules of a vehicle rescue management system based on cloud platform and edge computing, as shown in some embodiments of this specification. Figure 4 As shown, a vehicle rescue management system based on cloud platform and edge computing can include a resource configuration module, an information acquisition module, a dynamic scheduling module, and a rescue management module.

[0138] The resource configuration module is used to obtain historical vehicle rescue data of the target area, determine the vehicle rescue service station setting scheme and edge computing configuration scheme of the target area based on the historical vehicle rescue data of the target area, set up multiple vehicle rescue service stations in the target area based on the vehicle rescue service station setting scheme, and set up multiple edge computing stations in the target area based on the edge computing configuration scheme of the target area.

[0139] The information acquisition module is used to acquire real-time road information for the target area;

[0140] The dynamic scheduling module is used to determine the dynamic scheduling scheme for vehicle rescue resources and the edge computing dynamic scheduling scheme based on the real-time road information of the target area through the cloud platform, and to schedule vehicle rescue resources to multiple vehicle rescue service stations according to the dynamic scheduling scheme for vehicle rescue resources.

[0141] The rescue management module is used to schedule multiple edge computing stations based on a dynamic edge computing scheduling scheme. Based on real-time road information of the target area, it determines the vehicle rescue needs of the target area, receives vehicle rescue requests through multiple edge computing stations, and performs vehicle rescue based on the vehicle rescue needs and requests of the target area.

[0142] The vehicle rescue management system based on cloud platform and edge computing can be used to execute vehicle rescue management methods based on cloud platform and edge computing, which will not be elaborated here.

[0143] Finally, it should be understood that the embodiments described in this specification are merely illustrative of the principles of the embodiments described herein. Other variations may also fall within the scope of this specification. Therefore, alternative configurations of the embodiments described herein are intended to be illustrative rather than limiting, and should be considered consistent with the teachings of this specification. Accordingly, the embodiments described herein are not limited to those explicitly introduced and described herein.

Claims

1. A vehicle rescue management method based on cloud platform and edge computing, characterized in that, include: Obtain historical vehicle rescue data for the target area; Based on historical vehicle rescue data for the target area, determine the vehicle rescue service station setup plan and edge computing configuration plan for the target area; According to the vehicle rescue service station setup plan for the target area, multiple vehicle rescue service stations will be set up in the target area; Based on the edge computing configuration scheme of the target area, multiple edge computing stations are set up in the target area; Obtain real-time road information for the target area; Based on real-time road information in the target area, the cloud platform determines a dynamic dispatching scheme for vehicle rescue resources and a dynamic dispatching scheme for edge computing. According to the dynamic dispatch plan for vehicle rescue resources, vehicle rescue resources are dispatched to multiple vehicle rescue service stations; Based on the dynamic scheduling scheme of edge computing, multiple edge computing stations are scheduled to determine the vehicle rescue needs of the target area according to the real-time road information of the target area. Receive vehicle rescue requests through multiple edge computing stations; Vehicle rescue is carried out through multiple edge computing stations based on vehicle rescue needs and requests in the target area; Based on historical vehicle rescue data for the target area, determine the vehicle rescue service station setup plan for the target area, including: Based on historical vehicle rescue data for the target area, multiple risk locations within the target area and the risk coefficient for each risk location are determined. For any two risk locations, determine the driving distance between the two risk locations based on the electronic map of the target area; For any two risk locations, determine the accident correlation coefficient between the two risk locations based on historical vehicle rescue data of the target area; Based on the driving distance and accident correlation coefficient between any two risk locations, multiple risk locations in the target area are grouped to determine multiple first risk location groups; For each first risk location group, the first center location and rescue resource requirements corresponding to the first risk location group are determined based on the risk coefficient of each risk location included in the first risk location group and the accident correlation coefficient between any two risk locations included in the first risk location group. Based on the first center location and rescue resource needs corresponding to each first risk location group, determine the vehicle rescue service station setup plan for the target area; Based on historical vehicle rescue data for the target area, determine the edge computing configuration scheme for the target area, including: For any two risk locations, determine the straight-line distance between the two risk locations based on the electronic map of the target area; Based on the K-means algorithm, multiple risk locations in the target area are clustered according to the straight-line distance between any two risk locations and the maximum threshold of the straight-line distance, and multiple second risk location clusters are determined. For each second risk location cluster, the K-means algorithm is used to cluster multiple risk locations included in the second risk location cluster based on the accident correlation coefficient between any two risk locations and the maximum threshold of the accident correlation coefficient, to determine the second risk location group included in the second risk location cluster. For each second risk location group, the second center location and edge computing requirements corresponding to the second risk location group are determined based on the risk coefficient of each risk location included in the second risk location group and the accident correlation coefficient between any two risk locations included in the second risk location group. Based on the second center location and edge computing requirements corresponding to each second risk location group, determine the edge computing configuration scheme for the target area.

2. The vehicle rescue management method based on cloud platform and edge computing according to claim 1, characterized in that, Based on the driving distance and accident correlation coefficient between any two risk locations, multiple risk locations in the target area are grouped to determine multiple first-risk location groups, including: Based on the K-means algorithm, multiple risk locations in the target area are clustered according to the driving distance between any two risk locations and the maximum driving distance threshold, thus determining multiple first risk location clusters; For each first risk location cluster, the K-means algorithm is used to cluster multiple risk locations included in the first risk location cluster based on the accident correlation coefficient between any two risk locations and the maximum threshold of the accident correlation coefficient, thereby determining the first risk location group included in the first risk location cluster.

3. The vehicle rescue management method based on cloud platform and edge computing according to claim 1, characterized in that, Obtain real-time road information for the target area, including: Obtain meteorological data corresponding to historical vehicle rescue data in the target area; Based on the historical vehicle rescue data of the target area and the corresponding meteorological data, relevant meteorological elements are determined; Obtain traffic flow data corresponding to historical vehicle rescue data in the target area; For each risk location, the relevant roads corresponding to the risk location are determined based on the historical vehicle rescue data of the target area and the traffic flow data corresponding to the historical vehicle rescue data of the target area. For each risk location, the relevant risk locations are determined based on the historical vehicle rescue data of the target area; For each risk location, real-time road information for a single point is obtained based on relevant meteorological elements, relevant roads corresponding to the risk location, and relevant risk locations corresponding to the risk location. The real-time road information of the target area includes real-time road information for a single point corresponding to each risk location. The real-time road information for a single point corresponding to a risk location includes real-time meteorological information obtained based on relevant meteorological elements, real-time traffic flow information of relevant roads corresponding to the risk location, real-time vehicle rescue information of relevant risk locations corresponding to the risk location, and real-time status information of the risk location.

4. The vehicle rescue management method based on cloud platform and edge computing according to claim 3, characterized in that, Based on real-time road information of the target area, a dynamic dispatch plan for vehicle rescue resources is determined through a cloud platform, including: For each risk location, the single-point vehicle rescue demand parameters corresponding to the risk location are determined based on real-time meteorological information obtained from relevant meteorological elements, real-time traffic flow information of relevant roads corresponding to the risk location, and real-time vehicle rescue information of relevant risk locations. For each first risk location group, the group vehicle rescue demand parameters corresponding to the first risk location group are determined based on the single-point vehicle rescue demand parameters corresponding to each risk location included in the first risk location group. Based on the vehicle rescue demand parameters corresponding to each first-risk location group, a dynamic scheduling scheme for vehicle rescue resources is determined.

5. The vehicle rescue management method based on cloud platform and edge computing according to claim 3, characterized in that, Based on real-time road information of the target area, the cloud platform determines a dynamic edge computing scheduling scheme, including: For each risk location, the single-point computing power requirement parameters are determined based on real-time meteorological information obtained from relevant meteorological elements, real-time traffic flow information of the relevant roads corresponding to the risk location, and real-time vehicle rescue information of the relevant risk locations. For each second risk location group, the group computing power requirement parameter corresponding to the second risk location group is determined based on the single-point computing power requirement parameter corresponding to each risk location included in the second risk location group. Based on the group computing power requirement parameters corresponding to each second risk location group, a dynamic scheduling scheme for edge computing is determined, wherein the dynamic scheduling scheme for edge computing includes the risk location corresponding to each edge computing station.

6. The vehicle rescue management method based on cloud platform and edge computing according to claim 5, characterized in that, The real-time status information of the risk location includes at least the audio and image information of the risk location; Based on a dynamic scheduling scheme using edge computing, multiple edge computing stations are dispatched to determine the vehicle rescue needs in the target area according to real-time road information, including: For each edge computing station, the single-point vehicle rescue needs at the corresponding risk location are determined based on the sound and image information of the risk location. The single-point vehicle rescue needs at the risk location include the single-point vehicle rescue needs for each risk location.

7. The vehicle rescue management method based on cloud platform and edge computing according to claim 5, characterized in that, Vehicle rescue is conducted through multiple edge computing stations based on vehicle rescue needs and requests in the target area, including: For each edge computing station, based on the vehicle rescue needs and vehicle rescue requests at the corresponding risk location, the optimal vehicle rescue service station and target vehicle rescue resources are determined from multiple vehicle rescue service stations. The optimal vehicle rescue service station will provide vehicle rescue based on the available rescue resources for the target vehicle.

8. A vehicle rescue management system based on cloud platform and edge computing, characterized in that, The vehicle rescue management method based on cloud platform and edge computing according to any one of claims 1-7 includes: The resource configuration module is used to obtain historical vehicle rescue data of the target area, determine the vehicle rescue service station setting scheme and edge computing configuration scheme of the target area based on the historical vehicle rescue data of the target area, set up multiple vehicle rescue service stations in the target area based on the vehicle rescue service station setting scheme, and set up multiple edge computing stations in the target area based on the edge computing configuration scheme of the target area. The information acquisition module is used to acquire real-time road information for the target area; The dynamic scheduling module is used to determine the dynamic scheduling scheme for vehicle rescue resources and the edge computing dynamic scheduling scheme based on the real-time road information of the target area through the cloud platform, and to schedule vehicle rescue resources to multiple vehicle rescue service stations according to the dynamic scheduling scheme for vehicle rescue resources. The rescue management module is used to schedule multiple edge computing stations based on a dynamic edge computing scheduling scheme. Based on real-time road information of the target area, it determines the vehicle rescue needs of the target area, receives vehicle rescue requests through multiple edge computing stations, and performs vehicle rescue based on the vehicle rescue needs and requests of the target area.

Images